Doc. no. D????
Date: 2023-06-09
Project: Programming Language C++
Reply to: Jonathan Wakely <lwgchair@gmail.com>

C++ Standard Library Closed Issues List (Revision R124)

Revised 2023-06-09 at 14:25:43 UTC

Reference ISO/IEC IS 14882:2020(E)

Also see:

This document contains only library issues which have been closed by the Library Working Group as duplicates or not defects. That is, issues which have a status of Dup or NAD. See the Library Defect Reports and Accepted Issues for issues considered defects. See the Library Active Issues List for active issues and more information. The introductory material in that document also applies to this document.

Revision History

Closed Issues


2(i). Auto_ptr conversions effects incorrect

Section: 99 [auto.ptr.conv] Status: NAD Submitter: Nathan Myers Opened: 1997-12-04 Last modified: 2016-08-09

Priority: Not Prioritized

View all other issues in [auto.ptr.conv].

View all issues with NAD status.

Discussion:

Paragraph 1 in "Effects", says "Calls p->release()" where it clearly must be "Calls p.release()". (As it is, it seems to require using auto_ptr<>::operator-> to refer to X::release, assuming that exists.)

Proposed resolution:

Change 21.3.5.4 [meta.unary.prop] paragraph 1 Effects from "Calls p->release()" to "Calls p.release()".

Rationale:

Not a defect: the proposed change is already found in the standard. [Originally classified as a defect, later reclassified.]


4(i). basic_string size_type and difference_type should be implementation defined

Section: 23.4.3 [basic.string] Status: NAD Submitter: Beman Dawes Opened: 1997-11-16 Last modified: 2016-08-09

Priority: Not Prioritized

View other active issues in [basic.string].

View all other issues in [basic.string].

View all issues with NAD status.

Discussion:

In Morristown we changed the size_type and difference_type typedefs for all the other containers to implementation defined with a reference to 24.2 [container.requirements]. This should probably also have been done for strings.

Rationale:

Not a defect. [Originally classified as a defect, later reclassified.] basic_string, unlike the other standard library template containers, is severely constrained by its use of char_traits. Those types are dictated by the traits class, and are far from implementation defined.


6(i). File position not an offset unimplementable

Section: 31.5.3 [fpos] Status: NAD Submitter: Matt Austern Opened: 1997-12-15 Last modified: 2016-08-09

Priority: Not Prioritized

View all other issues in [fpos].

View all issues with NAD status.

Discussion:

Table 88, in I/O, is too strict; it's unimplementable on systems where a file position isn't just an offset. It also never says just what fpos<> is really supposed to be. [Here's my summary, which Jerry agrees is more or less accurate. "I think I now know what the class really is, at this point: it's a magic cookie that encapsulates an mbstate_t and a file position (possibly represented as an fpos_t), it has syntactic support for pointer-like arithmetic, and implementors are required to have real, not just syntactic, support for arithmetic." This isn't standardese, of course.]

Rationale:

Not a defect. The LWG believes that the Standard is already clear, and that the above summary is what the Standard in effect says.


10(i). Codecvt<>::do unclear

Section: 30.4.2.6 [locale.codecvt.byname] Status: Dup Submitter: Matt Austern Opened: 1998-01-14 Last modified: 2016-08-09

Priority: Not Prioritized

View all other issues in [locale.codecvt.byname].

View all issues with Dup status.

Duplicate of: 19

Discussion:

Section 22.2.1.5.2 says that codecvt<>::do_in and do_out should return the value noconv if "no conversion was needed". However, I don't see anything anywhere that defines what it means for a conversion to be needed or not needed. I can think of several circumstances where one might plausibly think that a conversion is not "needed", but I don't know which one is intended here.

Rationale:


12(i). Way objects hold allocators unclear

Section: 16.4.4.6 [allocator.requirements] Status: NAD Submitter: Angelika Langer Opened: 1998-02-23 Last modified: 2016-08-09

Priority: Not Prioritized

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with NAD status.

Discussion:

I couldn't find a statement in the standard saying whether the allocator object held by a container is held as a copy of the constructor argument or whether a pointer of reference is maintained internal. There is an according statement for compare objects and how they are maintained by the associative containers, but I couldn't find anything regarding allocators.

Did I overlook it? Is it an open issue or known defect? Or is it deliberately left unspecified?

Rationale:

Not a defect. The LWG believes that the Standard is already clear.  See 24.2 [container.requirements], paragraph 8.


43(i). Locale table correction

Section: 30.4.2.6 [locale.codecvt.byname] Status: Dup Submitter: Brendan Kehoe Opened: 1998-06-01 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.codecvt.byname].

View all issues with Dup status.

Duplicate of: 33

Discussion:

Rationale:


45(i). Stringstreams read/write pointers initial position unclear

Section: 31.8.4 [ostringstream] Status: NAD Submitter: Matthias Mueller Opened: 1998-05-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [ostringstream].

View all issues with NAD status.

Discussion:

In a comp.lang.c++.moderated Matthias Mueller wrote:

"We are not sure how to interpret the CD2 (see 31.3 [iostream.forward], 31.8.4.2 [ostringstream.cons], 31.8.2.2 [stringbuf.cons]) with respect to the question as to what the correct initial positions of the write and  read pointers of a stringstream should be."

"Is it the same to output two strings or to initialize the stringstream with the first and to output the second?"

[PJ Plauger, Bjarne Stroustrup, Randy Smithey, Sean Corfield, and Jerry Schwarz have all offered opinions; see reflector messages lib-6518, 6519, 6520, 6521, 6523, 6524.]

Rationale:

The LWG believes the Standard is correct as written. The behavior of stringstreams is consistent with fstreams, and there is a constructor which can be used to obtain the desired effect. This behavior is known to be different from strstreams.


58(i). Extracting a char from a wide-oriented stream

Section: 31.7.5.3.3 [istream.extractors] Status: NAD Submitter: Matt Austern Opened: 1998-07-01 Last modified: 2017-04-22

Priority: Not Prioritized

View all other issues in [istream.extractors].

View all issues with NAD status.

Discussion:

27.6.1.2.3 has member functions for extraction of signed char and unsigned char, both singly and as strings. However, it doesn't say what it means to extract a char from a basic_streambuf<charT, Traits>.

basic_streambuf, after all, has no members to extract a char, so basic_istream must somehow convert from charT to signed char or unsigned char. The standard doesn't say how it is to perform that conversion.

Rationale:

The Standard is correct as written. There is no such extractor and this is the intent of the LWG.


65(i). Underspecification of strstreambuf::seekoff

Section: D.15.2.4 [depr.strstreambuf.virtuals] Status: NAD Submitter: Matt Austern Opened: 1998-08-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [depr.strstreambuf.virtuals].

View all issues with NAD status.

Discussion:

The standard says how this member function affects the current stream position. (gptr or pptr) However, it does not say how this member function affects the beginning and end of the get/put area.

This is an issue when seekoff is used to position the get pointer beyond the end of the current read area. (Which is legal. This is implicit in the definition of seekhigh in D.7.1, paragraph 4.)

Rationale:

The LWG agrees that seekoff() is underspecified, but does not wish to invest effort in this deprecated feature.


67(i). Setw useless for strings

Section: 23.4.4.4 [string.io] Status: Dup Submitter: Steve Clamage Opened: 1998-07-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [string.io].

View all issues with Dup status.

Duplicate of: 25

Discussion:

In a comp.std.c++ posting Michel Michaud wrote: What should be output by:

   string text("Hello");
   cout << '[' << setw(10) << right << text << ']';

Shouldn't it be:

   [     Hello]

Another person replied: Actually, according to the FDIS, the width of the field should be the minimum of width and the length of the string, so the output shouldn't have any padding. I think that this is a typo, however, and that what is wanted is the maximum of the two. (As written, setw is useless for strings. If that had been the intent, one wouldn't expect them to have mentioned using its value.)

It's worth pointing out that this is a recent correction anyway; IIRC, earlier versions of the draft forgot to mention formatting parameters whatsoever.

Rationale:


72(i). Do_convert phantom member function

Section: 30.4.2.5 [locale.codecvt] Status: Dup Submitter: Nathan Myers Opened: 1998-08-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.codecvt].

View all issues with Dup status.

Duplicate of: 24

Discussion:

In 30.4.2.5 [locale.codecvt] par 3, and in 30.4.2.5.3 [locale.codecvt.virtuals] par 8, a nonexistent member function "do_convert" is mentioned. This member was replaced with "do_in" and "do_out", the proper referents in the contexts above.

Rationale:


73(i). is_open should be const

Section: 31.10 [file.streams] Status: NAD Submitter: Matt Austern Opened: 1998-08-27 Last modified: 2017-06-15

Priority: Not Prioritized

View all other issues in [file.streams].

View all issues with NAD status.

Discussion:

Classes basic_ifstream, basic_ofstream, and basic_fstream all have a member function is_open. It should be a const member function, since it does nothing but call one of basic_filebuf's const member functions.

Rationale:

Not a defect. This is a deliberate feature; const streams would be meaningless.


77(i). Valarray operator[] const returning value

Section: 28.6.2.4 [valarray.access] Status: Dup Submitter: Levente Farkas Opened: 1998-09-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [valarray.access].

View all issues with Dup status.

Duplicate of: 389

Discussion:

valarray:

    T operator[] (size_t) const;

why not

    const T& operator[] (size_t) const;

as in vector ???

One can't copy even from a const valarray eg:

    memcpy(ptr, &v[0], v.size() * sizeof(double));

[I] find this bug in valarray is very difficult.

Rationale:

The LWG believes that the interface was deliberately designed that way. That is what valarray was designed to do; that's where the "value array" name comes from. LWG members further comment that "we don't want valarray to be a full STL container." 28.6.2.4 [valarray.access] specifies properties that indicate "an absence of aliasing" for non-constant arrays; this allows optimizations, including special hardware optimizations, that are not otherwise possible.


81(i). Wrong declaration of slice operations

Section: 28.6.5 [template.slice.array], 28.6.7 [template.gslice.array], 28.6.8 [template.mask.array], 28.6.9 [template.indirect.array] Status: NAD Submitter: Nico Josuttis Opened: 1998-09-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [template.slice.array].

View all issues with NAD status.

Discussion:

Isn't the definition of copy constructor and assignment operators wrong?        Instead of

      slice_array(const slice_array&); 
      slice_array& operator=(const slice_array&);

IMHO they have to be

      slice_array(const slice_array<T>&); 
      slice_array& operator=(const slice_array<T>&);

Same for gslice_array.

Rationale:

Not a defect. The Standard is correct as written.


82(i). Missing constant for set elements

Section: 24.2.7 [associative.reqmts] Status: NAD Submitter: Nico Josuttis Opened: 1998-09-29 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with NAD status.

Discussion:

Paragraph 5 specifies:

For set and multiset the value type is the same as the key type. For map and multimap it is equal to pair<const Key, T>.

Strictly speaking, this is not correct because for set and multiset the value type is the same as the constant key type.

Rationale:

Not a defect. The Standard is correct as written; it uses a different mechanism (const &) for set and multiset. See issue 103 for a related issue.


84(i). Ambiguity with string::insert()

Section: 23.4.3.6 [string.access] Status: NAD Submitter: Nico Josuttis Opened: 1998-09-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [string.access].

View all issues with NAD status.

Discussion:

If I try

    s.insert(0,1,' ');

  I get an nasty ambiguity. It might be

    s.insert((size_type)0,(size_type)1,(charT)' ');

which inserts 1 space character at position 0, or

    s.insert((char*)0,(size_type)1,(charT)' ')

which inserts 1 space character at iterator/address 0 (bingo!), or

    s.insert((char*)0, (InputIterator)1, (InputIterator)' ')

which normally inserts characters from iterator 1 to iterator ' '. But according to 23.1.1.9 (the "do the right thing" fix) it is equivalent to the second. However, it is still ambiguous, because of course I mean the first!

Rationale:

Not a defect. The LWG believes this is a "genetic misfortune" inherent in the design of string and thus not a defect in the Standard as such .


85(i). String char types

Section: 23 [strings] Status: NAD Submitter: Nico Josuttis Opened: 1998-09-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [strings].

View all issues with NAD status.

Discussion:

The standard seems not to require that charT is equivalent to traits::char_type. So, what happens if charT is not equivalent to traits::char_type?

Rationale:

There is already wording in 23.2 [char.traits] paragraph 3 that requires them to be the same.


87(i). Error in description of string::compare()

Section: 23.4.3.7.8 [string.swap] Status: Dup Submitter: Nico Josuttis Opened: 1998-09-29 Last modified: 2016-11-12

Priority: Not Prioritized

View all other issues in [string.swap].

View all issues with Dup status.

Duplicate of: 5

Discussion:

The following compare() description is obviously a bug:

int compare(size_type pos, size_type n1, 
            charT *s, size_type n2 = npos) const;

because without passing n2 it should compare up to the end of the string instead of comparing npos characters (which throws an exception)

Rationale:


88(i). Inconsistency between string::insert() and string::append()

Section: 23.4.3.7.4 [string.insert], 23.4.3.7.2 [string.append] Status: NAD Submitter: Nico Josuttis Opened: 1998-09-29 Last modified: 2016-11-12

Priority: Not Prioritized

View all other issues in [string.insert].

View all issues with NAD status.

Discussion:

Why does

  template<class InputIterator> 
       basic_string& append(InputIterator first, InputIterator last);

return a string, while

  template<class InputIterator> 
       void insert(iterator p, InputIterator first, InputIterator last);

returns nothing ?

Rationale:

The LWG believes this stylistic inconsistency is not sufficiently serious to constitute a defect.


89(i). Missing throw specification for string::insert() and string::replace()

Section: 23.4.3.7.4 [string.insert], 23.4.3.7.6 [string.replace] Status: Dup Submitter: Nico Josuttis Opened: 1998-09-29 Last modified: 2016-11-12

Priority: Not Prioritized

View all other issues in [string.insert].

View all issues with Dup status.

Duplicate of: 83

Discussion:

All insert() and replace() members for strings with an iterator as first argument lack a throw specification. The throw specification should probably be: length_error if size exceeds maximum.

Rationale:

Considered a duplicate because it will be solved by the resolution of issue 83.


93(i). Incomplete Valarray Subset Definitions

Section: 28.6 [numarray] Status: NAD Submitter: Nico Josuttis Opened: 1998-09-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [numarray].

View all issues with NAD status.

Discussion:

You can easily create subsets, but you can't easily combine them with other subsets. Unfortunately, you almost always needs an explicit type conversion to valarray. This is because the standard does not specify that valarray subsets provide the same operations as valarrays.

For example, to multiply two subsets and assign the result to a third subset, you can't write the following:

va[slice(0,4,3)] = va[slice(1,4,3)] * va[slice(2,4,3)];

Instead, you have to code as follows:

va[slice(0,4,3)] = static_cast<valarray<double> >(va[slice(1,4,3)]) * 
                   static_cast<valarray<double> >(va[slice(2,4,3)]);

This is tedious and error-prone. Even worse, it costs performance because each cast creates a temporary objects, which could be avoided without the cast.

Proposed resolution:

Extend all valarray subset types so that they offer all valarray operations.

Rationale:

This is not a defect in the Standard; it is a request for an extension.


94(i). May library implementors add template parameters to Standard Library classes?

Section: 16.4.6 [conforming] Status: NAD Submitter: Matt Austern Opened: 1998-01-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [conforming].

View all issues with NAD status.

Discussion:

Is it a permitted extension for library implementors to add template parameters to standard library classes, provided that those extra parameters have defaults? For example, instead of defining template <class T, class Alloc = allocator<T> > class vector; defining it as template <class T, class Alloc = allocator<T>, int N = 1> class vector;

The standard may well already allow this (I can't think of any way that this extension could break a conforming program, considering that users are not permitted to forward-declare standard library components), but it ought to be explicitly permitted or forbidden.

comment from Steve Cleary via comp.std.c++:

I disagree [with the proposed resolution] for the following reason: consider user library code with template template parameters. For example, a user library object may be templated on the type of underlying sequence storage to use (deque/list/vector), since these classes all take the same number and type of template parameters; this would allow the user to determine the performance tradeoffs of the user library object. A similar example is a user library object templated on the type of underlying set storage (set/multiset) or map storage (map/multimap), which would allow users to change (within reason) the semantic meanings of operations on that object.

I think that additional template parameters should be forbidden in the Standard classes. Library writers don't lose any expressive power, and can still offer extensions because additional template parameters may be provided by a non-Standard implementation class:

 
   template <class T, class Allocator = allocator<T>, int N = 1>
   class __vector
   { ... };
   template <class T, class Allocator = allocator<T> >
   class vector: public __vector<T, Allocator>
   { ... };

Proposed resolution:

Add a new subclause [presumably 17.4.4.9] following 16.4.6.13 [res.on.exception.handling]:

17.4.4.9 Template Parameters

A specialization of a template class described in the C++ Standard Library behaves the same as if the implementation declares no additional template parameters.

Footnote: Additional template parameters with default values are thus permitted.

Add "template parameters" to the list of subclauses at the end of 16.4.6 [conforming] paragraph 1.

[Kona: The LWG agreed the standard needs clarification. After discussion with John Spicer, it seems added template parameters can be detected by a program using template-template parameters. A straw vote - "should implementors be allowed to add template parameters?" found no consensus ; 5 - yes, 7 - no.]

Rationale:

There is no ambiguity; the standard is clear as written. Library implementors are not permitted to add template parameters to standard library classes. This does not fall under the "as if" rule, so it would be permitted only if the standard gave explicit license for implementors to do this. This would require a change in the standard.

The LWG decided against making this change, because it would break user code involving template template parameters or specializations of standard library class templates.


95(i). Members added by the implementation

Section: 16.4.6.5 [member.functions] Status: NAD Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [member.functions].

View all issues with NAD status.

Discussion:

In 17.3.4.4/2 vs 17.3.4.7/0 there is a hole; an implementation could add virtual members a base class and break user derived classes.

Example:

// implementation code:
struct _Base { // _Base is in the implementer namespace
        virtual void foo ();
};
class vector : _Base // deriving from a class is allowed
{ ... };

// user code:
class vector_checking : public vector 
{
        void foo (); // don't want to override _Base::foo () as the 
                     // user doesn't know about _Base::foo ()
};

Proposed resolution:

Clarify the wording to make the example illegal.

Rationale:

This is not a defect in the Standard.  The example is already illegal.  See 16.4.6.5 [member.functions] paragraph 2.


96(i). Vector<bool> is not a container

Section: 24.3.11 [vector] Status: NAD Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [vector].

View all issues with NAD status.

Discussion:

vector<bool> is not a container as its reference and pointer types are not references and pointers.

Also it forces everyone to have a space optimization instead of a speed one.

See also: 99-0008 == N1185 Vector<bool> is Nonconforming, Forces Optimization Choice.

[In Santa Cruz the LWG felt that this was Not A Defect.]

[In Dublin many present felt that failure to meet Container requirements was a defect. There was disagreement as to whether or not the optimization requirements constituted a defect.]

[The LWG looked at the following resolutions in some detail:
     * Not A Defect.
     * Add a note explaining that vector<bool> does not meet Container requirements.
     * Remove vector<bool>.
     * Add a new category of container requirements which vector<bool> would meet.
     * Rename vector<bool>.

No alternative had strong, wide-spread, support and every alternative had at least one "over my dead body" response.

There was also mention of a transition scheme something like (1) add vector_bool and deprecate vector<bool> in the next standard. (2) Remove vector<bool> in the following standard.]

[Modifying container requirements to permit returning proxies (thus allowing container requirements conforming vector<bool>) was also discussed.]

[It was also noted that there is a partial but ugly workaround in that vector<bool> may be further specialized with a customer allocator.]

[Kona: Herb Sutter presented his paper J16/99-0035==WG21/N1211, vector<bool>: More Problems, Better Solutions. Much discussion of a two step approach: a) deprecate, b) provide replacement under a new name. LWG straw vote on that: 1-favor, 11-could live with, 2-over my dead body. This resolution was mentioned in the LWG report to the full committee, where several additional committee members indicated over-my-dead-body positions.]

Discussed at Lillehammer. General agreement that we should deprecate vector<bool> and introduce this functionality under a different name, e.g. bit_vector. This might make it possible to remove the vector<bool> specialization in the standard that comes after C++0x. There was also a suggestion that in C++0x we could additional say that it's implementation defined whether vector<bool> refers to the specialization or to the primary template, but there wasn't general agreement that this was a good idea.

We need a paper for the new bit_vector class.

[ Batavia: ]

The LWG feels we need something closer to SGI's bitvector to ease migration from vector<bool>. Although some of the funcitonality from N2050 could well be used in such a template. The concern is easing the API migration for those users who want to continue using a bit-packed container. Alan and Beman to work.

[ Post Summit Alisdair adds: ]

vector<bool> is now a conforming container under the revised terms of C++0x, which supports containers of proxies.

Recommend NAD.

Two issues remain:

i/ premature optimization in the specification. There is still some sentiment that deprecation is the correct way to go, although it is still not clear what it would mean to deprecate a single specialization of a template.

Recommend: Create a new issue for the discussion, leave as Open.

ii/ Request for a new bitvector class to guarantee the optimization, perhaps with a better tuned interface.

This is a clear extension request that may be handled via a future TR.

[ Batavia (2009-05): ]

We note that most of this issue has become moot over time, and agree with Alisdair's recommendations. Move to NAD Future for reconsideration of part (ii).

[ 2009-07-29 Alisdair reopens: ]

This infamous issue was closed as NAD Future when concepts introduced support for proxy iterators, so the only remaining requirement was to provide a better type to support bitsets of dynamic length. I fear we must re-open this issue until the post-concept form of iterators is available, and hopefully will support the necessary proxy functionality to allow us to close this issue as NAD.

I recommend we spawn a separate issue (1184) requesting a dynamic length bitset and pre-emptively file it as NAD Future. It is difficult to resolve #96 when it effectively contains two separate sub-issues.

[ 2009-10 Santa Cruz: ]

Mark as NAD, and give rationale.

Proposed resolution:

We now have: N2050 and N2160.

Rationale:

We want to support proxy iterators but that is going to be separate work. Don't want to see this issue come back in these kinds of terms. We're interested in a separate container, and proxy iterators, but both of those are separate issues.

We've looked at a lot of ways to fix this that would be close to this, but those things would break existing code. Attempts to fix this directly have not been tractable, and removing it has not been tractable. Therefore we are closing.


97(i). Insert inconsistent definition

Section: 24 [containers] Status: NAD Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [containers].

View all other issues in [containers].

View all issues with NAD status.

Discussion:

insert(iterator, const value_type&) is defined both on sequences and on set, with unrelated semantics: insert here (in sequences), and insert with hint (in associative containers). They should have different names (B.S. says: do not abuse overloading).

Rationale:

This is not a defect in the Standard. It is a genetic misfortune of the design, for better or for worse.


99(i). Reverse_iterator comparisons completely wrong

Section: 25.5.1.8 [reverse.iter.cmp] Status: NAD Submitter: AFNOR Opened: 1998-10-07 Last modified: 2021-06-06

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

The <, >, <=, >= comparison operator are wrong: they return the opposite of what they should.

Note: same problem in CD2, these were not even defined in CD1. SGI STL code is correct; this problem is known since the Morristown meeting but there it was too late

Rationale:

This is not a defect in the Standard. A careful reading shows the Standard is correct as written. A review of several implementations show that they implement exactly what the Standard says.


100(i). Insert iterators/ostream_iterators overconstrained

Section: 25.5.2 [insert.iterators], 25.6.5 [ostreambuf.iterator] Status: NAD Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [insert.iterators].

View all issues with NAD status.

Discussion:

Overspecified For an insert iterator it, the expression *it is required to return a reference to it. This is a simple possible implementation, but as the SGI STL documentation says, not the only one, and the user should not assume that this is the case.

Rationale:

The LWG believes this causes no harm and is not a defect in the standard. The only example anyone could come up with caused some incorrect code to work, rather than the other way around.


101(i). No way to free storage for vector and deque

Section: 24.3.11 [vector], 24.3.7 [array] Status: NAD Editorial Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [vector].

View all issues with NAD Editorial status.

Discussion:

Reserve can not free storage, unlike string::reserve

[ 2010-02-13 Alisdair adds: ]

This issue has been revisited and addressed (755, 850). This issues should be reclassified to NAD Editorial to reflect this action.

Rationale:

This is not a defect in the Standard. The LWG has considered this issue in the past and sees no need to change the Standard. Deque has no reserve() member function. For vector, shrink-to-fit can be expressed in a single line of code (where v is vector<T>):

vector<T>(v).swap(v);  // shrink-to-fit v


102(i). Bug in insert range in associative containers

Section: 24.2.7 [associative.reqmts] Status: Dup Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with Dup status.

Duplicate of: 264

Discussion:

Table 69 of Containers say that a.insert(i,j) is linear if [i, j) is ordered. It seems impossible to implement, as it means that if [i, j) = [x], insert in an associative container is O(1)!

Proposed resolution:

N+log (size()) if [i,j) is sorted according to value_comp()

Rationale:

Subsumed by issue 264.


104(i). Description of basic_string::operator[] is unclear

Section: 23.4.3.5 [string.capacity] Status: NAD Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [string.capacity].

View all issues with NAD status.

Discussion:

It is not clear that undefined behavior applies when pos == size () for the non const version.

Proposed resolution:

Rewrite as: Otherwise, if pos > size () or pos == size () and the non-const version is used, then the behavior is undefined.

Rationale:

The Standard is correct. The proposed resolution already appears in the Standard.


105(i). fstream ctors argument types desired

Section: 31.10 [file.streams] Status: Dup Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [file.streams].

View all issues with Dup status.

Duplicate of: 454

Discussion:

fstream ctors take a const char* instead of string.
fstream ctors can't take wchar_t

An extension to add a const wchar_t* to fstream would make the implementation non conforming.

Rationale:

This is not a defect in the Standard. It might be an interesting extension for the next Standard.


107(i). Valarray constructor is strange

Section: 28.6.2 [template.valarray] Status: NAD Submitter: AFNOR Opened: 1998-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [template.valarray].

View all issues with NAD status.

Discussion:

The order of the arguments is (elem, size) instead of the normal (size, elem) in the rest of the library. Since elem often has an integral or floating point type, both types are convertible to each other and reversing them leads to a well formed program.

Proposed resolution:

Inverting the arguments could silently break programs. Introduce the two signatures (const T&, size_t) and (size_t, const T&), but make the one we do not want private so errors result in a diagnosed access violation. This technique can also be applied to STL containers.

Rationale:

The LWG believes that while the order of arguments is unfortunate, it does not constitute a defect in the standard. The LWG believes that the proposed solution will not work for valarray<size_t> and perhaps other cases.


111(i). istreambuf_iterator::equal overspecified, inefficient

Section: 25.6.4.4 [istreambuf.iterator.ops] Status: NAD Submitter: Nathan Myers Opened: 1998-10-15 Last modified: 2017-11-29

Priority: Not Prioritized

View all other issues in [istreambuf.iterator.ops].

View all issues with NAD status.

Discussion:

The member istreambuf_iterator<>::equal is specified to be unnecessarily inefficient. While this does not affect the efficiency of conforming implementations of iostreams, because they can "reach into" the iterators and bypass this function, it does affect users who use istreambuf_iterators.

The inefficiency results from a too-scrupulous definition, which requires a "true" result if neither iterator is at eof. In practice these iterators can only usefully be compared with the "eof" value, so the extra test implied provides no benefit, but slows down users' code.

The solution is to weaken the requirement on the function to return true only if both iterators are at eof.

[ Summit: ]

Reopened by Alisdair.

[ Post Summit Daniel adds: ]

Recommend NAD. The proposed wording would violate the axioms of concept requirement EqualityComparable axioms as part of concept InputIterator and more specifically it would violate the explicit wording of 25.3.5.3 [input.iterators]/7:

If two iterators a and b of the same type are equal, then either a and b are both dereferenceable or else neither is dereferenceable.

[ 2009-07 Frankfurt ]

Agree NAD.

Proposed resolution:

Replace [istreambuf.iterator::equal], paragraph 1,

-1- Returns: true if and only if both iterators are at end-of-stream, or neither is at end-of-stream, regardless of what streambuf object they use.

with

-1- Returns: true if and only if both iterators are at end-of-stream, regardless of what streambuf object they use.

Rationale:

It is not clear that this is a genuine defect. Additionally, the LWG was reluctant to make a change that would result in operator== not being a equivalence relation. One consequence of this change is that an algorithm that's passed the range [i, i) would no longer treat it as an empty range.


113(i). Missing/extra iostream sync semantics

Section: 31.7.5.2 [istream], 31.7.5.4 [istream.unformatted] Status: NAD Submitter: Steve Clamage Opened: 1998-10-13 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [istream].

View all issues with NAD status.

Discussion:

In 27.6.1.1, class basic_istream has a member function sync, described in 27.6.1.3, paragraph 36.

Following the chain of definitions, I find that the various sync functions have defined semantics for output streams, but no semantics for input streams. On the other hand, basic_ostream has no sync function.

The sync function should at minimum be added to basic_ostream, for internal consistency.

A larger question is whether sync should have assigned semantics for input streams.

Classic iostreams said streambuf::sync flushes pending output and attempts to return unread input characters to the source. It is a protected member function. The filebuf version (which is public) has that behavior (it backs up the read pointer). Class strstreambuf does not override streambuf::sync, and so sync can't be called on a strstream.

If we can add corresponding semantics to the various sync functions, we should. If not, we should remove sync from basic_istream.

Rationale:

A sync function is not needed in basic_ostream because the flush function provides the desired functionality.

As for the other points, the LWG finds the standard correct as written.


116(i). bitset cannot be constructed with a const char*

Section: 22.9.2 [template.bitset] Status: Dup Submitter: Judy Ward Opened: 1998-11-06 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [template.bitset].

View all issues with Dup status.

Duplicate of: 778

Discussion:

The following code does not compile with the EDG compiler:

#include <bitset>
using namespace std;
bitset<32> b("111111111");

If you cast the ctor argument to a string, i.e.:

bitset<32> b(string("111111111"));

then it will compile. The reason is that bitset has the following templatized constructor:

template <class charT, class traits, class Allocator>
explicit bitset (const basic_string<charT, traits, Allocator>& str, ...);

According to the compiler vendor, Steve Adamcyk at EDG, the user cannot pass this template constructor a const char* and expect a conversion to basic_string. The reason is "When you have a template constructor, it can get used in contexts where type deduction can be done. Type deduction basically comes up with exact matches, not ones involving conversions."

I don't think the intention when this constructor became templatized was for construction from a const char* to no longer work.

Proposed resolution:

Add to 22.9.2 [template.bitset] a bitset constructor declaration

explicit bitset(const char*);

and in Section 22.9.2.2 [bitset.cons] add:

explicit bitset(const char* str);

Effects:
    Calls bitset((string) str, 0, string::npos);

Rationale:

Although the problem is real, the standard is designed that way so it is not a defect. Education is the immediate workaround. A future standard may wish to consider the Proposed Resolution as an extension.


121(i). Detailed definition for ctype<wchar_t> specialization

Section: 30.3.1.2.1 [locale.category] Status: NAD Submitter: Judy Ward Opened: 1998-12-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.category].

View all issues with NAD status.

Discussion:

Section 22.1.1.1.1 has the following listed in Table 51: ctype<char> , ctype<wchar_t>.

Also Section 30.4.2.2 [locale.ctype] says:

The instantiations required in Table 51 (22.1.1.1.1) namely ctype<char> and ctype<wchar_t> , implement character classing appropriate to the implementation's native character set.

However, Section 30.4.2.4 [facet.ctype.special] only has a detailed description of the ctype<char> specialization, not the ctype<wchar_t> specialization.

Proposed resolution:

Add the ctype<wchar_t> detailed class description to Section 30.4.2.4 [facet.ctype.special].

Rationale:

Specialization for wchar_t is not needed since the default is acceptable.


128(i). Need open_mode() function for file stream, string streams, file buffers, and string  buffers

Section: 31.8 [string.streams], 31.10 [file.streams] Status: NAD Submitter: Angelika Langer Opened: 1999-02-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [string.streams].

View all issues with NAD status.

Discussion:

The following question came from Thorsten Herlemann:

You can set a mode when constructing or opening a file-stream or filebuf, e.g. ios::in, ios::out, ios::binary, ... But how can I get that mode later on, e.g. in my own operator << or operator >> or when I want to check whether a file-stream or file-buffer object passed as parameter is opened for input or output or binary? Is there no possibility? Is this a design-error in the standard C++ library?

It is indeed impossible to find out what a stream's or stream buffer's open mode is, and without that knowledge you don't know how certain operations behave. Just think of the append mode.

Both streams and stream buffers should have a mode() function that returns the current open mode setting.

[ post Bellevue: Alisdair requested to re-Open. ]

[ 2009-07 Frankfurt ]

Neither Howard nor Bill has received a customer request for this.

No consensus for change. The programmer can save this information to the side.

Moved to NAD.

Proposed resolution:

For stream buffers, add a function to the base class as a non-virtual function qualified as const to 31.6.3 [streambuf]:

    openmode mode() const;

    Returns the current open mode.

With streams, I'm not sure what to suggest. In principle, the mode could already be returned by ios_base, but the mode is only initialized for file and string stream objects, unless I'm overlooking anything. For this reason it should be added to the most derived stream classes. Alternatively, it could be added to basic_ios and would be default initialized in basic_ios<>::init().

Rationale:

This might be an interesting extension for some future, but it is not a defect in the current standard. The Proposed Resolution is retained for future reference.


131(i). list::splice throws nothing

Section: 24.3.10.5 [list.ops] Status: NAD Submitter: Howard Hinnant Opened: 1999-03-06 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [list.ops].

View all issues with NAD status.

Discussion:

What happens if a splice operation causes the size() of a list to grow beyond max_size()?

Rationale:

Size() cannot grow beyond max_size(). 


135(i). basic_iostream doubly initialized

Section: 31.7.5.7.2 [iostream.cons] Status: NAD Submitter: Howard Hinnant Opened: 1999-03-06 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

-1- Effects Constructs an object of class basic_iostream, assigning initial values to the base classes by calling basic_istream<charT,traits>(sb) (lib.istream) and basic_ostream<charT,traits>(sb) (lib.ostream)

The called for basic_istream and basic_ostream constructors call init(sb). This means that the basic_iostream's virtual base class is initialized twice.

Proposed resolution:

Change 27.6.1.5.1, paragraph 1 to:

-1- Effects Constructs an object of class basic_iostream, assigning initial values to the base classes by calling basic_istream<charT,traits>(sb) (lib.istream).

Rationale:

The LWG agreed that the init() function is called twice, but said that this is harmless and so not a defect in the standard.


138(i). Class ctype_byname<char> redundant and misleading

Section: 30.4.2.5 [locale.codecvt] Status: NAD Submitter: Angelika Langer Opened: 1999-03-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.codecvt].

View all issues with NAD status.

Discussion:

Section 30.4.2.5 [locale.codecvt] specifies that ctype_byname<char> must be a specialization of the ctype_byname template.

It is common practice in the standard that specializations of class templates are only mentioned where the interface of the specialization deviates from the interface of the template that it is a specialization of. Otherwise, the fact whether or not a required instantiation is an actual instantiation or a specialization is left open as an implementation detail.

Clause 22.2.1.4 deviates from that practice and for that reason is misleading. The fact, that ctype_byname<char> is specified as a specialization suggests that there must be something "special" about it, but it has the exact same interface as the ctype_byname template. Clause 22.2.1.4 does not have any explanatory value, is at best redundant, at worst misleading - unless I am missing anything.

Naturally, an implementation will most likely implement ctype_byname<char> as a specialization, because the base class ctype<char> is a specialization with an interface different from the ctype template, but that's an implementation detail and need not be mentioned in the standard.

[ Summit: ]

Reopened by Alisdair.

[ 2009-07 Frankfurt ]

Moved to NAD.

Rationale:

The standard as written is mildly misleading, but the correct fix is to deal with the underlying problem in the ctype_byname base class, not in the specialization. See issue 228.


140(i). map<Key, T>::value_type does not satisfy the assignable requirement

Section: 24.4.4 [map] Status: NAD Editorial Submitter: Mark Mitchell Opened: 1999-04-14 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [map].

View all issues with NAD Editorial status.

Discussion:

24.2 [container.requirements]

expression         return type      pre/post-condition
-------------     -----------      -------------------
X::value_type    T                    T is assignable

24.4.4 [map]

A map satisfies all the requirements of a container.

For a map<Key, T> ... the value_type is pair<const Key, T>.

There's a contradiction here. In particular, `pair<const Key, T>' is not assignable; the `const Key' cannot be assigned to. So,  map<Key, T>::value_type does not satisfy the assignable requirement imposed by a container.

[See issue 103 for the slightly related issue of modification of set keys.]

Rationale:

The LWG believes that the standard is inconsistent, but that this is a design problem rather than a strict defect. May wish to reconsider for the next standard.


143(i). C .h header wording unclear

Section: 17.14 [support.c.headers] Status: NAD Submitter: Christophe de Dinechin Opened: 1999-05-04 Last modified: 2023-02-07

Priority: Not Prioritized

View all other issues in [support.c.headers].

View all issues with NAD status.

Discussion:

[depr.c.headers] paragraph 2 reads:

Each C header, whose name has the form name.h, behaves as if each name placed in the Standard library namespace by the corresponding cname header is also placed within the namespace scope of the namespace std and is followed by an explicit using-declaration (_namespace.udecl_)

I think it should mention the global name space somewhere...  Currently, it indicates that name placed in std is also placed in std...

I don't know what is the correct wording. For instance, if struct tm is defined in time.h, ctime declares std::tm. However, the current wording seems ambiguous regarding which of the following would occur for use of both ctime and time.h:

// version 1:
namespace std {
        struct tm { ... };
}
using std::tm;

// version 2:
struct tm { ... };
namespace std {
        using ::tm;
}

// version 3:
struct tm { ... };
namespace std {
        struct tm { ... };
}

I think version 1 is intended.

[Kona: The LWG agreed that the wording is not clear. It also agreed that version 1 is intended, version 2 is not equivalent to version 1, and version 3 is clearly not intended. The example below was constructed by Nathan Myers to illustrate why version 2 is not equivalent to version 1.

Although not equivalent, the LWG is unsure if (2) is enough of a problem to be prohibited. Points discussed in favor of allowing (2):

]

Example:

#include <time.h>
#include <utility>

int main() {
    std::tm * t;
    make_pair( t, t ); // okay with version 1 due to Koenig lookup
                       // fails with version 2; make_pair not found
    return 0;
}

Proposed resolution:

Replace [depr.c.headers] paragraph 2 with:

Each C header, whose name has the form name.h, behaves as if each name placed in the Standard library namespace by the corresponding cname header is also placed within the namespace scope of the namespace std by name.h and is followed by an explicit using-declaration (_namespace.udecl_) in global scope.

Rationale:

The current wording in the standard is the result of a difficult compromise that averted delay of the standard. Based on discussions in Tokyo it is clear that there is no still no consensus on stricter wording, so the issue has been closed. It is suggested that users not write code that depends on Koenig lookup of C library functions.


145(i). adjustfield lacks default value

Section: 31.5.4.2 [basic.ios.cons] Status: NAD Submitter: Angelika Langer Opened: 1999-05-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [basic.ios.cons].

View all issues with NAD status.

Discussion:

There is no initial value for the adjustfield defined, although many people believe that the default adjustment were right. This is a common misunderstanding. The standard only defines that, if no adjustment is specified, all the predefined inserters must add fill characters before the actual value, which is "as if" the right flag were set. The flag itself need not be set.

When you implement a user-defined inserter you cannot rely on right being the default setting for the adjustfield. Instead, you must be prepared to find none of the flags set and must keep in mind that in this case you should make your inserter behave "as if" the right flag were set. This is surprising to many people and complicates matters more than necessary.

Unless there is a good reason why the adjustfield should not be initialized I would suggest to give it the default value that everybody expects anyway.

Rationale:

This is not a defect. It is deliberate that the default is no bits set. Consider Arabic or Hebrew, for example. See 30.4.3.3.3 [facet.num.put.virtuals] paragraph 19, Table 61 - Fill padding.


157(i). Meaningless error handling for pword() and iword()

Section: 31.5.2.6 [ios.base.storage] Status: Dup Submitter: Dietmar Kühl Opened: 1999-07-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [ios.base.storage].

View all issues with Dup status.

Duplicate of: 41

Discussion:

According to paragraphs 2 and 4 of 31.5.2.6 [ios.base.storage], the functions iword() and pword() "set the badbit (which might throw an exception)" on failure. ... but what does it mean for ios_base to set the badbit? The state facilities of the IOStream library are defined in basic_ios, a derived class! It would be possible to attempt a down cast but then it would be necessary to know the character type used...

Rationale:


162(i). Really "formatted input functions"?

Section: 31.7.5.3.3 [istream.extractors] Status: Dup Submitter: Dietmar Kühl Opened: 1999-07-20 Last modified: 2017-04-22

Priority: Not Prioritized

View all other issues in [istream.extractors].

View all issues with Dup status.

Duplicate of: 60

Discussion:

It appears to be somewhat nonsensical to consider the functions defined in the paragraphs 1 to 5 to be "Formatted input function" but since these functions are defined in a section labeled "Formatted input functions" it is unclear to me whether these operators are considered formatted input functions which have to conform to the "common requirements" from 31.7.5.3.1 [istream.formatted.reqmts]: If this is the case, all manipulators, not just ws, would skip whitespace unless noskipws is set (... but setting noskipws using the manipulator syntax would also skip whitespace :-)

See also issue 166 for the same problem in formatted output

Rationale:


163(i). Return of gcount() after a call to gcount

Section: 31.7.5.4 [istream.unformatted] Status: Dup Submitter: Dietmar Kühl Opened: 1999-07-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [istream.unformatted].

View all issues with Dup status.

Duplicate of: 60

Discussion:

It is not clear which functions are to be considered unformatted input functions. As written, it seems that all functions in 31.7.5.4 [istream.unformatted] are unformatted input functions. However, it does not really make much sense to construct a sentry object for gcount(), sync(), ... Also it is unclear what happens to the gcount() if eg. gcount(), putback(), unget(), or sync() is called: These functions don't extract characters, some of them even "unextract" a character. Should this still be reflected in gcount()? Of course, it could be read as if after a call to gcount() gcount() return 0 (the last unformatted input function, gcount(), didn't extract any character) and after a call to putback() gcount() returns -1 (the last unformatted input function putback() did "extract" back into the stream). Correspondingly for unget(). Is this what is intended? If so, this should be clarified. Otherwise, a corresponding clarification should be used.

Rationale:


166(i). Really "formatted output functions"?

Section: 31.7.6.3.3 [ostream.inserters] Status: Dup Submitter: Dietmar Kühl Opened: 1999-07-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with Dup status.

Duplicate of: 60

Discussion:

From 31.7.6.3.1 [ostream.formatted.reqmts] it appears that all the functions defined in 31.7.6.3.3 [ostream.inserters] have to construct a sentry object. Is this really intended?

This is basically the same problem as issue 162 but for output instead of input.

Rationale:


177(i). Complex operators cannot be explicitly instantiated

Section: 28.4.6 [complex.ops] Status: NAD Submitter: Judy Ward Opened: 1999-07-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [complex.ops].

View all issues with NAD status.

Discussion:

A user who tries to explicitly instantiate a complex non-member operator will get compilation errors. Below is a simplified example of the reason why. The problem is that iterator_traits cannot be instantiated on a non-pointer type like float, yet when the compiler is trying to decide which operator+ needs to be instantiated it must instantiate the declaration to figure out the first argument type of a reverse_iterator operator.

namespace std {
template <class Iterator> 
struct iterator_traits
{
    typedef typename Iterator::value_type value_type;
};

template <class T> class reverse_iterator;

// reverse_iterator operator+
template <class T> 
reverse_iterator<T> operator+
(typename iterator_traits<T>::difference_type, const reverse_iterator<T>&);

template <class T> struct complex {};

// complex operator +
template <class T>
complex<T> operator+ (const T& lhs, const complex<T>& rhs) 
{ return complex<T>();} 
}

// request for explicit instantiation
template std::complex<float> std::operator+<float>(const float&, 
     const std::complex<float>&);

See also c++-stdlib reflector messages: lib-6814, 6815, 6816.

Rationale:

Implementors can make minor changes and the example will work. Users are not affected in any case.

According to John Spicer, It is possible to explicitly instantiate these operators using different syntax: change "std::operator+<float>" to "std::operator+".

The proposed resolution of issue 120 is that users will not be able to explicitly instantiate standard library templates. If that resolution is accepted then library implementors will be the only ones that will be affected by this problem, and they must use the indicated syntax.


178(i). Should clog and cerr initially be tied to cout?

Section: 31.4.3 [narrow.stream.objects] Status: NAD Submitter: Judy Ward Opened: 1999-07-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [narrow.stream.objects].

View all issues with NAD status.

Discussion:

Section 27.3.1 says "After the object cerr is initialized, cerr.flags() & unitbuf is nonzero. Its state is otherwise the same as required for ios_base::init (lib.basic.ios.cons). It doesn't say anything about the the state of clog. So this means that calling cerr.tie() and clog.tie() should return 0 (see Table 89 for ios_base::init effects).

Neither of the popular standard library implementations that I tried does this, they both tie cerr and clog to &cout. I would think that would be what users expect.

Rationale:

The standard is clear as written.

27.3.1/5 says that "After the object cerr is initialized, cerr.flags() & unitbuf is nonzero. Its state is otherwise the same as required for ios_base::init (27.4.4.1)." Table 89 in 27.4.4.1, which gives the postconditions of basic_ios::init(), says that tie() is 0. (Other issues correct ios_base::init to basic_ios::init().)


188(i). valarray helpers missing augmented assignment operators

Section: 28.6.2.7 [valarray.cassign] Status: NAD Submitter: Gabriel Dos Reis Opened: 1999-08-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [valarray.cassign].

View all issues with NAD status.

Discussion:

26.5.2.6 defines augmented assignment operators valarray<T>::op=(const T&), but fails to provide corresponding versions for the helper classes. Thus making the following illegal:

#include <valarray>

int main()
{
std::valarray<double> v(3.14, 1999);

v[99] *= 2.0; // Ok

std::slice s(0, 50, 2);

v[s] *= 2.0; // ERROR
}

I can't understand the intent of that omission. It makes the valarray library less intuitive and less useful.

Rationale:

Although perhaps an unfortunate design decision, the omission is not a defect in the current standard.  A future standard may wish to add the missing operators.


190(i). min() and max() functions should be std::binary_functions

Section: 27.8.9 [alg.min.max] Status: NAD Submitter: Mark Rintoul Opened: 1999-08-26 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.min.max].

View all issues with NAD status.

Discussion:

Both std::min and std::max are defined as template functions. This is very different than the definition of std::plus (and similar structs) which are defined as function objects which inherit std::binary_function.

This lack of inheritance leaves std::min and std::max somewhat useless in standard library algorithms which require a function object that inherits std::binary_function.

[ post Bellevue: Alisdair requested to re-Open. ]

[ 2009-07 Frankfurt ]

C++0x has lambdas to address this problem now.

Moved to NAD.

Rationale:

Although perhaps an unfortunate design decision, the omission is not a defect in the current standard.  A future standard may wish to consider additional function objects.


191(i). Unclear complexity for algorithms such as binary search

Section: 27.8.4 [alg.binary.search] Status: NAD Submitter: Nico Josuttis Opened: 1999-10-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.binary.search].

View all issues with NAD status.

Discussion:

The complexity of binary_search() is stated as "At most log(last-first) + 2 comparisons", which seems to say that the algorithm has logarithmic complexity. However, this algorithms is defined for forward iterators. And for forward iterators, the need to step element-by-element results into linear complexity. But such a statement is missing in the standard. The same applies to lower_bound(), upper_bound(), and equal_range(). 

However, strictly speaking the standard contains no bug here. So this might considered to be a clarification or improvement.

Rationale:

The complexity is expressed in terms of comparisons, and that complexity can be met even if the number of iterators accessed is linear. Paragraph 1 already says exactly what happens to iterators.


192(i). a.insert(p,t) is inefficient and overconstrained

Section: 24.2.7 [associative.reqmts] Status: NAD Submitter: Ed Brey Opened: 1999-06-06 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with NAD status.

Duplicate of: 233

Discussion:

As defined in 23.1.2, paragraph 7 (table 69), a.insert(p,t) suffers from several problems:

expression return type pre/post-condition complexity
a.insert(p,t) iterator inserts t if and only if there is no element with key equivalent to the key of t in containers with unique keys; always inserts t in containers with equivalent keys. always returns the iterator pointing to the element with key equivalent to the key of t . iterator p is a hint pointing to where the insert should start to search. logarithmic in general, but amortized constant if t is inserted right after p .

1. For a container with unique keys, only logarithmic complexity is guaranteed if no element is inserted, even though constant complexity is always possible if p points to an element equivalent to t.

2. For a container with equivalent keys, the amortized constant complexity guarantee is only useful if no key equivalent to t exists in the container. Otherwise, the insertion could occur in one of multiple locations, at least one of which would not be right after p.

3. By guaranteeing amortized constant complexity only when t is inserted after p, it is impossible to guarantee constant complexity if t is inserted at the beginning of the container. Such a problem would not exist if amortized constant complexity was guaranteed if t is inserted before p, since there is always some p immediately before which an insert can take place.

4. For a container with equivalent keys, p does not allow specification of where to insert the element, but rather only acts as a hint for improving performance. This negates the added functionality that p would provide if it specified where within a sequence of equivalent keys the insertion should occur. Specifying the insert location provides more control to the user, while providing no disadvantage to the container implementation.

Proposed resolution:

In 24.2.7 [associative.reqmts] paragraph 7, replace the row in table 69 for a.insert(p,t) with the following two rows:

expression return type pre/post-condition complexity
a_uniq.insert(p,t) iterator inserts t if and only if there is no element with key equivalent to the key of t. returns the iterator pointing to the element with key equivalent to the key of t. logarithmic in general, but amortized constant if t is inserted right before p or p points to an element with key equivalent to t.
a_eq.insert(p,t) iterator inserts t and returns the iterator pointing to the newly inserted element. t is inserted right before p if doing so preserves the container ordering. logarithmic in general, but amortized constant if t is inserted right before p.

Rationale:

Too big a change.  Furthermore, implementors report checking both before p and after p, and don't want to change this behavior.


194(i). rdbuf() functions poorly specified

Section: 31.5.4 [ios] Status: NAD Submitter: Steve Clamage Opened: 1999-09-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

In classic iostreams, base class ios had an rdbuf function that returned a pointer to the associated streambuf. Each derived class had its own rdbuf function that returned a pointer of a type reflecting the actual type derived from streambuf. Because in ARM C++, virtual function overrides had to have the same return type, rdbuf could not be virtual.

In standard iostreams, we retain the non-virtual rdbuf function design, and in addition have an overloaded rdbuf function that sets the buffer pointer. There is no need for the second function to be virtual nor to be implemented in derived classes.

Minor question: Was there a specific reason not to make the original rdbuf function virtual?

Major problem: Friendly compilers warn about functions in derived classes that hide base-class overloads. Any standard implementation of iostreams will result in such a warning on each of the iostream classes, because of the ill-considered decision to overload rdbuf only in a base class.

In addition, users of the second rdbuf function must use explicit qualification or a cast to call it from derived classes. An explicit qualification or cast to basic_ios would prevent access to any later overriding version if there was one.

What I'd like to do in an implementation is add a using- declaration for the second rdbuf function in each derived class. It would eliminate warnings about hiding functions, and would enable access without using explicit qualification. Such a change I don't think would change the behavior of any valid program, but would allow invalid programs to compile:

 filebuf mybuf;
 fstream f;
 f.rdbuf(mybuf); // should be an error, no visible rdbuf

I'd like to suggest this problem as a defect, with the proposed resolution to require the equivalent of a using-declaration for the rdbuf function that is not replaced in a later derived class. We could discuss whether replacing the function should be allowed.

Rationale:

For historical reasons, the standard is correct as written. There is a subtle difference between the base class rdbuf() and derived class rdbuf(). The derived class rdbuf() always returns the original streambuf, whereas the base class rdbuf() will return the "current streambuf" if that has been changed by the variant you mention.

Permission is not required to add such an extension. See 16.4.6.5 [member.functions].


196(i). Placement new example has alignment problems

Section: 17.6.3.4 [new.delete.placement] Status: Dup Submitter: Herb Sutter Opened: 1998-12-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [new.delete.placement].

View all issues with Dup status.

Duplicate of: 114

Discussion:

The example in 17.6.3.4 [new.delete.placement] paragraph 4 reads:

[Example: This can be useful for constructing an object at a known address:

   char place[sizeof(Something)];
   Something* p = new (place) Something();

end example]

This example has potential alignment problems.

Rationale:


197(i). max_size() underspecified

Section: 16.4.4.6 [allocator.requirements], 24.2 [container.requirements] Status: NAD Submitter: Andy Sawyer Opened: 1999-10-21 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with NAD status.

Discussion:

Must the value returned by max_size() be unchanged from call to call?

Must the value returned from max_size() be meaningful?

Possible meanings identified in lib-6827:

1) The largest container the implementation can support given "best case" conditions - i.e. assume the run-time platform is "configured to the max", and no overhead from the program itself. This may possibly be determined at the point the library is written, but certainly no later than compile time.

2) The largest container the program could create, given "best case" conditions - i.e. same platform assumptions as (1), but take into account any overhead for executing the program itself. (or, roughly "storage=storage-sizeof(program)"). This does NOT include any resource allocated by the program. This may (or may not) be determinable at compile time.

3) The largest container the current execution of the program could create, given knowledge of the actual run-time platform, but again, not taking into account any currently allocated resource. This is probably best determined at program start-up.

4) The largest container the current execution program could create at the point max_size() is called (or more correctly at the point max_size() returns :-), given it's current environment (i.e. taking into account the actual currently available resources). This, obviously, has to be determined dynamically each time max_size() is called.

Proposed resolution:

Rationale:

max_size() isn't useful for very many things, and the existing wording is sufficiently clear for the few cases that max_size() can be used for. None of the attempts to change the existing wording were an improvement.

It is clear to the LWG that the value returned by max_size() can't change from call to call.


203(i). basic_istream::sentry::sentry() is uninstantiable with ctype<user-defined type>

Section: 31.7.5.2.4 [istream.sentry] Status: NAD Submitter: Matt McClure and Dietmar Kühl Opened: 2000-01-01 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [istream.sentry].

View all issues with NAD status.

Discussion:

27.6.1.1.2 Paragraph 4 states:

To decide if the character c is a whitespace character, the constructor performs ''as if'' it executes the following code fragment: 

const ctype<charT>& ctype = use_facet<ctype<charT> >(is.getloc());
if (ctype.is(ctype.space,c)!=0)
// c is a whitespace character.

But Table 51 in 22.1.1.1.1 only requires an implementation to provide specializations for ctype<char> and ctype<wchar_t>. If sentry's constructor is implemented using ctype, it will be uninstantiable for a user-defined character type charT, unless the implementation has provided non-working (since it would be impossible to define a correct ctype<charT> specialization for an arbitrary charT) definitions of ctype's virtual member functions.

It seems the intent the standard is that sentry should behave, in every respect, not just during execution, as if it were implemented using ctype, with the burden of providing a ctype specialization falling on the user. But as it is written, nothing requires the translation of sentry's constructor to behave as if it used the above code, and it would seem therefore, that sentry's constructor should be instantiable for all character types.

Note: If I have misinterpreted the intent of the standard with respect to sentry's constructor's instantiability, then a note should be added to the following effect:

An implementation is forbidden from using the above code if it renders the constructor uninstantiable for an otherwise valid character type.

In any event, some clarification is needed.

Rationale:

It is possible but not easy to instantiate on types other than char or wchar_t; many things have to be done first. That is by intention and is not a defect.


204(i). distance(first, last) when "last" is before "first"

Section: 25.4.3 [iterator.operations] Status: NAD Submitter: Rintala Matti Opened: 2000-01-28 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [iterator.operations].

View all other issues in [iterator.operations].

View all issues with NAD status.

Discussion:

Section 24.3.4 describes the function distance(first, last) (where first and last are iterators) which calculates "the number of increments or decrements needed to get from 'first' to 'last'".

The function should work for forward, bidirectional and random access iterators, and there is a requirement 24.3.4.5 which states that "'last' must be reachable from 'first'".

With random access iterators the function is easy to implement as "last - first".

With forward iterators it's clear that 'first' must point to a place before 'last', because otherwise 'last' would not be reachable from 'first'.

But what about bidirectional iterators? There 'last' is reachable from 'first' with the -- operator even if 'last' points to an earlier position than 'first'. However, I cannot see how the distance() function could be implemented if the implementation does not know which of the iterators points to an earlier position (you cannot use ++ or -- on either iterator if you don't know which direction is the "safe way to travel").

The paragraph 24.3.4.1 states that "for ... bidirectional iterators they use ++ to provide linear time implementations". However, the ++ operator is not mentioned in the reachability requirement. Furthermore 24.3.4.4 explicitly mentions that distance() returns the number of increments _or decrements_, suggesting that it could return a negative number also for bidirectional iterators when 'last' points to a position before 'first'.

Is a further requirement is needed to state that for forward and bidirectional iterators "'last' must be reachable from 'first' using the ++ operator". Maybe this requirement might also apply to random access iterators so that distance() would work the same way for every iterator category?

Rationale:

"Reachable" is defined in the standard in 25.3.4 [iterator.concepts] paragraph 6. The definition is only in terms of operator++(). The LWG sees no defect in the standard.


205(i). numeric_limits unclear on how to determine floating point types

Section: 17.3.5.2 [numeric.limits.members] Status: NAD Submitter: Steve Cleary Opened: 2000-01-28 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [numeric.limits.members].

View all issues with NAD status.

Discussion:

In several places in 17.3.5.2 [numeric.limits.members], a member is described as "Meaningful for all floating point types." However, no clear method of determining a floating point type is provided.

In 17.3.5.3 [numeric.special], paragraph 1 states ". . . (for example, epsilon() is only meaningful if is_integer is false). . ." which suggests that a type is a floating point type if is_specialized is true and is_integer is false; however, this is unclear.

When clarifying this, please keep in mind this need of users: what exactly is the definition of floating point? Would a fixed point or rational representation be considered one? I guess my statement here is that there could also be types that are neither integer or (strictly) floating point.

Rationale:

It is up to the implementor of a user define type to decide if it is a floating point type.


207(i). ctype<char> members return clause incomplete

Section: 30.4.2.4.3 [facet.ctype.char.members] Status: Dup Submitter: Robert Klarer Opened: 1999-11-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [facet.ctype.char.members].

View all issues with Dup status.

Duplicate of: 153

Discussion:

The widen and narrow member functions are described in 22.2.1.3.2, paragraphs 9-11. In each case we have two overloaded signatures followed by a Returns clause. The Returns clause only describes one of the overloads.

Proposed resolution:

Change the returns clause in 30.4.2.4.3 [facet.ctype.char.members] paragraph 10 from:

    Returns: do_widen(low, high, to).

to:

    Returns: do_widen(c) or do_widen(low, high, to), respectively.

Change the returns clause in 30.4.2.4.3 [facet.ctype.char.members] paragraph 11 from:

    Returns: do_narrow(low, high, to).

to:

    Returns: do_narrow(c) or do_narrow(low, high, to), respectively.

Rationale:

Subsumed by issue 153, which addresses the same paragraphs.


213(i). Math function overloads ambiguous

Section: 28.7 [c.math] Status: NAD Submitter: Nico Josuttis Opened: 2000-02-26 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [c.math].

View all issues with NAD status.

Discussion:

Due to the additional overloaded versions of numeric functions for float and long double according to Section 26.5, calls such as int x; std::pow (x, 4) are ambiguous now in a standard conforming implementation. Current implementations solve this problem very different (overload for all types, don't overload for float and long double, use preprocessor, follow the standard and get ambiguities).

This behavior should be standardized or at least identified as implementation defined.

Rationale:

These math issues are an understood and accepted consequence of the design. They have been discussed several times in the past. Users must write casts or write floating point expressions as arguments.


215(i). Can a map's key_type be const?

Section: 24.2.7 [associative.reqmts] Status: NAD Submitter: Judy Ward Opened: 2000-02-29 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with NAD status.

Discussion:

A user noticed that this doesn't compile with the Rogue Wave library because the rb_tree class declares a key_allocator, and allocator<const int> is not legal, I think:

map < const int, ... > // legal?

which made me wonder whether it is legal for a map's key_type to be const. In email from Matt Austern he said:

I'm not sure whether it's legal to declare a map with a const key type. I hadn't thought about that question until a couple weeks ago. My intuitive feeling is that it ought not to be allowed, and that the standard ought to say so. It does turn out to work in SGI's library, though, and someone in the compiler group even used it. Perhaps this deserves to be written up as an issue too.

Rationale:

The "key is assignable" requirement from table 69 in 24.2.7 [associative.reqmts] already implies the key cannot be const.


216(i). setbase manipulator description flawed

Section: 31.7.7 [std.manip] Status: Dup Submitter: Hyman Rosen Opened: 2000-02-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [std.manip].

View all issues with Dup status.

Duplicate of: 193

Discussion:

31.7.7 [std.manip] paragraph 5 says:

smanip setbase(int base);

Returns: An object s of unspecified type such that if out is an (instance of) basic_ostream then the expression out<<s behaves as if f(s) were called, in is an (instance of) basic_istream then the expression in>>s behaves as if f(s) were called. Where f can be defined as:

ios_base& f(ios_base& str, int base)
{
  // set basefield
  str.setf(n == 8 ? ios_base::oct :
                n == 10 ? ios_base::dec :
                n == 16 ? ios_base::hex :
                  ios_base::fmtflags(0), ios_base::basefield);
  return str;
}

There are two problems here. First, f takes two parameters, so the description needs to say that out<<s and in>>s behave as if f(s,base) had been called. Second, f is has a parameter named base, but is written as if the parameter was named n.

Actually, there's a third problem. The paragraph has grammatical errors. There needs to be an "and" after the first comma, and the "Where f" sentence fragment needs to be merged into its preceding sentence. You may also want to format the function a little better. The formatting above is more-or-less what the Standard contains.

Rationale:

The resolution of this defect is subsumed by the proposed resolution for issue 193.

[Tokyo: The LWG agrees that this is a defect and notes that it occurs additional places in the section, all requiring fixes.]


218(i). Algorithms do not use binary predicate objects for default comparisons

Section: 27.8 [alg.sorting] Status: NAD Submitter: Pablo Halpern Opened: 2000-03-06 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.sorting].

View all issues with NAD status.

Discussion:

Many of the algorithms take an argument, pred, of template parameter type BinaryPredicate or an argument comp of template parameter type Compare. These algorithms usually have an overloaded version that does not take the predicate argument. In these cases pred is usually replaced by the use of operator== and comp is replaced by the use of operator<.

This use of hard-coded operators is inconsistent with other parts of the library, particularly the containers library, where equality is established using equal_to<> and ordering is established using less<>. Worse, the use of operator<, would cause the following innocent-looking code to have undefined behavior:

vector<string*> vec;
sort(vec.begin(), vec.end());

The use of operator< is not defined for pointers to unrelated objects. If std::sort used less<> to compare elements, then the above code would be well-defined, since less<> is explicitly specialized to produce a total ordering of pointers.

Rationale:

This use of operator== and operator< was a very deliberate, conscious, and explicitly made design decision; these operators are often more efficient. The predicate forms are available for users who don't want to rely on operator== and operator<.


219(i). find algorithm missing version that takes a binary predicate argument

Section: 27.6.6 [alg.find] Status: NAD Submitter: Pablo Halpern Opened: 2000-03-06 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.find].

View all issues with NAD status.

Discussion:

The find function always searches for a value using operator== to compare the value argument to each element in the input iterator range. This is inconsistent with other find-related functions such as find_end and find_first_of, which allow the caller to specify a binary predicate object to be used for determining equality. The fact that this can be accomplished using a combination of find_if and bind_1st or bind_2nd does not negate the desirability of a consistent, simple, alternative interface to find.

[ Summit: ]

Reopened by Alisdair.

[ 2009-07 Frankfurt ]

The same thing can be achieved using find_if (as noted in the issue).

Moved to NAD.

Proposed resolution:

In section 27.6.6 [alg.find], add a second prototype for find (between the existing prototype and the prototype for find_if), as follows:

    template<class InputIterator, class T, class BinaryPredicate>
      InputIterator find(InputIterator first, InputIterator last,
                         const T& value, BinaryPredicate bin_pred);

Change the description of the return from:

Returns: The first iterator i in the range [first, last) for which the following corresponding conditions hold: *i == value, pred(*i) != false. Returns last if no such iterator is found.

 to:

Returns: The first iterator i in the range [first, last) for which the following  corresponding condition holds: *i == value, bin_pred(*i,value) != false, pred(*) != false. Return last if no such iterator is found.

Rationale:

This is request for a pure extension, so it is not a defect in the current standard.  As the submitter pointed out, "this can be accomplished using a combination of find_if and bind_1st or bind_2nd".


236(i). ctype<char>::is() member modifies facet

Section: 30.4.2.4.3 [facet.ctype.char.members] Status: Dup Submitter: Dietmar Kühl Opened: 2000-04-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [facet.ctype.char.members].

View all issues with Dup status.

Duplicate of: 28

Discussion:

The description of the is() member in paragraph 4 of 30.4.2.4.3 [facet.ctype.char.members] is broken: According to this description, the second form of the is() method modifies the masks in the ctype object. The correct semantics if, of course, to obtain an array of masks. The corresponding method in the general case, ie. the do_is() method as described in 30.4.2.2.3 [locale.ctype.virtuals] paragraph 1 does the right thing.

Proposed resolution:

Change paragraph 4 from

The second form, for all *p in the range [low, high), assigns vec[p-low] to table()[(unsigned char)*p].

to become

The second form, for all *p in the range [low, high), assigns table()[(unsigned char)*p] to vec[p-low].

Rationale:


244(i). Must find's third argument be CopyConstructible?

Section: 27.6.6 [alg.find] Status: NAD Submitter: Andrew Koenig Opened: 2000-05-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.find].

View all issues with NAD status.

Discussion:

Is the following implementation of find acceptable?

        template<class Iter, class X>
        Iter find(Iter begin, Iter end, const X& x)
        {
            X x1 = x;           // this is the crucial statement
            while (begin != end && *begin != x1)
                ++begin;
            return begin;
        }

If the answer is yes, then it is implementation-dependent as to whether the following fragment is well formed:

        vector<string> v;

        find(v.begin(), v.end(), "foo");

At issue is whether there is a requirement that the third argument of find be CopyConstructible. There may be no problem here, but analysis is necessary.

Rationale:

There is no indication in the standard that find's third argument is required to be Copy Constructible. The LWG believes that no such requirement was intended. As noted above, there are times when a user might reasonably pass an argument that is not Copy Constructible.


245(i). Which operations on istream_iterator trigger input operations?

Section: 25.6.2 [istream.iterator] Status: NAD Submitter: Andrew Koenig Opened: 2000-05-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [istream.iterator].

View all issues with NAD status.

Discussion:

I do not think the standard specifies what operation(s) on istream iterators trigger input operations. So, for example:

        istream_iterator<int> i(cin);

        int n = *i++;

I do not think it is specified how many integers have been read from cin. The number must be at least 1, of course, but can it be 2? More?

Rationale:

The standard is clear as written: the stream is read every time operator++ is called, and it is also read either when the iterator is constructed or when operator* is called for the first time. In the example above, exactly two integers are read from cin.

There may be a problem with the interaction between istream_iterator and some STL algorithms, such as find. There are no guarantees about how many times find may invoke operator++.


246(i). a.insert(p,t) is incorrectly specified

Section: 24.2.7 [associative.reqmts] Status: Dup Submitter: Mark Rodgers Opened: 2000-05-19 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with Dup status.

Duplicate of: 233

Discussion:

Closed issue 192 raised several problems with the specification of this function, but was rejected as Not A Defect because it was too big a change with unacceptable impacts on existing implementations. However, issues remain that could be addressed with a smaller change and with little or no consequent impact.

  1. The specification is inconsistent with the original proposal and with several implementations.

    The initial implementation by Hewlett Packard only ever looked immediately before p, and I do not believe there was any intention to standardize anything other than this behavior. Consequently, current implementations by several leading implementors also look immediately before p, and will only insert after p in logarithmic time. I am only aware of one implementation that does actually look after p, and it looks before p as well. It is therefore doubtful that existing code would be relying on the behavior defined in the standard, and it would seem that fixing this defect as proposed below would standardize existing practice.

  2. The specification is inconsistent with insertion for sequence containers.

    This is difficult and confusing to teach to newcomers. All insert operations that specify an iterator as an insertion location should have a consistent meaning for the location represented by that iterator.

  3. As specified, there is no way to hint that the insertion should occur at the beginning of the container, and the way to hint that it should occur at the end is long winded and unnatural.

    For a container containing n elements, there are n+1 possible insertion locations and n+1 valid iterators. For there to be a one-to-one mapping between iterators and insertion locations, the iterator must represent an insertion location immediately before the iterator.

  4. When appending sorted ranges using insert_iterators, insertions are guaranteed to be sub-optimal.

    In such a situation, the optimum location for insertion is always immediately after the element previously inserted. The mechanics of the insert iterator guarantee that it will try and insert after the element after that, which will never be correct. However, if the container first tried to insert before the hint, all insertions would be performed in amortized constant time.

Proposed resolution:

In 23.1.2 [lib.associative.reqmts] paragraph 7, table 69, make the following changes in the row for a.insert(p,t):

assertion/note pre/post condition:
Change the last sentence from

"iterator p is a hint pointing to where the insert should start to search."

to

"iterator p is a hint indicating that immediately before p may be a correct location where the insertion could occur."

complexity:
Change the words "right after" to "immediately before".

Rationale:


249(i). Return Type of auto_ptr::operator=

Section: 99 [auto.ptr] Status: NAD Submitter: Joseph Gottman Opened: 2000-06-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [auto.ptr].

View all issues with NAD status.

Discussion:

According to section 20.4.5, the function auto_ptr::operator=() returns a reference to an auto_ptr. The reason that operator=() usually returns a reference is to facilitate code like

    int x,y,z;
    x = y = z = 1;

However, given analogous code for auto_ptrs,

    auto_ptr<int> x, y, z;
    z.reset(new int(1));
    x = y = z;

the result would be that z and y would both be set to NULL, instead of all the auto_ptrs being set to the same value. This makes such cascading assignments useless and counterintuitive for auto_ptrs.

Proposed resolution:

Change auto_ptr::operator=() to return void instead of an auto_ptr reference.

Rationale:

The return value has uses other than cascaded assignments: a user can call an auto_ptr member function, pass the auto_ptr to a function, etc. Removing the return value could break working user code.


255(i). Why do basic_streambuf<>::pbump() and gbump() take an int?

Section: 31.6.3 [streambuf] Status: NAD Submitter: Martin Sebor Opened: 2000-08-12 Last modified: 2017-06-06

Priority: Not Prioritized

View all other issues in [streambuf].

View all issues with NAD status.

Discussion:

The basic_streambuf members gbump() and pbump() are specified to take an int argument. This requirement prevents the functions from effectively manipulating buffers larger than std::numeric_limits<int>::max() characters. It also makes the common use case for these functions somewhat difficult as many compilers will issue a warning when an argument of type larger than int (such as ptrdiff_t on LLP64 architectures) is passed to either of the function. Since it's often the result of the subtraction of two pointers that is passed to the functions, a cast is necessary to silence such warnings. Finally, the usage of a native type in the functions signatures is inconsistent with other member functions (such as sgetn() and sputn()) that manipulate the underlying character buffer. Those functions take a streamsize argument.

[ 2009-07 Frankfurt ]

This is part of a bigger problem. If anyone cares enough, they should write a paper solving the bigger problem of offset types in iostreams.

This is related to the paper about large file sizes. Beman has already agreed to drop the section of that paper that deals with this.

int is big enough for reasonable buffers.

Move to NAD Future.

This is related to LWG 423.

[2017-02 in Kona, LEWG recommends NAD]

[2017-06-02 Issues Telecon]

Resolve as NAD

The previous rationale given suggested that LWG believes the change is too big for now. Actually, changing the parameter type is too big a change more or less forever, because that would break every custom streambuf type; there are too many such types in the wild to make a breaking change. The overload approach may be more plausible, but is not an entirely breakage-free solution; it can produce ambiguities, and can still break streambuf hierarchies.

Proposed resolution:

Change the signatures of these functions in the synopsis of template class basic_streambuf (27.5.2) and in their descriptions (27.5.2.3.1, p4 and 27.5.2.3.2, p4) to take a streamsize argument.

Although this change has the potential of changing the ABI of the library, the change will affect only platforms where int is different than the definition of streamsize. However, since both functions are typically inline (they are on all known implementations), even on such platforms the change will not affect any user code unless it explicitly relies on the existing type of the functions (e.g., by taking their address). Such a possibility is IMO quite remote.

Alternate Suggestion from Howard Hinnant, c++std-lib-7780:

This is something of a nit, but I'm wondering if streamoff wouldn't be a better choice than streamsize. The argument to pbump and gbump MUST be signed. But the standard has this to say about streamsize (27.4.1/2/Footnote):

[Footnote: streamsize is used in most places where ISO C would use size_t. Most of the uses of streamsize could use size_t, except for the strstreambuf constructors, which require negative values. It should probably be the signed type corresponding to size_t (which is what Posix.2 calls ssize_t). — end footnote]

This seems a little weak for the argument to pbump and gbump. Should we ever really get rid of strstream, this footnote might go with it, along with the reason to make streamsize signed.

Rationale:

The LWG believes this change is too big for now. We may wish to reconsider this for a future revision of the standard. One possibility is overloading pbump, rather than changing the signature.

[ [2006-05-04: Reopened at the request of Chris (Krzysztof Żelechowski)] ]


257(i). STL functional object and iterator inheritance.

Section: 99 [depr.base], 99 [iterator.basic] Status: NAD Submitter: Robert Dick Opened: 2000-08-17 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [depr.base].

View all issues with NAD status.

Discussion:

According to the November 1997 Draft Standard, the results of deleting an object of a derived class through a pointer to an object of its base class are undefined if the base class has a non-virtual destructor. Therefore, it is potentially dangerous to publicly inherit from such base classes.

Defect:
The STL design encourages users to publicly inherit from a number of classes which do nothing but specify interfaces, and which contain non-virtual destructors.

Attribution:
Wil Evers and William E. Kempf suggested this modification for functional objects.

Proposed resolution:

When a base class in the standard library is useful only as an interface specifier, i.e., when an object of the class will never be directly instantiated, specify that the class contains a protected destructor. This will prevent deletion through a pointer to the base class without performance, or space penalties (on any implementation I'm aware of).

As an example, replace...

    template <class Arg, class Result>
    struct unary_function {
            typedef Arg    argument_type;
            typedef Result result_type;
    };

... with...

    template <class Arg, class Result>
    struct unary_function {
            typedef Arg    argument_type;
            typedef Result result_type;
    protected:
            ~unary_function() {}
    };

Affected definitions:
 20.3.1 [lib.function.objects] -- unary_function, binary_function
 24.3.2 [lib.iterator.basic] -- iterator

Rationale:

The standard is clear as written; this is a request for change, not a defect in the strict sense. The LWG had several different objections to the proposed change. One is that it would prevent users from creating objects of type unary_function and binary_function. Doing so can sometimes be legitimate, if users want to pass temporaries as traits or tag types in generic code.


267(i). interaction of strstreambuf::overflow() and seekoff()

Section: D.15.2.4 [depr.strstreambuf.virtuals] Status: NAD Submitter: Martin Sebor Opened: 2000-10-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [depr.strstreambuf.virtuals].

View all issues with NAD status.

Discussion:

It appears that the interaction of the strstreambuf members overflow() and seekoff() can lead to undefined behavior in cases where defined behavior could reasonably be expected. The following program demonstrates this behavior:

    #include <strstream>

    int main ()
    {
         std::strstreambuf sb;
         sb.sputc ('c');

         sb.pubseekoff (-1, std::ios::end, std::ios::in);
         return !('c' == sb.sgetc ());
    }

D.7.1.1, p1 initializes strstreambuf with a call to basic_streambuf<>(), which in turn sets all pointers to 0 in 27.5.2.1, p1.

27.5.2.2.5, p1 says that basic_streambuf<>::sputc(c) calls overflow(traits::to_int_type(c)) if a write position isn't available (it isn't due to the above).

D.7.1.3, p3 says that strstreambuf::overflow(off, ..., ios::in) makes at least one write position available (i.e., it allows the function to make any positive number of write positions available).

D.7.1.3, p13 computes newoff = seekhigh - eback(). In D.7.1, p4 we see seekhigh = epptr() ? epptr() : egptr(), or seekhigh = epptr() in this case. newoff is then epptr() - eback().

D.7.1.4, p14 sets gptr() so that gptr() == eback() + newoff + off, or gptr() == epptr() + off holds.

If strstreambuf::overflow() made exactly one write position available then gptr() will be set to just before epptr(), and the program will return 0. Buf if the function made more than one write position available, epptr() and gptr() will both point past pptr() and the behavior of the program is undefined.

Proposed resolution:

Change the last sentence of D.15.2 [depr.strstreambuf] paragraph 4 from

Otherwise, seeklow equals gbeg and seekhigh is either pend, if pend is not a null pointer, or gend.

to become

Otherwise, seeklow equals gbeg and seekhigh is either gend if 0 == pptr(), or pbase() + max where max is the maximum value of pptr() - pbase() ever reached for this stream.

[ pre-Copenhagen: Dietmar provided wording for proposed resolution. ]

[ post-Copenhagen: Fixed a typo: proposed resolution said to fix 4.7.1, not D.7.1. ]

Rationale:

This is related to issue 65: it's not clear what it means to seek beyond the current area. Without resolving issue 65 we can't resolve this. As with issue 65, the library working group does not wish to invest time nailing down corner cases in a deprecated feature.


269(i). cstdarg and unnamed parameters

Section: 17.9 [support.exception] Status: NAD Submitter: J. Stephen Adamczyk Opened: 2000-10-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [support.exception].

View all issues with NAD status.

Discussion:

One of our customers asks whether this is valid C++:

   #include <cstdarg>

   void bar(const char *, va_list);

   void
   foo(const char *file, const char *, ...)
   {
     va_list ap;
     va_start(ap, file);
     bar(file, ap);
     va_end(ap);
   }

The issue being whether it is valid to use cstdarg when the final parameter before the "..." is unnamed. cstdarg is, as far as I can tell, inherited verbatim from the C standard. and the definition there (7.8.1.1 in the ISO C89 standard) refers to "the identifier of the rightmost parameter". What happens when there is no such identifier?

My personal opinion is that this should be allowed, but some tweak might be required in the C++ standard.

Rationale:

Not a defect, the C and C++ standards are clear. It is impossible to use varargs if the parameter immediately before "..." has no name, because that is the parameter that must be passed to va_start. The example given above is broken, because va_start is being passed the wrong parameter.

There is no support for extending varargs to provide additional functionality beyond what's currently there. For reasons of C/C++ compatibility, it is especially important not to make gratuitous changes in this part of the C++ standard. The C committee has already been requested not to touch this part of the C standard unless necessary.


277(i). Normative encouragement in allocator requirements unclear

Section: 16.4.4.6 [allocator.requirements] Status: NAD Submitter: Matt Austern Opened: 2000-11-07 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with NAD status.

Discussion:

In 20.1.5, paragraph 5, the standard says that "Implementors are encouraged to supply libraries that can accept allocators that encapsulate more general memory models and that support non-equal instances." This is intended as normative encouragement to standard library implementors. However, it is possible to interpret this sentence as applying to nonstandard third-party libraries.

Proposed resolution:

In 20.1.5, paragraph 5, change "Implementors" to "Implementors of the library described in this International Standard".

Rationale:

The LWG believes the normative encouragement is already sufficiently clear, and that there are no important consequences even if it is misunderstood.


279(i). const and non-const iterators should have equivalent typedefs

Section: 24.2 [container.requirements] Status: NAD Submitter: Steve Cleary Opened: 2000-11-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [container.requirements].

View all issues with NAD status.

Discussion:

This came from an email from Steve Cleary to Fergus in reference to issue 179. The library working group briefly discussed this in Toronto and believes it should be a separate issue.

Steve said: "We may want to state that the const/non-const iterators must have the same difference type, size_type, and category."

(Comment from Judy) I'm not sure if the above sentence should be true for all const and non-const iterators in a particular container, or if it means the container's iterator can't be compared with the container's const_iterator unless the above it true. I suspect the former.

Proposed resolution:

In Section: 24.2 [container.requirements], table 65, in the assertion/note pre/post condition for X::const_iterator, add the following:

typeid(X::const_iterator::difference_type) == typeid(X::iterator::difference_type)

typeid(X::const_iterator::size_type) == typeid(X::iterator::size_type)

typeid(X::const_iterator::category) == typeid(X::iterator::category)

Rationale:

Going through the types one by one: Iterators don't have a size_type. We already know that the difference types are identical, because the container requirements already say that the difference types of both X::iterator and X::const_iterator are both X::difference_type. The standard does not require that X::iterator and X::const_iterator have the same iterator category, but the LWG does not see this as a defect: it's possible to imagine cases in which it would be useful for the categories to be different.

It may be desirable to require X::iterator and X::const_iterator to have the same value type, but that is a new issue. (Issue 322.)


287(i). conflicting ios_base fmtflags

Section: 31.5.2.3 [fmtflags.state] Status: NAD Submitter: Judy Ward Opened: 2000-12-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [fmtflags.state].

View all issues with NAD status.

Discussion:

The Effects clause for ios_base::setf(fmtflags fmtfl) says "Sets fmtfl in flags()". What happens if the user first calls ios_base::scientific and then calls ios_base::fixed or vice-versa? This is an issue for all of the conflicting flags, i.e. ios_base::left and ios_base::right or ios_base::dec, ios_base::hex and ios_base::oct.

I see three possible solutions:

  1. Set ios_base::failbit whenever the user specifies a conflicting flag with one previously explicitly set. If the constructor is supposed to set ios_base::dec (see discussion below), then the user setting hex or oct format after construction will not set failbit.
  2. The last call to setf "wins", i.e. it clears any conflicting previous setting.
  3. All the flags that the user specifies are set, but when actually interpreting them, fixed always override scientific, right always overrides left, dec overrides hex which overrides oct.

Most existing implementations that I tried seem to conform to resolution #3, except that when using the iomanip manipulator hex or oct then that always overrides dec, but calling setf(ios_base::hex) doesn't.

There is a sort of related issue, which is that although the ios_base constructor says that each ios_base member has an indeterminate value after construction, all the existing implementations I tried explicitly set ios_base::dec.

Proposed resolution:

Rationale:

adjustfield, basefield, and floatfield are each multi-bit fields. It is possible to set multiple bits within each of those fields. (For example, dec and oct). These fields are used by locale facets. The LWG reviewed the way in which each of those three fields is used, and believes that in each case the behavior is well defined for any possible combination of bits. See for example Table 58, in 30.4.3.3.3 [facet.num.put.virtuals], noting the requirement in paragraph 6 of that section.

Users are advised to use manipulators, or else use the two-argument version of setf, to avoid unexpected behavior.


289(i). <cmath> requirements missing C float and long double versions

Section: 28.7 [c.math] Status: NAD Submitter: Judy Ward Opened: 2000-12-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [c.math].

View all issues with NAD status.

Discussion:

In ISO/IEC 9899:1990 Programming Languages C we find the following concerning <math.h>:

7.13.4 Mathematics <math.h>
The names of all existing functions declared in the <math.h> header, suffixed with f or l, are reserved respectively for corresponding functions with float and long double arguments are return values.

For example, float sinf(float) is reserved.

In the C99 standard, <math.h> must contain declarations for these functions.

So, is it acceptable for an implementor to add these prototypes to the C++ versions of the math headers? Are they required?

Proposed resolution:

Add these Functions to Table 80, section 26.5 and to Table 99, section C.2:

    acosf asinf atanf atan2f ceilf cosf coshf 
    expf fabsf floorf fmodf frexpf ldexpf 
    logf log10f modff powf sinf sinhf sqrtf 
    tanf tanhf 
    acosl asinl atanl atan2l ceill cosl coshl 
    expl fabsl floorl fmodl frexpl ldexpl 
    logl log10l modfl powl sinl sinhl sqrtl 
    tanl tanhl

There should probably be a note saying that these functions are optional and, if supplied, should match the description in the 1999 version of the C standard. In the next round of C++ standardization they can then become mandatory.

Rationale:

The C90 standard, as amended, already permits (but does not require) these functions, and the C++ standard incorporates the C90 standard by reference. C99 is not an issue, because it is never referred to by the C++ standard.


290(i). Requirements to for_each and its function object

Section: 27.6.5 [alg.foreach] Status: NAD Submitter: Angelika Langer Opened: 2001-01-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.foreach].

View all issues with NAD status.

Discussion:

The specification of the for_each algorithm does not have a "Requires" section, which means that there are no restrictions imposed on the function object whatsoever. In essence it means that I can provide any function object with arbitrary side effects and I can still expect a predictable result. In particular I can expect that the function object is applied exactly last - first times, which is promised in the "Complexity" section.

I don't see how any implementation can give such a guarantee without imposing requirements on the function object.

Just as an example: consider a function object that removes elements from the input sequence. In that case, what does the complexity guarantee (applies f exactly last - first times) mean?

One can argue that this is obviously a nonsensical application and a theoretical case, which unfortunately it isn't. I have seen programmers shooting themselves in the foot this way, and they did not understand that there are restrictions even if the description of the algorithm does not say so.

[Lillehammer: This is more general than for_each. We don't want the function object in transform invalidiating iterators either. There should be a note somewhere in clause 17 (17, not 25) saying that user code operating on a range may not invalidate iterators unless otherwise specified. Bill will provide wording.]

[ 2009-07 Frankfurt ]

Moved to NAD.

It was felt that the current description is adequate, and that there are limits to what the standard can reasonably say to prohibit perverse uses of the library.

Proposed resolution:


293(i). Order of execution in transform algorithm

Section: 27.7.4 [alg.transform] Status: NAD Submitter: Angelika Langer Opened: 2001-01-04 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.transform].

View all issues with NAD status.

Discussion:

This issue is related to issue 242. In case that the resolution proposed for issue 242 is accepted, we have have the following situation: The 4 numeric algorithms (accumulate and consorts) as well as transform would allow a certain category of side effects. The numeric algorithms specify that they invoke the functor "for every iterator i in the range [first, last) in order". transform, in contrast, would not give any guarantee regarding order of invocation of the functor, which means that the functor can be invoked in any arbitrary order.

Why would that be a problem? Consider an example: say the transformator that is a simple enumerator ( or more generally speaking, "is order-sensitive" ). Since a standard compliant implementation of transform is free to invoke the enumerator in no definite order, the result could be a garbled enumeration. Strictly speaking this is not a problem, but it is certainly at odds with the prevalent understanding of transform as an algorithms that assigns "a new _corresponding_ value" to the output elements.

All implementations that I know of invoke the transformator in definite order, namely starting from first and proceeding to last - 1. Unless there is an optimization conceivable that takes advantage of the indefinite order I would suggest to specify the order, because it eliminate the uncertainty that users would otherwise have regarding the order of execution of their potentially order-sensitive function objects.

Proposed resolution:

In section 25.2.3 - Transform [lib.alg.transform] change:

-1- Effects: Assigns through every iterator i in the range [result, result + (last1 - first1)) a new corresponding value equal to op(*(first1 + (i - result)) or binary_op(*(first1 + (i - result), *(first2 + (i - result))).

to:

-1- Effects: Computes values by invoking the operation op or binary_op for every iterator in the range [first1, last1) in order. Assigns through every iterator i in the range [result, result + (last1 - first1)) a new corresponding value equal to op(*(first1 + (i - result)) or binary_op(*(first1 + (i - result), *(first2 + (i - result))).

Rationale:

For Input Iterators an order is already guaranteed, because only one order is possible. If a user who passes a Forward Iterator to one of these algorithms really needs a specific order of execution, it's possible to achieve that effect by wrapping it in an Input Iterator adaptor.


299(i). Incorrect return types for iterator dereference

Section: 25.3.5.6 [bidirectional.iterators], 25.3.5.7 [random.access.iterators] Status: NAD Editorial Submitter: John Potter Opened: 2001-01-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [bidirectional.iterators].

View all issues with NAD Editorial status.

Discussion:

In section 25.3.5.6 [bidirectional.iterators], Table 75 gives the return type of *r-- as convertible to T. This is not consistent with Table 74 which gives the return type of *r++ as T&. *r++ = t is valid while *r-- = t is invalid.

In section 25.3.5.7 [random.access.iterators], Table 76 gives the return type of a[n] as convertible to T. This is not consistent with the semantics of *(a + n) which returns T& by Table 74. *(a + n) = t is valid while a[n] = t is invalid.

Discussion from the Copenhagen meeting: the first part is uncontroversial. The second part, operator[] for Random Access Iterators, requires more thought. There are reasonable arguments on both sides. Return by value from operator[] enables some potentially useful iterators, e.g. a random access "iota iterator" (a.k.a "counting iterator" or "int iterator"). There isn't any obvious way to do this with return-by-reference, since the reference would be to a temporary. On the other hand, reverse_iterator takes an arbitrary Random Access Iterator as template argument, and its operator[] returns by reference. If we decided that the return type in Table 76 was correct, we would have to change reverse_iterator. This change would probably affect user code.

History: the contradiction between reverse_iterator and the Random Access Iterator requirements has been present from an early stage. In both the STL proposal adopted by the committee (N0527==94-0140) and the STL technical report (HPL-95-11 (R.1), by Stepanov and Lee), the Random Access Iterator requirements say that operator[]'s return value is "convertible to T". In N0527 reverse_iterator's operator[] returns by value, but in HPL-95-11 (R.1), and in the STL implementation that HP released to the public, reverse_iterator's operator[] returns by reference. In 1995, the standard was amended to reflect the contents of HPL-95-11 (R.1). The original intent for operator[] is unclear.

In the long term it may be desirable to add more fine-grained iterator requirements, so that access method and traversal strategy can be decoupled. (See "Improved Iterator Categories and Requirements", N1297 = 01-0011, by Jeremy Siek.) Any decisions about issue 299 should keep this possibility in mind.

Further discussion: I propose a compromise between John Potter's resolution, which requires T& as the return type of a[n], and the current wording, which requires convertible to T. The compromise is to keep the convertible to T for the return type of the expression a[n], but to also add a[n] = t as a valid expression. This compromise "saves" the common case uses of random access iterators, while at the same time allowing iterators such as counting iterator and caching file iterators to remain random access iterators (iterators where the lifetime of the object returned by operator*() is tied to the lifetime of the iterator).

Note that the compromise resolution necessitates a change to reverse_iterator. It would need to use a proxy to support a[n] = t.

Note also there is one kind of mutable random access iterator that will no longer meet the new requirements. Currently, iterators that return an r-value from operator[] meet the requirements for a mutable random access iterator, even though the expression a[n] = t will only modify a temporary that goes away. With this proposed resolution, a[n] = t will be required to have the same operational semantics as *(a + n) = t.

[ 2009-07-28 Reopened by Alisdair. No longer solved by concepts. ]

[ 2009-09-18 Alisdair adds: ]

Why can't we write through the reference returned from operator[] on a random access iterator?

Recommended solution:

In table Table 104 — Random access iterator requirements, replace

a[n] : convertible to const T & T& if X is mutable, otherwise convertible to const T&

[ 2009-10 Santa Cruz: ]

Leave Open. Alisdair to spearhead a paper on revivification.

[ 2010 Pittsburgh: Moved to NAD Editorial. Rationale added below. ]

Rationale:

Solved by N3066.

Proposed resolution:

In section 24.1.4 [lib.bidirectdional.iterators], change the return type in table 75 from "convertible to T" to T&.

In section 24.1.5 [lib.random.access.iterators], change the operational semantics for a[n] to " the r-value of a[n] is equivalent to the r-value of *(a + n)". Add a new row in the table for the expression a[n] = t with a return type of convertible to T and operational semantics of *(a + n) = t.

[Lillehammer: Real problem, but should be addressed as part of iterator redesign]

Rationale:

[ San Francisco: ]

Solved by N2758.


302(i). Need error indication from codecvt<>::do_length

Section: 30.4.2.6 [locale.codecvt.byname] Status: NAD Submitter: Gregory Bumgardner Opened: 2001-01-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.codecvt.byname].

View all issues with NAD status.

Discussion:

The effects of codecvt<>::do_length() are described in 22.2.1.5.2, paragraph 10. As implied by that paragraph, and clarified in issue 75, codecvt<>::do_length() must process the source data and update the stateT argument just as if the data had been processed by codecvt<>::in(). However, the standard does not specify how do_length() would report a translation failure, should the source sequence contain untranslatable or illegal character sequences.

The other conversion methods return an "error" result value to indicate that an untranslatable character has been encountered, but do_length() already has a return value (the number of source characters that have been processed by the method).

Proposed resolution:

This issue cannot be resolved without modifying the interface. An exception cannot be used, as there would be no way to determine how many characters have been processed and the state object would be left in an indeterminate state.

A source compatible solution involves adding a fifth argument to length() and do_length() that could be used to return position of the offending character sequence. This argument would have a default value that would allow it to be ignored:

  int length(stateT& state, 
             const externT* from, 
             const externT* from_end, 
             size_t max,
             const externT** from_next = 0);

  virtual
  int do_length(stateT& state, 
                const externT* from, 
                const externT* from_end, 
                size_t max,
                const externT** from_next);

Then an exception could be used to report any translation errors and the from_next argument, if used, could then be used to retrieve the location of the offending character sequence.

Rationale:

The standard is already clear: the return value is the number of "valid complete characters". If it encounters an invalid sequence of external characters, it stops.


304(i). Must *a return an lvalue when a is an input iterator?

Section: 25.3.4 [iterator.concepts] Status: NAD Submitter: Dave Abrahams Opened: 2001-02-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iterator.concepts].

View all issues with NAD status.

Discussion:

We all "know" that input iterators are allowed to produce values when dereferenced of which there is no other in-memory copy.

But: Table 72, with a careful reading, seems to imply that this can only be the case if the value_type has no members (e.g. is a built-in type).

The problem occurs in the following entry:

  a->m     pre: (*a).m is well-defined
           Equivalent to (*a).m

*a.m can be well-defined if *a is not a reference type, but since operator->() must return a pointer for a->m to be well-formed, it needs something to return a pointer to. This seems to indicate that *a must be buffered somewhere to make a legal input iterator.

I don't think this was intentional.

Rationale:

The current standard is clear and consistent. Input iterators that return rvalues are in fact implementable. They may in some cases require extra work, but it is still possible to define an operator-> in such cases: it doesn't have to return a T*, but may return a proxy type. No change to the standard is justified.


309(i). Does sentry catch exceptions?

Section: 31.7 [iostream.format] Status: NAD Submitter: Martin Sebor Opened: 2001-03-19 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iostream.format].

View all issues with NAD status.

Discussion:

The descriptions of the constructors of basic_istream<>::sentry ( [istream::sentry]) and basic_ostream<>::sentry ( [ostream::sentry]) do not explain what the functions do in case an exception is thrown while they execute. Some current implementations allow all exceptions to propagate, others catch them and set ios_base::badbit instead, still others catch some but let others propagate.

The text also mentions that the functions may call setstate(failbit) (without actually saying on what object, but presumably the stream argument is meant). That may have been fine for basic_istream<>::sentry prior to issue 195, since the function performs an input operation which may fail. However, issue 195 amends [istream::sentry], p2 to clarify that the function should actually call setstate(failbit | eofbit), so the sentence in p3 is redundant or even somewhat contradictory.

The same sentence that appears in [ostream::sentry], p3 doesn't seem to be very meaningful for basic_istream<>::sentry which performs no input. It is actually rather misleading since it would appear to guide library implementers to calling setstate(failbit) when os.tie()->flush(), the only called function, throws an exception (typically, it's badbit that's set in response to such an event).

Additional comments from Martin, who isn't comfortable with the current proposed resolution (see c++std-lib-11530)

The istream::sentry ctor says nothing about how the function deals with exemptions (27.6.1.1.2, p1 says that the class is responsible for doing "exception safe"(*) prefix and suffix operations but it doesn't explain what level of exception safety the class promises to provide). The mockup example of a "typical implementation of the sentry ctor" given in 27.6.1.1.2, p6, removed in ISO/IEC 14882:2003, doesn't show exception handling, either. Since the ctor is not classified as a formatted or unformatted input function, the text in 27.6.1.1, p1 through p4 does not apply. All this would seem to suggest that the sentry ctor should not catch or in any way handle exceptions thrown from any functions it may call. Thus, the typical implementation of an istream extractor may look something like [1].

The problem with [1] is that while it correctly sets ios::badbit if an exception is thrown from one of the functions called from the sentry ctor, if the sentry ctor reaches EOF while extracting whitespace from a stream that has eofbit or failbit set in exceptions(), it will cause an ios::failure to be thrown, which will in turn cause the extractor to set ios::badbit.

The only straightforward way to prevent this behavior is to move the definition of the sentry object in the extractor above the try block (as suggested by the example in 22.2.8, p9 and also indirectly supported by 27.6.1.3, p1). See [2]. But such an implementation will allow exceptions thrown from functions called from the ctor to freely propagate to the caller regardless of the setting of ios::badbit in the stream object's exceptions().

So since neither [1] nor [2] behaves as expected, the only possible solution is to have the sentry ctor catch exceptions thrown from called functions, set badbit, and propagate those exceptions if badbit is also set in exceptions(). (Another solution exists that deals with both kinds of sentries, but the code is non-obvious and cumbersome -- see [3].)

Please note that, as the issue points out, current libraries do not behave consistently, suggesting that implementors are not quite clear on the exception handling in istream::sentry, despite the fact that some LWG members might feel otherwise. (As documented by the parenthetical comment here: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2003/n1480.html#309)

Also please note that those LWG members who in Copenhagen felt that "a sentry's constructor should not catch exceptions, because sentries should only be used within (un)formatted input functions and that exception handling is the responsibility of those functions, not of the sentries," as noted here http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2001/n1310.html#309 would in effect be either arguing for the behavior described in [1] or for extractors implemented along the lines of [3].

The original proposed resolution (Revision 25 of the issues list) clarifies the role of the sentry ctor WRT exception handling by making it clear that extractors (both library or user-defined) should be implemented along the lines of [2] (as opposed to [1]) and that no exception thrown from the callees should propagate out of either function unless badbit is also set in exceptions().

[1] Extractor that catches exceptions thrown from sentry:

struct S { long i; };

istream& operator>> (istream &strm, S &s)
{
    ios::iostate err = ios::goodbit;
    try {
        const istream::sentry guard (strm, false);
        if (guard) {
            use_facet<num_get<char> >(strm.getloc ())
                .get (istreambuf_iterator<char>(strm),
                      istreambuf_iterator<char>(),
                      strm, err, s.i);
        }
    }
    catch (...) {
        bool rethrow;
        try {
            strm.setstate (ios::badbit);
            rethrow = false;
        }
        catch (...) {
            rethrow = true;
        }
        if (rethrow)
            throw;
    }
    if (err)
        strm.setstate (err);
    return strm;
}

[2] Extractor that propagates exceptions thrown from sentry:

istream& operator>> (istream &strm, S &s)
{
    istream::sentry guard (strm, false);
    if (guard) {
        ios::iostate err = ios::goodbit;
        try {
            use_facet<num_get<char> >(strm.getloc ())
                .get (istreambuf_iterator<char>(strm),
                      istreambuf_iterator<char>(),
                      strm, err, s.i);
        }
        catch (...) {
            bool rethrow;
            try {
                strm.setstate (ios::badbit);
                rethrow = false;
            }
            catch (...) {
                rethrow = true;
            }
            if (rethrow)
                throw;
        }
        if (err)
            strm.setstate (err);
    }
    return strm;
}

[3] Extractor that catches exceptions thrown from sentry but doesn't set badbit if the exception was thrown as a result of a call to strm.clear().

istream& operator>> (istream &strm, S &s)
{
    const ios::iostate state = strm.rdstate ();
    const ios::iostate except = strm.exceptions ();
    ios::iostate err = std::ios::goodbit;
    bool thrown = true;
    try {
        const istream::sentry guard (strm, false);
        thrown = false;
        if (guard) {
            use_facet<num_get<char> >(strm.getloc ())
                .get (istreambuf_iterator<char>(strm),
                      istreambuf_iterator<char>(),
                      strm, err, s.i);
        }
    }
    catch (...) {
        if (thrown && state & except)
            throw;
        try {
            strm.setstate (ios::badbit);
            thrown = false;
        }
        catch (...) {
            thrown = true;
        }
        if (thrown)
            throw;
    }
    if (err)
        strm.setstate (err);

    return strm;
}

[Pre-Berlin] Reopened at the request of Paolo Carlini and Steve Clamage.

[Pre-Portland] A relevant newsgroup post:

The current proposed resolution of issue #309 is unacceptable. I write commerical software and coding around this makes my code ugly, non-intuitive, and requires comments referring people to this very issue. Following is the full explanation of my experience.

In the course of writing software for commercial use, I constructed std::ifstream's based on user-supplied pathnames on typical POSIX systems.

It was expected that some files that opened successfully might not read successfully -- such as a pathname which actually refered to a directory. Intuitively, I expected the streambuffer underflow() code to throw an exception in this situation, and recent implementations of libstdc++'s basic_filebuf do just that (as well as many of my own custom streambufs).

I also intuitively expected that the istream code would convert these exceptions to the "badbit' set on the stream object, because I had not requested exceptions. I refer to 27.6.1.1. P4.

However, this was not the case on at least two implementations -- if the first thing I did with an istream was call operator>>( T& ) for T among the basic arithmetic types and std::string. Looking further I found that the sentry's constructor was invoking the exception when it pre-scanned for whitespace, and the extractor function (operator>>()) was not catching exceptions in this situation.

So, I was in a situation where setting 'noskipws' would change the istream's behavior even though no characters (whitespace or not) could ever be successfully read.

Also, calling .peek() on the istream before calling the extractor() changed the behavior (.peek() had the effect of setting the badbit ahead of time).

I found this all to be so inconsistent and inconvenient for me and my code design, that I filed a bugzilla entry for libstdc++. I was then told that the bug cannot be fixed until issue #309 is resolved by the committee.

[ 2009-07 Frankfurt ]

Moved to NAD.

See the rationale in the issue. Paolo, who requested that the issue be reopened, agreed with the rationale.

Proposed resolution:

Rationale:

The LWG agrees there is minor variation between implementations, but believes that it doesn't matter. This is a rarely used corner case. There is no evidence that this has any commercial importance or that it causes actual portability problems for customers trying to write code that runs on multiple implementations.


313(i). set_terminate and set_unexpected question

Section: 17.9.5.4 [terminate] Status: NAD Submitter: Judy Ward Opened: 2001-04-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [terminate].

View all issues with NAD status.

Discussion:

According to section 18.7.3.3 of the standard, std::terminate() is supposed to call the terminate_handler in effect immediately after evaluating the throw expression.

Question: what if the terminate_handler in effect is itself std::terminate?

For example:

  #include <exception>

  int main () {
      std::set_terminate(std::terminate);
      throw 5;
      return 0;
  }

Is the implementation allowed to go into an infinite loop?

I think the same issue applies to std::set_unexpected.

Proposed resolution:

Rationale:

Infinite recursion is to be expected: users who set the terminate handler to terminate are explicitly asking for terminate to call itself.


314(i). Is the stack unwound when terminate() is called?

Section: 17.9.5.4 [terminate] Status: NAD Submitter: Detlef Vollmann Opened: 2001-04-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [terminate].

View all issues with NAD status.

Discussion:

The standard appears to contradict itself about whether the stack is unwound when the implementation calls terminate().

From 18.7.3.3p2:

Calls the terminate_handler function in effect immediately after evaluating the throw-expression (lib.terminate.handler), if called by the implementation [...]

So the stack is guaranteed not to be unwound.

But from 15.3p9:

[...]whether or not the stack is unwound before this call to terminate() is implementation-defined (except.terminate).

And 15.5.1 actually defines that in most cases the stack is unwound.

Proposed resolution:

Rationale:

There is definitely no contradiction between the core and library clauses; nothing in the core clauses says that stack unwinding happens after terminate is called. 18.7.3.3p2 does not say anything about when terminate() is called; it merely specifies which terminate_handler is used.


323(i). abs() overloads in different headers

Section: 28.7 [c.math] Status: NAD Submitter: Dave Abrahams Opened: 2001-06-04 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [c.math].

View all issues with NAD status.

Discussion:

Currently the standard mandates the following overloads of abs():

    abs(long), abs(int) in <cstdlib>

    abs(float), abs(double), abs(long double) in <cmath>

    template<class T> T abs(const complex<T>&) in <complex>

    template<class T> valarray<T> abs(const valarray<T>&); in <valarray>

The problem is that having only some overloads visible of a function that works on "implicitly inter-convertible" types is dangerous in practice. The headers that get included at any point in a translation unit can change unpredictably during program development/maintenance. The wrong overload might be unintentionally selected.

Currently, there is nothing that mandates the simultaneous visibility of these overloads. Indeed, some vendors have begun fastidiously reducing dependencies among their (public) headers as a QOI issue: it helps people to write portable code by refusing to compile unless all the correct headers are #included.

The same issue may exist for other functions in the library.

Redmond: PJP reports that C99 adds two new kinds of abs: complex, and int_max_abs.

Related issue: 343.

[ Bellevue: ]

The situation is not sufficiently severe to warrant a change.

Rationale:

The programs that could potentially be broken by this situation are already fragile, and somewhat contrived: For example, a user-defined class that has conversion overloads both to long and to float. If x is a value of such a class, then abs(x) would give the long version if the user included <cstdlib>, the float version if the user included <cmath>, and would be diagnosed as ambiguous at compile time if the user included both headers. The LWG couldn't find an example of a program whose meaning would be changed (as opposed to changing it from well-formed to ill-formed) simply by adding another standard header.

Since the harm seems minimal, and there don't seem to be any simple and noninvasive solutions, this is being closed as NAD. It is marked as "Future" for two reasons. First, it might be useful to define an <all> header that would include all Standard Library headers. Second, we should at least make sure that future library extensions don't make this problem worse.


326(i). Missing typedef in moneypunct_byname

Section: 30.4.7.5 [locale.moneypunct.byname] Status: NAD Submitter: Martin Sebor Opened: 2001-07-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

The definition of the moneypunct facet contains the typedefs char_type and string_type. Only one of these names, string_type, is defined in the derived facet, moneypunct_byname.

Proposed resolution:

For consistency with the numpunct facet, add a typedef for char_type to the definition of the moneypunct_byname facet in 30.4.7.5 [locale.moneypunct.byname].

Rationale:

The absence of the typedef is irrelevant. Users can still access the typedef, because it is inherited from the base class.


330(i). Misleading "exposition only" value in class locale definition

Section: 30.3.1 [locale] Status: NAD Submitter: Martin Sebor Opened: 2001-07-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale].

View all issues with NAD status.

Discussion:

The "exposition only" value of the std::locale::none constant shown in the definition of class locale is misleading in that it on many systems conflicts with the value assigned to one if the LC_XXX constants (specifically, LC_COLLATE on AIX, LC_ALL on HP-UX, LC_CTYPE on Linux and SunOS). This causes incorrect behavior when such a constant is passed to one of the locale member functions that accept a locale::category argument and interpret it as either the C LC_XXX constant or a bitmap of locale::category values. At least three major implementations adopt the suggested value without a change and consequently suffer from this problem.

For instance, the following code will (presumably) incorrectly copy facets belonging to the collate category from the German locale on AIX:

  std::locale l (std::locale ("C"), "de_DE", std::locale::none);

Rationale:

The LWG agrees that it may be difficult to implement locale member functions in such a way that they can take either category arguments or the LC_ constants defined in <cctype>. In light of this requirement (30.3.1.2.1 [locale.category], paragraph 2), and in light of the requirement in the preceding paragraph that it is possible to combine category bitmask elements with bitwise operations, defining the category elements is delicate, particularly if an implementor is constrained to work with a preexisting C library. (Just using the existing LC_ constants would not work in general.) There's no set of "exposition only" values that could give library implementors proper guidance in such a delicate matter. The non-normative example we're giving is no worse than any other choice would be.

See issue 347.


332(i). Consider adding increment and decrement operators to std::fpos< T >

Section: 31.5.3 [fpos] Status: NAD Submitter: PremAnand M. Rao Opened: 2001-08-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [fpos].

View all issues with NAD status.

Discussion:

Increment and decrement operators are missing from Table 88 -- Position type requirements in 31.5.3 [fpos].

Proposed resolution:

Table 88 (section 27.4.3) -- Position type requirements be updated to include increment and decrement operators.

expression        return type     operational    note

++p               fpos&           p += O(1)
p++               fpos            { P tmp = p;
                                    ++p;
                                    return tmp; }
--p               fpos&           p -= O(1)
p--               fpos            { P tmp = p;
                                    --p;
                                    return tmp; }

Rationale:

The LWG believes this is a request for extension, not a defect report. Additionally, nobody saw a clear need for this extension; fpos is used only in very limited ways.


342(i). seek and eofbit

Section: 31.7.5.4 [istream.unformatted] Status: NAD Submitter: Howard Hinnant Opened: 2001-10-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [istream.unformatted].

View all issues with NAD status.

Discussion:

I think we have a defect.

According to lwg issue 60 which is now a dr, the description of seekg in 31.7.5.4 [istream.unformatted] paragraph 38 now looks like:

Behaves as an unformatted input function (as described in 27.6.1.3, paragraph 1), except that it does not count the number of characters extracted and does not affect the value returned by subsequent calls to gcount(). After constructing a sentry object, if fail() != true, executes rdbuf()->pubseekpos( pos).

And according to lwg issue 243 which is also now a dr, 27.6.1.3, paragraph 1 looks like:

Each unformatted input function begins execution by constructing an object of class sentry with the default argument noskipws (second) argument true. If the sentry object returns true, when converted to a value of type bool, the function endeavors to obtain the requested input. Otherwise, if the sentry constructor exits by throwing an exception or if the sentry object returns false, when converted to a value of type bool, the function returns without attempting to obtain any input. In either case the number of extracted characters is set to 0; unformatted input functions taking a character array of non-zero size as an argument shall also store a null character (using charT()) in the first location of the array. If an exception is thrown during input then ios::badbit is turned on in *this'ss error state. If (exception()&badbit)!= 0 then the exception is rethrown. It also counts the number of characters extracted. If no exception has been thrown it ends by storing the count in a member object and returning the value specified. In any event the sentry object is destroyed before leaving the unformatted input function.

And finally 27.6.1.1.2/5 says this about sentry:

If, after any preparation is completed, is.good() is true, ok_ != false otherwise, ok_ == false.

So although the seekg paragraph says that the operation proceeds if !fail(), the behavior of unformatted functions says the operation proceeds only if good(). The two statements are contradictory when only eofbit is set. I don't think the current text is clear which condition should be respected.

Further discussion from Redmond:

PJP: It doesn't seem quite right to say that seekg is "unformatted". That makes specific claims about sentry that aren't quite appropriate for seeking, which has less fragile failure modes than actual input. If we do really mean that it's unformatted input, it should behave the same way as other unformatted input. On the other hand, "principle of least surprise" is that seeking from EOF ought to be OK.

Pre-Berlin: Paolo points out several problems with the proposed resolution in Ready state:

[ 2009-07 Frankfurt ]

Moved to NAD. Will reopen if proposed resolution is supplied.

Proposed resolution:

Change 31.7.5.4 [istream.unformatted] to:

Behaves as an unformatted input function (as described in 27.6.1.3, paragraph 1), except that it does not count the number of characters extracted, does not affect the value returned by subsequent calls to gcount(), and does not examine the value returned by the sentry object. After constructing a sentry object, if fail() != true, executes rdbuf()->pubseekpos(pos). In case of success, the function calls clear(). In case of failure, the function calls setstate(failbit) (which may throw ios_base::failure).

[Lillehammer: Matt provided wording.]

Rationale:

In C, fseek does clear EOF. This is probably what most users would expect. We agree that having eofbit set should not deter a seek, and that a successful seek should clear eofbit. Note that fail() is true only if failbit or badbit is set, so using !fail(), rather than good(), satisfies this goal.


344(i). grouping + showbase

Section: 30.4.3 [category.numeric] Status: NAD Submitter: Howard Hinnant Opened: 2001-10-13 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

When both grouping and showbase are active and the basefield is octal, does the leading 0 participate in the grouping or not? For example, should one format as: 0,123,456 or 0123,456?

An analogy can be drawn with hexadecimal. It appears that 0x123,456 is preferred over 0x,123,456. However, this analogy is not universally accepted to apply to the octal base. The standard is not clear on how to format (or parse) in this manner.

Proposed resolution:

Insert into 30.4.4.1.3 [facet.numpunct.virtuals] paragraph 3, just before the last sentence:

The leading hexadecimal base specifier "0x" does not participate in grouping. The leading '0' octal base specifier may participate in grouping. It is unspecified if the leading '0' participates in formatting octal numbers. In parsing octal numbers, the implementation is encouraged to accept both the leading '0' participating in the grouping, and not participating (e.g. 0123,456 or 0,123,456).

Rationale:

The current behavior may be unspecified, but it's not clear that it matters. This is an obscure corner case, since grouping is usually intended for the benefit of humans and oct/hex prefixes are usually intended for the benefit of machines. There is not a strong enough consensus in the LWG for action.


348(i). Minor issue with std::pair operator<

Section: 22.3 [pairs] Status: Dup Submitter: Andy Sawyer Opened: 2001-10-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [pairs].

View all issues with Dup status.

Duplicate of: 532

Discussion:

The current wording of 20.2.2 [lib.pairs] p6 precludes the use of operator< on any pair type which contains a pointer.

Proposed resolution:

In 22.3 [pairs] paragraph 6, replace:

    Returns: x.first < y.first || (!(y.first < x.first) && x.second <
        y.second).

With:

    Returns: std::less<T1>()( x.first, y.first ) ||
             (!std::less<T1>()( y.first, x.first) && 
             std::less<T2>()( x.second, y.second ) )

Rationale:

This is an instance of a much more general problem. If we want operator< to translate to std::less for pairs of pointers, where do we draw the line? The same issue applies to individual pointers, smart pointer wrappers, std::vector<T*>, and so on.

Andy Koenig suggests that the real issue here is that we aren't distinguishing adequately between two different orderings, a "useful ordering" and a "canonical ordering" that's used just because we sometimes need some ordering without caring much which ordering it is. Another example of the later is typeinfo's before.


350(i). allocator<>::address

Section: 20.2.10.2 [allocator.members], 16.4.4.6 [allocator.requirements], 16.4.2.2 [contents] Status: Dup Submitter: Nathan Myers Opened: 2001-10-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [allocator.members].

View all issues with Dup status.

Duplicate of: 634

Discussion:

See c++std-lib-9006 and c++std-lib-9007. This issue is taken verbatim from -9007.

The core language feature allowing definition of operator&() applied to any non-builtin type makes that operator often unsafe to use in implementing libraries, including the Standard Library. The result is that many library facilities fail for legal user code, such as the fragment

  class A { private: A* operator&(); };
  std::vector<A> aa;

  class B { };
  B* operator&(B&) { return 0; }
  std::vector<B> ba;

In particular, the requirements table for Allocator (Table 32) specifies no semantics at all for member address(), and allocator<>::address is defined in terms of unadorned operator &.

Proposed resolution:

In 20.6.1.1, Change the definition of allocator<>::address from:

Returns: &x

to:

Returns: The value that the built in operator&(x) would return if not overloaded.

In 20.1.6, Table 32, add to the Notes column of the a.address(r) and a.address(s) lines, respectively:

  allocator<T>::address(r)
  allocator<T>::address(s)

In addition, in clause 17.4.1.1, add a statement:

The Standard Library does not apply operator& to any type for which operator& may be overloaded.

Rationale:

The LWG believes both examples are ill-formed. The contained type is required to be CopyConstructible (16.4.4.2 [utility.arg.requirements]), and that includes the requirement that &t return the usual types and values. Since allocators are intended to be used in conjunction with containers, and since the CopyConstructible requirements appear to have been written to deal with the concerns of this issue, the LWG feels it is NAD unless someone can come up with a well-formed example exhibiting a problem.

It may well be that the CopyConstructible requirements are too restrictive and that either the container requirements or the CopyConstructive requirements should be relaxed, but that's a far larger issue. Marking this issue as "future" as a pointer to that larger issue.


351(i). unary_negate and binary_negate: struct or class?

Section: 22.10 [function.objects] Status: NAD Editorial Submitter: Dale Riley Opened: 2001-11-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [function.objects].

View all issues with NAD Editorial status.

Discussion:

In 22.10 [function.objects] the header <functional> synopsis declares the unary_negate and binary_negate function objects as struct. However in [negators] the unary_negate and binary_negate function objects are defined as class. Given the context, they are not "basic function objects" like negate, so this is either a typo or an editorial oversight.

[Taken from comp.std.c++]

Proposed resolution:

Change the synopsis to reflect the useage in [negators]

[Curaçao: Since the language permits "struct", the LWG views this as NAD. They suggest, however, that the Project Editor might wish to make the change as editorial.]


356(i). Meaning of ctype_base::mask enumerators

Section: 30.4.2 [category.ctype] Status: NAD Submitter: Matt Austern Opened: 2002-01-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [category.ctype].

View all issues with NAD status.

Discussion:

What should the following program print?

  #include <locale>
  #include <iostream>

  class my_ctype : public std::ctype<char>
  {
    typedef std::ctype<char> base;
  public:
    my_ctype(std::size_t refs = 0) : base(my_table, false, refs)
    {
      std::copy(base::classic_table(), base::classic_table() + base::table_size,
                my_table);
      my_table[(unsigned char) '_'] = (base::mask) (base::print | base::space);
    }
  private:
    mask my_table[base::table_size];
  };

  int main()
  {
    my_ctype ct;
    std::cout << "isspace: " << ct.is(std::ctype_base::space, '_') << "    "
              << "isalpha: " << ct.is(std::ctype_base::alpha, '_') << std::endl;
  }

The goal is to create a facet where '_' is treated as whitespace.

On gcc 3.0, this program prints "isspace: 1 isalpha: 0". On Microsoft C++ it prints "isspace: 1 isalpha: 1".

I believe that both implementations are legal, and the standard does not give enough guidance for users to be able to use std::ctype's protected interface portably.

The above program assumes that ctype_base::mask enumerators like space and print are disjoint, and that the way to say that a character is both a space and a printing character is to or those two enumerators together. This is suggested by the "exposition only" values in 30.4.2 [category.ctype], but it is nowhere specified in normative text. An alternative interpretation is that the more specific categories subsume the less specific. The above program gives the results it does on the Microsoft compiler because, on that compiler, print has all the bits set for each specific printing character class.

From the point of view of std::ctype's public interface, there's no important difference between these two techniques. From the point of view of the protected interface, there is. If I'm defining a facet that inherits from std::ctype<char>, I'm the one who defines the value that table()['a'] returns. I need to know what combination of mask values I should use. This isn't so very esoteric: it's exactly why std::ctype has a protected interface. If we care about users being able to write their own ctype facets, we have to give them a portable way to do it.

Related reflector messages: lib-9224, lib-9226, lib-9229, lib-9270, lib-9272, lib-9273, lib-9274, lib-9277, lib-9279.

Issue 339 is related, but not identical. The proposed resolution if issue 339 says that ctype_base::mask must be a bitmask type. It does not say that the ctype_base::mask elements are bitmask elements, so it doesn't directly affect this issue.

More comments from Benjamin Kosnik, who believes that that C99 compatibility essentially requires what we're calling option 1 below.

I think the C99 standard is clear, that isspace -> !isalpha.
--------

#include <locale>
#include <iostream>

class my_ctype : public std::ctype<char>
{
private:
  typedef std::ctype<char> base;
  mask my_table[base::table_size];

public:
  my_ctype(std::size_t refs = 0) : base(my_table, false, refs)
  {
    std::copy(base::classic_table(), base::classic_table() + base::table_size,
              my_table);
    mask both = base::print | base::space;
    my_table[static_cast<mask>('_')] = both;
  }
};

int main()
{
  using namespace std;
  my_ctype ct;
  cout << "isspace: " << ct.is(ctype_base::space, '_') << endl;
  cout << "isprint: " << ct.is(ctype_base::print, '_') << endl;

  // ISO C99, isalpha iff upper | lower set, and !space.
  // 7.5, p 193
  // -> looks like g++ behavior is correct.
  // 356 -> bitmask elements are required for ctype_base
  // 339 -> bitmask type required for mask
  cout << "isalpha: " << ct.is(ctype_base::alpha, '_') << endl;
}

Proposed resolution:

Informally, we have three choices:

  1. Require that the enumerators are disjoint (except for alnum and graph)
  2. Require that the enumerators are not disjoint, and specify which of them subsume which others. (e.g. mandate that lower includes alpha and print)
  3. Explicitly leave this unspecified, which the result that the above program is not portable.

Either of the first two options is just as good from the standpoint of portability. Either one will require some implementations to change.

Rationale:

The LWG agrees that this is a real ambiguity, and that both interpretations are conforming under the existing standard. However, there's no evidence that it's causing problems for real users. Users who want to define ctype facets portably can test the ctype_base masks to see which interpretation is being used.


357(i). <cmath> float functions cannot return HUGE_VAL

Section: 28.7 [c.math] Status: NAD Editorial Submitter: Ray Lischner Opened: 2002-02-26 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [c.math].

View all issues with NAD Editorial status.

Discussion:

The float versions of the math functions have no meaningful value to return for a range error. The long double versions have a value they can return, but it isn't necessarily the most reasonable value.

Section 26.5 [lib.c.math], paragraph 5, says that C++ "adds float and long double overloaded versions of these functions, with the same semantics," referring to the math functions from the C90 standard.

The C90 standard, in section 7.5.1, paragraph 3, says that functions return "the value of the macro HUGE_VAL" when they encounter a range error. Section 7.5, paragraph 2, defines HUGE_VAL as a macro that "expands to a positive double expression, not necessarily representable as a float."

Therefore, the float versions of the math functions have no way to signal a range error. [Curaçao: The LWG notes that this isn't strictly correct, since errno is set.] The semantics require that they return HUGE_VAL, but they cannot because HUGE_VAL might not be representable as a float.

The problem with long double functions is less severe because HUGE_VAL is representable as a long double. On the other hand, it might not be a "huge" long double value, and might fall well within the range of normal return values for a long double function. Therefore, it does not make sense for a long double function to return a double (HUGE_VAL) for a range error.

Proposed resolution:

Curaçao: C99 was faced with a similar problem, which they fixed by adding HUGE_VALF and HUGE_VALL in addition to HUGE_VAL.

C++ must also fix, but it should be done in the context of the general C99 based changes to C++, not via DR. Thus the LWG in Curaçao felt the resolution should be NAD, FUTURE, but the issue is being held open for one more meeting to ensure LWG members not present during the discussion concur.

Rationale:

Will be fixed as part of more general work in the TR.


361(i). num_get<>::do_get (..., void*&) checks grouping

Section: 30.4.3.3.3 [facet.num.put.virtuals] Status: NAD Submitter: Martin Sebor Opened: 2002-03-12 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [facet.num.put.virtuals].

View all other issues in [facet.num.put.virtuals].

View all issues with NAD status.

Discussion:

22.2.2.2.2, p12 specifies that thousands_sep is to be inserted only for integral types (issue 282 suggests that this should be done for all arithmetic types).

22.2.2.1.2, p12 requires that grouping be checked for all extractors including that for void*.

I don't think that's right. void* values should not be checked for grouping, should they? (Although if they should, then num_put needs to write them out, otherwise their extraction will fail.)

Proposed resolution:

Change the first sentence of 22.2.2.2.2, p12 from

Digit grouping is checked. That is, the positions of discarded separators is examined for consistency with use_facet<numpunct<charT> >(loc).grouping(). If they are not consistent then ios_base::failbit is assigned to err.

to

Except for conversions to void*, digit grouping is checked...

Rationale:

This would be a change: as it stands, the standard clearly specifies that grouping applies to void*. A survey of existing practice shows that most existing implementations do that, as they should.


366(i). Excessive const-qualification

Section: 31 [input.output] Status: NAD Submitter: Walter Brown, Marc Paterno Opened: 2002-05-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [input.output].

View all issues with NAD status.

Discussion:

The following member functions are declared const, yet return non-const pointers. We believe they are should be changed, because they allow code that may surprise the user. See document N1360 for details and rationale.

[Santa Cruz: the real issue is that we've got const member functions that return pointers to non-const, and N1360 proposes replacing them by overloaded pairs. There isn't a consensus about whether this is a real issue, since we've never said what our constness policy is for iostreams. N1360 relies on a distinction between physical constness and logical constness; that distinction, or those terms, does not appear in the standard.]

Proposed resolution:

In 27.4.4 and 27.4.4.2

Replace

  basic_ostream<charT,traits>* tie() const;

with

  basic_ostream<charT,traits>* tie();
  const basic_ostream<charT,traits>* tie() const;

and replace

  basic_streambuf<charT,traits>* rdbuf() const;

with

  basic_streambuf<charT,traits>* rdbuf();
  const basic_streambuf<charT,traits>* rdbuf() const;

In 27.5.2 and 27.5.2.3.1

Replace

  char_type* eback() const;

with

  char_type* eback();
  const char_type* eback() const;

Replace

  char_type gptr() const;

with

  char_type* gptr();
  const char_type* gptr() const;

Replace

  char_type* egptr() const;

with

  char_type* egptr();
  const char_type* egptr() const;

In 27.5.2 and 27.5.2.3.2

Replace

  char_type* pbase() const;

with

  char_type* pbase();
  const char_type* pbase() const;

Replace

  char_type* pptr() const;

with

  char_type* pptr();
  const char_type* pptr() const;

Replace

  char_type* epptr() const;

with

  char_type* epptr();
  const char_type* epptr() const;

In 27.7.2, 27.7.2.2, 27.7.3 27.7.3.2, 27.7.4, and 27.7.6

Replace

  basic_stringbuf<charT,traits,Allocator>* rdbuf() const;

with

  basic_stringbuf<charT,traits,Allocator>* rdbuf();
  const basic_stringbuf<charT,traits,Allocator>* rdbuf() const;

In 27.8.1.5, 27.8.1.7, 27.8.1.8, 27.8.1.10, 27.8.1.11, and 27.8.1.13

Replace

  basic_filebuf<charT,traits>* rdbuf() const;

with

  basic_filebuf<charT,traits>* rdbuf();
  const basic_filebuf<charT,traits>* rdbuf() const;

Rationale:

The existing specification is a bit sloppy, but there's no particular reason to change this other than tidiness, and there are a number of ways in which streams might have been designed differently if we were starting today. There's no evidence that the existing constness policy is harming users. We might consider a different constness policy as part of a full stream redesign.


367(i). remove_copy/remove_copy_if and Input Iterators

Section: 27.7.8 [alg.remove] Status: NAD Submitter: Anthony Williams Opened: 2002-05-13 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.remove].

View all issues with NAD status.

Discussion:

remove_copy and remove_copy_if (27.7.8 [alg.remove]) permit their input range to be marked with Input Iterators. However, since two operations are required against the elements to copy (comparison and assigment), when the input range uses Input Iterators, a temporary copy must be taken to avoid dereferencing the iterator twice. This therefore requires the value type of the InputIterator to be CopyConstructible. If the iterators are at least Forward Iterators, then the iterator can be dereferenced twice, or a reference to the result maintained, so the temporary is not required.

Proposed resolution:

Add "If InputIterator does not meet the requirements of forward iterator, then the value type of InputIterator must be copy constructible. Otherwise copy constructible is not required." to 27.7.8 [alg.remove] paragraph 6.

Rationale:

The assumption is that an input iterator can't be dereferenced twice. There's no basis for that assumption in the Standard.


368(i). basic_string::replace has two "Throws" paragraphs

Section: 23.4.3.7.6 [string.replace] Status: NAD Editorial Submitter: Beman Dawes Opened: 2002-06-03 Last modified: 2016-11-12

Priority: Not Prioritized

View all other issues in [string.replace].

View all issues with NAD Editorial status.

Discussion:

23.4.3.7.6 [string.replace] basic_string::replace, second signature, given in paragraph 1, has two "Throws" paragraphs (3 and 5).

In addition, the second "Throws" paragraph (5) includes specification (beginning with "Otherwise, the function replaces ...") that should be part of the "Effects" paragraph.

Proposed resolution:

Rationale:

This is editorial. Both "throws" statements are true. The bug is just that the second one should be a sentence, part of the "Effects" clause, not a separate "Throws". The project editor has been notified.


372(i). Inconsistent description of stdlib exceptions

Section: 16.4.6.13 [res.on.exception.handling], 17.7.3 [type.info] Status: NAD Submitter: Randy Maddox Opened: 2002-07-22 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [res.on.exception.handling].

View all other issues in [res.on.exception.handling].

View all issues with NAD status.

Discussion:

Paragraph 3 under clause 16.4.6.13 [res.on.exception.handling], Restrictions on Exception Handling, states that "Any other functions defined in the C++ Standard Library that do not have an exception-specification may throw implementation-defined exceptions unless otherwise specified." This statement is followed by a reference to footnote 178 at the bottom of that page which states, apparently in reference to the C++ Standard Library, that "Library implementations are encouraged (but not required) to report errors by throwing exceptions from (or derived from) the standard exceptions."

These statements appear to be in direct contradiction to clause 17.7.3 [type.info], which states "The class exception defines the base class for the types of objects thrown as exceptions by the C++ Standard library components ...".

Is this inconsistent?

Proposed resolution:

Rationale:

Clause 17 is setting the overall library requirements, and it's clear and consistent. This sentence from Clause 18 is descriptive, not setting a requirement on any other class.


374(i). moneypunct::frac_digits returns int not unsigned

Section: 30.4.7.4.2 [locale.moneypunct.members], 30.4.7.4.3 [locale.moneypunct.virtuals] Status: NAD Submitter: Ray Lischner Opened: 2002-08-08 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

In section 30.4.7.4.2 [locale.moneypunct.members], frac_digits() returns type "int". This implies that frac_digits() might return a negative value, but a negative value is nonsensical. It should return "unsigned".

Similarly, in section 30.4.7.4.3 [locale.moneypunct.virtuals], do_frac_digits() should return "unsigned".

Proposed resolution:

Rationale:

Regardless of whether the return value is int or unsigned, it's always conceivable that frac_digits might return a nonsensical value. (Is 4294967295 really any better than -1?) The clients of moneypunct, the get and put facets, can and do perform range checks.


377(i). basic_string::insert and length_error

Section: 23.4.3.7.4 [string.insert] Status: NAD Submitter: Ray Lischner Opened: 2002-08-16 Last modified: 2016-11-12

Priority: Not Prioritized

View all other issues in [string.insert].

View all issues with NAD status.

Discussion:

Section 23.4.3.7.4 [string.insert], paragraph 4, contains the following, "Then throws length_error if size() >= npos - rlen."

Related to DR 83, this sentence should probably be removed.

Proposed resolution:

Rationale:

This requirement is redundant but correct. No change is needed.


378(i). locale immutability and locale::operator=()

Section: 30.3.1 [locale] Status: Dup Submitter: Martin Sebor Opened: 2002-09-06 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale].

View all issues with Dup status.

Duplicate of: 31

Discussion:

I think there is a problem with 22.1.1, p6 which says that

    -6- An instance of locale is immutable; once a facet reference
        is obtained from it, that reference remains usable as long
        as the locale value itself exists.

and 22.1.1.2, p4:

    const locale& operator=(const locale& other) throw();

    -4- Effects: Creates a copy of other, replacing the current value.

How can a reference to a facet obtained from a locale object remain valid after an assignment that clearly must replace all the facets in the locale object? Imagine a program such as this

    std::locale loc ("de_DE");
    const std::ctype<char> &r0 = std::use_facet<std::ctype<char> >(loc);
    loc = std::locale ("en_US");
    const std::ctype<char> &r1 = std::use_facet<std::ctype<char> >(loc);

Is r0 really supposed to be preserved and destroyed only when loc goes out of scope?

Proposed resolution:

[Summer '04 mid-meeting mailing: Martin and Dietmar believe this is a duplicate of issue 31 and recommend that it be closed. ]


382(i). codecvt do_in/out result

Section: 30.4.2.5 [locale.codecvt] Status: NAD Submitter: Martin Sebor Opened: 2002-08-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.codecvt].

View all issues with NAD status.

Discussion:

It seems that the descriptions of codecvt do_in() and do_out() leave sufficient room for interpretation so that two implementations of codecvt may not work correctly with the same filebuf. Specifically, the following seems less than adequately specified:

  1. the conditions under which the functions terminate
  2. precisely when the functions return ok
  3. precisely when the functions return partial
  4. the full set of conditions when the functions return error
  1. 30.4.2.5.3 [locale.codecvt.virtuals], p2 says this about the effects of the function: ...Stops if it encounters a character it cannot convert... This assumes that there *is* a character to convert. What happens when there is a sequence that doesn't form a valid source character, such as an unassigned or invalid UNICODE character, or a sequence that cannot possibly form a character (e.g., the sequence "\xc0\xff" in UTF-8)?
  2. Table 53 says that the function returns codecvt_base::ok to indicate that the function(s) "completed the conversion." Suppose that the source sequence is "\xc0\x80" in UTF-8, with from pointing to '\xc0' and (from_end==from + 1). It is not clear whether the return value should be ok or partial (see below).
  3. Table 53 says that the function returns codecvt_base::partial if "not all source characters converted." With the from pointers set up the same way as above, it is not clear whether the return value should be partial or ok (see above).
  4. Table 53, in the row describing the meaning of error mistakenly refers to a "from_type" character, without the symbol from_type having been defined. Most likely, the word "source" character is intended, although that is not sufficient. The functions may also fail when they encounter an invalid source sequence that cannot possibly form a valid source character (e.g., as explained in bullet 1 above).

Finally, the conditions described at the end of 30.4.2.5.3 [locale.codecvt.virtuals], p4 don't seem to be possible:

"A return value of partial, if (from_next == from_end), indicates that either the destination sequence has not absorbed all the available destination elements, or that additional source elements are needed before another destination element can be produced."

If the value is partial, it's not clear to me that (from_next ==from_end) could ever hold if there isn't enough room in the destination buffer. In order for (from_next==from_end) to hold, all characters in that range must have been successfully converted (according to 30.4.2.5.3 [locale.codecvt.virtuals], p2) and since there are no further source characters to convert, no more room in the destination buffer can be needed.

It's also not clear to me that (from_next==from_end) could ever hold if additional source elements are needed to produce another destination character (not element as incorrectly stated in the text). partial is returned if "not all source characters have been converted" according to Table 53, which also implies that (from_next==from) does NOT hold.

Could it be that the intended qualifying condition was actually (from_next != from_end), i.e., that the sentence was supposed to read

"A return value of partial, if (from_next != from_end),..."

which would make perfect sense, since, as far as I understand it, partial can only occur if (from_next != from_end)?

[Lillehammer: Defer for the moment, but this really needs to be fixed. Right now, the description of codecvt is too vague for it to be a useful contract between providers and clients of codecvt facets. (Note that both vendors and users can be both providers and clients of codecvt facets.) The major philosophical issue is whether the standard should only describe mappings that take a single wide character to multiple narrow characters (and vice versa), or whether it should describe fully general N-to-M conversions. When the original standard was written only the former was contemplated, but today, in light of the popularity of utf8 and utf16, that doesn't seem sufficient for C++0x. Bill supports general N-to-M conversions; we need to make sure Martin and Howard agree.]

[ 2009-07 Frankfurt ]

codecvt is meant to be a 1-to-N to N-to-1 conversion. It does not work well for N-to-M conversions. wbuffer_convert now exists, and handles N-to-M cases. Also, there is a new specialization of codecvt that permits UTF-16 <-> UTF-8 conversions.

NAD without prejudice. Will reopen if proposed resolution is supplied.

Proposed resolution:


385(i). Does call by value imply the CopyConstructible requirement?

Section: 16 [library] Status: NAD Submitter: Matt Austern Opened: 2002-10-23 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Many function templates have parameters that are passed by value; a typical example is find_if's pred parameter in 27.6.6 [alg.find]. Are the corresponding template parameters (Predicate in this case) implicitly required to be CopyConstructible, or does that need to be spelled out explicitly?

This isn't quite as silly a question as it might seem to be at first sight. If you call find_if in such a way that template argument deduction applies, then of course you'll get call by value and you need to provide a copy constructor. If you explicitly provide the template arguments, however, you can force call by reference by writing something like find_if<my_iterator, my_predicate&>. The question is whether implementation are required to accept this, or whether this is ill-formed because my_predicate& is not CopyConstructible.

The scope of this problem, if it is a problem, is unknown. Function object arguments to generic algorithms in clauses 27 [algorithms] and 28 [numerics] are obvious examples. A review of the whole library is necessary.

[ This is really two issues. First, predicates are typically passed by value but we don't say they must be Copy Constructible. They should be. Second: is specialization allowed to transform value arguments into references? References aren't copy constructible, so this should not be allowed. ]

[ 2007-01-12, Howard: First, despite the note above, references are copy constructible. They just aren't assignable. Second, this is very closely related to 92 and should be consistent with that. That issue already says that implementations are allowed to copy function objects. If one passes in a reference, it is copyable, but susceptible to slicing if one passes in a reference to a base. Third, with rvalue reference in the language one only needs to satisfy MoveConstructible to pass an rvalue "by value". Though the function might still copy the function object internally (requiring CopyConstructible). Finally (and fwiw), if we wanted to, it is easy to code all of the std::algorithms such that they do not copy function objects internally. One merely passes them by reference internally if desired (this has been fully implemented and shipped for several years). If this were mandated, it would reverse 92, allowing function objects to reliably maintain state. E.g. the example in 92 would reliably remove only the third element. ]

Proposed resolution:

Recommend NAD.

Rationale:

Generic algorithms will be marked with concepts and these will imply a requirement of MoveConstructible (not CopyConstructible). The signature of the function will then precisely describe and enforce the precise requirements.


388(i). Use of complex as a key in associative containers

Section: 28.4 [complex.numbers] Status: NAD Submitter: Gabriel Dos Reis Opened: 2002-11-08 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [complex.numbers].

View all issues with NAD status.

Discussion:

Practice with std::complex<> and the associative containers occasionally reveals artificial and distracting issues with constructs resembling: std::set<std::complex<double> > s;

The main reason for the above to fail is the absence of an appropriate definition for std::less<std::complex<T> >. That in turn comes from the definition of the primary template std::less<> in terms of operator<.

The usual argument goes as follows: Since there is no ordering over the complex field compatible with field operations it makes little sense to define a function operator< operating on the datatype std::complex<T>. That is fine. However, that reasoning does not carry over to std::less<T> which is used, among other things, by associative containers as an ordering useful to meet complexity requirements.

Related issue: 348.

[ Pre Bellevue: Reopened at the request of Alisdair. ]

[ Bellevue: ]

This is a request for a design change, and not a defect in the standard. It is in scope to consider, but the group feels that it is not a change that we need to do. Is there a total ordering for floating point values, including NaN? There is not a clear enough solution or big enough problem for us to solve. Solving this problem would require solving the problem for floating point, which is equally unclear. The LWG noted that users who want to put objects into an associative container for which operator< isn't defined can simply provide their own comparison function object. NAD

Proposed resolution:

Informally: Add a specialization of std::less for std::complex.

Rationale:

Discussed in Santa Cruz. An overwhelming majority of the LWG believes this should not be treated a DR: it's a request for a design change, not a defect in the existing standard. Most people (10-3) believed that we probably don't want this change, period: as with issue 348, it's hard to know where to draw the line. The LWG noted that users who want to put objects into an associative container for which operator< isn't defined can simply provide their own comparison function object.


390(i). CopyConstructible requirements too strict

Section: 16.4.4.2 [utility.arg.requirements] Status: NAD Editorial Submitter: Doug Gregor Opened: 2002-10-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [utility.arg.requirements].

View all issues with NAD Editorial status.

Discussion:

The CopyConstructible requirements in Table 30 state that for an object t of type T (where T is CopyConstructible), the expression &t returns the address of t (with type T*). This requirement is overly strict, in that it disallows types that overload operator& to not return a value of type T*. This occurs, for instance, in the Boost.Lambda library, where operator& is overloaded for a Boost.Lambda function object to return another function object.

Example:

  std::vector<int> u, v;
  int x;
  // ...
  std::transform(u.begin(), u.end(), std::back_inserter(v), _1 * x);

_1 * x returns an unnamed function object with operator& overloaded to not return T* , therefore rendering the std::transform call ill-formed. However, most standard library implementations will compile this code properly, and the viability of such binder libraries is severely hindered by the unnecessary restriction in the CopyConstructible requirements.

For reference, the address of an object can be retrieved without using the address-of operator with the following function template:

  template <typename T> T* addressof(T& v)
  {
    return reinterpret_cast<T*>(
         &const_cast<char&>(reinterpret_cast<const volatile char &>(v)));
  }

Note: this relates directly to library issue 350, which will need to be reexamined if the CopyConstructible requirements change.

Proposed resolution:

Remove the last two rows of Table 30, eliminating the requirements that &t and &u return the address of t and u, respectively.

Rationale:

This was a deliberate design decision. Perhaps it should be reconsidered for C++0x.


392(i). 'equivalence' for input iterators

Section: 25.3.5.3 [input.iterators] Status: NAD Submitter: Corwin Joy Opened: 2002-12-11 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [input.iterators].

View all other issues in [input.iterators].

View all issues with NAD status.

Discussion:

In section 25.3.5.3 [input.iterators] table 72 - 'Input Iterator Requirements' we have as a postcondition of *a: "If a==b and (a, b) is in the domain of == then *a is equivalent to *b".

In section [istreambuf.iterator::equal] it states that "istreambuf_iterator::equal returns true if and only if both iterators are at end-of-stream, or neither is at end-of-stream, regardless of what streambuf object they use." (My emphasis).

The defect is that either 'equivalent' needs to be more precisely defined or the conditions for equality in [istreambuf.iterator::equal] are incorrect. (Or both).

Consider the following example:

   #include <iostream>
   #include <fstream>
   #include <iterator>
   using namespace std;

   int main() {
    ifstream file1("file1.txt"), file2("file2.txt");
    istreambuf_iterator<char> f1(file1), f2(file2);
    cout << "f1 == f2 : " << boolalpha << (f1 == f2) << endl;
    cout << "f1 = " << *f1 << endl;
    cout << "f2 = " << *f2 << endl;
    return 0;
   }

Now assuming that neither f1 or f2 are at the end-of-stream then f1 == f2 by [istreambuf.iterator::equal].

However, it is unlikely that *f1 will give the same value as *f2 except by accident.

So what does *f1 'equivalent' to *f2 mean? I think the standard should be clearer on this point, or at least be explicit that this does not mean that *f1 and *f2 are required to have the same value in the case of input iterators.

Proposed resolution:

Rationale:

The two iterators aer not in the domain of ==


393(i). do_in/do_out operation on state unclear

Section: 30.4.2.5.3 [locale.codecvt.virtuals] Status: NAD Editorial Submitter: Alberto Barbati Opened: 2002-12-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.codecvt.virtuals].

View all issues with NAD Editorial status.

Discussion:

this DR follows the discussion on the previous thread "codecvt::do_in not consuming external characters". It's just a clarification issue and not a request for a change.

Can do_in()/do_out() produce output characters without consuming input characters as a result of operation on state?

Proposed resolution:

Add a note at the end of 30.4.2.5.3 [locale.codecvt.virtuals], paragraph 3:

[Note: As a result of operations on state, it can return ok or partial and set from_next == from and to_next != to. --end note]

Rationale:

The submitter believes that standard already provides an affirmative answer to the question. However, the current wording has induced a few library implementors to make the incorrect assumption that do_in()/do_out() always consume at least one internal character when they succeed.

The submitter also believes that the proposed resolution is not in conflict with the related issue 76. Moreover, by explicitly allowing operations on state to produce characters, a codecvt implementation may effectively implement N-to-M translations without violating the "one character at a time" principle described in such issue. On a side note, the footnote in the proposed resolution of issue 76 that informally rules out N-to-M translations for basic_filebuf should be removed if this issue is accepted as valid.

[ Kona (2007): The proposed resolution is to add a note. Since this is non-normative, the issue is editorial, but we believe that the note is correct. Proposed Disposition: NAD, Editorial ]


394(i). behavior of formatted output on failure

Section: 31.7.6.3.1 [ostream.formatted.reqmts] Status: NAD Submitter: Martin Sebor Opened: 2002-12-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

There is a contradiction in Formatted output about what bit is supposed to be set if the formatting fails. On sentence says it's badbit and another that it's failbit.

27.6.2.5.1, p1 says in the Common Requirements on Formatted output functions:

     ... If the generation fails, then the formatted output function
     does setstate(ios::failbit), which might throw an exception.

27.6.2.5.2, p1 goes on to say this about Arithmetic Inserters:

... The formatting conversion occurs as if it performed the following code fragment:

     bool failed =
         use_facet<num_put<charT,ostreambuf_iterator<charT,traits>
         > >
         (getloc()).put(*this, *this, fill(), val). failed();

     ... If failed is true then does setstate(badbit) ...

The original intent of the text, according to Jerry Schwarz (see c++std-lib-10500), is captured in the following paragraph:

In general "badbit" should mean that the stream is unusable because of some underlying failure, such as disk full or socket closure; "failbit" should mean that the requested formatting wasn't possible because of some inconsistency such as negative widths. So typically if you clear badbit and try to output something else you'll fail again, but if you clear failbit and try to output something else you'll succeed.

In the case of the arithmetic inserters, since num_put cannot report failure by any means other than exceptions (in response to which the stream must set badbit, which prevents the kind of recoverable error reporting mentioned above), the only other detectable failure is if the iterator returned from num_put returns true from failed().

Since that can only happen (at least with the required iostream specializations) under such conditions as the underlying failure referred to above (e.g., disk full), setting badbit would seem to be the appropriate response (indeed, it is required in 27.6.2.5.2, p1). It follows that failbit can never be directly set by the arithmetic (it can only be set by the sentry object under some unspecified conditions).

The situation is different for other formatted output functions which can fail as a result of the streambuf functions failing (they may do so by means other than exceptions), and which are then required to set failbit.

The contradiction, then, is that ostream::operator<<(int) will set badbit if the disk is full, while operator<<(ostream&, char) will set failbit under the same conditions. To make the behavior consistent, the Common requirements sections for the Formatted output functions should be changed as proposed below.

[Kona: There's agreement that this is a real issue. What we decided at Kona: 1. An error from the buffer (which can be detected either directly from streambuf's member functions or by examining a streambuf_iterator) should always result in badbit getting set. 2. There should never be a circumstance where failbit gets set. That represents a formatting error, and there are no circumstances under which the output facets are specified as signaling a formatting error. (Even more so for string output that for numeric because there's nothing to format.) If we ever decide to make it possible for formatting errors to exist then the facets can signal the error directly, and that should go in clause 22, not clause 27. 3. The phrase "if generation fails" is unclear and should be eliminated. It's not clear whether it's intended to mean a buffer error (e.g. a full disk), a formatting error, or something else. Most people thought it was supposed to refer to buffer errors; if so, we should say so. Martin will provide wording.]

[ 2009-07 Frankfurt ]

NAD. This issue is already fixed.

Proposed resolution:

Rationale:


397(i). ostream::sentry dtor throws exceptions

Section: 31.7.6.2.4 [ostream.sentry] Status: NAD Editorial Submitter: Martin Sebor Opened: 2003-01-05 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [ostream.sentry].

View all issues with NAD Editorial status.

Discussion:

17.4.4.8, p3 prohibits library dtors from throwing exceptions.

27.6.2.3, p4 says this about the ostream::sentry dtor:

    -4- If ((os.flags() & ios_base::unitbuf) && !uncaught_exception())
        is true, calls os.flush().
    

27.6.2.6, p7 that describes ostream::flush() says:

    -7- If rdbuf() is not a null pointer, calls rdbuf()->pubsync().
        If that function returns ?-1 calls setstate(badbit) (which
        may throw ios_base::failure (27.4.4.3)).
    

That seems like a defect, since both pubsync() and setstate() can throw an exception.

[ The contradiction is real. Clause 17 says destructors may never throw exceptions, and clause 27 specifies a destructor that does throw. In principle we might change either one. We're leaning toward changing clause 17: putting in an "unless otherwise specified" clause, and then putting in a footnote saying the sentry destructor is the only one that can throw. PJP suggests specifying that sentry::~sentry() should internally catch any exceptions it might cause. ]

[ See 418 and 622 for related issues. ]

[ 2009-07 Frankfurt ]

Move to Review. Add "Throws: nothing" to the specification of ostream::sentry::~sentry().

[ 2009-10-13 Daniel adds: ]

The proposed resolution of 835 is written to match the outcome of this issue.

[ 2009 Santa Cruz: ]

Move to Open. Our intent is to solve this issue with 835.

[ 2010-03-06 Martin updates wording. ]

[ 2010 Pittsburgh: ]

Moved to NAD Editorial.

Rationale:

Solved by 835.

Proposed resolution:

Add after [ostream::sentry] p17:

~sentry();

-17- If (os.flags() & ios_base::unitbuf) is true, calls os.flush().

Throws: Nothing.


398(i). effects of end-of-file on unformatted input functions

Section: 31.7.6.2.4 [ostream.sentry] Status: NAD Submitter: Martin Sebor Opened: 2003-01-05 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [ostream.sentry].

View all issues with NAD status.

Discussion:

While reviewing unformatted input member functions of istream for their behavior when they encounter end-of-file during input I found that the requirements vary, sometimes unexpectedly, and in more than one case even contradict established practice (GNU libstdc++ 3.2, IBM VAC++ 6.0, STLPort 4.5, SunPro 5.3, HP aCC 5.38, Rogue Wave libstd 3.1, and Classic Iostreams).

The following unformatted input member functions set eofbit if they encounter an end-of-file (this is the expected behavior, and also the behavior of all major implementations):

    basic_istream<charT, traits>&
    get (char_type*, streamsize, char_type);
    

Also sets failbit if it fails to extract any characters.

    basic_istream<charT, traits>&
    get (char_type*, streamsize);
    

Also sets failbit if it fails to extract any characters.

    basic_istream<charT, traits>&
    getline (char_type*, streamsize, char_type);
    

Also sets failbit if it fails to extract any characters.

    basic_istream<charT, traits>&
    getline (char_type*, streamsize);
    

Also sets failbit if it fails to extract any characters.

    basic_istream<charT, traits>&
    ignore (int, int_type);
    
    basic_istream<charT, traits>&
    read (char_type*, streamsize);
    

Also sets failbit if it encounters end-of-file.

    streamsize readsome (char_type*, streamsize);
    

The following unformated input member functions set failbit but not eofbit if they encounter an end-of-file (I find this odd since the functions make it impossible to distinguish a general failure from a failure due to end-of-file; the requirement is also in conflict with all major implementation which set both eofbit and failbit):

    int_type get();
    
    basic_istream<charT, traits>&
    get (char_type&);
    

These functions only set failbit of they extract no characters, otherwise they don't set any bits, even on failure (I find this inconsistency quite unexpected; the requirement is also in conflict with all major implementations which set eofbit whenever they encounter end-of-file):

    basic_istream<charT, traits>&
    get (basic_streambuf<charT, traits>&, char_type);
    
    basic_istream<charT, traits>&
    get (basic_streambuf<charT, traits>&);
    

This function sets no bits (all implementations except for STLport and Classic Iostreams set eofbit when they encounter end-of-file):

    int_type peek ();
    

Informally, what we want is a global statement of intent saying that eofbit gets set if we trip across EOF, and then we can take away the specific wording for individual functions. A full review is necessary. The wording currently in the standard is a mishmash, and changing it on an individual basis wouldn't make things better. Dietmar will do this work.

[ 2009-07 Frankfurt ]

Moved to NAD. See 31.7.5.2 [istream] p3.

Proposed resolution:


399(i). volations of unformatted input function requirements

Section: 31.7.5.4 [istream.unformatted] Status: NAD Submitter: Martin Sebor Opened: 2003-01-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [istream.unformatted].

View all issues with NAD status.

Discussion:

The Effects clauses for the two functions below violate the general requirements on unformatted input functions outlined in 27.6.1.3: they do not begin by constructing a sentry object. Instead, they begin by calling widen ('\n'), which may throw an exception. The exception is then allowed to propagate from the unformatted input function irrespective of the setting of exceptions().

Note that in light of 27.6.1.1, p3 and p4, the fact that the functions allow exceptions thrown from widen() to propagate may not strictly speaking be a defect (but the fact that the functions do not start by constructing a sentry object still is). However, since an exception thrown from ctype<charT> ::widen() during any other input operation (say, from within a call to num_get<charT>::get()) will be caught and cause badbit to be set, these two functions should not be treated differently for the sake of consistency.

Proposed resolution:

Rationale:

Not a defect. The standard is consistent, and the behavior required by the standard is unambiguous. Yes, it's theoretically possible for widen to throw. (Not that this will happen for the default ctype facet or for most real-world replacement ctype facets.) Users who define ctype facets that can throw, and who care about this behavior, can use alternative signatures that don't call widen.


408(i). Is vector<reverse_iterator<char*> > forbidden?

Section: 25.3 [iterator.requirements] Status: NAD Editorial Submitter: Nathan Myers Opened: 2003-06-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iterator.requirements].

View all issues with NAD Editorial status.

Discussion:

I've been discussing iterator semantics with Dave Abrahams, and a surprise has popped up. I don't think this has been discussed before.

25.3.4 [iterator.concepts] says that the only operation that can be performed on "singular" iterator values is to assign a non-singular value to them. (It doesn't say they can be destroyed, and that's probably a defect.) Some implementations have taken this to imply that there is no need to initialize the data member of a reverse_iterator<> in the default constructor. As a result, code like

  std::vector<std::reverse_iterator<char*> > v(7);
  v.reserve(1000);

invokes undefined behavior, because it must default-initialize the vector elements, and then copy them to other storage. Of course many other vector operations on these adapters are also left undefined, and which those are is not reliably deducible from the standard.

I don't think that 24.1 was meant to make standard-library iterator types unsafe. Rather, it was meant to restrict what operations may be performed by functions which take general user- and standard iterators as arguments, so that raw pointers would qualify as iterators. However, this is not clear in the text, others have come to the opposite conclusion.

One question is whether the standard iterator adaptors have defined copy semantics. Another is whether they have defined destructor semantics: is

  { std::vector<std::reverse_iterator<char*> >  v(7); }

undefined too?

Note this is not a question of whether algorithms are allowed to rely on copy semantics for arbitrary iterators, just whether the types we actually supply support those operations. I believe the resolution must be expressed in terms of the semantics of the adapter's argument type. It should make clear that, e.g., the reverse_iterator<T> constructor is actually required to execute T(), and so copying is defined if the result of T() is copyable.

Issue 235, which defines reverse_iterator's default constructor more precisely, has some relevance to this issue. However, it is not the whole story.

The issue was whether

  reverse_iterator() { }

is allowed, vs.

  reverse_iterator() : current() { }

The difference is when T is char*, where the first leaves the member uninitialized, and possibly equal to an existing pointer value, or (on some targets) may result in a hardware trap when copied.

8.5 paragraph 5 seems to make clear that the second is required to satisfy DR 235, at least for non-class Iterator argument types.

But that only takes care of reverse_iterator, and doesn't establish a policy for all iterators. (The reverse iterator adapter was just an example.) In particular, does my function

  template <typename Iterator>
    void f() { std::vector<Iterator>  v(7); } 

evoke undefined behavior for some conforming iterator definitions? I think it does, now, because vector<> will destroy those singular iterator values, and that's explicitly disallowed.

24.1 shouldn't give blanket permission to copy all singular iterators, because then pointers wouldn't qualify as iterators. However, it should allow copying of that subset of singular iterator values that are default-initialized, and it should explicitly allow destroying any iterator value, singular or not, default-initialized or not.

Related issues: 407, 1012

[ We don't want to require all singular iterators to be copyable, because that is not the case for pointers. However, default construction may be a special case. Issue: is it really default construction we want to talk about, or is it something like value initialization? We need to check with core to see whether default constructed pointers are required to be copyable; if not, it would be wrong to impose so strict a requirement for iterators. ]

[ 2009-05-10 Alisdair provided wording. ]

The comments regarding destroying singular iterators have already been resolved. That just leaves copying (with moving implied).

[ 2009-07 Frankfurt ]

This is related to LWG 1012.

Note that there is a bug in the proposed resolution to LWG 1012. The change to [reverse.iter.con] should be modified so that the word "default" in the second sentence of the Effects clause is replaced by "value."

We believe that the proposed fix to LWG 1012 (now corrected) is sufficient to solve the problem for reverse_iterator. However, Alisdair pointed out that LWG 1012 does not solve the general problem for authors of iterator adaptors.

There are some problems with the proposed resolution. The phrase "safely copyable" is not a term of art. Also, it mentions a DefaultConstructible? concept.

Move to Review after Alisdair updates the wording.

[ 2009-07-31 Alisdair revised wording: ]

[ 2009-08-17 Alisdair and Daniel collaborate on slightly revised wording. This issue depends upon 724 ]

[ 2009-10-14 Daniel adds: ]

There is a clear dependency on 1213, because the term "singular", which is used as part of the resolution, is not properly defined yet.

[ 2009-10 Santa Cruz: ]

Moved to Open. Alisdair will provide improved wording to make this have "value semantics" and otherwise behave like a valid iterator.

[ 2010 Pittsburgh: Moved to NAD Editorial. Rationale added below. ]

Rationale:

Solved by N3066.

Proposed resolution:

Add a new paragrpah to Iterator concepts 25.3 [iterator.requirements] after para 5 (the one describing singular iterators)

Just as a regular pointer to an array guarantees that there is a pointer value pointing past the last element of the array, so for any iterator type there is an iterator value that points past the last element of a corresponding container. These values are called past-the-end values. Values of an iterator i for which the expression *i is defined are called dereferenceable. The library never assumes that past-the-end values are dereferenceable. Iterators can also have singular values that are not associated with any container. [Example: After the declaration of an uninitialized pointer x (as with int* x;), x must always be assumed to have a singular value of a pointer. — end example] Results of most expressions are undefined for singular values; the only exceptions are destroying an iterator that holds a singular value and the assignment of a non-singular value to an iterator that holds a singular value. In this case the singular value is overwritten the same way as any other value. Dereferenceable values are always non-singular.

After value-initialization, any iterator that satisfies the DefaultConstructible requirements ([defaultconstructible]) shall not introduce undefined behaviour when used as the source of a copy or move operation, even if it would otherwise be singular. [Note: This guarantee is not offered for default-initialization (9.4 [dcl.init]), although the distinction only matters for types with trivial default constructors such as pointers. — end note]


417(i). what does ctype::do_widen() return on failure

Section: 30.4.2.2.3 [locale.ctype.virtuals] Status: NAD Submitter: Martin Sebor Opened: 2003-09-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.ctype.virtuals].

View all issues with NAD status.

Discussion:

The Effects and Returns clauses of the do_widen() member function of the ctype facet fail to specify the behavior of the function on failure. That the function may not be able to simply cast the narrow character argument to the type of the result since doing so may yield the wrong value for some wchar_t encodings. Popular implementations of ctype<wchar_t> that use mbtowc() and UTF-8 as the native encoding (e.g., GNU glibc) will fail when the argument's MSB is set. There is no way for the the rest of locale and iostream to reliably detect this failure.

[Kona: This is a real problem. Widening can fail. It's unclear what the solution should be. Returning WEOF works for the wchar_t specialization, but not in general. One option might be to add a default, like narrow. But that's an incompatible change. Using traits::eof might seem like a good idea, but facets don't have access to traits (a recurring problem). We could have widen throw an exception, but that's a scary option; existing library components aren't written with the assumption that widen can throw.]

[ 2009-07 Frankfurt ]

NAD. The behavior is specified for all of the facets that an implementation is required to provide, for the basic character set.

Proposed resolution:


418(i). exceptions thrown during iostream cleanup

Section: 31.5.2.2.6 [ios.init] Status: NAD Submitter: Martin Sebor Opened: 2003-09-18 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [ios.init].

View all issues with NAD status.

Discussion:

The dtor of the ios_base::Init object is supposed to call flush() on the 6 standard iostream objects cout, cerr, clog, wcout, wcerr, and wclog. This call may cause an exception to be thrown.

17.4.4.8, p3 prohibits all library destructors from throwing exceptions.

The question is: What should this dtor do if one or more of these calls to flush() ends up throwing an exception? This can happen quite easily if one of the facets installed in the locale imbued in the iostream object throws.

[Kona: We probably can't do much better than what we've got, so the LWG is leaning toward NAD. At the point where the standard stream objects are being cleaned up, the usual error reporting mechanism are all unavailable. And exception from flush at this point will definitely cause problems. A quality implementation might reasonably swallow the exception, or call abort, or do something even more drastic.]

[ See 397 and 622 for related issues. ]

[ 2009-07 Frankfurt ]

Moved to NAD, no consensus for change.

Proposed resolution:


421(i). is basic_streambuf copy-constructible?

Section: 31.6.3.2 [streambuf.cons] Status: NAD Submitter: Martin Sebor Opened: 2003-09-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [streambuf.cons].

View all issues with NAD status.

Discussion:

The reflector thread starting with c++std-lib-11346 notes that the class template basic_streambuf, along with basic_stringbuf and basic_filebuf, is copy-constructible but that the semantics of the copy constructors are not defined anywhere. Further, different implementations behave differently in this respect: some prevent copy construction of objects of these types by declaring their copy ctors and assignment operators private, others exhibit undefined behavior, while others still give these operations well-defined semantics.

Note that this problem doesn't seem to be isolated to just the three types mentioned above. A number of other types in the library section of the standard provide a compiler-generated copy ctor and assignment operator yet fail to specify their semantics. It's believed that the only types for which this is actually a problem (i.e. types where the compiler-generated default may be inappropriate and may not have been intended) are locale facets. See issue 439.

[ 2009-07 Frankfurt ]

NAD. Option B is already in the Working Draft.

Proposed resolution:

27.5.2 [lib.streambuf]: Add into the synopsis, public section, just above the destructor declaration:

basic_streambuf(const basic_streambuf& sb);
basic_streambuf& operator=(const basic_streambuf& sb);

Insert after 27.5.2.1, paragraph 2:

basic_streambuf(const basic_streambuf& sb);

Constructs a copy of sb.

Postcondtions:

                eback() == sb.eback()
                gptr()  == sb.gptr()
                egptr() == sb.egptr()
                pbase() == sb.pbase()
                pptr()  == sb.pptr()
                epptr() == sb.epptr()
                getloc() == sb.getloc()
basic_streambuf& operator=(const basic_streambuf& sb);

Assigns the data members of sb to this.

Postcondtions:

                eback() == sb.eback()
                gptr()  == sb.gptr()
                egptr() == sb.egptr()
                pbase() == sb.pbase()
                pptr()  == sb.pptr()
                epptr() == sb.epptr()
                getloc() == sb.getloc()

Returns: *this.

27.7.1 [lib.stringbuf]:

Option A:

Insert into the basic_stringbuf synopsis in the private section:

basic_stringbuf(const basic_stringbuf&);             // not defined
basic_stringbuf& operator=(const basic_stringbuf&);  // not defined

Option B:

Insert into the basic_stringbuf synopsis in the public section:

basic_stringbuf(const basic_stringbuf& sb);
basic_stringbuf& operator=(const basic_stringbuf& sb);

27.7.1.1, insert after paragraph 4:

basic_stringbuf(const basic_stringbuf& sb);

Constructs an independent copy of sb as if with sb.str(), and with the openmode that sb was constructed with.

Postcondtions:

               str() == sb.str()
               gptr()  - eback() == sb.gptr()  - sb.eback()
               egptr() - eback() == sb.egptr() - sb.eback()
               pptr()  - pbase() == sb.pptr()  - sb.pbase()
               getloc() == sb.getloc()

Note: The only requirement on epptr() is that it point beyond the initialized range if an output sequence exists. There is no requirement that epptr() - pbase() == sb.epptr() - sb.pbase().

basic_stringbuf& operator=(const basic_stringbuf& sb);

After assignment the basic_stringbuf has the same state as if it were initially copy constructed from sb, except that the basic_stringbuf is allowed to retain any excess capacity it might have, which may in turn effect the value of epptr().

27.8.1.1 [lib.filebuf]

Insert at the bottom of the basic_filebuf synopsis:

private:
  basic_filebuf(const basic_filebuf&);             // not defined
  basic_filebuf& operator=(const basic_filebuf&);  // not defined

[Kona: this is an issue for basic_streambuf itself and for its derived classes. We are leaning toward allowing basic_streambuf to be copyable, and specifying its precise semantics. (Probably the obvious: copying the buffer pointers.) We are less sure whether the streambuf derived classes should be copyable. Howard will write up a proposal.]

[Sydney: Dietmar presented a new argument against basic_streambuf being copyable: it can lead to an encapsulation violation. filebuf inherits from streambuf. Now suppose you inherit a my_hijacking_buf from streambuf. You can copy the streambuf portion of a filebuf to a my_hijacking_buf, giving you access to the pointers into the filebuf's internal buffer. Perhaps not a very strong argument, but it was strong enough to make people nervous. There was weak preference for having streambuf not be copyable. There was weak preference for having stringbuf not be copyable even if streambuf is. Move this issue to open for now. ]

[ 2007-01-12, Howard: Rvalue Reference Recommendations for Chapter 27 recommends protected copy constructor and assignment for basic_streambuf with the same semantics as would be generated by the compiler. These members aid in derived classes implementing move semantics. A protected copy constructor and copy assignment operator do not expose encapsulation more so than it is today as each data member of a basic_streambuf is already both readable and writable by derived classes via various get/set protected member functions (eback(), setp(), etc.). Rather a protected copy constructor and copy assignment operator simply make the job of derived classes implementing move semantics less tedious and error prone. ]

Rationale:

27.5.2 [lib.streambuf]: The proposed basic_streambuf copy constructor and assignment operator are the same as currently implied by the lack of declarations: public and simply copies the data members. This resolution is not a change but a clarification of the current standard.

27.7.1 [lib.stringbuf]: There are two reasonable options: A) Make basic_stringbuf not copyable. This is likely the status-quo of current implementations. B) Reasonable copy semantics of basic_stringbuf can be defined and implemented. A copyable basic_streambuf is arguably more useful than a non-copyable one. This should be considered as new functionality and not the fixing of a defect. If option B is chosen, ramifications from issue 432 are taken into account.

27.8.1.1 [lib.filebuf]: There are no reasonable copy semantics for basic_filebuf.


424(i). normative notes

Section: 16.3.2.2 [structure.summary] Status: NAD Editorial Submitter: Martin Sebor Opened: 2003-09-18 Last modified: 2016-11-12

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

The text in 17.3.1.1, p1 says:
"Paragraphs labelled "Note(s):" or "Example(s):" are informative, other paragraphs are normative."
The library section makes heavy use of paragraphs labeled "Notes(s)," some of which are clearly intended to be normative (see list 1), while some others are not (see list 2). There are also those where the intent is not so clear (see list 3).

List 1 -- Examples of (presumably) normative Notes:
20.2.10.2 [allocator.members], p3,
20.2.10.2 [allocator.members], p10,
23.4.3.3 [string.cons], p11,
30.3.1.3 [locale.cons], p11,
24.3.8.4 [deque.modifiers], p2,
27.8.9 [alg.min.max], p3,
28.4.6 [complex.ops], p15,
31.6.3.5.3 [streambuf.virt.get], p7.

List 2 -- Examples of (presumably) informative Notes:
17.6.3.4 [new.delete.placement], p3,
23.4.3.7.6 [string.replace], p14,
30.4.2.5.3 [locale.codecvt.virtuals], p3,
27.6.5 [alg.foreach], p4,
28.4.5 [complex.member.ops], p1,
31.5.2.6 [ios.base.storage], p6.

List 3 -- Examples of Notes that are not clearly either normative or informative:
30.3.1.3 [locale.cons], p8,
30.3.1.6 [locale.statics], p6,
31.6.3.5.5 [streambuf.virt.put], p4.

None of these lists is meant to be exhaustive.

[Definitely a real problem. The big problem is there's material that doesn't quite fit any of the named paragraph categories (e.g. Effects). Either we need a new kind of named paragraph, or we need to put more material in unnamed paragraphs jsut after the signature. We need to talk to the Project Editor about how to do this. ]

[ Bellevue: Specifics of list 3: First 2 items correct in std (22.1.1.2, 22.1.1.5) Third item should be non-normative (27.5.2.4.5), which Pete will handle editorially. ]

[ post San Francisco: Howard: reopened, needs attention. ]

[Pete: I changed the paragraphs marked "Note" and "Notes" to use "Remark" and "Remarks". Fixed as editorial. This change has been in the WD since the post-Redmond mailing, in 2004. Recommend NAD.]

[ Batavia: We feel that the references in List 2 above should be changed from Remarks to Notes. We also feel that those items in List 3 need to be double checked for the same change. Alan and Pete to review. ]

[ Batavia (2009-05): ]

A spot-check of List 2 suggests the issue is still relevant, and a review of List 3 still seems called-for.

Move to NAD Editorial.

Proposed resolution:


429(i). typo in basic_ios::clear(iostate)

Section: 31.5.4.4 [iostate.flags] Status: Dup Submitter: Martin Sebor Opened: 2003-09-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iostate.flags].

View all issues with Dup status.

Duplicate of: 412

Discussion:

The Effects clause in 27.4.4.3, p5 describing the effects of a call to the ios_base member function clear(iostate state) says that the function only throws if the respective bits are already set prior to the function call. That's obviously not the intent. If it was, a call to clear(badbit) on an object for which (rdstate() == goodbit && exceptions() == badbit) holds would not result in an exception being thrown.

Proposed resolution:

The text ought to be changed from
"If (rdstate() & exceptions()) == 0, returns. ..."
to
"If (state & exceptions()) == 0, returns. ..."

Rationale:


433(i). Contradiction in specification of unexpected()

Section: 99 [unexpected] Status: NAD Submitter: Vyatcheslav Sysoltsev Opened: 2003-09-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Clause [except.unexpected] paragraph 1 says that "void unexpected(); is called (18.7.2) immediately after completing the stack unwinding for the former function", but 18.7.2.4 (Effects) says that "void unexpected(); . . . Calls the unexpected_handler function in effect immediately after evaluating the throwexpression (18.7.2.2),". Isn't here a contradiction: 15.5.2 requires stack have been unwound when in void unexpected() and therefore in unexpected_handler but 18.7.2.4 claims that unexpected_handler is called "in effect immediately" after evaluation of throw expression is finished, so there is no space left for stack to be unwound therefore? I think the phrase "in effect immediately" should be removed from the standard because it brings ambiguity in understanding.

Proposed resolution:

Rationale:

There is no contradiction. The phrase "in effect immediately" is just to clarify which handler is to be called.


437(i). Formatted output of function pointers is confusing

Section: 31.7.6.3.2 [ostream.inserters.arithmetic] Status: NAD Submitter: Ivan Godard Opened: 2003-10-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [ostream.inserters.arithmetic].

View all issues with NAD status.

Discussion:

Given:

void f(int) {}
void(*g)(int) = f;
cout << g;

(with the expected #include and usings), the value printed is a rather surprising "true". Rather useless too.

The standard defines:

ostream& operator<<(ostream&, void*);

which picks up all data pointers and prints their hex value, but does not pick up function pointers because there is no default conversion from function pointer to void*. Absent that, we fall back to legacy conversions from C and the function pointer is converted to bool.

There should be an analogous inserter that prints the address of a function pointer.

Proposed resolution:

Rationale:

This is indeed a wart, but there is no good way to solve it. C doesn't provide a portable way of outputting the address of a function point either.


439(i). Should facets be copyable?

Section: 30.4 [locale.categories] Status: NAD Submitter: Matt Austern Opened: 2003-11-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.categories].

View all issues with NAD status.

Discussion:

The following facets classes have no copy constructors described in the standard, which, according to the standard, means that they are supposed to use the compiler-generated defaults. Default copy behavior is probably inappropriate. We should either make these classes uncopyable or else specify exactly what their constructors do.

Related issue: 421.

        ctype_base
        ctype
        ctype_byname
        ctype<char>
        ctype_byname<char>
        codecvt_base
        codecvt
        codecvt_byname
        num_get
        num_put
        numpunct
        numpunct_byname
        collate
        collate_byname
        time_base
        time_get
        time_get_byname
        time_put
        time_put_byname
        money_get
        money_put
        money_base
        moneypunct
        moneypunct_byname
        messages_base
        messages
        messages_byname

Proposed resolution:

Rationale:

The copy constructor in the base class is private.


440(i). Should std::complex use unqualified transcendentals?

Section: 28.4.8 [complex.transcendentals] Status: NAD Submitter: Matt Austern Opened: 2003-11-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [complex.transcendentals].

View all issues with NAD status.

Discussion:

Operations like pow and exp on complex<T> are typically implemented in terms of operations like sin and cos on T. Should implementations write this as std::sin, or as plain unqualified sin?

The issue, of course, is whether we want to use argument-dependent lookup in the case where T is a user-defined type. This is similar to the issue of valarray transcendentals, as discussed in issue 226.

This issue differs from valarray transcendentals in two important ways. First, "the effect of instantiating the template complex for types other than float, double or long double is unspecified." (28.4.2 [complex.syn]) Second, the standard does not dictate implementation, so there is no guarantee that a particular real math function is used in the implementation of a particular complex function.

Proposed resolution:

Rationale:

If you instantiate std::complex for user-defined types, all bets are off.


446(i). Iterator equality between different containers

Section: 25.3 [iterator.requirements], 24.2 [container.requirements] Status: NAD Editorial Submitter: Andy Koenig Opened: 2003-12-16 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iterator.requirements].

View all issues with NAD Editorial status.

Discussion:

What requirements does the standard place on equality comparisons between iterators that refer to elements of different containers. For example, if v1 and v2 are empty vectors, is v1.end() == v2.end() allowed to yield true? Is it allowed to throw an exception?

The standard appears to be silent on both questions.

[Sydney: The intention is that comparing two iterators from different containers is undefined, but it's not clear if we say that, or even whether it's something we should be saying in clause 23 or in clause 24. Intuitively we might want to say that equality is defined only if one iterator is reachable from another, but figuring out how to say it in any sensible way is a bit tricky: reachability is defined in terms of equality, so we can't also define equality in terms of reachability. ]

[ 2009-07 Frankfurt ]

Daniel volunteered to work on this.

[ 2009-09-20 Daniel provided wording. ]

[ 2009-10 Santa Cruz: ]

Leave as Open. Alisdair has volunteered to refine the wording.

[ 2010 Pittsburgh: Moved to NAD Editorial. Rationale added below. ]

Rationale:

Solved by N3066.

Proposed resolution:

Insert a new paragraph between 25.3 [iterator.requirements]/7+8:

[..] The result of the application of functions in the library to invalid ranges is undefined.

The result of directly or indirectly evaluating any comparison function or the binary - operator with two iterator values as arguments that were obtained from two different ranges r1 and r2 (including their past-the-end values) which are not subranges of one common range is undefined, unless explicitly described otherwise.


447(i). Wrong template argument for time facets

Section: 30.3.1.2.1 [locale.category] Status: Dup Submitter: Pete Becker Opened: 2003-12-26 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.category].

View all issues with Dup status.

Duplicate of: 327

Discussion:

22.1.1.1.1/4, table 52, "Required Instantiations", lists, among others:

    time_get<char,InputIterator>
    time_get_byname<char,InputIterator>
    time_get<wchar_t,OutputIterator>
    time_get_byname<wchar_t,OutputIterator>

The second argument to the last two should be InputIterator, not OutputIterator.

Proposed resolution:

Change the second template argument to InputIterator.

Rationale:


450(i). set::find is inconsistent with associative container requirements

Section: 24.4.6 [set] Status: Dup Submitter: Bill Plauger Opened: 2004-01-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [set].

View all issues with Dup status.

Duplicate of: 214

Discussion:

map/multimap have:

    iterator find(const key_type& x) const;
    const_iterator find(const key_type& x) const;

which is consistent with the table of associative container requirements. But set/multiset have:

    iterator find(const key_type&) const;

set/multiset should look like map/multimap, and honor the requirements table, in this regard.

Proposed resolution:

Rationale:


451(i). Associative erase should return an iterator

Section: 24.2.7 [associative.reqmts], 24.4 [associative] Status: Dup Submitter: Bill Plauger Opened: 2004-01-30 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with Dup status.

Duplicate of: 130

Discussion:

map/multimap/set/multiset have:

    void erase(iterator);
    void erase(iterator, iterator);

But there's no good reason why these can't return an iterator, as for vector/deque/list:

    iterator erase(iterator);
    iterator erase(iterator, iterator);

Proposed resolution:

Informally: The table of associative container requirements, and the relevant template classes, should return an iterator designating the first element beyond the erased subrange.

Rationale:


452(i). locale::combine should be permitted to generate a named locale

Section: 30.3.1.4 [locale.members] Status: NAD Submitter: Bill Plauger Opened: 2004-01-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.members].

View all issues with NAD status.

Discussion:

template<class Facet>
    locale::combine(const locale&) const;

is obliged to create a locale that has no name. This is overspecification and overkill. The resulting locale should follow the usual rules -- it has a name if the locale argument has a name and Facet is one of the standard facets.

[ Sydney and post-Sydney (see c++std-lib-13439, c++std-lib-13440, c++std-lib-13443): agreed that it's overkill to say that the locale is obligated to be nameless. However, we also can't require it to have a name. At the moment, locale names are based on categories and not on individual facets. If a locale contains two different facets of different names from the same category, then this would not fit into existing naming schemes. We need to give implementations more freedom. Bill will provide wording. ]

Rationale:

After further discussion the LWG decided to close this as NAD. The fundamental problem is that names right now are per-category, not per-facet. The combine member function works at the wrong level of granularity.


454(i). basic_filebuf::open should accept wchar_t names

Section: 31.10.2.4 [filebuf.members] Status: NAD Submitter: Bill Plauger Opened: 2004-01-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [filebuf.members].

View all issues with NAD status.

Duplicate of: 105

Discussion:

    basic_filebuf *basic_filebuf::open(const char *, ios_base::open_mode);

should be supplemented with the overload:

    basic_filebuf *basic_filebuf::open(const wchar_t *, ios_base::open_mode);

Depending on the operating system, one of these forms is fundamental and the other requires an implementation-defined mapping to determine the actual filename.

[Sydney: Yes, we want to allow wchar_t filenames. Bill will provide wording.]

[ In Toronto we noted that this is issue 5 from N1569. ]

How does this interact with the newly-defined character types, and how do we avoid interface explosion considering std::string overloads that were added? Propose another solution that is different than the suggestion proposed by PJP.

Suggestion is to make a member template function for basic_string (for char, wchar_t, u16char, u32char instantiations), and then just keep a const char* member.

Goal is to do implicit conversion between character string literals to appropriate basic_string type. Not quite sure if this is possible.

Implementors are free to add specific overloads for non-char character types.

[ Martin adds pre-Sophia Antipolis: ]

Please see issue 454: problems and solutions.

[ Sophia Antipolis: ]

Beman is concerned that making these changes to basic_filebuf is not usefully changed unless fstream is also changed; this also only handles wchar_t and not other character types.

The TR2 filesystem library is a more complete solution, but is not available soon.

[ Martin adds: please reference N2683 for problems and solutions. ]

Proposed resolution:

Change from:

basic_filebuf<charT,traits>* open(
    const char* s,
    ios_base::openmode mode );

Effects: If is_open() != false, returns a null pointer. Otherwise, initializes the filebuf as required. It then opens a file, if possible, whose name is the NTBS s ("as if" by calling std::fopen(s,modstr)).

to:

basic_filebuf<charT,traits>* open(
    const char* s,
    ios_base::openmode mode );

basic_filebuf<charT,traits>* open(
    const wchar_t* ws,
    ios_base::openmode mode );

Effects: If is_open() != false, returns a null pointer. Otherwise, initializes the filebuf as required. It then opens a file, if possible, whose name is the NTBS s ("as if" by calling std::fopen(s,modstr)). For the second signature, the NTBS s is determined from the WCBS ws in an implementation-defined manner.

(NOTE: For a system that "naturally" represents a filename as a WCBS, the NTBS s in the first signature may instead be mapped to a WCBS; if so, it follows the same mapping rules as the first argument to open.)

Rationale:

Slightly controversial, but by a 7-1 straw poll the LWG agreed to move this to Ready. The controversy was because the mapping between wide names and files in a filesystem is implementation defined. The counterargument, which most but not all LWG members accepted, is that the mapping between narrow files names and files is also implemenation defined.

[Lillehammer: Moved back to "open" status, at Beman's urging. (1) Why just basic_filebuf, instead of also basic_fstream (and possibly other things too). (2) Why not also constructors that take std::basic_string? (3) We might want to wait until we see Beman's filesystem library; we might decide that it obviates this.]

[ post Bellevue: ]

Move again to Ready.

There is a timing issue here. Since the filesystem library will not be in C++0x, this should be brought forward. This solution would remain valid in the context of the proposed filesystem.

This issue has been kicking around for a while, and the wchar_t addition alone would help many users. Thus, we suggest putting this on the reflector list with an invitation for someone to produce proposed wording that covers basic_fstream. In the meantime, we suggest that the proposed wording be adopted as-is.

If more of the Lillehammer questions come back, they should be introduced as separate issues.

[ San Francisco: ]

Some existing implementations provide overload already. Expected filesystem "path" object overloads neatly, without surprises; implying NAD.


458(i). 24.1.5 contains unintended limitation for operator-

Section: 25.3.5.7 [random.access.iterators] Status: NAD Submitter: Daniel Frey Opened: 2004-02-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [random.access.iterators].

View all issues with NAD status.

Discussion:

In 24.1.5 [lib.random.access.iterators], table 76 the operational semantics for the expression "r -= n" are defined as "return r += -n". This means, that the expression -n must be valid, which is not the case for unsigned types.

[ Sydney: Possibly not a real problem, since difference type is required to be a signed integer type. However, the wording in the standard may be less clear than we would like. ]

[ Post Summit Alisdair adds: ]

This issue refers to a requirements table we have removed.

The issue might now relate to 25.3.5.7 [random.access.iterators] p5. However, the rationale in the issue already recognises that the difference_type must be signed, so this really looks NAD.

[ Batavia (2009-05): ]

We agree with Alisdair's observations.

Move to NAD.

[ 2009-07 Frankfurt: ]

Need to look at again without concepts.

There was a question about this phrase in the discussion: "the expression -n must be valid, which is not the case for unsigned types." If n is an object of the iterator difference_type (eg ptrdiff_t), then it is never unsigned.

[ 2009-10 Santa Cruz: ]

The group reviewed the wording in the draft and agreed that n is of difference type, the difference type is signed, and the current wording is correct. Moved to NAD.

Proposed resolution:

To remove this limitation, I suggest to change the operational semantics for this column to:

    { Distance m = n;
      if (m >= 0)
        while (m--) --r;
      else
        while (m++) ++r;
      return r; }

459(i). Requirement for widening in stage 2 is overspecification

Section: 30.4.3.2.3 [facet.num.get.virtuals] Status: NAD Submitter: Martin Sebor Opened: 2004-03-16 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [facet.num.get.virtuals].

View all other issues in [facet.num.get.virtuals].

View all issues with NAD status.

Discussion:

When parsing strings of wide-character digits, the standard requires the library to widen narrow-character "atoms" and compare the widened atoms against the characters that are being parsed. Simply narrowing the wide characters would be far simpler, and probably more efficient. The two choices are equivalent except in convoluted test cases, and many implementations already ignore the standard and use narrow instead of widen.

First, I disagree that using narrow() instead of widen() would necessarily have unfortunate performance implications. A possible implementation of narrow() that allows num_get to be implemented in a much simpler and arguably comparably efficient way as calling widen() allows, i.e. without making a virtual call to do_narrow every time, is as follows:

  inline char ctype<wchar_t>::narrow (wchar_t wc, char dflt) const
  {
      const unsigned wi = unsigned (wc);

      if (wi > UCHAR_MAX)
          return typeid (*this) == typeid (ctype<wchar_t>) ?
                 dflt : do_narrow (wc, dflt);

      if (narrow_ [wi] < 0) {
         const char nc = do_narrow (wc, dflt);
         if (nc == dflt)
             return dflt;
         narrow_ [wi] = nc;
      }

      return char (narrow_ [wi]);
  }

Second, I don't think the change proposed in the issue (i.e., to use narrow() instead of widen() during Stage 2) would be at all drastic. Existing implementations with the exception of libstdc++ currently already use narrow() so the impact of the change on programs would presumably be isolated to just a single implementation. Further, since narrow() is not required to translate alternate wide digit representations such as those mentioned in issue 303 to their narrow equivalents (i.e., the portable source characters '0' through '9'), the change does not necessarily imply that these alternate digits would be treated as ordinary digits and accepted as part of numbers during parsing. In fact, the requirement in 30.4.2.2.3 [locale.ctype.virtuals], p13 forbids narrow() to translate an alternate digit character, wc, to an ordinary digit in the basic source character set unless the expression (ctype<charT>::is(ctype_base::digit, wc) == true) holds. This in turn is prohibited by the C standard (7.25.2.1.5, 7.25.2.1.5, and 5.2.1, respectively) for charT of either char or wchar_t.

[Sydney: To a large extent this is a nonproblem. As long as you're only trafficking in char and wchar_t we're only dealing with a stable character set, so you don't really need either 'widen' or 'narrow': can just use literals. Finally, it's not even clear whether widen-vs-narrow is the right question; arguably we should be using codecvt instead.]

[ 2009-07 Frankfurt ]

NAD. The standard is clear enough as written.

Proposed resolution:

Change stage 2 so that implementations are permitted to use either technique to perform the comparison:

  1. call widen on the atoms and compare (either by using operator== or char_traits<charT>::eq) the input with the widened atoms, or
  2. call narrow on the input and compare the narrow input with the atoms
  3. do (1) or (2) only if charT is not char or wchar_t, respectively; i.e., avoid calling widen or narrow if it the source and destination types are the same

462(i). Destroying objects with static storage duration

Section: 6.9.3.4 [basic.start.term], 17.2.2 [cstdlib.syn] Status: NAD Submitter: Bill Plauger Opened: 2004-03-23 Last modified: 2023-02-07

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

3.6.3 Termination spells out in detail the interleaving of static destructor calls and calls to functions registered with atexit. To match this behavior requires intimate cooperation between the code that calls destructors and the exit/atexit machinery. The former is tied tightly to the compiler; the latter is a primitive mechanism inherited from C that traditionally has nothing to do with static construction and destruction. The benefits of intermixing destructor calls with atexit handler calls is questionable at best, and very difficult to get right, particularly when mixing third-party C++ libraries with different third-party C++ compilers and C libraries supplied by still other parties.

I believe the right thing to do is defer all static destruction until after all atexit handlers are called. This is a change in behavior, but one that is likely visible only to perverse test suites. At the very least, we should permit deferred destruction even if we don't require it.

[If this is to be changed, it should probably be changed by CWG. At this point, however, the LWG is leaning toward NAD. Implementing what the standard says is hard work, but it's not impossible and most vendors went through that pain years ago. Changing this behavior would be a user-visible change, and would break at least one real application.]

[ Batavia: Send to core with our recommendation that we should permit deferred destruction but not require it. ]

[ Howard: The course of action recommended in Batavia would undo LWG issue 3 and break current code implementing the "phoenix singleton". Search the net for "phoenix singleton atexit" to get a feel for the size of the adverse impact this change would have. Below is sample code which implements the phoenix singleton and would break if atexit is changed in this way: ]

#include <cstdlib>
#include <iostream>
#include <type_traits>
#include <new>

class A
{
    bool alive_;
    A(const A&);
    A& operator=(const A&);
public:
    A() : alive_(true) {std::cout << "A()\n";}
    ~A() {alive_ = false; std::cout << "~A()\n";}
    void use()
    {
        if (alive_)
            std::cout << "A is alive\n";
        else
            std::cout << "A is dead\n";
    }
};

void deallocate_resource();

// This is the phoenix singleton pattern
A& get_resource(bool create = true)
{
    static std::aligned_storage<sizeof(A), std::alignment_of<A>::value>::type buf;
    static A* a;
    if (create)
    {
        if (a != (A*)&buf)
        {
            a = ::new (&buf) A;
            std::atexit(deallocate_resource);
        }
    }
    else
    {
        a->~A();
        a = (A*)&buf + 1;
    }
    return *a;
}

void deallocate_resource()
{
    get_resource(false);
}

void use_A(const char* message)
{
    A& a = get_resource();
    std::cout << "Using A " << message << "\n";
    a.use();
}

struct B
{
    ~B() {use_A("from ~B()");}
};

B b;

int main()
{
    use_A("from main()");
}

The correct output is:

A()
Using A from main()
A is alive
~A()
A()
Using A from ~B()
A is alive
~A()

[ Bellevue: Confirmed no interaction with quick_exit. Strong feeling against mandating the change. Leaning towards NAD rather than permitting the change, as this would make common implementations of pheonix-singleton pattern implementation defined, as noted by Howard. Bill agrees issue is no longer serious, and accepts NAD. ]

Proposed resolution:


463(i). auto_ptr usability issues

Section: 99 [auto.ptr] Status: NAD Submitter: Rani Sharoni Opened: 2003-12-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [auto.ptr].

View all issues with NAD status.

Discussion:

TC1 CWG DR #84 effectively made the template<class Y> operator auto_ptr<Y>() member of auto_ptr (20.4.5.3/4) obsolete.

The sole purpose of this obsolete conversion member is to enable copy initialization base from r-value derived (or any convertible types like cv-types) case:

#include <memory>
using std::auto_ptr;

struct B {};
struct D : B {};

auto_ptr<D> source();
int sink(auto_ptr<B>);
int x1 = sink( source() ); // #1 EDG - no suitable copy constructor

The excellent analysis of conversion operations that was given in the final auto_ptr proposal (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/1997/N1128.pdf) explicitly specifies this case analysis (case 4). DR #84 makes the analysis wrong and actually comes to forbid the loophole that was exploited by the auto_ptr designers.

I didn't encounter any compliant compiler (e.g. EDG, GCC, BCC and VC) that ever allowed this case. This is probably because it requires 3 user defined conversions and in fact current compilers conform to DR #84.

I was surprised to discover that the obsolete conversion member actually has negative impact of the copy initialization base from l-value derived case:

auto_ptr<D> dp;
int x2 = sink(dp); // #2 EDG - more than one user-defined conversion applies

I'm sure that the original intention was allowing this initialization using the template<class Y> auto_ptr(auto_ptr<Y>& a) constructor (20.4.5.1/4) but since in this copy initialization it's merely user defined conversion (UDC) and the obsolete conversion member is UDC with the same rank (for the early overloading stage) there is an ambiguity between them.

Removing the obsolete member will have impact on code that explicitly invokes it:

int y = sink(source().operator auto_ptr<B>());

IMHO no one ever wrote such awkward code and the reasonable workaround for #1 is:

int y = sink( auto_ptr<B>(source()) );

I was even more surprised to find out that after removing the obsolete conversion member the initialization was still ill-formed: int x3 = sink(dp); // #3 EDG - no suitable copy constructor

This copy initialization semantically requires copy constructor which means that both template conversion constructor and the auto_ptr_ref conversion member (20.4.5.3/3) are required which is what was explicitly forbidden in DR #84. This is a bit amusing case in which removing ambiguity results with no candidates.

I also found exception safety issue with auto_ptr related to auto_ptr_ref:

int f(auto_ptr<B>, std::string);
auto_ptr<B> source2();

// string constructor throws while auto_ptr_ref
// "holds" the pointer
int x4 = f(source2(), "xyz"); // #4

The theoretic execution sequence that will cause a leak:

  1. call auto_ptr<B>::operator auto_ptr_ref<B>()
  2. call string::string(char const*) and throw

According to 20.4.5.3/3 and 20.4.5/2 the auto_ptr_ref conversion member returns auto_ptr_ref<Y> that holds *this and this is another defect since the type of *this is auto_ptr<X> where X might be different from Y. Several library vendors (e.g. SGI) implement auto_ptr_ref<Y> with Y* as member which is much more reasonable. Other vendor implemented auto_ptr_ref as defectively required and it results with awkward and catastrophic code: int oops = sink(auto_ptr<B>(source())); // warning recursive on all control paths

Dave Abrahams noticed that there is no specification saying that auto_ptr_ref copy constructor can't throw.

My proposal comes to solve all the above issues and significantly simplify auto_ptr implementation. One of the fundamental requirements from auto_ptr is that it can be constructed in an intuitive manner (i.e. like ordinary pointers) but with strict ownership semantics which yield that source auto_ptr in initialization must be non-const. My idea is to add additional constructor template with sole propose to generate ill-formed, diagnostic required, instance for const auto_ptr arguments during instantiation of declaration. This special constructor will not be instantiated for other types which is achievable using 14.8.2/2 (SFINAE). Having this constructor in hand makes the constructor template<class Y> auto_ptr(auto_ptr<Y> const&) legitimate since the actual argument can't be const yet non const r-value are acceptable.

This implementation technique makes the "private auxiliary class" auto_ptr_ref obsolete and I found out that modern C++ compilers (e.g. EDG, GCC and VC) consume the new implementation as expected and allow all intuitive initialization and assignment cases while rejecting illegal cases that involve const auto_ptr arguments.

The proposed auto_ptr interface:

namespace std {
    template<class X> class auto_ptr {
    public:
        typedef X element_type;

        // 20.4.5.1 construct/copy/destroy:
        explicit auto_ptr(X* p=0) throw();
        auto_ptr(auto_ptr&) throw();
        template<class Y> auto_ptr(auto_ptr<Y> const&) throw();
        auto_ptr& operator=(auto_ptr&) throw();
        template<class Y> auto_ptr& operator=(auto_ptr<Y>) throw();
        ~auto_ptr() throw();

        // 20.4.5.2 members:
        X& operator*() const throw();
        X* operator->() const throw();
        X* get() const throw();
        X* release() throw();
        void reset(X* p=0) throw();

    private:
        template<class U>
        auto_ptr(U& rhs, typename
unspecified_error_on_const_auto_ptr<U>::type = 0);
    };
}

One compliant technique to implement the unspecified_error_on_const_auto_ptr helper class is using additional private auto_ptr member class template like the following:

template<typename T> struct unspecified_error_on_const_auto_ptr;

template<typename T>
struct unspecified_error_on_const_auto_ptr<auto_ptr<T> const>
{ typedef typename auto_ptr<T>::const_auto_ptr_is_not_allowed type; };

There are other techniques to implement this helper class that might work better for different compliers (i.e. better diagnostics) and therefore I suggest defining its semantic behavior without mandating any specific implementation. IMO, and I didn't found any compiler that thinks otherwise, 14.7.1/5 doesn't theoretically defeat the suggested technique but I suggest verifying this with core language experts.

Further changes in standard text:

Remove section 20.4.5.3

Change 20.4.5/2 to read something like: Initializing auto_ptr<X> from const auto_ptr<Y> will result with unspecified ill-formed declaration that will require unspecified diagnostic.

Change 20.4.5.1/4,5,6 to read:

template<class Y> auto_ptr(auto_ptr<Y> const& a) throw();

4 Requires: Y* can be implicitly converted to X*.

5 Effects: Calls const_cast<auto_ptr<Y>&>(a).release().

6 Postconditions: *this holds the pointer returned from a.release().

Change 20.4.5.1/10

template<class Y> auto_ptr& operator=(auto_ptr<Y> a) throw();

10 Requires: Y* can be implicitly converted to X*. The expression delete get() is well formed.

LWG TC DR #127 is obsolete.

Notice that the copy constructor and copy assignment operator should remain as before and accept non-const auto_ptr& since they have effect on the form of the implicitly declared copy constructor and copy assignment operator of class that contains auto_ptr as member per 12.8/5,10:

struct X {
    // implicit X(X&)
    // implicit X& operator=(X&)
    auto_ptr<D> aptr_;
};

In most cases this indicates about sloppy programming but preserves the current auto_ptr behavior.

Dave Abrahams encouraged me to suggest fallback implementation in case that my suggestion that involves removing of auto_ptr_ref will not be accepted. In this case removing the obsolete conversion member to auto_ptr<Y> and 20.4.5.3/4,5 is still required in order to eliminate ambiguity in legal cases. The two constructors that I suggested will co exist with the current members but will make auto_ptr_ref obsolete in initialization contexts. auto_ptr_ref will be effective in assignment contexts as suggested in DR #127 and I can't see any serious exception safety issues in those cases (although it's possible to synthesize such). auto_ptr_ref<X> semantics will have to be revised to say that it strictly holds pointer of type X and not reference to an auto_ptr for the favor of cases in which auto_ptr_ref<Y> is constructed from auto_ptr<X> in which X is different from Y (i.e. assignment from r-value derived to base).

[Redmond: punt for the moment. We haven't decided yet whether we want to fix auto_ptr for C++-0x, or remove it and replace it with move_ptr and unique_ptr.]

[ Oxford 2007: Recommend NAD. We're just going to deprecate it. It still works for simple use cases and people know how to deal with it. Going forward unique_ptr is the recommended tool. ]

[ 2007-11-09: Reopened at the request of David Abrahams, Alisdair Meredith and Gabriel Dos Reis. ]

[ 2009-07 Frankfurt ]

This is a complicated issue, so we agreed to defer discussion until later in the week so that interested parties can read up on it.

[ 2009-10-04 Daniel adds: ]

I suggest to close this issue as NAD. The reasons are two-fold: First, the suggested proposed resolution uses no longer appropriate language means to solve this issue, which has the effect that the recommended resolution is another - but better - form of hack. Second, either following the suggested resolution or the now more natural alternative via the added member set

template<class Y> auto_ptr(auto_ptr<Y>&&) throw();
template<class Y> auto_ptr& operator=(auto_ptr<Y>&&) throw();

would still have a non-zero probability to break user-code that actively references auto_ptr_ref. This risk seems to indicate that a decision which would not touch the current spec of auto_ptr at all (but deprecating it) and instead recommending to use unique_ptr for new code instead might have the best cost-benefit ratio. IMO the current solution of 1100 can be considered as an active user-support for this transition.

[ 2009-10 Santa Cruz: ]

Mark as NAD. Alisdair will open a new issue (1247) with proposed wording to handle auto_ptr_ref.

Proposed resolution:

Change the synopsis in 99 [auto.ptr]:

namespace std { 
  template <class Y> struct auto_ptr_ref {};

  // exposition only
  template <class T> struct constant_object;

  // exposition only
  template <class T>
  struct cannot_transfer_ownership_from
    : constant_object<T> {};

  template <class X> class auto_ptr { 
  public: 
    typedef X element_type; 

    // D.9.1.1 construct/copy/destroy: 
    explicit auto_ptr(X* p =0) throw(); 
    auto_ptr(auto_ptr&) throw(); 
    template<class Y> auto_ptr(auto_ptr<Y> const&) throw(); 
    auto_ptr& operator=(auto_ptr&) throw(); 
    template<class Y> auto_ptr& operator=(auto_ptr<Y>&) throw();
    auto_ptr& operator=(auto_ptr_ref<X> r) throw();
    ~auto_ptr() throw(); 

    // D.9.1.2 members: 
    X& operator*() const throw();
    X* operator->() const throw();
    X* get() const throw();
    X* release() throw();
    void reset(X* p =0) throw();

    // D.9.1.3 conversions:
    auto_ptr(auto_ptr_ref<X>) throw();
    template<class Y> operator auto_ptr_ref<Y>() throw();
    template<class Y> operator auto_ptr<Y>() throw();

    // exposition only
    template<class U>
    auto_ptr(U& rhs, typename cannot_transfer_ownership_from<U>::error = 0);
  }; 

  template <> class auto_ptr<void> 
  { 
  public: 
    typedef void element_type; 
  }; 

}

Remove 99 [auto.ptr.conv].

Change 99 [auto.ptr], p3:

The auto_ptr provides a semantics of strict ownership. An auto_ptr owns the object it holds a pointer to. Copying an auto_ptr copies the pointer and transfers ownership to the destination. If more than one auto_ptr owns the same object at the same time the behavior of the program is undefined. Templates constant_object and cannot_transfer_ownership_from, and the final constructor of auto_ptr are for exposition only. For any types X and Y, initializing auto_ptr<X> from const auto_ptr<Y> is ill-formed, diagnostic required. [Note: The uses of auto_ptr include providing temporary exception-safety for dynamically allocated memory, passing ownership of dynamically allocated memory to a function, and returning dynamically allocated memory from a function. auto_ptr does not meet the CopyConstructible and Assignable requirements for Standard Library container elements and thus instantiating a Standard Library container with an auto_ptr results in undefined behavior. -- end note]

Change [auto.ptr.cons], p5:

template<class Y> auto_ptr(auto_ptr<Y> const& a) throw();

Requires: Y* can be implicitly converted to X*.

Effects: Calls const_cast<auto_ptr<Y>&>(a).release().

Postconditions: *this holds the pointer returned from a.release().

Change [auto.ptr.cons], p10:

template<class Y> auto_ptr& operator=(auto_ptr<Y>& a) throw();

Requires: Y* can be implicitly converted to X*. The expression delete get() is well formed.

Effects: Calls reset(a.release()).

Returns: *this.


466(i). basic_string ctor should prevent null pointer error

Section: 23.4.3.2 [string.require] Status: NAD Submitter: Daniel Frey Opened: 2004-06-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [string.require].

View all issues with NAD status.

Discussion:

Today, my colleagues and me wasted a lot of time. After some time, I found the problem. It could be reduced to the following short example:

  #include <string>
  int main() { std::string( 0 ); }

The problem is that the tested compilers (GCC 2.95.2, GCC 3.3.1 and Comeau online) compile the above without errors or warnings! The programs (at least for the GCC) resulted in a SEGV.

I know that the standard explicitly states that the ctor of string requires a char* which is not zero. STLs could easily detect the above case with a private ctor for basic_string which takes a single 'int' argument. This would catch the above code at compile time and would not ambiguate any other legal ctors.

[Redmond: No great enthusiasm for doing this. If we do, however, we want to do it for all places that take charT* pointers, not just the single-argument constructor. The other question is whether we want to catch this at compile time (in which case we catch the error of a literal 0, but not an expression whose value is a null pointer), at run time, or both. Recommend NAD. Relegate this functionality to debugging implementations.]

[ Post Summit: Alisdair requests this be re-opened as several new language facilities are designed to solve exactly this kind of problem. ]

[ Batavia (2009-05): ]

We are unable to achieve consensus on an approach to a resolution. There is some sentiment for treating this as a QOI matter. It is also possible that when string is brought into the concepts world, this issue might be addressed in that context.

[ 2009-07 Frankfurt ]

We considered three options:

The consensus was NAD.

Proposed resolution:

Add to the synopsis in 23.4.3 [basic.string]

basic_string( nullptr_t ) = delete;

470(i). accessing containers from their elements' special functions

Section: 24 [containers] Status: NAD Submitter: Martin Sebor Opened: 2004-06-28 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [containers].

View all other issues in [containers].

View all issues with NAD status.

Discussion:

The standard doesn't prohibit the destructors (or any other special functions) of containers' elements invoked from a member function of the container from "recursively" calling the same (or any other) member function on the same container object, potentially while the container is in an intermediate state, or even changing the state of the container object while it is being modified. This may result in some surprising (i.e., undefined) behavior.

Read email thread starting with c++std-lib-13637 for more.

Proposed resolution:

Add to Container Requirements the following new paragraph:

    Unless otherwise specified, the behavior of a program that
    invokes a container member function f from a member function
    g of the container's value_type on a container object c that
    called g from its mutating member function h, is undefined.
    I.e., if v is an element of c, directly or indirectly calling
    c.h() from v.g() called from c.f(), is undefined.

[Redmond: This is a real issue, but it's probably a clause 17 issue, not clause 23. We get the same issue, for example, if we try to destroy a stream from one of the stream's callback functions.]

Rationale:

Recommend NAD. We agree this is an issue, but not a defect. We believe that there is no wording we can put in the standard that will cover all cases without introducing unfortunate corner cases.


472(i). Missing "Returns" clause in std::equal_range

Section: 27.8.4.4 [equal.range] Status: Dup Submitter: Prateek R Karandikar Opened: 2004-06-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [equal.range].

View all issues with Dup status.

Duplicate of: 270

Discussion:

There is no "Returns:" clause for std::equal_range, which returns non-void.

Proposed resolution:

Rationale:

Fixed as part of issue 270.


476(i). Forward Iterator implied mutability

Section: 25.3.5.5 [forward.iterators] Status: NAD Submitter: Dave Abrahams Opened: 2004-07-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [forward.iterators].

View all issues with NAD status.

Discussion:

24.1/3 says:

Forward iterators satisfy all the requirements of the input and output iterators and can be used whenever either kind is specified

The problem is that satisfying the requirements of output iterator means that you can always assign *something* into the result of dereferencing it. That makes almost all non-mutable forward iterators non-conforming. I think we need to sever the refinement relationship between forward iterator and output iterator.

Related issue: 200. But this is not a dup.

Proposed resolution:

Rationale:

Yes, 24.1/3 does say that. But it's introductory material. The precise specification is in 24.1.3, and the requrements table there is right. We don't need to fine-tune introductory wording. (Especially since this wording is likely to be changed as part of the iterator overhaul.)


477(i). Operator-> for const forward iterators

Section: 25.3.5.5 [forward.iterators] Status: Dup Submitter: Dave Abrahams Opened: 2004-07-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [forward.iterators].

View all issues with Dup status.

Duplicate of: 478

Discussion:

The Forward Iterator requirements table contains the following:

 expression  return type         operational  precondition
                                  semantics
  ==========  ==================  ===========  ==========================
  a->m        U& if X is mutable, (*a).m       pre: (*a).m is well-defined.
              otherwise const U&

  r->m        U&                  (*r).m       pre: (*r).m is well-defined.

The first line is exactly right. The second line is wrong. Basically it implies that the const-ness of the iterator affects the const-ness of referenced members. But Paragraph 11 of [lib.iterator.requirements] says:

In the following sections, a and b denote values of type const X, n denotes a value of the difference type Distance, u, tmp, and m denote identifiers, r denotes a value of X&, t denotes a value of value type T, o denotes a value of some type that is writable to the output iterator.

AFAICT if we need the second line at all, it should read the same as the first line.

Related issue: 478

Proposed resolution:

Rationale:

The LWG agrees that this is a real problem. Marked as a DUP because the LWG chose to adopt the solution proposed in 478.


479(i). Container requirements and placement new

Section: 24.2 [container.requirements] Status: Dup Submitter: Herb Sutter Opened: 2004-08-01 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [container.requirements].

View all issues with Dup status.

Duplicate of: 580

Discussion:

Nothing in the standard appears to make this program ill-formed:

  struct C {
    void* operator new( size_t s ) { return ::operator new( s ); }
    // NOTE: this hides in-place and nothrow new
  };

  int main() {
    vector<C> v;
    v.push_back( C() );
  }

Is that intentional? We should clarify whether or not we intended to require containers to support types that define their own special versions of operator new.

[ Lillehammer: A container will definitely never use this overridden operator new, but whether it will fail to compile is unclear from the standard. Are containers supposed to use qualified or unqualified placement new? 20.4.1.1 is somewhat relevant, but the standard doesn't make it completely clear whether containers have to use Allocator::construct(). If containers don't use it, the details of how containers use placement new are unspecified. That is the real bug, but it needs to be fixed as part of the allocator overhaul. Weak support that the eventual solution should make this code well formed. ]

Proposed resolution:


480(i). unary_function and binary_function should have protected nonvirtual destructors

Section: 99 [depr.base] Status: NAD Submitter: Joe Gottman Opened: 2004-08-19 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [depr.base].

View all issues with NAD status.

Discussion:

The classes std::unary_function and std::binary_function are both designed to be inherited from but contain no virtual functions. This makes it too easy for a novice programmer to write code like binary_function<int, int, int> *p = new plus<int>; delete p;

There are two common ways to prevent this source of undefined behavior: give the base class a public virtual destructor, or give it a protected nonvirtual destructor. Since unary_function and binary_function have no other virtual functions, (note in particular the absence of an operator()() ), it would cost too much to give them public virtual destructors. Therefore, they should be given protected nonvirtual destructors.

Proposed resolution:

Change Paragraph 20.3.1 of the Standard from

    template <class Arg, class Result>
    struct unary_function {
        typedef Arg argument_type;
        typedef Result result_type;
    };

    template <class Arg1, class Arg2, class Result>
    struct binary_function {
        typedef Arg1 first_argument_type;
        typedef Arg2 second_argument_type;
        typedef Result result_type;
    };

to

    template <class Arg, class Result>
        struct unary_function {
        typedef Arg argument_type;
        typedef Result result_type;
    protected:
        ~unary_function() {}
    };

    template <class Arg1, class Arg2, class Result>
    struct binary_function {
        typedef Arg1 first_argument_type;
        typedef Arg2 second_argument_type;
        typedef Result result_type;
    protected:
        ~binary_function() {}
    };

Rationale:

The LWG doesn't believe the existing definition causes anybody any concrete harm.


481(i). unique's effects on the range [result, last)

Section: 27.7.9 [alg.unique] Status: NAD Submitter: Andrew Koenig Opened: 2004-08-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.unique].

View all issues with NAD status.

Discussion:

The standard says that unique(first, last) "eliminates all but the first element from every consecutive group of equal elements" in [first, last) and returns "the end of the resulting range". So a postcondition is that [first, result) is the same as the old [first, last) except that duplicates have been eliminated.

What postconditions are there on the range [result, last)? One might argue that the standard says nothing about those values, so they can be anything. One might also argue that the standard doesn't permit those values to be changed, so they must not be. Should the standard say something explicit one way or the other?

Proposed resolution:

Rationale:

We don't want to make many guarantees about what's in [result, end). Maybe we aren't being quite explicit enough about not being explicit, but it's hard to think that's a major problem.


483(i). Heterogeneous equality and EqualityComparable

Section: 27.6 [alg.nonmodifying], 27.7 [alg.modifying.operations] Status: Dup Submitter: Peter Dimov Opened: 2004-09-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with Dup status.

Duplicate of: 283

Discussion:

c++std-lib-14262

[lib.alg.find] requires T to be EqualityComparable:

template <class InputIterator, class T>
   InputIterator find(InputIterator first, InputIterator last,
                      const T& value);

However the condition being tested, as specified in the Effects clause, is actually *i == value, where i is an InputIterator.

The two clauses are in agreement only if the type of *i is T, but this isn't necessarily the case. *i may have a heterogeneous comparison operator that takes a T, or a T may be convertible to the type of *i.

Further discussion (c++std-lib-14264): this problem affects a number of algorithsm in clause 25, not just find. We should try to resolve this problem everywhere it appears.

Proposed resolution:

[lib.alg.find]:

Remove [lib.alg.find]/1.

[lib.alg.count]:

Remove [lib.alg.count]/1.

[lib.alg.search]:

Remove "Type T is EqualityComparable (20.1.1), " from [lib.alg.search]/4.

[lib.alg.replace]:

Remove [lib.alg.replace]/1. Replace [lb.alg.replace]/2 with:

For every iterator i in the range [first, last) for which *i == value or pred(*i) holds perform *i = new_value.

Remove the first sentence of /4. Replace the beginning of /5 with:

For every iterator i in the range [result, result + (last - first)), assign to *i either...

(Note the defect here, current text says assign to i, not *i).

[lib.alg.fill]:

Remove "Type T is Assignable (23.1), " from /1. Replace /2 with:

For every iterator i in the range [first, last) or [first, first + n), perform *i = value.

[lib.alg.remove]:

Remove /1. Remove the first sentence of /6.

Rationale:

Duplicate of (a subset of) issue 283.


486(i). min/max CopyConstructible requirement is too strict

Section: 27.8.9 [alg.min.max] Status: Dup Submitter: Dave Abrahams Opened: 2004-10-13 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.min.max].

View all issues with Dup status.

Duplicate of: 281

Discussion:

A straightforward implementation of these algorithms does not need to copy T.

Proposed resolution:

drop the the words "and CopyConstructible" from paragraphs 1 and 4

Rationale:


487(i). Allocator::construct is too limiting

Section: 16.4.4.6 [allocator.requirements] Status: NAD Submitter: Dhruv Matani Opened: 2004-10-17 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with NAD status.

Discussion:

The standard's version of allocator::construct(pointer, const_reference) severely limits what you can construct using this function. Say you can construct a socket from a file descriptor. Now, using this syntax, I first have to manually construct a socket from the fd, and then pass the constructed socket to the construct() function so it will just to an uninitialized copy of the socket I manually constructed. Now it may not always be possible to copy construct a socket eh! So, I feel that the changes should go in the allocator::construct(), making it:

    template<typename T>
    struct allocator{
      template<typename T1>
      void construct(pointer T1 const& rt1);
    };

Now, the ctor of the class T which matches the one that takes a T1 can be called! Doesn't that sound great?

Proposed resolution:

Rationale:

NAD. STL uses copying all the time, and making it possible for allocators to construct noncopyable objects is useless in the absence of corresponding container changes. We might consider this as part of a larger redesign of STL.


489(i). std::remove / std::remove_if wrongly specified

Section: 27.7.8 [alg.remove] Status: NAD Submitter: Thomas Mang Opened: 2004-12-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.remove].

View all issues with NAD status.

Discussion:

In Section 25.2.7 [lib.alg.remove], paragraphs 1 to 5 describe the behavior of the mutating sequence operations std::remove and std::remove_if. However, the wording does not reflect the intended behavior [Note: See definition of intended behavior below] of these algorithms, as it is known to the C++ community [1].

1) Analysis of current wording:

25.2.7 [lib.alg.remove], paragraph 2:

Current wording says: "Effects: Eliminates all the elements referred to by iterator i in the range [first, last) for which the following corresponding conditions hold: *i == value, pred(*i) != false."

This sentences expresses specifically that all elements denoted by the (original) range [first, last) for which the corresponding condition hold will be eliminated. Since there is no formal definition of the term "eliminate" provided, the meaning of "eliminate" in everyday language implies that as postcondition, no element in the range denoted by [first, last) will hold the corresponding condition on reiteration over the range [first, last).

However, this is neither the intent [Note: See definition of intended behavior below] nor a general possible approach. It can be easily proven that if all elements of the original range[first, last) will hold the condition, it is not possible to substitute them by an element for which the condition will not hold.

25.2.7 [lib.alg.remove], paragraph 3:

Current wording says: "Returns: The end of the resulting range."

The resulting range is not specified. In combination with 25.2.7 [lib.alg.remove], paragraph 2, the only reasonable interpretation of this so-called resulting range is the range [first,last) - thus returning always the ForwardIterator 'last' parameter.

25.2.7 [lib.alg.remove], paragraph 4:

Current wording says: "Notes: Stable: the relative order of the elements that are not removed is the same as their relative order in the original range"

This sentences makes use of the term "removed", which is neither specified, nor used in a previous paragraph (which uses the term "eliminate"), nor unamgiuously separated from the name of the algorithm.

2) Description of intended behavior:

For the rest of this Defect Report, it is assumed that the intended behavior was that all elements of the range [first, last) which do not hold the condition *i == value (std::remove) or pred(*i) != false (std::remove_if)], call them s-elements [Note: s...stay], will be placed into a contiguous subrange of [first, last), denoted by the iterators [first, return value). The number of elements in the resulting range [first, return value) shall be equal to the number of s-elements in the original range [first, last). The relative order of the elements in the resulting subrange[first, return value) shall be the same as the relative order of the corresponding elements in the original range. It is undefined whether any elements in the resulting subrange [return value, last) will hold the corresponding condition, or not.

All implementations known to the author of this Defect Report comply with this intent. Since the intent of the behavior (contrary to the current wording) is also described in various utility references serving the C++ community [1], it is not expected that fixing the paragraphs will influence current code - unless the code relies on the behavior as it is described by current wording and the implementation indeed reflects the current wording, and not the intent.

3) Proposed fixes:

Change 25.2.7 [lib.alg.remove], paragraph 2 to:

"Effect: Places all the elements referred to by iterator i in the range [first, last) for which the following corresponding conditions hold : !(*i == value), pred(*i) == false into the subrange [first, k) of the original range, where k shall denote a value of type ForwardIterator. It is undefined whether any elements in the resulting subrange [k, last) will hold the corresponding condition, or not."

Comments to the new wording:

a) "Places" has no special meaning, and the everyday language meaning should fit. b) The corresponding conditions were negated compared to the current wording, becaue the new wording requires it. c) The wording "of the original range" might be redundant, since any subrange starting at 'first' and containing no more elements than the original range is implicitly a subrange of the original range [first, last). d) The iterator k was introduced instead of "return value" in order to avoid a cyclic dependency on 25.2.7/3. The wording ", where k shall denote a value of type ForwardIterator" might be redundant, because it follows implicitly by 25.2.7/3. e) "Places" does, in the author's opinion, explicitly forbid duplicating any element holding the corresponding condition in the original range [first, last) within the resulting range [first, k). If there is doubt this term might be not unambiguous regarding this, it is suggested that k is specified more closely by the following wording: "k shall denote a value of type ForwardIterator [Note: see d)] so that k - first is equal to the number of elements in the original range [first, last) for which the corresponding condition did hold". This could also be expressed as a separate paragraph "Postcondition:" f) The senctence "It is undefined whether any elements in the resulting subrange [k, last) will hold the corresponding condition, or not." was added consciously so the term "Places" does not imply if the original range [first, last) contains n elements holding the corresponding condition, the identical range[first, last) will also contain exactly n elements holding the corresponding condition after application of the algorithm.

Change 25.2.7 [lib.alg.remove], paragraph 3 to: "Returns: The iterator k."

Change 25.2.7 [lib.alg.remove], paragraph 4 to: "Notes: Stable: the relative order of the elements that are placed into the subrange [first, return value) shall be the same as their relative order was in the original range [first, last) prior to application of the algorithm."

Comments to the new wording:

a) the wording "was ... prior to application of the algorithm" is used to explicitly distinguish the original range not only by means of iterators, but also by a 'chronological' factor from the resulting range [first, return value). It might be redundant.

[1]: The wording of these references is not always unambiguous, and provided examples partially contradict verbal description of the algorithms, because the verbal description resembles the problematic wording of ISO/IEC 14882:2003.

Proposed resolution:

Rationale:

The LWG believes that the standard is sufficiently clear, and that there is no evidence of any real-world confusion about this point.


490(i). std::unique wrongly specified

Section: 27.7.9 [alg.unique] Status: NAD Submitter: Thomas Mang Opened: 2004-12-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.unique].

View all issues with NAD status.

Discussion:

In Section 25.2.8 [lib.alg.unique], paragraphs 1 to 3 describe the behavior of the mutating sequence operation std::unique. However, the wording does not reflect the intended behavior [Note: See definition of intended behavior below] of these algorithms, as it is known to the C++ community [1].

1) Analysis of current wording:

25.2.8 [lib.alg.unique], paragraph 1:

Current wording says: "Effects: Eliminates all but the first element from every consecutive group of equal elements referred to by the iterator i in the range [first, last) for which the following corresponding conditions hold: *i == *(i - 1) or pred(*i, *(i -1)) != false"

This sentences expresses specifically that all elements denoted by the (original) range [first, last) which are not but the first element from a consecutive group of equal elements (where equality is defined as *i == *(i - 1) or pred(*i, *(i - 1)) ! = false) [Note: See DR 202], call them r-elements [Note: r...remove], will be eliminated. Since there is no formal definition of the term "eliminate" provided, it is undefined how this "elimination" takes place. But the meaning of "eliminate" in everyday language seems to disallow explicitly that after application of the algorithm, any r-element will remain at any position of the range [first, last) [2].

Another defect in the current wording concerns the iterators used to compare two elements for equality: The current wording contains the expression "(i - 1)", which is not covered by 25/9 [Note: See DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions].

25.2.8 [lib.alg.unique], paragraph 2:

Current wording says: "Returns: The end of the resulting range."

The resulting range is not specified. In combination with 25.2.8 [lib.alg.unique], paragraph 1, one reasonable interpretation (in the author's opinion even the only possible interpretation) of this so-called resulting range is the range [first, last) - thus returning always the ForwardIterator 'last' parameter.

2) Description of intended behavior:

For the rest of this Defect Report, it is assumed that the intended behavior was that all elements denoted by the original range [first, last) which are the first element from a consecutive group of elements for which the corresponding conditions: *(i-1) == *i (for the version of unique without a predicate argument) or pred(*(i-1), *i) ! = false (for the version of unique with a predicate argument) [Note: If such a group of elements consists of only a single element, this is also considered the first element] [Note: See resolutions of DR 202], call them s-elements [Note: s...stay], will be placed into a contiguous subrange of [first, last), denoted by the iterators [first, return value). The number of elements in the resulting range [first, return value) shall be equal to the number of s-elements in the original range [first, last). Invalid iterator arithmetic expressions are expected to be resolved as proposed in DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions. It is also assumed by the author that the relative order of the elements in the resulting subrange [first, return value) shall be the same as the relative order of the corresponding elements (the s-elements) in the original range [Note: If this was not intended behavior, the additional proposed paragraph about stable order will certainly become obsolete]. Furthermore, the resolutions of DR 202 are partially considered.

All implementations known to the author of this Defect Report comply with this intent [Note: Except possible effects of DR 202]. Since this intent of the behavior (contrary to the current wording) is also described in various utility references serving the C++ community [1], it is not expected that fixing the paragraphs will influence current code [Note: Except possible effects of DR 202] - unless the code relies on the behavior as it is described by current wording and the implementation indeed reflects the current wording, and not the intent.

3) Proposed fixes:

Change 25.2.8 [lib.alg.unique], paragraph 1 to:

"Effect: Places the first element from every consecutive group of elements, referred to by the iterator i in the range [first, last), for which the following conditions hold: *(i-1) == *i (for the version of unique without a predicate argument) or pred(*(i -1), *i) != false (for the version of unique with a predicate argument), into the subrange [first, k) of the original range, where k shall denote a value of type ForwardIterator."

Comments to the new wording:

a) The new wording was influenced by the resolutions of DR 202. If DR 202 is resolved in another way, the proposed wording need also additional review. b) "Places" has no special meaning, and the everyday language meaning should fit. c) The expression "(i - 1)" was left, but is expected that DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions will take this into account. d) The wording "(for the version of unique without a predicate argument)" and "(for the version of unique with a predicate argument)" was added consciously for clarity and is in resemblence with current 23.2.2.4 [lib.list.ops], paragraph 19. It might be considered redundant. e) The wording "of the original range" might be redundant, since any subrange starting at first and containing no more elements than the original range is implicitly a subrange of the original range [first, last). f) The iterator k was introduced instead of "return value" in order to avoid a cyclic dependency on 25.2.8 [lib.alg.unique], paragraph 2. The wording ", where k shall denote a value of type ForwardIterator" might be redundant, because it follows implicitly by 25.2.8 [lib.alg.unique], paragraph 2. g) "Places" does, in the author's opinion, explicitly forbid duplicating any s-element in the original range [first, last) within the resulting range [first, k). If there is doubt this term might be not unambiguous regarding this, it is suggested that k is specified more closely by the following wording: "k shall denote a value of type ForwardIterator [Note: See f)] so that k - first is equal to the number of elements in the original range [first, last) being the first element from every consecutive group of elements for which the corresponding condition did hold". This could also be expressed as a separate paragraph "Postcondition:". h) If it is considered that the wording is unclear whether it declares the element of a group which consists of only a single element implicitly to be the first element of this group [Note: Such an interpretation could eventually arise especially in case last - first == 1] , the following additional sentence is proposed: "If such a group of elements consists of only a single element, this element is also considered the first element."

Change 25.2.8 [lib.alg.unique], paragraph 2 to: "Returns: The iterator k."

Add a separate paragraph "Notes:" as 25.2.8 [lib.alg.unique], paragraph 2a or 3a, or a separate paragraph "Postcondition:" before 25.2.8 [lib.alg.unique], paragraph 2 (wording inside {} shall be eliminated if the preceding expressions are used, or the preceding expressions shall be eliminated if wording inside {} is used):

"Notes:{Postcondition:} Stable: the relative order of the elements that are placed into the subrange [first, return value {k}) shall be the same as their relative order was in the original range [first, last) prior to application of the algorithm."

Comments to the new wording:

a) It is assumed by the author that the algorithm was intended to be stable. In case this was not the intent, this paragraph becomes certainly obsolete. b) The wording "was ... prior to application of the algorithm" is used to explicitly distinguish the original range not only by means of iterators, but also by a 'chronological' factor from the resulting range [first, return value). It might be redundant.

25.2.8 [lib.alg.unique], paragraph 3:

See DR 239.

4) References to other DRs:

See DR 202, but which does not address any of the problems described in this Defect Report [Note: This DR is supposed to complement DR 202]. See DR 239. See DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions.

[1]: The wording of these references is not always unambiguous, and provided examples partially contradict verbal description of the algorithms, because the verbal description resembles the problematic wording of ISO/IEC 14882:2003.

[2]: Illustration of conforming implementations according to current wording:

One way the author of this DR considers how this "elimination" could be achieved by a conforming implementation according to current wording is by substituting each r-element by _any_ s-element [Note: s...stay; any non-r-element], since all r-elements are "eliminated".

In case of a sequence consisting of elements being all 'equal' [Note: See DR 202], substituting each r-element by the single s-element is the only possible solution according to current wording.

Proposed resolution:

Rationale:

The LWG believes the standard is sufficiently clear. No implementers get it wrong, and changing it wouldn't cause any code to change, so there is no real-world harm here.


491(i). std::list<>::unique incorrectly specified

Section: 24.3.10.5 [list.ops] Status: NAD Submitter: Thomas Mang Opened: 2004-12-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [list.ops].

View all issues with NAD status.

Discussion:

In Section 24.3.10.5 [list.ops], paragraphs 19 to 21 describe the behavior of the std::list<T, Allocator>::unique operation. However, the current wording is defective for various reasons.

1) Analysis of current wording:

24.3.10.5 [list.ops], paragraph 19:

Current wording says: "Effects: Eliminates all but the first element from every consecutive group of equal elements referred to by the iterator i in the range [first + 1, last) for which *i == *(i - 1) (for the version of unique with no argument) or pred(*i, *(i -1)) (for the version of unique with a predicate argument) holds."

This sentences makes use of the undefined term "Eliminates". Although it is, to a certain degree, reasonable to consider the term "eliminate" synonymous with "erase", using "Erase" in the first place, as the wording of 24.3.10.5 [list.ops], paragraph 15 does, would be clearer.

The range of the elements referred to by iterator i is "[first + 1, last)". However, neither "first" nor "last" is defined.

The sentence makes three times use of iterator arithmetic expressions ( "first + 1", "*i == *(i - 1)", "pred(*i, *(i -1))" ) which is not defined for bidirectional iterator [see DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions].

The same problems as pointed out in DR 202 (equivalence relation / order of arguments for pred()) apply to this paragraph.

24.3.10.5 [list.ops], paragraph 20:

Current wording says: "Throws: Nothing unless an exception in thrown by *i == *(i-1) or pred(*i, *(i - 1))"

The sentence makes two times use of invalid iterator arithmetic expressions ( "*i == *(i - 1)", "pred(*i, *(i -1))" ).

[Note: Minor typos: "in" / missing dot at end of sentence.]

24.3.10.5 [list.ops], paragraph 21:

Current wording says: "Complexity: If the range (last - first) is not empty, exactly (last - first) - 1 applications of the corresponding predicate, otherwise no application of the predicate.

See DR 315 regarding "(last - first)" not yielding a range.

Invalid iterator arithmetic expression "(last - first) - 1" left .

2) Description of intended behavior:

For the rest of this Defect Report, it is assumed that "eliminate" is supposed to be synonymous to "erase", that "first" is equivalent to an iterator obtained by a call to begin(), "last" is equivalent to an iterator obtained by a call to end(), and that all invalid iterator arithmetic expressions are resolved as described in DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions.

Furthermore, the resolutions of DR 202 are considered regarding equivalence relation and order of arguments for a call to pred.

All implementations known to the author of this Defect Report comply with these assumptions, apart from the impact of the alternative resolution of DR 202. Except for the changes implied by the resolutions of DR 202, no impact on current code is expected.

3) Proposed fixes:

Change 24.3.10.5 [list.ops], paragraph 19 to:

"Effect: Erases all but the first element from every consecutive group of elements, referred to by the iterator i in the range [begin(), end()), for which the following conditions hold: *(i-1) == *i (for the version of unique with no argument) or pred(*(i-1), *i) != false (for the version of unique with a predicate argument)."

Comments to the new wording:

a) The new wording was influenced by DR 202 and the resolutions presented there. If DR 202 is resolved in another way, the proposed wording need also additional review. b) "Erases" refers in the author's opinion unambiguously to the member function "erase". In case there is doubt this might not be unamgibuous, a direct reference to the member function "erase" is suggested [Note: This would also imply a change of 24.3.10.5 [list.ops], paragraph 15.]. c) The expression "(i - 1)" was left, but is expected that DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions will take this into account. d) The wording "(for the version of unique with no argument)" and "(for the version of unique with a predicate argument)" was kept consciously for clarity. e) "begin()" substitutes "first", and "end()" substitutes "last". The range need adjustment from "[first + 1, last)" to "[begin(), end())" to ensure a valid range in case of an empty list. f) If it is considered that the wording is unclear whether it declares the element of a group which consists of only a single element implicitly to be the first element of this group [Note: Such an interpretation could eventually arise especially in case size() == 1] , the following additional sentence is proposed: "If such a group of elements consists of only a single element, this element is also considered the first element."

Change 24.3.10.5 [list.ops], paragraph 20 to:

"Throws: Nothing unless an exception is thrown by *(i-1) == *i or pred(*(i-1), *i)."

Comments to the new wording:

a) The wording regarding the conditions is identical to proposed 24.3.10.5 [list.ops], paragraph 19. If 24.3.10.5 [list.ops], paragraph 19 is resolved in another way, the proposed wording need also additional review. b) The expression "(i - 1)" was left, but is expected that DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions will take this into account. c) Typos fixed.

Change 24.3.10.5 [list.ops], paragraph 21 to:

"Complexity: If empty() == false, exactly size() - 1 applications of the corresponding predicate, otherwise no applications of the corresponding predicate."

Comments to the new wording:

a) The new wording is supposed to also replace the proposed resolution of DR 315, which suffers from the problem of undefined "first" / "last".

5) References to other DRs:

See DR 202. See DR 239. See DR 315. See DR submitted by Thomas Mang regarding invalid iterator arithmetic expressions.

Proposed resolution:

Rationale:

"All implementations known to the author of this Defect Report comply with these assumption", and "no impact on current code is expected", i.e. there is no evidence of real-world confusion or harm.


492(i). Invalid iterator arithmetic expressions

Section: 16.3.2.4 [structure.specifications] Status: NAD Submitter: Thomas Mang Opened: 2004-12-12 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [structure.specifications].

View all other issues in [structure.specifications].

View all issues with NAD status.

Discussion:

Various clauses other than clause 25 make use of iterator arithmetic not supported by the iterator category in question. Algorithms in clause 25 are exceptional because of 25 [lib.algorithms], paragraph 9, but this paragraph does not provide semantics to the expression "iterator - n", where n denotes a value of a distance type between iterators.

1) Examples of current wording:

Current wording outside clause 25:

23.2.2.4 [lib.list.ops], paragraphs 19-21: "first + 1", "(i - 1)", "(last - first)" 23.3.1.1 [lib.map.cons], paragraph 4: "last - first" 23.3.2.1 [lib.multimap.cons], paragraph 4: "last - first" 23.3.3.1 [lib.set.cons], paragraph 4: "last - first" 23.3.4.1 [lib.multiset.cons], paragraph 4: "last - first" 24.4.1 [lib.reverse.iterators], paragraph 1: "(i - 1)"

[Important note: The list is not complete, just an illustration. The same issue might well apply to other paragraphs not listed here.]

None of these expressions is valid for the corresponding iterator category.

Current wording in clause 25:

25.1.1 [lib.alg.foreach], paragraph 1: "last - 1" 25.1.3 [lib.alg.find.end], paragraph 2: "[first1, last1 - (last2-first2))" 25.2.8 [lib.alg.unique], paragraph 1: "(i - 1)" 25.2.8 [lib.alg.unique], paragraph 5: "(i - 1)"

However, current wording of 25 [lib.algorithms], paragraph 9 covers neither of these four cases:

Current wording of 25 [lib.algorithms], paragraph 9:

"In the description of the algorithms operator + and - are used for some of the iterator categories for which they do not have to be defined. In these cases the semantics of a+n is the same as that of

{X tmp = a;
advance(tmp, n);
return tmp;
}

and that of b-a is the same as of return distance(a, b)"

This paragrpah does not take the expression "iterator - n" into account, where n denotes a value of a distance type between two iterators [Note: According to current wording, the expression "iterator - n" would be resolved as equivalent to "return distance(n, iterator)"]. Even if the expression "iterator - n" were to be reinterpreted as equivalent to "iterator + -n" [Note: This would imply that "a" and "b" were interpreted implicitly as values of iterator types, and "n" as value of a distance type], then 24.3.4/2 interfers because it says: "Requires: n may be negative only for random access and bidirectional iterators.", and none of the paragraphs quoted above requires the iterators on which the algorithms operate to be of random access or bidirectional category.

2) Description of intended behavior:

For the rest of this Defect Report, it is assumed that the expression "iterator1 + n" and "iterator1 - iterator2" has the semantics as described in current 25 [lib.algorithms], paragraph 9, but applying to all clauses. The expression "iterator1 - n" is equivalent to an result-iterator for which the expression "result-iterator + n" yields an iterator denoting the same position as iterator1 does. The terms "iterator1", "iterator2" and "result-iterator" shall denote the value of an iterator type, and the term "n" shall denote a value of a distance type between two iterators.

All implementations known to the author of this Defect Report comply with these assumptions. No impact on current code is expected.

3) Proposed fixes:

Change 25 [lib.algorithms], paragraph 9 to:

"In the description of the algorithms operator + and - are used for some of the iterator categories for which they do not have to be defined. In this paragraph, a and b denote values of an iterator type, and n denotes a value of a distance type between two iterators. In these cases the semantics of a+n is the same as that of

{X tmp = a;
advance(tmp, n);
return tmp;
}

,the semantics of a-n denotes the value of an iterator i for which the following condition holds: advance(i, n) == a, and that of b-a is the same as of return distance(a, b)".

Comments to the new wording:

a) The wording " In this paragraph, a and b denote values of an iterator type, and n denotes a value of a distance type between two iterators." was added so the expressions "b-a" and "a-n" are distinguished regarding the types of the values on which they operate. b) The wording ",the semantics of a-n denotes the value of an iterator i for which the following condition holds: advance(i, n) == a" was added to cover the expression 'iterator - n'. The wording "advance(i, n) == a" was used to avoid a dependency on the semantics of a+n, as the wording "i + n == a" would have implied. However, such a dependency might well be deserved. c) DR 225 is not considered in the new wording.

Proposed fixes regarding invalid iterator arithmetic expressions outside clause 25:

Either a) Move modified 25 [lib.algorithms], paragraph 9 (as proposed above) before any current invalid iterator arithmetic expression. In that case, the first sentence of 25 [lib.algorithms], paragraph 9, need also to be modified and could read: "For the rest of this International Standard, ...." / "In the description of the following clauses including this ...." / "In the description of the text below ..." etc. - anyways substituting the wording "algorithms", which is a straight reference to clause 25. In that case, 25 [lib.algorithms] paragraph 9 will certainly become obsolete. Alternatively, b) Add an appropiate paragraph similar to resolved 25 [lib.algorithms], paragraph 9, to the beginning of each clause containing invalid iterator arithmetic expressions. Alternatively, c) Fix each paragraph (both current wording and possible resolutions of DRs) containing invalid iterator arithmetic expressions separately.

5) References to other DRs:

See DR 225. See DR 237. The resolution could then also read "Linear in last - first".

[ Bellevue: ]

Keep open and ask Bill to provide wording.

[ 2009-05-09 Alisdair adds: ]

This issue is related to 997.

[ 2009-07 Frankfurt ]

Hinnant: this isn't going to change any user's code or any vendor's implementation.

No objection to "NAD without prejudice." If anyone proposes a resolution, the LWG will consider it.

Move to NAD.

Proposed resolution:

[Lillehammer: Minor issue, but real. We have a blanket statement about this in 25/11. But (a) it should be in 17, not 25; and (b) it's not quite broad enough, because there are some arithmetic expressions it doesn't cover. Bill will provide wording.]


493(i). Undefined Expression in Input Iterator Note Title

Section: 25.3.5.3 [input.iterators] Status: NAD Submitter: Chris Jefferson Opened: 2004-12-13 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [input.iterators].

View all other issues in [input.iterators].

View all issues with NAD status.

Discussion:

1) In 24.1.1/3, the following text is currently present.

"Note: For input iterators, a==b does not imply ++a=++b (Equality does not guarantee the substitution property or referential transparency)."

However, when in Table 72, part of the definition of ++r is given as:

"pre: r is dereferenceable. post: any copies of the previous value of r are no longer required either to be dereferenceable ..."

While a==b does not imply that b is a copy of a, this statement should perhaps still be made more clear.

2) There are no changes to intended behaviour

3) This Note should be altered to say "Note: For input iterators a==b, when its behaviour is defined ++a==++b may still be false (Equality does not guarantee the substitution property or referential transparency).

Proposed resolution:

Rationale:

This is descriptive text, not normative, and the meaning is clear.


494(i). Wrong runtime complexity for associative container's insert and delete

Section: 24.2.7 [associative.reqmts] Status: NAD Submitter: Hans B os Opened: 2004-12-19 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with NAD status.

Discussion:

According to [lib.associative.reqmts] table 69, the runtime comlexity of insert(p, t) and erase(q) can be done in amortized constant time.

It was my understanding that an associative container could be implemented as a balanced binary tree.

For inser(p, t), you 'll have to iterate to p's next node to see if t can be placed next to p. Furthermore, the insertion usually takes place at leaf nodes. An insert next to the root node will be done at the left of the root next node

So when p is the root node you 'll have to iterate from the root to its next node, which takes O(log(size)) time in a balanced tree.

If you insert all values with insert(root, t) (where root is the root of the tree before insertion) then each insert takes O(log(size)) time. The amortized complexity per insertion will be O(log(size)) also.

For erase(q), the normal algorithm for deleting a node that has no empty left or right subtree, is to iterate to the next (or previous), which is a leaf node. Then exchange the node with the next and delete the leaf node. Furthermore according to DR 130, erase should return the next node of the node erased. Thus erasing the root node, requires iterating to the next node.

Now if you empty a map by deleting the root node until the map is empty, each operation will take O(log(size)), and the amortized complexity is still O(log(size)).

The operations can be done in amortized constant time if iterating to the next node can be done in (non amortized) constant time. This can be done by putting all nodes in a double linked list. This requires two extra links per node. To me this is a bit overkill since you can already efficiently insert or erase ranges with erase(first, last) and insert(first, last).

Proposed resolution:

Rationale:

Only "amortized constant" in special circumstances, and we believe that's implementable. That is: doing this N times will be O(N), not O(log N).


499(i). Std. doesn't seem to require stable_sort() to be stable!

Section: 27.8.2.2 [stable.sort] Status: NAD Editorial Submitter: Prateek Karandikar Opened: 2005-04-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

17.3.1.1 Summary

1 The Summary provides a synopsis of the category, and introduces the first-level subclauses. Each subclause also provides a summary, listing the headers specified in the subclause and the library entities provided in each header.

2 Paragraphs labelled "Note(s):" or "Example(s):" are informative, other paragraphs are normative.

So this means that a "Notes" paragraph wouldn't be normative.

25.3.1.2 stable_sort

template<class RandomAccessIterator> 
void stable_sort(RandomAccessIterat or first, RandomAccessIterator last); 

template<class RandomAccessIterator, class Compare> 
void stable_sort(RandomAccessIterat or first, RandomAccessIterator last, Compare comp);

1 Effects: Sorts the elements in the range [first, last).

2 Complexity: It does at most N(log N)^2 (where N == last - first) comparisons; if enough extra memory is available, it is N log N.

3 Notes: Stable: the relative order of the equivalent elements is preserved.

The Notes para is informative, and nowhere else is stability mentioned above.

Also, I just searched for the word "stable" in my copy of the Standard. and the phrase "Notes: Stable: the relative order of the elements..." is repeated several times in the Standard library clauses for describing various functions. How is it that stability is talked about in the informative paragraph? Or am I missing something obvious?

Proposed resolution:

Rationale:

This change has already been made.


500(i). do_length cannot be implemented correctly

Section: 30.4.2.6 [locale.codecvt.byname] Status: NAD Submitter: Krzysztof Żelechowski Opened: 2005-05-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.codecvt.byname].

View all issues with NAD status.

Discussion:

  1. codecvt::do_length is of type int;
  2. it is assumed to be sort-of returning from_next - from of type ptrdiff_t;
  3. ptrdiff_t cannot be cast to an int without data loss.

Contradiction.

Proposed resolution:


501(i). Proposal: strengthen guarantees of lib.comparisons

Section: 99 [depr.base] Status: NAD Submitter: Me <anti_spam_email2003@yahoo.com> Opened: 2005-06-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [depr.base].

View all issues with NAD status.

Discussion:

"For templates greater, less, greater_equal, and less_equal, the specializations for any pointer type yield a total order, even if the built-in operators <, >, <=, >= do not."

The standard should do much better than guarantee that these provide a total order, it should guarantee that it can be used to test if memory overlaps, i.e. write a portable memmove. You can imagine a platform where the built-in operators use a uint32_t comparison (this tests for overlap on this platform) but the less<T*> functor is allowed to be defined to use a int32_t comparison. On this platform, if you use std::less with the intent of making a portable memmove, comparison on an array that straddles the 0x7FFFFFFF/0x8000000 boundary can give incorrect results.

Proposed resolution:

Add a footnote to 20.5.3/8 saying:

Given a p1 and p2 such that p1 points to N objects of type T and p2 points to M objects of type T. If [p1,p1+N) does not overlap [p2,p2+M), less returns the same value when comparing all pointers in [p1,p1+N) to all pointers in [p2,p2+M). Otherwise, there is a value Q and a value R such that less returns the same value when comparing all pointers in [p1,p1+Q) to all pointers in [p2,p2+R) and an opposite value when comparing all pointers in [p1+Q,p1+N) to all pointers in [p2+R,p2+M). For the sake of completeness, the null pointer value (4.10) for T is considered to be an array of 1 object that doesn't overlap with any non-null pointer to T. less_equal, greater, greater_equal, equal_to, and not_equal_to give the expected results based on the total ordering semantics of less. For T of void, treat it as having similar semantics as T of char i.e. less<cv T*>(a, b) gives the same results as less<cv void*>(a, b) which gives the same results as less<cv char*>((cv char*)(cv void*)a, (cv char*)(cv void*)b).

I'm also thinking there should be a footnote to 20.5.3/1 saying that if A and B are similar types (4.4/4), comp<A>(a,b) returns the same value as comp<B>(a,b) (where comp is less, less_equal, etc.). But this might be problematic if there is some really funky operator overloading going on that does different things based on cv (that should be undefined behavior if somebody does that though). This at least should be guaranteed for all POD types (especially pointers) that use the built-in comparison operators.

Rationale:

less is already required to provide a strict weak ordering which is good enough to detect overlapping memory situations.


502(i). Proposition: Clarification of the interaction between a facet and an iterator

Section: 30.3.1.2.1 [locale.category] Status: NAD Submitter: Christopher Conrade Zseleghovski Opened: 2005-06-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.category].

View all issues with NAD status.

Discussion:

Motivation:

This requirement seems obvious to me, it is the essence of code modularity. I have complained to Mr. Plauger that the Dinkumware library does not observe this principle but he objected that this behaviour is not covered in the standard.

[ 2009-07 Frankfurt ]

No objection to NAD, Fixed.

Move to NAD.

Proposed resolution:

Append the following point to 22.1.1.1.1:

6. The implementation of a facet of Table 52 parametrized with an InputIterator/OutputIterator should use that iterator only as character source/sink respectively. For a *_get facet, it means that the value received depends only on the sequence of input characters and not on how they are accessed. For a *_put facet, it means that the sequence of characters output depends only on the value to be formatted and not of how the characters are stored.

[ Berlin: Moved to Open, Need to clean up this area to make it clear locales don't have to contain open ended sets of facets. Jack, Howard, Bill. ]


503(i). more on locales

Section: 30.4 [locale.categories] Status: NAD Submitter: P.J. Plauger Opened: 2005-06-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.categories].

View all issues with NAD status.

Discussion:

a) In 22.2.1.1 para. 2 we refer to "the instantiations required in Table 51" to refer to the facet *objects* associated with a locale. And we almost certainly mean just those associated with the default or "C" locale. Otherwise, you can't switch to a locale that enforces a different mapping between narrow and wide characters, or that defines additional uppercase characters.

b) 22.2.1.5 para. 3 (codecvt) has the same issues.

c) 22.2.1.5.2 (do_unshift) is even worse. It *forbids* the generation of a homing sequence for the basic character set, which might very well need one.

d) 22.2.1.5.2 (do_length) likewise dictates that the default mapping between wide and narrow characters be taken as one-for-one.

e) 22.2.2 para. 2 (num_get/put) is both muddled and vacuous, as far as I can tell. The muddle is, as before, calling Table 51 a list of instantiations. But the constraint it applies seems to me to cover *all* defined uses of num_get/put, so why bother to say so?

f) 22.2.3.1.2 para. 1(do_decimal_point) says "The required instantiations return '.' or L'.'.) Presumably this means "as appropriate for the character type. But given the vague definition of "required" earlier, this overrules *any* change of decimal point for non "C" locales. Surely we don't want to do that.

g) 22.2.3.1.2 para. 2 (do_thousands_sep) says "The required instantiations return ',' or L','.) As above, this probably means "as appropriate for the character type. But this overrules the "C" locale, which requires *no* character ('\0') for the thousands separator. Even if we agree that we don't mean to block changes in decimal point or thousands separator, we should also eliminate this clear incompatibility with C.

h) 22.2.3.1.2 para. 2 (do_grouping) says "The required instantiations return the empty string, indicating no grouping." Same considerations as for do_decimal_point.

i) 22.2.4.1 para. 1 (collate) refers to "instantiations required in Table 51". Same bad jargon.

j) 22.2.4.1.2 para. 1 (do_compare) refers to "instantiations required in Table 51". Same bad jargon.

k) 22.2.5 para. 1 (time_get/put) uses the same muddled and vacuous as num_get/put.

l) 22.2.6 para. 2 (money_get/put) uses the same muddled and vacuous as num_get/put.

m) 22.2.6.3.2 (do_pos/neg_format) says "The instantiations required in Table 51 ... return an object of type pattern initialized to {symbol, sign, none, value}." This once again *overrides* the "C" locale, as well as any other locale."

3) We constrain the use_facet calls that can be made by num_get/put, so why don't we do the same for money_get/put? Or for any of the other facets, for that matter?

4) As an almost aside, we spell out when a facet needs to use the ctype facet, but several also need to use a codecvt facet and we don't say so.

[ Berlin: Bill to provide wording. ]

[ 2009-07 Frankfurt ]

No objection to NAD.

Move to NAD.

Proposed resolution:


504(i). Integer types in pseudo-random number engine requirements

Section: 28.5.3 [rand.req], 99 [tr.rand.req] Status: NAD Editorial Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.req].

View all issues with NAD Editorial status.

Discussion:

In [tr.rand.req], Paragraph 2 states that "... s is a value of integral type, g is an ... object returning values of unsigned integral type ..."

Proposed resolution:

In 5.1.1 [tr.rand.req], Paragraph 2 replace

... s is a value of integral type, g is an lvalue of a type other than X that defines a zero-argument function object returning values of unsigned integral type unsigned long int, ...

In 5.1.1 [tr.rand.seq], Table 16, replace in the line for X(s)

creates an engine with the initial internal state determined by static_cast<unsigned long>(s)

[ Mont Tremblant: Both s and g should be unsigned long. This should refer to the constructor signatures. Jens provided wording post Mont Tremblant. ]

[ Berlin: N1932 adopts the proposed resolution: see 26.3.1.3/1e and Table 3 row 2. Moved to Ready. ]

Rationale:

Jens: Just requiring X(unsigned long) still makes it possible for an evil library writer to also supply a X(int) that does something unexpected. The wording above requires that X(s) always performs as if X(unsigned long) would have been called. I believe that is sufficient and implements our intentions from Mont Tremblant. I see no additional use in actually requiring a X(unsigned long) signature. u.seed(s) is covered by its reference to X(s), same arguments.

[ Portland: Subsumed by N2111. ]


506(i). Requirements of Distribution parameter for variate_generator

Section: 28.5 [rand], 99 [tr.rand.var] Status: NAD Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand].

View all issues with NAD status.

Discussion:

Paragraph 3 requires that template argument U (which corresponds to template parameter Engine) satisfy all uniform random number generator requirements. However, there is no analogous requirement regarding the template argument that corresponds to template parameter Distribution. We believe there should be, and that it should require that this template argument satisfy all random distribution requirements.

Proposed resolution:

Consequence 1: Remove the precondition clauses [tr.rand.var]/16 and /18.

Consequence 2: Add max() and min() functions to those distributions that do not already have them.

[ Mont Tremblant: Jens reccommends NAD, min/max not needed everywhere. Marc supports having min and max to satisfy generic programming interface. ]

Rationale:

Berlin: N1932 makes this moot: variate_generator has been eliminated.


509(i). Uniform_int template parameters

Section: 28.5.9.2 [rand.dist.uni], 99 [tr.rand.dist.iunif] Status: NAD Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.dist.uni].

View all issues with NAD status.

Discussion:

In [tr.rand.dist.iunif] the uniform_int distribution currently has a single template parameter, IntType, used as the input_type and as the result_type of the distribution. We believe there is no reason to conflate these types in this way.

Proposed resolution:

We recommend that there be a second template parameter to reflect the distribution's input_type, and that the existing first template parameter continue to reflect (solely) the result_type:

template< class IntType = int, UIntType = unsigned int >
class uniform_int
{
public:
  // types
  typedef  UIntType  input_type;
  typedef  IntType   result_type;

[ Berlin: Moved to NAD. N1932 makes this moot: the input_type template parameter has been eliminated. ]


510(i). Input_type for bernoulli_distribution

Section: 28.5.9.3 [rand.dist.bern], 99 [tr.rand.dist.bern] Status: NAD Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

In [tr.rand.dist.bern] the distribution currently requires;

typedef  int  input_type;

Proposed resolution:

We believe this is an unfortunate choice, and recommend instead:

typedef  unsigned int  input_type;

[ Berlin: Moved to NAD. N1932 makes this moot: the input_type template parameter has been eliminated. ]


511(i). Input_type for binomial_distribution

Section: 28.5.9 [rand.dist], 99 [tr.rand.dist.bin] Status: NAD Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.dist].

View all issues with NAD status.

Discussion:

Unlike all other distributions in TR1, this binomial_distribution has an implementation-defined input_type. We believe this is an unfortunate choice, because it hinders users from writing portable code. It also hinders the writing of compliance tests. We recommend instead:

typedef  RealType  input_type;

While this choice is somewhat arbitrary (as it was for some of the other distributions), we make this particular choice because (unlike all other distributions) otherwise this template would not publish its RealType argument and so users could not write generic code that accessed this second template parameter. In this respect, the choice is consistent with the other distributions in TR1.

We have two reasons for recommending that a real type be specified instead. One reason is based specifically on characteristics of binomial distribution implementations, while the other is based on mathematical characteristics of probability distribution functions in general.

Implementations of binomial distributions commonly use Stirling approximations for values in certain ranges. It is far more natural to use real values to represent these approximations than it would be to use integral values to do so. In other ranges, implementations reply on the Bernoulli distribution to obtain values. While TR1's bernoulli_distribution::input_type is specified as int, we believe this would be better specified as double.

This brings us to our main point: The notion of a random distribution rests on the notion of a cumulative distribution function, which in turn mathematically depends on a continuous dependent variable. Indeed, such a distribution function would be meaningless if it depended on discrete values such as integers - and this remains true even if the distribution function were to take discrete steps.

Although this note is specifically about binomial_distribution::input_type, we intend to recommend that all of the random distributions input_types be specified as a real type (either a RealType template parameter, or double, as appropriate).

Of the nine distributions in TR1, four already have this characteristic (uniform_real, exponential_distribution, normal_distribution, and gamma_distribution). We have already argued the case for the binomial the remaining four distributions.

In the case of uniform_int, we believe that the calculations to produce an integer result in a specified range from an integer in a different specified range is best done using real arithmetic. This is because it involves a product, one of whose terms is the ratio of the extents of the two ranges. Without real arithmetic, the results become less uniform: some numbers become more (or less) probable that they should be. This is, of course, undesireable behavior in a uniform distribution.

Finally, we believe that in the case of the bernoulli_distribution (briefly mentioned earlier), as well as the cases of the geometric_distribution and the poisson_distribution, it would be far more natural to have a real input_type. This is because the most natural computation involves the random number delivered and the distribution's parameter p (in the case of bernoulli_distribution, for example, the computation is a comparison against p), and p is already specified in each case as having some real type.

Proposed resolution:

typedef  RealType  input_type;

[ Berlin: Moved to NAD. N1932 makes this moot: the input_type template parameter has been eliminated. ]


512(i). Seeding subtract_with_carry_01 from a single unsigned long

Section: 28.5.4 [rand.eng], 99 [tr.rand.eng.sub1] Status: NAD Editorial Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.eng].

View all issues with NAD Editorial status.

Discussion:

Paragraph 8 specifies the algorithm by which a subtract_with_carry_01 engine is to be seeded given a single unsigned long. This algorithm is seriously flawed in the case where the engine parameter w (also known as word_size) exceeds 31 [bits]. The key part of the paragraph reads:

sets x(-r) ... x(-1) to (lcg(1)*2**(-w)) mod 1

and so forth.

Since the specified linear congruential engine, lcg, delivers numbers with a maximum of 2147483563 (just a shade under 31 bits), then when w is, for example, 48, each of the x(i) will be less than 2**-17. The consequence is that roughly the first 400 numbers delivered will be conspicuously close to either zero or one.

Unfortunately, this is not an innocuous flaw: One of the predefined engines in [tr.rand.predef], namely ranlux64_base_01, has w = 48 and would exhibit this poor behavior, while the original N1378 proposal states that these pre-defined engines are intended to be of "known good properties."

Proposed resolution:

In 5.1.4.4 [tr.rand.eng.sub1], replace the "effects" clause for void seed(unsigned long value = 19780503) by

Effects: If value == 0, sets value to 19780503. In any case, with a linear congruential generator lcg(i) having parameters mlcg = 2147483563, alcg = 40014, clcg = 0, and lcg(0) = value, sets carry(-1) and x(-r) … x(-1) as if executing


linear_congruential<unsigned long, 40014, 0, 2147483563> lcg(value);
seed(lcg);

to (lcg(1) · 2-w) mod 1 … (lcg(r) · 2-w) mod 1, respectively. If x(-1) == 0, sets carry(-1) = 2-w, else sets carry(-1) = 0.

[ Jens provided revised wording post Mont Tremblant. ]

[ Berlin: N1932 adopts the originally-proposed resolution of the issue. Jens's supplied wording is a clearer description of what is intended. Moved to Ready. ]

Rationale:

Jens: I'm using an explicit type here, because fixing the prose would probably not qualify for the (with issue 504 even stricter) requirements we have for seed(Gen&).

[ Portland: Subsumed by N2111. ]


513(i). Size of state for subtract_with_carry_01

Section: 28.5.4 [rand.eng], 99 [tr.rand.eng.sub1] Status: NAD Editorial Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.eng].

View all issues with NAD Editorial status.

Discussion:

Paragraph 3 begins:

The size of the state is r.

However, this is not quite consistent with the remainder of the paragraph which specifies a total of nr+1 items in the textual representation of the state. We recommend the sentence be corrected to match:

The size of the state is nr+1.

To give meaning to the coefficient n, it may be also desirable to move n's definition from later in the paragraph. Either of the following seem reasonable formulations:

With n=..., the size of the state is nr+1.

The size of the state is nr+1, where n=... .

Proposed resolution:

[ Jens: I plead for "NAD" on the grounds that "size of state" is only used as an argument for big-O complexity notation, thus constant factors and additions don't count. ]

[ Berlin: N1932 adopts the proposed NAD. ]


514(i). Size of state for subtract_with_carry

Section: 28.5.4.4 [rand.eng.sub], 99 [tr.rand.eng.sub] Status: NAD Editorial Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.eng.sub].

View all issues with NAD Editorial status.

Discussion:

Paragraph 2 begins:

The size of the state is r.

However, the next sentence specifies a total of r+1 items in the textual representation of the state, r specific x's as well as a specific carry. This makes a total of r+1 items that constitute the size of the state, rather than r.

Proposed resolution:

We recommend the sentence be corrected to match:

The size of the state is r+1.

[ Jens: I plead for "NAD" on the grounds that "size of state" is only used as an argument for big-O complexity notation, thus constant factors and additions don't count. ]

[ Berlin: N1932 adopts the proposed NAD. ]


515(i). Random number engine traits

Section: 28.5.2 [rand.synopsis], 99 [tr.rand.synopsis] Status: NAD Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.synopsis].

View all issues with NAD status.

Discussion:

To accompany the concept of a pseudo-random number engine as defined in Table 17, we propose and recommend an adjunct template, engine_traits, to be declared in [tr.rand.synopsis] as:

template< class PSRE >
class engine_traits;

This template's primary purpose would be as an aid to generic programming involving pseudo-random number engines. Given only the facilities described in tr1, it would be very difficult to produce any algorithms involving the notion of a generic engine. The intent of this proposal is to provide, via engine_traits<>, sufficient descriptive information to allow an algorithm to employ a pseudo-random number engine without regard to its exact type, i.e., as a template parameter.

For example, today it is not possible to write an efficient generic function that requires any specific number of random bits. More specifically, consider a cryptographic application that internally needs 256 bits of randomness per call:

template< class Eng, class InIter, class OutIter >
void crypto( Eng& e, InIter in, OutIter out );

Without knowning the number of bits of randomness produced per call to a provided engine, the algorithm has no means of determining how many times to call the engine.

In a new section [tr.rand.eng.traits], we proposed to define the engine_traits template as:

template< class PSRE >
class engine_traits
{
  static  std::size_t  bits_of_randomness = 0u;
  static  std::string  name()  { return "unknown_engine"; }
  // TODO: other traits here
};

Further, each engine described in [tr.rand.engine] would be accompanied by a complete specialization of this new engine_traits template.

Proposed resolution:

[ Berlin: Walter: While useful for implementation per TR1, N1932 has no need for this feature. Recommend close as NAD. ]

Rationale:

Recommend NAD, N1932, N2111 covers this. Already in WP.


516(i). Seeding subtract_with_carry_01 using a generator

Section: 28.5.4 [rand.eng], 99 [tr.rand.eng.sub1] Status: NAD Editorial Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.eng].

View all issues with NAD Editorial status.

Discussion:

Paragraph 6 says:

... obtained by successive invocations of g, ...

We recommend instead:

... obtained by taking successive invocations of g mod 2**32, ...

as the context seems to require only 32-bit quantities be used here.

Proposed resolution:

Berlin: N1932 adopts the proposed resultion: see 26.3.3.4/7. Moved to Ready.

[ Portland: Subsumed by N2111. ]


517(i). Should include name in external representation

Section: 28.5.3 [rand.req], 99 [tr.rand.req] Status: NAD Submitter: Walter Brown Opened: 2005-07-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.req].

View all issues with NAD status.

Discussion:

The last two rows of Table 16 deal with the i/o requirements of an engine, specifying that the textual representation of an engine's state, appropriately formatted, constitute the engine's external representation.

This seems adequate when an engine's type is known. However, it seems inadequate in the context of generic code, where it becomes useful and perhaps even necessary to determine an engine's type via input.

Proposed resolution:

We therefore recommend that, in each of these two rows of Table 16, the text "textual representation" be expanded so as to read "engine name followed by the textual representation."

[ Berlin: N1932 considers this NAD. This is a QOI issue. ]


526(i). Is it undefined if a function in the standard changes in parameters?

Section: 24.2.4 [sequence.reqmts] Status: NAD Submitter: Chris Jefferson Opened: 2005-09-14 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [sequence.reqmts].

View all other issues in [sequence.reqmts].

View all issues with NAD status.

Discussion:

Problem: There are a number of places in the C++ standard library where it is possible to write what appear to be sensible ways of calling functions, but which can cause problems in some (or all) implementations, as they cause the values given to the function to be changed in a way not specified in standard (and therefore not coded to correctly work). These fall into two similar categories.

1) Parameters taken by const reference can be changed during execution of the function

Examples:

Given std::vector<int> v:

v.insert(v.begin(), v[2]);

v[2] can be changed by moving elements of vector

Given std::list<int> l:

l.remove(*l.begin());

Will delete the first element, and then continue trying to access it. This is particularily vicious, as it will appear to work in almost all cases.

2) A range is given which changes during the execution of the function: Similarly,

v.insert(v.begin(), v.begin()+4, v.begin()+6);

This kind of problem has been partly covered in some cases. For example std::copy(first, last, result) states that result cannot be in the range [first, last). However, does this cover the case where result is a reverse_iterator built from some iterator in the range [first, last)? Also, std::copy would still break if result was reverse_iterator(last + 1), yet this is not forbidden by the standard

Solution:

One option would be to try to more carefully limit the requirements of each function. There are many functions which would have to be checked. However as has been shown in the std::copy case, this may be difficult. A simpler, more global option would be to somewhere insert text similar to:

If the execution of any function would change either any values passed by reference or any value in any range passed to a function in a way not defined in the definition of that function, the result is undefined.

Such code would have to at least cover chapters 23 and 25 (the sections I read through carefully). I can see no harm on applying it to much of the rest of the standard.

Some existing parts of the standard could be improved to fit with this, for example the requires for 25.2.1 (Copy) could be adjusted to:

Requires: For each non-negative integer n < (last - first), assigning to *(result + n) must not alter any value in the range [first + n, last).

However, this may add excessive complication.

One other benefit of clearly introducing this text is that it would allow a number of small optimisations, such as caching values passed by const reference.

Matt Austern adds that this issue also exists for the insert and erase members of the ordered and unordered associative containers.

[ Berlin: Lots of controversey over how this should be solved. Lots of confusion as to whether we're talking about self referencing iterators or references. Needs a good survey as to the cases where this matters, for which implementations, and how expensive it is to fix each case. ]

Proposed resolution:

Rationale:

Recommend NAD.


528(i). const_iterator iterator issue when they are the same type

Section: 24.5 [unord], 99 [tr.unord.unord] Status: NAD Submitter: Paolo Carlini Opened: 2005-10-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [unord].

View all issues with NAD status.

Discussion:

while implementing the resolution of issue 6.19 I'm noticing the following: according to 6.3.4.3/2 (and 6.3.4.5/2), for unordered_set and unordered_multiset:

"The iterator and const_iterator types are both const types. It is unspecified whether they are the same type"

Now, according to the resolution of 6.19, we have overloads of insert with hint and erase (single and range) both for iterator and const_iterator, which, AFAICS, can be meaningful at the same time *only* if iterator and const_iterator *are* in fact different types.

Then, iterator and const_iterator are *required* to be different types? Or that is an unintended consequence? Maybe the overloads for plain iterators should be added only to unordered_map and unordered_multimap? Or, of course, I'm missing something?

Proposed resolution:

Add to 6.3.4.3p2 (and 6.3.4.5p2):

2 ... The iterator and const_iterator types are both const constant iterator types. It is unspecified whether they are the same type.

Add a new subsection to 17.4.4 [lib.conforming]:

An implementation shall not supply an overloaded function signature specified in any library clause if such a signature would be inherently ambiguous during overload resolution due to two library types referring to the same type.

[Note: For example, this occurs when a container's iterator and const_iterator types are the same. -- end note]

[ Post-Berlin: Beman supplied wording. ]

Rationale:

Toronto: The first issue has been fixed by N2350 (the insert and erase members are collapsed into one signature). Alisdair to open a separate issue on the chapter 17 wording.


529(i). The standard encourages redundant and confusing preconditions

Section: 99 [res.on.required] Status: NAD Editorial Submitter: David Abrahams Opened: 2005-10-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

17.4.3.8/1 says:

Violation of the preconditions specified in a function's Required behavior: paragraph results in undefined behavior unless the function's Throws: paragraph specifies throwing an exception when the precondition is violated.

This implies that a precondition violation can lead to defined behavior. That conflicts with the only reasonable definition of precondition: that a violation leads to undefined behavior. Any other definition muddies the waters when it comes to analyzing program correctness, because precondition violations may be routinely done in correct code (e.g. you can use std::vector::at with the full expectation that you'll get an exception when your index is out of range, catch the exception, and continue). Not only is it a bad example to set, but it encourages needless complication and redundancy in the standard. For example:

  21 Strings library 
  21.3.3 basic_string capacity

  void resize(size_type n, charT c);

  5 Requires: n <= max_size()
  6 Throws: length_error if n > max_size().
  7 Effects: Alters the length of the string designated by *this as follows:

The Requires clause is entirely redundant and can be dropped. We could make that simplifying change (and many others like it) even without changing 17.4.3.8/1; the wording there just seems to encourage the redundant and error-prone Requires: clause.

[ Batavia: Alan and Pete to work. ]

[ Bellevue: NAD Editorial, this group likes N2121, Pete agrees, accepting it is Pete's business. General agreement that precondition violations are synonymous with UB. ]

Proposed resolution:

1. Change 17.4.3.8/1 to read:

Violation of the preconditions specified in a function's Required behavior: paragraph results in undefined behavior unless the function's Throws: paragraph specifies throwing an exception when the precondition is violated.

2. Go through and remove redundant Requires: clauses. Specifics to be provided by Dave A.

[ Berlin: The LWG requests a detailed survey of part 2 of the proposed resolution. ]

[ Alan provided the survey N2121. ]


536(i). Container iterator constructor and explicit convertibility

Section: 24.2 [container.requirements] Status: Dup Submitter: Joaquín M López Muñoz Opened: 2005-12-17 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [container.requirements].

View all issues with Dup status.

Duplicate of: 589

Discussion:

The iterator constructor X(i,j) for containers as defined in 23.1.1 and 23.2.2 does only require that i and j be input iterators but nothing is said about their associated value_type. There are three sensible options:

  1. iterator's value_type is exactly X::value_type (modulo cv).
  2. iterator's value_type is *implicitly* convertible to X::value_type.
  3. iterator's value_type is *explicitly* convertible to X::value_type.

The issue has practical implications, and stdlib vendors have taken divergent approaches to it: Dinkumware follows 2, libstdc++ follows 3.

The same problem applies to the definition of insert(p,i,j) for sequences and insert(i,j) for associative contianers, as well as assign.

[ The following added by Howard and the example code was originally written by Dietmar. ]

Valid code below?

#include <vector> 
#include <iterator> 
#include <iostream> 

struct foo 
{ 
    explicit foo(int) {} 
}; 

int main() 
{ 
    std::vector<int> v_int; 
    std::vector<foo> v_foo1(v_int.begin(), v_int.end()); 
    std::vector<foo> v_foo2((std::istream_iterator<int>(std::cin)), 
                             std::istream_iterator<int>()); 
} 

[ Berlin: Some support, not universal, for respecting the explicit qualifier. ]

Proposed resolution:


544(i). minor NULL problems in C.2

Section: C.7 [diff.library] Status: NAD Editorial Submitter: Martin Sebor Opened: 2005-11-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [diff.library].

View all issues with NAD Editorial status.

Discussion:

According to C.2.2.3, p1, "the macro NULL, defined in any of <clocale>, <cstddef>, <cstdio>, <cstdlib>, <cstring>, <ctime>, or <cwchar>." This is consistent with the C standard.

However, Table 95 in C.2 fails to mention <clocale> and <cstdlib>.

In addition, C.2, p2 claims that "The C++ Standard library provides 54 standard macros from the C library, as shown in Table 95." While table 95 does have 54 entries, since a couple of them (including the NULL macro) are listed more than once, the actual number of macros defined by the C++ Standard Library may not be 54.

Proposed resolution:

I propose we add <clocale> and <cstdlib> to Table 96 and remove the number of macros from C.2, p2 and reword the sentence as follows:

The C++ Standard library provides 54 standard macros from defines a number macros corresponding to those defined by the C Standard library, as shown in Table 96.

[ Portland: Resolution is considered editorial. It will be incorporated into the WD. ]


546(i). [tr1] _Longlong and _ULonglong are integer types

Section: 5.1.1 [tr1::tr.rand.req] Status: NAD Submitter: Matt Austern Opened: 2006-01-10 Last modified: 2016-10-31

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

The TR sneaks in two new integer types, _Longlong and _Ulonglong, in [tr.c99]. The rest of the TR should use that type. I believe this affects two places. First, the random number requirements, 5.1.1/10-11, lists all of the types with which template parameters named IntType and UIntType may be instantiated. _Longlong (or "long long", assuming it is added to C++0x) should be added to the IntType list, and UIntType (again, or "unsigned long long") should be added to the UIntType list. Second, 6.3.2 lists the types for which hash<> is required to be instantiable. _Longlong and _Ulonglong should be added to that list, so that people may use long long as a hash key.

[ 2009-07 Frankfurt ]

We are not going to fix TR1.

The paper "long long goes to the library" addresses the integration of long long into the C++0x library.

Move to NAD.

Proposed resolution:


547(i). division should be floating-point, not integer

Section: 28.5 [rand], 99 [tr.rand] Status: NAD Submitter: Matt Austern Opened: 2006-01-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand].

View all issues with NAD status.

Discussion:

Paragraph 10 describes how a variate generator uses numbers produced by an engine to pass to a generator. The sentence that concerns me is: "Otherwise, if the value for engine_value_type::result_type is true and the value for Distribution::input_type is false [i.e. if the engine produces integers and the engine wants floating-point values], then the numbers in s_eng are divided by engine().max() - engine().min() + 1 to obtain the numbers in s_e." Since the engine is producing integers, both the numerator and the denominator are integers and we'll be doing integer division, which I don't think is what we want. Shouldn't we be performing a conversion to a floating-point type first?

Proposed resolution:

Rationale:

Recommend NAD as the affected section is now gone and so the issue is moot. N2111.


548(i). May random_device block?

Section: 28.5.7 [rand.device], 99 [tr.rand.device] Status: NAD Submitter: Matt Austern Opened: 2006-01-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.device].

View all issues with NAD status.

Discussion:

Class random_device "produces non-deterministic random numbers", using some external source of entropy. In most real-world systems, the amount of available entropy is limited. Suppose that entropy has been exhausted. What is an implementation permitted to do? In particular, is it permitted to block indefinitely until more random bits are available, or is the implementation required to detect failure immediately? This is not an academic question. On Linux a straightforward implementation would read from /dev/random, and "When the entropy pool is empty, reads to /dev/random will block until additional environmental noise is gathered." Programmers need to know whether random_device is permitted to (or possibly even required to?) behave the same way.

[ Berlin: Walter: N1932 considers this NAD. Does the standard specify whether std::cin may block? ]

See N2391 and N2423 for some further discussion.

Proposed resolution:

Adopt the proposed resolution in N2423 (NAD).


549(i). Undefined variable in binomial_distribution

Section: 28.5.9 [rand.dist], 99 [tr.rand.dist.bin] Status: NAD Editorial Submitter: Matt Austern Opened: 2006-01-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.dist].

View all issues with NAD Editorial status.

Discussion:

Paragraph 1 says that "A binomial distributon random distribution produces integer values i>0 with p(i) = (n choose i) * p*i * (1-p)^(t-i), where t and p are the parameters of the distribution. OK, that tells us what t, p, and i are. What's n?

Proposed resolution:

Berlin: Typo: "n" replaced by "t" in N1932: see 26.3.7.2.2/1.

[ Portland: Subsumed by N2111. ]


553(i). very minor editorial change intptr_t / uintptr_t

Section: 17.4.1 [cstdint.syn], 99 [tr.c99.cstdint.syn] Status: NAD Editorial Submitter: Paolo Carlini Opened: 2006-01-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [cstdint.syn].

View all issues with NAD Editorial status.

Discussion:

In the synopsis, some types are identified as optional: int8_t, int16_t, and so on, consistently with C99, indeed.

On the other hand, intptr_t and uintptr_t, are not marked as such and probably should, consistently with C99, 7.18.1.4.

Proposed resolution:

Change 17.4.1 [cstdint.syn]:

...
typedef signed integer type intptr_t;    // optional
...
typedef unsigned integer type uintptr_t;    // optional
...

Rationale:

Recommend NAD and fix as editorial with the proposed resolution.


554(i). Problem with lwg DR 184 numeric_limits<bool>

Section: 17.3.5.3 [numeric.special] Status: NAD Submitter: Howard Hinnant Opened: 2006-01-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [numeric.special].

View all issues with NAD status.

Discussion:

I believe we have a bug in the resolution of: 184 (WP status).

The resolution spells out each member of numeric_limits<bool>. The part I'm having a little trouble with is:

static const bool traps = false;

Should this not be implementation defined? Given:

int main()
{
     bool b1 = true;
     bool b2 = false;
     bool b3 = b1/b2;
}

If this causes a trap, shouldn't numeric_limits<bool>::traps be true?

Proposed resolution:

Change 18.2.1.5p3:

-3- The specialization for bool shall be provided as follows:

namespace std { 
   template <> class numeric_limits<bool> {
      ...
      static const bool traps = false implementation-defined;
      ...
   };
}

[ Redmond: NAD because traps refers to values, not operations. There is no bool value that will trap. ]


555(i). [tr1] 8.21/1: typo

Section: 8.21 [tr1::tr.c99.boolh] Status: NAD Editorial Submitter: Paolo Carlini Opened: 2006-02-02 Last modified: 2016-10-31

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

This one, if nobody noticed it yet, seems really editorial: s/cstbool/cstdbool/

Proposed resolution:

Change 8.21p1:

-1- The header behaves as if it defines the additional macro defined in <cstdbool> by including the header <cstdbool>.

[ Redmond: Editorial. ]


557(i). TR1: div(_Longlong, _Longlong) vs div(intmax_t, intmax_t)

Section: 17.4.1 [cstdint.syn], 99 [tr.c99.cstdint] Status: NAD Editorial Submitter: Paolo Carlini Opened: 2006-02-06 Last modified: 2023-02-07

Priority: Not Prioritized

View all other issues in [cstdint.syn].

View all issues with NAD Editorial status.

Discussion:

I'm seeing a problem with such overloads: when, _Longlong == intmax_t == long long we end up, essentially, with the same arguments and different return types (lldiv_t and imaxdiv_t, respectively). Similar issue with abs(_Longlong) and abs(intmax_t), of course.

Comparing sections 8.25 and 8.11, I see an important difference, however: 8.25.3 and 8.25.4 carefully describe div and abs for _Longlong types (rightfully, because not moved over directly from C99), whereas there is no equivalent in 8.11: the abs and div overloads for intmax_t types appear only in the synopsis and are not described anywhere, in particular no mention in 8.11.2 (at variance with 8.25.2).

I'm wondering whether we really, really, want div and abs for intmax_t...

Proposed resolution:

[ Portland: no consensus. ]

Rationale:

[ Batavia, Bill: The <cstdint> synopsis in [tr.c99.cinttypes.syn] contains: ]

intmax_t imaxabs(intmax_t i);
intmax_t abs(intmax_t i);

imaxdiv_t imaxdiv(intmax_t numer, intmax_t denom);
imaxdiv_t div(intmax_t numer, intmax_t denom);

[ and in [tr.c99.cinttypes.def]: ]

The header defines all functions, types, and macros the same as C99 subclause 7.8.

[ This is as much definition as we give for most other C99 functions, so nothing need change. We might, however, choose to add the footnote: ]

[Note: These overloads for abs and div may well be equivalent to those that take long long arguments. If so, the implementation is responsible for avoiding conflicting declarations. -- end note]

[ Bellevue: NAD Editorial. Pete must add a footnote, as described below. ]

[ Looks like a real problem. Dietmar suggests div() return a template type. Matt: looks like imaxdiv_t is loosly defined. Can it be a typedef for lldiv_t when _Longlong == intmax_t? PJP seems to agree. We would need a non-normative note declaring that the types lldiv_t and imaxdiv_t may not be unique if intmax_t==_longlong. ]


558(i). lib.input.iterators Defect

Section: 25.3.5.3 [input.iterators] Status: NAD Editorial Submitter: David Abrahams Opened: 2006-02-09 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [input.iterators].

View all other issues in [input.iterators].

View all issues with NAD Editorial status.

Discussion:

24.1.1 Input iterators [lib.input.iterators]

1 A class or a built-in type X satisfies the requirements of an input iterator for the value type T if the following expressions are valid, where U is the type of any specified member of type T, as shown in Table 73.

There is no capital U used in table 73. There is a lowercase u, but that is clearly not meant to denote a member of type T. Also, there's no description in 24.1.1 of what lowercase a means. IMO the above should have been...Hah, a and b are already covered in 24.1/11, so maybe it should have just been:

Proposed resolution:

Change 24.1.1p1:

-1- A class or a built-in type X satisfies the requirements of an input iterator for the value type T if the following expressions are valid, where U is the type of any specified member of type T, as shown in Table 73.

[ Portland: Editorial. ]


560(i). User-defined allocators without default constructor

Section: 16.4.4.6 [allocator.requirements] Status: NAD Submitter: Sergey P. Derevyago Opened: 2006-02-17 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with NAD status.

Discussion:

1. The essence of the problem.

User-defined allocators without default constructor are not explicitly supported by the standard but they can be supported just like std::vector supports elements without default constructor.

As a result, there exist implementations that work well with such allocators and implementations that don't.

2. The cause of the problem.

1) The standard doesn't explicitly state this intent but it should. In particular, 20.1.5p5 explicitly state the intent w.r.t. the allocator instances that compare non-equal. So it can similarly state the intent w.r.t. the user-defined allocators without default constructor.

2) Some container operations are obviously underspecified. In particular, 21.3.7.1p2 tells:

template<class charT, class traits, class Allocator>
  basic_string<charT,traits,Allocator> operator+(
    const charT* lhs,
    const basic_string<charT,traits,Allocator>& rhs
  );

Returns: basic_string<charT,traits,Allocator>(lhs) + rhs.

That leads to the basic_string<charT,traits,Allocator>(lhs, Allocator()) call. Obviously, the right requirement is:

Returns: basic_string<charT,traits,Allocator>(lhs, rhs.get_allocator()) + rhs.

It seems like a lot of DRs can be submitted on this "Absent call to get_allocator()" topic.

3. Proposed actions.

1) Explicitly state the intent to allow for user-defined allocators without default constructor in 20.1.5 Allocator requirements.

2) Correct all the places, where a correct allocator object is available through the get_allocator() call but default Allocator() gets passed instead.

4. Code sample.

Let's suppose that the following memory pool is available:

class mem_pool {
      // ...
      void* allocate(size_t size);
      void deallocate(void* ptr, size_t size);
};

So the following allocator can be implemented via this pool:

class stl_allocator {
      mem_pool& pool;

 public:
      explicit stl_allocator(mem_pool& mp) : pool(mp) {}
      stl_allocator(const stl_allocator& sa) : pool(sa.pool) {}
      template <class U>
      stl_allocator(const stl_allocator<U>& sa)  : pool(sa.get_pool()) {}
      ~stl_allocator() {}

      pointer allocate(size_type n, std::allocator<void>::const_pointer = 0)
      {
       return (n!=0) ? static_cast<pointer>(pool.allocate(n*sizeof(T))) : 0;
      }

      void deallocate(pointer p, size_type n)
      {
       if (n!=0) pool.deallocate(p, n*sizeof(T));
      }

      // ...
};

Then the following code works well on some implementations and doesn't work on another:

typedef basic_string<char, char_traits<char>, stl_allocator<char> > 
  tl_string;
mem_pool mp;
tl_string s1("abc", stl_allocator<int>(mp));
printf("(%s)\n", ("def"+s1).c_str());

In particular, on some implementations the code can't be compiled without default stl_allocator() constructor.

The obvious way to solve the compile-time problems is to intentionally define a NULL pointer dereferencing default constructor

stl_allocator() : pool(*static_cast<mem_pool*>(0)) {}

in a hope that it will not be called. The problem is that it really gets called by operator+(const char*, const string&) under the current 21.3.7.1p2 wording.

Proposed resolution:

Rationale:

Recommend NAD. operator+() with string already requires the desired semantics of copying the allocator from one of the strings (lhs when there is a choice).


568(i). [tr1] log2 overloads missing

Section: 8.16.4 [tr1::tr.c99.cmath.over] Status: NAD Submitter: Paolo Carlini Opened: 2006-03-07 Last modified: 2016-02-01

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

log2 is missing from the list of "additional overloads" in 8.16.4 [tr1::tr.c99.cmath.over] p1.

Hinnant: This is a TR1 issue only. It is fixed in the current (N2135) WD.

[ Batavia (2009-05): ]

We agree this has been fixed in the Working Draft. Move to NAD.

Proposed resolution:

Add log2 to the list of functions in 8.16.4 [tr1::tr.c99.cmath.over] p1.


569(i). Postcondition for basic_ios::clear(iostate) incorrectly stated

Section: 31.5.4.4 [iostate.flags] Status: Dup Submitter: Seungbeom Kim Opened: 2006-03-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iostate.flags].

View all issues with Dup status.

Duplicate of: 272

Discussion:

Section: 27.4.4.3 [lib.iostate.flags]

Paragraph 4 says:

void clear(iostate state = goodbit);

Postcondition: If rdbuf()!=0 then state == rdstate(); otherwise rdstate()==state|ios_base::badbit.

The postcondition "rdstate()==state|ios_base::badbit" is parsed as "(rdstate()==state)|ios_base::badbit", which is probably what the committee meant.

Rationale:


570(i). Request adding additional explicit specializations of char_traits

Section: 23.2 [char.traits] Status: NAD Submitter: Jack Reeves Opened: 2006-04-06 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [char.traits].

View all issues with NAD status.

Discussion:

Currently, the Standard Library specifies only a declaration for template class char_traits<> and requires the implementation provide two explicit specializations: char_traits<char> and char_traits<wchar_t>. I feel the Standard should require explicit specializations for all built-in character types, i.e. char, wchar_t, unsigned char, and signed char.

I have put together a paper (N1985) that describes this in more detail and includes all the necessary wording.

[ Portland: Jack will rewrite N1985 to propose a primary template that will work with other integral types. ]

[ Toronto: issue has grown with addition of char16_t and char32_t. ]

[ post Bellevue: ]

We suggest that Jack be asked about the status of his paper, and if it is not forthcoming, the work-item be assigned to someone else. If no one steps forward to do the paper before the next meeting, we propose to make this NAD without further discussion. We leave this Open for now, but our recommendation is NAD.

Note: the issue statement should be updated, as the Toronto comment has already been resolved. E.g., char_traits specializations for char16_t and char32_t are now in the working paper.

[ Sophia Antipolis: ]

Nobody has submitted the requested paper, so we move to NAD, as suggested by the decision at the last meeting.

Proposed resolution:


571(i). Update C90 references to C99?

Section: 2 [intro.refs] Status: NAD Editorial Submitter: Beman Dawes Opened: 2006-04-08 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [intro.refs].

View all issues with NAD Editorial status.

Discussion:

1.2 Normative references [intro.refs] of the WP currently refers to ISO/IEC 9899:1990, Programming languages - C. Should that be changed to ISO/IEC 9899:1999?

What impact does this have on the library?

Proposed resolution:

In 1.2/1 [intro.refs] of the WP, change:

Rationale:

Recommend NAD, fixed editorially.


572(i). Oops, we gave 507 WP status

Section: 28.5 [rand], 99 [tr.rand] Status: NAD Submitter: Howard Hinnant Opened: 2006-04-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand].

View all issues with NAD status.

Discussion:

In Berlin, as a working group, we voted in favor of N1932 which makes issue 507 moot: variate_generator has been eliminated. Then in full committee we voted to give this issue WP status (mistakenly).

Proposed resolution:

Strike the proposed resolution of issue 507.

[ post-Portland: Walter and Howard recommend NAD. The proposed resolution of 507 no longer exists in the current WD. ]

Rationale:

NAD. Will be moot once N2135 is adopted.


573(i). C++0x file positioning should handle modern file sizes

Section: 31.5.3 [fpos] Status: NAD Submitter: Beman Dawes Opened: 2006-04-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [fpos].

View all issues with NAD status.

Discussion:

There are two deficiencies related to file sizes:

  1. It doesn't appear that the Standard Library is specified in a way that handles modern file sizes, which are often too large to be represented by an unsigned long.
  2. The std::fpos class does not currently have the ability to set/get file positions.

The Dinkumware implementation of the Standard Library as shipped with the Microsoft compiler copes with these issues by:

  1. Defining fpos_t be long long, which is large enough to represent any file position likely in the foreseeable future.
  2. Adding member functions to class fpos. For example,
    fpos_t seekpos() const;
    

Because there are so many types relating to file positions and offsets (fpos_t, fpos, pos_type, off_type, streamoff, streamsize, streampos, wstreampos, and perhaps more), it is difficult to know if the Dinkumware extensions are sufficient. But they seem a useful starting place for discussions, and they do represent existing practice.

[ Kona (2007): We need a paper. It would be nice if someone proposed clarifications to the definitions of pos_type and off_type. Currently these definitions are horrible. Proposed Disposition: Open ]

[ 2009-07 Frankfurt ]

This is the subject of paper N2926.

If we choose to take any action, we will move the paper, so the issue can be closed.

Move to NAD.

Proposed resolution:


579(i). erase(iterator) for unordered containers should not return an iterator

Section: 24.2.8 [unord.req] Status: NAD Submitter: Joaquín M López Muñoz Opened: 2006-06-13 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with NAD status.

Discussion:

Addresses ES-2

See N2023 for full discussion.

[ 2009-12-11 Paolo opens: ]

I'm asking for DR 579 to be re-opened, basing on recent discussions on the library reflector, see Message c++std-lib-26040 and replies.

[ 2010-02-07 Paolo updates wording. ]

As pointed out by Chris in c++std-lib-26040, that an erase(unordered_container, iterator) returning an iterator can easily implemented in user code, if needed; that actually returning an iterator costs nothing for the overload taking two iterators, thus that proposed change is only for consistency; that forward_list::erase_after also returns void (for different reasons, granted, but isn't that any "erase" function in the containers uniformly returns an iterator); that, also in thread started by Chris' message, Alberto pointed out that the proxy idea isn't a good one; that users both of the GNU and Boost implementations are reporting serious performance problems with the current version returning an iterator.

[ 2010-02-07 Original wording saved here: ]

Option 1:

The problem can be eliminated by omitting the requirement that a.erase(q) return an iterator. This is, however, in contrast with the equivalent requirements for other standard containers.

Option 2:

a.erase(q) can be made to compute the next iterator only when explicitly requested: the technique consists in returning a proxy object implicitly convertible to iterator, so that

iterator q1=a.erase(q);

works as expected, while

a.erase(q);

does not ever invoke the conversion-to-iterator operator, thus avoiding the associated computation. To allow this technique, some sections of TR1 along the line "return value is an iterator..." should be changed to "return value is an unspecified object implicitly convertible to an iterator..." Although this trick is expected to work transparently, it can have some collateral effects when the expression a.erase(q) is used inside generic code.

[ 2010-03-27 Joaquín adds: ]

Signature of iterator erase(const_iterator) should be changed to void erase(const_iterator). If this is not viable an acceptable tradeoff could be to make the return type of erase(const_iterator) implementation defined.

The standard should allow implementations of unordered associative containers using either singly or doubly linked lists. N2023 proves that singly-linked lists implementations cannot provide the required complexity for iterator erase(const_iterator). Thus, some action is needed to allow both implementations.

Option 1: Changing the required complexity from O(1) to O(log n). This option merely masks a design flaw. Users are forcefully penalized for what they don't use (the returned iterator). Besides, they would have to learn about the pathological (yet very real) situations where using erase can lead to quadratic performance. Two out of these three objections remain even if some alternative member function like void quick_erase(const_iterator) is thrown in to the interface.

Some objections have been expressed to changing return type of erase to void, arguing that it would break current existing practice with standard library implementations based on doubly-linked lists, where the problem does not occur. However implementations based on drafts should not block the resolution of a serious design issue, more so when the issue will hurt future users of C++, as it's happening already.

Option 2: Make erase return type implementation defined. There's a possible tradeoff with the objectors above consisting in changing the signature to implementation defined erase(iterator), so that returning an iterator is indeed a valid extension. To this it can be argued that this would make implementantions returning an iterator look as somehow promoting proprietary extensions: this in my opinion is not a valid argument since those implementations are already extending the required interface by providing bidirectional iterators (just forward iterators are required).

[ 2010 Rapperswil: ]

The issue was lengthy discussed and implementation experience was demonstrated that a non-void return type is implementable for both single-linked and double-linked lists without loss of efficiency.

By a 12-1-1-0 poll voted to keep the return type of erase as iterator instead of void and a second 0-0-3-10 poll rejected the additional proposal to add a quick_erase returning void, thus LWG decided for NAD.

Rationale:

No consensus for a change.

Proposed resolution:


580(i). unused allocator members

Section: 24.2.2.1 [container.requirements.general] Status: NAD Editorial Submitter: Martin Sebor Opened: 2006-06-14 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [container.requirements.general].

View all other issues in [container.requirements.general].

View all issues with NAD Editorial status.

Duplicate of: 479

Discussion:

C++ Standard Library templates that take an allocator as an argument are required to call the allocate() and deallocate() members of the allocator object to obtain storage. However, they do not appear to be required to call any other allocator members such as construct(), destroy(), address(), and max_size(). This makes these allocator members less than useful in portable programs.

It's unclear to me whether the absence of the requirement to use these allocator members is an unintentional omission or a deliberate choice. However, since the functions exist in the standard allocator and since they are required to be provided by any user-defined allocator I believe the standard ought to be clarified to explictly specify whether programs should or should not be able to rely on standard containers calling the functions.

I propose that all containers be required to make use of these functions.

[ Batavia: We support this resolution. Martin to provide wording. ]

[ pre-Oxford: Martin provided wording. ]

[ 2009-04-28 Pablo adds: ]

N2554 (scoped allocators), N2768 (allocator concepts), and N2810 (allocator defects), address all of these points EXCEPT max_size(). So, I would add a note to that affect and re-class the defect as belonging to section 24.2.2.1 [container.requirements.general].

[ 2009-07 Frankfurt ]

The comment in the description of this issue that this "would be" rendered editorial by the adoption of N2257 is confusing. It appears that N2257 was never adopted.

[ 2009-10 Santa Cruz: ]

NAD Editorial. Addressed by N2982.

Proposed resolution:

Specifically, I propose to change 24.2 [container.requirements], p9 as follows:

-9- Copy constructors for all container types defined in this clause that are parametrized on Allocator copy anthe allocator argument from their respective first parameters. All other constructors for these container types take an const Allocator& argument (20.1.6), an allocator whose value_type is the same as the container's value_type. A copy of this argument isshall be used for any memory allocation and deallocation performed, by these constructors and by all member functions, during the lifetime of each container object. Allocation shall be performed "as if" by calling the allocate() member function on a copy of the allocator object of the appropriate type New Footnote), and deallocation "as if" by calling deallocate() on a copy of the same allocator object of the corresponding type. A copy of this argument shall also be used to construct and destroy objects whose lifetime is managed by the container, including but not limited to those of the container's value_type, and to obtain their address. All objects residing in storage allocated by a container's allocator shall be constructed "as if" by calling the construct() member function on a copy of the allocator object of the appropriate type. The same objects shall be destroyed "as if" by calling destroy() on a copy of the same allocator object of the same type. The address of such objects shall be obtained "as if" by calling the address() member function on a copy of the allocator object of the appropriate type. Finally, a copy of this argument shall be used by its container object to determine the maximum number of objects of the container's value_type the container may store at the same time. The container member function max_size() obtains this number from the value returned by a call to get_allocator().max_size(). In all container types defined in this clause that are parametrized on Allocator, the member get_allocator() returns a copy of the Allocator object used to construct the container.258)

New Footnote: This type may be different from Allocator: it may be derived from Allocator via Allocator::rebind<U>::other for the appropriate type U.

The proposed wording seems cumbersome but I couldn't think of a better way to describe the requirement that containers use their Allocator to manage only objects (regardless of their type) that persist over their lifetimes and not, for example, temporaries created on the stack. That is, containers shouldn't be required to call Allocator::construct(Allocator::allocate(1), elem) just to construct a temporary copy of an element, or Allocator::destroy(Allocator::address(temp), 1) to destroy temporaries.

[ Howard: This same paragraph will need some work to accommodate 431. ]

[ post Oxford: This would be rendered NAD Editorial by acceptance of N2257. ]


582(i). specialized algorithms and volatile storage

Section: 27.11.5 [uninitialized.copy] Status: NAD Submitter: Martin Sebor Opened: 2006-06-14 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [uninitialized.copy].

View all issues with NAD status.

Discussion:

Related to 1029

The specialized algorithms [lib.specialized.algorithms] are specified as having the general effect of invoking the following expression:


new (static_cast<void*>(&*i))
    typename iterator_traits<ForwardIterator>::value_type (x)

            

This expression is ill-formed when the type of the subexpression &*i is some volatile-qualified T.

[ Batavia: Lack of support for proposed resolution but agree there is a defect. Howard to look at wording. Concern that move semantics properly expressed if iterator returns rvalue. ]

[ 2009-06-17 Pablo adds: ]

Propose that Issue 582 be closed NAD.

Issue 582 asks that uninitialized_copy, uninitialized_fill, and uninitialized_fill_n should be well-formed if the result type is volatile. My feeling is that the standard does not, and should not, guarantee any useful behavior when constructors are invoked on volatile storage, so making it syntactically legal to call uninitialized_copy on volatile storage is not useful. A possible editorial change would be to put my previous sentence into a non-normative note.

Note that the three sections starting with 27.11.5 [uninitialized.copy] do not yet have concepts. Here's a first crack at the first one:

template <InputIterator InIter, OutputIterator OutIter>
requires ExplicitConvertible<HasDereference<OutIter::reference>::result,
                             OutIter::value_type&>
      && Convertible<OutIter::value_type*, void*>
      && ExplicitConvertible<OutIter::value_type, InIter::reference>
  OutIter uninitialized_copy(InIter first, InIter last, OutIter result);

Effects:

while (first != last) {
  typedef OutIter::value_type value_type;
  value_type& outRef = static_cast<value_type&>(*result++);
  ::new (static_cast<void*>(addressof(outRef))) value_type(*first++);
}

Notes:

  1. This definition is actually LESS constrained than in C++03 because there is no requirement that the result be a forward iterator.
  2. If OutIter returns a proxy type with an overloaded operator&, this definition probably won't compile. Lifting this limitation while allowing value_type to have an overloaded operator& would be hard, but is probably possible with careful overloading. I'm not sure it's worth it.
  3. This definition retains the prohibition on the use of volatile types for the result.

[ 2009-07 Frankfurt ]

We don't deal with volatile in the library.

Jim: should we state that explicitly somewhere?

Beman: you might argue that clause 17 should say something about volatile. However, if you want to raise we argument, we should open it as a separate issue and consult with experts on concurrency.

Hinnant: actually, some library components do handle volatile, so we'd need to be very careful about what we say in clause 17.

No objection to NAD.

Move to NAD.

Proposed resolution:

In order to allow these algorithms to operate on volatile storage I propose to change the expression so as to make it well-formed even for pointers to volatile types. Specifically, I propose the following changes to clauses 20 and 24. Change 20.6.4.1, p1 to read:


Effects:

typedef typename iterator_traits<ForwardIterator>::pointer    pointer;
typedef typename iterator_traits<ForwardIterator>::value_type value_type;

for (; first != last; ++result, ++first)
    new (static_cast<void*>(const_cast<pointer>(&*result))
        value_type (*first);

            

change 20.6.4.2, p1 to read


Effects:

typedef typename iterator_traits<ForwardIterator>::pointer    pointer;
typedef typename iterator_traits<ForwardIterator>::value_type value_type;

for (; first != last; ++result, ++first)
    new (static_cast<void*>(const_cast<pointer>(&*first))
        value_type (*x);

            

and change 20.6.4.3, p1 to read


Effects:

typedef typename iterator_traits<ForwardIterator>::pointer    pointer;
typedef typename iterator_traits<ForwardIterator>::value_type value_type;

for (; n--; ++first)
    new (static_cast<void*>(const_cast<pointer>(&*first))
        value_type (*x);

            

In addition, since there is no partial specialization for iterator_traits<volatile T*> I propose to add one to parallel such specialization for <const T*>. Specifically, I propose to add the following text to the end of 24.3.1, p3:

and for pointers to volatile as


namespace std {
template<class T> struct iterator_traits<volatile T*> {
typedef ptrdiff_t difference_type;
typedef T value_type;
typedef volatile T* pointer;
typedef volatile T& reference;
typedef random_access_iterator_tag iterator_category;
};
}

            

Note that the change to iterator_traits isn't necessary in order to implement the specialized algorithms in a way that allows them to operate on volatile strorage. It is only necesassary in order to specify their effects in terms of iterator_traits as is done here. Implementations can (and some do) achieve the same effect by means of function template overloading.


583(i). div() for unsigned integral types

Section: 28.7 [c.math] Status: NAD Submitter: Beman Dawes Opened: 2006-06-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [c.math].

View all issues with NAD status.

Discussion:

There is no div() function for unsigned integer types.

There are several possible resolutions. The simplest one is noted below. Other possibilities include a templated solution.

Proposed resolution:

Add to 26.7 [lib.c.math] paragraph 8:

struct udiv_t div(unsigned, unsigned);
struct uldiv_t div(unsigned long, unsigned long);
struct ulldiv_t div(unsigned long long, unsigned long long);

Rationale:

Toronto: C99 does not have these unsigned versions because the signed version exist just to define the implementation-defined behavior of signed integer division. Unsigned integer division has no implementation-defined behavior and thus does not need this treatment.


584(i). missing int pow(int,int) functionality

Section: 28.7 [c.math] Status: NAD Submitter: Beman Dawes Opened: 2006-06-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [c.math].

View all issues with NAD status.

Discussion:

There is no pow() function for any integral type.

Proposed resolution:

Add something like:

template< typename T>
T power( T x, int n );
// requires: n >=0

Rationale:

Toronto: We already have double pow(integral, integral) from 28.7 [c.math] p11.


585(i). facet error reporting

Section: 30.4 [locale.categories] Status: NAD Submitter: Martin Sebor, Paolo Carlini Opened: 2006-06-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.categories].

View all issues with NAD status.

Discussion:

Section 22.2, paragraph 2 requires facet get() members that take an ios_base::iostate& argument, err, to ignore the (initial) value of the argument, but to set it to ios_base::failbit in case of a parse error.

We believe there are a few minor problems with this blanket requirement in conjunction with the wording specific to each get() member function.

First, besides get() there are other member functions with a slightly different name (for example, get_date()). It's not completely clear that the intent of the paragraph is to include those as well, and at least one implementation has interpreted the requirement literally.

Second, the requirement to "set the argument to ios_base::failbit suggests that the functions are not permitted to set it to any other value (such as ios_base::eofbit, or even ios_base::eofbit | ios_base::failbit).

However, 22.2.2.1.2, p5 (Stage 3 of num_get parsing) and p6 (bool parsing) specifies that the do_get functions perform err |= ios_base::eofbit, which contradicts the earlier requirement to ignore err's initial value.

22.2.6.1.2, p1 (the Effects clause of the money_get facet's do_get member functions) also specifies that err's initial value be used to compute the final value by ORing it with either ios_base::failbit or withios_base::eofbit | ios_base::failbit.

[ 2009-07 Frankfurt ]

Move to NAD.

Proposed resolution:

We believe the intent is for all facet member functions that take an ios_base::iostate& argument to:

To that effect we propose to change 22.2, p2 as follows:

The put() members make no provision for error reporting. (Any failures of the OutputIterator argument must be extracted from the returned iterator.) Unless otherwise specified, the get() members that take an ios_base::iostate& argument whose value they ignore, but set to ios_base::failbit in case of a parse error., err, start by evaluating err = ios_base::goodbit, and may subsequently set err to either ios_base::eofbit, or ios_base::failbit, or ios_base::eofbit | ios_base::failbit in response to reaching the end-of-file or in case of a parse error, or both, respectively.

[ Kona (2007): We need to change the proposed wording to clarify that the phrase "the get members" actually denotes get(), get_date(), etc. Proposed Disposition: Open ]


587(i). iststream ctor missing description

Section: D.15.3.2 [depr.istrstream.cons] Status: NAD Editorial Submitter: Martin Sebor Opened: 2006-06-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

The iststream(char*, streamsize) ctor is in the class synopsis in D.7.2 but its signature is missing in the description below (in D.7.2.1).

Proposed resolution:

This seems like a simple editorial issue and the missing signature can be added to the one for const char* in paragraph 2.

[ post Oxford: Noted that it is already fixed in N2284 ]


588(i). requirements on zero sized tr1::arrays and other details

Section: 24.3.7 [array] Status: NAD Submitter: Gennaro Prota Opened: 2006-07-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [array].

View all issues with NAD status.

Discussion:

The wording used for section 23.2.1 [lib.array] seems to be subtly ambiguous about zero sized arrays (N==0). Specifically:

* "An instance of array<T, N> stores N elements of type T, so that [...]"

Does this imply that a zero sized array object stores 0 elements, i.e. that it cannot store any element of type T? The next point clarifies the rationale behind this question, basically how to implement begin() and end():

* 23.2.1.5 [lib.array.zero], p2: "In the case that N == 0, begin() == end() == unique value."

What does "unique" mean in this context? Let's consider the following possible implementations, all relying on a partial specialization:

a)
    template< typename T >
    class array< T, 0 > {
    
        ....

        iterator begin()
        { return iterator( reinterpret_cast< T * >( this ) ); }
        ....

    };

This has been used in boost, probably intending that the return value had to be unique to the specific array object and that array couldn't store any T. Note that, besides relying on a reinterpret_cast, has (more than potential) alignment problems.

b)
    template< typename T >
    class array< T, 0 > {
    
        T t;

        iterator begin()
        { return iterator( &t ); }
        ....

    };

This provides a value which is unique to the object and to the type of the array, but requires storing a T. Also, it would allow the user to mistakenly provide an initializer list with one element.

A slight variant could be returning *the* null pointer of type T

    return static_cast<T*>(0);

In this case the value would be unique to the type array<T, 0> but not to the objects (all objects of type array<T, 0> with the same value for T would yield the same pointer value).

Furthermore this is inconsistent with what the standard requires from allocation functions (see library issue 9).

c) same as above but with t being a static data member; again, the value would be unique to the type, not to the object.

d) to avoid storing a T *directly* while disallowing the possibility to use a one-element initializer list a non-aggregate nested class could be defined

    struct holder { holder() {} T t; } h;

and then begin be defined as

 iterator begin() { return &h.t; }

But then, it's arguable whether the array stores a T or not. Indirectly it does.

-----------------------------------------------------

Now, on different issues:

* what's the effect of calling assign(T&) on a zero-sized array? There seems to be only mention of front() and back(), in 23.2.1 [lib.array] p4 (I would also suggest to move that bullet to section 23.2.1.5 [lib.array.zero], for locality of reference)

* (minor) the opening paragraph of 23.2.1 [lib.array] wording is a bit inconsistent with that of other sequences: that's not a problem in itself, but compare it for instance with "A vector is a kind of sequence that supports random access iterators"; though the intent is obvious one might argue that the wording used for arrays doesn't tell what an array is, and relies on the reader to infer that it is what the <array> header defines.

* it would be desiderable to have a static const data member of type std::size_t, with value N, for usage as integral constant expression

* section 23.1 [lib.container.requirements] seem not to consider fixed-size containers at all, as it says: "[containers] control allocation and deallocation of these objects [the contained objects] through constructors, destructors, *insert and erase* operations"

* max_size() isn't specified: the result is obvious but, technically, it relies on table 80: "size() of the largest possible container" which, again, doesn't seem to consider fixed size containers

[ 2009-05-29 Daniel adds: ]

  1. star bullet 1 ("what's the effect of calling assign(T&) on a zero-sized array?[..]");

    assign has been renamed to fill and the semantic of fill is now defined in terms of the free algorithm fill_n, which is well-defined for this situation.

  2. star bullet 3 ("it would be desiderable to have a static const data member..."):

    It seems that tuple_size<array<T, N> >::value as of 24.3.7.7 [array.tuple] does provide this functionality now.

[ 2009-07 Frankfurt ]

Alisdair to address by the next meeting, or declare NAD.

Moved to Tentatively NAD.

[ 2009 Santa Cruz: ]

Moved to NAD.

Proposed resolution:

[ Kona (2007): requirements on zero sized tr1::arrays and other details Issue 617: std::array is a sequence that doesn't satisfy the sequence requirements? Alisdair will prepare a paper. Proposed Disposition: Open ]


590(i). Type traits implementation latitude should be removed for C++0x

Section: 21 [meta], 99 [tr.meta.req] Status: NAD Editorial Submitter: Beman Dawes Opened: 2006-08-10 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [meta].

View all other issues in [meta].

View all issues with NAD Editorial status.

Discussion:

20.4.9 [lib.meta.req], Implementation requirements, provides latitude for type traits implementers that is not needed in C++0x. It includes the wording:

[Note: the latitude granted to implementers in this clause is temporary, and is expected to be removed in future revisions of this document. -- end note]

Note: N2157: Minor Modifications to the type traits Wording also has the intent of removing this wording from the WP.

Proposed resolution:

Remove 20.4.9 [lib.meta.req] in its entirety from the WP.

[ post-Oxford: Recommend NAD Editorial. This resolution is now in the current working draft. ]


591(i). Misleading "built-in

Section: 17.3.5.2 [numeric.limits.members] Status: NAD Editorial Submitter: whyglinux Opened: 2006-08-08 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [numeric.limits.members].

View all issues with NAD Editorial status.

Discussion:

18.2.1.2 numeric_limits members [lib.numeric.limits.members] Paragraph 7:

"For built-in integer types, the number of non-sign bits in the representation."

26.1 Numeric type requirements [lib.numeric.requirements] Footnote:

"In other words, value types. These include built-in arithmetic types, pointers, the library class complex, and instantiations of valarray for value types."

Integer types (which are bool, char, wchar_t, and the signed and unsigned integer types) and arithmetic types (which are integer and floating types) are all built-in types and thus there are no non-built-in (that is, user-defined) integer or arithmetic types. Since the redundant "built-in" in the above 2 sentences can mislead that there may be built-in or user-defined integer and arithmetic types (which is not correct), the "built-in" should be removed.

Proposed resolution:

18.2.1.2 numeric_limits members [lib.numeric.limits.members] Paragraph 7:

"For built-in integer types, the number of non-sign bits in the representation."

26.1 Numeric type requirements [lib.numeric.requirements] Footnote:

"In other words, value types. These include built-in arithmetic types, pointers, the library class complex, and instantiations of valarray for value types."

Rationale:

Recommend NAD / Editorial. The proposed resolution is accepted as editorial.


592(i). Incorrect treatment of rdbuf()->close() return type

Section: 31.10.3.4 [ifstream.members] Status: NAD Editorial Submitter: Christopher Kohlhoff Opened: 2006-08-17 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [ifstream.members].

View all issues with NAD Editorial status.

Discussion:

I just spotted a minor problem in 27.8.1.7 [lib.ifstream.members] para 4 and also 27.8.1.13 [lib.fstream.members] para 4. In both places it says:

void close();

Effects: Calls rdbuf()->close() and, if that function returns false, ...

However, basic_filebuf::close() (27.8.1.2) returns a pointer to the filebuf on success, null on failure, so I think it is meant to say "if that function returns a null pointer". Oddly, it is correct for basic_ofstream.

Proposed resolution:

Change 31.10.3.4 [ifstream.members], p5:

Effects: Calls rdbuf()->close() and, if that function fails (returns false a null pointer), calls setstate(failbit) (which may throw ios_base::failure (27.4.4.3)).

Change 31.10.5.4 [fstream.members], p5:

Effects: Calls rdbuf()->close() and, if that function fails (returns false a null pointer), calls setstate(failbit) (which may throw ios_base::failure (27.4.4.3)).

[ Kona (2007): Proposed Disposition: NAD, Editorial ]


597(i). [dec.tr] The notion of 'promotion' cannot be emulated by user-defined types.

Section: 3.2 [dec.tr::trdec.types.types] Status: NAD Submitter: Daveed Vandevoorde Opened: 2006-04-05 Last modified: 2016-01-31

Priority: Not Prioritized

View all other issues in [dec.tr::trdec.types.types].

View all issues with NAD status.

Discussion:

In a private email, Daveed writes:

I am not familiar with the C TR, but my guess is that the class type approach still won't match a built-in type approach because the notion of "promotion" cannot be emulated by user-defined types.

Here is an example:


         struct S {
           S(_Decimal32 const&);  // Converting constructor
         };
         void f(S);

         void f(_Decimal64);

         void g(_Decimal32 d) {
           f(d);
         }

If _Decimal32 is a built-in type, the call f(d) will likely resolve to f(_Decimal64) because that requires only a promotion, whereas f(S) requires a user-defined conversion.

If _Decimal32 is a class type, I think the call f(d) will be ambiguous because both the conversion to _Decimal64 and the conversion to S will be user-defined conversions with neither better than the other.

Robert comments:

In general, a library of arithmetic types cannot exactly emulate the behavior of the intrinsic numeric types. There are several ways to tell whether an implementation of the decimal types uses compiler intrinisics or a library. For example:

                 _Decimal32 d1;
                 d1.operator+=(5);  // If d1 is a builtin type, this won't compile.

In preparing the decimal TR, we have three options:

  1. require that the decimal types be class types
  2. require that the decimal types be builtin types, like float and double
  3. specify a library of class types, but allow enough implementor latitude that a conforming implementation could instead provide builtin types

We decided as a group to pursue option #3, but that approach implies that implementations may not agree on the semantics of certain use cases (first example, above), or on whether certain other cases are well-formed (second example). Another potentially important problem is that, under the present definition of POD, the decimal classes are not POD types, but builtins will be.

Note that neither example above implies any problems with respect to C-to-C++ compatibility, since neither example can be expressed in C.

[ 2009-07 Frankfurt ]

Decimal numeric types may either be builtin types or library types. We only intend to specify the common subset of behaviors of the two implementation approaches. The front matter of the Decimal TR says this explicitly.

Move to NAD.

Proposed resolution:


606(i). [dec.tr] allow narrowing conversions

Section: 3.2 [dec.tr::trdec.types.types] Status: NAD Submitter: Martin Sebor Opened: 2006-06-15 Last modified: 2016-01-31

Priority: Not Prioritized

View all other issues in [dec.tr::trdec.types.types].

View all issues with NAD status.

Discussion:

In c++std-lib-17205, Martin writes:

...was it a deliberate design choice to make narrowing assignments ill-formed while permitting narrowing compound assignments? For instance:

      decimal32 d32;
      decimal64 d64;

      d32 = 64;     // error
      d32 += 64;    // okay

In c++std-lib-17229, Robert responds:

It is a vestige of an old idea that I forgot to remove from the paper. Narrowing assignments should be permitted. The bug is that the converting constructors that cause narrowing should not be explicit. Thanks for pointing this out.

[ 2009-07 Frankfurt ]

The current state of the Decimal TR is the result of a deliberate design decision that has been examined many times.

Move to NAD.

Proposed resolution:

1. In "3.2.2 Class decimal32" synopsis, remove the explicit specifier from the narrowing conversions:

                // 3.2.2.2 conversion from floating-point type:
                explicit decimal32(decimal64 d64);
                explicit decimal32(decimal128 d128);

2. Do the same thing in "3.2.2.2. Conversion from floating-point type."

3. In "3.2.3 Class decimal64" synopsis, remove the explicit specifier from the narrowing conversion:

                // 3.2.3.2 conversion from floating-point type:
                explicit decimal64(decimal128 d128);

4. Do the same thing in "3.2.3.2. Conversion from floating-point type."

[ Redmond: We prefer explicit conversions for narrowing and implicit for widening. ]


614(i). std::string allocator requirements still inconsistent

Section: 23.4.3 [basic.string] Status: NAD Submitter: Bo Persson Opened: 2006-12-05 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [basic.string].

View all other issues in [basic.string].

View all issues with NAD status.

Discussion:

This is based on N2134, where 21.3.1/2 states: "... The Allocator object used shall be a copy of the Allocator object passed to the basic_string object's constructor or, if the constructor does not take an Allocator argument, a copy of a default-constructed Allocator object."

Section 21.3.2/1 lists two constructors:

basic_string(const basic_string<charT,traits,Allocator>& str );

basic_string(const basic_string<charT,traits,Allocator>& str ,
             size_type pos , size_type n = npos,
             const Allocator& a = Allocator());

and then says "In the first form, the Allocator value used is copied from str.get_allocator().", which isn't an option according to 21.3.1.

[ Batavia: We need blanket statement to the effect of: ]

  1. If an allocator is passed in, use it, or,
  2. If a string is passed in, use its allocator.

[ Review constructors and functions that return a string; make sure we follow these rules (substr, operator+, etc.). Howard to supply wording. ]

[ Bo adds: The new container constructor which takes only a size_type is not consistent with 24.2 [container.requirements], p9 which says in part: ]

All other constructors for these container types take an Allocator& argument (20.1.2), an allocator whose value type is the same as the container's value type. A copy of this argument is used for any memory allocation performed, by these constructors and by all member functions, during the lifetime of each container object.

[ post Bellevue: We re-confirm that the issue is real. Pablo will provide wording. ]

[ 2009-07 Frankfurt ]

Move to NAD.

Proposed resolution:


615(i). Inconsistencies in Section 21.4

Section: 23.5 [c.strings] Status: NAD Editorial Submitter: Bo Persson Opened: 2006-12-11 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [c.strings].

View all other issues in [c.strings].

View all issues with NAD Editorial status.

Discussion:

In the current draft N2134, 21.4/1 says

"Tables 59,228) 60, 61, 62,and 63 229) 230) describe headers <cctype>, <cwctype>, <cstring>, <cwchar>, and <cstdlib> (character conversions), respectively."

Here footnote 229 applies to table 62, not table 63.

Also, footnote 230 lists the new functions in table 63, "atoll, strtoll, strtoull, strtof, and strtold added by TR1". However, strtof is not present in table 63.

Proposed resolution:

Rationale:

Recommend NAD, editorial. Send to Pete.


626(i). new Remark clauses not documented

Section: 16.3.2.4 [structure.specifications] Status: NAD Editorial Submitter: Martin Sebor Opened: 2007-01-20 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [structure.specifications].

View all other issues in [structure.specifications].

View all issues with NAD Editorial status.

Discussion:

The Remark clauses newly introduced into the Working Paper (N2134) are not mentioned in 16.3.2.4 [structure.specifications] where we list the meaning of Effects, Requires, and other clauses (with the exception of Notes which are documented as informative in 16.3.2.2 [structure.summary], p2, and which they replace in many cases).

Propose add a bullet for Remarks along with a brief description.

[ Batavia: Alan and Pete to work. ]

[ Bellevue: Already resolved in current working paper. ]

Proposed resolution:


627(i). Low memory and exceptions

Section: 17.6.3.2 [new.delete.single] Status: NAD Submitter: P.J. Plauger Opened: 2007-01-23 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [new.delete.single].

View all other issues in [new.delete.single].

View all issues with NAD status.

Discussion:

I recognize the need for nothrow guarantees in the exception reporting mechanism, but I strongly believe that implementors also need an escape hatch when memory gets really low. (Like, there's not enough heap to construct and copy exception objects, or not enough stack to process the throw.) I'd like to think we can put this escape hatch in 17.6.3.2 [new.delete.single], operator new, but I'm not sure how to do it. We need more than a footnote, but the wording has to be a bit vague. The idea is that if new can't allocate something sufficiently small, it has the right to abort/call terminate/call unexpected.

[ Bellevue: NAD. 1.4p2 specifies a program must behave correctly "within its resource limits", so no further escape hatch is necessary. ]

Proposed resolution:


631(i). conflicting requirements for BinaryPredicate

Section: 27 [algorithms] Status: NAD Submitter: James Kanze Opened: 2007-01-31 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [algorithms].

View all other issues in [algorithms].

View all issues with NAD status.

Discussion:

The general requirements for BinaryPredicate (in 27 [algorithms]/8) contradict the implied specific requirements for some functions. In particular, it says that:

[...] if an algorithm takes BinaryPredicate binary_pred as its argument and first1 and first2 as its iterator arguments, it should work correctly in the construct if (binary_pred (*first1 , *first2 )){...}. BinaryPredicate always takes the first iterator type as its first argument, that is, in those cases when T value is part of the signature, it should work correctly in the context of if (binary_pred (*first1 , value)){...}.

In the description of upper_bound (27.8.4.3 [upper.bound]/2), however, the use is described as "!comp(value, e)", where e is an element of the sequence (a result of dereferencing *first).

In the description of lexicographical_compare, we have both "*first1 < *first2" and "*first2 < *first1" (which presumably implies "comp( *first1, *first2 )" and "comp( *first2, *first1 )".

Logically, the BinaryPredicate is used as an ordering relationship, with the semantics of "less than". Depending on the function, it may be used to determine equality, or any of the inequality relationships; doing this requires being able to use it with either parameter first. I would thus suggest that the requirement be:

Alternatively, one could specify an order for each function. IMHO, this would be more work for the committee, more work for the implementors, and of no real advantage for the user: some functions, such as lexicographical_compare or equal_range, will still require both functions, and it seems like a much easier rule to teach that both functions are always required, rather than to have a complicated list of when you only need one, and which one.

[ Toronto: Moved to Open. ConceptGCC seems to get lower_bound and upper_bound to work withoutt these changes. ]

[ 2009-07-28 Reopened by Alisdair. No longer solved by concepts. ]

[ 2009-10 Santa Cruz: ]

Move to Review. The small problem with the "iterator type" will be fixed. The cited functions (lower_bound, uppwer_bound, equal_range) don't actually use BinaryPredicate , and where it is used, it is consistent with [algorithm]/8, so the main complaint of the issue is moot.

[ 2010-01-16 Beman clarified wording. ]

[ 2010-01-31: Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Rationale:

[ post San Francisco: ]

Solved by N2759.

2010-01-31: The draft standard is well specified as is, and this specification is desired. Issues 556 and 870 solve the remaining unclearness regarding the meaning of BinaryPredicate.

Proposed resolution:

Change 27 [algorithms] paragraph 8 as indicated:

8 The BinaryPredicate parameter is used whenever an algorithm expects a function object that when applied to the result of dereferencing two corresponding iterators or to dereferencing an iterator and type T when T is part of the signature returns a value testable as true. BinaryPredicate always takes the first iterator value_type as one of its arguments; which argument is unspecified. In other words, if If an algorithm takes BinaryPredicate binary_pred as its argument and first1 and first2 as its iterator arguments, it should work correctly both in the construct if (binary_pred(*first1, *first2)){...} and if (binary_pred (*first2, *first1)){...}. BinaryPredicate always takes the first iterator type as its first argument, that is, in In those cases when T value is part of the signature, it should work correctly in the context of if (binary_pred(*first1, value)){...} and of if (binary_pred (value, *first1)){...}. binary_pred shall not apply any non-constant function through the dereferenced iterators. [Note: if the two types are not identical, and neither is convertable to the other, this may require that the BinaryPredicate be a functional object with two overloaded operator()() functions. — end note]


632(i). Time complexity of size() for std::set

Section: 24.2 [container.requirements] Status: NAD Submitter: Lionel B Opened: 2007-02-01 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [container.requirements].

View all issues with NAD status.

Discussion:

A recent news group discussion:

Anyone know if the Standard has anything to say about the time complexity of size() for std::set? I need to access a set's size (not to know if it is empty!) heavily during an algorithm and was thus wondering whether I'd be better off tracking the size "manually" or whether that'd be pointless.

That would be pointless. size() is O(1).

Nit: the standard says "should" have constant time. Implementations may take license to do worse. I know that some do this for std::list<> as a part of some trade-off with other operation.

I was aware of that, hence my reluctance to use size() for std::set.

However, this reason would not apply to std::set<> as far as I can see.

Ok, I guess the only option is to try it and see...

If I have any recommendation to the C++ Standards Committee it is that implementations must (not "should"!) document clearly[1], where known, the time complexity of *all* container access operations.

[1] In my case (gcc 4.1.1) I can't swear that the time complexity of size() for std::set is not documented... but if it is it's certainly well hidden away.

[ Kona (2007): This issue affects all the containers. We'd love to see a paper dealing with the broad issue. We think that the complexity of the size() member of every container -- except possibly list -- should be O(1). Alan has volunteered to provide wording. ]

[ Bellevue: ]

Mandating O(1) size will not fly, too many implementations would be invalidated. Alan to provide wording that toughens wording, but that does not absolutely mandate O(1).

[ Batavia (2009-05): ]

We observed that the wording "should" (in note a) has no effect. Howard prefers that O(1) size be mandated. It is not clear that this issue can be resolved to everyone's satisfaction, but Alan will provide wording nonetheless.

[ 2009-07 Frankfurt ]

Fixed by paper N2923.

Proposed resolution:


633(i). Return clause mentions undefined "type()"

Section: 22.10.17.3.6 [func.wrap.func.targ] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-02-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [func.wrap.func.targ].

View all issues with NAD Editorial status.

Discussion:

22.10.17.3.6 [func.wrap.func.targ], p4 says:

Returns: If type() == typeid(T), a pointer to the stored function target; otherwise a null pointer.

  1. There exists neither a type, a typedef type, nor member function type() in class template function nor in the global or std namespace.
  2. Assuming that type should have been target_type(), this description would lead to false results, if T = cv void due to returns clause 22.10.17.3.6 [func.wrap.func.targ], p1.

Proposed resolution:

Change 22.10.17.3.6 [func.wrap.func.targ], p4:

Returns: If type() target_type() == typeid(T) && typeid(T) != typeid(void), a pointer to the stored function target; otherwise a null pointer.

[ Pete: Agreed. It's editorial, so I'll fix it. ]


636(i). 26.5.2.3 valarray::operator[]

Section: 28.6.2.4 [valarray.access] Status: NAD Editorial Submitter: Bo Persson Opened: 2007-02-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [valarray.access].

View all issues with NAD Editorial status.

Discussion:

The signature of the const operator[] has been changed to return a const reference.

The description in paragraph 1 still says that the operator returns by value.

[ Pete recommends editorial fix. ]

Proposed resolution:


637(i). §[c.math]/10 inconsistent return values

Section: 28.7 [c.math] Status: NAD Editorial Submitter: Bo Persson Opened: 2007-02-13 Last modified: 2016-02-01

Priority: Not Prioritized

View all other issues in [c.math].

View all issues with NAD Editorial status.

Discussion:

28.7 [c.math], paragraph 10 has long lists of added signatures for float and long double functions. All the signatures have float/long double return values, which is inconsistent with some of the double functions they are supposed to overload.

Proposed resolution:

Change 28.7 [c.math], paragraph 10,

float int ilogb(float);
float long lrint(float);
float long lround(float);
float long long llrint(float);
float long long llround(float);

long double int ilogb(long double);
long double long lrint(long double);
long double long lround(long double);
long double long long llrint(long double);
long double long long llround(long double);

639(i). Still problems with exceptions during streambuf IO

Section: 31.7.5.3.3 [istream.extractors], 31.7.6.3.3 [ostream.inserters] Status: NAD Submitter: Daniel Krügler Opened: 2007-02-17 Last modified: 2017-04-22

Priority: Not Prioritized

View all other issues in [istream.extractors].

View all issues with NAD status.

Discussion:

There already exist two active DR's for the wording of [istream::extractors]/13 from 14882:2003(E), namely 64 and 413.

Even with these proposed corrections, already maintained in N2134, I have the feeling, that the current wording does still not properly handle the "exceptional" situation. The combination of para 14

"[..] Characters are extracted and inserted until any of the following occurs:

[..]

- an exception occurs (in which case the exception is caught)."

and 15

"If the function inserts no characters, it calls setstate(failbit), which may throw ios_base::failure (27.4.4.3). If it inserted no characters because it caught an exception thrown while extracting characters from *this and failbit is on in exceptions() (27.4.4.3), then the caught exception is rethrown."

both in N2134 seems to imply that any exception, which occurs *after* at least one character has been inserted is caught and lost for ever. It seems that even if failbit is on in exceptions() rethrow is not allowed due to the wording "If it inserted no characters because it caught an exception thrown while extracting".

Is this behaviour by design?

I would like to add that its output counterpart in 31.7.6.3.3 [ostream.inserters]/7-9 (also N2134) does not demonstrate such an exception-loss-behaviour. On the other side, I wonder concerning several subtle differences compared to input::

1) Paragraph 8 says at its end:

"- an exception occurs while getting a character from sb."

Note that there is nothing mentioned which would imply that such an exception will be caught compared to [istream::extractors]/14.

2) Paragraph 9 says:

"If the function inserts no characters, it calls setstate(failbit) (which may throw ios_base::failure (27.4.4.3)). If an exception was thrown while extracting a character, the function sets failbit in error state, and if failbit is on in exceptions() the caught exception is rethrown."

The sentence starting with "If an exception was thrown" seems to imply that such an exception *should* be caught before.

Proposed resolution:

(a) In [istream::extractors]/15 (N2134) change the sentence

If the function inserts no characters, it calls setstate(failbit), which may throw ios_base::failure (27.4.4.3). If it inserted no characters because it caught an exception thrown while extracting characters from *this an exception was thrown while extracting a character from *this, the function sets failbit in error state, and failbit is on in exceptions() (27.4.4.3), then the caught exception is rethrown.

(b) In 31.7.6.3.3 [ostream.inserters]/8 (N2134) change the sentence:

Gets characters from sb and inserts them in *this. Characters are read from sb and inserted until any of the following occurs:

Rationale:

This extractor is described as a formatted input function so the exception behavior is already specified. There is additional behavior described in this section that applies to the case in which failbit is set. This doesn't contradict the usual exception behavior for formatted input functions because that applies to the case in which badbit is set.


641(i). Editorial fix for 27.6.4 (N2134)

Section: 31.7.8 [ext.manip] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-02-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [ext.manip].

View all issues with NAD Editorial status.

Discussion:

The function f in para 4 (31.7.8 [ext.manip]) references an unknown strm in the following line:

mg.get(Iter(str.rdbuf()), Iter(), intl, strm, err, mon);

Proposed resolution:

Change 31.7.8 [ext.manip], p4:

mg.get(Iter(str.rdbuf()), Iter(), intl, strm, err, mon);

[ Oxford: Editorial. ]


642(i). Invalidated fstream footnotes in N2134

Section: 31.10.3.4 [ifstream.members], 31.10.4.4 [ofstream.members] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-02-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [ifstream.members].

View all issues with NAD Editorial status.

Discussion:

The standard wording of N2134 has extended the 14882:2003(E) wording for the ifstream/ofstream/fstream open function to fix a long standing problem, see 409.

Now it's properly written as

"If that function does not return a null pointer calls clear(), otherwise calls setstate(failbit)[..]"

instead of the previous

"If that function returns a null pointer, calls setstate(failbit)[..]

While the old footnotes saying

"A successful open does not change the error state."

where correct and important, they are invalid now for ifstream and ofstream (because clear *does* indeed modify the error state) and should be removed (Interestingly fstream itself never had these, although they where needed for that time).

Proposed resolution:

In 31.10.3.4 [ifstream.members], remove footnote:

334) A successful open does not change the error state.

In 31.10.4.4 [ofstream.members], remove footnote:

335) A successful open does not change the error state.


644(i). Possible typos in 'function' description

Section: 22.10.17.3 [func.wrap.func] Status: NAD Submitter: Bo Persson Opened: 2007-02-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [func.wrap.func].

View all issues with NAD status.

Discussion:

22.10.17.3 [func.wrap.func]

The note in paragraph 2 refers to 'undefined void operators', while the section declares a pair of operators returning bool.

[ Post-Sophia Antipolis: ]

Changed from Pending WP to Open. This issue was voted to WP at the same time the operators were changed from private to deleted. The two issues stepped on each other. What do we want the return type of these deleted functions to be?

[ 2009-05-02 Daniel adds: ]

I suggest harmonizing this issue with similar classes. E.g. in 20.3.2.3 [util.smartptr.weak] bool return values for

template <class Y> bool operator<(weak_ptr<Y> const&) const = delete;
template <class Y> bool operator<=(weak_ptr<Y> const&) const = delete;
template <class Y> bool operator>(weak_ptr<Y> const&) const = delete;
template <class Y> bool operator>=(weak_ptr<Y> const&) const = delete;

are used and basically all newer provided deleted copy assignment operators of type X use the canonical return type X& instead of void. Since the note mentioned in the issue description has now already been changed to

deleted overloads close possible hole in the type system

it seems to be of even lesser need to perform the change. Therefore I recommend declaring the issue as NAD.

[ Batavia (2009-05): ]

We agree with Daniel's recommendation.

Move to NAD.

Proposed resolution:

Change 22.10.17.3 [func.wrap.func]

...
private:
   // 22.10.17.3 [func.wrap.func], undefined operators:
   template<class Function2> bool void operator==(const function<Function2>&);
   template<class Function2> bool void operator!=(const function<Function2>&);
};

Change 22.10.17.3 [func.wrap.func]

template<class Function2> bool void operator==(const function<Function2>&);
template<class Function2> bool void operator!=(const function<Function2>&);

645(i). Missing members in match_results

Section: 32.9 [re.results] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-02-26 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [re.results].

View all issues with NAD Editorial status.

Discussion:

According to the description given in 32.9 [re.results]/2 the class template match_results "shall satisfy the requirements of a Sequence, [..], except that only operations defined for const-qualified Sequences are supported". Comparing the provided operations from 32.9 [re.results]/3 with the sequence/container tables 80 and 81 one recognizes the following missing operations:

1) The members

const_iterator rbegin() const;
const_iterator rend() const;

should exists because 23.1/10 demands these for containers (all sequences are containers) which support bidirectional iterators. Aren't these supported by match_result? This is not explicitely expressed, but it's somewhat implied by two arguments:

(a) Several typedefs delegate to iterator_traits<BidirectionalIterator>.

(b) The existence of const_reference operator[](size_type n) const implies even random-access iteration. I also suggest, that match_result should explicitly mention, which minimum iterator category is supported and if this does not include random-access the existence of operator[] is somewhat questionable.

2) The new "convenience" members

const_iterator cbegin() const;
const_iterator cend() const;
const_iterator crbegin() const;
const_iterator crend() const;

should be added according to tables 80/81.

Proposed resolution:

Add the following members to the match_results synopsis after end() in 32.9 [re.results] para 3:

const_iterator cbegin() const; 
const_iterator cend() const;

In section 32.9.5 [re.results.acc] change:

const_iterator begin() const;
const_iterator cbegin() const;

-7- Returns: A starting iterator that enumerates over all the sub-expressions stored in *this.

const_iterator end() const;
const_iterator cend() const;

-8- Returns: A terminating iterator that enumerates over all the sub-expressions stored in *this.

[ Kona (2007): Voted to adopt proposed wording in N2409 except removing the entry in the table container requirements. Moved to Review. ]

[ Bellevue: Proposed wording now in the WP. ]


647(i). Inconsistent regex_search params

Section: 32.10.3 [re.alg.search] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-02-26 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

32.10.3 [re.alg.search]/5 declares

template <class iterator, class charT, class traits>
bool regex_search(iterator first, iterator last,
                  const basic_regex<charT, traits>& e,
                  regex_constants::match_flag_type flags =
                      regex_constants::match_default);

where it's not explained, which iterator category the parameter iterator belongs to. This is inconsistent to the preceding declaration in the synopsis section 32.3 [re.syn], which says:

template <class BidirectionalIterator, class charT, class traits>
bool regex_search(BidirectionalIterator first, BidirectionalIterator last,
                  const basic_regex<charT, traits>& e,
                  regex_constants::match_flag_type flags =
                      regex_constants::match_default);

Proposed resolution:

In 32.10.3 [re.alg.search]/5 replace all three occurences of param "iterator" with "BidirectionalIterator"

template <class iterator BidirectionalIterator, class charT, class traits>
  bool regex_search(iterator BidirectionalIterator first, iterator BidirectionalIterator last, 
                    const basic_regex<charT, traits>& e, 
                    regex_constants::match_flag_type flags = 
                      regex_constants::match_default);

-6- Effects: Behaves "as if" by constructing an object what of type match_results<iterator BidirectionalIterator> and then returning the result of regex_search(first, last, what, e, flags).

Rationale:

Applied to working paper while issue was still in New status.


648(i). regex_iterator c'tor needs clarification/editorial fix

Section: 32.11.1.2 [re.regiter.cnstr] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-03-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

In 32.11.1.2 [re.regiter.cnstr]/2 the effects paragraph starts with:

Effects: Initializes begin and end to point to the beginning and the end of the target sequence, sets pregex to &re, sets flags to f,[..]

There are two issues with this description:

  1. The meaning of very first part of this quote is unclear, because there is no target sequence provided, instead there are given two parameters a and b, both of type BidirectionalIterator. The mentioned part does not explain what a and b represent.
  2. There does not exist any parameter f, but instead a parameter m in the constructor declaration, so this is actually an editorial fix.

Proposed resolution:

In 32.11.1.2 [re.regiter.cnstr]/2 change the above quoted part by

Effects: Initializes begin and end to point to the beginning and the end of the target sequence designated by the iterator range [a, b), sets pregex to &re, sets flags to f m, then calls regex_search(begin, end, match, *pregex, flags). If this call returns false the constructor sets *this to the end-of-sequence iterator.


649(i). Several typos in regex_token_iterator constructors

Section: 32.11.2.2 [re.tokiter.cnstr] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-03-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [re.tokiter.cnstr].

View all issues with NAD Editorial status.

Discussion:

In 32.11.2.2 [re.tokiter.cnstr]/1+2 both the constructor declaration and the following text shows some obvious typos:

1) The third constructor form is written as

template <std::size_t N>
  regex_token_iterator(BidirectionalIterator a, BidirectionalIterator b, 
                       const regex_type& re, 
                       const int (&submatches)[R], 
                       regex_constants::match_flag_type m = 
                         regex_constants::match_default);

where the dimensions of submatches are specified by an unknown value R, which should be N.

2) Paragraph 2 of the same section says in its last sentence:

The third constructor initializes the member subs to hold a copy of the sequence of integer values pointed to by the iterator range [&submatches, &submatches + R).

where again R must be replaced by N.

3) Paragraph 3 of the same section says in its first sentence:

Each constructor then sets N to 0, and position to position_iterator(a, b, re, f).

where a non-existing parameter "f" is mentioned, which must be replaced by the parameter "m".

Proposed resolution:

Change 32.11.2.2 [re.tokiter.cnstr]/1:

template <std::size_t N>
  regex_token_iterator(BidirectionalIterator a, BidirectionalIterator b, 
                       const regex_type& re, 
                       const int (&submatches)[R N], 
                       regex_constants::match_flag_type m = 
                         regex_constants::match_default);

Change 32.11.2.2 [re.tokiter.cnstr]/2:

Effects: The first constructor initializes the member subs to hold the single value submatch. The second constructor initializes the member subs to hold a copy of the argument submatches. The third constructor initializes the member subs to hold a copy of the sequence of integer values pointed to by the iterator range [&submatches, &submatches + R N).

Change 32.11.2.2 [re.tokiter.cnstr]/3:

Each constructor then sets N to 0, and position to position_iterator(a, b, re, f m). If position is not an end-of-sequence iterator the constructor sets result to the address of the current match. Otherwise if any of the values stored in subs is equal to -1 the constructor sets *this to a suffix iterator that points to the range [a, b), otherwise the constructor sets *this to an end-of-sequence iterator.


653(i). Library reserved names

Section: 2 [intro.refs] Status: NAD Submitter: Alisdair Meredith Opened: 2007-03-08 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [intro.refs].

View all issues with NAD status.

Discussion:

2 [intro.refs] Normative references

The following standards contain provisions which, through reference in this text, constitute provisions of this Interna- tional Standard. At the time of publication, the editions indicated were valid. All standards are subject to revision, and parties to agreements based on this International Standard are encouraged to investigate the possibility of applying the most recent editions of the standards indicated below. Members of IEC and ISO maintain registers of currently valid International Standards.

I'm not sure how many of those reserve naming patterns that might affect us, but I am equally sure I don't own a copy of any of these to check!

The point is to list the reserved naming patterns, rather than the individual names themselves - although we may want to list C keywords that are valid identifiers in C++ but likely to cause trouble in shared headers (e.g. restrict)

[ Kona (2007): Recommend NAD. No one has identified a specific defect, just the possibility of one. ]

[ Post-Kona: Alisdair request Open. A good example of the problem was a discussion of the system error proposal, where it was pointed out an all-caps identifier starting with a capital E conflicted with reserved macro names for both Posix and C. I had absolutely no idea of this rule, and suspect I was not the only one in the room.

Resolution will require someone with access to all the listed documents to research their respective name reservation rules, or people with access to specific documents add their rules to this issue until the list is complete. ]

[ Bellevue: Wording is aleady present in various standards, and no-one has come forward with wording. Suggest a formal paper rather than a defect report is the correct way to proceed. ]

Proposed resolution:


656(i). Typo in subtract_with_carry_engine declaration

Section: 28.5.2 [rand.synopsis] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-03-08 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.synopsis].

View all issues with NAD Editorial status.

Discussion:

28.5.2 [rand.synopsis] the header <random> synopsis contains an unreasonable closing curly brace inside the subtract_with_carry_engine declaration.

Proposed resolution:

Change the current declaration in 28.5.2 [rand.synopsis]

template <class UIntType, size_t w}, size_t s, size_t r>
class subtract_with_carry_engine;

[ Pete: Recommends editorial. ]


657(i). unclear requirement about header inclusion

Section: 16.4.3.2 [using.headers] Status: NAD Submitter: Gennaro Prota Opened: 2007-03-14 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [using.headers].

View all issues with NAD status.

Discussion:

16.4.3.2 [using.headers] states:

A translation unit shall include a header only outside of any external declaration or definition, [...]

I see three problems with this requirement:

  1. The C++ standard doesn't define what an "external declaration" or an "external definition" are (incidentally the C99 standard does, and has a sentence very similar to the above regarding header inclusion).

    I think the intent is that the #include directive shall lexically appear outside *any* declaration; instead, when the issue was pointed out on comp.std.c++ at least one poster interpreted "external declaration" as "declaration of an identifier with external linkage". If this were the correct interpretation, then the two inclusions below would be legal:

      // at global scope
      static void f()
      {
    # include <cstddef>
      }
    
      static void g()
      {
    # include <stddef.h>
      }
    

    (note that while the first example is unlikely to compile correctly, the second one may well do)

  2. as the sentence stands, violations will require a diagnostic; is this the intent? It was pointed out on comp.std.c++ (by several posters) that at least one way to ensure a diagnostic exists:

    [If there is an actual file for each header,] one simple way to implement this would be to insert a reserved identifier such as __begin_header at the start of each standard header. This reserved identifier would be ignored for all other purposes, except that, at the appropriate point in phase 7, if it is found inside an external definition, a diagnostic is generated. There's many other similar ways to achieve the same effect.

    --James Kuyper, on comp.std.c++

  3. is the term "header" meant to be limited to standard headers? Clause 17 is all about the library, but still the general question is interesting and affects one of the points in the explicit namespaces proposal (n1691):

    Those seeking to conveniently enable argument-dependent lookups for all operators within an explicit namespace could easily create a header file that does so:

        namespace mymath::
        {
            #include "using_ops.hpp"
        }
    

Proposed resolution:

Rationale:

We believe that the existing language does not cause any real confusion and any new formulation of the rules that we could come up with are unlikely to be better than what's already in the standard.


662(i). Inconsistent handling of incorrectly-placed thousands separators

Section: 30.4.3.2.3 [facet.num.get.virtuals] Status: NAD Submitter: Cosmin Truta Opened: 2007-04-05 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [facet.num.get.virtuals].

View all other issues in [facet.num.get.virtuals].

View all issues with NAD status.

Discussion:

From Section 30.4.3.2.3 [facet.num.get.virtuals], paragraphs 11 and 12, it is implied that the value read from a stream must be stored even if the placement of thousands separators does not conform to the grouping() specification from the numpunct facet. Since incorrectly-placed thousands separators are flagged as an extraction failure (by the means of failbit), we believe it is better not to store the value. A consistent strategy, in which any kind of extraction failure leaves the input item intact, is conceptually cleaner, is able to avoid corner-case traps, and is also more understandable from the programmer's point of view.

Here is a quote from "The C++ Programming Language (Special Edition)" by B. Stroustrup (Section D.4.2.3, pg. 897):

"If a value of the desired type could not be read, failbit is set in r. [...] An input operator will use r to determine how to set the state of its stream. If no error was encountered, the value read is assigned through v; otherwise, v is left unchanged."

This statement implies that rdstate() alone is sufficient to determine whether an extracted value is to be assigned to the input item val passed to do_get. However, this is in disagreement with the current C++ Standard. The above-mentioned assumption is true in all cases, except when there are mismatches in digit grouping. In the latter case, the parsed value is assigned to val, and, at the same time, err is assigned to ios_base::failbit (essentially "lying" about the success of the operation). Is this intentional? The current behavior raises both consistency and usability concerns.

Although digit grouping is outside the scope of scanf (on which the virtual methods of num_get are based), handling of grouping should be consistent with the overall behavior of scanf. The specification of scanf makes a distinction between input failures and matching failures, and yet both kinds of failures have no effect on the input items passed to scanf. A mismatch in digit grouping logically falls in the category of matching failures, and it would be more consistent, and less surprising to the user, to leave the input item intact whenever a failure is being signaled.

The extraction of bool is another example outside the scope of scanf, and yet consistent, even in the event of a successful extraction of a long but a failed conversion from long to bool.

Inconsistency is further aggravated by the fact that, when failbit is set, subsequent extraction operations are no-ops until failbit is explicitly cleared. Assuming that there is no explicit handling of rdstate() (as in cin>>i>>j) it is counter-intuitive to be able to extract an integer with mismatched digit grouping, but to be unable to extract another, properly-formatted integer that immediately follows.

Moreover, setting failbit, and selectively assigning a value to the input item, raises usability problems. Either the strategy of scanf (when there is no extracted value in case of failure), or the strategy of the strtol family (when there is always an extracted value, and there are well-defined defaults in case of a failure) are easy to understand and easy to use. On the other hand, if failbit alone cannot consistently make a difference between a failed extraction, and a successful but not-quite-correct extraction whose output happens to be the same as the previous value, the programmer must resort to implementation tricks. Consider the following example:

    int i = old_i;
    cin >> i;
    if (cin.fail())
        // can the value of i be trusted?
        // what does it mean if i == old_i?
        // ...

Last but not least, the current behvaior is not only confusing to the casual reader, but it has also been confusing to some book authors. Besides Stroustrup's book, other books (e.g. "Standard C++ IOStreams and Locales" by Langer and Kreft) are describing the same mistaken assumption. Although books are not to be used instead of the standard reference, the readers of these books, as well as the people who are generally familiar to scanf, are even more likely to misinterpret the standard, and expect the input items to remain intact when a failure occurs.

Proposed resolution:

Change 30.4.3.2.3 [facet.num.get.virtuals]:

Stage 3: The result of stage 2 processing can be one of

In the first case, Ddigit grouping is checked. That is, the positions of discarded separators is examined for consistency with use_facet<numpunct<charT> >(loc).grouping(). If they are not consistent then ios_base::failbit is assigned to err. Otherwise, the value that was converted in stage 2 is stored in val and ios_base::goodbit is stored in err.

Rationale:

post-Toronto: Changed from New to NAD at the request of the author. The preferred solution of N2327 makes this resolution obsolete.


663(i). Complexity Requirements

Section: 16.3.2.4 [structure.specifications] Status: NAD Submitter: Thomas Plum Opened: 2007-04-16 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [structure.specifications].

View all other issues in [structure.specifications].

View all issues with NAD status.

Discussion:

16.3.2.4 [structure.specifications] para 5 says

-5- Complexity requirements specified in the library clauses are upper bounds, and implementations that provide better complexity guarantees satisfy the requirements.

The following objection has been raised:

The library clauses suggest general guidelines regarding complexity, but we have been unable to discover any absolute hard-and-fast formulae for these requirements. Unless or until the Library group standardizes specific hard-and-fast formulae, we regard all the complexity requirements as subject to a "fudge factor" without any intrinsic upper bound.

[Plum ref _23213Y31 etc]

Proposed resolution:

Rationale:

Kona (2007): No specific instances of underspecification have been identified, and big-O notation always involves constant factors.


667(i). money_get's widened minus sign

Section: 30.4.7.2.2 [locale.money.get.virtuals] Status: NAD Submitter: Thomas Plum Opened: 2007-04-16 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.money.get.virtuals].

View all issues with NAD status.

Discussion:

30.4.7.2.2 [locale.money.get.virtuals], para 1 says:

The result is returned as an integral value stored in units or as a sequence of digits possibly preceded by a minus sign (as produced by ct.widen(c) where c is '-' or in the range from '0' through '9', inclusive) stored in digits.

The following objection has been raised:

Some implementations interpret this to mean that a facet derived from ctype<wchar_t> can provide its own member do_widen(char) which produces e.g. L'@' for the "widened" minus sign, and that the '@' symbol will appear in the resulting sequence of digits. Other implementations have assumed that one or more places in the standard permit the implementation to "hard-wire" L'-' as the "widened" minus sign. Are both interpretations permissible, or only one?

[Plum ref _222612Y14]

Furthermore: if ct.widen('9') produces L'X' (a non-digit), does a parse fail if a '9' appears in the subject string? [Plum ref _22263Y33]

[ Kona (2007): Bill and Dietmar to provide proposed wording. ]

[ post Bellevue: Bill adds: ]

The Standard is clear that the minus sign stored in digits is ct.widen('-'). The subject string must contain characters c in the set [-0123456789] which are translated by ct.widen(c) calls before being stored in digits; the widened characters are not relevant to the parsing of the subject string.

[ Batavia (2009-05): ]

We agree with Bill's comment above, in line with the first of the interpretations offered in the issue. Move to NAD.

Proposed resolution:


668(i). money_get's empty minus sign

Section: 30.4.7.2.2 [locale.money.get.virtuals] Status: NAD Submitter: Thomas Plum Opened: 2007-04-16 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.money.get.virtuals].

View all issues with NAD status.

Discussion:

30.4.7.2.2 [locale.money.get.virtuals], para 3 says:

If pos or neg is empty, the sign component is optional, and if no sign is detected, the result is given the sign that corresponds to the source of the empty string.

The following objection has been raised:

A negative_sign of "" means "there is no way to write a negative sign" not "any null sequence is a negative sign, so it's always there when you look for it".

[Plum ref _222612Y32]

[ Kona (2007): Bill to provide proposed wording and interpretation of existing wording. ]

Related to 669.

[ 2009-05-17 Howard adds: ]

I disagree that a negative_sign of "" means "there is no way to write a negative sign". The meaning requires the sentences of 30.4.7.2.2 [locale.money.get.virtuals] p3 following that quoted above to be taken into account:

-3- ... If pos or neg is empty, the sign component is optional, and if no sign is detected, the result is given the sign that corresponds to the source of the empty string. Otherwise, the character in the indicated position must match the first character of pos or neg, and the result is given the corresponding sign. If the first character of pos is equal to the first character of neg, or if both strings are empty, the result is given a positive sign.

So a negative_sign of "" means "there is no way to write a negative sign" only when positive_sign is also "". However when negative_sign is "" and postive_sign.size() > 0, then one writes a negative value by not writing the postive_sign in the position indicated by money_base::sign. For example:

pattern = {symbol, sign, value, none}
positive_sign = "+"
negative_sign = ""
$123   // a negative value, using optional sign
$+123  // a positive value
$-123  // a parse error

And:

pattern = {symbol, sign, value, none}
positive_sign = ""
negative_sign = ""
$123   // a positive value, no sign possible
$+123  // a parse error
$-123  // a parse error

And (regarding 669):

pattern = {symbol, sign, value, none}
positive_sign = "-"
negative_sign = "-"
$123   // a parse error, sign is mandatory
$+123  // a parse error
$-123  // a positive value

The text seems both unambiguous and clear to me. I recommend NAD for both this issue and 669. However I would have no objection to adding examples such as those above.

[ Batavia (2009-05): ]

This discussion applies equally to issue 669 (q.v.). Howard has added examples above, and recommends either NAD or a resolution that adds his (or similar) examples to the Working Paper.

Alan would like to rewrite paragraph 3.

We recommend moving to NAD. Anyone who feels strongly about adding the examples is invited to submit corresponding wording. We further recommend issue 669 be handled identically.

[ 2009-07-14 Alan reopens with improved wording. ]

[ 2009-07 Frankfurt ]

No consensus for closing as NAD. Leave in Review.

[ 2009-10 Santa Cruz: ]

NAD. Agreed that the original assessment as NAD was correct.

Proposed resolution:

Change 30.4.7.2.2 [locale.money.get.virtuals] p3:

-3- If the first character (if any) in the string pos returned by mp.positive_sign() or the string neg returned by mp.negative_sign() is recognized in the position indicated by sign in the format pattern, it is consumed and any remaining characters in the string are required after all the other format components. [Example: If showbase is off, then for a neg value of "()" and a currency symbol of "L", in "(100 L)" the "L" is consumed; but if neg is "-", the "L" in "-100 L" is not consumed. -- end example] If pos or neg is empty, the sign component is optional, and if no sign is detected, the result is given the sign that corresponds to the source of the empty string. Otherwise, the character in the indicated position must match the first character of pos or neg, and the result is given the corresponding sign. If the first character of pos is equal to the first character of neg, or if both strings are empty, the result is given a positive sign. The sign pattern strings pos and neg are returned by mp.positive_sign() and mp.negative_sign() respectively. A sign pattern is matched if its first character is recognized in s in the position indicated by sign in the format pattern, or if the pattern is empty and there is no sign recognized in s. A match is required to occur. If both patterns are matched, the result is given a positive sign, otherwise the result is given the sign corresponding to the matched pattern. If the pattern contains more than one character, the characters after the first must be matched in s after all other format components. If any sign characters are matched, s is consumed up to and including those characters. [Example: If showbase is off, then for a neg value of "()" and a currency symbol of "L", in "(100 L)" the entire string is consumed; but for a neg value of "-", in "-100 L", the string is consumed through the second "0" (the space and "L" are not consumed). — end example]


669(i). Equivalent postive and negative signs in money_get

Section: 30.4.7.2.2 [locale.money.get.virtuals] Status: NAD Submitter: Thomas Plum Opened: 2007-04-16 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.money.get.virtuals].

View all issues with NAD status.

Discussion:

30.4.7.2.2 [locale.money.get.virtuals], para 3 sentence 4 says:

If the first character of pos is equal to the first character of neg, or if both strings are empty, the result is given a positive sign.

One interpretation is that an input sequence must match either the positive pattern or the negative pattern, and then in either event it is interpreted as positive. The following objections has been raised:

The input can successfully match only a positive sign, so the negative pattern is an unsuccessful match.

[Plum ref _222612Y34, 222612Y51b]

[ Bill to provide proposed wording and interpretation of existing wording. ]

[ 2009-05-17 See Howard's comments in related issue 668. ]

[ Batavia (2009-05): ]

This discussion applies equally to issue 668 (q.v.). Howard has added examples there, and recommends either NAD or a resolution that adds his (or similar) examples to the Working Paper.

We recommend moving to NAD. Anyone who feels strongly about adding the examples is invited to submit corresponding wording. We further recommend issue 668 be handled identically.

Proposed resolution:


670(i). money_base::pattern and space

Section: 30.4.7.4 [locale.moneypunct] Status: Dup Submitter: Thomas Plum Opened: 2007-04-16 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [locale.moneypunct].

View all issues with Dup status.

Duplicate of: 836

Discussion:

30.4.7.4 [locale.moneypunct], para 2 says:

The value space indicates that at least one space is required at that position.

The following objection has been raised:

Whitespace is optional when matching space. (See 30.4.7.2.2 [locale.money.get.virtuals], para 2.)

[Plum ref _22263Y22]

[ Kona (2007): Bill to provide proposed wording. We agree that C++03 is ambiguous, and that we want C++0X to say "space" means 0 or more whitespace characters on input. ]

Proposed resolution:


683(i). regex_token_iterator summary error

Section: 32.11.2 [re.tokiter] Status: NAD Editorial Submitter: Eric Niebler Opened: 2007-06-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [re.tokiter].

View all issues with NAD Editorial status.

Discussion:

32.11.2 [re.tokiter], p3 says:

After it is constructed, the iterator finds and stores a value match_results<BidirectionalIterator> position and sets the internal count N to zero.

Should read:

After it is constructed, the iterator finds and stores a value match_resultsregex_iterator<BidirectionalIterator, charT, traits> position and sets the internal count N to zero.

[ John adds: ]

Yep, looks like a typo/administrative fix to me.

Proposed resolution:


684(i). Unclear which members of match_results should be used in comparison

Section: 32.9 [re.results] Status: NAD Editorial Submitter: Nozomu Katoo Opened: 2007-05-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [re.results].

View all issues with NAD Editorial status.

Discussion:

In 32.3 [re.syn] of N2284, two template functions are declared here:

// 28.10, class template match_results: 
  <snip>
// match_results comparisons 
  template <class BidirectionalIterator, class Allocator> 
    bool operator== (const match_results<BidirectionalIterator, Allocator>& m1, 
                     const match_results<BidirectionalIterator, Allocator>& m2); 
  template <class BidirectionalIterator, class Allocator> 
    bool operator!= (const match_results<BidirectionalIterator, Allocator>& m1, 
                     const match_results<BidirectionalIterator, Allocator>& m2); 

// 28.10.6, match_results swap:

But the details of these two bool operator functions (i.e., which members of match_results should be used in comparison) are not described in any following sections.

[ John adds: ]

That looks like a bug: operator== should return true only if the two objects refer to the same match - ie if one object was constructed as a copy of the other.

[ Kona (2007): Bill and Pete to add minor wording to that proposed in N2409. ]

Proposed resolution:

Add a new section after 32.9.8 [re.results.swap], which reads:

28.10.7 match_results non-member functions.

template<class BidirectionalIterator, class Allocator> 
  bool operator==(const match_results<BidirectionalIterator, Allocator>& m1, 
                  const match_results<BidirectionalIterator, Allocator>& m2);

Returns: true only if the two objects refer to the same match.

template<class BidirectionalIterator, class Allocator> 
  bool operator!=(const match_results<BidirectionalIterator, Allocator>& m1, 
                  const match_results<BidirectionalIterator, Allocator>& m2);

Returns: !(m1 == m2).

template<class BidirectionalIterator, class Allocator> 
  void swap(match_results<BidirectionalIterator, Allocator>& m1, 
            match_results<BidirectionalIterator, Allocator>& m2);

Returns: m1.swap(m2).

[ Bellevue: Proposed wording now in WP. ]


686(i). unique_ptr and shared_ptr fail to specify non-convertibility to int for unspecified-bool-type

Section: 20.3.1.3.5 [unique.ptr.single.observers], 20.3.2.2.6 [util.smartptr.shared.obs] Status: NAD Submitter: Beman Dawes Opened: 2007-06-14 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [unique.ptr.single.observers].

View all issues with NAD status.

Discussion:

The standard library uses the operator unspecified-bool-type() const idiom in five places. In three of those places (22.10.17.3.4 [func.wrap.func.cap], function capacity for example) the returned value is constrained to disallow unintended conversions to int. The standardese is

The return type shall not be convertible to int.

This constraint is omitted for unique_ptr and shared_ptr. It should be added for those.

[ Bellevue: ]

Close as NAD. Accepting paper N2435 makes it irrelevant.

Proposed resolution:

To the Returns paragraph for operator unspecified-bool-type() const of 20.3.1.3.5 [unique.ptr.single.observers] paragraph 11 and 20.3.2.2.6 [util.smartptr.shared.obs] paragraph 16, add the sentence:

The return type shall not be convertible to int.

[ Kona (2007): Uncertain if nullptr will address this issue. ]


690(i). abs(long long) should return long long

Section: 28.7 [c.math] Status: NAD Editorial Submitter: Niels Dekker Opened: 2007-06-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [c.math].

View all issues with NAD Editorial status.

Discussion:

Quoting the latest draft (n2135), 28.7 [c.math]:

The added signatures are:

long abs(long); // labs()
long abs(long long); // llabs()

Shouldn't abs(long long) have long long as return type?

Proposed resolution:

Change 28.7 [c.math]:

long long abs(long long); // llabs()

Rationale:

Had already been fixed in the WP by the time the LWG reviewed this.


701(i). [tr1] assoc laguerre poly's

Section: 5.2.1.1 [tr1::tr.num.sf.Lnm] Status: NAD Submitter: Christopher Crawford Opened: 2007-06-30 Last modified: 2016-02-01

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

I see that the definition the associated Laguerre polynomials 5.2.1.1 [tr1::tr.num.sf.Lnm] has been corrected since N1687. However, the draft standard only specifies ranks of integer value m, while the associated Laguerre polynomials are actually valid for real values of m > -1. In the case of non-integer values of m, the definition Ln(m) = (1/n!)exx-m (d/dx)n (e-xxm+n) must be used, which also holds for integer values of m. See Abramowitz & Stegun, 22.11.6 for the general case, and 22.5.16-17 for the integer case. In fact fractional values are most commonly used in physics, for example to m = +/- 1/2 to describe the harmonic oscillator in 1 dimension, and 1/2, 3/2, 5/2, ... in 3 dimensions.

If I am correct, the calculation of the more general case is no more difficult, and is in fact the function implemented in the GNU Scientific Library. I would urge you to consider upgrading the standard, either adding extra functions for real m or switching the current ones to double.

[ Batavia (2009-05): ]

We understand the issue, and have opted not to extend as recommended.

Move to NAD.

Proposed resolution:


702(i). [tr1] Restriction in associated Legendre functions

Section: 5.2.1.2 [tr1::tr.num.sf.Plm] Status: NAD Submitter: Christopher Crawford Opened: 2007-06-30 Last modified: 2016-02-01

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

One other small thing, in 5.2.1.2 [tr1::tr.num.sf.Plm], the restriction should be |x| <= 1, not x >= 0.

[ Batavia (2009-05): ]

The error has been corrected in the pending IS.

Move to NAD.

Proposed resolution:


707(i). null pointer constant for exception_ptr

Section: 17.9.7 [propagation] Status: NAD Submitter: Jens Maurer Opened: 2007-07-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [propagation].

View all issues with NAD status.

Discussion:

From the Toronto Core wiki:

What do you mean by "null pointer constant"? How do you guarantee that exception_ptr() == 1 doesn't work? Do you even want to prevent that? What's the semantics? What about void *p = 0; exception_ptr() == p? Maybe disallow those in the interface, but how do you do that with portable C++? Could specify just "make it work".

Peter's response:

null pointer constant as defined in 7.3.12 [conv.ptr]. Intent is "just make it work", can be implemented as assignment operator taking a unique pointer to member, as in the unspecified bool type idiom.

[ Bellevue: ]

Original implementation was possible using the "unspecified-null-pointer" idiom, similar to unspecified-bool.

Even simpler now with nullptr_t.

NAD Rationale : null pointer constant is a perfectly defined term, and while API is clearly implementable there is no need to spell out implementation details.

Proposed resolution:


708(i). Locales need to be per thread and updated for POSIX changes

Section: 30 [localization] Status: NAD Submitter: Peter Dimov Opened: 2007-07-28 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [localization].

View all issues with NAD status.

Discussion:

The POSIX "Extended API Set Part 4,"

http://www.opengroup.org/sib/details.tpl?id=C065

introduces extensions to the C locale mechanism that allow multiple concurrent locales to be used in the same application by introducing a type locale_t that is very similar to std::locale, and a number of _l functions that make use of it.

The global locale (set by setlocale) is now specified to be per- process. If a thread does not call uselocale, the global locale is in effect for that thread. It can install a per-thread locale by using uselocale.

There is also a nice querylocale mechanism by which one can obtain the name (such as "de_DE") for a specific facet, even for combined locales, with no std::locale equivalent.

std::locale should be harmonized with the new POSIX locale_t mechanism and provide equivalents for uselocale and querylocale.

[ Kona (2007): Bill and Nick to provide wording. ]

[ San Francisco: Bill and Nick still intend to provide wording, but this is a part of the task to be addressed by the group that will look into issue 860. ]

[ 2009-07 Frankfurt: ]

It's our intention to stay in sync with WG14. If WG14 makes a decision that requires a change in WG21 the issue will be reopened.

Move to NAD Future.

[LEWG Kona 2017]

Recommend NAD: uselocale() is bad; pass locales around as objects

[Kona 2019]

Jonathan points out: The standard already allows per-thread locales, see [locale] p9.

Proposed resolution:


717(i). Incomplete valarray::operator[] specification in [valarray.access]

Section: 28.6.2.4 [valarray.access] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-08-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [valarray.access].

View all issues with NAD Editorial status.

Discussion:

Since the return type of valarray's operator[] const overload has been changed to const T& as described in 389 several paragraphs of the section 28.6.2.4 [valarray.access] are now incompletely specified, because many requirements and guarantees should now also apply to the const overload. Most notably, the address and reference guarantees should be extended to the const overload case.

Proposed resolution:

Change 28.6.2.4 [valarray.access]:

-1- When applied to a constant array, the subscript operator returns a reference to the corresponding element of the array. When applied to a non-constant array, tThe subscript operator returns a reference to the corresponding element of the array.

-3- The expression &a[i+j] == &a[i] + j evaluates as true for all size_t i and size_t j such that i+j is less than the length of the non-constant array a.

-4- Likewise, the expression &a[i] != &b[j] evaluates as true for any two non-constant arrays a and b and for any size_t i and size_t j such that i is less than the length of a and j is less than the length of b. This property indicates an absence of aliasing and may be used to advantage by optimizing compilers.281)

-5- The reference returned by the subscript operator for an non-constant array is guaranteed to be valid until the member function resize(size_t, T) (26.5.2.7) is called for that array or until the lifetime of that array ends, whichever happens first.


718(i). basic_string is not a sequence

Section: 23.4.3 [basic.string] Status: NAD Editorial Submitter: Bo Persson Opened: 2007-08-18 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [basic.string].

View all other issues in [basic.string].

View all issues with NAD Editorial status.

Discussion:

Paragraph 23.4.3 [basic.string]/3 states:

The class template basic_string conforms to the requirements for a Sequence (23.1.1) and for a Reversible Container (23.1).

First of all, 24.2.4 [sequence.reqmts] is no longer "Sequence" but "Sequence container". Secondly, after the resent changes to containers (emplace, push_back, const_iterator parameters to insert and erase), basic_string is not even close to conform to the current requirements.

[ Bellevue: ]

General consensus is to suggest option 2.

[ 2009-07 Frankfurt: ]

Move to NAD Editorial

Proposed resolution:

Remove this sentence, in recognition of the fact that basic_string is not just a vector-light for literal types, but something quite different, a string abstraction in its own right.


721(i). wstring_convert inconsistensies

Section: D.27.2 [depr.conversions.string] Status: NAD Submitter: Bo Persson Opened: 2007-08-27 Last modified: 2017-04-22

Priority: Not Prioritized

View other active issues in [depr.conversions.string].

View all other issues in [depr.conversions.string].

View all issues with NAD status.

Discussion:

Paragraph 3 says that the Codecvt template parameter shall meet the requirements of std::codecvt, even though std::codecvt itself cannot be used (because of a protected destructor).

How are we going to explain this code to beginning programmers?

template<class I, class E, class S>
struct codecvt : std::codecvt<I, E, S>
{
    ~codecvt()
    { }
};

void main()
{
    std::wstring_convert<codecvt<wchar_t, char, std::mbstate_t> > compiles_ok;
    
    std::wstring_convert<std::codecvt<wchar_t, char, std::mbstate_t> >   not_ok;
}

[ San Francisco: ]

Bill will propose a resolution.

[ 2009-07 Frankfurt: ]

codecvt isn't intended for beginning programmers. This is a regrettable consequence of the original design of the facet.

Move to NAD.

Proposed resolution:


725(i). Optional sequence container requirements column label

Section: 24.2.4 [sequence.reqmts] Status: NAD Editorial Submitter: David Abrahams Opened: 2007-09-16 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [sequence.reqmts].

View all other issues in [sequence.reqmts].

View all issues with NAD Editorial status.

Discussion:

Table 90: (Optional sequence container operations) states the "assertion note pre/post-condition" of operator[] to be

*(a.begin() + n)

Surely that's meant to be "operational semantics?"

Proposed resolution:

Table 90: Optional sequence container operations
expression return type assertion/note
pre/post-condition

operational semantics
container

726(i). Missing regex_replace() overloads

Section: 32.10.4 [re.alg.replace] Status: NAD Submitter: Stephan T. Lavavej Opened: 2007-09-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [re.alg.replace].

View all issues with NAD status.

Discussion:

Two overloads of regex_replace() are currently provided:

template <class OutputIterator, class BidirectionalIterator, 
    class traits, class charT> 
  OutputIterator 
  regex_replace(OutputIterator out, 
                BidirectionalIterator first, BidirectionalIterator last, 
                const basic_regex<charT, traits>& e, 
                const basic_string<charT>& fmt, 
                regex_constants::match_flag_type flags = 
                  regex_constants::match_default);
 
template <class traits, class charT> 
  basic_string<charT> 
  regex_replace(const basic_string<charT>& s, 
                const basic_regex<charT, traits>& e, 
                const basic_string<charT>& fmt, 
                regex_constants::match_flag_type flags = 
                  regex_constants::match_default);
  1. Overloads taking const charT * are provided for regex_match() and regex_search(), but not regex_replace(). This is inconsistent.
  2. The absence of const charT * overloads prevents ordinary-looking code from compiling, such as:

    const string s("kitten");
    const regex r("en");
    cout << regex_replace(s, r, "y") << endl;
    

    The compiler error message will be something like "could not deduce template argument for 'const std::basic_string<_Elem> &' from 'const char[1]'".

    Users expect that anything taking a basic_string<charT> can also take a const charT *. In their own code, when they write a function taking std::string (or std::wstring), they can pass a const char * (or const wchar_t *), thanks to basic_string's implicit constructor. Because the regex algorithms are templated on charT, they can't rely on basic_string's implicit constructor (as the compiler error message indicates, template argument deduction fails first).

    If a user figures out what the compiler error message means, workarounds are available - but they are all verbose. Explicit template arguments could be given to regex_replace(), allowing basic_string's implicit constructor to be invoked - but charT is the last template argument, not the first, so this would be extremely verbose. Therefore, constructing a basic_string from each C string is the simplest workaround.

  3. There is an efficiency consideration: constructing basic_strings can impose performance costs that could be avoided by a library implementation taking C strings and dealing with them directly. (Currently, for replacement sources, C strings can be converted into iterator pairs at the cost of verbosity, but for format strings, there is no way to avoid constructing a basic_string.)

[ Sophia Antipolis: ]

We note that Boost already has these overloads. However, the proposed wording is provided only for 32.10.4 [re.alg.replace]; wording is needed for the synopsis as well. We also note that this has impact on match_results::format, which may require further overloads.

[ 2009-07 Frankfurt: ]

Daniel to tweak for us.

[ 2009-07-25 Daniel tweaks both this issue and 727. ]

This is solved by the proposed resolution of 727.

[ 2009-10 Santa Cruz: ]

Leave Open. Though we believe this is solved by the proposed resolution to 727.

[ 2010-01-27 Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Rationale:

Solved by 727.

Proposed resolution:

Provide additional overloads for regex_replace(): one additional overload of the iterator-based form (taking const charT* fmt), and three additional overloads of the convenience form (one taking const charT* str, another taking const charT* fmt, and the third taking both const charT* str and const charT* fmt). 32.10.4 [re.alg.replace]:

template <class OutputIterator, class BidirectionalIterator, 
    class traits, class charT> 
  OutputIterator 
  regex_replace(OutputIterator out, 
                BidirectionalIterator first, BidirectionalIterator last, 
                const basic_regex<charT, traits>& e, 
                const basic_string<charT>& fmt, 
                regex_constants::match_flag_type flags = 
                  regex_constants::match_default);

template <class OutputIterator, class BidirectionalIterator, 
    class traits, class charT> 
  OutputIterator 
  regex_replace(OutputIterator out, 
                BidirectionalIterator first, BidirectionalIterator last, 
                const basic_regex<charT, traits>& e, 
                const charT* fmt, 
                regex_constants::match_flag_type flags = 
                  regex_constants::match_default);

...

template <class traits, class charT> 
  basic_string<charT> 
  regex_replace(const basic_string<charT>& s, 
                const basic_regex<charT, traits>& e, 
                const basic_string<charT>& fmt, 
                regex_constants::match_flag_type flags = 
                  regex_constants::match_default);

template <class traits, class charT> 
  basic_string<charT> 
  regex_replace(const basic_string<charT>& s, 
                const basic_regex<charT, traits>& e, 
                const charT* fmt, 
                regex_constants::match_flag_type flags = 
                  regex_constants::match_default);

template <class traits, class charT> 
  basic_string<charT> 
  regex_replace(const charT* s, 
                const basic_regex<charT, traits>& e, 
                const basic_string<charT>& fmt, 
                regex_constants::match_flag_type flags = 
                  regex_constants::match_default);

template <class traits, class charT> 
  basic_string<charT> 
  regex_replace(const charT* s, 
                const basic_regex<charT, traits>& e, 
                const charT* fmt, 
                regex_constants::match_flag_type flags = 
                  regex_constants::match_default);

729(i). Problem in [rand.req.eng]/3

Section: 28.5.3.4 [rand.req.eng] Status: NAD Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.req.eng].

View all issues with NAD status.

Discussion:

The 3rd table row in 28.5.3.4 [rand.req.eng]/3 requires random number engines to accept any arithmetic type as a seed, which is then casted to the engine's result_type and subsequently used for seeding the state of the engine. The requirement stated as "Creates an engine with initial state determined by static_cast<X::result_type>(s)" forces random number engines to either use a seeding method that completely depends on the result_type (see the discussion of seeding for the mersenne_twister_engine in point T2 above) or at least to throw away "bits of randomness" in the seed value if the result_type is smaller than the seed type. This seems to be inappropriate for many modern random number generators, in particular F2-linear or cryptographic ones, which operate on an internal bit array that in principle is independent of the type of numbers returned.

Possible resolution: I propose to change the wording to a version similar to "Creates an engine with initial state determined by static_cast<UintType>(s), where UintType is an implementation specific unsigned integer type."

Additionally, the definition of s in 28.5.3.4 [rand.req.eng]/1 c) could be restricted to unsigned integer types.

Similarly, the type of the seed in 28.5.3.5 [rand.req.adapt]/3 e) could be left unspecified.

See N2424 for further discussion.

[ Stephan Tolksdorf adds pre-Bellevue: ]

In reply to the discussion in N2424 regarding this issue:

The descriptions of all engines and engine adaptors given in sections 28.5.4 [rand.eng] and 28.5.5 [rand.adapt] already specify the concrete types of the integer arguments for seeding. Hence, relaxing the general requirement in 28.5.3.4 [rand.req.eng] would not affect portability and reproducibility of the standard library. Furthermore, it is not clear to me what exactly the guarantee "with initial state determined by static_cast<X::result_type>(s)" is useful for. On the other hand, relaxing the requirement would allow developers to implement other random number engines that do not have to cast all arithmetic seed arguments to their result_types.

[ Bellevue: ]

Propose close NAD for the reasons given in N2424.

Proposed resolution:

See N2424 for further discussion.

[ Stephan Tolksdorf adds pre-Bellevue: ]

Change row 3 of table 105 "Random number engine requirements" in 28.5.3.4 [rand.req.eng]/3

Creates an engine with initial state determined by static_cast<X::result_type>(s)

Similarly, change 28.5.3.5 [rand.req.adapt]/3 e)

When X::X is invoked with an X::result_type value s of arithmetic type (3.9.1), ...


730(i). Comment on [rand.req.adapt]/3 e)

Section: 28.5.3.5 [rand.req.adapt] Status: NAD Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

If an engine adaptor is invoked with an argument of type seed_seq, then all base engines are specified to be seeded with this seed_seq. As seed_seq's randomization method is qualified as constant, this procedure will ef fectively initialize all base engines with the same seed (though the resulting state might still dif fer to a certain degree if the engines are of different types). It is not clear whether this mode of operation is in general appropriate, hence -- as far as the stated requirements are of general nature and not just specific to the engine adaptors provided by the library -- it might be better to leave the behaviour unspecified, since the current definition of seed_seq does not allow for a generally satisfying specification.

Posssible resolution: [As above]

See N2424 for further discussion.

[ Bellevue: ]

Close NAD for the reasons given in N2424.

Proposed resolution:

See N2424 for the proposed resolution.


731(i). proposal for a customizable seed_seq

Section: 28.5.8.1 [rand.util.seedseq] Status: NAD Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.util.seedseq].

View all issues with NAD status.

Discussion:

The proper way to seed random number engines seems to be the most frequently discussed issue of the 28.5 [rand] proposal. While the new seed_seq approach is already rather general and probably sufficient for most situations, it is unlikely to be optimal in every case (one problem was pointed out in point T5 above). In some situations it might, for instance, be better to seed the state with a cryptographic generator.

In my opinion this is a pretty strong argument for extending the standard with a simple facility to customize the seeding procedure. This could, for example, be done with the following minimal changes:

Possible resolution:

  1. Turn the interface specification of 28.5.8.1 [rand.util.seedseq]/2 into a "SeedSeq" requirement, where the exact behaviour of the constructors and the randomize method are left unspecified and where the const qualification for randomize is removed. Classes implementing this interface are additionally required to specialize the traits class in c).
  2. Provide the class seed_seq as a default implementation of the SeedSeq interface.
  3. Supplement the seed_seq with a traits class

    template <typename T> 
    struct is_seed_seq { static const bool value = false; }
    

    and the specialization

    template <> 
    struct is_seed_seq<seed_seq> { static const bool value = true; }
    

    which users can supplement with further specializations.

  4. Change 28.5.3.4 [rand.req.eng]/1 d) to "q is an lvalue of a type that fulfils the SeedSeq requirements", and modify the constructors and seed methods in 28.5.4 [rand.eng] appropriately (the actual implementation could be done using the SFINAE technique).

[ Bellevue: ]

See N2424. Close NAD but note that "conceptizing" the library may cause this problem to be solved by that route.

Proposed resolution:

See N2424 for the proposed resolution.


733(i). Comment on [rand.req.dist]/9

Section: 28.5.3.6 [rand.req.dist] Status: NAD Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.req.dist].

View all issues with NAD status.

Discussion:

The requirement "P shall have a declaration of the form typedef X distribution_- type" effectively makes the use of inheritance for implementing distributions very inconvenient, because the child of a distribution class in general will not satisfy this requirement. In my opinion the benefits of having a typedef in the parameter class pointing back to the distribution class are not worth the hassle this requirement causes. [In my code base I never made use of the nested typedef but on several occasions could have profited from being able to use simple inheritance for the implementation of a distribution class.]

Proposed resolution: I propose to drop this requirement.

[ Bellevue: ]

Close NAD for the reasons given in N2424. In practice it is not inconvenient to meet these requirements.

Proposed resolution:

See N2424 for the proposed resolution.


735(i). Unfortunate naming

Section: 28.5.9.3.2 [rand.dist.bern.bin], 28.5.9.3.4 [rand.dist.bern.negbin] Status: NAD Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

In my opinion the choice of name for the t parameter of the binomial_distribution is very unfortunate. In virtually every internet reference, book and software implementation this parameter is called n instead, see for example Wikipedia, Mathworld, Evans et al. (1993) Statistical Distributions, 2nd E., Wiley, p. 38, the R statistical computing language, p. 926, Mathematica and Matlab.

Similarly, the choice of k for the parameter of the negative binomial distributions is rather unusual. The most common choice for the negative binomial distribution seems to be r instead.

Choosing unusual names for the parameters causes confusion among users and makes the interface unnecessarily inconvenient to use.

Possible resolution: For these reasons, I propose to change the name of the respective parameters to n and r.

[ Bellevue: ]

In N2424. NAD It has been around for a while. It is hardly universal, there is prior art, and this would confuse people.

Proposed resolution:

See N2424 for the proposed resolution.


736(i). Comment on [rand.dist.samp.discrete]

Section: 28.5.9.6.1 [rand.dist.samp.discrete] Status: NAD Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.dist.samp.discrete].

View all issues with NAD status.

Discussion:

  1. The specification for discrete_distribution requires the member probabilities() to return a vector of standardized probabilities, which forces the implementation every time to divide each probability by the sum of all probabilities, as the sum will in practice almost never be exactly 1.0. This is unnecessarily inef ficient as the implementation would otherwise not need to compute the standardized probabilities at all and could instead work with the non-standardized probabilities and the sum. If there was no standardization the user would just get back the probabilities that were previously supplied to the distribution object, which to me seems to be the more obvious solution.
  2. The behaviour of discrete_distribution is not specified in case the number of given probabilities is larger than the maximum number representable by the IntType.

Possible resolution: I propose to change the specification such that the non-standardized probabilities need to be returned and that an additional requirement is included for the number of probabilities to be smaller than the maximum of IntType.

[ Stephan Tolksdorf adds pre-Bellevue: ]

In reply to the discussion in N2424 of this issue:

Rescaled floating-point parameter vectors can not be expected to compare equal because of the limited precision of floating-point numbers. My proposal would at least guarantee that a parameter vector (of type double) passed into the distribution would compare equal with the one returned by the probabilities() method. Furthermore, I do not understand why "the changed requirement would lead to a significant increase in the amount of state in the distribution object". A typical implementation's state would increase by exactly one number: the sum of all probabilities. The textual representation for serialization would not need to grow at all. Finally, the proposed replacement "0 < n <= numeric_limits<IntType>::max() + 1" makes the implementation unnecessarily complicated, "0 < n <= numeric_limits<IntType>::max()" would be better.

[ Bellevue: ]

In N2424. We agree with the observation and the proposed resolution to part b). We recommend the wording n > 0 be replaced with 0 < n numeric_limits::max() + 1. However, we disagree with part a), as it would interfere with the definition of parameters' equality. Further, the changed requirement would lead to a significant increase in the amount of state of the distribution object.

As it stands now, it is convenient, and the changes proposed make it much less so.

NAD. Part a the current behavior is desirable. Part b, any constructor can fail, but the rules under which it can fail do not need to be listed here.

Proposed resolution:

See N2424 for the proposed resolution.

[ Stephan Tolksdorf adds pre-Bellevue: ]

In 28.5.9.6.1 [rand.dist.samp.discrete]:

Proposed wording a):

Change in para. 2

Constructs a discrete_distribution object with n=1 and p0 = w0 = 1

and change in para. 5

Returns: A vector<double> whose size member returns n and whose operator[] member returns pk the weight wk as a double value when invoked with argument k for k = 0, ..., n-1

Proposed wording b):

Change in para. 3:

If firstW == lastW, let the sequence w have length n = 1 and consist of the single value w0 = 1. Otherwise, [firstW,lastW) shall form a sequence w of length n > 0 such that 0 < n <= numeric_limits<IntType>::max(), and *firstW shall yield a value w0 convertible to double. [Note: The values wk are commonly known as the weights . -- end note]


737(i). Comment on [rand.dist.samp.pconst]

Section: 28.5.9.6.2 [rand.dist.samp.pconst] Status: NAD Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.dist.samp.pconst].

View all issues with NAD status.

Discussion:

  1. The discussion in point T11 above regarding probabilities() similarly applies to the method densities() of piecewise_constant_distribution.
  2. The design of the constructor

    template <class InputIteratorB, class InputIteratorW> 
    piecewise_constant_distribution( InputIteratorB firstB, InputIteratorB lastB, 
                                     InputIteratorW firstW);
    

    is unnecessarily unsafe, as there is no separate end-iterator given for the weights. I can't see any performance or convenience reasons that would justify the risks inherent in such a function interface, in particular the risk that input error might go unnoticed.

Possible resolution: I propose to add an InputIteratorW lastW argument to the interface.

[ Stephan Tolksdorf adds pre-Bellevue: ]

In reply to the discussion in N2424 I'd like to make the same comments as for 736.

[ Bellevue: ]

In N2424. There is already precedent elsewhere in the library. Follows existing convention. NAD.

Proposed resolution:

See N2424 for the proposed resolution.

[ Stephan Tolksdorf adds pre-Bellevue: ]

In 28.5.9.6.2 [rand.dist.samp.pconst]:

Proposed wording a)

Change in para. 2

Constructs a piecewise_constant_distribution object with n = 1, p0 = w0 = 1, b0 = 0, and b1 = 1

and change in para. 5

A vector<result_type> whose size member returns n and whose operator[] member returns pk the weight wk as a double value when invoked with argument k for k = 0, ..., n-1

Proposed wording b)

Change both occurrences of

"piecewise_constant_distribution(InputIteratorB firstB, InputIteratorB lastB,
                                 InputIteratorW firstW, InputIteratorW lastW)

and change in para. 3

the length of the sequence w starting from firstW shall be at least n, *firstW shall return a value w0 that is convertible to double, and any wk for k >= n shall be ignored by the distribution [firstW, lastW) shall form a sequence w of length n whose leading element w0 shall be convertible to double


738(i). Editorial issue in [rand.adapt.disc]/3

Section: 28.5.5.2 [rand.adapt.disc] Status: NAD Editorial Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.adapt.disc].

View all issues with NAD Editorial status.

Discussion:

Since the template parameter p and r are of type size_t, the member n in the class exposition should have type size_t, too.

Proposed resolution:

See N2424 for the proposed resolution.


739(i). Defect in [rand.util.canonical]/3

Section: 28.5.8.2 [rand.util.canonical] Status: NAD Submitter: Stephan Tolksdorf Opened: 2007-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.util.canonical].

View all issues with NAD status.

Discussion:

The complexity of generate_canonical is specified to be "exactly k=max(1, ceil(b/log2 R)) invocations of g". This terms involves a logarithm that is not rounded and hence can not (in general) be computed at compile time. As this function template is performance critical, I propose to replace ceil(b/log2 R) with ceil(b/floor(log2 R)).

See N2424 for further discussion.

[ Bellevue: ]

In N2424. Close NAD as described there.

Proposed resolution:

See N2424 for the proposed resolution.


741(i). Const-incorrect get_deleter function for shared_ptr

Section: 20.3.2.2.11 [util.smartptr.getdeleter] Status: NAD Submitter: Daniel Krügler Opened: 2007-09-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [util.smartptr.getdeleter].

View all issues with NAD status.

Discussion:

The following issue was raised by Alf P. Steinbach in c.l.c++.mod:

According to the recent draft N2369, both the header memory synopsis of 20.2 [memory] and 20.3.2.2.11 [util.smartptr.getdeleter] declare:

template<class D, class T> D* get_deleter(shared_ptr<T> const& p);

This allows to retrieve the pointer to a mutable deleter of a const shared_ptr (if that owns one) and therefore contradicts the usual philosophy that associated functors are either read-only (e.g. key_comp or value_comp of std::map) or do at least reflect the mutability of the owner (as seen for the both overloads of unique_ptr::get_deleter). Even the next similar counter-part of get_deleter - the two overloads of function::target in the class template function synopsis 22.10.17.3 [func.wrap.func] or in 22.10.17.3.6 [func.wrap.func.targ] - do properly mirror the const-state of the owner.

Possible proposed resolutions:

Replace the declarations of get_deleter in the header <memory> synopsis of 20.2 [memory] and in 20.3.2.2.11 [util.smartptr.getdeleter] by one of the following alternatives (A) or (B):

  1. Provide only the immutable variant. This would reflect the current praxis of container::get_allocator(), map::key_comp(), or map::value_comp.
    template<class D, class T> const D* get_deleter(shared_ptr<T> const& p);
    
  2. Just remove the function.

Alberto Ganesh Barbati adds:

  1. Replace it with two functions:

    template <class D, class T> D get_deleter(shared_ptr<T> const&);
    template <class D, class T> bool has_deleter(shared_ptr<T> const&);
    

    The first one would throw if D is the wrong type, while the latter would never throw. This approach would reflect the current praxis of use_facet/has_facet, with the twist of returning the deleter by value as container::get_allocator() do.

Peter Dimov adds:

My favorite option is "not a defect". A, B and C break useful code.

[ Bellevue: ]

Concern this is similar to confusing "pointer to const" with "a constant pointer".

Proposed resolution:


745(i). copy_exception API slices.

Section: 17.9.7 [propagation] Status: NAD Submitter: Alisdair Meredith Opened: 2007-10-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [propagation].

View all issues with NAD status.

Discussion:

It could be I did not understand the design rationale, but I thought copy_exception would produce an exception_ptr to the most-derived (dynamic) type of the passed exception. Instead it slices, which appears to be less useful, and a likely source of FAQ questions in the future.

(Peter Dimov suggests NAD)

[ Bellevue: ]

How could this be implemented in a way that the dynamic type is cloned?

The feature is designed to create an exception_ptr from an object whose static type is identical to the dynamic type and thus there is no slicing involved.

Proposed resolution:


747(i). We have 3 separate type traits to identify classes supporting no-throw operations

Section: 21.3.5.4 [meta.unary.prop] Status: NAD Submitter: Alisdair Meredith Opened: 2007-10-10 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [meta.unary.prop].

View all other issues in [meta.unary.prop].

View all issues with NAD status.

Discussion:

We have 3 separate type traits to identify classes supporting no-throw operations, which are very useful when trying to provide exception safety guarantees. However, I'm not entirely clear on what the current wording requires of a conforming implementation. To quote from has_nothrow_default_constructor:

or T is a class type with a default constructor that is known not to throw any exceptions

What level of magic do we expect to deduce if this is known?

E.g.

struct test{
 int x;
 test() : x() {}
};

Should I expect a conforming compiler to assert( has_nothrow_constructor<test>::value )

Is this a QoI issue?

Should I expect to 'know' only if-and-only-if there is an inline definition available?

Should I never expect that to be true, and insist that the user supplies an empty throw spec if they want to assert the no-throw guarantee?

It would be helpful to maybe have a footnote explaining what is required, but right now I don't know what to suggest putting in the footnote.

(agreement since is that trivial ops and explicit no-throws are required. Open if QoI should be allowed to detect further)

[ Bellevue: ]

This looks like a QoI issue. In the case of trivial and nothrow it is known. Static analysis of the program is definitely into QoI. Move to OPEN. Need to talk to Core about this.

[ 2009-07 Frankfurt: ]

This is QoI.

Move to NAD.

Proposed resolution:


748(i). The is_abstract type trait is defined by reference to 10.4.

Section: 21.3.5.4 [meta.unary.prop] Status: NAD Submitter: Alisdair Meredith Opened: 2007-10-10 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [meta.unary.prop].

View all other issues in [meta.unary.prop].

View all issues with NAD status.

Discussion:

I am trying to decide is a pure virtual function is a necessary as well as sufficient requirement to be classified as abstract?

For instance, is the following (non-polymorphic) type considered abstract?

struct abstract {
protected:
 abstract(){}
 abstract( abstract const & ) {}
 ~abstract() {}
};

(Suggested that this may be NAD, with an editorial fix-up from Pete on the core wording to make clear that abstract requires a pure virtual function)

Proposed resolution:

Core has clarified that the definition abstract is adequate. Issue withdrawn by submitter. NAD.


750(i). The current definition for is_convertible requires that the type be implicitly convertible, so explicit constructors are ignored.

Section: 21.3.7 [meta.rel] Status: Dup Submitter: Alisdair Meredith Opened: 2007-10-10 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [meta.rel].

View all other issues in [meta.rel].

View all issues with Dup status.

Duplicate of: 719

Discussion:

With the pending arrival of explicit conversion functions though, I'm wondering if we want an additional trait, is_explictly_convertible?

[ Bellevue: ]

Alisdair is considering preparing a paper listing a number of missing type traits, and feels that it might be useful to handle them all together rather than piecemeal. This would affect issue 719 and 750. These two issues should move to OPEN pending AM paper on type traits.

[ 2009-07 Frankfurt: ]

Duplicate of 719 (for our purposes).

[ Addressed in N2947. ]

Proposed resolution:


751(i). change pass-by-reference members of vector<bool> to pass-by-value?

Section: 24.3.12 [vector.bool] Status: NAD Submitter: Alisdair Meredith Opened: 2007-10-10 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [vector.bool].

View all other issues in [vector.bool].

View all issues with NAD status.

Discussion:

A number of vector<bool> members take const bool& as arguments. Is there any chance we could change them to pass-by-value or would I be wasting everyone's time if wrote up an issue?

[ post Bellevue: ]

As we understand it, the original requester (Martin Sebor) would like for implementations to be permitted to pass-by-value. Alisdair suggests that if this is to be resolved, it should be resolved more generally, e.g. in other containers as well.

We note that this would break ABI. However, we also suspect that this might be covered under the "as-if" rule in section 1.9.

Many in the group feel that for vector<bool>, this is a "don't care", and that at this point in the process it's not worth the bandwidth.

Issue 679 -- which was in ready status pre-Bellevue and is now in the working paper -- is related to this, though not a duplicate.

Moving to Open with a task for Alisdair to craft a informative note to be put whereever appropriate in the WP. This note would clarify places where pass-by-const-ref can be transformed to pass-by-value under the as-if rule.

[ San Francisco: ]

This is really a clause 17 issue, rather than something specific to vector<bool>.

Move to Open. Alisdair to provide a resolution. Alternately, Howard can close this as NAD and then open a new issue to handle the general issue (rather than the vector<bool> one).

Howard: Haven't yet opened new issue. Lacking wording for it.

[ 2009-07 Frankfurt: ]

NAD. Insufficient motivation to make any changes.

Proposed resolution:


754(i). Ambiguous return clause for std::uninitialized_copy

Section: 27.11.5 [uninitialized.copy] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2007-10-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [uninitialized.copy].

View all issues with NAD Editorial status.

Discussion:

14882-2003, [lib.uninitialized.copy] is currently written as follows:

template <class InputIterator, class ForwardIterator>
  ForwardIterator uninitialized_copy(InputIterator first, InputIterator last,
                                     ForwardIterator result);

-1- Effects:

for (; first != last; ++result, ++first)
  new (static_cast<void*>(&*result))
    typename iterator_traits<ForwardIterator>::value_type(*first);

-2- Returns: result

similarily for N2369, and its corresponding section 27.11.5 [uninitialized.copy].

It's not clear to me what the return clause is supposed to mean, I see two possible interpretations:

  1. The notion of result is supposed to mean the value given by the function parameter result [Note to the issue editor: Please use italics for result]. This seems somewhat implied by recognizing that both the function parameter and the name used in the clause do have the same italic font.
  2. The notion of "result" is supposed to mean the value of result after the preceding effects clause. This is in fact what all implementations I checked do (and which is probably it's intend, because it matches the specification of std::copy).

The problem is: I see nothing in the standard which grants that this interpretation is correct, specifically [lib.structure.specifications] or 16.3.2.4 [structure.specifications] resp. do not clarify which "look-up" rules apply for names found in the elements of the detailed specifications - Do they relate to the corresponding synopsis or to the effects clause (or possibly other elements)? Fortunately most detailed descriptions are unambigious in this regard, e.g. this problem does not apply for std::copy.

Proposed resolution:

Change the wording of the return clause to say (27.11.5 [uninitialized.copy]):

-2- Returns: The value of result after effects have taken place.

[ Bellevue: ]

Resolution: NAD editorial -- project editor to decide if change is worthwhile. Concern is that there are many other places this might occur.


757(i). Typo in the synopsis of vector

Section: 24.3.11 [vector] Status: NAD Editorial Submitter: Paolo Carlini Opened: 2007-11-04 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [vector].

View all issues with NAD Editorial status.

Discussion:

In the synopsis 24.3.11 [vector], there is the signature:

void insert(const_iterator position, size_type n, T&& x);

instead of:

iterator insert(const_iterator position, T&& x);

24.3.11.5 [vector.modifiers] is fine.

Proposed resolution:

Change the synopsis in 24.3.11 [vector]:

iterator insert(const_iterator position, const T& x); 
iterator insert(const_iterator position, T&& x);
void     insert(const_iterator position, size_type n, const T& x); 
void     insert(const_iterator position, size_type n, T&& x);

760(i). The emplace issue

Section: 24.2 [container.requirements] Status: NAD Submitter: Paolo Carlini Opened: 2007-11-11 Last modified: 2017-07-17

Priority: 2

View all other issues in [container.requirements].

View all issues with NAD status.

Discussion:

In an emplace member function the function parameter pack may be bound to a priori unlimited number of objects: some or all of them can be elements of the container itself. Apparently, in order to conform to the blanket statement 24.2 [container.requirements]/11, the implementation must check all of them for that possibility. A possible solution can involve extending the exception in 24.2 [container.requirements]/12 also to the emplace member. As a side note, the push_back and push_front member functions are luckily not affected by this problem, can be efficiently implemented anyway.

[ Related to 767 and to 2164 ]

[ Bellevue: ]

The proposed addition (13) is partially redundant with the existing paragraph 12. Why was the qualifier "rvalues" added to paragraph 12? Why does it not cover subelements and pointers?

Resolution: Alan Talbot to rework language, then set state to Review.

[ 2009-07 Frankfurt ]

The problem is broader than emplace. The LWG doesn't feel that it knows how to write wording that prohibits all of the problematic use cases at this time.

NAD Future.

[2015-02 Cologne]

LWG believes that 2164 addresses this issue and therefore considers 760 as NAD.

Proposed resolution:

Add after 24.2 [container.requirements]/12:

-12- Objects passed to member functions of a container as rvalue references shall not be elements of that container. No diagnostic required.

-13- Objects bound to the function parameter pack of the emplace member function shall not be elements or sub-objects of elements of the container. No diagnostic required.


763(i). Renaming emplace() overloads

Section: 24.2.7 [associative.reqmts] Status: NAD Submitter: Sylvain Pion Opened: 2007-12-04 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with NAD status.

Discussion:

The associative containers provide 2 overloads of emplace():

template <class... Args> pair<iterator, bool> emplace(Args&&... args);
template <class... Args> iterator emplace(const_iterator position, Args&&... args);

This is a problem if you mean the first overload while passing a const_iterator as first argument.

[ Related to 767 ]

[ Bellevue: ]

This can be disambiguated by passing "begin" as the first argument in the case when the non-default choice is desired. We believe that desire will be rare.

[For related discussion see 1302]

LWG N2680 renamed one of the overloads to emplace_hint.

Resolution: Change state to NAD.

Proposed resolution:

Rename one of the two overloads. For example to emplace_here, hint_emplace...


764(i). equal_range on unordered containers should return a pair of local_iterators

Section: 24.2.8 [unord.req] Status: NAD Submitter: Joe Gottman Opened: 2007-11-29 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with NAD status.

Discussion:

A major attribute of the unordered containers is that iterating though them inside a bucket is very fast while iterating between buckets can be much slower. If an unordered container has a low load factor, iterating between the last iterator in one bucket and the next iterator, which is in another bucket, is O(bucket_count()) which may be much larger than O(size()).

If b is an non-const unordered container of type B and k is an object of it's key_type, then b.equal_range(k) currently returns pair<B::iterator, B::iterator>. Consider the following code:

B::iterator lb, ub;
tie(lb, ub) = b.equal_range(k);
for (B::iterator it = lb; it != ub; ++it) {
        // Do something with *it
}

If b.equal_range(k) returns a non-empty range (i.e. b contains at least on element whose key is equivalent to k), then every iterator in the half-open range [lb, ub) will be in the same bucket, but ub will likely either be in a different bucket or be equal to b.end(). In either case, iterating between ub - 1 and ub could take a much longer time than iterating through the rest of the range.

If instead of returning pair<iterator, iterator>, equal_range were to return pair<local_iterator, local_iterator>, then ub (which, like lb, would now be a local_iterator) could be guaranteed to always be in the same bucket as lb. In the cases where currently ub is equal to b.end() or is in a different bucket, ub would be equal to b.end(b.bucket(key)). This would make iterating between lb and ub much faster, as every iteration would be constant time.

[ Bellevue: ]

The proposed resolution breaks consistency with other container types for dubious benefit, and iterators are already constant time.

Proposed resolution:

Change the entry for equal_range in Table 93 (24.2.8 [unord.req]) as follows:

expression return type assertion/note pre/post-condition complexity
b.equal_range(k) pair<local_iterator,local_iterator>; pair<const_local_iterator,const_local_iterator> for const b. Returns a range containing all elements with keys equivalent to k. Returns make_pair(b.end(b.bucket(key)),b.end(b.bucket(key))) if no such elements exist. Average case Θ(b.count(k)). Worst case Θ(b.size()).

773(i). issues with random

Section: 28.5.9.2 [rand.dist.uni] Status: NAD Submitter: P.J. Plauger Opened: 2008-01-14 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.dist.uni].

View all issues with NAD status.

Discussion:

  1. 28.5.9.2.1 [rand.dist.uni.int] uniform_int constructor has changed the default max constructor parameter from 9 (in TR1) to max(). The value is arbitrary at best and shouldn't be lightly changed because it breaks backward compatibility.
  2. 28.5.9.2.1 [rand.dist.uni.int] uniform_int has a parameter param that you can provide on construction or operator(), set, and get. But there is not even a hint of what this might be for.
  3. 28.5.9.2.2 [rand.dist.uni.real] uniform_real. Same issue as #2.

[ Bellevue: ]

NAD. Withdrawn.

Proposed resolution:


784(i). unique_lock::release

Section: 33.6.5.4.4 [thread.lock.unique.mod] Status: NAD Submitter: Constantine Sapuntzakis Opened: 2008-02-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

unique_lock::release will probably lead to many mistakes where people call release instead of unlock. I just coded such a mistake using the boost pre-1.35 threads library last week.

In many threading libraries, a call with release in it unlocks the lock (e.g. ReleaseMutex in Win32, java.util.concurrent.Semaphore).

I don't call unique_lock::lock much at all, so I don't get to see the symmetry between ::lock and ::unlock. I usually use the constructor to lock the mutex. So I'm left to remember whether to call release or unlock during the few times I need to release the mutex before the scope ends. If I get it wrong, the compiler doesn't warn me.

An alternative name for release may be disown.

This might be a rare case where usability is hurt by consistency with the rest of the C++ standard (e.g. std::auto_ptr::release).

[ Bellevue: ]

Change a name from release to disown. However prior art uses the release name. Compatibility with prior art is more important that any possible benefit such a change might make. We do not see the benefit for changing. NAD

Proposed resolution:

Change the synopsis in 33.6.5.4 [thread.lock.unique]:

template <class Mutex> 
class unique_lock 
{ 
public:
   ...
   mutex_type* release disown();
   ...
};

Change 33.6.5.4.4 [thread.lock.unique.mod]:

mutex_type *release disown();

785(i). [tr1] Random Number Requirements in TR1

Section: 5.1.4.5 [tr1::tr.rand.eng.disc], 5.1.4.6 [tr1::tr.rand.eng.xor] Status: NAD Submitter: John Maddock Opened: 2008-01-15 Last modified: 2016-02-01

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Table 16 of TR1 requires that all Pseudo Random Number generators have a

seed(integer-type s)

member function that is equivalent to:

mygen = Generator(s)

But the generators xor_combine and discard_block have no such seed member, only the

template <class Gen>
seed(Gen&);

member, which will not accept an integer literal as an argument: something that appears to violate the intent of Table 16.

So... is this a bug in TR1?

This is a real issue BTW, since the Boost implementation does adhere to the requirements of Table 16, while at least one commercial implementation does not and follows a strict adherence to sections 5.1.4.5 and 5.1.4.6 instead.

[ Jens adds: ]

Both engines do have the necessary constructor, therefore the omission of the seed() member functions appears to be an oversight.

[ Post Summit Daniel adds: ]

Recommend NAD: xor_combine does no longer exist and discard_block[_engine] has now the required seed overload accepting a result_type, which shall be an unsigned integral type.

[ Batavia (2009-05): ]

Move to NAD as recommended.

Proposed resolution:

NAD Recommended.


790(i). xor_combine::seed not specified

Section: 99 [rand.adapt.xor] Status: NAD Submitter: P.J. Plauger Opened: 2008-02-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.adapt.xor].

View all issues with NAD status.

Discussion:

xor_combine::seed(result_type) and seed(seed_seq&) don't say what happens to each of the sub-engine seeds. (Should probably do the same to both, unlike TR1.)

[ Bellevue: ]

Overcome by the previous proposal. NAD mooted by resolution of 789.

Proposed resolution:


791(i). piecewise_constant_distribution::densities has wrong name

Section: 28.5.9.6.2 [rand.dist.samp.pconst] Status: NAD Submitter: P.J. Plauger Opened: 2008-02-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.dist.samp.pconst].

View all issues with NAD status.

Discussion:

piecewise_constant_distribution::densities() should be probabilities(), just like discrete_distribution. (There's no real use for weights divided by areas.)

[ Bellevue: ]

Fermilab does not agree with this summary. As defined in the equation in 26.4.8.5.2/4, the quantities are indeed probability densities not probabilities. Because we view this distribution as a parameterization of a *probability density function*, we prefer to work in terms of probability densities.

We don't think this should be changed.

If there is a technical argument about why the implementation dealing with these values can't be as efficient as one dealing with probabilities, we might reconsider. We don't care about this one member function being somewhat more or less efficient; we care about the size of the distribution object and the speed of the calls to generate variates.

Proposed resolution:

Change synopsis in 28.5.9.6.2 [rand.dist.samp.pconst]:

template <class RealType = double> 
class piecewise_constant_distribution 
{ 
public:
    ...
    vector<double> densities probabilities() const;
    ...
};

Change 28.5.9.6.2 [rand.dist.samp.pconst]/6:

vector<double> densities probabilities() const;

795(i). general_pdf_distribution should be dropped

Section: 99 [rand.dist.samp.genpdf] Status: Dup Submitter: P.J. Plauger Opened: 2008-02-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.dist.samp.genpdf].

View all issues with Dup status.

Duplicate of: 732

Discussion:

general_pdf_distribution should be dropped. (It's a research topic in adaptive numerical integration.)

[ Stephan Tolksdorf notes: ]

This appears to be a duplicate of 732.

Proposed resolution:


796(i). ranlux48_base returns wrong value

Section: 28.5.6 [rand.predef] Status: NAD Submitter: P.J. Plauger Opened: 2008-02-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.predef].

View all issues with NAD status.

Discussion:

The 10,000th value returned by ranlux48_base is supposed to be 61839128582725. We get 192113843633948. (Note that the underlying generator was changed in Kona.)

[ Bellevue: ]

Submitter withdraws defect.

Proposed resolution:

Change 28.5.6 [rand.predef]/p5:

typedef subtract_with_carry_engine<uint_fast64_t, 48, 5, 12> 
        ranlux48_base; 

Required behavior: The 10000th consecutive invocation of a default-constructed object of type ranlux48_base shall produce the value 61839128582725 192113843633948.


797(i). ranlux48 returns wrong value

Section: 28.5.6 [rand.predef] Status: NAD Submitter: P.J. Plauger Opened: 2008-02-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.predef].

View all issues with NAD status.

Discussion:

The 10,000th value returned by ranlux48 is supposed to be 249142670248501. We get 88229545517833. (Note that this depends on ranlux48_base.)

[ Bellevue: ]

Submitter withdraws defect.

Proposed resolution:

Change 28.5.6 [rand.predef]/p6:

typedef discard_block_engine<ranlux48_base, 389, 11> 
        ranlux48

Required behavior: The 10000th consecutive invocation of a default-constructed object of type ranlux48 shall produce the value 249142670248501 88229545517833.


799(i). Mersenne twister equality overspecified

Section: 28.5.4.3 [rand.eng.mers], 99 [tr.rand.eng.mers] Status: NAD Submitter: Stephan Tolksdorf Opened: 2008-02-18 Last modified: 2023-05-24

Priority: Not Prioritized

View all other issues in [rand.eng.mers].

View all issues with NAD status.

Discussion:

99 [tr.rand.eng.mers](10) requires that operator== for the mersenne_twister returns true if and only if the states of two mersenne_twisters, consisting each of n integers between 0 and 2w - 1, are completely equal. This is a contradiction with 99 [tr.rand.req](3) because the given definition of the state also includes the lower r bits of x(i-n), which will never be used to generate a random number. If two mersenne_twisters only differ in the lower bits of x(i-n) they will not compare equal, although they will produce an identical sequence of random numbers.

28.5.4.3 [rand.eng.mers] in the latest C++ draft does not specify the behaviour of operator== but uses a similar definition of the state and, just like 99 [tr.rand.eng.mers], requires the textual representation of a mersenne_twister_engine to consist of Xi-n to Xi-1, including the lower bits of Xi-n. This leads to two problems: First, the unsuspecting implementer is likely to erroneously compare the lower r bits of Xi-n in operator==. Second, if only the lower r bits differ, two mersenne_twister_engines will compare equal (if correctly implemented) but have different textual representations, which conceptually is a bit ugly.

I propose that a paragraph or footnote is added to 28.5.4.3 [rand.eng.mers] which clarifies that the lower r bits of Xi-n are not to be compared in operator== and operator!=. It would only be consequent if furthermore the specification for the textual respresentation was changed to Xi-n bitand ((2w - 1) - (2r - 1)), Xi-(n-1), ..., Xi-1 or something similar.

These changes would likely have no practical effect, but would allow an implementation that does the right thing to be standard-conformant.

[ Bellevue: ]

Fermi Lab has no objection to the proposed change. However it feels that more time is needed to check the details, which would suggest a change to REVIEW.

Bill feels that this is NAD, not enough practical importance to abandon the simple definition of equality, and someone would have to do a lot more study to ensure that all cases are covered for a very small payback. The submitter admits that "These changes would likely have no practical effect,", and according to Plum's razor this means that it is not worth the effort!

Revisted: Agree that the fact that there is no practical difference means that no change can be justified.

Proposed resolution:

In 28.5.4.3 [rand.eng.mers]:

Insert at the end of para 2.:

[Note: The lower r bits of Xi-n do not influence the state transition and hence should not be compared when comparing two mersenne_twister_engine objects. -- end note]

In para 5. change:

The textual representation of xi consists of the values of Xi-n bitand ((2w - 1) - (2r - 1)), Xi-(n-1), ..., Xi-1, in that order.


802(i). knuth_b returns wrong value

Section: 28.5.6 [rand.predef] Status: NAD Submitter: P.J. Plauger Opened: 2008-02-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.predef].

View all issues with NAD status.

Discussion:

The 10,000th value returned by knuth_b is supposed to be 1112339016. We get 2126698284.

Proposed resolution:

Change 28.5.6 [rand.predef]/p8:

typedef shuffle_order_engine<minstd_rand0, 256> 
        knuth_b; 

Required behavior: The 10000th consecutive invocation of a default-constructed object of type knuth_b shall produce the value 1112339016 2126698284.

[ Bellevue: Submitter withdraws defect. "We got the wrong value for entirely the right reasons". NAD. ]


812(i). unsolicited multithreading considered harmful?

Section: 27.8.2 [alg.sort] Status: NAD Editorial Submitter: Paul McKenney Opened: 2008-02-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

Multi-threading is a good thing, but unsolicited multi-threading can potentially be harmful. For example, sort() performance might be greatly increased via a multithreaded implementation. However, such a multithreaded implementation could result in concurrent invocations of the user-supplied comparator. This would in turn result in problems given a caching comparator that might be written for complex sort keys. Please note that this is not a theoretical issue, as multithreaded implementations of sort() already exist.

Having a multithreaded sort() available is good, but it should not be the default for programs that are not explicitly multithreaded. Users should not be forced to deal with concurrency unless they have asked for it.

[ This may be covered by N2410 Thread-Safety in the Standard Library (Rev 1). ]

Proposed resolution:

Rationale:

This is already covered by 17.6.5.6/20 in N2723.


822(i). Object with explicit copy constructor no longer CopyConstructible

Section: 16.4.4.2 [utility.arg.requirements] Status: NAD Submitter: James Kanze Opened: 2008-04-01 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [utility.arg.requirements].

View all issues with NAD status.

Discussion:

I just noticed that the following program is legal in C++03, but is forbidden in the current draft:

#include <vector>
#include <iostream>

class Toto
{
public:
    Toto() {}
    explicit Toto( Toto const& ) {}
} ;

int
main()
{
    std::vector< Toto > v( 10 ) ;
    return 0 ;
}

Is this change intentional? (And if so, what is the justification? I wouldn't call such code good, but I don't see any reason to break it unless we get something else in return.)

[ San Francisco: ]

The subgroup that looked at this felt this was a good change, but it may already be handled by incoming concepts (we're not sure).

[ Post Summit: ]

Alisdair: Proposed resolution kinda funky as these tables no longer exist. Move from direct init to copy init. Clarify with Doug, recommends NAD.

Walter: Suggest NAD via introduction of concepts.

Recommend close as NAD.

[ 2009-07 Frankfurt: ]

Need to look at again without concepts.

[ 2009-07 Frankfurt: ]

Move to Ready with original proposed resolution.

[Howard: Original proposed resolution restored.]

[ 2010-11 Batavia: ]

This issue was re-reviewed in relation to [another issue, number to follow], and the verdict was reversed. Explicit copy and move constructors are rare beasts, and the ripple effect of this fix was far more difficult to contain than simply saying such types do not satisfy the MoveConstructible and CopyConstructible requirements.

In 16.4.4.2 [utility.arg.requirements] change Table 33: MoveConstructible requirements [moveconstructible]:

expressionpost-condition
T t(rv) = rvt is equivalent to the value of rv before the construction
...

In 16.4.4.2 [utility.arg.requirements] change Table 34: CopyConstructible requirements [copyconstructible]:

expressionpost-condition
T t(u) = uthe value of u is unchanged and is equivalent to t
...

Proposed resolution:

Resolved by n3215.


826(i). Equivalent of %'d, or rather, lack thereof?

Section: 30.4.3.3 [locale.nm.put] Status: NAD Submitter: Peter Dimov Opened: 2008-04-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

In the spirit of printf vs iostream...

POSIX printf says that %'d should insert grouping characters (and the implication is that in the absence of ' no grouping characters are inserted). The num_put facet, on the other hand, seems to always insert grouping characters. Can this be considered a defect worth fixing for C++0x? Maybe ios_base needs an additional flag?

[ Pablo Halpern: ]

I'm not sure it constitutes a defect, but I would be in favor of adding another flag (and corresponding manipulator).

[ Martin Sebor: ]

I don't know if it qualifies as a defect but I agree that there should be an easy way to control whether the thousands separator should or shouldn't be inserted. A new flag would be in line with the current design of iostreams (like boolalpha, showpos, or showbase).

[ Sophia Antipolis: ]

This is not a part of C99. LWG suggests submitting a paper may be appropriate.

Proposed resolution:


830(i). Incomplete list of char_traits specializations

Section: 23.2 [char.traits] Status: NAD Editorial Submitter: Dietmar Kühl Opened: 2008-04-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [char.traits].

View all issues with NAD Editorial status.

Discussion:

Paragraph 4 of 23.2 [char.traits] mentions that this section specifies two specializations (char_traits<char> and (char_traits<wchar_t>). However, there are actually four specializations provided, i.e. in addition to the two above also char_traits<char16_t> and char_traits<char32_t>). I guess this was just an oversight and there is nothing wrong with just fixing this.

[ Alisdair adds: ]

char_traits< char16/32_t > should also be added to <ios_fwd> in 31.3 [iostream.forward], and all the specializations taking a char_traits parameter in that header.

[ Sophia Antipolis: ]

Idea of the issue is ok.

Alisdair to provide wording, once that wording arrives, move to review.

[ 2009-05-04 Alisdair adds: ]

The main point of the issue was resolved editorially in N2723, so we are close to NAD Editorial. However, exploring the issue we found a second tweak was necessary for <iosfwd> and that is still outstanding, so here are the words I am long overdue delivering:

[ Howard: I've put Alisdair's words into the proposed wording section and moved the issue to Review. ]

[ Original proposed wording. ]

Replace paragraph 4 of 23.2 [char.traits] by:

This subclause specifies a struct template, char_traits<charT>, and four explicit specializations of it, char_traits<char>, char_traits<char16_t>, char_traits<char32_t>, and char_traits<wchar_t>, all of which appear in the header <string> and satisfy the requirements below.

[ Batavia (2009-05): ]

We agree. Move to NAD Editorial.

Proposed resolution:

Change Forward declarations 31.3 [iostream.forward]:

Header <iosfwd> synopsis

namespace std {
   template<class charT> class char_traits;
   template<> class char_traits<char>;
   template<> class char_traits<char16_t>;
   template<> class char_traits<char32_t>;
   template<> class char_traits<wchar_t>;
...
}

831(i). wrong type for not_eof()

Section: 23.2.4 [char.traits.specializations] Status: NAD Editorial Submitter: Dietmar Kühl Opened: 2008-04-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [char.traits.specializations].

View all issues with NAD Editorial status.

Discussion:

In Table 56 (Traits requirements) the not_eof() member function is using an argument of type e which denotes an object of type X::int_type. However, the specializations in 23.2.4 [char.traits.specializations] all use char_type. This would effectively mean that the argument type actually can't represent EOF in the first place. I'm pretty sure that the type used to be int_type which is quite obviously the only sensible argument.

This issue is close to being editorial. I suspect that the proposal changing this section to include the specializations for char16_t and char32_t accidentally used the wrong type.

Proposed resolution:

In 23.2.4.2 [char.traits.specializations.char], [char.traits.specializations.char16_t], [char.traits.specializations.char32_t], and [char.traits.specializations.wchar_t] correct the argument type from char_type to int_type.

Rationale:

Already fixed in WP.


832(i). Applying constexpr to System error support

Section: 19.5 [syserr] Status: NAD Submitter: Beman Dawes Opened: 2008-05-14 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [syserr].

View all issues with NAD status.

Discussion:

Initialization of objects of class error_code (19.5.4 [syserr.errcode]) and class error_condition (19.5.5 [syserr.errcondition]) can be made simpler and more reliable by use of the new constexpr feature [N2349] of C++0x. Less code will need to be generated for both library implementations and user programs when manipulating constant objects of these types.

This was not proposed originally because the constant expressions proposal was moving into the standard at about the same time as the Diagnostics Enhancements proposal [N2241], and it wasn't desirable to make the later depend on the former. There were also technical concerns as to how constexpr would apply to references. Those concerns are now resolved; constexpr can't be used for references, and that fact is reflected in the proposed resolution.

Thanks to Jens Maurer, Gabriel Dos Reis, and Bjarne Stroustrup for clarification of constexpr requirements.

LWG issue 804 is related in that it raises the question of whether the exposition only member cat_ of class error_code (19.5.4 [syserr.errcode]) and class error_condition (19.5.5 [syserr.errcondition]) should be presented as a reference or pointer. While in the context of 804 that is arguably an editorial question, presenting it as a pointer becomes more or less required with this proposal, given constexpr does not play well with references. The proposed resolution thus changes the private member to a pointer, which also brings it in sync with real implementations.

[ Sophia Antipolis: ]

On going question of extern pointer vs. inline functions for interface.

[ Pre-San Francisco: ]

Beman Dawes reports that this proposal is unimplementable, and thus NAD.

Implementation would require constexpr objects of classes derived from class error_category, which has virtual functions, and that is not allowed by the core language. This was determined when trying to implement the proposal using a constexpr enabled compiler provided by Gabriel Dos Reis, and subsequently verified in discussions with Gabriel and Jens Maurer.

Proposed resolution:

The proposed wording assumes the LWG 805 proposed wording has been applied to the WP, resulting in the former posix_category being renamed generic_category. If 805 has not been applied, the names in this proposal must be adjusted accordingly.

Change 19.5.3.1 [syserr.errcat.overview] Class error_category overview error_category synopsis as indicated:

const error_category& get_generic_category();
const error_category& get_system_category();

static extern const error_category&* const generic_category = get_generic_category();
static extern const error_category&* const native_category system_category = get_system_category();

Change 19.5.3.5 [syserr.errcat.objects] Error category objects as indicated:

extern const error_category&* const get_generic_category();

Returns: A reference generic_category shall point to an a statically initialized object of a type derived from class error_category.

Remarks: The object's default_error_condition and equivalent virtual functions shall behave as specified for the class error_category. The object's name virtual function shall return a pointer to the string "GENERIC".

extern const error_category&* const get_system_category();

Returns: A reference system_category shall point to an a statically initialized object of a type derived from class error_category.

Remarks: The object's equivalent virtual functions shall behave as specified for class error_category. The object's name virtual function shall return a pointer to the string "system". The object's default_error_condition virtual function shall behave as follows:

If the argument ev corresponds to a POSIX errno value posv, the function shall return error_condition(posv, generic_category). Otherwise, the function shall return error_condition(ev, system_category). What constitutes correspondence for any given operating system is unspecified. [Note: The number of potential system error codes is large and unbounded, and some may not correspond to any POSIX errno value. Thus implementations are given latitude in determining correspondence. -- end note]

Change 19.5.4.1 [syserr.errcode.overview] Class error_code overview as indicated:

class error_code {
public:
  ...;
  constexpr error_code(int val, const error_category&* cat);
  ...
  void assign(int val, const error_category&* cat);
  ...
  const error_category&* category() const;
  ...
private:
  int val_;                    // exposition only
  const error_category&* cat_; // exposition only

Change 19.5.4.2 [syserr.errcode.constructors] Class error_code constructors as indicated:

constexpr error_code(int val, const error_category&* cat);

Effects: Constructs an object of type error_code.

Postconditions: val_ == val and cat_ == cat.

Throws: Nothing.

Change 19.5.4.3 [syserr.errcode.modifiers] Class error_code modifiers as indicated:

void assign(int val, const error_category&* cat);

Postconditions: val_ == val and cat_ == cat.

Throws: Nothing.

Change 19.5.4.4 [syserr.errcode.observers] Class error_code observers as indicated:

const error_category&* category() const;

Returns: cat_.

Throws: Nothing.

Change 19.5.5.1 [syserr.errcondition.overview] Class error_condition overview as indicated:

class error_condition {
public:
  ...;
  constexpr error_condition(int val, const error_category&* cat);
  ...
  void assign(int val, const error_category&* cat);
  ...
  const error_category&* category() const;
  ...
private:
  int val_;                    // exposition only
  const error_category&* cat_; // exposition only

Change 19.5.5.2 [syserr.errcondition.constructors] Class error_condition constructors as indicated:

constexpr error_condition(int val, const error_category&* cat);

Effects: Constructs an object of type error_condition.

Postconditions: val_ == val and cat_ == cat.

Throws: Nothing.

Change 19.5.5.3 [syserr.errcondition.modifiers] Class error_condition modifiers as indicated:

void assign(int val, const error_category&* cat);

Postconditions: val_ == val and cat_ == cat.

Throws: Nothing.

Change 19.5.5.4 [syserr.errcondition.observers] Class error_condition observers as indicated:

const error_category&* category() const;

Returns: cat_.

Throws: Nothing.

Throughout 19.5 [syserr] System error support, change "category()." to "category()->". Appears approximately six times.

[Partially Editorial] In 19.5.6 [syserr.compare] Comparison operators, paragraphs 2 and 4, change "category.equivalent(" to "category()->equivalent(".

Change 19.5.8.1 [syserr.syserr.overview] Class system_error overview as indicated:

public:
  system_error(error_code ec, const string& what_arg);
  system_error(error_code ec);
  system_error(int ev, const error_category&* ecat,
      const string& what_arg);
  system_error(int ev, const error_category&* ecat);

Change 19.5.8.2 [syserr.syserr.members] Class system_error members as indicated:

system_error(int ev, const error_category&* ecat, const string& what_arg);

Effects: Constructs an object of class system_error.

Postconditions: code() == error_code(ev, ecat) and strcmp(runtime_error::what(), what_arg.c_str()) == 0.

system_error(int ev, const error_category&* ecat);

Effects: Constructs an object of class system_error.

Postconditions: code() == error_code(ev, ecat) and strcmp(runtime_error::what(), "") == 0.

Rationale:

[ San Francisco: ]

NAD because Beman said so.


833(i). Freestanding implementations header list needs review for C++0x

Section: 16.4.2.5 [compliance] Status: NAD Submitter: Beman Dawes Opened: 2008-05-14 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [compliance].

View all issues with NAD status.

Discussion:

Once the C++0x standard library is feature complete, the LWG needs to review 16.4.2.5 [compliance] Freestanding implementations header list to ensure it reflects LWG consensus.

[ San Francisco: ]

This is a placeholder defect to remind us to review the table once we've stopped adding headers to the library.

Three new headers that need to be added to the list:

<initializer_list> <concept> <iterator_concepts>

<iterator_concepts>, in particular, has lots of stuff that isn't needed, so maybe the stuff that is needed should be broken out into a separate header.

Robert: What about reference_closure? It's currently in <functional>.

[ Post Summit Daniel adds: ]

  1. The comment regarding reference_closure seems moot since it was just recently decided to remove that.
  2. A reference to proposal N2814 ("Fixing freestanding") should be added. This paper e.g. proposes to add only <initializer_list> to the include list of freestanding.

[ 2009-07 Frankfurt: ]

Addressed by paper N2814.

Move to NAD.

Proposed resolution:


837(i). basic_ios::copyfmt() overly loosely specified

Section: 31.5.4.3 [basic.ios.members] Status: NAD Editorial Submitter: Martin Sebor Opened: 2008-05-17 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [basic.ios.members].

View all issues with NAD Editorial status.

Discussion:

The basic_ios::copyfmt() member function is specified in 31.5.4.3 [basic.ios.members] to have the following effects:

Effects: If (this == &rhs) does nothing. Otherwise assigns to the member objects of *this the corresponding member objects of rhs, except that

Since the rest of the text doesn't specify what the member objects of basic_ios are this seems a little too loose.

[ Batavia (2009-05): ]

We agree with the proposed resolution. Move to NAD Editorial.

Proposed resolution:

I propose to tighten things up by adding a Postcondition clause to the function like so:

Postconditions:

copyfmt() postconditions
Element Value
rdbuf() unchanged
tie() rhs.tie()
rdstate() unchanged
exceptions() rhs.exceptions()
flags() rhs.flags()
width() rhs.width()
precision() rhs.precision()
fill() rhs.fill()
getloc() rhs.getloc()

The format of the table follows Table 117 (as of N2588): basic_ios::init() effects.

The intent of the new table is not to impose any new requirements or change existing ones, just to be more explicit about what I believe is already there.


840(i). pair default template argument

Section: 22.3 [pairs] Status: NAD Submitter: Thorsten Ottosen Opened: 2008-05-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [pairs].

View all issues with NAD status.

Discussion:

I have one issue with std::pair. Well, it might just be a very annoying historical accident, but why is there no default template argument for the second template argument? This is so annoying when the type in question is looong and hard to write (type deduction with auto won't help those cases where we use it as a return or argument type).

Proposed resolution:

Change the synopsis in 22.2 [utility] to read:

template <class T1, class T2 = T1> struct pair;

Change 22.3 [pairs] to read:

namespace std {
 template <class T1, class T2 = T1>
 struct pair {
   typedef T1 first_type;
   typedef T2 second_type;
   ...

Rationale:

std::pair is a heterogeneous container.


841(i). cstdint.syn inconsistent with C99

Section: 17.4.1 [cstdint.syn] Status: NAD Editorial Submitter: Martin Sebor Opened: 2008-05-17 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [cstdint.syn].

View all issues with NAD Editorial status.

Discussion:

In specifying the names of macros and types defined in header <stdint.h>, C99 makes use of the symbol N to accommodate unusual platforms with word sizes that aren't powers of two. C99 permits N to take on any positive integer value (including, for example, 24).

In [stdint.sy] Header <cstdint> synopsis, C++ on the other hand, fixes the value of N to 8, 16, 32, and 64, and specifies only types with these exact widths.

In addition, paragraph 1 of the same section makes use of a rather informal shorthand notation to specify sets of macros. When interpreted strictly, the notation specifies macros such as INT_8_MIN that are not intended to be specified.

Finally, the section is missing the usual table of symbols defined in that header, making it inconsistent with the rest of the specification.

Proposed resolution:

I propose to use the same approach in the C++ spec as C99 uses, that is, to specify the header synopsis in terms of "exposition only" types that make use of the symbol N to denote one or more of a theoretically unbounded set of widths.

Further, I propose to add a new table to section listing the symbols defined in the header using a more formal notation that avoids introducing inconsistencies.

To this effect, in [stdint.sy] Header <cstdint> synopsis, replace both the synopsis and paragraph 1 with the following text:

  1. In the names defined in the <cstdint> header, the symbol N represents a positive decimal integer with no leading zeros (e.g., 8 or 24, but not 0, 04, or 048). With the exception of exact-width types, macros and types for values of N in the set of 8, 16, 32, and 64 are required. Exact-width types, and any macros and types for values of N other than 8, 16, 32, and 64 are optional. However, if an implementation provides integer types with widths of 8, 16, 32, or 64 bits, the corresponding exact-width types and macros are required.
namespace std {

   // required types

   // Fastest minimum-width integer types
   typedef signed integer type   int_fast8_t;
   typedef signed integer type   int_fast16_t;
   typedef signed integer type   int_fast32_t;
   typedef signed integer type   int_fast64_t;

   typedef unsigned integer type uint_fast8_t;
   typedef unsigned integer type uint_fast16_t;
   typedef unsigned integer type uint_fast32_t;
   typedef unsigned integer type uint_fast64_t;

   // Minimum-width integer types
   typedef signed integer type   int_least8_t;
   typedef signed integer type   int_least16_t;
   typedef signed integer type   int_least32_t;
   typedef signed integer type   int_least64_t;

   typedef unsigned integer type uint_least8_t;
   typedef unsigned integer type uint_least16_t;
   typedef unsigned integer type uint_least32_t;
   typedef unsigned integer type uint_least64_t;

   // Greatest-width integer types
   typedef signed integer type   intmax_t;
   typedef unsigned integer type uintmax_t;

   // optionally defined types

   // Exact-width integer types
   typedef signed integer type   intN_t;
   typedef unsigned integer type uintN_t;

   // Fastest minimum-width integer types for values
   // of N other than 8, 16, 32, and 64
   typedef signed integer type   uint_fastN_t;
   typedef unsigned integer type uint_fastN_t;

   // Minimum-width integer types for values
   // of N other than 8, 16, 32, and 64
   typedef signed integer type   uint_leastN_t;
   typedef unsigned integer type uint_leastN_t;

   // Integer types capable of holding object pointers
   typedef signed integer type   intptr_t;
   typedef signed integer type   intptr_t;

}

[Note to editor: Remove all of the existing paragraph 1 from [stdint.sy].]

Table ??: Header <cstdint> synopsis

Type Name(s)
Macros: INTN_MIN INTN_MAX UINTN_MAX
INT_FASTN_MIN INT_FASTN_MAX UINT_FASTN_MAX
INT_LEASTN_MIN INT_LEASTN_MAX UINT_LEASTN_MAX
INTPTR_MIN INTPTR_MAX UINTPTR_MAX
INTMAX_MIN INTMAX_MAX UINTMAX_MAX
PTRDIFF_MIN PTRDIFF_MAX PTRDIFF_MAX
SIG_ATOMIC_MIN SIG_ATOMIC_MAX SIZE_MAX
WCHAR_MIN WCHAR_MAX
WINT_MIN WINT_MAX
INTN_C() UINTN_C()
INTMAX_C() UINTMAX_C()
Types: intN_t uintN_t
int_fastN_t uint_fastN_t
int_leastN_t uint_leastN_t
intptr_t uintptr_t
intmax_t uintmax_t

849(i). missing type traits to compute root class and derived class of types in a class hierachy

Section: 21.3.8.7 [meta.trans.other] Status: NAD Submitter: Thorsten Ottosen Opened: 2008-06-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [meta.trans.other].

View all issues with NAD status.

Discussion:

The type traits library contains various traits to dealt with polymorphic types, e.g. std::has_virtual_destructor, std::is_polymorphic and std::is_base_of. However, there is no way to compute the unique public base class of a type if such one exists. Such a trait could be very useful if one needs to instantiate a specialization made for the root class whenever a derived class is passed as parameter. For example, imagine that you wanted to specialize std::hash for a class hierarchy---instead of specializing each class, you could specialize the std::hash<root_class> and provide a partial specialization that worked for all derived classes.

This ability---to specify operations in terms of their equivalent in the root class---can be done with e.g. normal functions, but there is, AFAIK, no way to do it for class templates. Being able to access compile-time information about the type-hierachy can be very powerful, and I therefore also suggest traits that computes the directly derived class whenever that is possible.

If the computation can not be done, the traits should fall back on an identity transformation. I expect this gives the best overall usability.

Proposed resolution:

Add the following to the synopsis in 21.3.3 [meta.type.synop] under "other transformations":

template< class T > struct direct_base_class;
template< class T > struct direct_derived_class;
template< class T > struct root_base_class;

Add three new entries to table 51 (21.3.8.7 [meta.trans.other]) with the following content

TemplateConditionComments
template< class T > struct direct_base_class; T shall be a complete type. The member typedef type shall equal the accessible unambiguous direct base class of T. If no such type exists, the member typedef type shall equal T.
template< class T > struct direct_derived_class; T shall be a complete type. The member typedef type shall equal the unambiguous type which has T as an accessible unambiguous direct base class. If no such type exists, the member typedef type shall equal T.
template< class T > struct root_base_class; T shall be a complete type. The member typedef type shall equal the accessible unambiguous most indirect base class of T. If no such type exists, the member typedef type shall equal T.

Rationale:

2008-9-16 San Francisco: Issue pulled by author prior to being reviewed by the LWG.


851(i). simplified array construction

Section: 24.3.7 [array] Status: NAD Submitter: Benjamin Kosnik Opened: 2008-06-05 Last modified: 2017-06-06

Priority: Not Prioritized

View all other issues in [array].

View all issues with NAD status.

Discussion:

This is an issue that came up on the libstdc++ list, where a discrepancy between "C" arrays and C++0x's std::array was pointed out.

In "C," this array usage is possible:

int ar[] = {1, 4, 6};

But for C++,

std::array<int> a = { 1, 4, 6 }; // error

Instead, the second parameter of the array template must be explicit, like so:

std::array<int, 3> a = { 1, 4, 6 };

Doug Gregor proposes the following solution, that assumes generalized initializer lists.

template<typename T, typename... Args>
inline array<T, sizeof...(Args)> 
make_array(Args&&... args) 
{ return { std::forward<Args>(args)... };  }

Then, the way to build an array from a list of unknown size is:

auto a = make_array<T>(1, 4, 6);

[ San Francisco: ]

Benjamin: Move to Ready?

Bjarne: I'm not convinced this is useful enough to add, so I'd like us to have time to reflect on it.

Alisdair: the constraints are wrong, they should be

template<ValueType T, ValueType... Args>
requires Convertible<Args, T>...
array<T, sizeof...(Args)> make_array(Args&&... args);

Alidair: this would be useful if we had a constexpr version.

Bjarne: this is probably useful for arrays with a small number of elements, but it's not clearly useful otherwise.

Consensus is to move to Open.

[ 2009-06-07 Daniel adds: ]

I suggest a fix and a simplification of the current proposal: Recent prototyping by Howard showed, that a fix is required because narrowing conversion 9.4.5 [dcl.init.list]/6 b.3 would severely limit the possible distribution of argument types, e.g. the expression make_array<double>(1, 2.0) is ill-formed, because the narrowing happens inside the function body where no constant expressions exist anymore. Furthermore given e.g.

int f();
double g();

we probably want to support

make_array<double>(f(), g());

as well. To make this feasible, the currently suggested expansion

{ std::forward<Args>(args)... }

needs to be replaced by

{ static_cast<T>(std::forward<Args>(args))... }

which is safe, because we already ensure convertibility via the element-wise Convertible<Args, T> requirement. Some other fixes are necessary: The ValueType requirement for the function parameters is invalid, because all lvalue arguments will deduce to an lvalue-reference, thereby no longer satisfying this requirement.

The suggested simplification is to provide a default-computed effective type for the result array based on common_type and decay, in unconstrained form:

template<typename... Args>
array<typename decay<typename common_type<Args...>::type>::type,
sizeof...(Args)>
make_array(Args&&... args);

The approach used below is similar to that of make_pair and make_tuple using a symbol C to represent the decayed common type [Note: Special handling of reference_wrapper types is intentionally not provided, because our target has so satisfy ValueType, thus under the revised proposal only an all-reference_wrapper-arguments would be well-formed and an array of reference_wrapper will be constructed]. I do currently not suggest to add new concepts reflecting decay and common_type, but an implementor will need something like this to succeed. Note that we use a similar fuzziness for make_pair and make_tuple currently. This fuzziness is not related to the currently missing Constructible<Vi, Ti&&> requirement for those functions. The following proposal fixes that miss for make_array. If the corresponding C type deduction is explicitly wanted for standardization, here the implementation

auto concept DC<typename... T> {
  typename type = typename decay<typename common_type<T...>::type>::type;
}

where C is identical to DC<Args...>::type in the proposed resolution below.

I intentionally added no further type relation between type and the concept template parameters, but instead added this requirement below to make the specification as transparent as possible. As written this concept is satisfied, if the corresponding associated type exists.

Suggested Resolution:

  1. Add to the array synopsis in 24.3 [sequences]:

    
    template<ReferentType... Args>
    requires ValueType<C> && IdentityOf<Args> && Constructible<C, Args&&>...
    array<C, sizeof...(Args)>
    make_array(Args&&... args);
    
    
  2. Append after 24.3.7.7 [array.tuple] Tuple interface to class template array the following new section:

    23.4.1.7 Array creation functions [array.creation]

    
    template<ReferentType... Args>
    requires ValueType<C> && IdentityOf<Args> && Constructible<C, Args&&>...
    array<C, sizeof...(Args)>
    make_array(Args&&... args);
    

    Let C be decay<common_type<Args...>::type>::type.

    Returns: an array<C, sizeof...(Args)> initialized with { static_cast<C>(std::forward<Args>(args))... }.

[ 2009-07 Frankfurt: ]

The proposed resolution uses concepts.

Daniel to rewrite the proposed resolution.

Leave Open.

[ 2009-07-25 Daniel provides rewritten proposed resolution. ]

[ 2009-10 Santa Cruz: ]

Argument for NAD future: everything about this could be added on. This does not require changes to the existing text.

[2015-11-29, Alisdair comments]

N4391 was adopted for Fundamentals 2 at the Lenexa meeting.

[2017-02 in Kona, LEWG recommends NAD]

[2017-06-02 Issues Telecon]

It seems as if deduction guides can solve most of the problem. In addition to that, make_array is available in Library Fundamentals TS v2. If it's desired to be able to specify the type but not the extent, make_array can do that, and if make_array isn't acceptable for that, we are talking about a language extension in deduction guides, which needs a proposal paper.

Resolve as NAD

Proposed resolution:

  1. Add to the array synopsis in 24.3 [sequences]:

    template<class... Args>
      array<CT, sizeof...(Args)>
      make_array(Args&&... args);
    
  2. Append after 24.3.7.7 [array.tuple] "Tuple interface to class template array" the following new section:

    XX.X.X.X Array creation functions [array.creation]

    
    template<class... Args>
    array<CT, sizeof...(Args)>
    make_array(Args&&... args)
    

    Let CT be decay<common_type<Args...>::type>::type.

    Returns: An array<CT, sizeof...(Args)> initialized with { static_cast<CT>(std::forward<Args>(args))... }.

    [Example:

    
    int i = 0; int& ri = i;
    make_array(42u, i, 2.78, ri);
    

    returns an array of type

    
    array<double, 4>
    

    end example]


855(i). capacity() and reserve() for deque?

Section: 24.3.8.3 [deque.capacity] Status: NAD Submitter: Hervé Brönnimann Opened: 2008-06-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [deque.capacity].

View all issues with NAD status.

Discussion:

The main point is that capacity can be viewed as a mechanism to guarantee the validity of iterators when only push_back/pop_back operations are used. For vector, this goes with reallocation. For deque, this is a bit more subtle: capacity() of a deque may shrink, whereas that of vector doesn't. In a circular buffer impl. of the map, as Howard did, there is very similar notion of capacity: as long as size() is less than B * (total size of the map - 2), it is guaranteed that no iterator is invalidated after any number of push_front/back and pop_front/back operations. But this does not hold for other implementations.

Still, I believe, capacity() can be defined by size() + how many push_front/back minus pop_front/back that can be performed before terators are invalidated. In a classical impl., capacity() = size() + the min distance to either "physical" end of the deque (i.e., counting the empty space in the last block plus all the blocks until the end of the map of block pointers). In Howard's circular buffer impl., capacity() = B * (total size of the map - 2) still works with this definition, even though the guarantee could be made stronger.

A simple picture of a deque:

A-----|----|-----|---F+|++++|++B--|-----|-----Z

(A,Z mark the beginning/end, | the block boundaries, F=front, B=back, and - are uninitialized, + are initialized) In that picture: capacity = size() + min(dist(A,F),dist(B,Z)) = min (dist(A,B),dist(F,Z)).

Reserve(n) can grow the map of pointers and add possibly a number of empty blocks to it, in order to guarantee that the next n-size() push_back/push_front operations will not invalidate iterators, and also will not allocate (i.e. cannot throw). The second guarantee is not essential and can be left as a QoI. I know well enough existing implementations of deque (sgi/stl, roguewave, stlport, and dinkumware) to know that either can be implemented with no change to the existing class layout and code, and only a few modifications if blocks are pre-allocated (instead of always allocating a new block, check if the next entry in the map of block pointers is not zero).

Due to the difference with vector, wording is crucial. Here's a proposed wording to make things concrete; I tried to be reasonably careful but please double-check me:

[ San Francisco: ]

Hans: should the Returns clause for capacity read "1 Returns: A lower bound..." rather than "1 Returns: An upper bound..."

Howard: maybe what's needed is capacity_front and capacity_back. In fact, I think I implemented a deque that had these members as implementation details.

Proposed resolution:

Add new signatures to synopsis in 24.3.8 [deque]:

size_type capacity() const;
bool reserve(size_type n);

Add new signatures to 24.3.8.3 [deque.capacity]:

size_type capacity() const;

1 Returns: An upper bound on n + max(n_f - m_f, n_b - m_b) such that, for any sequence of n_f push_front, m_f pop_front, n_b push_back, and m_b pop_back operations, interleaved in any order, starting with the current deque of size n, the deque does not invalidate any of its iterators except to the erased elements.

2 Remarks: Unlike a vector's capacity, the capacity of a deque can decrease after a sequence of insertions at both ends, even if none of the operations caused the deque to invalidate any of its iterators except to the erased elements.

bool reserve(size_type n);

2 Effects: A directive that informs a deque of a planned sequence of push_front, pop_front, push_back, and pop_back operations, so that it can manage iterator invalidation accordingly. After reserve(), capacity() is greater or equal to the argument of reserve if this operation returns true; and equal to the previous value of capacity() otherwise. If an exception is thrown, there are no effects.

3 Returns: true if iterators are invalidated as a result of this operation, and false otherwise.

4 Complexity: It does not change the size of the sequence and takes at most linear time in n.

5 Throws: length_error if n > max_size().

6 Remarks: It is guaranteed that no invalidation takes place during a sequence of insert or erase operations at either end that happens after a call to reserve() except to the erased elements, until the time when an insertion would make max(n_f-m_f, n_b-m_b) larger than capacity(), where n_f is the number of push_front, m_f of pop_front, n_b of push_back, and m_b of pop_back operations since the call to reserve().

7 An implementation is free to pre-allocate buffers so as to offer the additional guarantee that no exception will be thrown during such a sequence other than by the element constructors.

And 24.3.8.4 [deque.modifiers] para 1, can be enhanced:

1 Effects: An insertion in the middle of the deque invalidates all the iterators and references to elements of the deque. An insertion at either end of the deque invalidates all the iterators to the deque, unless provisions have been made with reserve, but has no effect on the validity of references to elements of the deque.

Rationale:

Complication outweighs the benefit.


862(i). Impossible complexity for 'includes'

Section: 27.8.7.2 [includes] Status: NAD Editorial Submitter: Alisdair Meredith Opened: 2008-07-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [includes].

View all issues with NAD Editorial status.

Discussion:

In 27.8.7.2 [includes] the complexity is "at most -1 comparisons" if passed two empty ranges. I don't know how to perform a negative number of comparisions!

This same issue also applies to:

[ 2009-03-30 Beman adds: ]

Suggest NAD. The complexity of empty ranges is -1 in other places in the standard. See 27.8.6 [alg.merge] merge and inplace_merge, and forward_list merge, for example. The time and effort to find and fix all places in the standard where empty range[s] result in negative complexity isn't worth the very limited benefit.

[ 2009-05-09 Alisdair adds: ]

I'm not happy with NAD if we can find a simple solution.

How about adding a rider somewhere in clause 17 suggesting that complexities that specify a negative number of operations are treated as specifying zero operations? That should generically solve the issue without looking for further cases.

[ Batavia (2009-05): ]

Pete to provide "straightforward" wording. Move to NAD Editorial.

Proposed resolution:

Recommend NAD.


863(i). What is the state of a stream after close() succeeds

Section: 31.10 [file.streams] Status: NAD Submitter: Steve Clamage Opened: 2008-07-08 Last modified: 2017-06-15

Priority: Not Prioritized

View all other issues in [file.streams].

View all issues with NAD status.

Discussion:

Suppose writing to an [o]fstream fails and you later close the stream. The overflow() function is called to flush the buffer (if it exists). Then the file is unconditionally closed, as if by calling flcose.

If either overflow or fclose fails, close() reports failure, and clearly the stream should be in a failed or bad state.

Suppose the buffer is empty or non-existent (so that overflow() does not fail), and fclose succeeds. The close() function reports success, but what is the state of the stream?

[ Batavia (2009-05): ]

Tom's impression is that the issue is about the failbit, etc.

Bill responds that the stream is now closed, and any status bits remain unchanged.

See the description of close() in 31.10.5.4 [fstream.members].

We prefer not to add wording to say that nothing changes. Move to NAD.

Proposed resolution:


864(i). Defect in atomic wording

Section: 33.5.8.2 [atomics.types.operations] Status: NAD Editorial Submitter: Anthony Williams Opened: 2008-07-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.operations].

View all issues with NAD Editorial status.

Discussion:

There's an error in 33.5.8.2 [atomics.types.operations]/p9:

C atomic_load(const volatile A * object);
C atomic_load_explicit(const volatile A * object, memory_order);
C A ::load(memory_order order = memory_order_seq_cst) const volatile;

Requires: The order argument shall not be memory_order_acquire nor memory_order_acq_rel.

I believe that this should state

shall not be memory_order_release.

There's also an error in 33.5.8.2 [atomics.types.operations]/p17:

... When only one memory_order argument is supplied, the value of success is order, and the value of failure is order except that a value of memory_order_acq_rel shall be replaced by the value memory_order_require ...

I believe this should state

shall be replaced by the value memory_order_acquire ...

Proposed resolution:

Change 33.5.8.2 [atomics.types.operations]/p9:

C atomic_load(const volatile A * object);
C atomic_load_explicit(const volatile A * object, memory_order);
C A ::load(memory_order order = memory_order_seq_cst) const volatile;

Requires: The order argument shall not be memory_order_acquire memory_order_release nor memory_order_acq_rel.

Change 33.5.8.2 [atomics.types.operations]/p17:

... When only one memory_order argument is supplied, the value of success is order, and the value of failure is order except that a value of memory_order_acq_rel shall be replaced by the value memory_order_require memory_order_acquire ...

Rationale:

Already fixed by the time the LWG processed it.


867(i). Valarray and value-initialization

Section: 28.6.2.2 [valarray.cons] Status: NAD Editorial Submitter: Alberto Ganesh Barbati Opened: 2008-07-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [valarray.cons].

View all issues with NAD Editorial status.

Discussion:

From 28.6.2.2 [valarray.cons], paragraph 2:

explicit  valarray(size_t);

The array created by this constructor has a length equal to the value of the argument. The elements of the array are constructed using the default constructor for the instantiating type T.

The problem is that the most obvious Ts for valarray are float and double, they don't have a default constructor. I guess the intent is to value-initialize the elements, so I suggest replacing:

The elements of the array are constructed using the default constructor for the instantiating type T.

with

The elements of the array are value-initialized.

There is another reference to the default constructor of T in the non-normative note in paragraph 9. That reference should also be replaced. (The normative wording in paragraph 8 refers to T() and so it doesn't need changes).

[ Batavia (2009-05): ]

We agree with the proposed resolution. Move to NAD Editorial.

Proposed resolution:

Change 28.6.2.2 [valarray.cons], paragraph 2:

explicit  valarray(size_t);

The array created by this constructor has a length equal to the value of the argument. The elements of the array are constructed using the default constructor for the instantiating type T value-initialized (9.4 [dcl.init]).

Change 28.6.2.8 [valarray.members], paragraph 9:

[Example: If the argument has the value -2, the first two elements of the result will be constructed using the default constructor value-initialized (9.4 [dcl.init]); the third element of the result will be assigned the value of the first element of the argument; etc. -- end example]


873(i). signed integral type and unsigned integral type are not clearly defined

Section: 6.8.2 [basic.fundamental] Status: NAD Editorial Submitter: Travis Vitek Opened: 2008-06-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

Neither the term "signed integral type" nor the term "unsigned integral type" is defined in the core language section of the standard, therefore the library section should avoid its use. The terms signed integer type and unsigned integer type are indeed defined (in 6.8.2 [basic.fundamental]), thus the usages should be replaced accordingly.

Note that the key issue here is that "signed" + "integral type" != "signed integral type". The types bool, char, char16_t, char32_t and wchar_t are all listed as integral types, but are neither of signed integer type or unsigned integer type. According to 6.8 [basic.types] p7, a synonym for integral type is integer type. Given this, one may choose to assume that an integral type that can represent values less than zero is a signed integral type. Unfortunately this can cause ambiguities. As an example, if T is unsigned char, the expression make_signed<T>::type, is supposed to name a signed integral type. There are potentially two types that satisfy this requirement, namely signed char and char (assuming CHAR_MIN < 0).

[ San Francisco: ]

Plum, Sebor to review.

[ Post Summit Daniel adds: ]

The proposed resolution needs to be "conceptualized". Currently we have in [concept.support] only concept IntegralType for all "integral types", thus indeed the current Container concept and Iterator concepts are sufficiently satisfied with "integral types". If the changes are applied, we might ask core for concept BilateralIntegerType and add proper restrictions to the library concepts.

Proposed resolution:

I propose to use the terms "signed integer type" and "unsigned integer type" in place of "signed integral type" and "unsigned integral type" to eliminate such ambiguities.

The proposed change makes it absolutely clear that the difference between two pointers cannot be char or wchar_t, but could be any of the signed integer types. 7.6.6 [expr.add] paragraph 6...

  1. When two pointers to elements of the same array object are subtracted, the result is the difference of the subscripts of the two array elements. The type of the result is an implementation-defined signed integral typesigned integer type; this type shall be the same type that is defined as std::ptrdiff_t in the <cstdint> header (18.1)...

The proposed change makes it clear that X::size_type and X::difference_type cannot be char or wchar_t, but could be one of the signed or unsigned integer types as appropriate. 16.4.4.6 [allocator.requirements] table 40...

Table 40: Allocator requirements

expression return type assertion/note/pre/post-condition
X::size_type unsigned integral type unsigned integer type a type that can represent the size of the largest object in the allocation model.
X::difference_type signed integral type signed integer type a type that can represent the difference between any two pointers in the allocation model.

The proposed change makes it clear that make_signed<T>::type must be one of the signed integer types as defined in 3.9.1. Ditto for make_unsigned<T>type and unsigned integer types. 21.3.8.4 [meta.trans.sign] table 48...

Table 48: Sign modifications

Template Comments
template <class T> struct make_signed; If T names a (possibly cv-qualified) signed integral typesigned integer type (3.9.1) then the member typedef type shall name the type T; otherwise, if T names a (possibly cv-qualified) unsigned integral typeunsigned integer type then type shall name the corresponding signed integral typesigned integer type, with the same cv-qualifiers as T; otherwise, type shall name the signed integral typesigned integer type with the smallest rank (4.13) for which sizeof(T) == sizeof(type), with the same cv-qualifiers as T. Requires: T shall be a (possibly cv-qualified) integral type or enumeration but not a bool type.
template <class T> struct make_unsigned; If T names a (possibly cv-qualified) unsigned integral typeunsigned integer type (3.9.1) then the member typedef type shall name the type T; otherwise, if T names a (possibly cv-qualified) signed integral typesigned integer type then type shall name the corresponding unsigned integral typeunsigned integer type, with the same cv-qualifiers as T; otherwise, type shall name the unsigned integral typeunsigned integer type with the smallest rank (4.13) for which sizeof(T) == sizeof(type), with the same cv-qualifiers as T. Requires: T shall be a (possibly cv-qualified) integral type or enumeration but not a bool type.

Note: I believe that the basefield values should probably be prefixed with ios_base:: as they are in 30.4.3.3.3 [facet.num.put.virtuals] The listed virtuals are all overloaded on signed and unsigned integer types, the new wording just maintains consistency. 30.4.3.2.3 [facet.num.get.virtuals] table 78...

Table 78: Integer Conversions

State stdio equivalent
basefield == oct %o
basefield == hex %X
basefield == 0 %i
signed integral typesigned integer type %d
unsigned integral typeunsigned integer type %u

Rationale is same as above. 30.4.3.3.3 [facet.num.put.virtuals] table 80...

Table 80: Integer Conversions

State stdio equivalent
basefield == ios_base::oct %o
(basefield == ios_base::hex) && !uppercase %x
(basefield == ios_base::hex) %X
basefield == 0 %i
for a signed integral typesigned integer type %d
for a unsigned integral typeunsigned integer type %u

24.2 [container.requirements] table 80...

Table 89: Container requirements

expression return type operational semantics assertion/note/pre/post-condition complexity
X::difference_type signed integral typesigned integer type   is identical to the difference type of X::iterator and X::const_iterator compile time
X::size_type unsigned integral typeunsigned integer type   size_type can represent any non-negative value of difference_type compile time

25.3.4 [iterator.concepts] paragraph 1...

Iterators are a generalization of pointers that allow a C++ program to work with different data structures (containers) in a uniform manner. To be able to construct template algorithms that work correctly and efficiently on different types of data structures, the library formalizes not just the interfaces but also the semantics and complexity assumptions of iterators. All input iterators i support the expression *i, resulting in a value of some class, enumeration, or built-in type T, called the value type of the iterator. All output iterators support the expression *i = o where o is a value of some type that is in the set of types that are writable to the particular iterator type of i. All iterators i for which the expression (*i).m is well-defined, support the expression i->m with the same semantics as (*i).m. For every iterator type X for which equality is defined, there is a corresponding signed integral type signed integer type called the difference type of the iterator.

I'm a little unsure of this change. Previously this paragraph would allow instantiations of linear_congruential_engine on char, wchar_t, bool, and other types. The new wording prohibits this. 28.5.4.2 [rand.eng.lcong] paragraph 2...

The template parameter UIntType shall denote an unsigned integral typeunsigned integer type large enough to store values as large as m - 1. If the template parameter m is 0, the modulus m used throughout this section 26.4.3.1 is numeric_limits<result_type>::max() plus 1. [Note: The result need not be representable as a value of type result_type. --end note] Otherwise, the following relations shall hold: a < m and c < m.

Same rationale as the previous change. 99 [rand.adapt.xor] paragraph 6...

Both Engine1::result_type and Engine2::result_type shall denote (possibly different) unsigned integral typesunsigned integer types. The member result_type shall denote either the type Engine1::result_type or the type Engine2::result_type, whichever provides the most storage according to clause 3.9.1.

28.5.8.1 [rand.util.seedseq] paragraph 7...

Requires:RandomAccessIterator shall meet the requirements of a random access iterator (24.1.5) such that iterator_traits<RandomAccessIterator>::value_type shall denote an unsigned integral typeunsigned integer type capable of accomodating 32-bit quantities.

By making this change, integral types that happen to have a signed representation, but are not signed integer types, would no longer be required to use a two's complement representation. This may go against the original intent, and should be reviewed. 33.5.8.2 [atomics.types.operations] paragraph 24...

Remark: For signed integral typessigned integer types, arithmetic is defined using two's complement representation. There are no undefined results. For address types, the result may be an undefined address, but the operations otherwise have no undefined behavior.


877(i). to throw() or to Throw: Nothing.

Section: 16 [library] Status: NAD Submitter: Martin Sebor Opened: 2008-08-23 Last modified: 2018-06-23

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Recent changes to the working draft have introduced a gratuitous inconsistency with the C++ 2003 version of the specification with respect to exception guarantees provided by standard functions. While the C++ 2003 standard consistenly uses the empty exception specification, throw(), to declare functions that are guaranteed not to throw exceptions, the current working draft contains a number of "Throws: Nothing." clause to specify essentially the same requirement. The difference between the two approaches is that the former specifies the behavior of programs that violate the requirement (std::unexpected() is called) while the latter leaves the behavior undefined.

A survey of the working draft reveals that there are a total of 209 occurrences of throw() in the library portion of the spec, the majority in clause 18, a couple (literally) in 19, a handful in 20, a bunch in 22, four in 24, one in 27, and about a dozen in D.9.

There are also 203 occurrences of "Throws: Nothing." scattered throughout the spec.

While sometimes there are good reasons to use the "Throws: Nothing." approach rather than making use of throw(), these reasons do not apply in most of the cases where this new clause has been introduced and the empty exception specification would be a better approach.

First, functions declared with the empty exception specification permit compilers to generate better code for calls to such functions. In some cases, the compiler might even be able to eliminate whole chunks of user-written code when instantiating a generic template on a type whose operations invoked from the template specialization are known not to throw. The prototypical example are the std::uninitialized_copy() and std::uninitialized_fill() algorithms where the entire catch(...) block can be optimized away.

For example, given the following definition of the std::uninitialized_copy function template and a user-defined type SomeType:

template <class InputIterator, class ForwardIterator>
ForwardIterator
uninitialized_copy (InputIterator first, InputIterator last, ForwardIterator res)
{
   typedef iterator_traits<ForwardIterator>::value_type ValueType;

   ForwardIterator start = res;

   try {
       for (; first != last; ++first, ++res)
           ::new (&*res) ValueType (*first);
   }
   catch (...) {
       for (; start != res; --start)
           (&*start)->~ValueType ();
       throw;
   }
   return res;
}

struct SomeType {
   SomeType (const SomeType&) throw ();
}

compilers are able to emit the following efficient specialization of std::uninitialized_copy<const SomeType*, SomeType*> (note that the catch block has been optimized away):

template <> SomeType*
uninitialized_copy (const SomeType *first, const SomeType *last, SomeType *res)
{
   for (; first != last; ++first, ++res)
       ::new (res) SomeType (*first);

   return res;
}

Another general example is default constructors which, when decorated with throw(), allow the compiler to eliminate the implicit try and catch blocks that it otherwise must emit around each the invocation of the constructor in new-expressions.

For example, given the following definitions of class MayThrow and WontThrow and the two statements below:

struct MayThrow {
   MayThrow ();
};

struct WontThrow {
   WontThrow () throw ();
};

MayThrow  *a = new MayThrow [N];
WontThrow *b = new WontThrow [N];

the compiler generates the following code for the first statement:

MayThrow *a;
{
   MayThrow *first = operator new[] (N * sizeof (*a));
   MayThrow *last  = first + N;
   MayThrow *next  = first;
   try {
       for ( ; next != last; ++next)
           new (next) MayThrow;
   }
   catch (...) {
       for ( ; first != first; --next)
           next->~MayThrow ();
       operator delete[] (first);
       throw;
   }
   a = first;
}

but it is can generate much more compact code for the second statement:

WontThrow *b    = operator new[] (N * sizeof (*b));
WontThrow *last = b + N;
for (WontThrow *next = b; next != last; ++next)
   new (next) WontThrow;

Second, in order for users to get the maximum benefit out of the new std::has_nothrow_xxx traits when using standard library types it will be important for implementations to decorate all non throwing copy constructors and assignment operators with throw(). Note that while an optimizer may be able to tell whether a function without an explicit exception specification can throw or not based on its definition, it can only do so when it can see the source code of the definition. When it can't it must assume that the function may throw. To prevent violating the One Definition Rule, the std::has_nothrow_xxx trait must return the most pessimistic guess across all translation units in the program, meaning that std::has_nothrow_xxx<T>::value must evaluate to false for any T whose xxx (where xxx is default or copy ctor, or assignment operator) is defined out-of-line.

Counterarguments:

During the discussion of this issue on lib@lists.isocpp.org (starting with post c++std-lib-21950) the following arguments in favor of the "Throws: Nothing." style have been made.

  1. Decorating functions that cannot throw with the empty exception specification can cause the compiler to generate suboptimal code for the implementation of the function when it calls other functions that aren't known to the compiler not to throw (i.e., that aren't decorated with throw() even if they don't actually throw). This is a common situation when the called function is a C or POSIX function.
  2. Alternate, proprietary mechanisms exist (such as GCC __attribute__((nothrow)) or Visual C++ __declspec(nothrow)) that let implementers mark up non-throwing functions, often without the penalty mentioned in (1) above. The C++ standard shouldn't preclude the use of these potentially more efficient mechanisms.
  3. There are functions, especially function templates, that invoke user-defined functions that may or may not be declared throw(). Declaring such functions with the empty exception specification will cause compilers to generate suboptimal code when the user-defined function isn't also declared not to throw.

The answer to point (1) above is that implementers can (and some have) declare functions with throw() to indicate to the compiler that calls to the function can safely be assumed not to throw in order to allow it to generate efficient code at the call site without also having to define the functions the same way and causing the compiler to generate suboptimal code for the function definition. That is, the function is declared with throw() in a header but it's defined without it in the source file. The throw() declaration is suppressed when compiling the definition to avoid compiler errors. This technique, while strictly speaking no permitted by the language, is safe and has been employed in practice. For example, the GNU C library takes this approach. Microsoft Visual C++ takes a similar approach by simply assuming that no function with C language linkage can throw an exception unless it's explicitly declared to do so using the language extension throw(...).

Our answer to point (2) above is that there is no existing practice where C++ Standard Library implementers have opted to make use of the proprietary mechanisms to declare functions that don't throw. The language provides a mechanism specifically designed for this purpose. Avoiding its use in the specification itself in favor of proprietary mechanisms defeats the purpose of the feature. In addition, making use of the empty exception specification inconsistently, in some areas of the standard, while conspicuously avoiding it and making use of the "Throws: Nothing." form in others is confusing to users.

The answer to point (3) is simply to exercise caution when declaring functions and especially function templates with the empty exception specification. Functions that required not to throw but that may call back into user code are poor candidates for the empty exception specification and should instead be specified using "Throws: Nothing." clause.

[ 2009-07 Frankfurt ]

We need someone to do an extensive review.

NAD Future.

[2017-02 in Kona, LEWG recommends NAD]

The discussed discrepancy isn't relevant any longer: now we have noexcept and have deprecated throw(). Additionally, the guidance on narrow vs. wide contracts, Requires clauses, and noexcept/Throws means that the proposed resolution is more subtle even if updated in terms of noexcept().

[2017-06-02 Issues Telecon]

Resolve as NAD

Proposed resolution:

We propose two possible solutions. Our recommendation is to adopt Option 1 below.

Option 1:

Except for functions or function templates that make calls back to user-defined functions that may not be declared throw() replace all occurrences of the "Throws: Nothing." clause with the empty exception specification. Functions that are required not to throw but that make calls back to user code should be specified to "Throw: Nothing."

Option 2:

For consistency, replace all occurrences of the empty exception specification with a "Throws: Nothing." clause.


879(i). Atomic load const qualification

Section: 33.5 [atomics] Status: NAD Editorial Submitter: Alexander Chemeris Opened: 2008-08-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics].

View all issues with NAD Editorial status.

Discussion:

The atomic_address type and atomic<T*> specialization provide atomic updates to pointers. However, the current specification requires that the types pointer be to non-const objects. This restriction is unnecessary and unintended.

[ Summit: ]

Move to review. Lawrence will first check with Peter whether the current examples are sufficient, or whether they need to be expanded to include all cases.

[ 2009-07 Frankfurt ]

Lawrence will handle all issues relating to atomics in a single paper.

LWG will defer discussion on atomics until that paper appears.

Move to Open.

[ 2009-08-17 Handled by N2925. ]

[ 2009-10 Santa Cruz: ]

NAD Editorial. Solved by N2992.

Proposed resolution:

Add const qualification to the pointer values of the atomic_address and atomic<T*> specializations. E.g.

typedef struct atomic_address {
   void store(const void*, memory_order = memory_order_seq_cst) volatile;
   void* exchange( const void*, memory_order = memory_order_seq_cst) volatile;
   bool compare_exchange( const void*&, const void*,
                          memory_order, memory_order) volatile;
   bool compare_exchange( const void*&, const void*,
                          memory_order = memory_order_seq_cst ) volatile;
   void* operator=(const void*) volatile;
} atomic_address;

void atomic_store(volatile atomic_address*, const void*);
void atomic_store_explicit(volatile atomic_address*, const void*,
                          memory_order);
void* atomic_exchange(volatile atomic_address*, const void*);
void* atomic_exchange_explicit(volatile atomic_address*, const void*,
                              memory_order);
bool atomic_compare_exchange(volatile atomic_address*,
                            const void**, const void*);
bool atomic_compare_exchange_explicit(volatile atomic_address*,
                                     const void**, const void*,
                                     memory_order, memory_order);

887(i). issue with condition::wait_...

Section: 33.7.4 [thread.condition.condvar] Status: NAD Submitter: Lawrence Crowl Opened: 2008-09-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.condition.condvar].

View all issues with NAD status.

Discussion:

The Posix/C++ working group has identified an inconsistency between Posix and the C++ working draft in that Posix requires the clock to be identified at creation, whereas C++ permits identifying the clock at the call to wait. The latter cannot be implemented with the former.

[ San Francisco: ]

Howard recommends NAD with the following explanation:

The intent of the current wording is for the condtion_variable::wait_until be able to handle user-defined clocks as well as clocks the system knows about. This can be done by providing overloads for the known clocks, and another overload for unknown clocks which synchs to a known clock before waiting. For example:

template <class Duration>
bool
condition_variable::wait_until(unique_lock<mutex>& lock,
                               const chrono::time_point<chrono::system_clock, Duration>& abs_time)
{
    using namespace chrono;
    nanoseconds d = __round_up<nanoseconds>(abs_time.time_since_epoch());
    __do_timed_wait(lock.mutex()->native_handle(), time_point<system_clock, nanoseconds>(d));
    return system_clock::now() < abs_time;
}

template <class Clock, class Duration>
bool
condition_variable::wait_until(unique_lock<mutex>& lock,
                               const chrono::time_point<Clock, Duration>& abs_time)
{
    using namespace chrono;
    system_clock::time_point    s_entry = system_clock::now();
    typename Clock::time_point  c_entry = Clock::now();
    nanoseconds dn = __round_up<nanoseconds>(abs_time.time_since_epoch() -
                                              c_entry.time_since_epoch());
    __do_timed_wait(lock.mutex()->native_handle(), s_entry + dn);
    return Clock::now() < abs_time;
}

In the above example, system_clock is the only clock which the underlying condition variable knows how to deal with. One overload just passes that clock through. The second overload (approximately) converts the unknown clock into a system_clock time_point prior to passing it down to the native condition variable.

On Posix systems vendors are free to add implementation defined constructors which take a clock. That clock can be stored in the condition_variable, and converted to (or not as necessary) as shown above.

If an implementation defined constructor takes a clock (for example), then part of the semantics for that implementation defined ctor might include that a wait_until using a clock other than the one constructed with results in an error (exceptional condition) instead of a conversion to the stored clock. Such a design is up to the vendor as once an implementation defined ctor is used, the vendor is free to specifiy the behavior of waits and/or notifies however he pleases (when the cv is constructed in an implementation defined manner).

[ Post Summit: ]

"POSIX people will review the proposed NAD resolution at their upcoming NY meeting.

See the minutes at: http://wiki.dinkumware.com/twiki/bin/view/Posix/POSIX-CppBindingWorkingGroupNewYork2009.

[ 2009-07 Frankfurt ]

Move to NAD.

[ 2009-07-18 Detlef reopens the issue: ]

On Friday afternoon in Frankfurt is was decided that 887 is NAD. This decision was mainly based on a sample implementation presented by Howard that implemented one clock on top of another. Unfortunately this implementation doesn't work for the probably most important case where a system has a monotonic clock and a real-time clock (or "wall time" clock):

If the underlying "system_clock" is a monotonic clock, and the program waits on the real-time clock, and the real-time clock is set forward, the wait will unblock too late.

If the underlying "system_clock" is a real-time clock, and the program waits on the monotonic clock, and the real-time clock is set back, the wait again will unblock too late.

Sorry that I didn't remember this on Friday, but it was Friday afternoon after a busy week...

So as the decision was made on a wrong asumption, I propose to re-open the issue.

[ 2009-07-26 Howard adds: ]

Detlef correctly argues that condition_variable::wait_until could return "too late" in the context of clocks being adjusted during the wait. I agree with his logic. But I disagree that this makes this interface unimplementable on POSIX.

The POSIX spec also does not guarantee that pthread_cond_timedwait does not return "too late" when clocks are readjusted during the wait. Indeed, the POSIX specification lacks any requirements at all concerning how soon pthread_cond_timedwait returns after a time out. This is evidently a QOI issue by the POSIX standard. Here is a quote of the most relevant normative text concerning pthread_cond_timedwait found here.

The pthread_cond_timedwait() function shall be equivalent to pthread_cond_wait(), except that an error is returned if the absolute time specified by abstime passes (that is, system time equals or exceeds abstime) before the condition cond is signaled or broadcasted, or if the absolute time specified by abstime has already been passed at the time of the call.

I.e. the POSIX specification speaks of the error code returned in case of a time out, but not on the timeliness of that return.

Might this simply be an oversight, or minor defect in the POSIX specification?

I do not believe so. This same section goes on to say in non-normative text:

For cases when the system clock is advanced discontinuously by an operator, it is expected that implementations process any timed wait expiring at an intervening time as if that time had actually occurred.

Here is non-normative wording encouraging the implementation to ignore an advancing underlying clock and subsequently causing an early (spurious) return. There is no wording at all which addresses Detlef's example of a "late return". With pthread_cond_timedwait this would be caused by setting the system clock backwards. It seems reasonable to assume, based on the wording that is already in the POSIX spec, that again, the discontinuously changed clock would be ignored by pthread_cond_timedwait.

A noteworthy difference between pthread_cond_timedwait and condition_variable::wait_until is that the POSIX spec appears to say that ETIMEDOUT should be returned if pthread_cond_timedwait returns because of timeout signal, whether or not the system clock was discontinuously advanced during the wait. In contrast condition_variable::wait_until always returns:

Clock::now() < abs_time

That is, the C++ spec requires that the clock be rechecked (detecting discontinuous adjustments during the wait) at the time of return. condition_variable::wait_until may indeed return early or late. But regardless it will return a value reflecting timeout status at the time of return (even if clocks have been adjusted). Of course the clock may be adjusted after the return value is computed but before the client has a chance to read the result of the return. Thus there are no iron-clad guarantees here.

condition_variable::wait_until (and pthread_cond_timedwait) is little more than a convenience function for making sure condition_variable::wait doesn't hang for an unreasonable amount of time (where the client gets to define "unreasonable"). I do not think it is in anyone's interest to try to make it into anything more than that.

I maintain that this is a useful and flexible specification in the spirit of C++, and is implementable on POSIX. The implementation technique described above is a reasonable approach. There may also be higher quality approaches. This specification, like the POSIX specification, gives a wide latitude for QOI.

I continue to recommend NAD, but would not object to a clarifying note regarding the behavior of condition_variable::wait_until. At the moment, I do not have good wording for such a note, but welcome suggestions.

[ 2009-09-30: See N2969. ]

[ 2009-10 Santa Cruz: ]

The LWG is in favor of Detlef to supply revision which adopts Option 2 from N2969 but is modified by saying that system_clock must be available for wait_until.

[ 2010-02-11 Anthony provided wording. ]

[ 2010-02-22 Anthony adds: ]

I am strongly against N2999.

Firstly, I think that the most appropriate use of a timed wait on a condition variable is with a monotonic clock, so it ought to be guaranteed to be available on systems that support such a clock. Also, making the set of supported clocks implementation defined essentially kills portability around the use of user-defined clocks.

I also think that wait_for is potentially useful, and trivially implementable given a working templated wait_until and a monotonic clock.

I also disagree with many of Detlef's points in the rationale. In a system with hard latency limits there is likely to be a monotonic clock, otherwise you have no way of measuring against these latency limits since the system_clock may change arbitrarily. In such systems, you want to be able to use wait_for, or wait_until with a monotonic clock.

I disagree that the wait_* functions cannot be implemented correctly on top of POSIX: I have done so. The only guarantee in the working draft is that when the function returns certain properties are true; there is no guarantee that the function will return immediately that the properties are true. My resolution to issue 887 makes this clear. How small the latency is is QoI.

On systems without a monotonic clock, you cannot measure the problem since the system clock can change arbitrarily so any timing calculations you make may be wrong due to clock changes.

On systems with a monotonic clock, you can choose to use it for your condition variables. If you are waiting against a system_clock::time_point then you can check the clock when waking, and either return as a timeout or spurious wake depending on whether system_clock::now() is before or after the specified time_point.

Windows does provide condition variables from Vista onwards. I choose not to use them, but they are there. If people are concerned about implementation difficulty, the Boost implementation can be used for most purposes; the Boost license is pretty liberal in that regard.

My preferred resolution to issue 887 is currently the PR in the issues list.

[ 2010 Pittsburgh: ]

There is no consensus for moving the related paper N2999 into the WP.

There was support for moving this issue as proposed to Ready, but the support was insufficient to call a consensus.

There was consensus for moving this issue to NAD as opposed to leaving it open. Rationale added.

Rationale:

The standard as written is sufficiently implementable and self consistent.

Proposed resolution:

Add a new paragraph after 33.2.4 [thread.req.timing]p3:

3 The resolution of timing provided by an implementation depends on both operating system and hardware. The finest resolution provided by an implementation is called the native resolution.

If a function in this clause takes a timeout argument, and the time point or elapsed time specified passes before the function returns, the latency between the timeout occurring and the function returning is unspecified [Note: Implementations should strive to keep such latency as small as possible, but portable code should not rely on any specific upper limits — end note]


892(i). Forward_list issues...

Section: 24.3.9.6 [forward.list.ops] Status: NAD Editorial Submitter: Ed Smith-Rowland Opened: 2008-09-15 Last modified: 2023-02-07

Priority: Not Prioritized

View all other issues in [forward.list.ops].

View all issues with NAD Editorial status.

Discussion:

I was looking at the latest draft on forward_list. Especially the splice methods.

The first one splices a whole list after a given iterator in this. The name is splice_after. I think in [forwardlist.ops] paragraph 40 change:

Effect: Insert the contents of x before after position, ...

A deeper issue involves the complexity. forward_list has no size and we don't know when we've reached the end except to walk up to it. To splice we would need to hook the end of the source list to the item after position in this list. This would involve walking length of the source list until we got to the last dereference-able element in source. There's no way we could do this in O(1) unless we stored a bogus end in forward_list.

OTOH, the last version of splice_after with iterator ranges we could do in O(1) because we know how to hook the end of the source range to ...

Unless I'm misconceiving the whole thing. Which is possible. I'll look at it again.

I'm pretty sure about the first part though.

[ San Francisco: ]

This issue is more complicated than it looks.

paragraph 47: replace each (first, last) with (first, last]

add a statement after paragraph 48 that complexity is O(1)

remove the complexity statement from the first overload of splice_after

We may have the same problems with other modifiers, like erase_after. Should it require that all iterators in the range (position, last] be dereferenceable?

We do, however, like the proposed changes and consider them Editorial. Move to NAD Editorial, Pending. Howard to open a new issue to handle the problems with the complexity requirements.

Opened 897.

Proposed resolution:

In [forwardlist.ops] paragraph 40 change:

Effect: Insert the contents of x before after position, ...


895(i). "Requires:" on std::string::at et al

Section: 16.3.2.4 [structure.specifications] Status: Dup Submitter: James Dennett Opened: 2008-09-16 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [structure.specifications].

View all other issues in [structure.specifications].

View all issues with Dup status.

Duplicate of: 625

Discussion:

Per discussion, we need an issue open to cover looking at "Requires" clauses which are not constraints on user code, such as that on std::basic_string::at.

[ 2009-07 Frankfurt ]

Alan to address in paper.

Proposed resolution:


901(i). insert iterators can move from lvalues

Section: 25.5.2.4 [insert.iterator] Status: NAD Submitter: Alisdair Meredith Opened: 2008-09-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses UK 282

The requires clause on the const T & overloads in back_insert_iterator/front_insert_iterator/insert_iterator mean that the assignment operator will implicitly move from lvalues of a move-only type.

Suggested resolutions are:

  1. Add another overload with a negative constraint on copy-constructible and flag it "= delete".
  2. Drop the copy-constructible overload entirely and rely on perfect forwarding to catch move issues one level deeper.
  3. This is a fundamental problem in move-syntax that relies on the presence of two overloads, and we need to look more deeply into this area as a whole - do not solve this issue in isolation.

[ Post Summit, Alisdair adds: ]

Both comment and issue have been resolved by the adoption of N2844 (rvalue references safety fix) at the last meeting.

Suggest resolve as NAD Editorial with a reference to the paper.

[ Batavia (2009-05): ]

We agree that this has been resolved in the latest Working Draft. Move to NAD.

Proposed resolution:

Recommend NAD, addressed by N2844.


902(i). Regular is the wrong concept to constrain numeric_limits

Section: 17.3.5 [numeric.limits] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2008-09-24 Last modified: 2017-06-15

Priority: Not Prioritized

View other active issues in [numeric.limits].

View all other issues in [numeric.limits].

View all issues with NAD Concepts status.

Discussion:

Addresses FR 32 and DE 16

numeric_limits has functions specifically designed to return NaNs, which break the model of Regular (via its axioms.) While floating point types will be acceptible in many algorithms taking Regular values, it is not appopriate for this specific API and we need a less refined constraint.

FR 32:

The definition of numeric_limits<> as requiring a regular type is both conceptually wrong and operationally illogical. As we pointed before, this mistake needs to be corrected. For example, the template can be left unconstrained. In fact this reflects a much more general problem with concept_maps/axioms and their interpretations. It appears that the current text heavily leans toward experimental academic type theory.

DE 16:

The class template numeric_limits should not specify the Regular concept requirement for its template parameter, because it contains functions returning NaN values for floating-point types; these values violate the semantics of EqualityComparable.

[ Summit: ]

Move to Open. Alisdair and Gaby will work on a solution, along with the new treatment of axioms in clause 14.

Proposed resolution:


903(i). back_insert_iterator issue

Section: 25.5.2.2 [back.insert.iterator] Status: NAD Submitter: Dave Abrahams Opened: 2008-09-19 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [back.insert.iterator].

View all issues with NAD status.

Discussion:

I just noticed this; don't know how far the problem(?) extends or whether it's new or existing: back_insert_iterator's operator* is not const, so you can't dereference a const one.

[ Post Summit Daniel adds: ]

If done, this change should be applied for front_insert_iterator, insert_iterator, ostream_iterator, and ostreambuf_iterator as well.

[ Batavia (2009-05): ]

Alisdair notes that these all are output iterators. Howard points out that ++*i would no longer work if we made this change.

Move to NAD.

[ 2009-05-25 Daniel adds: ]

  1. If 1009 is accepted, OutputIterator does no longer support post increment.
  2. To support backward compatibility a second overload of operator* can be added. Note that the HasDereference concept (and the HasDereference part of concept Iterator) was specifically refactored to cope with optional const qualification and to properly reflect the dual nature of built-in operator* as of 12.6 [over.literal]/6.

Proposed resolution:


905(i). Mutex specification questions

Section: 33.6.4.2.2 [thread.mutex.class] Status: Dup Submitter: Herb Sutter Opened: 2008-09-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.mutex.class].

View all issues with Dup status.

Duplicate of: 893

Discussion:

A few questions on the current WP, N2723:

33.6.4 [thread.mutex.requirements]/24 says an expression mut.unlock() "Throws: Nothing." I'm assuming that, per 99 [res.on.required], errors that violate the precondition "The calling thread shall own the mutex" opens the door for throwing an exception anyway, such as to report unbalanced unlock operations and unlocking from a thread that does not have ownership. Right?

33.6.4.2.2 [thread.mutex.class]/3 (actually numbered paragraph "27" in the WP; this is just a typo I think) says

The behavior of a program is undefined if:

As already discussed, I think the second bullet should be removed, and such a lock() or try_lock() should fail with an exception or returning false, respectively.

A potential addition to the list would be

but without that the status quo text endorses the technique of the program logically transferring ownership of a mutex to another thread with correctness enforced by programming discipline. Was that intended?

[ Summit: ]

Two resolutions: "not a defect" and "duplicate", as follows:

Proposed resolution:


906(i). ObjectType is the wrong concept to constrain initializer_list

Section: 17.10 [support.initlist] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2008-09-26 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [support.initlist].

View all other issues in [support.initlist].

View all issues with NAD Concepts status.

Discussion:

The currently proposed constraint on initializer_list's element type E is that is has to meet ObjectType. This is an underspecification, because both core language and library part of initializer_list make clear, that it references an implicitly allocated array:

9.4.5 [dcl.init.list]/4:

When an initializer list is implicitly converted to a std::initializer_list<E>, the object passed is constructed as if the implementation allocated an array of N elements of type E, where N is the number of elements in the initializer list.[..]

17.10 [support.initlist]/2.

An object of type initializer_list<E> provides access to an array of objects of type const E.[..]

Therefore, E needs to fulfill concept ValueType (thus excluding abstract class types). This stricter requirement should be added to prevent deep instantiation errors known from the bad old times, as shown in the following example:

// Header A: (Should concept-check even in stand-alone modus)

template <DefaultConstructible T>
requires MoveConstructible<T>
void generate_and_do_3(T a) {
  std::initializer_list<T> list{T(), std::move(a), T()};
  ...
}

void do_more();
void do_more_or_less();

template <DefaultConstructible T>
requires MoveConstructible<T>
void more_generate_3() {
  do_more();
  generate_and_do_3(T());
}

template <DefaultConstructible T>
requires MoveConstructible<T>
void something_and_generate_3() {
  do_more_or_less();
  more_generate_3();
}

// Test.cpp

#include "A.h"

class Abstract {
public:
  virtual ~Abstract();
  virtual void foo() = 0; // abstract type
  Abstract(Abstract&&){} // MoveConstructible
  Abstract(){} // DefaultConstructible
};

int main() {
  // The restricted template *accepts* the argument, but
  // causes a deep instantiation error in the internal function
  // generate_and_do_3:
  something_and_generate_3<Abstract>();
}

The proposed stricter constraint does not minimize the aim to support more general containers for which ObjectType would be sufficient. If such an extended container (lets assume it's still a class template) provides a constructor that accepts an initializer_list only this constructor would need to be restricted on ValueType:

template<ObjectType T>
class ExtContainer {
public:
  requires ValueType<T>
  ExtContainer(std::initializer_list<T>);
  ...
};

[ Batavia (2009-05): ]

Move to Tentatively Ready.

[ 2009-07 Frankfurt: ]

Need to look at again without concepts.

Proposed resolution:

  1. In 17.10 [support.initlist]/p.1 replace in "header <initializer_list> synopsis" the constraint "ObjectType" in the template parameter list by the constraint "ValueType".

910(i). Effects of MoveAssignable

Section: 16.4.4.2 [utility.arg.requirements] Status: NAD Concepts Submitter: Alberto Ganesh Barbati Opened: 2008-09-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [utility.arg.requirements].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 150

The description of the effect of operator= in the MoveAssignable concept, given in paragraph 7 is:

result_type  T::operator=(T&&  rv);  // inherited from HasAssign<T, T&&>

Postconditions: the constructed T object is equivalent to the value of rv before the assignment. [Note: there is no requirement on the value of rv after the assignment. --end note]

The sentence contains a typo (what is the "constructed T object"?) probably due to a cut&paste from MoveConstructible. Moreover, the discussion of LWG issue 675 shows that the postcondition is too generic and might not reflect the user expectations. An implementation of the move assignment that just calls swap() would always fulfill the postcondition as stated, but might have surprising side-effects in case the source rvalue refers to an object that is not going to be immediately destroyed. See LWG issue 900 for another example. Due to the sometimes intangible nature of the "user expectation", it seems difficult to have precise normative wording that could cover all cases without introducing unnecessary restrictions. However a non-normative clarification could be a very helpful warning sign that swapping is not always the correct thing to do.

[ 2009-05-09 Alisdair adds: ]

Issue 910 is exactly the reason BSI advanced the Editorial comment UK-150.

The post-conditions after assignment are at a minimum that the object referenced by rv must be safely destructible, and the transaction should not leak resources. Ideally it should be possible to simply assign rv a new valid state after the call without invoking undefined behaviour, but any other use of the referenced object would depend upon additional guarantees made by that type.

[ 2009-05-09 Howard adds: ]

The intent of the rvalue reference work is that the moved from rv is a valid object. Not one in a singular state. If, for example, the moved from object is a vector, one should be able to do anything on that moved-from vector that you can do with any other vector. However you would first have to query it to find out what its current state is. E.g. it might have capacity, it might not. It might have a non-zero size, it might not. But regardless, you can push_back on to it if you want.

That being said, most standard code is now conceptized. That is, the concepts list the only operations that can be done with templated types - whether or not the values have been moved from.

Here is user-written code which must be allowed to be legal:

#include <vector>
#include <cstdio>

template <class Allocator>
void
inspect(std::vector<double, Allocator>&& v)
{
    std::vector<double, Allocator> result(move(v));
    std::printf("moved from vector has %u size and %u capacity\n", v.size(), v.capacity());
    std::printf("The contents of the vector are:\n");
    typedef typename std::vector<double, Allocator>::iterator I;
    for (I i = v.begin(), e = v.end(); i != e; ++i)
        printf("%f\n", *i);
}

int main()
{
    std::vector<double> v1(100, 5.5);
    inspect(move(v1));
}

The above program does not treat the moved-from vector as singular. It only treats it as a vector with an unknown value.

I believe the current proposed wording is consistent with my view on this.

[ Batavia (2009-05): ]

We agree that the proposed resolution is an improvement over the current wording.

[ 2009-07 Frankfurt: ]

Need to look at again without concepts.

[ 2009-07 Frankfurt: ]

Walter will consult with Dave and Doug.

[ 2009-10 Santa Cruz: ]

We believe this is handled by the resolution to issue 1204, but there is to much going on in this area to be sure. Defer for now.

[ 2010-01-23 Moved to Tentatively NAD Concepts after 5 positive votes on c++std-lib. Rationale added below. ]

Rationale:

The current MoveAssignable requirements say everything that can be said in general. Each std-defined type has a more detailed specification of move assignment.

Proposed resolution:

In [concept.copymove], replace the postcondition in paragraph 7 with:

Postconditions: *this is equivalent to the value of rv before the assignment. [Note: there is no requirement on the value of rv after the assignment, but the effect should be unsurprising to the user even in case rv is not immediately destroyed. This may require that resources previously owned by *this are released instead of transferred to rv. -- end note]


912(i). Array swap needs to be conceptualized

Section: 27.7.3 [alg.swap] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2008-10-01 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.swap].

View all issues with NAD Concepts status.

Discussion:

With the adaption of 809 we have a new algorithm swap for C-arrays, which needs to be conceptualized.

[ Post Summit Daniel adds: ]

Recommend as NAD Editorial: The changes have already been applied to the WP N2800.

[ Batavia (2009-05): ]

Move to NAD; the changes have already been made.

Proposed resolution:

Replace in 27.7.3 [alg.swap] before p. 3 until p. 4 by

template <class ValueType T, size_t N>
requires Swappable<T>
void swap(T (&a)[N], T (&b)[N]);

Requires: T shall be Swappable.

Effects: swap_ranges(a, a + N, b);


913(i). Superfluous requirements for replace algorithms

Section: 27.7.5 [alg.replace] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2008-10-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.replace].

View all issues with NAD Concepts status.

Discussion:

(A) 27.7.5 [alg.replace]/1:

Requires: The expression *first = new_value shall be valid.

(B) 27.7.5 [alg.replace]/4:

Requires: The results of the expressions *first and new_value shall be writable to the result output iterator.[..]

Since conceptualization, the quoted content of these clauses is covered by the existing requirements

(A) OutputIterator<Iter, const T&>

and

(B) OutputIterator<OutIter, InIter::reference> && OutputIterator<OutIter, const T&>

resp, and thus should be removed.

[ Batavia (2009-05): ]

We agree with the proposed resolution.

Move to Tentatively Ready.

Proposed resolution:

  1. Remove 27.7.5 [alg.replace]/1.

    template<ForwardIterator Iter, class T> 
      requires OutputIterator<Iter, Iter::reference> 
            && OutputIterator<Iter, const T&> 
            && HasEqualTo<Iter::value_type, T> 
      void replace(Iter first, Iter last, 
                   const T& old_value, const T& new_value); 
    
    template<ForwardIterator Iter, Predicate<auto, Iter::value_type> Pred, class T> 
      requires OutputIterator<Iter, Iter::reference> 
            && OutputIterator<Iter, const T&> 
            && CopyConstructible<Pred> 
      void replace_if(Iter first, Iter last, 
                      Pred pred, const T& new_value);
    

    1 Requires: The expression *first = new_value shall be valid.

  2. 27.7.5 [alg.replace]/4: Remove the sentence "The results of the expressions *first and new_value shall be writable to the result output iterator.".

    template<InputIterator InIter, typename OutIter, class T> 
      requires OutputIterator<OutIter, InIter::reference> 
            && OutputIterator<OutIter, const T&> 
            && HasEqualTo<InIter::value_type, T> 
      OutIter replace_copy(InIter first, InIter last, 
                           OutIter result, 
                           const T& old_value, const T& new_value);
    
    template<InputIterator InIter, typename OutIter,
             Predicate<auto, InIter::value_type> Pred, class T> 
      requires OutputIterator<OutIter, InIter::reference> 
            && OutputIterator<OutIter, const T&> 
            && CopyConstructible<Pred> 
      OutIter replace_copy_if(InIter first, InIter last, 
                              OutIter result, 
                              Pred pred, const T& new_value);
    

    4 Requires: The results of the expressions *first and new_value shall be writable to the result output iterator. The ranges [first,last) and [result,result + (last - first)) shall not overlap.


914(i). Superfluous requirement for unique

Section: 27.7.9 [alg.unique] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2008-10-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.unique].

View all issues with NAD Concepts status.

Discussion:

27.7.9 [alg.unique]/2: "Requires: The comparison function shall be an equivalence relation."

The essence of this is already covered by the given requirement

EquivalenceRelation<auto, Iter::value_type> Pred

and should thus be removed.

[ Batavia (2009-05): ]

We agree with the proposed resolution. Move to Tentatively Ready.

Proposed resolution:

Remove 27.7.9 [alg.unique]/2

template<ForwardIterator Iter>
  requires OutputIterator<Iter, Iter::reference>
        && EqualityComparable<Iter::value_type>
  Iter unique(Iter first, Iter last);

template<ForwardIterator Iter, EquivalenceRelation<auto, Iter::value_type> Pred>
  requires OutputIterator<Iter, RvalueOf<Iter::reference>::type>
        && CopyConstructible<Pred>
  Iter unique(Iter first, Iter last,
               Pred pred);

1 Effects: ...

2 Requires: The comparison function shall be an equivalence relation.


915(i). minmax with initializer_list should return pair of T, not pair of const T&

Section: 27.8.9 [alg.min.max] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2008-10-04 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.min.max].

View all issues with NAD Editorial status.

Discussion:

It seems that the proposed changes for N2772 were not clear enough in this point:

27.8.9 [alg.min.max], before p.23 + p.24 + before p. 27 + p. 28 say that the return type of the minmax overloads with an initializer_list is pair<const T&, const T&>, which is inconsistent with the decision for the other min/max overloads which take a initializer_list as argument and return a T, not a const T&. Doing otherwise for minmax would easily lead to unexpected life-time problems by using minmax instead of min and max separately.

[ Batavia (2009-05): ]

We agree with the proposed resolution. Move to Tentatively Ready.

[ 2009-07 Frankfurt ]

Moved from Tentatively Ready to Open only because the wording needs to be tweaked for concepts removal.

[ 2009-08-18 Daniel adds: ]

Recommend NAD since the proposed changes have already been performed as part of editorial work of N2914.

[ 2009-10 Santa Cruz: ]

Can't find initializer_list form of minmax anymore, only variadic version. Seems like we had an editing clash with concepts. Leave Open, at least until editorial issues resolved. Bring this to Editor's attention.

[ 2010 Pittsburgh: Pete to reapply N2772. ]

Rationale:

Solved by reapplying N2772.

Proposed resolution:

  1. In 27 [algorithms]/2, header <algorithm> synopsis change as indicated:

    template<classLessThanComparable T>
    requires CopyConstructible<T>
    pair<const T&, const T&>
    minmax(initializer_list<T> t);
    
    template<class T, classStrictWeakOrder<auto, T> Compare>
    requires CopyConstructible<T>
    pair<const T&, const T&>
    minmax(initializer_list<T> t, Compare comp);
    
  2. In 27.8.9 [alg.min.max] change as indicated (Begin: Just before p.20):

    template<classLessThanComparable T>
      requires CopyConstructible<T>
      pair<const T&, const T&>
      minmax(initializer_list<T> t);
    

    -20- Requires: T is LessThanComparable and CopyConstructible.

    -21- Returns: pair<const T&, const T&>(x, y) where x is the smallest value and y the largest value in the initializer_list.

    [..]

    template<class T, classStrictWeakOrder<auto, T> Compare>
      requires CopyConstructible<T>
      pair<const T&, const T&>
      minmax(initializer_list<T> t, Compare comp);
    

    -24- Requires: type T is LessThanComparable and CopyConstructible.

    -25- Returns: pair<const T&, const T&>(x, y) where x is the smallest value and y largest value in the initializer_list.


916(i). Redundant move-assignment operator of pair should be removed

Section: 22.3 [pairs] Status: NAD Submitter: Daniel Krügler Opened: 2008-10-04 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [pairs].

View all issues with NAD status.

Discussion:

see also 917.

The current WP provides the following assignment operators for pair in 22.3 [pairs]/1:

  1. template<class U , class V>
    requires HasAssign<T1, const U&> && HasAssign<T2, const V&>
    pair& operator=(const pair<U , V>& p);
    
  2. requires MoveAssignable<T1> && MoveAssignable<T2> pair& operator=(pair&& p );
    
  3. template<class U , class V>
    requires HasAssign<T1, RvalueOf<U>::type> && HasAssign<T2, RvalueOf<V>::type>
    pair& operator=(pair<U , V>&& p);
    

It seems that the functionality of (2) is completely covered by (3), therefore (2) should be removed.

[ Batavia (2009-05): ]

Bill believes the extra assignment operators are necessary for resolving ambiguities, but that does not mean it needs to be part of the specification.

Move to Open. We recommend this be looked at in the context of the ongoing work related to the pair templates.

[ 2009-07 Frankfurt: ]

Leave this open pending the removal of concepts from the WD.

[ 2009-10 Santa Cruz: ]

Mark as NAD, see issue 801.

Proposed resolution:

  1. In 22.3 [pairs] p. 1, class pair and just before p. 13 remove the declaration:

    requires MoveAssignable<T1> && MoveAssignable<T2> pair& operator=(pair&& p );
    
  2. Remove p.13+p.14

917(i). Redundant move-assignment operator of tuple should be removed

Section: 22.4.4.1 [tuple.cnstr] Status: NAD Submitter: Daniel Krügler Opened: 2008-10-04 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [tuple.cnstr].

View all other issues in [tuple.cnstr].

View all issues with NAD status.

Discussion:

see also 916.

N2770 (and thus now the WP) removed the non-template move-assignment operator from tuple's class definition, but the latter individual member description does still provide this operator. Is this (a) an oversight and can it (b) be solved as part of an editorial process?

[ Post Summit Daniel provided wording. ]

[ Batavia (2009-05): ]

We believe that the proposed resolution's part 1 is editorial.

Regarding part 2, we either remove the specification as proposed, or else add back the declaration to which the specification refers. Alisdair and Bill prefer the latter. It is not immediately obvious whether the function is intended to be present.

We recommend that the Project Editor restore the missing declaration and that we keep part 2 of the issue alive.

Move to Open.

[ 2009-07 Frankfurt: ]

Leave this open pending the removal of concepts from the WD.

[ 2009-10 Santa Cruz: ]

Mark as NAD, see issue 801.

Proposed resolution:

  1. In 22.4.4 [tuple.tuple], class tuple just before member swap please change as indicated:

    [ This fixes an editorial loss between N2798 to N2800 ]

    template <class... UTypes>
    requires HasAssign<Types, const UTypes&>...
    tuple& operator=(const pair<UTypes...>&);
    
    template <class... UTypes>
    requires HasAssign<Types, RvalueOf<UTypes>::type>...
    tuple& operator=(pair<UTypes...>&&);
    
  2. In 22.4.4.1 [tuple.cnstr], starting just before p. 11 please remove as indicated:

    requires MoveAssignable<Types>... tuple& operator=(tuple&& u);
    

    -11- Effects: Move-assigns each element of u to the corresponding element of *this.

    -12- Returns: *this.


918(i). Swap for tuple needs to be conceptualized

Section: 22.4.4.3 [tuple.swap] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2008-10-04 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

Issue 522 was accepted after tuple had been conceptualized, therefore this step needs to be completed.

[ Post Summit Daniel adds ]

This is now NAD Editorial (addressed by N2844) except for item 3 in the proposed wording.

[ 2009-05-01 Daniel adds: ]

As of the recent WP (N2857), this issue is now completely covered by editorial changes (including the third bullet), therefore I unconditionally recommend NAD.

[ Batavia (2009-05): ]

We observed that all the proposed changes have already been applied to the Working Draft, rendering this issue moot.

Move to NAD.

Proposed resolution:

  1. In both 22.4.1 [tuple.general]/2 and 22.4.12 [tuple.special] change

    template <class Swappable... Types>
    void swap(tuple<Types...>& x, tuple<Types...>& y);
    
  2. In 22.4.4 [tuple.tuple], class tuple definition and in 22.4.4.3 [tuple.swap], change

    requires Swappable<Types>...void swap(tuple&);
    
  3. In 22.4.4.3 [tuple.swap] remove the current requires-clause, which says:

    Requires: Each type in Types shall be Swappable


919(i). (forward_)list specialized remove algorithms are over constrained

Section: 24.3.9.6 [forward.list.ops], 24.3.10.5 [list.ops] Status: NAD Submitter: Daniel Krügler Opened: 2008-10-06 Last modified: 2023-02-07

Priority: Not Prioritized

View all other issues in [forward.list.ops].

View all issues with NAD status.

Discussion:

The signatures of forwardlist::remove and list::remove defined in [forwardlist.ops] before 11 + 24.3.10.5 [list.ops] before 15:

requires EqualityComparable<T> void remove(const T& value);

are asymmetric to their predicate variants (which only require Predicate, not EquivalenceRelation) and with the free algorithm remove (which only require HasEqualTo). Also, nothing in the pre-concept WP N2723 implies that EqualityComparable should be the intended requirement.

[ Batavia (2009-05): ]

We agree with the proposed resolution, but would like additional input from concepts experts.

Move to Review.

[ 2009-07-21 Alisdair adds: ]

Current rationale and wording for this issue is built around concepts. I suggest the issue reverts to Open status. I believe there is enough of an issue to review after concepts are removed from the WP to re-examine the issue in Santa Cruz, rather than resolve as NAD Concepts.

[ 2009-10-10 Daniel adds: ]

Recommend NAD: The concept-free wording as of N2960 has no longer the over-specified requirement EqualityComparable for the remove function that uses ==. In fact, now the same test conditions exists as for the free algorithm remove (27.7.8 [alg.remove]). The error was introduced in the process of conceptifying.

[ 2009-10 Santa Cruz: ]

NAD, solved by the removal of concepts.

Proposed resolution:

  1. Replace in [forwardlist.ops] before 11 and in 24.3.10.5 [list.ops] before 15

    requires EqualityComparable<T> HasEqualTo<T, T> void remove(const T& value);
    

926(i). Sequentially consistent fences, relaxed operations and modification order

Section: 33.5.4 [atomics.order] Status: NAD Editorial Submitter: Anthony Williams Opened: 2008-10-19 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [atomics.order].

View all other issues in [atomics.order].

View all issues with NAD Editorial status.

Discussion:

Addresses UK 313

There was an interesting issue raised over on comp.programming.threads today regarding the following example

// Thread 1:
x.store(1, memory_order_relaxed);           // SX
atomic_thread_fence(memory_order_seq_cst);  // F1
y.store(1, memory_order_relaxed);           // SY1
atomic_thread_fence(memory_order_seq_cst);  // F2
r1 = y.load(memory_order_relaxed);          // RY

// Thread 2:
y.store(0, memory_order_relaxed);          // SY2
atomic_thread_fence(memory_order_seq_cst); // F3
r2 = x.load(memory_order_relaxed);         // RX

is the outcome r1 == 0 and r2 == 0 possible?

I think the intent is that this is not possible, but I am not sure the wording guarantees that. Here is my analysis:

Since all the fences are SC, there must be a total order between them. F1 must be before F2 in that order since they are in the same thread. Therefore F3 is either before F1, between F1 and F2 or after F2.

If F3 is after F2, then we can apply 33.5.4 [atomics.order]p5 from N2798:

For atomic operations A and B on an atomic object M, where A modifies M and B takes its value, if there are memory_order_seq_cst fences X and Y such that A is sequenced before X, Y is sequenced before B, and X precedes Y in S, then B observes either the effects of A or a later modification of M in its modification order.

In this case, A is SX, B is RX, the fence X is F2 and the fence Y is F3, so RX must see 1.

If F3 is before F2, this doesn't apply, but F3 can therefore be before or after F1.

If F3 is after F1, the same logic applies, but this time the fence X is F1. Therefore again, RX must see 1.

Finally we have the case that F3 is before F1 in the SC ordering. There are now no guarantees about RX, and RX can see r2==0.

We can apply 33.5.4 [atomics.order]p5 again. This time, A is SY2, B is RY, X is F3 and Y is F1. Thus RY must observe the effects of SY2 or a later modification of y in its modification order.

Since SY1 is sequenced before RY, RY must observe the effects of SY1 or a later modification of y in its modification order.

In order to ensure that RY sees (r1==1), we must see that SY1 is later in the modification order of y than SY2.

We're now skating on thin ice. Conceptually, SY2 happens-before F3, F3 is SC-ordered before F1, F1 happens-before SY1, so SY1 is later in the modification order M of y, and RY must see the result of SY1 (r1==1). However, I don't think the words are clear on that.

[ Post Summit Hans adds: ]

In my (Hans') view, our definition of fences will always be weaker than what particular hardware will guarantee. Memory_order_seq_cst fences inherently don't guarantee sequential consistency anyway, for good reasons (e.g. because they can't enforce a total order on stores). Hence I don't think the issue demonstrates a gross failure to achieve what we intended to achieve. The example in question is a bit esoteric. Hence, in my view, living with the status quo certainly wouldn't be a disaster either.

In any case, we should probably add text along the lines of the following between p5 and p6 in 33.5.4 [atomics.order]:

[Note: Memory_order_seq_cst only ensures sequential consistency for a data-race-free program that uses exclusively memory_order_seq_cst operations. Any use of weaker ordering will invalidate this guarantee unless extreme care is used. In particular, memory_order_seq_cst fences only ensure a total order for the fences themselves. They cannot, in general, be used to restore sequential consistency for atomic operations with weaker ordering specifications.]

Also see thread beginning at c++std-lib-23271.

[ Herve's correction: ]

Minor point, and sorry for the knee jerk reaction: I admit to having no knowledge of Memory_order_seq_cst, but my former boss (John Lakos) has ingrained an automatic introspection on the use of "only". I think you meant:

[Note: Memory_order_seq_cst ensures sequential consistency only for . . . . In particular, memory_order_seq_cst fences ensure a total order only for . . .

Unless, of course, Memory_order_seq_cst really do nothing but ensure sequential consistency for a data-race-free program that uses exclusively memory_order_seq_cst operations.

[ 2009-10 Santa Cruz: ]

NAD Editorial. Solved by N2992.

Proposed resolution:

Add a new paragraph after 33.5.4 [atomics.order]p5 that says

For atomic operations A and B on an atomic object M, where A and B modify M, if there are memory_order_seq_cst fences X and Y such that A is sequenced before X, Y is sequenced before B, and X precedes Y in S, then B occurs later than A in the modifiction order of M.


927(i). Dereferenceable should be HasDereference

Section: 99 [allocator.concepts] Status: NAD Concepts Submitter: Pablo Halpern Opened: 2008-10-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

99 [allocator.concepts] contains a reference to a concept named Dereferenceable. No such concept exists.

[ Daniel adds 2009-02-14: ]

The proposal given in the paper N2829 would automatically resolve this issue.

[ Batavia (2009-05): ]

This particular set of changes has already been made. There are two related changes later on (and possibly also an earlier Example); these can be handled editorially.

Move to NAD Editorial.

Proposed resolution:

Change all uses of the concept Dereferenceable to HasDereference in 99 [allocator.concepts].


928(i). Wrong concepts used for tuple's comparison operators

Section: 22.4.9 [tuple.rel] Status: NAD Concepts Submitter: Joe Gottman Opened: 2008-10-28 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [tuple.rel].

View all other issues in [tuple.rel].

View all issues with NAD Concepts status.

Discussion:

In the latest working draft for C++0x, tuple's operator== and operator< are declared as

template<class... TTypes, class... UTypes> 
  requires EqualityComparable<TTypes, UTypes>... 
  bool operator==(const tuple<TTypes...>& t, const tuple<UTypes...>& u);

and

template<class... TTypes, class... UTypes> 
  requires LessThanComparable<TTypes, UTypes>... 
  bool operator<(const tuple<TTypes...>& t, const tuple<UTypes...>& u);

But the concepts EqualityComparable and LessThanComparable only take one parameter, not two. Also, even if LessThanComparable could take two parameters, the definition of tuple::operator<() should also require

LessThanComparable<UTypes, TTypes>... // (note the order) 

since the algorithm for tuple::operator< is the following (pseudo-code)

for (size_t N = 0; N < sizeof...(TTypes); ++N) { 
    if (get<N>(t) < get<N>(u) return true; 
    else if ((get<N>(u) < get<N>(t)) return false; 
} 

return false; 

Similar problems hold for tuples's other comparison operators.

[ Post Summit: ]

Recommend Tentatively Ready.

Proposed resolution:

In 22.4.1 [tuple.general] and 22.4.9 [tuple.rel] change:

template<class... TTypes, class... UTypes>
  requires EqualityComparableHasEqualTo<TTypes, UTypes>...
  bool operator==(const tuple<TTypes...>&, const tuple<UTypes...>&);

template<class... TTypes, class... UTypes>
  requires LessThanComparableHasLess<TTypes, UTypes>... && HasLess<UTypes, TTypes>...
  bool operator<(const tuple<TTypes...>&, const tuple<UTypes...>&);

template<class... TTypes, class... UTypes>
  requires EqualityComparableHasEqualTo<TTypes, UTypes>...
  bool operator!=(const tuple<TTypes...>&, const tuple<UTypes...>&);

template<class... TTypes, class... UTypes>
  requires LessThanComparableHasLess<UTTypes, TUTypes>... && HasLess<UTypes, TTypes>...
  bool operator>(const tuple<TTypes...>&, const tuple<UTypes...>&);

template<class... TTypes, class... UTypes>
  requires LessThanComparableHasLess<UTTypes, TUTypes>... && HasLess<UTypes, TTypes>...
  bool operator<=(const tuple<TTypes...>&, const tuple<UTypes...>&);

template<class... TTypes, class... UTypes>
  requires LessThanComparableHasLess<TTypes, UTypes>... && HasLess<UTypes, TTypes>...
  bool operator>=(const tuple<TTypes...>&, const tuple<UTypes...>&);

930(i). Access to std::array data as built-in array type

Section: 24.3.7 [array] Status: NAD Submitter: Niels Dekker Opened: 2008-11-17 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [array].

View all issues with NAD status.

Discussion:

The Working Draft (N2798) allows access to the elements of std::array by its data() member function:

23.2.1.4 array::data [array.data]
 T *data();
 const T *data() const;
  1. Returns: elems.

Unfortunately, the result of std::array::data() cannot be bound to a reference to a built-in array of the type of array::elems. And std::array provides no other way to get a reference to array::elems. This hampers the use of std::array, for example when trying to pass its data to a C style API function:

 // Some C style API function. 
 void set_path( char (*)[MAX_PATH] );

 std::array<char,MAX_PATH> path;
 set_path( path.data() );  // error
 set_path( &(path.data()) );  // error

Another example, trying to pass the array data to an instance of another C++ class:

 // Represents a 3-D point in space.
 class three_d_point {
 public:
   explicit three_d_point(const double (&)[3]); 
 };

 const std::array<double,3> coordinates = { 0, 1, 2 };
 three_d_point point1( coordinates.data() );  // error.
 three_d_point point2( *(coordinates.data()) );  // error.

A user might be tempted to use std::array::elems instead, but doing so isn't recommended, because std::array::elems is "for exposition only". Note that Boost.Array users might already use boost::array::elems, as its documentation doesn't explicitly state that boost::array::elems is for exposition only: http://www.boost.org/doc/libs/1_36_0/doc/html/boost/array.html

I can think of three options to solve this issue:

  1. Remove the words "exposition only" from the definition of std::array::elems, as well as the note saying that "elems is shown for exposition only."
  2. Change the signature of std::array::data(), so that it would return a reference to the built-in array, instead of a pointer to its first element.
  3. Add extra member functions, returning a reference to the built-in array.

Lawrence Crowl wrote me that it might be better to leave std::array::elems "for exposition only", to allow alternate representations to allocate the array data dynamically. This might be of interest to the embedded community, having to deal with very limited stack sizes.

The second option, changing the return type of std::array::data(), would break backward compatible to current Boost and TR1 implementations, as well as to the other contiguous container (vector and string) in a very subtle way. For example, the following call to std::swap currently swap two locally declared pointers (data1, data2), for any container type T that has a data() member function. When std::array::data() is changed to return a reference, the std::swap call may swap the container elements instead.

 template <typename T>
 void func(T& container1, T& container2)
 {
   // Are data1 and data2 pointers or references?
   auto data1 = container1.data();
   auto data2 = container2.data();

   // Will this swap two local pointers, or all container elements?
   std::swap(data1, data2);
 }

The following concept is currently satisfied by all contiguous containers, but it no longer is for std::array, when array::data() is changed to return a reference (tested on ConceptGCC Alpha 7):

 auto concept ContiguousContainerConcept<typename T>
 {
   typename value_type = typename T::value_type;
   const value_type * T::data() const;
 }

Still it's worth considering having std::array::data() return a reference, because it might be the most intuitive option, from a user's point of view. Nicolai Josuttis (who wrote boost::array) mailed me that he very much prefers this option.

Note that for this option, the definition of data() would also need to be revised for zero-sized arrays, as its return type cannot be a reference to a zero-sized built-in array. Regarding zero-sized array, data() could throw an exception. Or there could be a partial specialization of std::array where data() returns T* or gets removed.

Personally I prefer the third option, adding a new member function to std::array, overloaded for const and non-const access, returning a reference to the built-in array, to avoid those compatible issues. I'd propose naming the function std::array::c_array(), which sounds intuitive to me. Note that boost::array already has a c_array() member, returning a pointer, but Nicolai told me that this one is only there for historical reasons. (Otherwise a name like std::array::native_array() or std::array::builtin_array() would also be fine with me.) According to my proposed resolution, a zero-sized std::array does not need to have c_array(), while it is still required to have data() functions.

[ Post Summit: ]

Alisdair: Don't like p4 suggesting implementation-defined behaviour.

Walter: What about an explicit conversion operator, instead of adding the new member function?

Alisdair: Noodling about:

template<size_t N, ValueType T>
struct array
{
  T elems[N];

// fantasy code starts here

// crazy decltype version for grins only
//requires True<(N>0)>
//explict operator decltype(elems) & () { return elems; }

// conversion to lvalue ref
requires True<(N>0)>
explict operator T(&)[N] () & { return elems; }

// conversion to const lvalue ref
requires True<(N>0)>
explict operator const T(&)[N] () const & { return elems; }

// conversion to rvalue ref using ref qualifiers
requires True<(N>0)>
explict operator T(&&)[N] () && { return elems; }

// fantasy code ends here

explicit operator bool() { return true; }
};

This seems legal but odd. Jason Merrill says currently a CWG issue 613 on the non-static data member that fixes the error that current G++ gives for the non-explicit, non-conceptualized version of this. Verdict from human compiler: seems legal.

Some grumbling about zero-sized arrays being allowed and supported.

Walter: Would this address the issue? Are we inclined to go this route?

Alan: What would usage look like?

// 3-d point in space
struct three_d_point
{
  explicit three_d_point(const double (&)[3]);
};

void sink(double*);

const std::array<double, 3> coordinates = { 0, 1, 2 };
three_d_point point1( coordinates.data() ); //error
three_d_point point2( *(coordinates.data()) ); // error
three_d_point point3( coordinates ); // yay!

sink(cooridinates); // error, no conversion

Recommended Open with new wording. Take the required clause and add the explicit conversion operators, not have a typedef. At issue still is use decltype or use T[N]. In favour of using T[N], even though use of decltype is specially clever.

[ Post Summit, further discussion in the thread starting with c++std-lib-23215. ]

[ 2009-07 post-Frankfurt (Saturday afternoon group): ]

The idea to resolve the issue by adding explicit conversion operators was abandoned, because it would be inconvenient to use, especially when passing the array to a template function, as mentioned by Daniel. So we reconsidered the original proposed resolution, which appeared acceptable, except for its proposed changes to 24.3.7.5 [array.zero], which allowed c_array_type and c_array() to be absent for a zero-sized array. Alisdair argued that such wording would disallow certain generic use cases. New wording for 24.3.7.5 [array.zero] was agreed upon (Howard: and is reflected in the proposed resolution).

Move to Review

[ 2009-07-31 Alisdair adds: ]

I will be unhappy voting the proposed resolution for 930 past review until we have implementation experience with reference qualifiers. Specifically, I want to understand the impact of the missing overload for const && (if any.)

If we think the issue is important enough it might be worthwhile stripping the ref qualifiers for easy progress next meeting, and opening yet another issue to put them back with experience.

Recommend deferring any decision on splitting the issue until we get LWG feedback next meeting - I may be the lone dissenting voice if others are prepared to proceed without it.

[ 2009-10 Santa Cruz: ]

Mark as NAD. There was not enough consensus that this was sufficiently useful. There are known other ways to do this, such as small inline conversion functions.

Proposed resolution:

Add to the template definition of array, 24.3.7 [array]/3:


typedef T c_array_type[N];
c_array_type & c_array() &;
c_array_type && c_array() &&;
const c_array_type & c_array() const &;

Add the following subsection to 24.3.7 [array], after [array.data]:

23.2.1.5 array::c_array [array.c_array]

c_array_type & c_array() &;
c_array_type && c_array() &&;
const c_array_type & c_array() const &;

Returns: elems.

Change Zero sized arrays 24.3.7.5 [array.zero]:

-2- ...

The type c_array_type is unspecified for a zero-sized array.

-3- The effect of calling c_array(), front(), or back() for a zero-sized array is implementation defined.


933(i). Unique_ptr defect

Section: 20.3.1.3.6 [unique.ptr.single.modifiers] Status: NAD Submitter: Alisdair Meredith Opened: 2008-11-27 Last modified: 2018-06-23

Priority: Not Prioritized

View all other issues in [unique.ptr.single.modifiers].

View all issues with NAD status.

Discussion:

If we are supporting stateful deleters, we need an overload for reset that takes a deleter as well.

void reset(pointer p, deleter_type d);

We probably need two overloads to support move-only deleters, and this sounds uncomfortably like the two constructors I have been ignoring for now...

[ Batavia (2009-05): ]

Howard comments that we have the functionality via move-assigment.

Move to Open.

[ 2009-10 Santa Cruz: ]

Mark as NAD Future.

[2017-02 in Kona, LEWG recommends NAD]

Pointer-accepting overloads of reset() are discouraged — the same functionality is available via move-assignment and that is more idiomatic.

[2017-06-02 Issues Telecon]

Resolve as NAD

Proposed resolution:


935(i). clock error handling needs to be specified

Section: 29.7 [time.clock] Status: NAD Submitter: Beman Dawes Opened: 2008-11-24 Last modified: 2019-02-26

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Each of the three clocks specified in Clocks 29.7 [time.clock] provides the member function:

static time_point now();

The semantics specified by Clock requirements 29.3 [time.clock.req] make no mention of error handling. Thus the function may throw bad_alloc or an implementation-defined exception (16.4.6.13 [res.on.exception.handling] paragraph 4).

Some implementations of these functions on POSIX, Windows, and presumably on other operating systems, may fail in ways only detectable at runtime. Some failures on Windows are due to supporting chipset errata and can even occur after successful calls to a clock's now() function.

These functions are used in cases where exceptions are not appropriate or where the specifics of the exception or cause of error need to be available to the user. See N2828, Library Support for hybrid error handling (Rev 1), for more specific discussion of use cases. Thus some change in the interface of now is required.

The proposed resolution has been implemented in the Boost version of the chrono library. No problems were encountered.

[ Batavia (2009-05): ]

We recommend this issue be deferred until the next Committee Draft has been issued and the prerequisite paper has been accepted.

Move to Open.

[ 2009-10 Santa Cruz: ]

Mark as NAD future. Too late to make this change without having already accepted the hybrid error handling proposal.

[LEWG Kona 2017]

Recommend NAD. Needs a paper. Proposed resolution no longer applies.

Proposed resolution:

Accept the proposed wording of N2828, Library Support for hybrid error handling (Rev 1).

Change Clock requirements 29.3 [time.clock.req] as indicated:

-2- In Table 55 C1 and C2 denote clock types. t1 and t2 are values returned by C1::now() where the call returning t1 happens before (1.10) the call returning t2 and both of these calls happen before C1::time_point::max(). ec denotes an object of type error_code (19.5.4.1 [syserr.errcode.overview]).

Table 55 — Clock requirements
ExpressionReturn typeOperational semantics
... ... ...
C1::now() C1::time_point Returns a time_point object representing the current point in time.
C1::now(ec) C1::time_point Returns a time_point object representing the current point in time.

Change class system_clock 29.7.2 [time.clock.system] as indicated:

static time_point now(error_code& ec=throws());

Change class monotonic_clock 99 [time.clock.monotonic] as indicated:

static time_point now(error_code& ec=throws());

Change class high_resolution_clock 29.7.8 [time.clock.hires] as indicated:

static time_point now(error_code& ec=throws());

937(i). Atomics for standard typedef types

Section: 33.5 [atomics] Status: NAD Editorial Submitter: Clark Nelson Opened: 2008-12-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics].

View all issues with NAD Editorial status.

Discussion:

Addresses US 89

The types in the table "Atomics for standard typedef types" should be typedefs, not classes. These semantics are necessary for compatibility with C.

Change the classes to typedefs.

N2427 specified different requirements for atomic analogs of fundamental integer types (such as atomic_int) and for atomic analogs of <cstdint> typedefs (such as atomic_size_t). Specifically, atomic_int et al. were specified to be distinct classes, whereas atomic_size_t et al. were specified to be typedefs. Unfortunately, in applying N2427 to the WD, that distinction was erased, and the atomic analog of every <cstdint> typedef is required to be a distinct class.

It shouldn't be required that the atomic analog of every <cstdint> typedef be a typedef for some fundamental integer type. After all, <cstdint> is supposed to provide standard names for extended integer types. So there was a problem in N2427, which certainly could have been interpreted to require that. But the status quo in the WD is even worse, because it's unambiguously wrong.

What is needed are words to require the existence of a bunch of type names, without specifying whether they are class names or typedef names.

[ Summit: ]

Change status to NAD, editorial. See US 89 comment notes above.

Direct the editor to turn the types into typedefs as proposed in the comment. Paper approved by committee used typedefs, this appears to have been introduced as an editorial change. Rationale: for compatibility with C.

Proposed resolution:


941(i). Ref-qualifiers for assignment operators

Section: 16 [library] Status: NAD Submitter: Niels Dekker Opened: 2008-12-18 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

The assignment and equality operators = and == are easily confused, just because of their visual similarity, and in this case a simple typo can cause a serious bug. When the left side of an operator= is an rvalue, it's highly unlikely that the assignment was intended by the programmer:

if ( func() = value )  // Typical typo: == intended!

Built-in types don't support assignment to an rvalue, but unfortunately, a lot of types provided by the Standard Library do.

Fortunately the language now offers a syntax to prevent a certain member function from having an rvalue as *this: by adding a ref-qualifier (&) to the member function declaration. Assignment operators are explicitly mentioned as a use case of ref-qualifiers, in "Extending Move Semantics To *this (Revision 1)", N1821 by Daveed Vandevoorde and Bronek Kozicki

Hereby I would like to propose adding ref-qualifiers to all appropriate assignment operators in the library.

[ Batavia (2009-05): ]

Move to Open. We recommend this be deferred until after the next Committee Draft.

[ Frankfurt 2009-07: ]

The LWG declined to move forward with N2819.

Moved to NAD.

Proposed resolution:

A proposed resolution is provided by the paper on this subject, N2819, Ref-qualifiers for assignment operators of the Standard Library


942(i). Atomics synopsis typo

Section: 33.5 [atomics] Status: Dup Submitter: Holger Grund Opened: 2008-12-19 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics].

View all issues with Dup status.

Duplicate of: 880

Discussion:

I'm looking at 33.5 [atomics] and can't really make sense of a couple of things.

Firstly, there appears to be a typo in the <cstdatomic> synopsis:

The atomic_exchange overload taking an atomic_address is missing the second parameter:

void* atomic_exchange(volatile atomic_address*);

should be

void* atomic_exchange(volatile atomic_address*, void*);

Note, that this is not covered by 880 "Missing atomic exchange parameter", which only talks about the atomic_bool.

Proposed resolution:

Change the synopsis in 33.5 [atomics]/2:

void* atomic_exchange(volatile atomic_address*, void*);

945(i). system_clock::rep not specified

Section: 29.7.2 [time.clock.system] Status: NAD Editorial Submitter: Pete Becker Opened: 2008-12-19 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [time.clock.system].

View all issues with NAD Editorial status.

Discussion:

In 29.7.2 [time.clock.system], the declaration of system_clock::rep says "see below", but there is nothing below that describes it.

[ Howard adds: ]

This note refers to:

-2- system_clock::duration::min() < system_clock::duration::zero() shall be true.

I.e. this is standardeze for "system_clock::rep is signed". Perhaps an editorial note along the lines of:

-2- system_clock::duration::min() < system_clock::duration::zero() shall be true. [Note: system_clock::rep shall be signed. -- end note].

?

[ Batavia (2009-05): ]

We agree with the direction of the proposed resolution. Move to NAD Editorial.

Proposed resolution:

Add a note to 29.7.2 [time.clock.system], p2:

-2- system_clock::duration::min() < system_clock::duration::zero() shall be true. [Note: system_clock::rep shall be signed. -- end note].


946(i). duration_cast improperly specified

Section: 29.5.8 [time.duration.cast] Status: NAD Submitter: Pete Becker Opened: 2008-12-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [time.duration.cast].

View all issues with NAD status.

Discussion:

29.5.8 [time.duration.cast]/3:

.... All intermediate computations shall be carried out in the widest possible representation... .

So ignoring floating-point types for the moment, all this arithmetic has to be done using the implementation's largest integral type, even if both arguments use int for their representation. This seems excessive. And it's not at all clear what this means if we don't ignore floating-point types.

This issue is related to 952.

[ Howard adds: ]

The intent of this remark is that intermediate computations are carried out using:

common_type<typename ToDuration::rep, Rep, intmax_t>::type

The Remark was intended to be clarifying prose supporting the rather algorithmic description of the previous paragraph. I'm open to suggestions. Perhaps the entire paragraph 3 (Remarks) would be better dropped?

[ Batavia (2009-05): ]

We view this as a specific case of issue 952, and should be resolved when that issue is resolved.

Move to NAD.

Proposed resolution:


952(i). Various threading bugs #2

Section: 29.5.8 [time.duration.cast] Status: NAD Editorial Submitter: Pete Becker Opened: 2009-01-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [time.duration.cast].

View all issues with NAD Editorial status.

Discussion:

29.5.8 [time.duration.cast] specifies an implementation and imposes requirements in text (and the implementation doesn't satisfy all of the text requirements). Pick one.

This issue is related to 946.

[ 2009-05-10 Howard adds: ]

The Remarks paragraph is an English re-statement of the preceeding Returns clause. It was meant to be clarifying and motivating, not confusing. I'm not aware with how the Remarks contradicts the Returns clause but I'm ok with simply removing the Remarks.

[ Batavia (2009-05): ]

Pete suggests that this could be resolved by rephrasing the Remarks to Notes.

Move to NAD Editorial.

Proposed resolution:


955(i). Various threading bugs #5

Section: 29.3 [time.clock.req] Status: NAD Submitter: Pete Becker Opened: 2009-01-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [time.clock.req].

View all issues with NAD status.

Discussion:

29.3 [time.clock.req] requires that a clock type have a member typedef named time_point that names an instantiation of the template time_point, and a member named duration that names an instantiation of the template duration. This mixing of levels is confusing. The typedef names should be different from the template names.

[ Post Summit, Anthony provided proposed wording. ]

[ 2009-05-04 Howard adds: ]

The reason that the typedef names were given the same name as the class templates was so that clients would not have to stop and think about whether they were using the clock's native time_point / duration or the class template directly. In this case, one person's confusion is another person's encapsulation. The detail that sometimes one is referring to the clock's native types, and sometimes one is referring to an independent type is purposefully "hidden" because it is supposed to be an unimportant detail. It can be confusing to have to remember when to type duration and when to type duration_type, and there is no need to require the client to remember something like that.

For example, here is code that I once wrote in testing out the usability of this facility:

template <class Clock, class Duration>
void do_until(const std::chrono::time_point<Clock, Duration>& t)
{
    typename Clock::time_point now = Clock::now();
    if (t > now)
    {
        typedef typename std::common_type
        <
            Duration,
            typename std::chrono::system_clock::duration
        >::type CD;
        typedef std::chrono::duration<double, std::nano> ID;

        CD d = t - now;
        ID us = duration_cast<ID>(d);
        if (us < d)
            ++us;
        ...
    }
}

I see no rationale to require the client to append _type to some of those declarations. It seems overly burdensome on the author of do_until:

template <class Clock, class Duration>
void do_until(const std::chrono::time_point<Clock, Duration>& t)
{
    typename Clock::time_point_type now = Clock::now();
    if (t > now)
    {
        typedef typename std::common_type
        <
            Duration,
            typename std::chrono::system_clock::duration_type
        >::type CD;
        typedef std::chrono::duration<double, std::nano> ID;

        CD d = t - now;
        ID us = duration_cast<ID>(d);
        if (us < d)
            ++us;
        ...
    }
}

Additionally I'm fairly certain that this suggestion hasn't been implemented. If it had, it would have been discovered that it is incomplete. time_point also has a nested type (purposefully) named duration.

That is, the current proposed wording would put the WP into an inconsistent state.

In contrast, the current WP has been implemented and I've received very favorable feedback from people using this interface in real-world code.

[ Batavia (2009-05): ]

Bill agrees that distinct names should be used for distinct kinds of entities.

Walter would prefer not to suffix type names, especially for such well-understood terms as "duration".

Howard reminds us that the proposed resolution is incomplete, per his comment in the issue.

Move to Open.

[ 2009-06-07 Howard adds: ]

Not meaning to be argumentative, but we have a decade of positive experience with the precedent of using the same name for the nested type as an external class representing an identical concept.

template<class Category, class T, class Distance = ptrdiff_t,
         class Pointer = T*, class Reference = T&>
struct iterator
{
    ...
};

template <BidirectionalIterator Iter>
class reverse_iterator
{
    ...
};

template <ValueType T, Allocator Alloc = allocator<T> >
    requires NothrowDestructible<T>
class list
{
public:
    typedef implementation-defined     iterator;
    ...
    typedef reverse_iterator<iterator> reverse_iterator;
    ...
};

I am aware of zero complaints regarding the use of iterator and reverse_iterator as nested types of the containers despite these names also having related meaning at namespace std scope.

Would we really be doing programmers a favor by renaming these nested types?

template <ValueType T, Allocator Alloc = allocator<T> >
    requires NothrowDestructible<T>
class list
{
public:
    typedef implementation-defined     iterator_type;
    ...
    typedef reverse_iterator<iterator> reverse_iterator_type;
    ...
};

I submit that such design contributes to needless verbosity which ends up reducing readability.

[ 2009-10 Santa Cruz: ]

Mark as NAD. No concensus for changing the WP.

Proposed resolution:

Change 29 [time]:

...
template <class Clock, class Duration = typename Clock::duration_type> class time_point;
...

Change 29.3 [time.clock.req]:

Table 45 -- Clock requirements
Expression Return type Operational semantics
... ... ...
C1::duration_type chrono::duration<C1::rep, C1::period> The native duration type of the clock.
C1::time_point_type chrono::time_point<C1> or chrono::time_point<C2, C1::duration_type< The native time_point type of the clock. Different clocks may share a time_point_type definition if it is valid to compare their time_point_types by comparing their respective duration_types. C1 and C2 shall refer to the same epoch.
... ... ...
C1::now() C1::time_point_type Returns a time_point_type object representing the current point in time.

Change 29.7.2 [time.clock.system]:

-1- Objects of class system_clock represent wall clock time from the system-wide realtime clock.

class system_clock { 
public: 
  typedef see below rep; 
  typedef ratio<unspecified, unspecified> period; 
  typedef chrono::duration<rep, period> duration_type; 
  typedef chrono::time_point<system_clock> time_point_type; 
  static const bool is_monotonic = unspecified ; 

  static time_point_type now(); 

  // Map to C API 
  static time_t to_time_t (const time_point_type& t); 
  static time_point_type from_time_t(time_t t); 
};

-2- system_clock::duration_type::min() < system_clock::duration_type::zero() shall be true.

time_t to_time_t(const time_point_type& t);

-3- Returns: A time_t object that represents the same point in time as t when both values are truncated to the coarser of the precisions of time_t and time_point_type.

time_point_type from_time_t(time_t t);

-4- Returns: A time_point_type object that represents the same point in time as t when both values are truncated to the coarser of the precisions of time_t and time_point_type.

Change 99 [time.clock.monotonic]:

class monotonic_clock { 
public: 
  typedef unspecified                                rep; 
  typedef ratio<unspecified , unspecified>           period; 
  typedef chrono::duration<rep, period>              duration_type; 
  typedef chrono::time_point<unspecified , duration_type> time_point_type; 
  static const bool is_monotonic =                   true; 

  static time_point_type now();
};

Change 29.7.8 [time.clock.hires]:

class high_resolution_clock { 
public: 
  typedef unspecified                                rep; 
  typedef ratio<unspecified , unspecified>           period; 
  typedef chrono::duration<rep, period>              duration_type; 
  typedef chrono::time_point<unspecified , duration_type> time_point_type; 
  static const bool is_monotonic =                   true; 

  static time_point_type now();
};

959(i). Various threading bugs #9

Section: 33.7.4 [thread.condition.condvar] Status: NAD Submitter: Pete Becker Opened: 2009-01-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.condition.condvar].

View all issues with NAD status.

Discussion:

33.7.4 [thread.condition.condvar]: condition_variable::wait_for is required to compute the absolute time by adding the duration value to chrono::monotonic_clock::now(), but monotonic_clock is not required to exist.

[ Summit: ]

Move to open. Associate with LWG 859 and any other monotonic-clock related issues.

[ 2009-08-01 Howard adds: ]

I believe that 859 (currently Ready) addresses this issue, and that this issue should be marked NAD, solved by 859 (assuming it moves to WP).

[ 2009-10 Santa Cruz: ]

Leave open, but expect to be fixed by N2969 revision that Detlef is writing.

[ 2009-11-18 Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Proposed resolution:

Rationale:

condition_variable::wait_for no longer refers to monotonic_clock, so this issue is moot.


969(i). What happened to Library Issue 475?

Section: 27.6.5 [alg.foreach] Status: NAD Editorial Submitter: Stephan T. Lavavej Opened: 2009-01-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.foreach].

View all issues with NAD Editorial status.

Discussion:

Library Issue 475 has CD1 status, but the non-normative note in N2723 was removed in N2798 (27.6.5 [alg.foreach] in both drafts).

[ Batavia (2009-05): ]

Move to NAD Editorial.

Proposed resolution:

Restore the non-normative note. It might need to be expressed in terms of concepts.


971(i). Spurious diagnostic conversion function

Section: 19.5.4.5 [syserr.errcode.nonmembers] Status: NAD Submitter: Beman Dawes Opened: 2009-01-19 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Anthony Williams raised the question in c++std-lib-22987 "why is there std::make_error_code(std::errc)? What purpose does this serve?"

The function make_error_code(errc e) is not required, since make_error_condition(errc e) is the function that is needed for errc conversions. make_error_code(errc e) appears to be a holdover from my initial confusion over the distinction between POSIX and operating systems that conform to the POSIX spec.

[ Post Summit: ]

Recommend Review.

[ Batavia (2009-05): ]

The designer of the facility (Christopher Kohlhoff) strongly disagrees that there is an issue here, and especially disagrees with the proposed resolution. Bill would prefer to be conservative and not apply this proposed resolution. Move to Open, and recommend strong consideration for NAD status.

[ 2009-05-21 Beman adds: ]

My mistake. Christopher and Bill are correct and the issue should be NAD. The function is needed by users.

[ 2009-07-21 Christopher Kohlhoff adds rationale for make_error_code: ]

Users (and indeed library implementers) may need to use the errc codes in portable code. For example:

void do_foo(error_code& ec)
{
#if defined(_WIN32)
  // Windows implementation ...
#elif defined(linux)
  // Linux implementation ...
#else
  // do_foo not supported on this platform
  ec = make_error_code(errc::not_supported);
#endif
}

[ 2009 Santa Cruz: ]

Moved to NAD.

Proposed resolution:

Change System error support 19.5 [syserr], Header <system_error> synopsis, as indicated:

error_code make_error_code(errc e);
error_condition make_error_condition(errc e);

Delete from Class error_code non-member functions 19.5.4.5 [syserr.errcode.nonmembers]:

error_code make_error_code(errc e);

Returns: error_code(static_cast<int>(e), generic_category).


972(i). The term "Assignable" undefined but still in use

Section: 16 [library] Status: NAD Editorial Submitter: Niels Dekker Opened: 2009-01-07 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD Editorial status.

Discussion:

Previous versions of the Draft had a table, defining the Assignable requirement. For example N2134 Table 79, "Assignable requirements". But I guess the term "Assignable" is outdated by now, because the current Committee Draft provides MoveAssignable, CopyAssignable, and TriviallyCopyAssignable concepts instead. And as far as I can see, it no longer has a definition of Assignable. (Please correct me if I'm wrong.) Still the word "Assignable" is used in eight places in the Draft, N2800.

Are all of those instances of "Assignable" to be replaced by "CopyAssignable"?

[ Batavia (2009-05): ]

Move to NAD Editorial.

Proposed resolution:

Change Exception Propagation 17.9.7 [propagation]:

exception_ptr shall be DefaultConstructible, CopyConstructible, CopyAssignable and EqualityComparable.

Change Class template reference_wrapper 22.10.6 [refwrap]:

reference_wrapper<T> is a CopyConstructible and CopyAssignable wrapper around a reference to an object of type T.

Change Placeholders 22.10.15.5 [func.bind.place]:

It is implementation defined whether placeholder types are CopyAssignable. CopyAssignable placeholders' copy assignment operators shall not throw exceptions.

Change Class template shared_ptr 20.3.2.2 [util.smartptr.shared]:

Specializations of shared_ptr shall be CopyConstructible, CopyAssignable, and LessThanComparable...

Change Class template weak_ptr 20.3.2.3 [util.smartptr.weak]:

Specializations of weak_ptr shall be CopyConstructible, CopyAssignable, and LessThanComparable...

Change traits typedefs 23.2.3 [char.traits.typedefs] (note: including deletion of reference to 23.1!):

Requires: state_type shall meet the requirements of CopyAssignable (23.1), CopyConstructible (20.1.8), and DefaultConstructible types.

Change Class seed_seq 28.5.8.1 [rand.util.seedseq] (note again: including deletion of reference to 23.1!):

In addition to the requirements set forth below, instances of seed_seq shall meet the requirements of CopyConstructible (20.1.8) and of CopyAssignable (23.1).

Note: The proposed resolution of this issue does not deal with the instance of the term "Assignable" in 99 [auto.ptr], as this is dealt with more specifically by LWG 973, "auto_ptr characteristics", submitted by Maarten Hilferink.


973(i). auto_ptr characteristics

Section: 99 [auto.ptr] Status: NAD Editorial Submitter: Maarten Hilferink Opened: 2009-01-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [auto.ptr].

View all issues with NAD Editorial status.

Discussion:

I think that the Note of 99 [auto.ptr], paragraph 3 needs a rewrite since "Assignable" is no longer defined as a concept. The relationship of auto_ptr with the new CopyAssignable, MoveAssignable, and MoveConstructible concepts should be clarified. Furthermore, since the use of auto_ptr is depreciated anyway, we can also omit a description of its intended use.

[ Batavia (2009-05): ]

We agree with the intent of the proposed resolution. Move to NAD Editorial.

Proposed resolution:

Change 99 [auto.ptr], paragraph 3:

The auto_ptr provides a semantics of strict ownership. An auto_ptr owns the ob ject it holds a pointer to. Copying an auto_ptr copies the pointer and transfers ownership to the destination. If more than one auto_ptr owns the same ob ject at the same time the behavior of the program is undefined. [Note: The uses of auto_ptr include providing temporary exception-safety for dynamically allocated memory, passing ownership of dynamically allocated memory to a function, and returning dynamically allocated memory from a function. auto_ptr does not meet the CopyConstructible and Assignable requirements for standard library container elements and thus instantiating a standard library container with an auto_ptr results in undefined behavior.

Instances of auto_ptr shall meet the MoveConstructible and MoveAssignable requirements, but do not meet the CopyConstructible and CopyAssignable requirements. -- end note]


977(i). insert iterators inefficient for expensive to move types

Section: 25.5.2 [insert.iterators] Status: NAD Submitter: Howard Hinnant Opened: 2009-02-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [insert.iterators].

View all issues with NAD status.

Discussion:

The new concepts for the insert iterators mandate an extra copy when inserting an lvalue:

requires CopyConstructible<Cont::value_type>
  back_insert_iterator<Cont>& 
  operator=(const Cont::value_type& value);

-1- Effects: push_back(*container, Cont::value_type(value));

The reason is to convert value into an rvalue because the current BackInsertionContainer concept only handles push_back-ing rvalues:

concept BackInsertionContainer<typename C> : Container<C> { 
  void push_back(C&, value_type&&); 
}

Without the conversion of value to an rvalue, the assignment operator fails to concept check.

A solution is to modify the BackInsertionContainer concept so that the client can pass in the parameter type for push_back similar to what is already done for the OutputIterator concept:

concept BackInsertionContainer<typename C, typename Value = C::value_type&&>
  : Container<C> { 
     void push_back(C&, Value); 
}

This allows the assignment operator to be adjusted appropriately:

requires BackInsertionContainer<Cont, Cont::value_type const&> &&
         CopyConstructible<Cont::value_type>
  back_insert_iterator<Cont>& 
  operator=(const Cont::value_type& value);

-1- Effects: push_back(*container, value);

[ We may want to propagate this fix to other concepts such as StackLikeContainer. ]

[ Solution and wording collaborated on by Doug and Howard. ]

[ Batavia (2009-05): ]

Howard notes that "these operations behaved efficiently until concepts were added."

Alisdair is uncertain that the proposed resolution is syntactically correct.

Move to Open, and recommend the issue be deferred until after the next Committee Draft is issued.

[ 2009-10 Santa Cruz: ]

NAD, solved by the removal of concepts.

Proposed resolution:

Change [container.concepts.free]:

concept FrontInsertionContainer<typename C, typename Value = C::value_type&&>
    : Container<C> { 
  void push_front(C&, value_type&& Value); 

  axiom FrontInsertion(C c, value_type Value x) { 
    x == (push_front(c, x), front(c)); 
  } 
}

...

concept BackInsertionContainer<typename C, typename Value = C::value_type&&>
    : Container<C> { 
  void push_back(C&, value_type&& Value); 
}

...

concept InsertionContainer<typename C, typename Value = C::value_type&&>
    : Container<C> { 
  iterator insert(C&, const_iterator, value_type&& Value); 

  axiom Insertion(C c, const_iterator position, value_type Value v) { 
    v == *insert(c, position, v); 
  } 
}

Change [container.concepts.member]:

auto concept MemberFrontInsertionContainer<typename C, typename Value = C::value_type&&>
    : MemberContainer<C> { 
  void C::push_front(value_type&& Value); 

  axiom MemberFrontInsertion(C c, value_type Value x) { 
    x == (c.push_front(x), c.front()); 
  } 
}

...

auto concept MemberBackInsertionContainer<typename C, typename Value = C::value_type&&>
    : MemberContainer<C> { 
  void C::push_back(value_type&& Value); 
}

...

auto concept MemberInsertionContainer<typename C, typename Value = C::value_type&&>
    : MemberContainer<C> { 
  iterator C::insert(const_iterator, value_type&& Value); 

  axiom MemberInsertion(C c, const_iterator position, value_type Value v) { 
    v == *c.insert(position, v); 
  } 
}

Change [container.concepts.maps]:

template <MemberFrontInsertionContainer C, typename Value = C::value_type&&> 
concept_map FrontInsertionContainer<C, Value> { 
  typedef Container<C>::value_type value_type;

  void push_front(C& c, value_type&& Value v) { c.push_front(static_cast<value_type&& Value>(v)); } 
}

...

template <MemberBackInsertionContainer C, typename Value = C::value_type&&> 
concept_map BackInsertionContainer<C, Value> { 
  typedef Container<C>::value_type value_type;

  void push_back(C& c, value_type&& Value v) { c.push_back(static_cast<value_type&& Value>(v)); } 
}

...

template <MemberInsertionContainer C, typename Value = C::value_type&&> 
concept_map InsertionContainer<C, Value> { 
  typedef Container<C>::value_type value_type;
  Container<C>::iterator insert(C& c, Container<C>::const_iterator i, value_type&& Value v) 
  { return c.insert(i, static_cast<value_type&& Value>(v)); } 
}

Change 25.5.2.2 [back.insert.iterator]:

template <BackInsertionContainer Cont> 
class back_insert_iterator {
  ...
  requires BackInsertionContainer<Cont, const Cont::value_type&>
           CopyConstructible<Cont::value_type>
    back_insert_iterator<Cont>& 
      operator=(const Cont::value_type& value);
  ...

Change [back.insert.iter.op=]:

requires BackInsertionContainer<Cont, const Cont::value_type&>
         CopyConstructible<Cont::value_type>
  back_insert_iterator<Cont>& 
    operator=(const Cont::value_type& value);

-1- Effects: push_back(*container, Cont::value_type(value));

Change 25.5.2.3 [front.insert.iterator]:

template <FrontInsertionContainer Cont> 
class front_insert_iterator {
  ...
  requires FrontInsertionContainer<Cont, const Cont::value_type&>
           CopyConstructible<Cont::value_type>
    front_insert_iterator<Cont>& 
      operator=(const Cont::value_type& value);
  ...

Change [front.insert.iter.op=]:

requires FrontInsertionContainer<Cont, const Cont::value_type&>
         CopyConstructible<Cont::value_type>
  front_insert_iterator<Cont>& 
    operator=(const Cont::value_type& value);

-1- Effects: push_front(*container, Cont::value_type(value));

Change 25.5.2.4 [insert.iterator]:

template <InsertionContainer Cont> 
class insert_iterator {
  ...
  requires InsertionContainer<Cont, const Cont::value_type&>
           CopyConstructible<Cont::value_type>
    insert_iterator<Cont>& 
      operator=(const Cont::value_type& value);
  ...

Change [insert.iter.op=]:

requires InsertionContainer<Cont, const Cont::value_type&>
         CopyConstructible<Cont::value_type>
  insert_iterator<Cont>& 
    operator=(const Cont::value_type& value);

-1- Effects:

iter = insert(*container, iter, Cont::value_type(value)); 
++iter;

979(i). Bad example

Section: 25.5.4 [move.iterators] Status: NAD Editorial Submitter: Howard Hinnant Opened: 2009-02-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [move.iterators].

View all issues with NAD Editorial status.

Discussion:

25.5.4 [move.iterators] has an incorrect example:

-2- [Example:

set<string> s; 
// populate the set s 
vector<string> v1(s.begin(), s.end());          // copies strings into v1 
vector<string> v2(make_move_iterator(s.begin()), 
                  make_move_iterator(s.end())); // moves strings into v2

-- end example]

One can not move from a set because the iterators return const references.

[ Batavia (2009-05): ]

We agree with the proposed resolution. Move to NAD Editorial.

Proposed resolution:

Change 25.5.4 [move.iterators]/2:

-2- [Example:

setlist<string> s; 
// populate the setlist s 
vector<string> v1(s.begin(), s.end());          // copies strings into v1 
vector<string> v2(make_move_iterator(s.begin()), 
                  make_move_iterator(s.end())); // moves strings into v2

-- end example]


980(i). mutex lock() missing error conditions

Section: 33.6.4 [thread.mutex.requirements] Status: NAD Submitter: Ion Gaztañaga Opened: 2009-02-07 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [thread.mutex.requirements].

View all other issues in [thread.mutex.requirements].

View all issues with NAD status.

Discussion:

POSIX 2008 adds two return values for pthread_mutex_xxxlock(): EOWNERDEAD (owner_dead) and ENOTRECOVERABLE (state_not_recoverable). In the first case the mutex is locked, in the second case the mutex is not locked.

Throwing an exception in the first case can be incompatible with the use of Locks, since the Lock::owns_lock() will be false when the lock is being destroyed.

Consider:

//Suppose mutex.lock() throws "owner_dead"
unique_lock ul(&mutex);
//mutex left locked if "owner_dead" is thrown

Throwing an exception with owner_dead might be also undesirable if robust-mutex support is added to C++ and the user has the equivalent of pthread_mutex_consistent() to notify the user has fixed the corrupted data and the mutex state should be marked consistent.

  1. For state_not_recoverable add it to the list of Error conditions:
  2. For owner_dead, no proposed resolution.

[ Summit: ]

Not a defect. Handling these error conditions is an implementation detail and must be handled below the C++ interface.

Proposed resolution:

Add to 33.6.4 [thread.mutex.requirements], p12:

-12- Error conditions:


988(i). Reflexivity meaningless?

Section: 99 [concept.comparison] Status: NAD Submitter: Alisdair Meredith Opened: 2009-02-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [concept.comparison].

View all issues with NAD status.

Discussion:

99 [concept.comparison] p2:

Due to the subtle meaning of == inside axioms, the Reflexivity axiom does not do anything as written. It merely states that a value is substitutable with itself, rather than asserting a property of the == operator.

Original proposed resolution:

Change the definition of Reflexivity in 99 [concept.comparison]:

axiom Reflexivity(T a) { (a == a) == true; }

[ Post Summit: ]

Alisdair: I was wrong.

Recommend NAD.

Proposed resolution:

NAD.


989(i). late_check and library

Section: 16 [library] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-02-24 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD Concepts status.

Discussion:

The example in 6.9p2 shows how late_check blocks inhibit concept_map lookup inside a constrained context, and so inhibit concept map adaption by users to meet template requirements.

Do we need some text in clause 17 prohibitting use of late_check in library template definitions unless otherwise documented?

[ Doug adds: ]

We need something like this, but it should be a more general statement about implementations respecting the concept maps provided by the user. Use of late_check is one way in which implementations can subvert the concept maps provided by the user, but there are other ways as well ("pattern-based" overloading, tricks with "auto" concept maps and defaulted associated type arguments).

[ Batavia (2009-05): ]

Move to Open, pending proposed wording from Alisdair and/or Doug for further review.

Proposed resolution:


992(i). Allow implementations to implement C library in the global namespace

Section: 16.4.2.2 [contents] Status: NAD Submitter: P.J. Plauger Opened: 2009-03-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [contents].

View all issues with NAD status.

Discussion:

Addresses UK-169 [CD1]

This phrasing contradicts later freedom to implement the C standard library portions in the global namespace as well as std. (17.6.2.3p4)

[ Batavia (2009-05): ]

The proposed wording seems to go too far. Move back to Open.

[ 2009-07 Frankfurt: ]

Howard to add NB reference to the description of this issue.

Move to NAD. This comment is informative and not normative by the use of the word "are" instead of the word "shall."

A note linking to Annex D would help clarify the intention, here.

Robert to Open a separate issue proposing that the standard C headers be undeprecated, for the purpose of clarifying the standard.

[ 2009-07-22 Bill modified the proposed wording with a clarifying footnote. ]

Proposed resolution:

Add a footnote to 16.4.2.2 [contents], p2:

-2- All library entities except macros, operator new and operator delete are defined within the namespace std or namespaces nested within namespace std*.

*The C standard library headers [depr.c.headers] also define names within the global namespace, while the C++ headers for C library facilities 16.4.2.3 [headers] may also define names within the global namespace.


995(i). Operational Semantics Unclear

Section: 16.3.2.3 [structure.requirements] Status: NAD Submitter: David Abrahams Opened: 2009-03-06 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [structure.requirements].

View all issues with NAD status.

Discussion:

As a practical matter there's disagreement on the meaning of operational semantics. If the text in 16.3.2.3 [structure.requirements]p4 isn't clear, it should be clarified. However, it's not clear whether the disagreement is merely due to people not being aware of the text.

[ Batavia (2009-05): ]

Agree with the recommended NAD resolution.

Proposed resolution:

Recommend NAD. The text in 16.3.2.3 [structure.requirements] is perfectly clear.


996(i). Move operation not well specified

Section: 16 [library] Status: NAD Submitter: David Abrahams Opened: 2009-03-06 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

There are lots of places in the standard where we talk about "the move constructor" but where we mean "the move operation," i.e. T( move( x ) ).

We also don't account for whether that operation modifies x or not, and we need to.

[ Batavia (2009-05): ]

Move to Open, pending proposed wording from Dave for further review.

[ 2010 Rapperswil: ]

Move to NAD. We define what we expect from a moved-from object in Table 34 [movesconstructible].

Proposed resolution:


1000(i). adjacent_find is over-constrained

Section: 27.6.10 [alg.adjacent.find] Status: NAD Concepts Submitter: Chris Jefferson Opened: 2009-03-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.adjacent.find].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 296

adjacent_find in C++03 allows an arbitrary predicate, but in C++0x EqualityComparable/EquivalenceRelation is required. This forbids a number of use cases, including:

adjacent_find(begin, end, less<double>) Find the first place where a range is not ordered in decreasing order - in use to check for sorted ranges.
adjacent_find(begin, end, DistanceBiggerThan(6) ) ) Find the first place in a range where values differ by more than a given value - in use to check an algorithm which produces points in space does not generate points too far apart.

A number of books use predicate which are not equivalence relations in examples, including "Thinking in C++" and "C++ Primer".

Adding the requirement that the predicate is an EquivalenceRelation does not appear to open up any possibility for a more optimised algorithm.

Proposed resolution:

Change the definition of adjacent_find in the synopsis of 27 [algorithms] and 27.6.10 [alg.adjacent.find] to:

template<ForwardIterator Iter> 
  requires EqualityComparableHasEqualTo<Iter::value_type, Iter::value_type>
  Iter adjacent_find(Iter first, Iter last);

template<ForwardIterator Iter, EquivalenceRelationPredicate<auto, Iter::value_type, Iter::value_type> Pred> 
  requires CopyConstructible<Pred> 
  Iter adjacent_find(Iter first, Iter last, Pred pred);

1001(i). Pointers, concepts and headers

Section: 16 [library] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-10 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 78

Related to 1063.

This is effectively an extension of LWG issue 343.

We know there is an increasing trend (encouraged by conformance testers and some users) that each library header should supply no more than required to satisfy the synopsis in the standard. This is typically achieved by breaking larger headers into smaller subsets, and judicious use of forward declarations.

If we apply this policy to C++0x (per N2800) it will be very surprising for people using library algorithms over ranges defined by pointers that they must #include <iterator_concepts> for their code to compile again. That is because pointers do not satisfy any of the iterator concepts without the concept_map supplied in this header.

Therefore, I suggest we should require all library headers that make use of iterator concepts are specifically required to #include <iterator_concepts>.

At a minimum, the list of headers would be: (assuming all are constrained by concepts)

algorithm
array
deque
forward_list
initializer_list
iterator
locale
list
map
memory          // if 1029 is adopted
memory_concepts
numeric
random
regex
set
string
tuple
unordered_map
unordered_set
utility
vector

[ Ganesh adds: ]

The same problems exists for <memory_concepts> and <container_concepts>.

In order to compile <vector> you just need the definitions of the concepts in <memory_concepts>, the concept maps defined there are not necessary. Yet, from the user point of view, if the concept map template for AllocatableElement are not in scope, <vector> is pretty useless. Same for <tuple> and ConstructibleWithAllocator.

Similarly, <queue> is not very useful if the concept map template for QueueLikeContainer is not in scope, although the definition of concept alone is theoretically sufficient.

There's a pattern here: if a concept has concept maps "attached", they should never be separated.

[ Beman provided the proposed resolution for the May 2009 mailing. He comments: ]

Initially I tried to specify exactly what header should include what other headers. This was verbose, error prone, hard to maintain, and appeared to add little value compared to just stating the general rule.

[ Batavia (2009-05): ]

Pete believes the proposed wording overconstrains implementers. Instead of specifying the mechanism, he prefers a solution that spells out what needs to be declared, rather than how those declarations are to be provided, e.g.,

A C++ header shall provide the names that are required to be defined in that header.

Bill suggests approaching the wording from a programmer's perspective. We may want to consider promising that certain widely-used headers (e.g., the concept headers) are included when needed by other headers. He feels, however, there is nothing broken now, although we may want to consider "something nicer."

Move to Open status.

[ 2009-06-16 Beman updated the proposed resolution: ]

[ 2009-07-15 Beman updated the proposed resolution: ]

[ 2009-07-17 Beman updated the proposed resolution based on feedback from the LWG in Frankfurt: ]

[ 2009-07 Frankfurt: ]

Revised Proposed Resolution:

A C++ header may include other C++ headers. A C++ header shall provide the declarations and definitions that appear in its synopsis (3.2 [basic.def.odr]). A C++ header shown in its synopsis as including other C++ headers shall provide the declarations and definitions that appear in the synopses of those other headers.

Alisdair: Does this address the BSI comment?

Beman: There were several overlapping comments. I tried to handle them all with one resolution.

Alisdair: I'd prefer to see this closed as NAD and have this resolution be the subject of some other, new issue.

Move to NAD Concepts. Howard to open a new issue (1178) in Ready state with the Proposed Resolution above. Beman will write up a discussion for the new issue.

Proposed resolution:

Change 16.4.6.2 [res.on.headers], Headers, paragraph 1, as indicated:

A C++ header may include other C++ headers.[footnote] A C++ header shall provide the declarations and definitions that appear in its synopsis (6.3 [basic.def.odr]). A C++ header shown in its synopsis as including other C++ headers shall provide the same declarations and definitions as if those other headers were included.

[footnote] C++ headers must include a C++ header that contains any needed definition (3.2).


1002(i). Provide bulk include headers

Section: 16.4.2.3 [headers] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [headers].

View all issues with NAD status.

Discussion:

Addresses UK-170 [CD1]

One of goals of C++0x is to make language easier to teach and for 'incidental' programmers. The fine-grained headers of the C++ library are valuable in large scale systems for managing dependencies and optimising build times, but overcomplicated for simple development and tutorials. Add additional headers to support the whole library through a single include statement.

[ Batavia (2009-05): ]

We do not all agree that this is an issue, but we agree that if it needs solving this is the right way to do it. Move to Tentatively Ready.

[ 2009-07-06 Beman notes: ]

This issue adds a header <std>.

There is a paper to be looked at, N2905 Aggregation headers, that adds a header <std-all> that is the same thing except it excludes deprecated headers. N2905 also proposes a second aggregation header.

Seems like this issue should be held in abeyance until the LWG has had a chance to look at N2905.

[ 2009-07-06 Howard: I've pulled this issue back to Review. ]

[ 2009-07 Frankfurt ]

No consensus for change.

Proposed resolution:

Insert a new paragraph in 16.4.2.3 [headers] between p4 and p5

An additional header <std> shall have the effect of supplying the entire standard library. [Note: for example, it might be implemented as a file with an #include statement for each of the headers listed in tables 13 and 14. -- end note]


1003(i). Require more useful headers for freestanding implementations

Section: 16.4.2.5 [compliance] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [compliance].

View all issues with NAD status.

Discussion:

Addresses JP 23 [CD1]

There is a freestanding implementation including <type_traits>, <array>, <ratio>, lately added to Table 13, C++ library headers. Programmers think them useful and hope that these headers are also added to Table 15, C++ headers for freestanding implementations, that shows the set of headers which a freestanding implementation shall include at least.

Original proposed resolution

Add <type_traits>, <array>, <ratio> to Table 15.

[ Summit: ]

The <array> header has far too many dependencies to require for a free-standing implementation.

The <ratio> header would be useful, has no dependencies, but is not strictly necessary.

The <type_traits> header is fundamentally a core language facility with a library interface, so should be supported.

(it is anticipated the resolution will come via an update to paper N2814) (see also LWG 833)

[ Batavia (2009-05): ]

Leave in Review status pending a paper on freestanding implementations by Martin Tasker.

[ 2009-07 Frankfurt: ]

Move this to NAD.

We considered all of the listed headers, and found a compelling case only for the inclusion of <type_traits> in the list of headers required of a freestanding implementation.

See Martin Tasker's paper Fixing Freestanding which provides the wording to include <type_traits> into freestanding implementations.

Proposed resolution:

Add <type_traits> to Table 15.


1005(i). numeric_limits partial specializations not concept enabled

Section: 17.3.5 [numeric.limits] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [numeric.limits].

View all other issues in [numeric.limits].

View all issues with NAD Concepts status.

Discussion:

Addresses JP 26

numeric_limits [partial specializations] does not use concept.

[ Summit: ]

Alisdair will provide a solution as part of treatment of axioms and LWG 902.

[ Post Summit: ]

Alisdair recommends NAD as the partial specializations are already constrained by requirements on the primary template.

[ Batavia (2009-05): ]

The Working Draft does not in general repeat a primary template's constraints in any specializations. Move to NAD.

[ 2009-05-25 Howard adds: ]

A c++std-lib thread starting at c++std-lib-23880 has cast doubt that NAD is the correct resolution of this issue. Indeed the discussion also casts doubt that the current proposed wording is the correct resolution as well. Personally I'm inclined to reset the status to Open. However I'm reverting the status to that which it had prior to the Batavia recommendation. I'm setting back to Review.

Proposed resolution:

Change 17.3.5 [numeric.limits]:

template<class Regular T> class numeric_limits<const T>;
template<class Regular T> class numeric_limits<volatile T>;
template<class Regular T> class numeric_limits<const volatile T>;

1007(i). throw_with_nested not concept enabled

Section: 17.9.8 [except.nested] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [except.nested].

View all issues with NAD Concepts status.

Discussion:

Addresses JP 29

throw_with_nested does not use concept.

[ Summit: ]

Agreed.

Proposed resolution:

Alisdair initially proposed wording in N2619.

We are awaiting an updated paper based on feedback from the San Francisco review.


1008(i). nested_exception wording unclear

Section: 17.9.8 [except.nested] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [except.nested].

View all issues with NAD status.

Discussion:

Addresses JP 31

It is difficult to understand in which case nested_exception is applied.

[ Summit: ]

Alisdair will add an example in an update to N2619.

[ 2009-10 Santa Cruz: ]

It doesn't appear that N2619 really addresses this. Alisdair to propose wording.

[ 2010 Pittsburgh: ]

Mark issue 1008 as NAD, the type is adequately described.

Rationale:

nested_exception is intended to be inherited from by exception classes that are to be thrown during the handling of another exception, i.e. when translating from one exception type to another. nested_exception allows the originally thrown exception to be easily retained in that scenario.

Proposed resolution:


1009(i). InputIterator post-increment dangerous

Section: 25.3.5.2 [iterator.iterators] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iterator.iterators].

View all issues with NAD status.

Discussion:

Addresses UK 251

The post-increment operator is dangerous for a general InputIterator. The multi-pass guarantees that make it meaningful are defined as part of the ForwardIterator refinement. Any change will affect only constrained templates that have not yet been written, so should not break existing user iterators which remain free to add these operations. This change will also affect the generalised OutputIterator, although there is no percieved need for the post-increment operator in this case either.

[ 2009-07-28 Alisdair adds: ]

We still think the issue is relevant, but needs totally rewording in non-concept language. We would like to see the issue retained as Open, rather than deferred as NAD Concepts. Review status is no longer appropriate.

[ 2009-10 Santa Cruz: ]

NAD. Without concepts we do not feel that input iterator post increment is broken.

Proposed resolution:

Change 25.3.5.2 [iterator.iterators]:

concept Iterator<typename X> : Semiregular<X> { 
  MoveConstructible reference = typename X::reference; 
  MoveConstructible postincrement_result;

  requires HasDereference<postincrement_result>;

  reference operator*(X&&); 
  X& operator++(X&); 
  postincrement_result operator++(X&, int);
}

...

postincrement_result operator++(X& r, int);

-3- Effects: equivalent to { X tmp = r; ++r; return tmp; }.

Change 25.3.5.3 [input.iterators]:

concept InputIterator<typename X> : Iterator<X>, EqualityComparable<X> { 
  ObjectType value_type = typename X::value_type; 
  MoveConstructible pointer = typename X::pointer; 

  SignedIntegralLike difference_type = typename X::difference_type; 

  requires IntegralType<difference_type> 
        && Convertible<reference, const value_type &>; 
        && Convertible<pointer, const value_type*>; 

  requires Convertible<HasDereference<postincrement_result>::result_type, const value_type&>;

  pointer operator->(const X&); 
}

Change 25.3.5.4 [output.iterators]:

auto concept OutputIterator<typename X, typename Value> { 
  requires Iterator<X>; 

  typename reference = Iterator<X>::reference; 
  typename postincrement_result = Iterator<X>::postincrement_result;
  requires SameType<reference, Iterator<X>::reference> 
        && SameType<postincrement_result, Iterator<X>::postincrement_result>
        && Convertible<postincrement_result, const X&>
        && HasAssign<reference, Value> 
        && HasAssign<HasDereference<postincrement_result>::result_type, Value>;
}

Change 25.3.5.5 [forward.iterators]:

[ See 1084 which is attempting to change this same area in a compatible way. ]

concept ForwardIterator<typename X> : InputIterator<X>, Regular<X> { 
  requires Convertible<postincrement_result, const X&>;

  MoveConstructible postincrement_result;
  requires HasDereference<postincrement_result>
        && Convertible<HasDereference<postincrement_result>::result_type, const value_type&>;

  postincrement_result operator++(X&, int);

  axiom MultiPass(X a, X b) { 
    if (a == b) *a == *b; 
    if (a == b) ++a == ++b; 
  } 
}

-4- ...

postincrement_result operator++(X& r, int);

-5- Effects: equivalent to { X tmp = r; ++r; return tmp; }.


1010(i). operator-= should use default in concept

Section: 25.3.5.7 [random.access.iterators] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [random.access.iterators].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 263

This requirement on operator-= would be better expressed as a default implementation in the concept, with a matching axiom.

[ Batavia (2009-05): ]

The proposed resolution should also remove paragraph 5 and the declaration that precedes it. Further, we should provide an axiom that captures the desired semantics. This may be a broader policy to be applied. Move to Open.

Proposed resolution:

Change 25.3.5.7 [random.access.iterators]:

concept RandomAccessIterator<typename X> : BidirectionalIterator<X>, LessThanComparable<X> {
  ...
  X& operator-=(X& x, difference_type n) { return x += -n; }
  ...
}

1013(i). Remove IsSameType hold-over constraints

Section: 27.8.9 [alg.min.max] Status: NAD Editorial Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.min.max].

View all issues with NAD Editorial status.

Discussion:

Addresses UK 305

The negative requirement on IsSameType is a hold-over from an earlier draught with a variadic template form of min/max algorith. It is no longer necessary.

[ Batavia (2009-05): ]

We agree with the proposed resolution. Move to Tentatively Ready.

[ 2009-07 Frankfurt ]

We believe this is NAD, but this needs to be reviewed against the post-remove-concepts draft.

Proposed resolution:

Change 27 [algorithms]:

template<class T, StrictWeakOrder<auto, T> Compare>
  requires !SameType<T, Compare> && CopyConstructible<Compare>
  const T& min(const T& a, const T& b, Compare comp);
...
template<class T, StrictWeakOrder<auto, T> Compare>
  requires !SameType<T, Compare> && CopyConstructible<Compare>
  const T& max(const T& a, const T& b, Compare comp);
...
template<class T, StrictWeakOrder<auto, T> Compare>
  requires !SameType<T, Compare> && CopyConstructible<Compare>
  pair<const T&, const T&> minmax(const T& a, const T& b, Compare comp);

Change 27.8.9 [alg.min.max], p1, p9 and p17:

template<class T, StrictWeakOrder<auto, T> Compare>
  requires !SameType<T, Compare> && CopyConstructible<Compare>
  const T& min(const T& a, const T& b, Compare comp);
...
template<class T, StrictWeakOrder<auto, T> Compare>
  requires !SameType<T, Compare> && CopyConstructible<Compare>
  const T& max(const T& a, const T& b, Compare comp);
...
template<class T, StrictWeakOrder<auto, T> Compare>
  requires !SameType<T, Compare> && CopyConstructible<Compare>
  pair<const T&, const T&> minmax(const T& a, const T& b, Compare comp);

1015(i). C++ programs - but not users - need to provide support concept_maps

Section: 99 [concept.transform] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [concept.transform].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 199

The requirement that programs do not supply concept_maps should probably be users do not supply their own concept_map specializations. The program will almost certainly supply concept_maps - the standard itself supplies a specialization for RvalueOf references. Note that the term program is defined in 6.6 [basic.link]p1 and makes no account of the standard library being treated differently to user written code.

[ 2009-05-09 Alisdair adds: ]

The same problem is present in the words added for the LvalueReference/RvalueReference concepts last meeting.

With three subsections requiring the same constraint, I'm wondering if there is a better way to organise this section. Possible 20.2.1 -> 20.2.3 belong in the fundamental concepts clause in [concept.support]? While they can be implemented purely as a library feature without additional compiler support, they are pretty fundamental and we want the same restriction on user-concept maps as is mandated there.

[ Batavia (2009-05): ]

We agree with the issue, but believe the wording needs further improvement. We want to investigate current definitions for nomenclature such as "user" and "program." Move to Open pending the recommended investigation.

Proposed resolution:

Change 99 [concept.transform] p2:

-2- A program user shall not provide concept maps for any concept in 20.1.1.

Change [concept.true] p2:

-2- Requires: a program user shall not provide a concept map for the True concept.

Change [concept.classify] p2:

-2- Requires: a programuser shall not provide concept maps for any concept in this section.


1016(i). Provide LessThanComparable and EqualityComparable for FloatingPointType

Section: 99 [concept.comparison] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [concept.comparison].

View all issues with NAD Concepts status.

Discussion:

Addresses JP 33

LessThanComparable and EqualityComparable don't correspond to NaN.

Original proposed resolution:

Apply concept_map to these concepts at FloatingPointType.

[ Post Summit, Alisdair adds: ]

I don't understand the proposed resolution - there is no such thing as a 'negative' concept_map, and these concepts are auto concepts that match float/double etc. Also not clear how we are supposed to match values to concepts.

Recommend NAD and treat as a subset of issue 902.

Proposed resolution:

Recommend NAD.


1017(i). Floating-point types should not satisfy Regular

Section: 99 [concept.regular] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

Addresses US 66

Application of the Regular concept to floating-point types appears to be controversial (see long discussion on std-lib reflector).

Original proposed resolution:

State that the Regular concept does not apply to floating-point types.

[ Summit: ]

Recommend that we handle the same as JP 33 / 1016.

[ Post Summit, Alisdair adds: ]

Recommend Open, and review after resolution of 902 and revised axiom feature.

Proposed resolution:


1018(i). Trait specifications should be expressed in terms of concepts

Section: 21 [meta] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [meta].

View all other issues in [meta].

View all issues with NAD Concepts status.

Discussion:

Addresses US 70 [CD1]

Specifications now expressed via narrative text are more accurately and clearly expressed via executable code.

Wherever concepts are available that directly match this section's type traits, express the traits in terms of the concepts instead of via narrative text. Where the type traits do not quite match the corresponding concepts, bring the two into alignment so as to avoid two nearly-identical notions.

[ Summit: ]

We think that this is a good idea, but it requires a lot of work. If someone submits a paper proposing specific changes, we would be happy to review it at the next meeting.

Proposed resolution:


1020(i). Restore aligned_union

Section: 21.3.8.7 [meta.trans.other] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [meta.trans.other].

View all issues with NAD status.

Discussion:

Addresses UK 204 [CD1]

It is not possible to create a variant union based on a parameter pack expansion, e.g. to implement a classic discriminated union template.

Original proposed resolutuion:

Restore aligned_union template that was removed by LWG issue 856.

[ Summit: ]

Agree. The need for aligned_union is compelling enough to reinstate.

[ Post Summit, Alisdair adds: ]

Paper N2843 proposes an extension to the [[align]] attribute that further diminishes the need for this template. Recommend NAD.

[ 2009-10 Santa Cruz: ]

Mark NAD as suggested.

Proposed resolution:


1022(i). Pointer-safety API has nothing to do with smart pointers

Section: 99 [util.dynamic.safety] Status: NAD Editorial Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [util.dynamic.safety].

View all issues with NAD Editorial status.

Discussion:

Addresses UK 212 [CD1]

The pointer-safety API is nothing to do with smart pointers, so does not belong in [util.smartptr]. In fact it is a set of language support features are really belongs in clause [language.support], with the contents declared in a header that deals with language-support of memory management.

[ Summit: ]

Agree in principle, but not with the proposed resolution. We believe it belongs either a subsection of either 22 [utilities] or 20.2 [memory] as part of the general reorganization of 22 [utilities]. The declaration should stay in <memory>.

Proposed resolution:


1023(i). Unclear inheritance relation for std::function

Section: 22.10.17.3 [func.wrap.func] Status: NAD Editorial Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [func.wrap.func].

View all issues with NAD Editorial status.

Discussion:

Addresses DE 22 [CD1]

Related to issue 1114.

The conditions for deriving from std::unary_function and std::binary_function are unclear: The condition would also be satisfied if ArgTypes were std::vector<T1>, because it (arguably) "contains" T1.

[ Summit: ]

Agree. std::reference_wrapper has the same structure, and we suggest that std::function be presented in the same way as std::reference_wrapper.

[ 2009-05-09 Alisdair adds: ]

Phrasing should be "publicly and unambiguously derived from" and probably back in reference_wrapper too. Updated wording supplied.

[ Batavia (2009-05): ]

We agree with the proposed wording. Move to NAD Editorial.

Proposed resolution:

(no changes to <functional> synopsis required)

Change synopsis in Class template function 22.10.17.3 [func.wrap.func]:

template<Returnable R, CopyConstructible... ArgTypes> 
class function<R(ArgTypes...)> 
  : public unary_function<T1, R>      // iff sizeof...(ArgTypes) == 1 and see below
                                      // ArgTypes contains T1
  : public binary_function<T1, T2, R> // iff sizeof...(ArgTypes) == 2 and see below
                                      // ArgTypes contains T1 and T2
{
   ...

Add new p1/p2 before 22.10.17.3.2 [func.wrap.func.con]:

The template instantiation function<R(T1)> shall be publicly and unambiguously derived from std::unary_function<T1,R> if and only if the template type parameter is a function type taking one argument of type T1 and returning R.

The template instantiation function<R(T1,T2)> shall be publicly and unambiguously derived from std::binary_function<T1,T2,R> if and only if the template type parameter is a function type taking two arguments of type T1 and T2 and returning R.

explicit function();

1024(i). std::function constructors overly generous

Section: 22.10.17.3 [func.wrap.func] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [func.wrap.func].

View all issues with NAD Concepts status.

Discussion:

Addresses JP 39 [CD1]

There are no requires corresponding to F of std::function.

[ 2009-05-01 Daniel adds: ]

1070 removes the second constructor.

[ Batavia (2009-05): ]

We agree with the proposed resolution. Move to Tentatively Ready. If issue 1070 is accepted, the changes to the second constructor in this issue are moot.

[ 2009-07 Frankfurt: ]

Constructors have no definition.

Proposed resolution:

Correct as follows in 22.10.17.3 [func.wrap.func] (class definition)

 template<class F, Allocator Alloc>
   requires ConstructibleWithAllocator<F, Alloc>
     && call=Callable<F, ArgTypes...>
     && Convertible<call::result_type, R>
   function(allocator_arg_t, const Alloc&, F);
 template<class F, Allocator Alloc>
   requires ConstructibleWithAllocator<F,Alloc>
     && call=Callable<F, ArgTypes...>
     && Convertible<call::result_type, R>
   function(allocator_arg_t, const Alloc&, F&&);

1025(i). The library should provide more specializations for std::hash

Section: 22.10.19 [unord.hash] Status: NAD Future Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2020-08-21

Priority: Not Prioritized

View all other issues in [unord.hash].

View all issues with NAD Future status.

Discussion:

Addresses UK 208 [CD1]

std::hash should be implemented for much more of the standard library. In particular for pair, tuple and all the standard containers.

[2020-05-28; LEWG issue reviewing]

At the Kona 2017 LEWG issue processing session, this issue was marked as NAD. The formatting of the LWGissuesInLEWGStatus page ran 1025 into other text. A search of the page yields:

1025 20.14.15 [unord.hash] The library should provide more specializations for std::hash

NAD: We don't want to do this before solving LEWG21.

Status change to Open.

[2020-08-21 Issue processing telecon: follow LEWG recommendation. Status changed: Open → NAD Future.]

Proposed resolution:


1026(i). Smart pointers need to be concept-constrained templates

Section: 20.2 [memory] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [memory].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 209 [CD1]

Smart pointers cannot be used in constrained templates.

[ Summit: ]

We look forward to a paper on this topic. We recommend no action until a paper is available. We understand that a paper is forthcoming.

[ Peter Dimov adds: ]

shared_ptr<T> and weak_ptr<T> support all types T for which T* is valid. In other words, a possible (partial) resolution is to change class T to PointeeType T for shared_ptr, weak_ptr and possibly enable_shared_from_this.

Proposed resolution:


1027(i). std::allocator needs to be a concept-constrained template

Section: 20.2.10 [default.allocator] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [default.allocator].

View all other issues in [default.allocator].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 213 [CD1]

std::allocator should be constrained to simplify its use on constrained contexts. This library component models allocation from free store via the new operator so choose constraints to match. The Allocator concept allows for a wider variety of allocators that users may choose to supply if their allocation model does not require operator new, without impacting the requirements of this template.

Suggested direction:

The primary allocator template should be constrained to require ObjectType<T> and FreeStoreAllocatable<T>. Further operations to be constrained as required.

[ Summit: ]

Agree as stated. A future paper will address additional related issues.

Proposed resolution:


1028(i). raw_storage_iterator needs to be a concept-constrained template

Section: 99 [depr.storage.iterator] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2017-06-15

Priority: Not Prioritized

View all other issues in [depr.storage.iterator].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 214 [CD1]

raw_storage_iterator needs constraining as an iterator adaptor to be safely used in constrained templates

[ Summit: ]

We look forward to a paper on this topic. We recommend no action until a paper is available.

[ Post Summit Alisdair provided wording and rationale. ]

Proposed resolution:

20.2 [memory] p2

Update the synopsis for <memory>

// 20.7.8, raw storage iterator:
template <class ForwardIterator OutputIterator, class ObjectType T> 
  requires OutputIterator< OutIter, T >
    class raw_storage_iterator;

template <ForwardIterator OutIter, ObjectType T> 
  requires OutputIterator< OutIter, T >
  concept_map Iterator<raw_storage_iterator< OutIter, T > > { }

[storage.iterator] p1

Replace class template definition with:

namespace std { 
  template <class ForwardIterator OutputIterator, class ObjectType T> 
    requires OutputIterator< OutIter, T >
  class raw_storage_iterator 
    : public iterator<output_iterator_tag,void,void,void,void> { 
  public: 
    explicit raw_storage_iterator(OutputIterator x); 

    raw_storage_iterator<OutputIterator,T>& operator*(); 
    raw_storage_iterator<OutputIterator,T>& operator=(const T& element); 
    raw_storage_iterator<OutputIterator,T>& operator++(); 
    raw_storage_iterator<OutputIterator,T> operator++(int); 
  }; 

  template <ForwardIterator OutIter, ObjectType T> 
    requires OutputIterator< OutIter, T >
    concept_map Iterator<raw_storage_iterator< OutIter, T > > { }
}

Rationale:

raw_storage_iterator has to adapt a ForwardIterator, rather than just an InputIterator for two reasons:

  1. The initial iterator passed by value is expected to remain valid, pointing to the initialized region of memory.
  2. to avoid breaking the declaration of post-increment operator which would require some kind of proxy formulation to support generalised InputIterators.

1029(i). Specialized algorithms for memory management need to be concept-constrained templates

Section: 27.11 [specialized.algorithms] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [specialized.algorithms].

View all other issues in [specialized.algorithms].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 210 [CD1]

Related to 582

Specialized algorithms for memory management need requirements to be easily usable in constrained templates.

[ Summit: ]

We look forward to a paper on this topic. We recommend no action until a paper is available.

[ Post Summit Alisdair provided wording. ]

[ Post Summit: ]

Daniel adds:

  1. I suggest Size should require IntegralLike and not UnsignedIntegralLike, because otherwise simple int-literals could not be provided as arguments and it would conflict with other algorithms that only require IntegralLike.
  2. The current for-loop-test relies on evaluation in boolean context which is not provided by ArithmeticLike and it's refinements. I propose to change the corresponding for-loop-headers to:

    1. for uninitialized_copy_n: for ( ; n > Size(0); ++result, ++first, --n) {
    2. for uninitialized_fill_n: for (; n > Size(0); ++first, --n) {

Alisdair adds:

For the record I agree with Daniel's suggestion.

Proposed resolution:

20.2 [memory] p2

Update the synopsis for <memory>

template <class InputIterator InIter,
         class ForwardIterator OutputIterator<auto, InIter::reference> OutIter> 
   requires ForwardIterator<OutIter>
   ForwardIterator OutIter
   uninitialized_copy(InputIterator InIter first, InputIterator InIter last, 
                      ForwardIterator OutIter result);

template <class InputIterator InIter,
          class IntegralLike Size,
          class ForwardIterator OutputIterator<auto, InIter::reference> OutIter> 
  requires ForwardIterator<OutIter>
  ForwardIterator OutIter
  uninitialized_copy_n(InputIterator InIter first, Size n, 
                       ForwardIterator OutIter result);

template <class ForwardIterator Iter, class ObjectType T>
  requires Constructible< Iter::value_type, const T& >
  void uninitialized_fill(ForwardIterator Iter first, ForwardIterator Iter last, 
                          const T& x);

template <class ForwardIterator Iter, class IntegralLike Size, class ObjectType T> 
  requires Constructible< Iter::value_type, const T& >
  void
  uninitialized_fill_n(ForwardIterator Iter first, Size n, const T& x);

Update as follows:

uninitialized_copy 27.11.5 [uninitialized.copy]

template <class InputIterator InIter,
         class ForwardIterator OutputIterator<auto, InIter::reference> OutIter> 
   requires ForwardIterator<OutIter>
   ForwardIterator OutIter
   uninitialized_copy(InputIterator InIter first, InputIterator InIter last, 
                      ForwardIterator OutIter result);

-1- Effects:

for (; first != last; ++result, ++first)  {
   new (static_cast<void*>(&*result))
       typename iterator_traits<ForwardIterator> OutIter::value_type(*first);
}

-2- Returns: result

template <class InputIterator InIter,
          class IntegralLike Size,
          class ForwardIterator OutputIterator<auto, InIter::reference> OutIter> 
  requires ForwardIterator<OutIter>
  ForwardIterator OutIter
  uninitialized_copy_n(InputIterator InIter first, Size n, 
                       ForwardIterator OutIter result);

-3- Effects:

for ( ; n > Size(0); ++result, ++first, --n) {
   new (static_cast<void*>(&*result))
       typename iterator_traits<ForwardIterator> OutIter::value_type(*first);
}

-4- Returns: result

uninitialized_fill 27.11.7 [uninitialized.fill]

template <class ForwardIterator Iter, class ObjectType T>
  requires Constructible< Iter::value_type, const T& >
  void uninitialized_fill(ForwardIterator Iter first, ForwardIterator Iter last, 
                          const T& x);

-1- Effects:

for (; first != last; ++first) {
   new ( static_cast<void*>( &*first) ) 
       typename iterator_traits<ForwardIterator> Iter::value_type(x);
}

uninitialized_fill_n [uninitialized.fill.n]

template <class ForwardIterator Iter, class IntegralLike Size, class ObjectType T> 
  requires Constructible< Iter::value_type, const T& >
  void
  uninitialized_fill_n(ForwardIterator Iter first, Size n, const T& x);

-1- Effects:

for (; n-- > Size(0); ++first, --n) {
   new ( static_cast<void*>( &*first) ) 
       typename iterator_traits<ForwardIterator> Iter::value_type(x);
}

1031(i). Need shared_ptr conversion to a unique_ptr

Section: 20.3.2.2 [util.smartptr.shared] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2018-06-23

Priority: Not Prioritized

View all other issues in [util.smartptr.shared].

View all issues with NAD status.

Discussion:

Addresses US 78 [CD1]

There is presently no way to convert directly from a shared_ptr to a unique_ptr. Add an interface that performs the conversion.

[ Summit: ]

We look forward to a paper on this topic. We recommend no action until a paper is available. We believe that the shared pointer must use the default deleter for the conversion to succeed.

[ Peter Dimov adds: ]

This is basically a request for shared_ptr<>::release in disguise, with all the associated problems. Not a good idea.

[ 2009-07 post-Frankfurt: ]

The rationale for the omission of a release() member function from shared_ptr is given in: http://www.boost.org/doc/libs/1_39_0/libs/smart_ptr/shared_ptr.htm

The implementation of such a member is non-trivial (and maybe impossible), because it would need to account for the deleter.

[ 2009-07-26 Howard sets to Tentatively NAD Future. ]

I took an online poll and got 3 votes for NAD and 3 for NAD Future. Personally I prefer NAD Future as this does refer to an extension that could conceivably be considered beyond C++0X.

However such an extension would need to solve a couple of problems:

  1. What is the interface for such a conversion when the shared_ptr does not have unique ownership? Throw an exception? Create a null unique_ptr? Undefined behavior?
  2. How does one handle custom deleters given to the shared_ptr constructor?

    I do not believe it is possible to implement a general answer to this question. The shared_ptr deleter is a run time (or construction time) characteristic. The unique_ptr deleter is a compile time characteristic. In general one can not know to what type of unqiue_ptr you are converting to.

    One answer is for the user of the conversion to specify the deleter type and perhaps throw an exception if the specification turns out to be incorrect.

    Another answer is for the conversion to only be valid when the underlying deleter is default_delete. We would probalby need to specify that this is indeed the underlying deleter of a shared_ptr when a custom deleter is not given in the constructor.

At any rate, there are non-trivial design issues which would need to be implemented and tested in the field for usability prior to standardization.

[ 2009 Santa Cruz: ]

Moved to NAD Future.

[2017-02 in Kona, LEWG recommends NAD]

We just removed std::shared_ptr<>::unique() because it's misleading at best (especially in the face of weak_ptr). Resolving this is far larger than issues-processing — it would require significant design discussion and potentially an ABI break (stashing a mutex in the shared_ptr control block).

[2017-06-02 Issues Telecon]

Resolve as NAD

Proposed resolution:


1032(i). Tome utility templates need to be concept-constrained

Section: 29 [time] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [time].

View all issues with NAD Concepts status.

Discussion:

Addresses JP 45 [CD1]

Rep, Period, Clock and Duration don't correspond to concept.

template <class Rep, class Period = ratio<1>> class duration; 
template <class Clock, class Duration = typename Clock::duration> class time_point; 

Make concept for Rep, Period, Clock and Duration. Fix 29 [time] and wait_until and wait_for's template parameter at 33 [thread].

[ Summit: ]

We agree that this section needs concepts. We look forward to a paper on this topic. We recommend no action until a paper is available.

Proposed resolution:


1035(i). <array>::swap can invalidate references, pointers, and iterators

Section: 24.2.2.1 [container.requirements.general] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-12 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [container.requirements.general].

View all other issues in [container.requirements.general].

View all issues with NAD status.

Discussion:

Addresses UK 226 [CD1]

<array> must be added to this list. In particular it doesn't satisfy: - no swap() function invalidates any references, pointers, or iterators referring to the elements of the containers being swapped. and probably doesn't satisfy: - no swap() function throws an exception.

If <array> remains a container, this will have to also reference array, which will then have to say which of these points it satisfies.

[ Summit: ]

Agree. The proposed resolution is incomplete. Further work required.

[ 2009-05-01 Daniel adds: ]

Issue 1099 also suggests adding move constructor to this.

[ 2009-07 post-Frankfurt: ]

Howard is to draft a note that explains what happens to references.

[ 2009-10 Santa Cruz: ]

Mark as NAD. No consensus for change.

[ 2009-08-01 Howard provided wording. ]

Proposed resolution:

Add a paragraph to 24.3.7.4 [array.special]:

template <Swappable T, size_t N> void swap(array<T,N>& x, array<T,N>& y);

Effects:

swap_ranges(x.begin(), x.end(), y.begin());

[Note: Outstanding iterators, references and pointers may be invalidated. — end note]


1036(i). Remove iterator specification that is redundant due to concept constraints

Section: 24.2.4 [sequence.reqmts] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-12 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [sequence.reqmts].

View all other issues in [sequence.reqmts].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 231 [CD1]

p9-p11 are redundant now that Concepts define what it means to be an Iterator and guide overload resolution accordingly.

[ Summit: ]

Agree with issue and change to 24.2.4 [sequence.reqmts]. The changes required to 23 [strings] will be part of the general concept support for that clause.

Proposed resolution:

Strike 24.2.4 [sequence.reqmts]p9-11. Make sure std::basic_string has constraints similar to std::vector to meet this old guarantee.


1042(i). Provide ContiguousStorage concept and apply it to corresponding containers

Section: 24.3 [sequences] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [sequences].

View all issues with NAD status.

Discussion:

Addresses UK 244 [CD1]

The validity of the expression &a[n] == &a[0] + n is contingent on operator& doing the "right thing" (as captured by the CopyConstructible requirements in table 30 in C++2003). However this constraint has been lost in the Concepts of C++0x. This applies to vector and array (it actually applies to string also, but that's a different chapter, so I'll file a separate comment there and cross-reference).

Suggested solution:

Define a ContiguousStorage and apply it to vector, array and string.

[ Summit: ]

Agree with the issue but not the details of the proposed solution. Walter to provide wording for the new concept.

[ Post Summit Alisdair adds: ]

Another LWG subgroup wondered if this concept should extend to complex<T>, and so not be built on the container concept at all?

[ 2009-07 post-Frankfurt: ]

Leave Open, pending a post-Concepts Working Draft.

[ 2009-10 Santa Cruz: ]

Mark issue 1042 as NAD, in rationale state that this was solved by removal of concepts.

Proposed resolution:

Add to <container_concepts> synopsis in [container.concepts]

concept< typename C > ContiguousStorageContainer see below;

Add a new section to the end of [container.concepts]

23.1.6.x ContiguousStorageContainer concept [container.concepts.contiguous]

concept ContiguousStorageContainer< typename C >
  : Container<C>
{
  value_type* data(C&);

  axiom Contiguity(C& c, size_type i) {
    if( i < size(c) ) {
         addressof( * (data(c) + i) )
      == addressof( * advance(data(c), i) );
    }
  }
}

The ContiguousStorageContainer concept describes a container whose elements are allocated in a single region of memory, and are stored sequentially without intervening padding other than to meet alignment requirements. For example, the elements may be stored in a single array of suitable length.

value_type * data( C& );

Returns: a pointer to the first element in the region of storage. Result is unspecified for an empty container.

Change 24.3.7 [array] p1:

-1- The header <array> defines a class template for storing fixed-size sequences of objects. An array supports random access iterators. An instance of array<T, N> stores N elements of type T, so that size() == N is an invariant. The elements of an array are stored contiguously, meaning that if a is an array<T, N> then it obeys the identity &a[n] == &a[0] + n for all 0 <= n < N satisfies the concept ContiguousStorageContainer< array<T, N>>.

Add to the synopsis in 24.3.7 [array]:

    ...
    T * data(); 
    const T * data() const; 
  };

  template< typename T, size_t N >
    concept_map ContiguousStorageContainer< array<T, N>> {};
} 

Change 24.3.11 [vector] p1:

A vector is a sequence container that supports random access iterators. In addition, it supports (amortized) constant time insert and erase operations at the end; insert and erase in the middle take linear time. Storage management is handled automatically, though hints can be given to improve efficiency. The elements of a vector are stored contiguously, meaning that if v is a vector<T, Alloc> (where T is some type other than bool), then it obeys the identity &v[n] == &v[0] + n for all 0 <= n < v.size() satisfies the concept ContiguousStorageContainer< vector< T, Alloc>>.

Add at the end of the synopsis in 24.3.11 [vector] p2:

template< typename T, typename A >
  requires !SameType< T, bool >
  concept_map ContiguousStorageContainer< vector<T, A>> {};

Rationale:

Solved by removal of concepts.


1051(i). Specify subscript operation return types of reverse_iterator and move_iterator

Section: 25.5.1.6 [reverse.iter.elem], 25.5.4.6 [move.iter.elem] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-12 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [reverse.iter.elem].

View all issues with NAD status.

Discussion:

Addresses UK 279 [CD1]

The reason the return type became unspecified is LWG issue 386. This reasoning no longer applies as there are at least two ways to get the right return type with the new language facilities added since the previous standard.

Proposal: Specify the return type using either decltype or the Iter concept_map.

[ Summit: ]

Under discussion. This is a general question about all iterator adapters.

[ Howard adds post Summit: ]

I am requesting test cases to demonstrate a position.

[ 2009-07-24 Daniel adds: ]

I recommend NAD. Without concepts we can no longer restrict this member in a trivial way. Using decltype the declaration would be along the lines of

static const Iter& __base(); // not defined
auto operator[](difference_type n) const -> decltype(__base()[-n-1]);

but once reverse_iterator is instantiated for some given type Iter which cannot form a well-formed expression __base()[-n-1] this would cause an ill-formed function declaration, diagnostic required, and no silent SFINAE elimination.

[ 2009-10 Santa Cruz: ]

Moved to NAD.

[ 2009-10-22 Daniel adds: ]

IMO, my original comment regarding ill-formedness of the described construction is still correct, but I must add that I should weaken my assertion "Without concepts we can no longer restrict this member in a trivial way".

In fact with the existence of default template arguments for function templates it is not too hard to implement this like as follows, which shows that we can indeed simulate to some sense constrained member functions in C++0x.

My example does not really proof that the specification is easy, but it should be possible. I assume that the implementation would not be ABI compatible, though.

It is now your own decision how to proceed ;-)

#include <type_traits>
#include <cstddef>

template<class T>
typename std::add_rvalue_reference<T>::type declval();

template<class It>
struct reverse_iterator {
    It base;

    typedef std::ptrdiff_t difference_type;

    template<class U = It, class Res =
     decltype(declval<const U&>()[declval<difference_type>()])
    >
    Res operator[](difference_type n) const  {
        return base[-n-1];
    }
};

struct MyIter {
};

int main() {
    reverse_iterator<int*> ri;
    ri[0] = 2;
    reverse_iterator<MyIter> ri2;
}

The above declaration could be simplified, but the ideal solution

template<class U = It>
  decltype(declval<const U&>()[declval<difference_type>()])
     operator[](difference_type n) const;

does not work yet on gcc 4.4.1.

Proposed resolution:


1053(i). Unify algorithms with operator and function object variants

Section: 27 [algorithms] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-12 Last modified: 2017-06-05

Priority: Not Prioritized

View other active issues in [algorithms].

View all other issues in [algorithms].

View all issues with NAD status.

Discussion:

Addresses UK 295 [CD1]

There is a level of redundancy in the library specification for many algorithms that can be eliminated with the combination of concepts and default parameters for function templates. Eliminating redundancy simplified specification and reduces the risk of introducing accidental inconsistencies.

Proposed resolution: Adopt N2743.

[ Summit: ]

NAD, this change would break code that takes the address of an algorithm.

[ Post Summit Alisdair adds: ]

Request 'Open'. The issues in the paper go beyond just reducing the number of signatures, but cover unifying the idea of the ordering operation used by algorithms, containers and other library components. At least, it takes a first pass at the problem.

For me (personally) that was the more important part of the paper, and not clearly addressed by the Summit resolution.

[ 2009-10 Santa Cruz: ]

Too inventive, too late, would really need a paper. Moved to NAD Future.

[2017-02 in Kona, LEWG recommends NAD (NAD-ABI)]

Concept specific, might be solved by Ranges. Get it right in STD2 or with the next big ABI break.

[2017-06-02 Issues Telecon]

Resolve as NAD

Proposed resolution:


1056(i). Must all Engines and Distributions be Streamable?

Section: 28.5 [rand] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand].

View all issues with NAD status.

Discussion:

Both the concepts RandomNumberEngine and RandomNumberDistribution have requirements to be InputStreamable and OutputStreamable.

I have no problems leaving the WP in an inconsistent state on the best-faith assumption these concepts will be provided later, however disagree with the proposers that these constraints are not separable, orthogonal to the basic concepts of generating random number distributions.

These constraints should be dropped, and applied to specific algorithms as needed.

If a more refined concept (certainly deemed useful by the proposers) is proposed there is no objection, but the basic concept should not require persistence via streaming.

[ Batavia (2009-05): ]

Move to Open.

[ 2009-05-31 Alisdair adds: ]

Working on constraining the stream iterators, I have a few more observations to make on the concepts proposed while constraining the random number facility.

While I still believe the concerns are orthogonal, I don't believe the existing constraints go far enough either! The goal we want to achieve is not that a RandomNumberEngine / RandomNumberDistribution supports the stream operators, but that it is Serializable. I.e. there is a relationship between the insert and extract operations that guarantees to restore the state of the original object. This implies a coupling of the concepts together in a broader concept (Serializable) with at least one axiom to assert the semantics.

One problem is that istream and ostream may be fundamentally different types, although we can hook a relation if we are prepared to drop down to the char type and char_traits template parameters. Doing so ties us to a form of serialization that demands implementation via the std iostreams framework, which seems overly prescriptive. I believe the goal is generally to support serialization without regard to how it is expressed - although this is getting even more inventive in terms of concepts we do not have today.

[ 2009-11-03 Alisdair adds: ]

I can't find the record in the wiki minutes, but it was agreed at both Frankfurt and Santa Cruz that this issue is NAD.

The agreement in SC was that I would provide you with the rationale (see below) to include when moving to NAD.

[ 2009-11-03 Howard adds: ]

Moved to Tentatively NAD after 5 positive votes on c++std-lib.

Proposed resolution:

Rationale:

The issue suggests a more refined concept should be used if we want to require streaming, to separate concerns from the basic RandomNumberEngine behaviour. In Frankfurt it was observed that RandomNumberEngine is that more refined concept, and the basic concept used in the framework is UniformRandomNumberGenerator, which it refines.

We concur, and expect this to have no repurcussions re-writing this clause now concepts are removed.


1057(i). RandomNumberEngineAdaptor

Section: 28.5 [rand] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand].

View all issues with NAD Concepts status.

Discussion:

The RandomNumberEngineAdaptor concept breaks precedent in the way the library has been specified by grouping requirements into a concept that is never actually used in the library.

This is undoubtedly a very helpful device for documentation, but we are not comfortable with the precedent - especially as we have rejected national body comments on the same grounds.

Suggest either removing the concept, or providing an algorithm/type that requires this concept in their definition (such as a factory function to create new engines).

The preference is to create a single new algorithm and retain the value of the existing documentation.

[ Batavia (2009-05): ]

Walter points out that it is unlikely that any algorithm would ever require this concept, but that the concept nonetheless is useful as documentation, and (via concept maps) as a means of checking specific adapters.

Alisdair disagrees as to the concept's value as documentation.

Marc points out that the RandomNumberDistribution is also a concept not used elsewhere in the Standard.

Pete agrees that a policy of not inventing concepts that aren't used in the Standard is a good starting point, but should not be used as a criterion for rejecting a concept.

Move to Open.

Proposed resolution:


1058(i). New container issue

Section: 24.2.4 [sequence.reqmts] Status: NAD Editorial Submitter: Alisdair Meredith Opened: 2009-03-12 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [sequence.reqmts].

View all other issues in [sequence.reqmts].

View all issues with NAD Editorial status.

Discussion:

Sequence containers 24.2.4 [sequence.reqmts]:

The return value of new calls added to table 83 are not specified.

[ Batavia (2009-05): ]

We agree with the proposed resolution.

Move to NAD Editorial.

Proposed resolution:

Add after p6 24.2.4 [sequence.reqmts]:

-6- ...

The iterator returned from a.insert(p,rv) points to the copy of rv inserted into a.

The iterator returned from a.emplace(p, args) points to the new element constructed from args inserted into a.


1059(i). Usage of no longer existing FunctionType concept

Section: 22.10.17.3 [func.wrap.func] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2009-03-13 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [func.wrap.func].

View all issues with NAD Concepts status.

Discussion:

Due to a deliberate core language decision, the earlier called "foundation" concept std::FunctionType had been removed in N2773 shortly before the first "conceptualized" version of the WP (N2798) had been prepared. This caused a break of the library, which already used this concept in the adapted definition of std::function (22.10 [function.objects]/2, header <functional> synopsis and 22.10.17.3 [func.wrap.func]).

A simple fix would be to either (a) make std::function's primary template unconstrained or to (b) add constraints based on existing (support) concepts. A more advanced fix would (c) introduce a new library concept.

The big disadvantage of (a) is, that users can define templates which cause compiler errors during instantiation time because of under-constrainedness and would thus violate the basic advantage of constrained code.

For (b), the ideal constraints for std::function's template parameter would be one which excludes everything else but the single provided partial specialization that matches every "free function" type (i.e. any function type w/o cv-qualifier-seq and w/o ref-qualifier). Expressing such a type as as single requirement would be written as

template<typename T>
requires ReferentType<T> // Eliminate cv void and function types with cv-qual-seq
                         //   or ref-qual (depending on core issue #749)
      && PointeeType<T>  // Eliminate reference types
      && !ObjectType<T>  // Eliminate object types

Just for completeness approach (c), which would make sense, if the library has more reasons to constrain for free function types:

auto concept FreeFunctionType<typename T>
  : ReferentType<T>, PointeeType<T>, MemberPointeeType<T>
{
  requires !ObjectType<T>;
}

I mention that approach because I expect that free function types belong to the most natural type categories for every days coders. Potential candidates in the library are addressof and class template packaged_task.

[ Batavia (2009-05): ]

Alisdair would prefer to have a core-supported FunctionType concept in order that any future changes be automatically correct without need for a library solution to catch up; he points to type traits as a precedent. Further, he believes that a published concept can't in the future be changed.

Bill feels this category of entity would change sufficiently slowly that he would be willing to take the risk.

Of the discussed solutions, we tend toward option (c). We like the idea of having a complete taxonomy of native types, and perhaps erred in trimming the set.

We would like to have this issue reviewed by Core and would like their feedback. Move to Open.

Proposed resolution:

  1. Change in 22.10 [function.objects]/2, Header <functional> synopsis:

    // 20.6.16 polymorphic function wrappers:
    class bad_function_call;
    template<FunctionTypeReferentType F>
    requires PointeeType<F> && !ObjectType<F>
    class function; // undefined
    
  2. Change in 22.10.17.3 [func.wrap.func]:

    namespace std {
    template<FunctionTypeReferentType F>
    requires PointeeType<F> && !ObjectType<F>
    class function; // undefined
    

1060(i). Embedded nulls in NTBS

Section: 16.3.3.3.4.2 [byte.strings] Status: NAD Editorial Submitter: Alisdair Meredith Opened: 2009-03-13 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

Definition of null-terminated sequences allow for embedded nulls. This is surprising, and probably not supportable with the intended use cases.

[ Batavia (2009-05): ]

We agree with the issue, but believe this can be handled editorially. Move to NAD Editorial.

Proposed resolution:


1061(i). Bad indexing for tuple access to pair (Editorial?)

Section: 22.3.4 [pair.astuple] Status: NAD Editorial Submitter: Alisdair Meredith Opened: 2009-03-13 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [pair.astuple].

View all issues with NAD Editorial status.

Discussion:

The definition of get implies that get must return the second element if given a negative integer.

[ Batavia (2009-05): ]

Move to NAD Editorial.

Proposed resolution:

22.3.4 [pair.astuple] p5:

template<int size_t I, class T1, class T2> 
  requires True<(I < 2)> 
  const P& get(const pair<T1, T2>&);

1062(i). Missing insert_iterator for stacks/queues

Section: 25.5.2 [insert.iterators] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-13 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [insert.iterators].

View all issues with NAD status.

Discussion:

It is odd that we have an iterator to insert into a vector, but not an iterator to insert into a vector that is adapted as a stack. The standard container adapters all have a common interface to push and pop so it should be simple to create an iterator adapter to complete the library support.

We should provide an AdaptedContainer concept supporting push and pop operations. Create a new insert iterator and factory function that inserts values into the container by calling push.

[ Batavia (2009-05): ]

Walter recommends NAD Future.

Move to Open, and recommend deferring the issue until after the next Committee Draft is issued.

[ 2009-07-29 Howard moves to Tentatively NAD Future. ]

A poll on the LWG reflector voted unanimously to move this issue to Tentatively NAD Future.

[ 2009 Santa Cruz: ]

Moved to NAD. The intent of these adapters are to restrict the interfaces.

Proposed resolution:


1063(i). 03 iterator compatibilty

Section: 99 [iterator.backward] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

Which header must a user #include to obtain the library-supplied concept_maps declared in this paragraph?

This is important information, as existing user code will break if this header is not included, and we should make a point of mandating this header is #include-d by library headers likely to make use of it, notably <algorithm>. See issue 1001 for more details.

[ Batavia (2009-05): ]

We agree with the direction of the proposed resolution. Move to Tentatively Ready.

[ 2009-07 Frankfurt ]

We believe this is NAD Concepts, but this needs to be reviewed against the post-remove-concepts draft.

Proposed resolution:

Change [depr.lib.iterator.primitives], Iterator primitives, as indicated:

To simplify the use of iterators and provide backward compatibility with previous C++ Standard Libraries, the library provides several classes and functions. Unless otherwise specified, these classes and functions shall be defined in header <iterator>.

Change 99 [iterator.backward], Iterator backward compatibility, as indicated:

The library provides concept maps that allow iterators specified with iterator_traits to interoperate with algorithms that require iterator concepts. These concept maps shall be defined in the same header that defines the iterator. [Example:


1064(i). Term "object state" should not refer to classes

Section: 99 [defns.obj.state] Status: NAD Submitter: Howard Hinnant Opened: 2009-03-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [defns.obj.state].

View all issues with NAD status.

Discussion:

Addresses UK 152 [CD1]

Object state is using a definition of object (instance of a class) from outside the standard, rather than the 'region of storage' definiton in 6.7.2 [intro.object]p1

[ Summit: ]

We think we're removing this; See 99 [func.referenceclosure.cons].

[ 2009-10 Santa Cruz: ]

Mark as NAD. This will not affect user or implementer code

Proposed resolution:


1067(i). simplified wording for inner_product

Section: 27.10 [numeric.ops] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-17 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [numeric.ops].

View all issues with NAD Concepts status.

Discussion:

One of the motivating examples for introducing requirements-aliases was to simplify the wording of the inner_product requirements. As the paper adopting the feature and constrained wording for the library went through in the same meeting, it was not possible to make the change at the time. The simpler form should be adopted now though. Similarly, most the other numerical algorithms can benefit from a minor cleanup.

Note that in each case, the second more generalised form of the algorithm does not benefit, as there are already named constraints supplied by the template type parameters.

[ 2009-05-02 Daniel adds: ]

one part of the suggested resolution suggests the removal of the MoveConstructible<T> requirement from inner_product. According to 27.10.5 [inner.product]

Computes its result by initializing the accumulator acc with the initial value init

this step requires at least MoveConstructible.

Therefore I strongly suggest to take this removal back (Note also that the corresponding overload with a functor argument still has the same MoveConstructible<T> requirement).

[ Batavia (2009-05): ]

We agree with the proposed resolution as amended by Daniel's suggestion to restore MoveConstructible, reflected in the updated proposed resolution below.

Move to Tentatively Ready.

Proposed resolution:

Change in 27.10 [numeric.ops] and 27.10.3 [accumulate]:

template <InputIterator Iter, MoveConstructible T>
 requires add = HasPlus<T, Iter::reference>
       && HasAssign<T, HasPlus<T, Iter::reference> add::result_type>
 T accumulate(Iter first, Iter last, T init);

Change in 27.10 [numeric.ops] and 27.10.5 [inner.product]:

template <InputIterator Iter1, InputIterator Iter2, MoveConstructible T>
  requires mult = HasMultiply<Iter1::reference, Iter2::reference>
        && add = HasPlus<T, HasMultiply<Iter1::reference, Iter2::reference> mult::result_type>
        && HasAssign< 
             T,
             HasPlus<T,
                     HasMultiply<Iter1::reference, Iter2::reference>::result_type> add::result_type>
  T inner_product(Iter1 first1, Iter1 last1, Iter2 first2, T init);

Change in 27.10 [numeric.ops] and 27.10.7 [partial.sum]:

template <InputIterator InIter, OutputIterator<auto, const InIter::value_type&> OutIter>
  requires add = HasPlus<InIter::value_type, InIter::reference>
        && HasAssign<InIter::value_type,
                     HasPlus<InIter::value_type, InIter::reference> add::result_type>
        && Constructible<InIter::value_type, InIter::reference>
  OutIter partial_sum(InIter first, InIter last, OutIter result);

Change in 27.10 [numeric.ops] and 27.10.12 [adjacent.difference]:

template <InputIterator InIter, OutputIterator<auto, const InIter::value_type&> OutIter>
  requires sub = HasMinus<InIter::value_type, InIter::value_type>
        && Constructible<InIter::value_type, InIter::reference>
        && OutputIterator<OutIter, HasMinus<InIter::value_type, InIter::value_type> sub::result_type>
        && MoveAssignable<InIter::value_type>
  OutIter adjacent_difference(InIter first, InIter last, OutIter result);

1068(i). class random_device should be movable

Section: 28.5.7 [rand.device] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.device].

View all issues with NAD status.

Discussion:

class random_device should be movable.

[ Batavia (2009-05): ]

Move to Open, and recommend this issue be deferred until after the next Committee Draft is issued.

[ 2009-10 post-Santa Cruz: ]

Leave open. Walter to provide drafting as part of his planned paper.

[ 2010 Pittsburgh: Moved to NAD. ]

Rationale:

WP is correct as written.

Proposed resolution:


1069(i). class seed_seq should support efficient move operations

Section: 28.5.8.1 [rand.util.seedseq] Status: NAD Submitter: Alisdair Meredith Opened: 2009-03-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.util.seedseq].

View all issues with NAD status.

Discussion:

class seed_seq should support efficient move operations.

[ Batavia (2009-05): ]

Move to Open, and recommend this issue be deferred until after the next Committee Draft is issued.

[ 2009-10 post-Santa Cruz: ]

Leave open. Walter to provide drafting as part of his planned paper.

[ 2010 Pittsburgh: ]

seed_seq is explicitly not copyable, so, much like LWG issue 1068, LWG issue 1069 could be marked NAD to be consistent with this.

Proposed resolution:


1072(i). Is std::hash a constrained template or not?

Section: 22.10.19 [unord.hash] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-19 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [unord.hash].

View all issues with NAD Concepts status.

Discussion:

Is std::hash a constrained template or not?

According to class template hash 22.10.19 [unord.hash], the definition is:

template <class T>
struct hash : public std::unary_function<T, std::size_t> {
  std::size_t operator()(T val) const;
};

And so unconstrained.

According to the <functional> synopsis in p2 Function objects 22.10 [function.objects] the template is declared as:

template <ReferentType T> struct hash;

which would make hash a constrained template.

[ 2009-03-22 Daniel provided wording. ]

[ Batavia (2009-05): ]

Alisdair is not certain that Daniel's proposed resolution is sufficient, and recommends we leave the hash template unconstrained for now.

Recommend that the Project Editor make the constrained declaration consistent with the definition in order to make the Working Paper internally consistent, and that the issue then be revisited.

Move to Open.

Proposed resolution:

[To the editor: This resolution is merge-compatible to the resolution of 1078]

  1. In 22.10 [function.objects]/2, header <functional> synopsis, change as indicated:

    // 20.6.17, hash function base template:
    template <ReferentType T> struct hash; // undefined
    
  2. In 22.10.19 [unord.hash]/1 change as indicated:

    namespace std {
     template <class T>
     struct hash : public std::unary_function<T, std::size_t> {
     std::size_t operator()(T val) const;
     };
     template <ReferentType T> struct hash; // undefined
    }
    
  3. In 22.10.19 [unord.hash]/2 change as indicated:

    -2- For all library-provided specializations, the template instantiation hash<T> shall provide a public operator() with return type std::size_t to satisfy the concept requirement Callable<const hash<T>, const T&>. If T is an object type or reference to object, hash<T> shall be publicly derived from std::unary_function<T, std::size_t>. The return value of operator() is unspecified, except that equal arguments shall yield the same result. operator() shall not throw exceptions.

  4. In 17.7 [support.rtti]/1, header <typeinfo> synopsis change as indicated:

    namespace std {
      class type_info;
      class type_index;
      template <classReferentType T> struct hash;
    

1074(i). concept map broken by N2840

Section: 99 [allocator.element.concepts] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-19 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

p7 Allocator-related element concepts 99 [allocator.element.concepts]

The changes to the AllocatableElement concept mean this concept_map specialization no longer matches the original concept:

template <Allocator Alloc, class T, class ... Args>
  requires HasConstructor<T, Args...>
    concept_map AllocatableElement<Alloc, T, Args&&...> {
      void construct_element(Alloc& a, T* t, Args&&... args) {
        Alloc::rebind<T>(a).construct(t, forward(args)...);
      }
    }

[ 2009-03-23 Pablo adds: ]

Actually, this is incorrect, N2840 says. "In section 99 [allocator.element.concepts] paragraph 8, modify the definition of the AllocatableElement concept and eliminate the related concept map:" but then neglects to include the red-lined text of the concept map that was to be eliminated. Pete also missed this, but I caught it he asked me to review his edits. Pete's updated WP removes the concept map entirely, which was the original intent. The issue is, therefore, moot. Note, as per my presentation of N2840 in summit, construct() no longer has a default implementation. This regrettable fact was deemed (by David Abrahams, Doug, and myself) to be preferable to the complexity of providing a default implementation that would not under-constrain a more restrictive allocator (like the scoped allocators).

[ 2009-05-01 Daniel adds: ]

it seems to me that #1074 should be resolved as a NAD, because the current WP has already removed the previous AllocatableElement concept map. It introduced auto concept AllocatableElement instead, but as of 99 [allocator.element.concepts]/7 this guy contains now

requires FreeStoreAllocatable<T>;
void Alloc::construct(T*, Args&&...);

[ Batavia (2009-05): ]

The affected code is no longer part of the Working Draft.

Move to NAD.

Proposed resolution:

Change 99 [allocator.element.concepts]:

template <Allocator Alloc, class T, class ... Args>
  requires HasConstructor<T, Args...>
    concept_map AllocatableElement<Alloc, T, Args&&...> {
      void construct_element(Alloc& a, T* t, Args&&... args) {
        Alloc::rebind<T>(a).construct(t, forward(args)...);
      }
    }

1076(i). unary/binary_negate need constraining and move support

Section: 99 [depr.negators] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-03-20 Last modified: 2017-06-15

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

The class templates unary/binary_negate need constraining and move support.

Ideally these classes would be deprecated, allowing unary/binary_function to also be deprecated. However, until a generic negate adaptor is introduced that can negate any Callable type, they must be supported so should be constrained. Likewise, they should be movable, and support adopting a move-only predicate type.

In order to preserve ABI compatibility, new rvalue overloads are supplied in preference to changing the existing pass-by-const-ref to pass-by-value.

Do not consider the issue of forwarding mutable lvalues at this point, although remain open to another issue on the topic.

[ 2009-05-01 Daniel adds: ]

IMO the currently proposed resolution needs some updates because it is ill-formed at several places:

  1. In concept AdaptableUnaryFunction change

    typename X::result_type;
    typename X::argument_type;
    

    to

    Returnable result_type = typename X::result_type;
    typename argument_type = typename X::argument_type;
    

    [The replacement "Returnable result_type" instead of "typename result_type" is non-editorial, but maybe you prefer that as well]

  2. In concept AdaptableBinaryFunction change

    typename X::result_type;
    typename X::first_argument_type;
    typename X::second_argument_type;
    

    to

    Returnable result_type = typename X::result_type;
    typename first_argument_type = typename X::first_argument_type;
    typename second_argument_type = typename X::second_argument_type;
    

    [The replacement "Returnable result_type" instead of "typename result_type" is non-editorial, but maybe you prefer that as well.]

  3. In class unary/binary_function

    1. I suggest to change "ReturnType" to "Returnable" in both cases.
    2. I think you want to replace the remaining occurrences of "Predicate" by "P" (in both classes in copy/move from a predicate)
  4. I think you need to change the proposed signatures of not1 and not2, because they would still remain unconstrained: To make them constrained at least a single requirement needs to be added to enable requirement implication. This could be done via a dummy ("requires True<true>") or just explicit as follows:

    1. template <AdaptableUnaryFunction P>
      requires Predicate< P, P::argument_type>
      unary_negate<P> not1(const P&& pred);
      template <AdaptableUnaryFunction P>
      requires Predicate< P, P::argument_type >
      unary_negate<P> not1(P&& pred);
      

      -3- Returns: unary_negate<P>(pred).

      [Don't we want a move call for the second overload as in

      unary_negate<P>(std::move(pred))
      

      in the Returns clause ?]

    2. template <AdaptableBinaryFunction P>
      requires Predicate< P, P::first_argument_type, P::second_argument_type >
      binary_negate<P> not2(const P& pred);
      template <AdaptableBinaryFunction P>
      requires Predicate< P, P::first_argument_type, P::second_argument_type >
      binary_negate<P> not2(P&& pred);
      

      -5- Returns: binary_negate<P>(pred).

      [Don't we want a move call for the second overload as in

      binary_negate<P>(std::move(pred))
      

      in the Returns clause ?]

[ Batavia (2009-05): ]

There is concern that complicating the solution to preserve the ABI seems unnecessary, since we're not in general preserving the ABI.

We would prefer a separate paper consolidating all Clause 20 issues that are for the purpose of providing constrained versions of the existing facilities.

Move to Open.

[ 2009-10 post-Santa Cruz: ]

Leave open pending the potential move constructor paper. Note that we consider the "constraining" part NAD Concepts.

[ 2010-01-31 Alisdair removes the current proposed wording from the proposed wording section because it is based on concepts. That wording is proposed here: ]

Add new concepts where appropriate::

auto concept AdaptableUnaryFunction< typename X > {
  typename X::result_type;
  typename X::argument_type;
}

auto concept AdaptableBinaryFunction< typename X > {
  typename X::result_type;
  typename X::first_argument_type;
  typename X::second_argument_type;
}

Revise as follows:

Base [base] (Only change is constrained Result)

-1- The following classes are provided to simplify the typedefs of the argument and result types:

namespace std {
  template <class Arg, class ReturnType Result>
  struct unary_function {
     typedef Arg    argument_type;
     typedef Result result_type;
  };

  template <class Arg1, class Arg2, class ReturnType Result>
  struct binary_function {
     typedef Arg1   first_argument_type;
     typedef Arg2   second_argument_type;
     typedef Result result_type;
  };
}

Negators [negators]:

-1- Negators not1 and not2 take a unary and a binary predicate, respectively, and return their complements (5.3.1).

template <class AdaptableUnaryFunction Predicate>
  requires Predicate< P, P::argument_type >
  class unary_negate
    : public unary_function<typename Predicate::argument_type,bool> {
  public:
    unary_negate(const unary_negate & ) = default;
    unary_negate(unary_negate && );

    requires CopyConstructible< P >
       explicit unary_negate(const Predicate& pred); 
    requires MoveConstructible< P >
       explicit unary_negate(Predicate && pred);

    bool operator()(const typename Predicate::argument_type& x) const;
  };

-2 operator() returns !pred(x).

template <class Predicate>
  unary_negate<Predicate> not1(const Predicate&amp; pred);
template <class Predicate>
  unary_negate<Predicate> not1(Predicate&& pred);

-3- Returns: unary_negate<Predicate>(pred).

template <class AdaptableBinaryFunction Predicate >
  requires Predicate< P, P::first_argument_type, P::second_argument_type >
  class binary_negate
    : public binary_function<typename Predicate::first_argument_type,
                              typename Predicate::second_argument_type, bool> {
  public:
    biary_negate(const binary_negate & ) = default;
    binary_negate(binary_negate && );

    requires CopyConstructible< P >
       explicit binary_negate(const Predicate& pred);
    requires MoveConstructible< P >
       explicit binary_negate(const Predicate& pred);

    bool operator()(const typename Predicate::first_argument_type& x,
                    const typename Predicate::second_argument_type& y) const;
  };

-4- operator() returns !pred(x,y).

template <class Predicate>
  binary_negate<Predicate> not2(const Predicate& pred);
template <class Predicate>
  binary_negate<Predicate> not2(Predicate&& pred);

-5- Returns: binary_negate<Predicate>(pred).

[ 2010 Rapperswil: ]

Move to NAD Concepts. The move-semantic part has been addressed by a core language change, which implicitly generates appropriate move constructors and move-assignment operators.

Proposed resolution:


1077(i). Nonesense tuple declarations

Section: 22.4.4 [tuple.tuple] Status: NAD Editorial Submitter: Pete Becker Opened: 2009-03-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [tuple.tuple].

View all issues with NAD Editorial status.

Discussion:

Class template tuple 22.4.4 [tuple.tuple]:

template <class... UTypes>
  requires Constructible<Types, const UTypes&>...
template <class... UTypes>
  requires Constructible<Types, RvalueOf<UTypes>::type>...

Somebody needs to look at this and say what it should be.

[ 2009-03-21 Daniel provided wording. ]

[ Batavia (2009-05): ]

The resolution looks correct; move to NAD Editorial.

Proposed resolution:

In 22.4.4 [tuple.tuple], class tuple, change as indicated:

template <class... UTypes>
  requires Constructible<Types, const UTypes&>...
  tuple(const pair<UTypes...>&);
template <class... UTypes>
  requires Constructible<Types, RvalueOf<UTypes>::type>...
  tuple(pair<UTypes...>&&);

[NB.: The corresponding prototypes do already exist in 22.4.4.1 [tuple.cnstr]/7+8]


1078(i). DE-17: Remove class type_index

Section: 22.11 [type.index] Status: NAD Concepts Submitter: Doug Gregor Opened: 2009-03-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [type.index].

View all issues with NAD Concepts status.

Discussion:

Addresses DE 17

DE-17:

The class type_index should be removed; it provides no additional functionality beyond providing appropriate concept maps.

[ 2009-03-31 Peter adds: ]

It is not true, in principle, that std::type_index provides no utility compared to bare std::type_info*.

std::type_index can avoid the lifetime issues with type_info when the DLL that has produced the type_info object is unloaded. A raw type_info* does not, and cannot, provide any protection in this case. A type_index can (if the implementor so chooses) because it can wrap a smart (counted or even cloning) pointer to the type_info data that is needed for name() and before() to work.

Proposed resolution:

Modify the header <typeinfo> synopsis in 17.7 [support.rtti]p1 as follows:

namespace std { 
  class type_info; 
  class type_index;
  template <class T> struct hash;
  template<> struct hash<type_indexconst type_info *> : public std::unary_function<type_indexconst type_info *, size_t> {
    size_t operator()(type_indexconst type_info * indext) const;
  };
  concept_map LessThanComparable<const type_info *> see below
  class bad_cast; 
  class bad_typeid;
}

Add the following new subsection

18.7.1.1 Template specialization hash<const type_info *> [type.info.hash]

size_t operator()(const type_info *x) const;
  1. Returns: x->hash_code()

Add the following new subsection

18.7.1.2 type_info concept map [type.info.concepts]

concept_map LessThanComparable<const type_info *> {
  bool operator<(const type_info *x, const type_info *y) { return x->before(*y); }
  bool operator<=(const type_info *x, const type_info *y) { return !y->before(*x); }
  bool operator>(const type_info *x, const type_info *y) { return y->before(*x); }
  bool operator>=(const type_info *x, const type_info *y) { return !x->before(*y); }
}
  1. Note: provides a well-defined ordering among type_info const pointers, which makes such pointers usable in associative containers (23.4).

Remove section 22.11 [type.index]


1080(i). Concept ArithmeticLike should provide explicit boolean conversion

Section: 99 [concept.arithmetic] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2009-03-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

Astonishingly, the current concept ArithmeticLike as specified in 99 [concept.arithmetic] does not provide explicit conversion to bool although this is a common property of arithmetic types (7.3.15 [conv.bool]). Recent proposals that introduced such types (integers of arbitrary precision, n2143, decimals n2732 indirectly via conversion to long long) also took care of such a feature.

Adding such an explicit conversion associated function would also partly solve a currently invalid effects clause in library, which bases on this property, 25.3.5.7 [random.access.iterators]/2:

{ difference_type m = n;
 if (m >= 0) while (m--) ++r;
 else while (m++) --r;
 return r; }

Both while-loops take advantage of a contextual conversion to bool (Another problem is that the >= comparison uses the no longer supported existing implicit conversion from int to IntegralLike).

Original proposed resolution:

  1. In 99 [concept.arithmetic], add to the list of less refined concepts one further concept:

    concept ArithmeticLike<typename T>
      : Regular<T>, LessThanComparable<T>, HasUnaryPlus<T>, HasNegate<T>,
        HasPlus<T, T>, HasMinus<T, T>, HasMultiply<T, T>, HasDivide<T, T>,
        HasPreincrement<T>, HasPostincrement<T>, HasPredecrement<T>,
        HasPostdecrement<T>,
        HasPlusAssign<T, const T&>, HasMinusAssign<T, const T&>,
        HasMultiplyAssign<T, const T&>,
        HasDivideAssign<T, const T&>, ExplicitlyConvertible<T, bool> {
    
  2. In 25.3.5.7 [random.access.iterators]/2 change the current effects clause as indicated [The proposed insertion fixes the problem that the previous implicit construction from integrals has been changed to an explicit constructor]:

    { difference_type m = n;
     if (m >= difference_type(0)) while (m--) ++r;
     else while (m++) --r;
     return r; }
    

[ Batavia (2009-05): ]

We agree that arithmetic types ought be convertible to bool, and we therefore agree with the proposed resolution's paragraph 1.

We do not agree that the cited effects clause is invalid, as it expresses intent rather than specific code.

Move to Review, pending input from concepts experts.

Proposed resolution:

In 99 [concept.arithmetic], add to the list of less refined concepts one further concept:

concept ArithmeticLike<typename T>
  : Regular<T>, LessThanComparable<T>, HasUnaryPlus<T>, HasNegate<T>,
    HasPlus<T, T>, HasMinus<T, T>, HasMultiply<T, T>, HasDivide<T, T>,
    HasPreincrement<T>, HasPostincrement<T>, HasPredecrement<T>,
    HasPostdecrement<T>,
    HasPlusAssign<T, const T&>, HasMinusAssign<T, const T&>,
    HasMultiplyAssign<T, const T&>,
    HasDivideAssign<T, const T&>, ExplicitlyConvertible<T, bool> {

1081(i). basic_string needs to be a concept-constrained template

Section: 23 [strings] Status: NAD Concepts Submitter: Howard Hinnant Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [strings].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 216, JP 46, JP 48 [CD1]

All the containers use concepts for their iterator usage, exect for basic_string. This needs fixing.

Use concepts for iterator template parameters throughout the chapter.

[ Summit: ]

NB comments to be handled by Dave Abrahams and Howard Hinnant with advice from PJP: UK216 (which duplicates) JP46, JP48. JP46 supplies extensive proposed wording; start there.

Proposed resolution:


1082(i). codecvt needs to be a concept-constrained template

Section: 30 [localization] Status: NAD Concepts Submitter: Howard Hinnant Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [localization].

View all issues with NAD Concepts status.

Discussion:

Addresses JP 49 [CD1]

codecvt does not use concept. For example, create CodeConvert concept and change as follows.

template<CodeConvert Codecvt, class Elem = wchar_t>
  class wstring_convert {

[ Summit: ]

To be handled by Howard Hinnant, Dave Abrahams, Martin Sebor, PJ Plauger.

Proposed resolution:


1083(i). InputIterator and OutputIterator template parameters need to be concept constraints

Section: 30 [localization] Status: NAD Concepts Submitter: Howard Hinnant Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [localization].

View all issues with NAD Concepts status.

Discussion:

Addresses JP 52, JP 53 [CD1]

InputIterator does not use concept.

OutputIterator does not use concept.

Comments include proposed wording.

[ Summit: ]

To be handled by Howard Hinnant, Dave Abrahams, Martin Sebor, PJ Plauger.

Proposed resolution:


1084(i). Concept ForwardIterator should provide default implementation for post-increment

Section: 25.3.5.5 [forward.iterators] Status: NAD Concepts Submitter: Howard Hinnant Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [forward.iterators].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 250 [CD1]

A default implementation should be supplied for the post-increment operator to simplify implementation of iterators by users.

Copy the Effects clause into the concept description as the default implementation. Assumes a default value for postincrement_result.

[ Summit: ]

Howard will open an issue.

[ 2009-06-07 Daniel adds: ]

This issue cannot currently be resolved as suggested, because that would render auto-detection of the return type postincrement_result invalid, see [concept.map.assoc]/4+5. The best fix would be to add a default type to that associated type, but unfortunately any default type will prevent auto-deduction of types of associated functions as quoted above. A corresponding core issue is in preparation.

Proposed resolution:

[ This wording assumes the acceptance of UK 251 / 1009. Both wordings change the same paragraphs. ]

Change 25.3.5.5 [forward.iterators]:

concept ForwardIterator<typename X> : InputIterator<X>, Regular<X> { 

  MoveConstructible postincrement_result;
  requires HasDereference<postincrement_result>
        && Convertible<HasDereference<postincrement_result>::result_type, const value_type&>;

  postincrement_result operator++(X& r, int); {
     X tmp = r;
     ++r;
     return tmp;
  }

  axiom MultiPass(X a, X b) { 
    if (a == b) *a == *b; 
    if (a == b) ++a == ++b; 
  } 
}

1085(i). BidirectionalIterator concept should provide default implementation for post-decrement

Section: 25.3.5.6 [bidirectional.iterators] Status: NAD Concepts Submitter: Howard Hinnant Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [bidirectional.iterators].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 258 [CD1]

A default implementation should be supplied for the post-decrement operator to simplify implementation of iterators by users.

Copy the Effects clause into the concept description as the default implementation. Assumes a default value for postincrement_result.

[ Summit: ]

Howard will open an issue.

[ 2009-06-07 Daniel adds: ]

This issue cannot currently be resolved as suggested, because that would render auto-detection of the return type postdecrement_result invalid, see 1084.

Proposed resolution:

Change 25.3.5.6 [bidirectional.iterators]:

concept BidirectionalIterator<typename X> : ForwardIterator<X> { 
  MoveConstructible postdecrement_result; 
  requires HasDereference<postdecrement_result> 
        && Convertible<HasDereference<postdecrement_result>::result_type, const value_type&> 
        && Convertible<postdecrement_result, const X&>; 
  X& operator--(X&); 
  postdecrement_result operator--(X& r, int); {
     X tmp = r;
     --r;
     return tmp;
  }
}

1086(i). Stream iterators need to be concept-constrained templates

Section: 25.6 [stream.iterators] Status: NAD Concepts Submitter: Howard Hinnant Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [stream.iterators].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 284 [CD1]

The stream iterators need constraining with concepts/requrires clauses.

[ Summit: ]

We agree. To be handled by Howard, Martin and PJ.

Proposed resolution:


1087(i). Incorrect OutputIterator concept requirements for replace algorithms

Section: 27.7.5 [alg.replace] Status: NAD Concepts Submitter: Howard Hinnant Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.replace].

View all issues with NAD Concepts status.

Discussion:

Addresses UK 301 [CD1]

replace and replace_if have the requirement: OutputIterator<Iter, Iter::reference> Which implies they need to copy some values in the range the algorithm is iterating over. This is not however the case, the only thing that happens is const T&s might be copied over existing elements (hence the OutputIterator<Iter, const T&>.

Remove OutputIterator<Iter, Iter::reference> from replace and replace_if.

[ Summit: ]

We agree. To be handled by Howard.

Proposed resolution:

Change in [algorithms.syn] and 27.7.5 [alg.replace]:

template<ForwardIterator Iter, class T> 
  requires OutputIterator<Iter, Iter::reference> 
        && OutputIterator<Iter, const T&> 
        && HasEqualTo<Iter::value_type, T> 
  void replace(Iter first, Iter last, 
               const T& old_value, const T& new_value); 

template<ForwardIterator Iter, Predicate<auto, Iter::value_type> Pred, class T> 
  requires OutputIterator<Iter, Iter::reference> 
        && OutputIterator<Iter, const T&> 
        && CopyConstructible<Pred> 
  void replace_if(Iter first, Iter last,
                  Pred pred, const T& new_value);

1091(i). Multimap description confusing

Section: 24.4.5.3 [multimap.modifiers] Status: NAD Submitter: LWG Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses UK 246

The content of this sub-clause is purely trying to describe in words the effect of the requires clauses on these operations, now that we have Concepts. As such, the description is more confusing than the signature itself. The semantic for these functions is adequately covered in the requirements tables in 24.2.7 [associative.reqmts].

[ Beman adds: ]

Pete is clearly right that this one is technical rather than editorial.

[ Batavia (2009-05): ]

We agree with the proposed resolution.

Move to Review.

[ 2009-10 Santa Cruz: ]

Mark as NAD, solved by removing concepts.

Proposed resolution:

Strike 24.4.5.3 [multimap.modifiers] entirely (but do NOT strike these signatures from the class template definition!).


1092(i). Class template integral_constant should be a constrained template

Section: 21.3.4 [meta.help] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2009-03-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [meta.help].

View all issues with NAD Concepts status.

Discussion:

A first step to change the type traits predicates to constrained templates is to constrain their common base template integral_constant. This can be done, without enforcing depending classes to be constrained as well, but not vice versa without brute force late_check usages. The following proposed resolution depends on the resolution of LWG issue 1019.

[ Batavia (2009-05): ]

Move to Open, pending a paper that looks at constraints for the entirety of the type traits and their relationship to the foundation concepts. We recommend this be deferred until after the next Committee Draft is issued.

Proposed resolution:

  1. In 21.3.3 [meta.type.synop], Header <type_traits> synopsis change as indicated:

    namespace std {
    // 20.5.3, helper class:
    template <classIntegralConstantExpressionType T, T v> struct integral_constant;
    
  2. In 21.3.4 [meta.help] change as indicated:

    template <classIntegralConstantExpressionType T, T v>
    struct integral_constant {
      static constexpr T value = v;
      typedef T value_type;
      typedef integral_constant<T,v> type;
      constexpr operator value_type() { return value; }
    };
    

1096(i). unconstrained rvalue ref parameters

Section: 16 [library] Status: NAD Concepts Submitter: David Abrahams Opened: 2009-03-21 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD Concepts status.

Discussion:

TODO: Look at all cases of unconstrained rvalue ref parameters and check that concept req'ts work when T deduced as reference.

We found some instances where that was not done correctly and we figure the possibility of deducing T to be an lvalue reference was probably overlooked elsewhere.

[ Batavia (2009-05): ]

Move to Open, pending proposed wording from Dave for further review.

Proposed resolution:


1099(i). Various issues

Section: 16 [library] Status: NAD Submitter: David Abrahams Opened: 2009-03-21 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Notes

[2009-03-21 Sat] p. 535 at the top we need MoveConstructible V1, MoveConstructible V2 (where V1,V2 are defined on 539). Also make_tuple on 550

CD-1 reads:

template <MoveConstructible T1, MoveConstructible T2> 
pair<V1, V2> make_pair(T1&&, T2&&); 

Actually I'm guessing we need something like MoveConstructible<V1,T1>, i.e. "V1 can be constructed from an rvalue of type T1."

Ditto for make_tuple

[2009-03-21 Sat] p1183 thread ctor, and in general, we need a way to talk about "copiable from generalized rvalue ref argument" for cases where we're going to forward and copy.

This issue may well be quite large. Language in para 4 about "if an lvalue" is wrong because types aren't expressions.

Maybe we should define the term "move" so we can just say in the effects, "f is moved into the newly-created thread" or something, and agree (and ideally document) that saying "f is moved" implies

F x(move(f))

is required to work. That would cover both ctors at once.

p1199, call_once has all the same issues.

[2009-03-21 Sat] p869 InputIterator pointer type should not be required to be convertible to const value_type*, rather it needs to have a operator-> of its own that can be used for the value type.

This one is serious and unrelated to the move issue.

[2009-03-21 Sat] p818 stack has the same problem with default ctor.

[2009-03-21 Sat] p816 priority_queue has the same sorts of problems as queue, only more so

   requires MoveConstructible<Cont> 
     explicit priority_queue(const Compare& x = Compare(), Cont&& = Cont()); 

Don't require MoveConstructible when default constructing Cont. Also missing semantics for move ctor.

[2009-03-21 Sat] Why are Allocators required to be CopyConstructible as opposed to MoveConstructible?

[2009-03-21 Sat] p813 queue needs a separate default ctor (Cont needn't be MoveConstructible). No documented semantics for move c'tor. Or *any* of its 7 ctors!

[2009-03-21 Sat] std::array should have constructors for C++0x, consequently must consider move construction.

[ 2009-05-01 Daniel adds: ]

This could be done as part of 1035, which already handles deviation of std::array from container tables.

[2009-03-21 Sat] p622 all messed up.

para 8 "implementation-defined" is the wrong term; should be "see below" or something.

para 12 "will be selected" doesn't make any sense because we're not talking about actual arg types.

paras 9-13 need to be totally rewritten for concepts.

[2009-03-21 Sat] Null pointer comparisons (p587) have all become unconstrained. Need to fix that

[2009-03-21 Sat] mem_fun_t etc. definition doesn't match declaration. We think CopyConstructible is the right reqt.

make_pair needs Constructible<V1, T1&&> requirements!

make_tuple needs something similar

tuple bug in synopsis:

   template <class... UTypes>
   requires Constructible<Types, const UTypes&>...
   template <class... UTypes>
   requires Constructible<Types, RvalueOf<UTypes>::type>...

Note: removal of MoveConstructible requirements in std::function makes these routines unconstrained!

[ 2009-05-02 Daniel adds: ]

This part of the issue is already covered by 1077.

these unique_ptr constructors are broken [ I think this is covered in "p622 all messed up" ]

 unique_ptr(pointer p, implementation-defined d);
 unique_ptr(pointer p, implementation-defined d);

multimap range constructor should not have MoveConstructible<value_type> requirement.

same with insert(..., P&&); multiset has the same issue, as do unordered_multiset and unordered_multimap. Review these!

[ Batavia (2009-05): ]

Move to Open, pending proposed wording from Dave for further review.

[ 2009-10 post-Santa Cruz: ]

Tentatively NAD. We are not sure what has been addressed and what hasn't. Recommend closing unless someone sorts this out into something more readable.

Rationale:

The issue(s) at hand not adequately communicated.

Proposed resolution:


1101(i). unique requirements

Section: 27.7.9 [alg.unique] Status: NAD Editorial Submitter: Howard Hinnant Opened: 2009-04-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.unique].

View all issues with NAD Editorial status.

Discussion:

From Message c++std-core-14160 Howard wrote:

It was the intent of the rvalue reference proposal for unique to only require MoveAssignable: N1860.

And Pete replied:

That was overridden by the subsequent changes made for concepts in N2573, which reimposed the C++03 requirements.

My impression is that this overwrite was a simple (unintentional) mistake. Wording below to correct it.

[ Batavia (2009-05): ]

Howard notes this issue resolves a discrepancy between the synopsis and the description.

Move to NAD Editorial.

Proposed resolution:

Change 27.7.9 [alg.unique]:

template<ForwardIterator Iter> 
  requires OutputIterator<Iter, RvalueOf<Iter::reference>::type> 
        && EqualityComparable<Iter::value_type> 
  Iter unique(Iter first, Iter last); 

template<ForwardIterator Iter, EquivalenceRelation<auto, Iter::value_type> Pred> 
  requires OutputIterator<Iter, RvalueOf<Iter::reference>::type> 
        && CopyConstructible<Pred> 
  Iter unique(Iter first, Iter last, Pred pred);

Note that the synopsis in [algorithms.syn] is already correct.


1105(i). Shouldn't Range be an auto concept

Section: 99 [iterator.concepts.range] Status: NAD Concepts Submitter: David Abrahams Opened: 2009-04-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

[ 2009-04-26 Herb adds: ]

Here's a common example: We have many ISV customers who have built lots of in-house STL-like containers. Imagine that, for the past ten years, the user has been happily using his XYZCorpContainer<T> that has begin() and end() and an iterator typedef, and indeed satisfies nearly all of Container, though maybe not quite all just like valarray. The user upgrades to a range-enabled version of a library, and now lib_algo( xyz.begin(), xyz.end()); no longer works -- compiler error.

Even though XYZCorpContainer matches the pre-conceptized version of the algorithm, and has been working for years, it appears the user has to write at least this:

template<class T> concept_map Range<XYZCorpContainer<T>> {};

template<class T> concept_map Range<const XYZCorpContainer<T>> {};

Is that correct?

But he may actually have to write this as we do for initializer list:

template<class T>
concept_map Range<XYZCorpContainer<T>> {
   typedef T* iterator;
   iterator begin(XYZCorpContainer<T> c) { return c.begin(); }
   iterator end(XYZCorpContainer<T> c) { return c.end(); }
};

template<class T>
concept_map Range<const XYZCorpContainer<T>> {
   typedef T* iterator;
   iterator begin(XYZCorpContainer<T> c) { return c.begin(); }
   iterator end(XYZCorpContainer<T> c) { return c.end(); }
};

[ 2009-04-28 Alisdair adds: ]

I recommend NAD, although remain concerned about header organisation.

A user container will satisfy the MemberContainer concept, which IS auto. There is a concept_map for all MemberContainers to Container, and then a further concept_map for all Container to Range, so the stated problem is not actually true. User defined containers will automatically match the Range concept without explicitly declaring a concept_map.

The problem is that they should now provide an additional two headers, <iterator_concepts> and <container_concepts>. The only difference from making Range an auto concept would be this reduces to a single header, <iterator_concepts>.

I am strongly in favour of any resolution that tackles the issue of explicitly requiring concept headers to make these concept maps available.

[ Batavia (2009-05): ]

We observe there is a recent paper by Bjarne that overlaps this issue.

Alisdair continues to recommend NAD.

Move to Open, and recommend the issue be deferred until after the next Committee Draft is issued.

Proposed resolution:


1107(i). constructor shared_future(unique_future) by value?

Section: 33.10.8 [futures.shared.future] Status: NAD Editorial Submitter: Thomas J. Gritzan Opened: 2009-04-03 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [futures.shared.future].

View all issues with NAD Editorial status.

Discussion:

In the shared_future class definition in [futures.shared_future] the move constructor that constructs a shared_future from an unique_future receives the parameter by value. In paragraph 3, the same constructor receives it as const value.

I think that is a mistake and the constructor should take a r-value reference:

shared_future(unique_future<R>&& rhs);

[ Batavia (2009-05): ]

We agree with the proposed resolution.

Move to Tentatively Ready.

[ 2009-07-05 Daniel notes: ]

The proposed change has already been incorported into the current working draft N2914.

Proposed resolution:

Change the synopsis in [futures.shared_future]:

shared_future(unique_future<R>&& rhs);

Change the definition of the constructor in [futures.shared_future]:

shared_future(const unique_future<R>&& rhs);

1109(i). std::includes should require CopyConstructible predicate

Section: 27.8.7.2 [includes] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-04-28 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [includes].

View all issues with NAD Concepts status.

Discussion:

All the set operation algorithms require a CopyConstructible predicate, with the exception of std::includes. This looks like a typo as much as anything, given the general library requirement that predicates are copy constructible, and wording style of other set-like operations.

[ Batavia (2009-05): ]

We agree with the proposed resolution. Move to NAD Editorial.

Proposed resolution:

Change [algorithms.syn] and 27.8.7.2 [includes]:

template<InputIterator Iter1, InputIterator Iter2,
         typename CopyConstructible Compare>
  requires Predicate<Compare, Iter1::value_type, Iter2::value_type>
        && Predicate<Compare, Iter2::value_type, Iter1::value_type>
  bool includes(Iter1 first1, Iter1 last1,
                Iter2 first2, Iter2 last2,
                Compare comp);

1111(i). associative containers underconstrained

Section: 24.4 [associative] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-04-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [associative].

View all issues with NAD Concepts status.

Discussion:

According to table 87 (n2857) the expression X::key_equal for an unordered container shall return a value of type Pred, where Pred is an equivalence relation.

However, all 4 containers constrain Pred to be merely a Predicate, and not EquivalenceRelation.

[ Batavia (2009-05): ]

We agree with the proposed resolution.

Move to Review.

Proposed resolution:

For ordered containers, replace

Predicate<auto, Key, Key> Compare = less<Key>

with

StrictWeakOrder<auto, Key, Key> Compare = less<Key>

For unordered containers, replace

Predicate<auto, Key, Key> Compare = less<Key>

with

EquivalenceRelation<auto, Key, Key> Compare = less<Key>

As in the following declarations:

Associative containers 24.4 [associative]

1 Headers <map> and <set>:

Header <map> synopsis

   namespace std {
     template <ValueType Key, ValueType T,
               PredicateStrictWeakOrder<auto, Key, Key> Compare = less<Key>,
               Allocator Alloc = allocator<pair<const Key, T> > >
       requires NothrowDestructible<Key> && NothrowDestructible<T>
             && CopyConstructible<Compare>
             && AllocatableElement<Alloc, Compare, const Compare&>
             && AllocatableElement<Alloc, Compare, Compare&&>
     class map;

     ...

     template <ValueType Key, ValueType T,
               PredicateStrictWeakOrder<auto, Key, Key> Compare = less<Key>,
               Allocator Alloc = allocator<pair<const Key, T> > >
       requires NothrowDestructible<Key> && NothrowDestructible<T>
             && CopyConstructible<Compare>
             && AllocatableElement<Alloc, Compare, const Compare&>
             && AllocatableElement<Alloc, Compare, Compare&&>
     class multimap;

     ...

   }

Header <set> synopsis

   namespace std {
     template <ValueType Key, PredicateStrictWeakOrder<auto, Key, Key> Compare = less<Key>,
               Allocator Alloc = allocator<Key> >
       requires NothrowDestructible<Key> && CopyConstructible<Compare>
             && AllocatableElement<Alloc, Compare, const Compare&>
             && AllocatableElement<Alloc, Compare, Compare&&>
     class set;

     ...

     template <ValueType Key, PredicateStrictWeakOrder<auto, Key, Key> Compare = less<Key>,
               Allocator Alloc = allocator<Key> >
       requires NothrowDestructible<Key> && CopyConstructible<Compare>
             && AllocatableElement<Alloc, Compare, const Compare&>
             && AllocatableElement<Alloc, Compare, Compare&&>
     class multiset;

     ...

   }

23.4.1p2 Class template map [map]

 namespace std {
   template <ValueType Key, ValueType T,
             PredicateStrictWeakOrder<auto, Key, Key> Compare = less<Key>,
             Allocator Alloc = allocator<pair<const Key, T> > >
     requires NothrowDestructible<Key> && NothrowDestructible<T>
           && CopyConstructible<Compare>
           && AllocatableElement<Alloc, Compare, const Compare&>
           && AllocatableElement<Alloc, Compare, Compare&&>
   class map {
     ...
   };
 }

23.4.2p2 Class template multimap [multimap]

 namespace std {
   template <ValueType Key, ValueType T,
             PredicateStrictWeakOrder<auto, Key, Key> Compare = less<Key>,
             Allocator Alloc = allocator<pair<const Key, T> > >
     requires NothrowDestructible<Key> && NothrowDestructible<T>
           && CopyConstructible<Compare>
           && AllocatableElement<Alloc, Compare, const Compare&>
           && AllocatableElement<Alloc, Compare, Compare&&>
   class multimap {
     ...
   };
 }

23.4.3p2 Class template set [set]

 namespace std {
   template <ValueType Key, PredicateStrictWeakOrder<auto, Key, Key> Compare = less<Key>,
             Allocator Alloc = allocator<Key> >
     requires NothrowDestructible<Key> && CopyConstructible<Compare>
           && AllocatableElement<Alloc, Compare, const Compare&>
           && AllocatableElement<Alloc, Compare, Compare&&>
   class set {
     ...
   };
 }

23.4.4p2 Class template multiset [multiset]

 namespace std {
   template <ValueType Key, PredicateStrictWeakOrder<auto, Key, Key> Compare = less<Key>,
             Allocator Alloc = allocator<Key> >
     requires NothrowDestructible<Key> && CopyConstructible<Compare>
           && AllocatableElement<Alloc, Compare, const Compare&>
           && AllocatableElement<Alloc, Compare, Compare&&>
   class multiset {
     ...
   };
 }

23.5 Unordered associative containers [unord]

1 Headers <unordered_map> and <unordered_set>:

Header <unordered_map> synopsis

 namespace std {
   // 23.5.1, class template unordered_map:
   template <ValueType Key,
             ValueType T,
             Callable<auto, const Key&> Hash = hash<Key>,
             PredicateEquivalenceRelation<auto, Key, Key> Pred = equal_to<Key>,
             Allocator Alloc = allocator<pair<const Key, T> > >
     requires NothrowDestructible<Key> && NothrowDestructible<T>
           && SameType<Hash::result_type, size_t>
           && CopyConstructible<Hash> && CopyConstructible<Pred>
           && AllocatableElement<Alloc, Pred, const Pred&>
           && AllocatableElement<Alloc, Pred, Pred&&>
           && AllocatableElement<Alloc, Hash, const Hash&>
           && AllocatableElement<Alloc, Hash, Hash&&>
     class unordered_map;

   // 23.5.2, class template unordered_multimap:
   template <ValueType Key,
             ValueType T,
             Callable<auto, const Key&> Hash = hash<Key>,
             PredicateEquivalenceRelation<auto, Key, Key> Pred = equal_to<Key>,
             Allocator Alloc = allocator<pair<const Key, T> > >
     requires NothrowDestructible<Key> && NothrowDestructible<T>
           && SameType<Hash::result_type, size_t>
           && CopyConstructible<Hash> && CopyConstructible<Pred>
           && AllocatableElement<Alloc, Pred, const Pred&>
           && AllocatableElement<Alloc, Pred, Pred&&>
           && AllocatableElement<Alloc, Hash, const Hash&>
           && AllocatableElement<Alloc, Hash, Hash&&>
     class unordered_multimap;

   ...
 }

Header <unordered_set> synopsis

 namespace std {
   // 23.5.3, class template unordered_set:
   template <ValueType Value,
             Callable<auto, const Value&> Hash = hash<Value>,
             PredicateEquivalenceRelation<auto, Value, Value> class Pred = equal_to<Value>,
             Allocator Alloc = allocator<Value> >
     requires NothrowDestructible<Value>
           && SameType<Hash::result_type, size_t>
           && CopyConstructible<Hash> && CopyConstructible<Pred>
           && AllocatableElement<Alloc, Pred, const Pred&>
           && AllocatableElement<Alloc, Pred, Pred&&>
           && AllocatableElement<Alloc, Hash, const Hash&>
           && AllocatableElement<Alloc, Hash, Hash&&>
     class unordered_set;

   // 23.5.4, class template unordered_multiset:
   template <ValueType Value,
             Callable<auto, const Value&> Hash = hash<Value>,
             PredicateEquivalenceRelation<auto, Value, Value> class Pred = equal_to<Value>,
             Allocator Alloc = allocator<Value> >
     requires NothrowDestructible<Value>
           && SameType<Hash::result_type, size_t>
           && CopyConstructible<Hash> && CopyConstructible<Pred>
           && AllocatableElement<Alloc, Pred, const Pred&>
           && AllocatableElement<Alloc, Pred, Pred&&>
           && AllocatableElement<Alloc, Hash, const Hash&>
           && AllocatableElement<Alloc, Hash, Hash&&>
     class unordered_multiset;

   ...
 }

23.5.1p3 Class template unordered_map [unord.map]

 namespace std {
   template <ValueType Key,
             ValueType T,
             Callable<auto, const Key&> Hash = hash<Key>,
             PredicateEquivalenceRelation<auto, Key, Key> Pred = equal_to<Key>,
             Allocator Alloc = allocator<pair<const Key, T> > >
     requires NothrowDestructible<Key> && NothrowDestructible<T>
           && SameType<Hash::result_type, size_t>
           && CopyConstructible<Hash> && CopyConstructible<Pred>
           && AllocatableElement<Alloc, Pred, const Pred&>
           && AllocatableElement<Alloc, Pred, Pred&&>
           && AllocatableElement<Alloc, Hash, const Hash&>
           && AllocatableElement<Alloc, Hash, Hash&&>
   class unordered_map
   {
     ...
   };
 }

23.5.2p3 Class template unordered_multimap [unord.multimap]

 namespace std {
   template <ValueType Key,
             ValueType T,
             Callable<auto, const Key&> Hash = hash<Key>,
             PredicateEquivalenceRelation<auto, Key, Key> Pred = equal_to<Key>,
             Allocator Alloc = allocator<pair<const Key, T> > >
     requires NothrowDestructible<Key> && NothrowDestructible<T>
           && SameType<Hash::result_type, size_t>
           && CopyConstructible<Hash> && CopyConstructible<Pred>
           && AllocatableElement<Alloc, Pred, const Pred&>
           && AllocatableElement<Alloc, Pred, Pred&&>
           && AllocatableElement<Alloc, Hash, const Hash&>
           && AllocatableElement<Alloc, Hash, Hash&&>
   class unordered_multimap
   {
     ...
   };
 }

23.5.3p3 Class template unordered_set [unord.set]

 namespace std {
   template <ValueType Value,
             Callable<auto, const Value&> Hash = hash<Value>,
             PredicateEquivalenceRelation<auto, Value, Value> class Pred = equal_to<Value>,
             Allocator Alloc = allocator<Value> >
     requires NothrowDestructible<Value>
           && SameType<Hash::result_type, size_t>
           && CopyConstructible<Hash> && CopyConstructible<Pred>
           && AllocatableElement<Alloc, Pred, const Pred&>
           && AllocatableElement<Alloc, Pred, Pred&&>
           && AllocatableElement<Alloc, Hash, const Hash&>
           && AllocatableElement<Alloc, Hash, Hash&&>
   class unordered_set
   {
     ...
   };
 }

23.5.4p3 Class template unordered_multiset [unord.multiset]

 namespace std {
   template <ValueType Value,
             Callable<auto, const Value&> Hash = hash<Value>,
             PredicateEquivalenceRelation<auto, Value, Value> class Pred = equal_to<Value>,
             Allocator Alloc = allocator<Value> >
     requires NothrowDestructible<Value>
           && SameType<Hash::result_type, size_t>
           && CopyConstructible<Hash> && CopyConstructible<Pred>
           && AllocatableElement<Alloc, Pred, const Pred&>
           && AllocatableElement<Alloc, Pred, Pred&&>
           && AllocatableElement<Alloc, Hash, const Hash&>
           && AllocatableElement<Alloc, Hash, Hash&&>
   class unordered_multiset
   {
     ...
   };
 }

1112(i). bitsets and new style for loop

Section: 22.9.2 [template.bitset] Status: NAD Submitter: Alisdair Meredith Opened: 2009-05-06 Last modified: 2018-06-23

Priority: Not Prioritized

View all other issues in [template.bitset].

View all issues with NAD status.

Discussion:

std::bitset is a homogeneous container-like sequence of bits, yet it does not model the Range concept so cannot be used with the new for-loop syntax. It is the only such type in the library that does NOT support the new for loop.

The obvious reason is that bitset does not support iterators.

At least two reasonable solutions are available:

  1. Add an iterator interface to bitset, bringing its interface close to that of std::array
  2. Provide an unspecified concept_map for Range<bitset>.

The latter will still need some kind of iterator-like adapter for bitset, but gives implementers greater freedom on the details. E.g. begin/end return some type that simply invokes operator[] on the object it wraps, and increments its index on operator++. A vendor can settle for InputIterator support, rather than wrapping up a full RandomAccessIterator.

I have a mild preference for option (ii) as I think it is less work to specify at this stage of the process, although (i) is probably more useful in the long run.

Hmm, my wording looks a little woolly, as it does not say what the element type of the range is. Do I get a range of bool, bitset<N>::reference, or something else entirely?

I guess most users will assume the behaviour of reference, but expect to work with bool. Bool is OK for read-only traversal, but you really need to take a reference to a bitset::reference if you want to write back.

[ Batavia (2009-05): ]

Move to Open. We further recommend this be deferred until after the next Committee Draft.

[ 2009-05-25 Alisdair adds: ]

I just stumbled over the Range concept_map for valarray and this should probably set the precedent on how to write the wording.

[ Howard: I've replaced the proposed wording with Alisdair's suggestion. ]

[ 2009-07-24 Daniel modifies the proposed wording for non-concepts. ]

[ 2009-10 post-Santa Cruz: ]

Mark as Tentatively NAD Future due to the loss of concepts.

[2017-02 in Kona, LEWG recommends NAD]

There are better APIs for bits coming, and seems to be consensus in LEWG not to polish bitset any further.

[2017-06-02 Issues Telecon]

Resolve as NAD

Rationale:

All concepts-related text has been removed from the draft.

Proposed resolution:

  1. Modify the section 22.9.2 [template.bitset] <bitset> synopsis by adding the following at the end of the synopsis:

    
    // XX.X.X bitset range access [bitset.range]
    template<size_t N> unspecified-1 begin(bitset<N>&);
    template<size_t N> unspecified-2 begin(const bitset<N>&);
    template<size_t N> unspecified-1 end(bitset<N>&);
    template<size_t N> unspecified-2 end(const bitset<N>&);
    
    
  2. Add a new section "bitset range access" [bitset.range] after the current section 22.9.4 [bitset.operators] with the following series of paragraphs:

    1. In the begin and end function templates that follow, unspecified-1 is a type that meets the requirements of a mutable random access iterator (25.3.5.7 [random.access.iterators]) whose value_type is bool and whose reference type is bitset<N>::reference. unspecified-2 is a type that meets the requirements of a constant random access iterator (25.3.5.7 [random.access.iterators]) whose value_type is bool and whose reference type is bool.

    
    template<size_t N> unspecified-1 begin(bitset<N>&);
    template<size_t N> unspecified-2 begin(const bitset<N>&);
    
    

    2. Returns: an iterator referencing the first bit in the bitset.

    
    template<size_t N> unspecified-1 end(bitset<N>&);
    template<size_t N> unspecified-2 end(const bitset<N>&);
    

    3. Returns: an iterator referencing one past the last bit in the bitset.


1115(i). va_copy missing from Standard macros table

Section: C.7 [diff.library] Status: NAD Editorial Submitter: Miles Zhao Opened: 2009-05-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [diff.library].

View all issues with NAD Editorial status.

Discussion:

In "Table 122 — Standard macros" of C.7 [diff.library], which lists the 56 macros inherited from C library, va_copy seems to be missing. But in "Table 21 — Header <cstdarg> synopsis" (17.13 [support.runtime]), there is.

[ 2009-10 post-Santa Cruz: ]

Mark as Tentatively NAD Editorial, if Pete disagrees, Howard will move to Tentatively Ready

Proposed resolution:

Add va_copy to Table 122 -- Standard macros in C.7 [diff.library].


1119(i). tuple query APIs do not support references

Section: 22.4.7 [tuple.helper] Status: NAD Submitter: Alisdair Meredith Opened: 2009-05-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [tuple.helper].

View all issues with NAD status.

Discussion:

The tuple query APIs tuple_size and tuple_element do not support references-to-tuples. This can be annoying when a template deduced a parameter type to be a reference, which must be explicitly stripped with remove_reference before calling these APIs.

I am not proposing a resolution at this point, as there is a combinatorial explosion with lvalue/rvalue references and cv-qualification (see previous issue) that suggests some higher refactoring is in order. This might be something to kick back over to Core/Evolution.

Note that we have the same problem in numeric_limits.

[ 2009-10 post-Santa Cruz: ]

Move to Open. Alisdair to provide wording.

[ 2010 Rapperswil: ]

Move to NAD. This is an extension after the FCD, without a clear motivation. May consider as NAD Future if motivating examples come forward.

Proposed resolution:


1120(i). New type trait - remove_all

Section: 21 [meta] Status: NAD Submitter: Alisdair Meredith Opened: 2009-05-23 Last modified: 2020-09-06

Priority: Not Prioritized

View other active issues in [meta].

View all other issues in [meta].

View all issues with NAD status.

Discussion:

Sometimes it is necessary to remove all qualifiers from a type before passing on to a further API. A good example would be calling the tuple query APIs tuple_size or tuple_element with a deduced type inside a function template. If the deduced type is cv-qualified or a reference then the call will fail. The solution is to chain calls to remove_cv<remove_reference<T>::type>::type, and note that the order matters.

Suggest it would be helpful to add a new type trait, remove_all, that removes all top-level qualifiers from a type i.e. cv-qualification and any references. Define the term in such a way that if additional qualifiers are added to the language, then remove_all is defined as stripping those as well.

[ 2009-10-14 Daniel adds: ]

remove_all seems too generic, a possible alternative matching the current naming style could be remove_cv_reference or remove_reference_cv. It should also be considered whether this trait should also remove 'extents', or pointer 'decorations'. Especially if the latter situations are considered as well, it might be easier to chose the name not in terms of what it removes (which might be a lot), but in terms of it creates. In this case I could think of e.g. extract_value_type.

[ 2009-10 Santa Cruz: ]

NAD Future.

[2017-02 in Kona, LEWG recommends NAD]

Send a paper to reopen: we'll need to bikeshed names if we decide to pursue such a design. It'll need to discuss difference between this and decay, and make choices about array types, and discuss why references are removed but not pointers.

[2017-06-02 Issues Telecon]

Resolve as NAD; note that P0550 addresses this

Proposed resolution:


1121(i). Support for multiple arguments

Section: 21.4.4 [ratio.arithmetic] Status: NAD Submitter: Alisdair Meredith Opened: 2009-05-25 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [ratio.arithmetic].

View all issues with NAD status.

Discussion:

Both add and multiply could sensibly be called with more than two arguments. The variadic template facility makes such declarations simple, and is likely to be frequently wrapped by end users if we do not supply the variant ourselves.

We deliberately ignore divide at this point as it is not transitive. Likewise, subtract places special meaning on the first argument so I do not suggest extending that immediately. Both could be supported with analogous wording to that for add/multiply below.

Note that the proposed resolution is potentially incompatible with that proposed for 921, although the addition of the typedef to ratio would be equally useful.

[ 2009-10-30 Alisdair adds: ]

The consensus of the group when we reviewed this in Santa Cruz was that 921 would proceed to Ready as planned, and the multi-paramater add/multiply templates should be renamed as ratio_sum and ratio_product to avoid the problem mixing template aliases with partial specializations.

It was also suggested to close this issue as NAD Future as it does not correspond directly to any NB comment. NBs are free to submit a specific comment (and re-open) in CD2 though.

Walter Brown also had concerns on better directing the order of evaluation to avoid overflows if we do proceed for 0x rather than TR1, so wording may not be complete yet.

[ Alisdair updates wording. ]

[ 2009-10-30 Howard: ]

Moved to Tentatively NAD Future after 5 positive votes on c++std-lib.

[LEWG Kona 2017]

PR for ratio_product is wrong, uses ratio_add instead of ratio_multiply. Recommend NAD: Doesn't meet the bar for standardization: hasn't been requested again in 7 years, easy to implement yourself.

Rationale:

Does not have sufficient support at this time. May wish to reconsider for a future standard.

Proposed resolution:

Add the following type traits to p3 21.4 [ratio]

// ratio arithmetic
template <class R1, class R2> struct ratio_add;
template <class R1, class R2> struct ratio_subtract;
template <class R1, class R2> struct ratio_multiply;
template <class R1, class R2> struct ratio_divide;
template <class R1, class ... RList> struct ratio_sum;
template <class R1, class ... RList> struct ratio_product;

after 21.4.4 [ratio.arithmetic] p1: add

template <class R1, class ... RList> struct ratio_sum; // declared, never defined

template <class R1> struct ratio_sum<R1> : R1 {};

Requires: R1 is a specialization of class template ratio

template <class R1, class R2, class ... RList> 
 struct ratio_sum<R1, R2, RList...>
   : ratio_add< R1, ratio_sum<R2, RList...>> {
};

Requires: R1 and each element in parmater pack RList is a specialization of class template ratio

after 21.4.4 [ratio.arithmetic] p3: add

template <class R1, class ... RList> struct ratio_product; // declared, never defined

template <class R1> struct ratio_product<R1> : R1 {};

Requires: R1 is a specialization of class template ratio

template <class R1, class R2, class ... RList> 
 struct ratio_sum<R1, R2, RList...>
   : ratio_add< R1, ratio_product<R2, RList...>> {
};

Requires: R1 and each element in parmater pack RList is a specialization of class template ratio


1124(i). Invalid definition of concept RvalueOf

Section: 99 [concept.transform] Status: NAD Concepts Submitter: Daniel Krügler Opened: 2009-05-28 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [concept.transform].

View all issues with NAD Concepts status.

Discussion:

A recent news group article points to several defects in the specification of reference-related concepts.

One problem of the concept RvalueOf as currently defined in 99 [concept.transform]:

concept RvalueOf<typename T> {
 typename type = T&&;
 requires ExplicitlyConvertible<T&,type> && Convertible<T&&,type>;
}

template<typename T> concept_map RvalueOf<T&> {
 typedef T&& type;
}

is that if T is an lvalue-reference, the requirement Convertible<T&&,type> isn't satisfied for lvalue-references, because after reference-collapsing in the concept definition we have Convertible<T&,type> in this case, which isn't satisfied in the concept map template and also is not the right constraint either. I think that the reporter is right that SameType requirements should do the job and that we also should use the new RvalueReference concept to specify a best matching type requirement.

Proposed resolution:

In 99 [concept.transform] before p. 4 change as indicated:

auto concept RvalueOf<typename T> {
  typenameRvalueReference type = T&&;
  requires ExplicitlyConvertible<T&, type> && Convertible<T&&, type>SameType<T&, type&>;
}

1125(i). ostream_iterator does not work with movable types

Section: 25.6.3.3 [ostream.iterator.ops] Status: NAD Submitter: Alisdair Meredith Opened: 2009-05-28 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

ostream_iterator has not been updated to support moveable types, in a similar manner to the insert iterators. Note that this is not a problem for ostreambuf_iterator, as the types it is restricted to dealing with do not support extra-efficient moving.

[ 2009-11-10 Howard adds: ]

Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below.

Proposed resolution:

Add second operator= overload to class template ostream_iterator in 25.6.3 [ostream.iterator], para 2:

ostream_iterator<T,charT,traits>& operator=(const T& value);
ostream_iterator<T,charT,traits>& operator=(T&& value);

Add a new paragraph: in 25.6.3.3 [ostream.iterator.ops]:

ostream_iterator& operator=(T&& value);

-2- Effects:

*out_stream << std::move(value);
if(delim != 0)
  *out_stream << delim;
return (*this);

Rationale:

Several objections to move forward with this issue were voiced in the thread starting with c++std-lib-25438. Among them is that we know of no motivating use case to make streaming rvalues behave differently than streaming const lvalues.


1127(i). rvalue references and iterator traits

Section: 25.3.2.3 [iterator.traits] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-05-28 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iterator.traits].

View all issues with NAD Concepts status.

Discussion:

The deprecated support for iterator_traits and legacy (unconstrained) iterators features the (exposition only) concept:

concept IsReference<typename T> { } // exposition only
template<typename T> concept_map IsReference<T&> { }

Now this looks exactly like the LvalueReference concept recently added to clause 20, so I wonder if we should use that instead? Then I consider the lack of rvalue-reference support, which means that move_iterator would always flag as merely supporting the input_iterator_tag category. This suggests we retain the exposition concept, but add a second concept_map to support rvalue references.

I would suggest adding the extra concept_map is the right way forward, but still wonder if the two exposition-only concepts in this clause might be worth promoting to clause 20. That question might better be answered with a fuller investigation of type_trait/concept unification though.

Proposed resolution:

In Iterator traits 25.3.2.3 [iterator.traits] para 4 add:

concept IsReference<typename T> { } // exposition only
template<typename T> concept_map IsReference<T&> { }
template<typename T> concept_map IsReference<T&&> { }

1128(i). Missing definition of iterator_traits<T*>

Section: 25.2 [iterator.synopsis] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-05-28 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [iterator.synopsis].

View all issues with NAD Concepts status.

Discussion:

The <iterator> header synopsis declares a partial specialization of iterator_traits to support pointers, [iterator.syn]. The implication is that specialization will be described in D10, yet it did not follow the rest of the deprecated material into this clause.

However, this is not as bad as it first seems! There are partial specializations of iterator_traits for types that satisfy the various Iterator concepts, and there are concept_maps for pointers to explicitly support the RandomAccessIterator concept, so the required template will be present - just not in the manner advertised.

I can see two obvious solutions:

  1. Restore the iterator_traits<T*> partial specialization in D.10
  2. Remove the declaration of iterator_traits<T*> from 24.3 synopsis

I recommend option (ii) in the wording below

Option (ii) could be extended to strike all the declarations of deprecated material from the synopsis, as it is effectively duplicating D.10 anyway. This is the approach taken for deprecated library components in the 98/03 standards. This is probably a matter best left to the Editor though.

Proposed resolution:

In [iterator.syn] strike:

template<class T> struct iterator_traits<T*>;

1132(i). JP-30: nested exceptions

Section: 17.9.8 [except.nested] Status: NAD Submitter: Seiji Hayashida Opened: 2009-06-01 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [except.nested].

View all issues with NAD status.

Discussion:

Addresses JP 30

C++0x nested_exception cannot handle a structured exception well. The following codes show two types of tree structured exception handling.

The first one is based on nested_exception in C++0x, while the second one is based on my library trickerr.h (in Japanese). http://tricklib.com/cxx/dagger/trickerr.h

Assume that Function A() calls two sub functions A_a() and A_b(), both might throw tree structured exceptions, and A_b() must be called even if A_a() throws an exception.

List A (code of tree structured exception handling based on nested_exception in C++0x)

void A()
{
    try
    {
        std::vector<exception_ptr> exception_list;
        try
        {
            // A_a() does a similar processing as A().
            A_a();
        }
        catch(...)
        {
            exception_list.push_back(current_exception());
        }

        // ***The processing A() has to do even when A_a() fails. ***
        try
        {
            // A_b() does a similar processing as A().
            A_b();
        }
        catch(...)
        {
            exception_list.push_back(current_exception());
        }
        if (!exception_list.empty())
        {
            throw exception_list;
        }
    }
    catch(...)
    {
        throw_with_nested(A_exception("someone error"));
    }
}
void print_tree_exception(exception_ptr e, const std::string & indent ="")
{
    const char * indent_unit = " ";
    const char * mark = "- ";
    try
    {
        rethow_exception(e);
    }
    catch(const std::vector<exception_ptr> e)
    {
        for(std::vector<exception_ptr>::const_iterator i = e.begin(); i!=e.end(); ++i)
        {
            print_tree_exception(i, indent);
        }
    }
    catch(const std::nested_exception  e)
    {
        print_tree_exception(evil_i(e), indent +indent_unit);
    }
    catch(const std::exception e)
    {
        std::cout << indent << mark << e.what() << std::endl;
    }
    catch(...)
    {
        std::cout << indent << mark << "unknown exception" << std::endl;
    }
}
int main(int, char * [])
{
    try
    {
        A();
    }
    catch()
    {
        print_tree_exception(current_exception());
    }
    return EXIT_SUCCESS;
}

List B ( code of tree structured exception handling based on trickerr.h. ) "trickerr.h" (in Japanese), refer to: http://tricklib.com/cxx/dagger/trickerr.h.

void A()
{
    tricklib::error_listener_type error_listener;
    // A_a() is like A(). A_a() can throw tree structured exception.
    A_a();

    // *** It must do process so that A_a() throws exception in A(). ***
    // A_b() is like A(). A_b() can throw tree structured exception.
    A_b();

    if (error_listener.has_error()) // You can write this "if block" in destructor
                                    //  of class derived from error_listener_type.
    {
        throw_error(new A_error("someone error",error_listener.listener_off().extract_pending_error()));
    }
}
void print_tree_error(const tricklib::error_type &a_error, const std::string & indent = "")
{
    const char * indent_unit = " ";
    const char * mark = "- ";

    tricklib::error_type error = a_error;
    while(error)
    {
        std::cout << indent << mark << error->message << std::endl;
        if (error->children)
        {
            print_tree_error(error->children, indent +indent_unit);
        }
        error = error->next;
    }
}
int main(int, char * [])
{
    tricklib::error_thread_power error_thread_power_on; // This object is necessary per thread.

    try
    {
        A();
    }
    catch(error_type error)
    {
        print_tree_error(error);
    }
    catch(...)
    {
        std::cout << "- unknown exception" << std::endl;
    }
    return EXIT_SUCCESS;
}

Prospect

We will focus on the method A() since the other methods, also main(), occur only once respectively.

According to the above observation, we cannot help concluding that it is not so easy to use the nested_exception handling as a tree structured exception handling mechanism in a practical sense.

This text is based on the web page below (in Japanese). http://d.hatena.ne.jp/wraith13/20081231/1230715424

[ 2009-10 Santa Cruz: ]

Mark as NAD. The committee agrees that nested_exception is not a good match for this usage model. The committee did not see a way of improving this within the timeframe allowed.

Proposed resolution:


1139(i). Thread support library not concept enabled

Section: 33 [thread] Status: NAD Concepts Submitter: LWG Opened: 2009-06-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread].

View all issues with NAD Concepts status.

Discussion:

Addresses US 93, JP 79, UK 333, JP 81

The thread chapter is not concept enabled.

Proposed resolution:


1140(i). Numerics library not concept enabled

Section: 28 [numerics] Status: NAD Concepts Submitter: LWG Opened: 2009-06-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [numerics].

View all issues with NAD Concepts status.

Discussion:

Addresses US 84

The numerics chapter is not concept enabled.

The portion of this comment dealing with random numbers was resolved by N2836, which was accepted in Summit.

Proposed resolution:


1141(i). Input/Output library not concept enabled

Section: 31 [input.output] Status: NAD Concepts Submitter: LWG Opened: 2009-06-15 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [input.output].

View all issues with NAD Concepts status.

Discussion:

Addresses US 85, JP 67, JP 68, JP 69, JP 72, UK 308

The input/output chapter is not concept enabled.

Proposed resolution:


1142(i). Regular expressions library not concept enabled

Section: 32 [re] Status: NAD Concepts Submitter: LWG Opened: 2009-06-15 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [re].

View all other issues in [re].

View all issues with NAD Concepts status.

Discussion:

Addresses US 86, UK 309, UK 310

The regular expressions chapter is not concept enabled.

Proposed resolution:


1148(i). Wrong argument type of I/O stream manipulators setprecision() and setw()

Section: 31.7 [iostream.format] Status: NAD Submitter: Marc Steinbach Opened: 2009-06-20 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iostream.format].

View all issues with NAD status.

Discussion:

The header <iomanip> synopsis in 31.7 [iostream.format] specifies

T5 setprecision(int n);
T6 setw(int n);

The argument types should be streamsize, as in class ios_base (see 31.5.2 [ios.base]):

streamsize precision() const;
streamsize precision(streamsize prec);
streamsize width() const;
streamsize width(streamsize wide);

(Editorial: 'wide' should probably be renamed as 'width', or maybe just 'w'.)

[ 2009-07-29 Daniel clarified wording. ]

[ 2009 Santa Cruz: ]

No concensus for this change. There was some interest in doing the opposite fix: Change the streamsize in <ios> to int. But ultimately there was no concensus for that change either.

Proposed resolution:

  1. In 31.7 [iostream.format], header <iomanip> synopsis change as indicated:

    T5 setprecision(intstreamsize n);
    T6 setw(intstreamsize n);
    
  2. In 31.7.7 [std.manip], just before p. 6 change as indicated:

    unspecified setprecision(intstreamsize n);
    
  3. In 31.7.7 [std.manip], just before p. 7 change as indicated:

    unspecified setw(intstreamsize n);
    

1149(i). Reformulating NonemptyRange axiom

Section: 99 [rand.concept.urng] Status: NAD Concepts Submitter: Walter Brown Opened: 2009-06-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Concepts status.

Discussion:

In 99 [rand.concept.urng], we have the following:

concept UniformRandomNumberGenerator<typename G> : Callable<G> {
  ...
  axiom NonemptyRange(G& g) {
    G::min() < G::max();
  }
  ...
}

Since the parameter G is in scope throughout the concept, there is no need for the axiom to be further parameterized, and so the axiom can be slightly simplified as:

axiom NonemptyRange()  {
  G::min() < G::max();
}

We can further reformulate so as to avoid any axiom machinery as:

requires True< G::min() < G::max() >;

This is not only a simpler statement of the same requirement, but also forces the requirement to be checked.

[ Post-Rapperswil: ]

Moved to Tentatively Ready after 5 positive votes on c++std-lib.

Proposed resolution:

In 99 [rand.concept.urng], replace the NonemptyRange axiom by:

axiom NonemptyRange(G& g) { 
   G::min() < G::max(); 
}
requires True< G::min() < G::max() >;

1153(i). Standard library needs review for constructors to be explicit to avoid treatment as initializer-list constructor

Section: 16 [library], 33 [thread], D [depr] Status: NAD Submitter: LWG Opened: 2009-06-28 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Addresses DE 2

Description

Marking a constructor with explicit has semantics even for a constructor with zero or several parameters: Such a constructor cannot be used with list-initialization in a copy-initialization context, see 12.2.2.8 [over.match.list]. The standard library apparently has not been reviewed for marking non-single-parameter constructors as explicit.

Suggestion

Consider marking zero-parameter and multi-parameter constructors explicit in classes that have at least one constructor marked explicit and that do not have an initializer-list constructor.

Notes

Robert Klarer to address this one.

[ 2009 Santa Cruz: ]

Move to "Open". Robert Klarer has promised to provide wording.

[ 2010 Pittsburgh: Moved to NAD, rationale added below. ]

Rationale:

We are unaware of any cases where initializer lists cause problem in this context, but if problems arise in the future the issue can be reopened.

Proposed resolution:


1154(i). complex should accept integral types

Section: 28.4 [complex.numbers] Status: NAD Submitter: LWG Opened: 2009-06-28 Last modified: 2020-09-06

Priority: Not Prioritized

View all other issues in [complex.numbers].

View all issues with NAD status.

Discussion:

Addresses FR 35

Description

Instantiations of the class template complex<> have to be allowed for integral types, to reflect existing practice and ISO standards (LIA-III).

Suggestion

[ 2009-10-26 Proposed wording in N3002. ]

[ 2010 Pittsburgh: ]

Moved to NAD Future. Rationale added.

[LEWG Kona 2017]

Recommend SG6

[2017-03-03, Kona]

SG6 suggests this issue is a new feature, not a problem with the existing standard, and should therefore be closed NAD. However, SG6 invites papers that bring the proposal up to date with the current standard.

Rationale:

There is no consensus for making this change at this time.

Proposed resolution:

Adopt N3002.


1155(i). Reference should be to C99

Section: C.7 [diff.library] Status: NAD Editorial Submitter: LWG Opened: 2009-06-28 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [diff.library].

View all issues with NAD Editorial status.

Discussion:

Addresses FR 38

Description

What is ISO/IEC 1990:9899/DAM 1? My guess is that's a typo for ISO/IEC 9899/Amd.1:1995 which I'd have expected to be referenced here (the tables make reference to things which were introduced by Amd.1).

Suggestion

One need probably a reference to the document which introduce char16_t and char32_t in C (ISO/IEC TR 19769:2004?).

Notes

Create issue. Document in question should be C99, not C90+amendment1. The rest of the section requires careful review for completeness. Example <cstdint> 17.4.1 [cstdint.syn]. Assign to C liasons.

[ 2009-10 Santa Cruz: ]

NAD Editorial. Already fixed.

Proposed resolution:


1156(i). Constraints on bitmask and enumeration types to be tightened

Section: 16.3.3.3.2 [enumerated.types], 16.3.3.3.3 [bitmask.types] Status: NAD Submitter: LWG Opened: 2009-06-28 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses UK 165

Description

Constraints on bitmask and enumeration types were supposed to be tightened up as part of the motivation for the constexpr feature - see paper N2235 for details

Suggestion

Adopt wording in line with the motivation described in N2235

Notes

Robert Klarer to review

[ 2009 Santa Cruz: ]

Move to Open. Ping Robert Klarer to provide wording, using N2235 as guidance.

[ 2010 Pittsburgh: ]

Moved to NAD. Rationale added.

Rationale:

UK NB did not sufficiently describe how to resolve their comment, and therefore we cannot make a change for the FCD. If a resolution were provided in the future, we would be happy to apply it.

Proposed resolution:


1164(i). promise::swap should pass by rvalue reference

Section: 33.10.6 [futures.promise] Status: NAD Submitter: LWG Opened: 2009-06-28 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [futures.promise].

View all other issues in [futures.promise].

View all issues with NAD status.

Discussion:

Addresses UK 341

Description

promise::swap accepts its parameter by lvalue reference. This is inconsistent with other types that provide a swap member function, where those swap functions accept an rvalue reference

Suggestion

Change promise::swap to take an rvalue reference.

Notes

Create an issue. Detlef will look into it. Probably ready as it.

[ 2009-07 Frankfurt ]

NAD, by virtue of the changed rvalue rules and swap signatures from Summit.

Proposed resolution:


1167(i). pair<T,U> doesn't model LessThanComparable in unconstrained code even if T and U do.

Section: 22.3 [pairs] Status: NAD Concepts Submitter: Dave Abrahams Opened: 2009-07-01 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [pairs].

View all issues with NAD Concepts status.

Discussion:

LessThanComparable requires (and provides default implementations for) <=,>, and >=. However, the defaults don't take effect in unconstrained code.

Still, it's a problem to have types acting one way in constrained code and another in unconstrained code, except in cases of syntax adaptation. It's also inconsistent with the containers, which supply all those operators.

Totally Unbiased Suggested Resolution:

accept the exported concept maps proposal and change the way this stuff is handled to use an explicit exported concept map rather than nested function templates

e.g., remove from the body of std::list

template <LessThanComparable T, class Allocator> 
bool operator< (const list<T,Allocator>& x, const list<T,Allocator>& y); 
template <LessThanComparable T, class Allocator> 
bool operator> (const list<T,Allocator>& x, const list<T,Allocator>& y); 
template <LessThanComparable T, class Allocator> 
bool operator>=(const list<T,Allocator>& x, const list<T,Allocator>& y); 
template <LessThanComparable T, class Allocator> 
bool operator<=(const list<T,Allocator>& x, const list<T,Allocator>& y); 

and add this concept_map afterwards:

template <LessThanComparable T, class Allocator> 
export concept_map LessThanComparable<list<T,Allocator> >
{
    bool operator<(const list<T,Allocator>& x, const list<T,Allocator>& y);
}

do similarly for std::pair. While you're at it, do the same for operator== and != everywhere, and seek out other such opportunities.

Alternative Resolution: keep the ugly, complex specification and add the missing operators to std::pair.

Proposed resolution:


1168(i). Odd wording for bitset equality operators

Section: 22.9.2.3 [bitset.members] Status: NAD Editorial Submitter: Alisdair Meredith Opened: 2009-07-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [bitset.members].

View all issues with NAD Editorial status.

Discussion:

The following wording seems a little unusual to me:

p42/43 22.9.2.3 [bitset.members]

bool operator==(const bitset<N>& rhs) const;

-42- Returns: A nonzero value if the value of each bit in *this equals the value of the corresponding bit in rhs.

bool operator!=(const bitset<N>& rhs) const;

-43- Returns: A nonzero value if !(*this == rhs).

"A nonzero value" may be well defined as equivalent to the literal 'true' for Booleans, but the wording is clumsy. I suggest replacing "A nonzero value" with the literal 'true' (in appropriate font) in each case.

[ 2009-07-24 Alisdair recommends NAD Editorial. ]

[ 2009-07-27 Pete adds: ]

It's obviously editorial. There's no need for further discussion.

[ 2009-07-27 Howard sets to NAD Editorial. ]

Proposed resolution:

Change 22.9.2.3 [bitset.members] p42-43:

bool operator==(const bitset<N>& rhs) const;

-42- Returns: A nonzero value true if the value of each bit in *this equals the value of the corresponding bit in rhs.

bool operator!=(const bitset<N>& rhs) const;

-43- Returns: A nonzero value true if !(*this == rhs).


1173(i). "Equivalence" wishy-washiness

Section: 16 [library] Status: NAD Submitter: David Abrahams Opened: 2009-07-14 Last modified: 2016-11-21

Priority: 3

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Issue: The CopyConstructible requirements are wishy-washy. It requires that the copy is "equivalent" to the original, but "equivalent" is never defined.

I believe this to be an example of a more general lack of rigor around copy and assignment, although I haven't done the research to dig up all the instances.

It's a problem because if you don't know what CopyConstructible means, you also don't know what it means to copy a pair of CopyConstructible types. It doesn't prevent us from writing code, but it is a hole in our ability to understand the meaning of copy.

Furthermore, I'm pretty sure that vector's copy constructor doesn't require the elements to be EqualityComparable, so that table is actually referring to some ill-defined notion of equivalence when it uses ==.

[ 2009 Santa Cruz: ]

Move to "Open". Dave is right that this is a big issue. Paper D2987 ("Defining Move Special Member Functions", Bjarne Stroustrup and Lawrence Crowl) touches on this but does not solve it. This issue is discussed in Elements of Programming.

[ 2010 Rapperswil: ]

This issue is quite vague, so it is difficult to know if and when it has been resolved. John Lakos wrote a paper covering this area a while back, and there is a real interest in providing some sort of clean-up in the future. We need a more clearly draughted issues with an addressable set of concerns, ideally with a paper proposing a resolution, but for a future revision of the standard. Move to Tentatively NAD Future.

[ Moved to NAD Future at 2010-11 Batavia ]

[2016-11-12, Issaquah]

Sat PM: Status: NAD

It was said in 2010: "We need a more clearly draughted issue with an addressable set of concerns, ideally with a paper proposing a resolution, but for a future revision of the standard."

Proposed resolution:


1176(i). Make thread constructor non-variadic

Section: 33.4.3.3 [thread.thread.constr] Status: NAD Submitter: Howard Hinnant Opened: 2009-07-18 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.thread.constr].

View all issues with NAD status.

Discussion:

The variadic thread constructor is causing controversy, e.g. N2901. This issue has been created as a placeholder for this course of action.

template <class F, class ...Args> thread(F&& f, Args&&... args);

See 929 for wording which specifies an rvalue-ref signature but with "decay behavior", but using variadics.

[ 2009-11-17 Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Proposed resolution:

Rationale:

The (tentative) concensus of the LWG is to keep the variadic thread constructor.


1179(i). Probably editorial in [structure.specifications]

Section: 16.3.2.4 [structure.specifications] Status: NAD Editorial Submitter: Robert Klarer Opened: 2009-07-21 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [structure.specifications].

View all other issues in [structure.specifications].

View all issues with NAD Editorial status.

Discussion:

While reviewing 971 I noted that 16.3.2.4 [structure.specifications]/7 says:

-7- Error conditions specify conditions where a function may fail. The conditions are listed, together with a suitable explanation, as the enum class errc constants (19.5) that could be used as an argument to function make_error_condition (19.5.3.6).

This paragraph should mention make_error_code or the text "that could be used as an argument to function make_error_condition (19.5.3.6)" should be deleted. I believe this is editorial.

[ 2009-07-21 Chris adds: ]

I'm not convinced there's a problem there, because as far as the "Error conditions" clauses are concerned, make_error_condition() is used by a user to test for the condition, whereas make_error_code is not. For example:

void foobar(error_code& ec = throws());

Error conditions:

permission_denied - Insufficient privilege to perform operation.

When a user writes:

error_code ec;
foobar(ec);
if (ec == errc::permission_denied)
   ...

the implicit conversion errc->error_condition makes the if-test equivalent to:

if (ec == make_error_condition(errc::permission_denied))

On the other hand, if the user had written:

if (ec == make_error_code(errc::permission_denied))

the test is now checking for a specific error code. The test may evaluate to false even though foobar() failed due to the documented error condition "Insufficient privilege".

[ 2009 Santa Cruz: ]

NAD Editorial.

What the WP says right now is literally true: these codes can be used as an argument to make_error_condition. (It is also true that they can be used as an argument to make_error_code, which the WP doesn't say.) Maybe it would be clearer to just delete "that could be used as an argument to function make_error_condition", since that fact is already implied by other things that we say. We believe that this is editorial.

Proposed resolution:


1184(i). Feature request: dynamic bitset

Section: 24.3.11 [vector] Status: NAD Submitter: Alisdair Meredith Opened: 2009-07-29 Last modified: 2020-09-06

Priority: Not Prioritized

View all other issues in [vector].

View all issues with NAD status.

Discussion:

Opened at Alisdair's request, steming from 96. Alisdair recommends NAD Future.

[ 2009-10 Santa Cruz: ]

NAD Future. We want a heap allocated bitset, but we don't have one today and don't have time to add one.

[2017-02 in Kona, LEWG recommends NAD]

Solution in process: Presented at Olu (Vincent Reverdy) P0237r5 Value reference pointer and iterator. The reference should be put into the issue.

[2017-06-02 Issues Telecon]

This is a feature request, not a defect.

Resolve as NAD

Proposed resolution:


1186(i). Forward list could model a stack

Section: 24.6.8 [stack] Status: NAD Concepts Submitter: Alisdair Meredith Opened: 2009-07-31 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [stack].

View all issues with NAD Concepts status.

Discussion:

The library template forward_list could easily model the idea of a stack, where the operations work on the front of the list rather than the back. However, the standard library stack adaptor cannot support this.

It would be relatively easy to write a partial specialization for stack to support forward_list, but that opens the question of which header to place it in. A much better solution would be to add a concept_map for the StackLikeContainer concept to the <forward_list> header and then everything just works, including a user's own further uses in a stack-like context.

Therefore while I am submitting the issue now so that it is on record, I strongly recommend we resolve as "NAD Concepts" as any non-concepts based solution will be inferior to the final goal, and the feature is not so compelling it must be supported ahead of the concepts-based library.

[ 2009-11-02 Howard adds: ]

Moved to Tentatively NAD Concepts after 5 positive votes on c++std-lib.

Rationale:

Any non-concepts based solution will be inferior to the final goal, and the feature is not so compelling it must be supported ahead of the concepts-based library.

Proposed resolution:


1188(i). Unordered containers should have a minimum load factor as well as a maximum

Section: 24.2.8 [unord.req], 24.5 [unord] Status: NAD Submitter: Matt Austern Opened: 2009-08-10 Last modified: 2019-02-26

Priority: Not Prioritized

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with NAD status.

Discussion:

Unordered associative containers have a notion of a maximum load factor: when the number of elements grows large enough, the containers automatically perform a rehash so that the number of elements per bucket stays below a user-specified bound. This ensures that the hash table's performance characteristics don't change dramatically as the size increases.

For similar reasons, Google has found it useful to specify a minimum load factor: when the number of elements shrinks by a large enough, the containers automatically perform a rehash so that the number of elements per bucket stays above a user-specified bound. This is useful for two reasons. First, it prevents wasting a lot of memory when an unordered associative container grows temporarily. Second, it prevents amortized iteration time from being arbitrarily large; consider the case of a hash table with a billion buckets and only one element. (This was discussed even before TR1 was published; it was TR issue 6.13, which the LWG closed as NAD on the grounds that it was a known design feature. However, the LWG did not consider the approach of a minimum load factor.)

The only interesting question is when shrinking is allowed. In principle the cleanest solution would be shrinking on erase, just as we grow on insert. However, that would be a usability problem; it would break a number of common idioms involving erase. Instead, Google's hash tables only shrink on insert and rehash.

The proposed resolution allows, but does not require, shrinking in rehash, mostly because a postcondition for rehash that involves the minimum load factor would be fairly complicated. (It would probably have to involve a number of special cases and it would probably have to mention yet another parameter, a minimum bucket count.)

The current behavior is equivalent to a minimum load factor of 0. If we specify that 0 is the default, this change will have no impact on backward compatibility.

[ 2010 Rapperswil: ]

This seems to a useful extension, but is too late for 0x. Move to Tentatively NAD Future.

[ Moved to NAD Future at 2010-11 Batavia ]

[LEWG Kona 2017]

Should there be a shrink_to_fit()? Is it too surprising to shrink on insert()? (We understand that shrinking on erase() is not an option.) Maybe make people call rehash(0) to shrink to the min_load_factor? On clear(), the load factor goes to 0 or undefined (0/0), which is likely to violate min_load_factor() min_load_factor(z)'s wording should match max_load_factor(z)'s, e.g. "May change the container’s maximum load factor" Want a paper exploring whether shrink-on-insert has been surprising. From Titus: Google's experience is that maps don't shrink in the way this would help with. NAD, not worth the time. Write a paper if you can demonstrate a need for this.

Proposed resolution:

Add two new rows, and change rehash's postcondition in the unordered associative container requirements table in 24.2.8 [unord.req]:

Table 87 — Unordered associative container requirements (in addition to container)
ExpressionReturn typeAssertion/note pre-/post-condition Complexity
a.min_load_factor() float Returns a non-negative number that the container attempts to keep the load factor greater than or equal to. The container automatically decreases the number of buckets as necessary to keep the load factor above this number. constant
a.min_load_factor(z) void Pre: z shall be non-negative. Changes the container's minimum load factor, using z as a hint. [Footnote: the minimum load factor should be significantly smaller than the maximum. If z is too large, the implementation may reduce it to a more sensible value.] constant
a.rehash(n) void Post: a.bucket_count() >= n, and a.size() <= a.bucket_count() * a.max_load_factor(). [Footnote: It is intentional that the postcondition does not mention the minimum load factor. This member function is primarily intended for cases where the user knows that the container's size will increase soon, in which case the container's load factor will temporarily fall below a.min_load_factor().] a.bucket_cout > a.size() / a.max_load_factor() and a.bucket_count() >= n. Average case linear in a.size(), worst case quadratic.

Add a footnote to 24.2.8 [unord.req] p12:

The insert members shall not affect the validity of references to container elements, but may invalidate all iterators to the container. The erase members shall invalidate only iterators and references to the erased elements.

[A consequence of these requirements is that while insert may change the number of buckets, erase may not. The number of buckets may be reduced on calls to insert or rehash.]

Change paragraph 13:

The insert members shall not affect the validity of iterators if (N+n) < z * B zmin * B <= (N+n) <= zmax * B, where N is the number of elements in the container prior to the insert operation, n is the number of elements inserted, B is the container's bucket count, zmin is the container's minimum load factor, and zmax is the container's maximum load factor.

Add to the unordered_map class synopsis in section 24.5.4 [unord.map], the unordered_multimap class synopsis in 24.5.5 [unord.multimap], the unordered_set class synopsis in 24.5.6 [unord.set], and the unordered_multiset class synopsis in 24.5.7 [unord.multiset]:


float min_load_factor() const;
void min_load_factor(float z);

In 24.5.4.2 [unord.map.cnstr], 24.5.5.2 [unord.multimap.cnstr], 24.5.6.2 [unord.set.cnstr], and 24.5.7.2 [unord.multiset.cnstr], change:

... max_load_factor() returns 1.0 and min_load_factor() returns 0.


1190(i). Setting the maximum load factor should return the previous value

Section: 24.2.8 [unord.req], 24.5 [unord] Status: NAD Submitter: Matt Austern Opened: 2009-08-10 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with NAD status.

Discussion:

The unordered associative container requirements table specifies that a.set_max_load_factor(z) has return type void. However, there is a useful piece of information to return: the previous value. Users who don't need it can always ignore it.

[ 2010 Rapperswil: ]

The benefit seems minor, while breaking with the getter/setter idiom these overloads support. Move to Tentatively NAD.

[ Moved to NAD at 2010-11 Batavia ]

Proposed resolution:

In the unordered associative container requirements table, change:

Table 87 — Unordered associative container requirements (in addition to container)
ExpressionReturn typeAssertion/note pre-/post-condition Complexity
a.max_load_factor(z) void float Pre: z shall be positive. Changes the container's maximum load load factor, using z as a hint. Returns: the previous value of a.max_load_factor(). constant

Change the return type of set_max_load_factor in the class synopses in 24.5.4 [unord.map], 24.5.5 [unord.multimap], 24.5.6 [unord.set], and 24.5.7 [unord.multiset].

If issue 1188 is also accepted, make the same changes for min_load_factor.


1200(i). "surprising" char_traits<T>::int_type requirements

Section: 23.2.3 [char.traits.typedefs] Status: NAD Submitter: Sean Hunt Opened: 2009-09-03 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [char.traits.typedefs].

View all issues with NAD status.

Discussion:

The footnote for int_type in 23.2.3 [char.traits.typedefs] says that

If eof() can be held in char_type then some iostreams implementations may give surprising results.

This implies that int_type should be a superset of char_type. However, the requirements for char16_t and char32_t define int_type to be equal to int_least16_t and int_least32_t respectively. int_least16_t is likely to be the same size as char_16_t, which may lead to surprising behavior, even if eof() is not a valid UTF-16 code unit. The standard should not prescribe surprising behavior, especially without saying what it is (it's apparently not undefined, just surprising). The same applies for 32-bit types.

I personally recommend that behavior be undefined if eof() is a member of char_type, and another type be chosen for int_type (my personal favorite has always been a struct {bool eof; char_type c;}). Alternatively, the exact results of such a situation should be defined, at least so far that I/O could be conducted on these types as long as the code units remain valid. Note that the argument that no one streams char16_t or char32_t is not really valid as it would be perfectly reasonable to use a basic_stringstream in conjunction with UTF character types.

[ 2009-10-28 Ganesh provides two possible resolutions and expresses a preference for the second: ]

  1. Replace [char.traits.specializations.char16_t] para 3 with:

    The member eof() shall return an implementation-defined constant that cannot appear as a valid UTF-16 code unit UINT_LEAST16_MAX [Note: this value is guaranteed to be a permanently reserved UCS-2 code position if UINT_LEAST16_MAX == 0xFFFF and it's not a UCS-2 code position otherwise — end note].

    Replace [char.traits.specializations.char32_t] para 3 with:

    The member eof() shall return an implementation-defined constant that cannot appear as a Unicode code point UINT_LEAST32_MAX [Note: this value is guaranteed to be a permanently reserved UCS-4 code position if UINT_LEAST32_MAX == 0xFFFFFFFF and it's not a UCS-4 code position otherwise — end note].

  2. In [char.traits.specializations.char16_t], in the definition of char_traits<char16_t> replace the definition of nested typedef int_type with:

    namespace std {
      template<> struct char_traits<char16_t> {
        typedef char16_t         char_type;
        typedef uint_least16_t uint_fast16_t int_type;
         ...
    

    Replace [char.traits.specializations.char16_t] para 3 with:

    The member eof() shall return an implementation-defined constant that cannot appear as a valid UTF-16 code unit UINT_FAST16_MAX [Note: this value is guaranteed to be a permanently reserved UCS-2 code position if UINT_FAST16_MAX == 0xFFFF and it's not a UCS-2 code position otherwise — end note].

    In [char.traits.specializations.char32_t], in the definition of char_traits<char32_t> replace the definition of nested typedef int_type with:

    namespace std {
      template<> struct char_traits<char32_t> {
        typedef char32_t         char_type;
        typedef uint_least32_t uint_fast32_t int_type;
         ...
    

    Replace [char.traits.specializations.char32_t] para 3 with:

    The member eof() shall return an implementation-defined constant that cannot appear as a Unicode code point UINT_FAST32_MAX [Note: this value is guaranteed to be a permanently reserved UCS-4 code position if UINT_FAST32_MAX == 0xFFFFFFFF and it's not a UCS-4 code position otherwise — end note].

[ 2010 Rapperswil: ]

This seems an overspecification, and it is not clear what problem is being solved - these values can be used portably by using the named functions; there is no need for the value itself to be portable. Move to Tentatively NAD.

[ Moved to NAD at 2010-11 Batavia ]

Proposed resolution:


1202(i). integral_constant needs a spring clean

Section: 21.3.4 [meta.help] Status: NAD Submitter: Alisdair Meredith Opened: 2009-09-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [meta.help].

View all issues with NAD status.

Discussion:

The specification of integral_constant has been inherited essentially unchanged from TR1:

template <class T, T v>
struct integral_constant {
  static const T value = v;
  typedef T value_type;
  typedef integral_constant<T,v> type;
};

In light of 0x language changes there are several things we might consider changing, notably the form of specification for value.

The current form requires a static data member have storage allocated for it, where we could now implement without this using the new enum syntax:

template <class T, T v>
struct integral_constant {
  enum : T { value = v };
  typedef T value_type;
  typedef integral_constant type;
};

The effective difference between these two implementation is:

  1. No requirement to allocate storage for data member (which we hope but do not guarantee compilers strip today)
  2. You can no longer take the address of the constant as &integral_constant<T,v>::value;

Also note the editorial change to drop the explicit qualification of integral_constant in the typedef type. This makes it quite clear we mean the current instantiation, and cannot be mistaken for a recursive metaprogram.

Even if we don't mandate this implementation, it would be nice to give vendors freedom under QoI to choose their preferred representation.

The other side of this issue is if we choose to retain the static constant form. In that case we should go further and insist on constexpr, much like we did throughout numeric_limits:

template <class T, T v>
struct integral_constant {
  static constexpr T value = v;
  typedef T value_type;
  typedef integral_constant type;
};

[Footnote] It turns out constexpr is part of the Tentatively Ready resolution for 1019. I don't want to interfere with that issue, but would like a new issue to consider if the fixed-base enum implementation should be allowed.

[ 2009-09-05 Daniel adds: ]

I think that the suggested resolution is incomplete and may have some possible unwanted side-effects. To understand why, note that integral_constant is completely specified by code in 21.3.4 [meta.help]. While this is usually considered as a good thing, let me give a possible user-defined specialization that would break given the suggested changes:

enum NodeColor { Red, Black };

std::integral_constant<NodeColor, Red> red;

The reason why that breaks is due to the fact that current core language rules does only allow integral types as enum-bases, see 9.7.1 [dcl.enum]/2.

So, I think that we cannot leave the implementation the freedom to decide which way they would like to provide the implementation, because that is easily user-visible (I don't speak of addresses, but of instantiation errors), therefore if applied, this should be either specified or wording must be added that gives a note about this freedom of implementation.

Another possible disadvantage seems to me that user-expectations are easy to disappoint if they see a failure of the test

assert(typeid(std::integral_constant<int, 0>::value) == typeid(int));

or of

static_assert(std::is_same<decltype(std::integral_constant<int, 0>::value), const int>::value, "Bad library");

[ 2010-01-14 Moved to Tentatively NAD after 5 positive votes on c++std-lib. ]

Rationale:

We think that the suggested resolution is incomplete and may have some possible unwanted side-effects. (see Daniel's 2009-09-05 comment for details).

Proposed resolution:


1217(i). Quaternion support

Section: 28.4 [complex.numbers] Status: NAD Submitter: Ted Shaneyfelt Opened: 2009-09-26 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [complex.numbers].

View all issues with NAD status.

Discussion:

Concerning mathematically proper operation of the type:

complex<complex<T> >

Generally accepted mathematical semantics of such a construct correspond to quaternions through Cayly-Dickson construct

(w+xi) + (y+zi) j

The proper implementation seems straightforward by adding a few declarations like those below. I have included operator definition for combining real scalars and complex types, as well, which seems appropriate, as algebra of complex numbers allows mixing complex and real numbers with operators. It also allows for constructs such as complex<double> i=(0,1), x = 12.34 + 5*i;

Quaternions are often used in areas such as computer graphics, where, for example, they avoid the problem of Gimbal lock when rotating objects in 3D space, and can be more efficient than matrix multiplications, although I am applying them to a different field.

/////////////////////////ALLOW OPERATORS TO COMBINE REAL SCALARS AND COMPLEX VALUES /////////////////////////
template<typename T,typename S> complex<T> operator+(const complex<T> x,const S a) {
    complex<T> result(x.real()+a, x.imag());
    return result;
}
template<typename T,typename S> complex<T> operator+(const S a,const complex<T> x) {
    complex<T> result(a+x.real(), x.imag());
    return result;
}
template<typename T,typename S> complex<T> operator-(const complex<T> x,const S a) {
    complex<T> result(x.real()-a, x.imag());
    return result;
}
template<typename T,typename S> complex<T> operator-(const S a,const complex<T> x) {
    complex<T> result(a-x.real(), x.imag());
    return result;
}
template<typename T,typename S> complex<T> operator*(const complex<T> x,const S a) {
    complex<T> result(x.real()*a, x.imag()*a);
    return result;
}
template<typename T,typename S> complex<T> operator*(const S a,const complex<T> x) {
    complex<T> result(a*x.real(), a*x.imag());
    return result;
}

/////////////////////////PROPERLY IMPLEMENT QUATERNION SEMANTICS/////////////////////////
template<typename T> double normSq(const complex<complex<T> >q) {
    return q.real().real()*q.real().real()
         + q.real().imag()*q.real().imag()
         + q.imag().real()*q.imag().real()
         + q.imag().imag()*q.imag().imag();
}
template<typename T> double norm(const complex<complex<T> >q) {
    return sqrt(normSq(q));
}
/////// Cayley-Dickson Construction
template<typename T> complex<complex<T> > conj(const complex<complex<T> > x) {
    complex<complex<T> > result(conj(x.real()),-x.imag());
    return result;
}
template<typename T> complex<complex<T> > operator*(const complex<complex<T> > ab,const complex<complex<T> > cd) {
    complex<T> re(ab.real()*cd.real()-conj(cd.imag())*ab.imag());
    complex<T> im(cd.imag()*ab.real()+ab.imag()*conj(cd.real()));
    complex<complex<double> > q(re,im);
    return q;
}
//// Quaternion division
template<typename S,typename T> complex<complex<T> > operator/(const complex<complex<T> > q,const S a) {
    return q * (1/a);
}
template<typename S,typename T> complex<complex<T> > operator/(const S a,const complex<complex<T> > q) {
    return a*conj(q)/normSq(q);
}
template<typename T> complex<complex<T> > operator/(const complex<complex<T> > n, const complex<complex<T> > d) {
    return n * (conj(d)/normSq(d));
}

[ 2009-10 Santa Cruz: ]

NAD Future. There is no consensus or time to move this into C++0X.

[LEWG Kona 2017]

Recommend SG6 - We note that complex<complex> is the wrong way to spell this

[2017-03-03, Kona]

SG6 suggests this issue is a new feature, not a problem with the existing standard, and should therefore be closed NAD. However, SG6 invites papers that bring the proposal up to date with the current standard.

Proposed resolution:


1219(i). unique_lock::lock and resource_deadlock_would_occur

Section: 33.6.5.4.3 [thread.lock.unique.locking] Status: Dup Submitter: Jeffrey Yasskin Opened: 2009-09-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.lock.unique.locking].

View all issues with Dup status.

Duplicate of: 1159

Discussion:

unique_lock::lock and friends raise "resource_deadlock_would_occur -- if the current thread already owns the mutex (i.e., on entry, owns is true)." 1) The current thread owning a mutex is not the same as any particular unique_lock::owns being true. 2) There's no need to raise this exception for a recursive_mutex if owns is false. 3) If owns is true, we need to raise some exception or the unique_lock will lose track of whether to unlock itself on destruction, but "deadlock" isn't it. For (3), s/bool owns/int ownership_level/ would fix it.

[ 2009-11-11 Alisdair notes that this issue is very closely related to 1159, if not a dup. ]

[ 2009-11-14 Moved to Tentatively Dup after 5 positive votes on c++std-lib. ]

Proposed resolution:


1223(i). condition_variable_any lock matching?

Section: 33.7.5 [thread.condition.condvarany] Status: NAD Submitter: Jeffrey Yasskin Opened: 2009-09-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.condition.condvarany].

View all issues with NAD status.

Discussion:

For condition_variable_any, must all lock arguments to concurrent wait calls "match" in some way, similar to the requirement in 33.7.4 [thread.condition.condvar] that lock.mutex() returns the same value for each of the lock arguments supplied by all concurrently waiting threads (via wait or timed_wait)?

[ 2010-02-12 Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Rationale:

The rationale is that it doesn't matter, and you can't check: the lock types may be different, or the same and user-defined, so the implementation must provide internal synchronization anyway.

Proposed resolution:


1224(i). condition_variable_any support for recursive mutexes?

Section: 33.7.5 [thread.condition.condvarany] Status: NAD Submitter: Jeffrey Yasskin Opened: 2009-09-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.condition.condvarany].

View all issues with NAD status.

Discussion:

For condition_variable_any, are recursive mutexes allowed? (I think "no")

[ 2009-11-17 Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Proposed resolution:

Rationale:

condition_variable_any::wait accepts any type of mutex. It calls unlock precisely once on entry and lock precisely once on exit. It is up to the user to ensure that this provides the required synchronization. Use of a recursive mutex is safe if either its lock count is 1, so after the single unlock it can be acquired by another thread, or another mechanism is used to synchronize the data.


1228(i). User-specialized nothrow type traits

Section: 21.3.5.4 [meta.unary.prop] Status: NAD Submitter: Alisdair Meredith Opened: 2009-10-07 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [meta.unary.prop].

View all other issues in [meta.unary.prop].

View all issues with NAD status.

Discussion:

According to p1 21.3.3 [meta.type.synop]:

The behavior of a program that adds specializations for any of the class templates defined in this subclause is undefined unless otherwise specified.

I believe we should 'otherwise specify' for the nothrow traits, are these are exactly the use cases where the end user actually has more information than the compiler.

[ 2009-10 Santa Cruz: ]

Moved to Open. Definitely need to give the users the ability to ensure that the traits give the right answers. Unsure we want to give them the ability to say this in more than one way. Believes the noexcept proposal already gives this.

[ 2010 Pittsburgh: Moved to NAD, rationale added below. ]

Rationale:

We believe the solution offered by N3050 is superior.

Proposed resolution:

Add the following comment:

user specialization permitted to derive from std::true_type when the operation is known not to throw.

to the following traits in 21.3.5.4 [meta.unary.prop] Table 43 Type property predicates.

[ This may require a new Comments column ]

has_nothrow_default_constructor
has_nothrow_copy_constructor
has_nothrow_assign

1230(i). mem_fn and variadic templates

Section: 22.10.16 [func.memfn] Status: Dup Submitter: Alisdair Meredith Opened: 2009-10-09 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [func.memfn].

View all issues with Dup status.

Duplicate of: 920

Discussion:

Since we have removed the entry in B [implimits] for the library-specific limit for number of arguments passed to function/tuple/etc. I believe we need to update the spec for mem_fn to reflect this.

The "Remarks: Implementations may implement mem_fn as a set of overloaded function templates." no longer holds, as we cannot create an arbitrary number of such overloads. I believe we should strike the remark and add a second signature:

template<class R, class T, typename ... ArgTypes>
  unspecified mem_fn(R (T::*pm)(ArgTypes...));

I believe we need two signatures as pointer-to-data-member and pointer-to-member-function-taking-no-args appear to use subtly different syntax.

[ 920 as a similar proposed resolution. ]

Proposed resolution:

Add to 22.10 [function.objects] and 22.10.16 [func.memfn]:

template<class R, class T> unspecified mem_fn(R T::* pm)

template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...));
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) const);
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) volatile);
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) const volatile);

template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...)&);
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) const&);
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) volatile&);
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) const volatile&);

template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...)&&);
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) const&&);
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) volatile&&);
template<class R, class T, class ...Args> unspecified mem_fn(R (T::* pm)(Args...) const volatile&&);

Strike 22.10.16 [func.memfn], p5:

Remarks: Implementations may implement mem_fn as a set of overloaded function templates.


1232(i). Still swap's with rvalue-references

Section: 16 [library] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2009-10-11 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD Editorial status.

Discussion:

The current library contains still rvalue reference-swaps that seem to be overlooked in the process of switching back to lvalue-ref swaps.

[ 2009-10 Santa Cruz: ]

Editor accepts as NAD Editorial.

Proposed resolution:

  1. Change 22.3 [pairs]/1 as indicated:

    template <class T1, class T2>
    struct pair {
      ...
      void swap(pair&& p);
    };
    
  2. Change 22.3 [pairs] before p. 17 as indicated:

    void swap(pair&& p);
    
  3. Change 22.3 [pairs] before p. 21 as indicated:

    template<class T1, class T2> void swap(pair<T1, T2>& x, pair<T1, T2>& y);
    template<class T1, class T2> void swap(pair<T1, T2>&& x, pair<T1, T2>& y);
    template<class T1, class T2> void swap(pair<T1, T2>& x, pair<T1, T2>&& y);
    
  4. Change 22.4.1 [tuple.general]/2, header <tuple> synopsis, as indicated:

    // 20.5.2.9, specialized algorithms:
    template <class... Types>
    void swap(tuple<Types...>& x, tuple<Types...>& y);
    template <class... Types>
    void swap(tuple<Types...>&& x, tuple<Types...>& y);
    template <class... Types>
    void swap(tuple<Types...>& x, tuple<Types...>&& y);
    
  5. Change 22.4.4 [tuple.tuple] as indicated:

    // 20.5.2.3, tuple swap
    void swap(tuple&&)
    
  6. Change 22.4.4.3 [tuple.swap] before 1 as indicated:

    void swap(tuple&& rhs);
    
  7. Change 22.10 [function.objects]/2, header <functional> synopsis, as indicated:

    template<class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&, function<R(ArgTypes...)>&);
    template<class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&&, function<R(ArgTypes...)>&);
    template<class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&, function<R(ArgTypes...)&&);
    
  8. Change 22.10.17.3 [func.wrap.func], as indicated:

    // 20.7.15.2.2, function modifiers:
    void swap(function&&);
    template<class F, class A> void assign(F, const A&);
    
    [..]
    
    // 20.7.15.2.7, specialized algorithms:
    template <class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&, function<R(ArgTypes...)>&);
    template <class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&&, function<R(ArgTypes...)>&);
    template <class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&, function<R(ArgTypes...)>&&);
    
  9. Change 22.10.17.3.8 [func.wrap.func.alg] before 1 as indicated:

    template<class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>& f1, function<R(ArgTypes...)>& f2);
    template<class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>&& f1, function<R(ArgTypes...)>& f2);
    template<class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>& f1, function<R(ArgTypes...)>&& f2);
    
  10. Change 20.3.2.2 [util.smartptr.shared]/1 as indicated:

    // 20.8.12.2.4, modifiers:
    void swap(shared_ptr&& r);
    
    [..]
    
    // 20.8.12.2.9, shared_ptr specialized algorithms:
    template<class T> void swap(shared_ptr<T>& a, shared_ptr<T>& b);
    template<class T> void swap(shared_ptr<T>&& a, shared_ptr<T>& b);
    template<class T> void swap(shared_ptr<T>& a, shared_ptr<T>&& b);
    
  11. Change 23.4 [string.classes]/1, header <string> synopsis, as indicated:

    // 21.4.8.8: swap
    template<class charT, class traits, class Allocator>
    void swap(basic_string<charT,traits,Allocator>& lhs, basic_string<charT,traits,Allocator>& rhs);
    template<class charT, class traits, class Allocator>
    void swap(basic_string<charT,traits,Allocator>&& lhs, basic_string<charT,traits,Allocator>& rhs);
    template<class charT, class traits, class Allocator>
    void swap(basic_string<charT,traits,Allocator>& lhs, basic_string<charT,traits,Allocator>&& rhs);
    
  12. Change 24.3 [sequences]/1, header <deque> synopsis, as indicated:

    template <class T, class Allocator>
    void swap(deque<T,Allocator>& x, deque<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(deque<T,Allocator>&& x, deque<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(deque<T,Allocator>& x, deque<T,Allocator>&& y);
    
  13. Change 24.3 [sequences]/1, header <list> synopsis, as indicated:

    template <class T, class Allocator>
    void swap(list<T,Allocator>& x, list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(list<T,Allocator>&& x, list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(list<T,Allocator>& x, list<T,Allocator>&& y);
    
  14. Change 24.3 [sequences]/1, header <queue> synopsis, as indicated:

    template <class T, class Allocator>
    void swap(queue<T, Container>& x, queue<T, Container>& y);
    template <class T, class Container>
    void swap(queue<T, Container>&& x, queue<T, Container>& y);
    template <class T, class Container>
    void swap(queue<T, Container>& x, queue<T, Container>&& y);
    
    template <class T, class Container = vector<T>, class Compare = less<typename Container::value_type> >
    class priority_queue;
    template <class T, class Container, class Compare>
    void swap(priority_queue<T, Container, Compare>& x, priority_queue<T, Container, Compare>& y);
    template <class T, class Container, class Compare>
    void swap(priority_queue<T, Container, Compare>&& x, priority_queue<T, Container, Compare>& y);
    template <class T, class Container, class Compare>
    void swap(priority_queue<T, Container, Compare>& x, priority_queue<T, Container, Compare>&& y);
    
  15. Change 24.3 [sequences]/1, header <stack> synopsis, as indicated:

    template <class T, class Container>
    void swap(stack<T, Container>& x, stack<T, Container>& y);
    template <class T, class Container>
    void swap(stack<T, Container>&& x, stack<T, Container>& y);
    template <class T, class Container>
    void swap(stack<T, Container>& x, stack<T, Container>&& y);
    
  16. Change 24.3 [sequences]/1, header <vector> synopsis, as indicated:

    template <class T, class Allocator>
    void swap(vector<T,Allocator>& x, vector<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(vector<T,Allocator>&& x, vector<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(vector<T,Allocator>& x, vector<T,Allocator>&& y);
    
  17. Change 24.3.8 [deque]/2 as indicated:

    iterator erase(const_iterator position);
    iterator erase(const_iterator first, const_iterator last);
    void swap(deque<T,Allocator>&&);
    void clear();
    
    [..]
    
    // specialized algorithms:
    template <class T, class Allocator>
    void swap(deque<T,Allocator>& x, deque<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(deque<T,Allocator>&& x, deque<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(deque<T,Allocator>& x, deque<T,Allocator>&& y);
    
  18. Change [deque.special] as indicated:

    template <class T, class Allocator>
    void swap(deque<T,Allocator>& x, deque<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(deque<T,Allocator>&& x, deque<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(deque<T,Allocator>& x, deque<T,Allocator>&& y);
    
  19. Change [forwardlist]/2 as indicated:

    iterator erase_after(const_iterator position);
    iterator erase_after(const_iterator position, iterator last);
    void swap(forward_list<T,Allocator>&&);
    
    [..]
    
    // 23.3.3.6 specialized algorithms:
    template <class T, class Allocator>
    void swap(forward_list<T,Allocator>& x, forward_list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(forward_list<T,Allocator>&& x, forward_list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(forward_list<T,Allocator>& x, forward_list<T,Allocator>&& y);
    
  20. Change [forwardlist.spec] as indicated:

    template <class T, class Allocator>
    void swap(forward_list<T,Allocator>& x, forward_list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(forward_list<T,Allocator>&& x, forward_list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(forward_list<T,Allocator>& x, forward_list<T,Allocator>&& y);
    
  21. Change 24.3.10 [list]/2 as indicated:

    iterator erase(const_iterator position);
    iterator erase(const_iterator position, const_iterator last);
    void swap(list<T,Allocator>&&);
    void clear();
    
    [..]
    
    // specialized algorithms:
    template <class T, class Allocator>
    void swap(list<T,Allocator>& x, list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(list<T,Allocator>&& x, list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(list<T,Allocator>& x, list<T,Allocator>&& y);
    
  22. Change [list.special] as indicated:

    template <class T, class Allocator>
    void swap(list<T,Allocator>& x, list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(list<T,Allocator>&& x, list<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(list<T,Allocator>& x, list<T,Allocator>&& y);
    
  23. Change 24.6.6.1 [queue.defn] as indicated:

    void swap(queue&& q) { c.swap(q.c); }
    
    [..]
    
    template <class T, class Container>
    void swap(queue<T, Container>& x, queue<T, Container>& y);
    template <class T, class Container>
    void swap(queue<T, Container>&& x, queue<T, Container>& y);
    template <class T, class Container>
    void swap(queue<T, Container>& x, queue<T, Container>&& y);
    
  24. Change 24.6.6.6 [queue.special] as indicated:

    template <class T, class Container>
    void swap(queue<T, Container>& x, queue<T, Container>& y);
    template <class T, class Container>
    void swap(queue<T, Container>&& x, queue<T, Container>& y);
    template <class T, class Container>
    void swap(queue<T, Container>& x, queue<T, Container>&& y);
    
  25. Change 24.6.7 [priority.queue]/1 as indicated:

    void swap(priority_queue&&);
    
    // no equality is provided
    template <class T, class Container, class Compare>
    void swap(priority_queue<T, Container, Compare>& x, priority_queue<T, Container, Compare>& y);
    template <class T, class Container, class Compare>
    void swap(priority_queue<T, Container, Compare>&& x, priority_queue<T, Container, Compare>& y);
    template <class T, class Container, class Compare>
    void swap(priority_queue<T, Container, Compare>& x, priority_queue<T, Container, Compare>&& y);
    
  26. Change 24.6.7.5 [priqueue.special] as indicated:

    template <class T, class Container, Compare>
    void swap(priority_queue<T, Container, Compare>& x, priority_queue<T, Container, Compare>& y);
    template <class T, class Container, Compare>
    void swap(priority_queue<T, Container, Compare>&& x, priority_queue<T, Container, Compare>& y);
    template <class T, class Container, Compare>
    void swap(priority_queue<T, Container, Compare>& x, priority_queue<T, Container, Compare>&& y);
    
  27. Change 24.6.8.2 [stack.defn] as indicated:

    void swap(stack&& s) { c.swap(s.c); }
    
    [..]
    
    template <class T, class Allocator>
    void swap(stack<T,Allocator>& x, stack<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(stack<T,Allocator>&& x, stack<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(stack<T,Allocator>& x, stack<T,Allocator>&& y);
    
  28. Change 24.6.8.7 [stack.special] as indicated:

    template <class T, class Container>
    void swap(stack<T, Container>& x, stack<T, Container>& y);
    template <class T, class Container>
    void swap(stack<T, Container>&& x, stack<T, Container>& y);
    template <class T, class Container>
    void swap(stack<T, Container>& x, stack<T, Container>&& y);
    
  29. Change 24.3.11 [vector]/2 as indicated:

    void swap(vector<T,Allocator>&&);
    void clear();
    
    [..]
    
    // specialized algorithms:
    template <class T, class Allocator>
    void swap(vector<T,Allocator>& x, vector<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(vector<T,Allocator>&& x, vector<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(vector<T,Allocator>& x, vector<T,Allocator>&& y);
    
  30. Change 24.3.11.3 [vector.capacity] before p. 8 as indicated:

    void swap(vector<T,Allocator>&& x);
    
  31. Change [vector.special] as indicated:

    template <class T, class Allocator>
    void swap(vector<T,Allocator>& x, vector<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(vector<T,Allocator>&& x, vector<T,Allocator>& y);
    template <class T, class Allocator>
    void swap(vector<T,Allocator>& x, vector<T,Allocator>&& y);
    
  32. Change 24.3.12 [vector.bool]/1 as indicated:

    iterator erase(const_iterator first, const_iterator last);
    void swap(vector<bool,Allocator>&&);
    static void swap(reference x, reference y);
    
  33. Change 24.4 [associative]/1, header <map> synopsis as indicated:

    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator>& x, map<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator&& x, map<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator& x, map<Key,T,Compare,Allocator>&& y);
    
    [..]
    
    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator>& x, multimap<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator&& x, multimap<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator& x, multimap<Key,T,Compare,Allocator>&& y);
    
  34. Change 24.4 [associative]/1, header <set> synopsis as indicated:

    template <class Key, class Compare, class Allocator>
    void swap(set<Key,Compare,Allocator>& x, set<Key,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(set<Key,T,Compare,Allocator&& x, set<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(set<Key,T,Compare,Allocator& x, set<Key,T,Compare,Allocator>&& y);
    
    [..]
    
    template <class Key, class Compare, class Allocator>
    void swap(multiset<Key,Compare,Allocator>& x, multiset<Key,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(multiset<Key,T,Compare,Allocator&& x, multiset<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(multiset<Key,T,Compare,Allocator& x, multiset<Key,T,Compare,Allocator>&& y);
    
  35. Change 24.4.4 [map]/2 as indicated:

    iterator erase(const_iterator first, const_iterator last);
    void swap(map<Key,T,Compare,Allocator>&&);
    void clear();
    
    [..]
    
    // specialized algorithms:
    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator>& x, map<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator&& x, map<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator& x, map<Key,T,Compare,Allocator>&& y);
    
  36. Change [map.special] as indicated:

    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator>& x, map<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator>&& x, map<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(map<Key,T,Compare,Allocator>& x, map<Key,T,Compare,Allocator>&& y);
    
  37. Change 24.4.5 [multimap]/2 as indicated:

    iterator erase(const_iterator first, const_iterator last);
    void swap(multimap<Key,T,Compare,Allocator>&&);
    void clear();
    
    [..]
    
    // specialized algorithms:
    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator>& x, multimap<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator&& x, multimap<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator& x, multimap<Key,T,Compare,Allocator>&& y);
    
  38. Change [multimap.special] as indicated:

    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator>& x, multimap<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator>&& x, multimap<Key,T,Compare,Allocator>& y);
    template <class Key, class T, class Compare, class Allocator>
    void swap(multimap<Key,T,Compare,Allocator>& x, multimap<Key,T,Compare,Allocator>&& y);
    
  39. Change 24.4.6 [set]/2 and [set.special] as indicated: (twice!)

    // specialized algorithms:
    template <class Key, class Compare, class Allocator>
    void swap(set<Key,Compare,Allocator>& x, set<Key,Compare,Allocator>& y);
    template <class Key, class Compare, class Allocator>
    void swap(set<Key,Compare,Allocator&& x, set<Key,Compare,Allocator>& y);
    template <class Key, class Compare, class Allocator>
    void swap(set<Key,Compare,Allocator& x, set<Key,Compare,Allocator>&& y);
    
  40. Change 24.4.7 [multiset]/2 as indicated:

    iterator erase(const_iterator first, const_iterator last);
    void swap(multiset<Key,Compare,Allocator>&&);
    void clear();
    
    [..]
    
    // specialized algorithms:
    template <class Key, class Compare, class Allocator>
    void swap(multiset<Key,Compare,Allocator>& x, multiset<Key,Compare,Allocator>& y);
    template <class Key, class Compare, class Allocator>
    void swap(multiset<Key,Compare,Allocator&& x, multiset<Key,Compare,Allocator>& y);
    template <class Key, class Compare, class Allocator>
    void swap(multiset<Key,Compare,Allocator& x, multiset<Key,Compare,Allocator>&& y);
    
  41. Change [multiset.special] as indicated:

    template <class Key, class Compare, class Allocator>
    void swap(multiset<Key,Compare,Allocator>& x, multiset<Key,Compare,Allocator>& y);
    template <class Key, class Compare, class Allocator>
    void swap(multiset<Key,Compare,Allocator>&& x, multiset<Key,Compare,Allocator>& y);
    template <class Key, class Compare, class Allocator>
    void swap(multiset<Key,Compare,Allocator>& x, multiset<Key,Compare,Allocator>&& y);
    

1233(i). Missing unique_ptr signatures in synopsis

Section: 20.2 [memory] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2009-10-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [memory].

View all issues with NAD Editorial status.

Discussion:

Related to 296. Some unique_ptr signatures are missing from the synopsis in 20.2 [memory].

[ 2009-11-04 Howard adds: ]

Moved to Tentatively NAD Editorial. The editor has adopted the fix.

Proposed resolution:

Add in 20.2 [memory], Header <memory> synopsis missing declarations as shown below:

// 20.8.11 Class unique_ptr:
template <class X> class default_delete;
template<class T> struct default_delete<T[]>;
template <class X, class D = default_delete<T>> class unique_ptr;
template<class T, class D> class unique_ptr<T[], D>;

template<class T, class D> void swap(unique_ptr<T, D>& x, unique_ptr<T, D>& y);

template<class T1, class D1, class T2, class D2>
bool operator==(const unique_ptr<T1, D1>& x, const unique_ptr<T2, D2>& y);
template<class T1, class D1, class T2, class D2>
bool operator!=(const unique_ptr<T1, D1>& x, const unique_ptr<T2, D2>& y);
template<class T1, class D1, class T2, class D2>
bool operator<(const unique_ptr<T1, D1>& x, const unique_ptr<T2, D2>& y);
template<class T1, class D1, class T2, class D2>
bool operator<=(const unique_ptr<T1, D1>& x, const unique_ptr<T2, D2>& y);
template<class T1, class D1, class T2, class D2>
bool operator>(const unique_ptr<T1, D1>& x, const unique_ptr<T2, D2>& y);
template<class T1, class D1, class T2, class D2>
bool operator>=(const unique_ptr<T1, D1>& x, const unique_ptr<T2, D2>& y);

1235(i). Issue with C++0x random number proposal

Section: 28.5.3.6 [rand.req.dist] Status: NAD Submitter: Matthias Troyer Opened: 2009-10-12 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [rand.req.dist].

View all issues with NAD status.

Discussion:

There exist optimized, vectorized vendor libraries for the creation of random number generators, such as Intel's MKL [1] and AMD's ACML [2]. In timing tests we have seen a performance gain of a factor of up to 80 (eighty) compared to a pure C++ implementation (in Boost.Random) when using these generator to generate a sequence of normally distributed random numbers. In codes dominated by the generation of random numbers (we have application codes where random number generation is more than 50% of the CPU time) this factor 80 is very significant.

To make use of these vectorized generators, we use a C++ class modeling the RandomNumberEngine concept and forwarding the generation of random numbers to those optimized generators. For example:

namespace mkl {
 class mt19937 {.... };
}

For the generation of random variates we also want to dispatch to optimized vectorized functions in the MKL or ACML libraries. See this example:

mkl::mt19937 eng;
std::normal_distribution<double> dist;

double n = dist(eng);

Since the variate generation is done through the operator() of the distribution there is no customization point to dispatch to Intel's or AMD's optimized functions to generate normally distributed numbers based on the mt19937 generator. Hence, the performance gain of 80 cannot be achieved.

Contrast this with TR1:

mkl::mt19937 eng;
std::tr1::normal_distribution<double> dist;
std::tr1::variate_generator<mkl::mt19937,std::tr1::normal_distribution<double> > rng(eng,dist);
double n = rng();

This - admittedly much uglier from an aestethic point of view - design allowed optimization by specializing the variate_generator template for mkl::mt19937:

namespace std { namespace tr1 {

template<>
class variate_generator<mkl::mt19937,std::tr1::normal_distribution<double> > { .... };

} }

A similar customization point is missing in the C++0x design and prevents the optimized vectorized version to be used.

Suggested resolution:

Add a customization point to the distribution concept. Instead of the variate_generator template this can be done through a call to a free function generate_variate found by ADL instead of operator() of the distribution:

template <RandomNumberDistribution, class RandomNumberEngine>
typename RandomNumberDistribution ::result_type
generate_variate(RandomNumberDistribution const& dist, RandomNumberEngine& eng);

This function can be overloaded for optimized enginges like mkl::mt19937.

[ 2009-10 Santa Cruz: ]

NAD Future. No time to add this feature for C++0X.

[LEWG Kona 2017]

Recommend NAD: The standard has changed enough that the issue doesn't make sense anymore. Write a paper proposing a way to get this performance as changes to the current library.

[Kona 2019: Jonathan notes:]

Libstdc++ has the following non-standard extensions for more efficient generation of large numbers of random numbers:

template<typename ForwardIterator, typename UniformRandomNumberGenerator>
void __generate(ForwardIterator, ForwardIterator, 
                UniformRandomNumberGenerator&);

template<typename ForwardIterator, typename UniformRandomNumberGenerator>
void __generate(ForwardIterator, ForwardIterator, 
                UniformRandomNumberGenerator&, const param_type&);

template<typename UniformRandomNumberGenerator>
void __generate(result_type*, result_type*,
                UniformRandomNumberGenerator&, const param_type&);

Proposed resolution:


1236(i). reserved identifiers in programs not using the library

Section: 16 [library] Status: NAD Submitter: Sean Hunt Opened: 2009-10-13 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

I wasn't sure whether to consider this a library or a language issue, because the issue is I think it's incorrectly categorized as being part of the library, so I thought I'd send a message to both of you and let you sort it out.

Most reserved identifiers are treated as unilaterally available to the implementation, such as to implement language extensions, or provide macros documenting its functionality. However, the requirements for reserved identifers are in 16.4.5.3 [reserved.names], which are a subsection of 16.4.5 [constraints]. 16.4.5.1 [constraints.overview] appears only to apply to "C++ programs that use the facilities of the C++ standard library", meaning that, in theory, all implementations are erroneous in having any non-standard identifiers predefined for programs that do not, at some point, include a standard library header.

Furthermore, it's unclear whether the use of certain identifiers is UB or results in an ill-formed program. In particular, 16.4.5.3.3 [macro.names] uses a "shall not", where [global.names] says that names are "reserved to the implementation". 16.4.5.3 [reserved.names] seems only to cover the instance of a name being described as "reserved", so are implementations required to diagnose a program that performs, as an example, "#undef get"?

[ 2009 Santa Cruz: ]

Move to NAD. There may in theory be multiple interpretations possible, but there's no evidence that this causes any genuine problems or uncertainty about what implementations are allowed to do. We do not believe this rises to the level of a defect.

Proposed resolution:


1239(i). Defect report

Section: 21.3.5.4 [meta.unary.prop] Status: NAD Editorial Submitter: David Abrahams Opened: 2009-10-16 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [meta.unary.prop].

View all other issues in [meta.unary.prop].

View all issues with NAD Editorial status.

Discussion:

Table 43 defines a number of traits that yield true for arrays of class types with the trait's property, but not arrays of other types with that property. For example, has_trivial_default_constructor:

T is a trivial type (3.9) or a class type with a trivial default constructor (12.1) or an array of such a class type.

[ 2009-10 post-Santa Cruz: ]

An array of a trivial type is a trivial type.

Mark as Tentatively NAD Editorial. The wording is OK as is, since an array of a trivial type is a trivial type, but the wording as proposed might be clearer.

Rationale:

The wording is OK as is, since an array of a trivial type is a trivial type. Project editor may wish to accept the suggested wording as editorial.

Proposed resolution:

Change all the traits in question following this pattern:

T is a trivial type (3.9) or a class type with a trivial default constructor (12.1), or an array of such a class type.

i.e., add a comma and delete a "class."


1242(i). Enable SCARY iterators

Section: 24 [containers] Status: NAD Submitter: Herb Sutter Opened: 2009-10-21 Last modified: 2020-09-06

Priority: Not Prioritized

View other active issues in [containers].

View all other issues in [containers].

View all issues with NAD status.

Discussion:

See N2980.

[ 2009-10 Santa Cruz ]

The paper was lengthy discussed but considerable concern remained to add this feature to C++0x. The LWG does not wish to make a change at this time. Strong consensus was found to consider it for C++1x, though.

[LEWG Kona 2017]

Recommend Open: seems to be existing practice now

[2018-11 San Diego Thursday night issue processing]

Status to NAD; Alisdair to provide rationale.

Rationale:

Insufficient motivation to make a change now. This is existing practice in the major implementations, and they are motivated to do this even if not mandated to do so, because it improves compilation time which benefits their users. If the issue submitter feels strongly about this they are invited to write an updated paper with suggested wording to mandate it.

Proposed resolution:


1243(i). Missing operator+= (initializer_list<T>) for valarray

Section: 28.6.2.7 [valarray.cassign] Status: NAD Submitter: Daniel Krügler Opened: 2009-10-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [valarray.cassign].

View all issues with NAD status.

Discussion:

Addresses JP 64

During the additions of initializer_list overloads basic_string added:

basic_string& operator+=(initializer_list<charT>);

but

valarray<T>& operator+= (initializer_list<T>);

was not defined.

[ Daniel adds on opening: ]

Recommend NAD. The operator+= overload of basic_string behaves as-if calling append, which is completely different in meaning as the existing operator+= overloads in valarray which just sum the value or values to the existing elements. The suggestion to add a corresponding append function to valarray was not considered as appropriate and the request was withdrawn (c++std-lib-24968).

[ 2009-10 Santa Cruz: ]

Mark as NAD. Request has been withdrawn by NB.

Proposed resolution:

Add to 28.6.2.7 [valarray.cassign]:

valarray<T>& operator+= (initializer_list<T>);

1246(i). vector::resize() missing efficiency guarantee

Section: 24.3.11.3 [vector.capacity] Status: NAD Submitter: David Abrahams Opened: 2009-10-24 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [vector.capacity].

View all other issues in [vector.capacity].

View all issues with NAD status.

Discussion:

If v is a vector, I think repeated calls to v.resize( v.size() + 1 ) should be amortized O(1), but it's not clear that's true from the text of the standard:

void resize(size_type sz);

Effects: If sz < size(), equivalent to erase(begin() + sz, end());. If size() < sz, appends sz - size() default constructed elements to the sequence.

Seems to me if we used push_back instead of appends, we might be giving the guarantee I'd like. Thoughts?

[ 2009-11-10 Howard adds: ]

Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below.

Proposed resolution:

In 24.3.11.3 [vector.capacity]/10, change

void resize(size_type sz);

Effects: If sz < size(), equivalent to erase(begin() + sz, end());. If size() < sz, appends sz - size() default constructed elements to the sequence equivalent to sz - size() consecutive evaluations of push_back(T()).

Rationale:

The description in terms of push_back led some to believe that one could expect the exact same growth pattern from both resize and push_back (e.g.) which could lead to sub-optimal implementations. Additionally, 24.3.11 [vector], p1 includes a statement that this container "supports (amortized) constant time insert and erase operations at the end;", therefore addressing the concern of this issue.


1251(i). move constructing basic_stringbuf

Section: 31.8.2.2 [stringbuf.cons] Status: NAD Submitter: Martin Sebor Opened: 2009-10-29 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [stringbuf.cons].

View all issues with NAD status.

Discussion:

I just came across issue 1204 -- Global permission to move, which seems to address the concern raised by the example in c++std-lib-25030.

IIUC, the example violates the permission to assume that arguments bound to rvalue references are unnamed temporaries granted to implementations by the resolution of issue 1204 - Global permission to move.

I.e., the ostringstream(ostringstream &&rhs) ctor can leave the rhs pointers pointing to the newly constructed object's buffer just as long as the dtor doesn't change or invalidate the buffer. The caller may not make any assumptions about rhs after the move beyond it being safe to destroy or reassign.

So unless I misunderstood something, I still think the basic_stringbuf move ctor is overspecified. Specifically, I think the third sentence in the Effects clause and the last 6 bullets in the Postconditions clause can, and IMO should, be stricken.

[ 2010-01-31 Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Rationale:

The sense of 1251 appears to be that the basic_stringbuf move constructor offers more guarantees than the minimum. This is true, and quite correct. The additional words guarantee that the internal buffer has genuinely transferred from one object to another, and further operations on the original will not affect the buffer of the newly created object. This is a very important guarantee, much as we see that a moved-from unique_ptr is guaranteed to be empty.

Proposed resolution:

Strike from 31.8.2.2 [stringbuf.cons]:

basic_stringbuf(basic_stringbuf&& rhs);

Effects: Move constructs from the rvalue rhs. It is implementation-defined whether the sequence pointers in *this (eback(), gptr(), egptr(), pbase(), pptr(), epptr()) obtain the values which rhs had. Whether they do or not, *this and rhs reference separate buffers (if any at all) after the construction. The openmode, locale and any other state of rhs is also copied.

Postconditions: Let rhs_p refer to the state of rhs just prior to this construction and let rhs_a referto the state of rhs just after this construction.


1259(i). Should initializer-list constructors move elements?

Section: 24.2.4 [sequence.reqmts] Status: NAD Submitter: Sean Hunt Opened: 2009-11-05 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [sequence.reqmts].

View all other issues in [sequence.reqmts].

View all issues with NAD status.

Discussion:

According to 24.2.4 [sequence.reqmts], X(il) is equivalent to X(il.begin(), il.end()). Should it instead be equivalent to X(move_iterator(il.begin()), move_iterator(il.end())) so that needless copies are not made? This doesn't seem ideal either - it may make more sense to provide two overloads for the constructor, one for move and one for copy.

[ 2009-11-10 Howard adds: ]

I've moved this issue to Tentatively NAD after 5 positive votes on c++std-lib, and added a rationale below.

Proposed resolution:

Rationale:

There is no consensus at this time within EWG or CWG to make the required language changes. Therefore this is not something that the LWG can even consider. Should such language changes be made for a future standard, no doubt there would need to be an accompanying library impact survey.


1263(i). missing swap overloads for regex

Section: 32.3 [re.syn] Status: NAD Submitter: Howard Hinnant Opened: 2009-11-12 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [re.syn].

View all issues with NAD status.

Discussion:

Addresses: UK 314

In Message c++std-lib-25529, Alisdair writes:

UK comment 314 requests rvalue swap overloads in a couple of places they were missed.

We have in general reverted to the single swap signature taking lvalue references, which could be seen as the alternative solution to UK 314, bringing consistency to the standard <g>

Either way, I no longer expect to see any work to resolve this comment - the work is complete and it should be either marked Rejected, or Accepted with Modifications (namely, removing all other rvalue swaps!)

[ Moved to Tentatively NAD after 5 positive votes on c++std-lib. ]

Proposed resolution:

Rationale:

We have in general reverted to the single swap signature taking lvalue references, which could be seen as the alternative solution to UK 314, bringing consistency to the standard.


1265(i). longjmp and destructors

Section: 17.13 [support.runtime] Status: NAD Submitter: Sean Hunt Opened: 2009-11-16 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [support.runtime].

View all issues with NAD status.

Discussion:

17.13 [support.runtime]/4 says that longjmp is undefined if unwinding by the mechanism used by catch and throw would invoke any nontrivial destructors. However, the text as written is rather vague, in particular when dealing with catch(...):

void foo() {
  jump_buf buf;
  non_trivial_dtor n1; // 1
  if (!setjmp(buf)) {
    non_trivial_dtor n2; // 2
    try {
      longjmp(buf, 1);
    } catch (...) {
    }
  }
}

My interpretation of the meaning of 17.13 [support.runtime]/4 is that declaration 2, but not 1, would cause the longjmp to be undefined behavior. However, it's not entirely clear from the text. Arguably, replacing the setjmp and longjmp with catch would still cause the destructor for n1 to be called after the unwinding, which would lead to undefined behavior. This is clearly not an intended consequence of the wording. However, it is probably still UB, as n1 now has "indeterminate" value, and running its destructor on foo's exit will cause Bad Things.

Declarations 2 has a more interesting issue. The catch(...) muddles up the definition that uses throw and catch - if longjmp() were indeed a throw, control would never return to the setjmp. As such, n2's destructor wouldn't be called (except by the argument for n1, which is that the destructor would be called later as the frame was left in the normal control flow).

I suggest that paragraph 4 of 17.13 [support.runtime] should be replaced with the following, or something that reads better but has the same effect:

The function signature longjmp(jmp_buf jbuf, int val) has more restricted behavior in this International Standard. A call to longjmp has undefined behavior if any non-trivial destructors would be called were the longjmp call replaced with a throw-expression whose nearest matching handler were a (possibly imaginary) function-try-block on the function containing the corresponding setjmp call.

[ 2009-11-17 Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Proposed resolution:

Change 17.13 [support.runtime]/4:

The function signature longjmp(jmp_buf jbuf, int val) has more restricted behavior in this International Standard. A setjmp/longjmp call pair has undefined behavior if replacing the setjmp and longjmp by catch and throw would invoke any non-trivial destructors for any automatic objects. A call to longjmp has undefined behavior if any non-trivial destructors would be called were the longjmp call replaced with a throw-expression whose nearest matching handler were a (possibly imaginary) function-try-block on the function containing the corresponding setjmp call.

Rationale:

In the given example, it is clear that it is only n2 and not n1 that is destroyed by the longjmp.

At this late stage in the standards process, we are focusing on issues that impact users or implementers. Trying to rewrite complex wording just for the sake of improved clarity is likely to do more harm than good.


1282(i). A proposal to add std::split algorithm

Section: 27 [algorithms] Status: NAD Submitter: Igor Semenov Opened: 2009-12-07 Last modified: 2019-02-26

Priority: Not Prioritized

View other active issues in [algorithms].

View all other issues in [algorithms].

View all issues with NAD status.

Discussion:

  1. Motivation and Scope

    Splitting strings into parts by some set of delimiters is an often task, but there is no simple and generalized solution in C++ Standard. Usually C++ developers use std::basic_stringstream<> to split string into parts, but there are several inconvenient restrictions:

  2. Impact on the Standard

    This algorithm doesn't interfere with any of current standard algorithms.

  3. Design Decisions

    This algorithm is implemented in terms of input/output iterators. Also, there is one additional wrapper for const CharType * specified delimiters.

  4. Example implementation

    template< class It, class DelimIt, class OutIt >
    void split( It begin, It end, DelimIt d_begin, DelimIt d_end, OutIt out )
    {
       while ( begin != end )
       {
           It it = std::find_first_of( begin, end, d_begin, d_end );
           *out++ = std::make_pair( begin, it );
           begin = std::find_first_of( it, end, d_begin, d_end,
               std::not2( std::equal_to< typename It::value_type >() ) );
       }
    }
    
    template< class It, class CharType, class OutIt >
    void split( It begin, It end, const CharType * delim, OutIt out )
    {
       split( begin, end, delim, delim + std::strlen( delim ), out );
    }
    
  5. Usage

    std::string ss( "word1 word2 word3" );
    std::vector< std::pair< std::string::const_iterator, std::string::const_iterator > > v;
    split( ss.begin(), ss.end(), " ", std::back_inserter( v ) );
    
    for ( int i = 0; i < v.size(); ++i )
    {
       std::cout << std::string( v[ i ].first, v[ i ].second ) << std::endl;
    }
    // word1
    // word2
    // word3
    

[ 2010-01-22 Moved to Tentatively NAD Future after 5 positive votes on c++std-lib. Rationale added below. ]

[LEWG Kona 2017]

Recommend NAD: Paper encouraged. Have papers for this; LEWG259.

Rationale:

The LWG is not considering completely new features for standardization at this time. We would like to revisit this good suggestion for a future TR and/or standard.

Proposed resolution:

Add to the synopsis in 27.1 [algorithms.general]:

template< class ForwardIterator1, class ForwardIterator2, class OutputIterator >
  void split( ForwardIterator1 first, ForwardIterator1 last,
              ForwardIterator2 delimiter_first, ForwardIterator2 delimiter_last,
              OutputIterator result );

template< class ForwardIterator1, class CharType, class OutputIterator >
  void split( ForwardIterator1 first, ForwardIterator1 last,
              const CharType * delimiters, OutputIterator result );

Add a new section [alg.split]:

template< class ForwardIterator1, class ForwardIterator2, class OutputIterator >
  void split( ForwardIterator1 first, ForwardIterator1 last,
              ForwardIterator2 delimiter_first, ForwardIterator2 delimiter_last,
              OutputIterator result );

1. Effects: splits the range [first, last) into parts, using any element of [delimiter_first, delimiter_last) as a delimiter. Results are pushed to output iterator in the form of std::pair<ForwardIterator1, ForwardIterator1>. Each of these pairs specifies a maximal subrange of [first, last) which does not contain a delimiter.

2. Returns: nothing.

3. Complexity: Exactly last - first assignments.

template< class ForwardIterator1, class CharType, class OutputIterator >
  void split( ForwardIterator1 first, ForwardIterator1 last,
              const CharType * delimiters, OutputIterator result );

1. Effects: split the range [first, last) into parts, using any element of delimiters (interpreted as zero-terminated string) as a delimiter. Results are pushed to output iterator in the form of std::pair<ForwardIterator1, ForwardIterator1>. Each of these pairs specifies a maximal subrange of [first, last) which does not contain a delimiter.

2. Returns: nothing.

3. Complexity: Exactly last - first assignments.


1289(i). Generic casting requirements for smart pointers

Section: 22.2 [utility] Status: NAD Submitter: Ion Gaztañaga Opened: 2009-12-14 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [utility].

View all issues with NAD status.

Discussion:

In section 16.4.4.6 [allocator.requirements], Table 40 — Allocator requirements, the following expression is required for allocator pointers:

Table 40 — Allocator requirements
Expression Return type Assertion/note
pre-/post-condition
Default
static_cast<X::pointer>(w) X::pointer static_cast<X::pointer>(w) == p  

To achieve this expression, a smart pointer writer must introduce an explicit conversion operator from smart_ptr<void> to smart_ptr<T> so that static_cast<pointer>(void_ptr) is a valid expression. Unfortunately this explicit conversion weakens the safety of a smart pointer since the following expression (invalid for raw pointers) would become valid:

smart_ptr<void> smart_v = ...;
smart_ptr<T> smart_t(smart_v);

On the other hand, shared_ptr also defines its own casting functions in 20.3.2.2.10 [util.smartptr.shared.cast], and although it's unlikely that a programmer will use shared_ptr as allocator::pointer, having two different ways to do the same cast operation does not seem reasonable. A possible solution would be to replace static_cast<X::pointer>(w) expression with a user customizable (via ADL) static_pointer_cast<value_type>(w), and establish the xxx_pointer_cast functions introduced by shared_ptr as the recommended generic casting utilities of the standard.

Unfortunately, we've experienced problems in Boost when trying to establish xxx_pointer_cast as customization points for generic libraries (http://objectmix.com/c/40424-adl-lookup-explicit-template-parameters.html) because these casting functions are called with explicit template parameters and the standard says in 13.10.2 [temp.arg.explicit] p.8 "Explicit template argument specification":

8 ...But when a function template with explicit template arguments is used, the call does not have the correct syntactic form unless there is a function template with that name visible at the point of the call. If no such name is visible, the call is not syntactically well-formed and argument-dependent lookup does not apply.

So we can do this:

template<class BasePtr>
void generic_ptr_swap(BasePtr p)
{
  //ADL customization point
  swap(p, p);
  //...
}

but not the following:

template<class BasePtr>
void generic_ptr_algo(BasePtr p)
{
  typedef std::pointer_traits<BasePtr>::template
     rebind<Derived> DerivedPtr;
  DerivedPtr dp = static_pointer_cast<Derived>(p);
}

The solution to make static_pointer_cast a customization point is to add a generic declaration (no definition) of static_pointer_cast in a namespace (like std) and apply "using std::static_pointer_cast" declaration to activate ADL:

namespace std{

template<typename U, typename T>
unspecified
static_pointer_cast(T&&) = delete;

}

template<class BasePtr>
void generic_ptr_algo(BasePtr p)
{
  typedef std::pointer_traits<BasePtr>::template
     rebind<Derived> DerivedPtr;

  //ADL applies because static_pointer_cast is made
  //  visible according to [temp.arg.explicit]/8
  using std::static_pointer_cast;

  DerivedPtr dp = static_pointer_cast<Derived>(p);

  //...
}

A complete solution will need also the definition of static_pointer_cast for raw pointers, and this definition has been present in Boost (http://www.boost.org/boost/ pointer_cast.hpp) for years.

[ 2010-03-26 Daniel made editorial adjustments to the proposed wording. ]

[ Moved to NAD Future at 2010-11 Batavia ]

This is a new feature rather than a defect. It can be added later: "this is such a hairy area that people will put up with changes"

[LEWG Kona 2017]

Recommend NAD: NAD. Should bring a paper as a proposal for 2020.

Proposed resolution:

Add to section 22.2 [utility] Utility components, Header <utility> synopsis:

// 20.3.X, generic pointer cast functions

template<typename U, typename T>
unspecified
static_pointer_cast(T&&) = delete;

template<typename U, typename T>
unspecified
dynamic_pointer_cast(T&&) = delete;

template<typename U, typename T>
unspecified
const_pointer_cast(T&&) = delete;

//Overloads for raw pointers
template<typename U, typename T>
auto static_pointer_cast(T* t) -> decltype(static_cast<U*>(t));

template<typename U, typename T>
auto dynamic_pointer_cast(T* t) -> decltype(dynamic_cast<U*>(t));

template<typename U, typename T>
auto const_pointer_cast(T* t) -> decltype(const_cast<U*>(t));

Add to section 22.2 [utility] Utility components, a new subclause 20.3.X Pointer cast utilities [pointer.cast]:

20.3.X Pointer cast utilities [pointer.cast]

1 The library defines generic pointer casting function templates so that template code can explicitly make these names visible and activate argument-dependent lookup for pointer cast calls.

//Generic declarations
template<typename U, typename T>
unspecified
static_pointer_cast(T&&) = delete;

template<typename U, typename T>
unspecified
dynamic_pointer_cast(T&&) = delete;

template<typename U, typename T>
unspecified
const_pointer_cast(T&&) = delete;

2 The library also defines overloads of these functions for raw pointers.

//Overloads for raw pointers
template<typename U, typename T>
auto static_pointer_cast(T* t) -> decltype(static_cast<U*>(t));

Returns: static_cast<U*>(t)

template<typename U, typename T>
auto dynamic_pointer_cast(T* t) -> decltype(dynamic_cast<U*>(t));

Returns: dynamic_cast<U*>(t)

template<typename U, typename T>
auto const_pointer_cast(T* t) -> decltype(const_cast<U*>(t));

Returns: const_cast<U*>(t)

[Example:

#include <utility> //static_pointer_cast
#include <memory>  //pointer_traits

class Base{};
class Derived : public Base{};

template<class BasePtr>
void generic_pointer_code(BasePtr b)
{
   typedef std::pointer_traits<BasePtr>::template
      rebind<Derived> DerivedPtr;

   using std::static_pointer_cast;
   //ADL applies now that static_pointer_cast is visible
   DerivedPtr d = static_pointer_cast<Derived>(b);
}

end example]

Replace in section 16.4.4.6 [allocator.requirements] Table 40 — Allocator requirements, the following table entries for allocator pointers:

Table 40 — Allocator requirements
Expression Return type Assertion/note
pre-/post-condition
Default
static_pointer_cast<X::pointerT>(w) X::pointer static_pointer_cast<X::pointerT>(w) == p  
static_pointer_cast<X::const_pointerconst T>(w) X::const_pointer static_pointer_cast<X::const_pointerconst T>(z) == q  

1296(i). map and multimap value_compare overspecified

Section: 24.4.4 [map] Status: NAD Submitter: Alisdair Meredith Opened: 2009-12-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [map].

View all issues with NAD status.

Discussion:

The container class templates map and multimap both contain a nested type called value_compare, that is used to compare the value_type pair elements, an adaptor of the user-supplied comparison function-like object.

I believe these types are over-specified, as we require a distinct type for each template, even though the allocator plays no part in the comparator, and map and multimap value_compare classes could easily be shared. The benefits are similar to the SCARY iterator proposal (although on a much smaller scale!) but unlike SCARY, this is not a QoI issue today but actively prohibited.

If the value_compare classes were marked 'exposition only', a vendor would be free to experiment with implementations that do not produce so many template instantiations with negligible impact on conforming programs. (There is a minor risk that programs could no longer portably overload functions taking value_compare types. This scenario is extremely unlikely outside conformance suites.)

(Note that there are no similar problems for unordered maps, nor any of the set variants)

[ 2010-01-31 Moved to Tentatively NAD after 5 positive votes on c++std-lib. Rationale added below. ]

Rationale:

The value_compare specification is an unfortunate bit from the past that we have to live with. Fortunately vendors can work around the problems mentioned in this issue.

Proposed resolution:

p2 24.4.4 [map]: Above the declaration of class value_compare in the map synopsis, add:

template <class Key, class T, class Compare = less<Key>,
          class Allocator = allocator<pair<const Key, T> > >
class map {
public:
  // types:
  ...
  // exposition only.
  class value_compare
    : public binary_function<value_type,value_type,bool> {
    ...

p2 24.4.5 [multimap]: Above the declaration of class value_compare in the map synopsis, add:

template <class Key, class T, class Compare = less<Key>,
          class Allocator = allocator<pair<const Key, T> > >
class multimap {
public:
  // types:
  ...
  // exposition only.
  class value_compare
    : public binary_function<value_type,value_type,bool> {
    ...

1301(i). clear() and assignment

Section: 24.2.4 [sequence.reqmts] Status: NAD Editorial Submitter: Nicolai Josuttis Opened: 2010-01-01 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [sequence.reqmts].

View all other issues in [sequence.reqmts].

View all issues with NAD Editorial status.

Discussion:

I propose that clear() be defined to be equivalent to erase(begin(),end()) except not using copy or move of elements.

To: C++ libraries mailing list
Message c++std-lib-26465

and specifiying as post: size()==0 might also not be appropriate because forward-Lists provide no size(), this it should be: post: empty()==true

Bjarne Stroustrup schrieb/wrote:

To: C++ libraries mailing list
Message c++std-lib-26458

in table 94 we define clear() as:

a.clear() void erase(begin(), end())
post: size() == 0

Now erase requires assignment (MoveAssignable) which makes sense if we have to move an element, but why should that be required from clear() where all elements are destroyed?

[ 2010-01-23 Alisdiar provides wording. ]

[ 2010-01-30 Moved to Tentatively Ready after 5 positive votes on c++std-lib. ]

[ 2010-01-30 Daniel opens: ]

First, I read the newly proposed spec for clear() that it does in general not invalidate a previous past-the-end iterator value, but deque says in 24.3.8.4 [deque.modifiers] for the semantics of erase that erasures at the end will invalidate the past-the-end iterator. With removal of a direct binding between clear() and erase() there seem to be some fixes necessary. One way to fix that would be to mention in Table 94 that this "may also invalidate the past-the-end iterator" and then to mention for all specific containers where this does not happen, the exception, [1] e.g. in std::vector. std::vector has no own specification of clear() and one aspect of the closed issue 1102 was to realize just that (indirectly via erase). IMO we should now add an extra specification for clear(). Btw.: std::vector::erase reads to me that it would invalidate previous past-the-end values (and that seems correct in general).

Before I will provide explicit wording, I would like to discuss these points.

[1] std::list does fortunately specify that clear does not invalidate the past-the-end iterator.

[ 2010-02-08 Moved to Tentatively NAD Editorial after 5 positive votes on c++std-lib. Rationale added below. ]

Rationale:

Solved as proposed by LWG 704.

Proposed resolution:

Change 24.2.2.1 [container.requirements.general]/10:

Unless otherwise specified (see 23.2.4.1, 23.2.5.1, 23.3.2.3, and 23.3.6.4) all container types defined in this Clause meet the following additional requirements:

Replace the following words from Table 94 — Sequence container requirements (in addition to container) in 24.2.4 [sequence.reqmts]:

Table 94 — Sequence container requirements (in addition to container)
Expression Return type Assertion/note
pre-/post-condition
a.clear() void erase(begin(), end())
Destroys all elements in the container a. Invalidates all references, pointers, and iterators referring to the elements of a and may invalidate the past-the-end iterator.
post: size() == 0 a.empty() == true.

Add a new paragraph after [forwardlist.modifiers]/23:

void clear();

23 Effects: Erases all elements in the range [begin(),end()).

Remarks: Does not invalidate past-the-end iterators.


1302(i). different emplace semantics for sequence and associated containers

Section: 24.2.7 [associative.reqmts], 24.2.8 [unord.req] Status: NAD Submitter: Nicolai Josuttis Opened: 2010-01-03 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with NAD status.

Discussion:

According to the new naming scheme introduced with N2680

vector<T> v;
v.emplace(v.begin(),x,y,z)

now has a different semantics than

set<T> s;
s.emplace(s.begin(),x,y,z);

While the version for vectors takes the first argument as position and the remaining for construction, the version for sets takes all arguments for construction.

IMO, this is a serious design mistake for a couple of reasons:

I consider this to be a serious design penalty because once this is specified we can't fix that without breaking backward compatibility.

However, we have two choices for a fix:

[ 2010 Pittsburgh: Moved to NAD, rationale added below. ]

Rationale:

There was no consensus to make this change.

Proposed resolution:

In 24.2.8 [unord.req], change:

Table 96 — Associative container requirements (in addition to container)
expression Return type Assertion/note pre-/post-condition Post-condition
...
a_uniq.emplace_value(args) pair<iterator, bool> inserts a T object t constructed with std::forward<Args>(args)...
if and only if there is no element in the container with key equivalent to the key of t.
The bool component of the returned pair is true if and only if the insertion takes place, and the iterator component of the pair points to the element with key equivalent to the key of t.
logarithmic
a_eq.emplace_value(args) iterator inserts a T object t constructed with std::forward<Args>(args)... and returns the iterator pointing to the newly inserted element. logarithmic
a.emplace_hint(p,args) iterator equivalent to a.emplace_value(std::forward<Args>(args)...). Return value is an iterator pointing to the element with the key equivalent to the newly inserted element. The const_iterator p is a hint pointing to where the search should start. Implementations are permitted to ignore the hint. logarithmic in general, but amortized constant if the element is inserted right after p
...

In 24.2.8 [unord.req], change:

Table 98 — Unordered associative container requirements (in addition to container)
expression Return type Assertion/note pre-/post-condition Post-condition
...
a_uniq.emplace_value(args) pair<iterator, bool> inserts a T object t constructed with std::forward<Args>(args)... if and only if there is no element in the container with key equivalent to the key of t. The bool component of the returned pair is true if and only if the insertion takes place, and the iterator component of the pair points to the element with key equivalent to the key of t. Average case O(1), worst case O(a_uniq.size()).
a_eq.emplace_value(args) iterator inserts a T object t constructed with std::forward<Args>(args)... and returns the iterator pointing to the newly inserted element. Average case O(1), worst case O(a_eq.size()).
a.emplace_hint(p,args) iterator equivalent to a.emplace_value(std::forward<Args>(args)...). Return value is an iterator pointing to the element with the key equivalent to the newly inserted element. The const_iterator p is a hint pointing to where the search should start. Implementations are permitted to ignore the hint. Average case O(1), worst case O(a.size()).
...

In 24.4.4 [map], 24.4.6 [set], 24.5.4 [unord.map], 24.5.6 [unord.set], change:

// modifiers:
template <class... Args> pair<iterator, bool> emplace_value(Args&&... args);
template <class... Args> iterator emplace_hint(const_iterator position, Args&&... args);

In 24.4.5 [multimap], 24.4.7 [multiset], 24.5.5 [unord.multimap], 24.5.7 [unord.multiset], change:

// modifiers:
template <class... Args> iterator emplace_value(Args&&... args);
template <class... Args> iterator emplace_hint(const_iterator position, Args&&... args);


1308(i). Concerns about initializer_list overloads of min, max, and minmax

Section: 27.8.9 [alg.min.max] Status: NAD Submitter: Niels Dekker Opened: 2010-02-02 Last modified: 2017-09-07

Priority: Not Prioritized

View all other issues in [alg.min.max].

View all issues with NAD status.

Discussion:

In San Francisco, June 2008, N2722 was adopted, replacing the variadic templates min, max, and minmax by overloads that have an initializer_list<T> parameter. The paper showed benchmark results wherein initializer_list versions of min appeared to outperform the corresponding variadic template. Unfortunately, in October 2009 a very serious error was detected in the benchmark. (c++std-lib-25210). In fact, an initializer_list<T> version of min often appears to perform worse than the corresponding variadic template, especially when T has an expensive copy constructor (c++std-lib-25253, http://www.xs4all.nl/~nd/dekkerware/issues/n2772_fix).

IMO, the biggest problem of the initializer_list overloads is that they pass and return T objects by value. Which has the following consequences:

  1. They require that T is CopyConstructible. IMO that is too much of a constraint for a generic, general purpose function like std::min<T>.
  2. They potentially throw an exception, even if T's less-than-operator throws nothing. (And of course, less-than typically throws nothing.)
  3. They are inconsistent with C++03 std::min and std::max. Consider the subtle difference between const T& c1 = min(a,b); and const T& c2 = min({a,b}); (c++std-lib-25265)
  4. They do not conveniently support use cases that need to have a reference to the minimum or maximum object itself, rather than just a copy.
  5. They potentially perform badly: possibly O(n), when the arguments themselves have a size of n.

In the future, this problem might be solvable by using an initializer_list of const references, instead:

const T& min(initializer_list<const T&>);
const T& max(initializer_list<const T&>);
pair<const T&, const T&> minmax(initializer_list<const T&>);

It is unlikely that C++0x will support initializer_list<const T&>, but technically it seems possible to add such a language feature after C++0x (c++std-core-15428).

Variadic templates of min, max, and minmax, as proposed by N2551 (Sylvain Pion), do have some other advantages over initializer_list overloads:

  1. It is likely that those variadic templates can be declared constexpr, now that CWG issue #991 is in drafting status.
  2. They provide complete compile-time protection against accidentally passing zero arguments.

Unfortunately, the variadic templates of min, max, and minmax may still need further improvement, before having them in the Standard Library. Especially the optional Compare parameter appears to be a concern. So for this moment I recommend to keep both versions out of C++0x, and postpone further discussion until after C++0x.

[ 2010 Pittsburgh: Discussed and the LWG still prefers the initializer list solution of N2772. ]

Rationale:

We prefer the solution of N2772 which will be reapplied.

Proposed resolution:

Remove both variadic templates and initializer_list overloads of min, max, and minmax from the synopsis in 27.1 [algorithms.general] and from 27.8.9 [alg.min.max].

[ Note: This proposed resolution will resolve LWG 915 as NAD. ]


1313(i). Seed sequence's param function not useful for pure output iterator

Section: 28.5.8.1 [rand.util.seedseq] Status: NAD Submitter: Daniel Krügler Opened: 2010-02-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.util.seedseq].

View all issues with NAD status.

Discussion:

The Seed sequence requirements (28.5.3.2 [rand.req.seedseq]) require the existence of a member function

template<typename OutputIterator>
void param(OutputIterator ob);

The fact that this function returns void instead of the value of ob after accepting the sequence data leads to the same problem as in issue 865 - In case of pure output iterators there is no way to serialize further data into that data sink.

[ 2010-02-07 Howard adds: ]

At the time this issue was opened, the suggested changes are with respect to an anticipated draft which does not yet exist.

[ 2010 Pittsburgh: ]

No technical counterarguments, but it is simply too late in the process to make this change at this point.

Proposed resolution:

  1. In Table 109 — Seed sequence requirements, expression "r.param(ob)" change the
    Return type entry:

    voidOutputIterator
    
  2. In 28.5.8.1 [rand.util.seedseq], class seed_seq synopsis change

    template<class OutputIterator>
    voidOutputIterator param(OutputIterator dest) const;
    

1314(i). NULL and nullptr

Section: 17.2 [support.types] Status: NAD Submitter: Sean Hunt Opened: 2010-02-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [support.types].

View all issues with NAD status.

Discussion:

Currently, the 17.2 [support.types]/3 allows NULL to be any null pointer constant. The footnote marks that 0 or 0L might be appropriate. However, this definition also allows the implementation to define NULL to be nullptr. This may lead to overload and conversion issues more serious than with the C++98 version:

void f(void*);
void f(int);

void g()
{
 // calls f(int) if NULL is integral
 // calls f(void*) if NULL is nullptr
 f(NULL);
}

Possible resolutions:

Making NULL nullptr would improve code correctness, and breaking backwards compatibility shouldn't be a huge concern as NULL shouldn't be used except as a null pointer constant anyways.

[ 2010-02-10 Chris provided wording. ]

[ 2010 Pittsburgh: Moved to NAD, rationale added below. ]

Rationale:

The LWG discussed the proposed resolution and several other options. There was no concensus to make this or any other changes.

Proposed resolution:

17.2 [support.types]

3 The macro NULL is defined to be nullptr. is an implementation-defined C++ null pointer constant in this International Standard (4.10).196

196) Possible definitions include 0 and 0L, but not (void*)0.

20.2.12 [c.malloc]

7 The contents are the same as the Standard C library header <string.h>, with the change to memchr() specified in 21.6 and the macro NULL defined to be nullptr.

[date.time]

2 The contents are the same as the Standard C library header <time.h>.232 except the macro NULL, which is defined to be nullptr. The functions asctime, ctime, gmtime, and localtime are not required to avoid data races (17.6.4.8).

30.5 [c.locales]

2 The contents are the same as the Standard C library header <locale.h> except the macro NULL, which is defined to be nullptr.

C.7.3.7 [diff.null]

1 The macro NULL, defined in any of <clocale>, <cstddef>, <cstdio>, <cstdlib>, <cstring>, <ctime>, or <cwchar>, is nullptr an implementation-defined C++ null pointer constant in this International Standard (18.2).


1315(i). return type of async

Section: 33.10.9 [futures.async] Status: NAD Editorial Submitter: Jonathan Wakely Opened: 2009-02-09 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [futures.async].

View all other issues in [futures.async].

View all issues with NAD Editorial status.

Discussion:

Both overloads of async return future<typename F::result_type> which requires that F has a nested type. This prevents async being used with function pointers and makes the example in 33.10.9 [futures.async] invalid. I believe this is unintentional.

The proposed resolution also addresses editorial issues with the launch_policy function parameter.

For the first overload it is not sufficient to return future<typename result_of<F(ArgTypes...)>::type>. Calling async(launch::xxx, foo, bar) performs argument deduction on both async overloads, which for the first overload attempts to instantiate result_of<launch(F, ArgTypes...)>, which is invalid. SFINAE must be used to prevent that.

[ 2010-02-12 Moved to Tentatively Ready after 5 positive votes on c++std-lib. ]

[ 2010-02-12 Daniel opens: ]

[..] if decay<F>::type is of type std::launch.

or

[..] if remove_cv<remove_reference<F>::type>::type is of type std::launch.

The latter is the more specific form, but the former is equivalent to the latter for all cases that can occur here. I suggest to use the former for simplicity, but expect that implementations can effectively use the latter.

[ 2010-02-12 Moved to Tentatively Ready after 5 positive votes on c++std-lib. ]

[ 2010 Pittsburgh: ]

Moved to NAD Editorial. Rationale added below.

Rationale:

Solved by N3058.

Proposed resolution:

In 33.10.1 [futures.overview] paragraph 1:

template <class F, class... Args>
  future<typename F::result_type>
  future<typename result_of<F(Args...)>::type>
  async(F&& f, Args&&... args);
template <class F, class... Args>
  future<typename F::result_type>
  future<typename result_of<F(Args...)>::type>
  async(launch policy, F&& f, Args&&... args);

In 33.10.9 [futures.async] before paragraph 1

template <class F, class... Args>
  future<typename F::result_type>
  future<typename result_of<F(Args...)>::type>
  async(F&& f, Args&&... args);
template <class F, class... Args>
  future<typename F::result_type>
  future<typename result_of<F(Args...)>::type>
  async(launch policy, F&& f, Args&&... args);

...

Remarks: The first signature shall not participate in overload resolution if decay<F>::type is std::launch.


1317(i). make_hash

Section: 22.10.19 [unord.hash] Status: NAD Submitter: Nicolai M. Josuttis Opened: 2010-02-10 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [unord.hash].

View all issues with NAD status.

Discussion:

Currently, the library lacks a convenient way to provide a hash function that can be used with the provided unordered containers to allow the usage of non trivial element types.

While we can easily declare an

std::unordered_set<int>

or

std::unordered_set<std::string>

we have no easy way to declare an unordered_set for a user defined type. IMO, this is a big obstacle to use unordered containers in practice. Note that in Java, the wide usage of HashMap is based on the fact that there is always a default hash function provided.

Of course, a default hash function implies the risk to provide poor hash functions. But often even poor hash functions are good enough.

While I really would like to see a default hash function, I don't propose it here because this would probably introduce a discussion that's too big for this state of C++0x.

However, I strongly suggest at least to provide a convenience variadic template function make_hash<>() to allow an easy definition of a (possibly poor) hash function.

As a consequence for a user-defined type such as

class Customer {
   friend class CustomerHash;
   private:
     string firstname;
     string lastname;
     long   no;
   ...
 };

would allow to specify:

class CustomerHash : public std::unary_function<Customer, std::size_t>
{
  public:
    std::size_t operator() (const Customer& c) const  {
       return make_hash(c.firstname,c.lastname,c.no);
    }
};

instead of:

class CustomerHash : public std::unary_function<Customer, std::size_t>
{
  public:
    std::size_t operator() (const Customer& c) const  {
       return std::hash<std::string>()(c.firstname) +
              std::hash<std::string>()(c.lastname) +
              std::hash<long>()(c.no);
    }
};

Note that, in principle, we can either specify that

make_hash returns the sum of a call of std::hash<T>()(x) for each argument x of type T

or we can specify that

make_hash provides a hash value for each argument, for which a std::hash() function is provided

with the possible note that the hash value may be poor or only a good hash value if the ranges of all passed arguments is equally distributed.

For my convenience, I propose wording that describes the concrete implementation.

[ 2010 Pittsburgh: Moved to NAD Editorial, rationale added below. ]

[LEWG Kona 2017]

Recommend NAD: Feature? Needs a paper. (This is LEWG21)

Rationale:

There is no consensus to make this change at this time.

Proposed resolution:

In Function objects 22.10 [function.objects] in paragraph 2 at the end of the Header <functional> synopsis insert:

// convenience functions
template <class T>
  size_t make_hash (const T&);
template <class T, class... Types>
  size_t make_hash (const T&, const Types&...);

In Class template hash 22.10.19 [unord.hash] add:

20.7.16.1 Hash creation functions [hash.creation]

template <class T>
  size_t make_hash (const T& val);

Returns: hash<T>()(val);

template <class T, class... Types>
  size_t make_hash (const T& val, const Types&... args);

Returns: hash<T>()(val) + std::make_hash(args...)


1318(i). N2982 removes previous allocator capabilities

Section: 20.2.9.2 [allocator.traits.types] Status: NAD Submitter: Pete Becker Opened: 2010-02-11 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [allocator.traits.types].

View all issues with NAD status.

Duplicate of: 1375

Discussion:

Addresses US-87

N2982 says that containers should have a nested typedef that defines their reference_type as value_type&; the previous standard deferred to the allocator to define its reference_type, and containers simply passed the allocator's typedef on. This change is a mistake. Allocators should define both a pointer type and a reference type. That's essential for their original purpose, which was to make different memory models transparent. If an allocator defines a pointer type that isn't compatible with a normal pointer it also has to define a corresponding reference type. For example (and please forgive a Windows-ism), if an allocator's pointer is T __far*, then it's reference has to be T __far&. Otherwise everything crashes (under the hood, references are pointers and have to have the same memory access mechanics). Extensions such as this for more general memory models were explicitly encouraged by C++03, and the allocator's pointer and reference typedefs were the hooks for such extensions. Removing the allocator's reference and const_reference typedefs makes those extensions unimplementable and breaks existing implementations that rely on those hooks.

[ 2010-02-25 Alisdair adds: ]

vector<bool>::reference is a nested class, and not a typedef. It should be removed from the list of containers when this change is made.

In general, I am uncomfortable placing this reference requirement on each container, as I would prefer to require:

is_same<Container::reference, Container::iterator::reference>

This distinction is important, if we intend to support proxy iterators. The iterator paper in the pre-Pittsburgh mailing (N3046) does not make this proposal, but organises clause 24 in such a way this will be much easier to specify.

The changes to clause 20 remain important for all the reasons Pete highlights.

[ 2010 Batavia ]

Removed vector from list of templates that should be adjusted as of meeting outcome.

[ 2010 post-Batavia ]

Replaced vector<bool> reference by vector reference because of misinterpreting meeting typo. Additional corrected numbering in P/R to N3225 wording.

[ 2010-12-06 Daniel reopens ]

Unfortunately, the current P/R is defective for several reasons:

  1. Table 43 — Descriptive variable definitions still contains three references to T&, namely in:
    t a value of type const T&
    r a value of type T& obtained by the expression *p
    s a value of type const T& obtained by the expression *q or by conversion from a value r
    Especially the second and third items are misses in the 1318 P/R, e.g. in N2723 or in C++03 these were referring to X::reference and X::const_reference, resp. None of them is referenced anywhere in the allocator requirements table: r and s where historically needed to define the expressions a.address(r) and a.address(s) which are gone now, and t was needed to define the expression a.construct(p, t) which has been replaced by a.construct(p,args).

    The easiest fix seems to be to remove all three rows from Table 43.

  2. Further-on, the current P/R suggests to replace the the current definitions of the adaptor classes
    stack
    priority_queue
    queue
    
    similar to the other container types, i.e. to define reference and const_reference now as
    typedef typename allocator_traits<Allocator>::reference reference;
    typedef typename allocator_traits<Allocator>::const_reference const_reference;
    
    This would not only be an ill-formed definition (because there is no name Allocator in scope), but it would also introduce a breakage compared to C++03, where these definitions where already referring to the definition of the wrapped containers. So, the adaptor class templates should be removed from the current list.
  3. Then the current P/R wording leads to one further unexpected and unwanted change due to the general formular used: match_result::reference is currently defined as
    typedef const_reference reference;
    
    because it is an immutable container (And we had this definition already in N2723). The application of the rule would change this silently.
  4. Finally the suggested wording for the unordered_ containers is incomplete. The reason is a current inconsistency between these containers and the rest: While normally the definition of the pointer types is
    typedef typename allocator_traits<Allocator>::pointer pointer;
    typedef typename allocator_traits<Allocator>::const_pointer const_pointer;
    
    for the unordered containers they are
    typedef typename allocator_type::pointer pointer;
    typedef typename allocator_type::const_pointer const_pointer;
    
    These definitions are not equivalent, because allocators are no longer required to define typedefs pointer and const_pointer, the allocator_traits were invented as a further indirection to cope with that. I.e. for the unordered containers we need to bring both the definition of references and pointers in sync.

[ 2011-02-23 Daniel updates the proposed wording with support from Pablo ]

The update attempts to fix the backward-compatibility problem that we have introduced by ignoring the C++03 member function overloads address of allocator types in C++0x completely. The resolution attempts to fix that by adding these functions as optional members of allocators that are considered first before falling back to pointer_traits::pointer_to. This still allows us to remove the unused symbol t from the table, but we adapt the symbols r and s to purely refer to the typenames reference and const_reference.

[2011-03-06 Daniel adapts numbering to N3242]

[2011-03-11 Daniel removes noexcept specifiers from address functions]

[2011-03-12 Further wording improvements by Daniel and Pablo]

[2011-03-22 Madrid]

Closed as NAD, no consensus to make a change

Rationale:

No consensus to make a change

Proposed resolution:


1320(i). Header for iter_swap

Section: 25.2 [iterator.synopsis] Status: NAD Submitter: Alisdair Meredith Opened: 2010-02-16 Last modified: 2017-06-05

Priority: Not Prioritized

View all other issues in [iterator.synopsis].

View all issues with NAD status.

Discussion:

The iter_swap function template appears in the <algorithm> header, yet its main use is in building further algorithms, not calling existing ones. The main clients are implementers of data structures and their iterators, so it seems most appropriate to place the template in the <iterator> header instead.

Note that this is not an issue for implementers of the standard library, as they rarely use the standard headers directly, designing a more fine-grained set of headers for their own internal use. This option is not available to customers of the standard library.

Note that we cannot remove iter_swap from <algorithm> without breaking code, but there is no reason we cannot offer the same declaration via two standard headers. Alternatively, require <algorithm> to #include <iterator>, but introducing the dependency on the iterator adaptors seems un-necessary.

[ ]

Discussed possibly moving to <utility> but don't like that. Some not seeing this as a defect, and want to keep it in <algorithm>. No one seems to feel strongly about moving to <iterator>.

[2017-02 in Kona, LEWG recommends NAD]

[2017-06-02 Issues Telecon]

Resolve as NAD

Proposed resolution:

Add the declaration of iter_swap to the <iterator> header synopsis (25.2 [iterator.synopsis]), with a note that it is documented in clause 27 [algorithms].

...
template <class T, size_t N> T* end(T (&array)[N]);

// documented in 27 [algorithms]
template<class ForwardIterator1, class ForwardIterator2>
  void iter_swap(ForwardIterator1 a, ForwardIterator2 b);

1330(i). Move container requirements into requirements tables

Section: 24.2 [container.requirements] Status: NAD Submitter: Nicolai Josuttis Opened: 2010-03-10 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [container.requirements].

View all issues with NAD status.

Discussion:

Abstract:

In general, it seems that in a couple of places container behavior is not described in requirement tables although it is a general behavior.

History:

Issue 676 added move semantics to unordered containers. For the added insert functions the Editor requested to put their semantic description into a requirements table rather than describing them for each container individually. The text however was taken from the associative containers, where we also have the semantics for each container described. Also, 1034 is to some extend requesting a clarification of the requirement tables and it turned out that in other places we have the same problem (e.g. we have no general requirement for type pointer and const_pointer although each container has them with issue 1306).

From my personal list of functions in requirement tables and containers, the following types/functions are missing in requirement tables:

As a special case, we lack the following requirements for all sequence containers BUT array (so special wording or a new container category is required):

Note that we also might have to add additional requirements on other places for sequence containers because having an allocator requires additional statements for the treatment of the allocators. E.g. swap for containers with allocators is not specified in any requirement table.

And finally, if we have the requirements in the requirements tables, we can remove the corresponding descriptions for the individual container. However, note that sequence container requirements have NO complexity column, so that we still need container specific descriptions for the functions listed there.

[ 2010 Batavia ]

While there is consensus that further cleaning up the container requirement tables would be a good thing, there is no feeling that this must be done in time for 0x. The issue remains open, but Deferred.

[ 2011 Bloomington ]

Closes as NAD. There are a number of deficiencies in the way the container requirements tables are presented, and the LWG welcomes further papers that will help clear up this presentation.

Proposed resolution:


1331(i). incorporate move special member functions into library

Section: 16 [library] Status: NAD Submitter: Martin Sebor Opened: 2010-03-10 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Review the library portion of the spec and incorporate the newly added core feature Move Special Member Functions (N3044).

Rationale:

2010 Batavia: This has now been done to a large extent.

Proposed resolution:


1348(i). Exception safety of unspecified types

Section: 16 [library] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Addresses GB-64

There are a number of unspecified types used throughout the library, such as the container iterators. Many of these unspecified types have restrictions or expectations on their behaviour in terms of exceptions. Are they permitted or required to use exception specifications, more specifically the new noexcept specification? For example, if vector<T>::iterator is implemented as a native pointer, all its operations will have an (effective) noexcept specification. If the implementation uses a class type to implement this iterator, is it permitted or required to support that same guarantee?

[ Resolution proposed by ballot comment ]

Clearly state the requirements for exception specifications on all unspecified library types. For example, all container iterator operations should be conditionally noexcept, with the condition matching the same operation applied to the allocator's pointer_type, a certain subset of which are already required not to throw.

[2011-03-24 Madrid meeting]

Discussion:

Alisdair: Probably an NAD Future at least but could be NAD.

Daniel K: Approach is not consistent with what we've decided with noexcept.

Alisdair: Any objection to marking 1348 as NAD?

No objections.

Rationale:

Standard is correct as written

Proposed resolution:


1350(i). Implicit contructors accidentally made some library types move-only

Section: 16 [library] Status: Dup Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with Dup status.

Duplicate of: 1421

Discussion:

Addresses CH-15

Due to the new rules about implicit copy and move constructors some library facilities are now move-only.

[ Resolution proposed by ballot comment ]

Make them copyable again.

Proposed resolution:


1351(i). Replace dynamic exception specifications with noexcept

Section: 16 [library] Status: Dup Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with Dup status.

Duplicate of: 1344

Discussion:

Addresses CH-16

Dynamic exception specifications are deprecated. Deprecated features shouldn't be used in the Standard.

[ Resolution proposed by ballot comment ]

Replace dynamic exception specifications with noexcept.

Proposed resolution:


1352(i). Apply noexcept where library specification says "Throws: Nothing"

Section: 16 [library] Status: Dup Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with Dup status.

Duplicate of: 1346

Discussion:

Addresses CH-17

The introduction of noexcept makes "Throws: Nothing" clauses looking strange.

[ Resolution proposed by ballot comment ]

Consider replacing "Throws: Nothing." clause by the respective noexcept specification.

Proposed resolution:


1358(i). Add <chrono> and <ratio> to freestanding implementations

Section: 16.4.2.5 [compliance] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [compliance].

View all issues with NAD status.

Discussion:

Addresses GB-55

The <thread> header uses duration types, found in the <chrono> header, and which rely on the ratio types declared in the <ratio> header.

[ Extracts from lengthy Rapperswil discussion: ]

There is a concern that this issue is a misunderstanding of the actual requirements of a free-standing implementation to support the <thread> header. In general, a free-standanding implementation will provide an empty header, specifically so that a user can test for the absence of the _ _ STDCPP_THREADS _ _ macro. This idiom as used as there is no portable way to test for the lack of a header.

At this point, it was suggested the NB comment is trying to solve the wrong problem, and that _ _ STDCPP_THREADS _ _ should be a pre-defined macro in clause 16 that can be tested before including <thread>. That would remove the need to add additional headers to the free-standanding requirements.

It is worth noting that Japan requested <ratio> as a free-standing header in their CD1 comments. No-one seemed keen to require clocks of a free-standing implementation though.

Detlef volunteers to look at a way to redraft 17.6.1.3 p3.

[ Original resolution proposed by NB comment: ]

Add the <chrono> and <ratio> headers to the freestanding requirements.

It might be necessary to address scaled-down expectations of clock support in a freestanding environment, much like <thread>.

[2011-02-25: Alberto drafts wording]

[2011-03-06: Daniel observes:]

Accepting the proposal n3256 would solve this issue.

[2011-03-24 Madrid meeting]

Freestanding no longer requires <thread> header

Rationale:

We are not adding new headers to freestanding at this point.

Proposed resolution:

  1. Add a new entry in Table 14 — C++ library headers:

    Table 14 — C++ library headers
    <iterator>
    <library_support>
    <limits>
  2. Remove the last row 33.4 [thread.threads] <threads> from Table 16 — C++ headers for freestanding implementations and insert a new one instead (To the editor: For the actual target Clause please see the comment in bullet 5 of this proposed resolution):

    Table 16 — C++ headers for freestanding implementations
    Subclause Header(s)
    33.4 [thread.threads] Threads <thread>
    ?? Library support <library_support>
  3. Modify paragraph 16.4.2.5 [compliance] p. 3:

    3 The supplied version of the header <cstdlib> shall declare at least the functions abort, atexit, at_quick_exit, exit, and quick_exit (18.5). The supplied version of the header <thread> shall meet the same requirements as for a hosted implementation or including it shall have no effect. The other headers listed in this table shall meet the same requirements as for a hosted implementation. A program can detect the presence of standard headers not listed in Table 16 using the facilities provided by the <library_support> header.

  4. Remove the following line from the header <thread> synopsis in 33.4 [thread.threads] p. 1:

    namespace std {
      #define __STDCPP_THREADS__ __cplusplus
    
      class thread;
      [...]
    }
    
  5. Add a new section in Clause 18 or 20 (or any other suitable place at the editor's discretion):

    ?? Library support [library.support]

    The header <library_support> defines an implementation-defined set of macros to allow a program detect the presence of standard headers in freestanding implementations. [Note: Hosted implementations shall provide all standard headers, thus shall provide all macros. — end note]

    For each standard header listed in Tables 14 (C++ library headers) and 15 (C++ headers for C library facilities) that is provided by the implementation, <library_support> shall define a macro with name _ _HAS_XXX_HEADER_ _ where XXX is replaced by the uppercase version of the name of the header. Each such macro shall expand to the value _ _cplusplus. [Example:

    #include <library_support>
    
    #ifdef _ _HAS_THREADS_HEADER_ _
      #include <threads>
      // code that exploit the presence of threads
    #else
      // fallback code that doesn't rely on threads
    #endif
    

    end example]

    No other standard header shall define macros with a name beginning with _ _HAS_ and ending with _HEADER_ _.


1359(i). Add <tuple> and <utility> to freestanding implementations

Section: 16.4.2.5 [compliance] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [compliance].

View all issues with NAD status.

Discussion:

Addresses GB-56

The <utility> header provides support for several important C++ idioms with move, forward and swap. Likewise, declval will be frequently used like a type trait. In order to complete cycles introduced by std::pair, the <tuple> header should also be made available. This is a similarly primitive set of functionality, with no dependency of a hosted environment, but does go beyond the minimal set of functionality otherwise suggested by the freestanding libraries.

Alternatively, split the move/forward/swap/declval functions out of <utility> and into a new primitive header, requiring only that of freestanding implementation.

[ Summary of Rapperswil discusions ]

The preference of the meeting was to extract the rvalue-reference related utilities and swap into a freestanding header, but there was no clear preference for a name. Howard suggested simply dropping them into <type_traits> as both these utilities and type traits are used pretty much everywhere in the library implementation, it is the most convenient place to keep them (from an implementer's perspective).

Poll: Two-way: New header for forward, move, swap, move_with_noexcept and declval vs. calling out forward, move, swap, move_with_noexcept and declval as freestanding explicitly? SF new header: 4 WF new header: 3 WF call out as freestanding: 1 SF call out as freestanding: 2 Alisdair: Willing to write up both solutions, give us some time to think on it. Action: Need an issue and proposed wording for GB 56 - Alisdair to draft both options as in the last poll.

[ Resolution proposed by ballot comment ]

Add <utility> and <tuple> to table 15, headers required for a free-standing implementation.

[ 2010-Batavia: ]

Closed as NAD, reversing the decision at Rapperswil.

The consensus was that any freestanding implementation is going to feel compelled to offer the important features of <utility> even if we do not make them a freestanding requirement; breaking out additional small headers may have additional costs at compile time, and while the critical move-related functions could migrate to <type_traits>, the header name is far from appealing; adding the whole of <utility> starts to drag in dependencies on <tuple> and <memory>, so we prefer to place the burden of slicing or supporting this whole header on free-standing vendors.

Proposed resolution:

Rationale:

No consensus for a change at this time.


1361(i). Does use of std::size_t in a header imply that typedef name is available to users?

Section: 16.4.3 [using] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses GB-58

It is not clear whether a library header specified in terms of a typedef name makes that same typedef name available for use, or if it simply requires that the specified type is an alias of the same type, and so the typedef name cannot be used without including the specific header that defines it. For example, is the following code required to be accepted:

#include <vector>
std::size_t x = 0;

Most often, this question concerns the typedefs defined in header <cstddef>

[ Resolution proposed by ballot comment: ]

Add a paragraph under 16.4.3 [using] clarifying whether or not headers specified in terms of std::size_t can be used to access the typedef size_t, or whether the header <cstddef> must be included to reliably use this name.

[Batavia: NAD - see rationale below]

Proposed resolution:

Rationale:

The standard is correct as written.


1369(i). rethrow_exception may introduce data races

Section: 17.9.7 [propagation] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [propagation].

View all issues with NAD status.

Discussion:

Addresses GB-74

One idea for the exception_ptr type was that a reference-counted implementation could simply 'reactivate' the same exception object in the context of a call to rethrow_exception. Such an implementation would allow the same exception object to be active in multiple threads (such as when multiple threads join on a shared_future) and introduce potential data races in any exception handler that catches exceptions by reference - notably existing library code written before this capability was added. rethrow_exception should always make a copy of the target exception object.

[Resolution suggested by NB comment]

Add the following to 18.8.5, [propogation]

Throws: a copy of the exception object to which p refers.

[2011-03-15: Anthony drafts wording]

[2011-03-23 Madrid meeting]

No consensus for a change

Rationale:

It would break too many existing implementations

Proposed resolution:

  1. Alter 14.2 [except.throw] p. 5 as follows:

    5 When the thrown object is a class object, the copy/move constructor and the destructor shall be accessible, even if the copy/move operation is elided ( [class.copy]). The copy constructor shall be accessible, and is odr-used (6.3 [basic.def.odr]), even if the copy operation is elided, or a move constructor used to construct the exception object.

  2. Alter 17.9.7 [propagation] p. 7 as follows:

    exception_ptr current_exception() noexcept;
    

    7 Returns: An exception_ptr object that refers to the currently handled exception (14.4 [except.handle]) or a copy of the currently handled exception, or a null exception_ptr object if no exception is being handled. The referenced object shall remain valid at least as long as there is an exception_ptr object that refers to it. If the function needs to allocate memory and the attempt fails, it returns an exception_ptr object that refers to an instance of bad_alloc. It is unspecified whether the return values of two successive calls to current_exception refer to the same exception object. [ Note: That is, it is unspecified whether current_exception creates a new copy each time it is called. — end note ] If the attempt to copy the current exception object throws an exception, or is otherwise not possible, the function returns an exception_ptr object that refers to the thrown exception if any or, if this is not possible, to an instance of bad_exception. [ Note: The copy constructor of the thrown exception may also fail, so the implementation is allowed to substitute a bad_exception object to avoid infinite recursion. — end note ]

  3. Alter 17.9.7 [propagation] p. 9 and add a new paragraph after p. 9 as follows:

    [[noreturn]] void rethrow_exception(exception_ptr p);
    

    8 Requires: p shall not be a null pointer.

    9 Throws: a copy of the exception object to which p refers, or any exception thrown by the attempt to copy the exception object to which p refers.

    ? Synchronization: Calls to rethrow_exception on exception_ptr objects that refer to the same exception object shall appear to occur in a single total order. The completion of each call shall synchronize with (6.9.2 [intro.multithread]) the next call in that total order.


1371(i). Standard exceptions require stronger no-throw guarantees

Section: 19 [diagnostics] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses GB-75

None of the exception types defined in clause 19 are allowed to throw an exception on copy or move operations, but there is no clear specification that the operations have an exception specification to prove it. Note that the implicitly declared constructors, taking the exception specification from their base class (ultimately std::exception) will implicitly generate a noexcept exception specification if all of their data members similarly declare noexcept operations. As the representation is unspecified, we cannot assume nonthrowing operations unless we explicitly state this as a constraint on the implementation.

[ Resolution proposed by ballot comment: ]

Add a global guarantee that all exception types defined in clause 19 that rely on implicitly declared operations have a non-throwing exception specification on those operations.

[ 2010 Batavia: ]

This is addressed by the current words in 18.8.1 [exception], p2

Each standard library class T that derives from class exception shall have a publicly accessible copy constructor and a publicly accessible copy assignment operator that do not exit with an exception.

Proposed resolution:


1373(i). Customizable traits should have their own headers

Section: 22.2 [utility] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [utility].

View all issues with NAD status.

Discussion:

Addresses GB-79

The library provides several traits mechanisms intended a customization points for users. Typically, they are declared in headers that are growing quite large. This is not a problem for standard library vendors, who can manage their internal file structure to avoid large dependencies, but can be a problem for end users who have no option but to include these large headers.

[ 2010 Rapperswil ]

There was no enthusiasm for touching char_traits or regex_traits. Consensus to move iterator_traits, allocator_traits and pointer_traits to their own respective headers once wording supplied.

[ 2010 Rapperswil ]

After some discussion, consensus is that moving these features into separate headers does not buy much in practice, as the larger headers will inevitably be included anyway. Resolve as NAD.

[ Resolution proposed in ballot comment ]

Move the following traits classes into their own headers, and require the existing header to #include the traits header to support backwards compatibility:

iterator_traits (plus the iterator tag-types)
allocator_traits
pointer_traits
char_traits
regex_traits

[ 2010 Batavia: ]

Closed as NAD with the rationale below.

Rationale:

This suggest is not a defect, as the likely benefit is small, if any, compared to the cost of not just implementating the feature, but also explaining/teaching it.

Proposed resolution:


1374(i). Clarify moved-from objects are "toxic"

Section: 16.4.4.2 [utility.arg.requirements] Status: NAD Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [utility.arg.requirements].

View all issues with NAD status.

Discussion:

Addresses US-85

20.2.1 Table 34 "MoveConstructible requirements" says "Note: rv remains a valid object. Its state is unspecified". Some components give stronger guarantees. For example, moved-from shared_ptrs are guaranteed empty (20.9.11.2.1/25). In general, what the standard really should say (preferably as a global blanket statement) is that moved-from objects can be destroyed and can be the destination of an assignment. Anything else is radioactive. For example, containers can be "emptier than empty". This needs to be explicit and required generally.

Note: The last time that one of us mentioned "emptier than empty" (i.e. containers missing sentinel nodes, etc.) the objection was that containers can store sentinel nodes inside themselves in order to avoid dynamically allocating them. This is unacceptable because

(a) it forces existing implementations (i.e. Dinkumware's, Microsoft's, IBM's, etc.) to change for no good reason (i.e. permitting more operations on moved-from objects), and

(b) it invalidates end-iterators when swapping containers. (The Working Paper currently permits end-iterator invalidation, which we consider to be wrong, but that's a separate argument. In any event, mandating end-iterator invalidation is very different from permitting it.)

[ Resolution proposed in ballot comment ]

State as a general requirement that moved-from objects can be destroyed and can be the destination of an assignment. Any other use is undefined behavior.

Proposed resolution:

Resolved by N3241


1375(i). reference_type should not have been removed from the allocator requirements

Section: 16.4.4.6 [allocator.requirements] Status: Dup Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with Dup status.

Duplicate of: 1318

Discussion:

Addresses US-87

reference_type should not have been removed from the allocator requirements. Even if it is always the same as value_type&, it is an important customization point for extensions and future features.

Proposed resolution:

In [allocator.requirements] Table 42 - Allocotor Requirements, Add a row (after value_type) with columns:

Expression: X::reference_type
Return type: T&
Assertion/note...: (empty)
Default: T&

[allocator.traits]:

namespace std {
  template <class Alloc> struct allocator_traits {
    typedef Alloc allocator_type;
    
    typedef typename Alloc::value_type value_type;

    typedef see below   pointer;
    typedef see below   const_pointer;
    typedef see below   void_pointer;
    typedef see below   const_void_pointer;
    typedef value_type& reference_type;

Add reference_type to allocator_traits template, defaulted to value_type&.


1376(i). Allocator interface is not backward compatible

Section: 16.4.4.6 [allocator.requirements] Status: NAD Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with NAD status.

Discussion:

Addresses US-88

Allocator interface is not backward compatible.

[ Resolution proposed by ballot comment ]

See Appendix 1 - Additional Details

[ 2010-10-24 Daniel adds: ]

n3165 provides an alternative resolution.

[ 2910 Batavia: ]

Closed as NAD - withdrawn by the submitter.

Proposed resolution:

See n3165

Rationale:

Withdrawn by the submitter.


1395(i). Inconsistent reference links should be unified

Section: 21.3.7 [meta.rel] Status: NAD Editorial Submitter: Japan Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [meta.rel].

View all other issues in [meta.rel].

View all issues with NAD Editorial status.

Discussion:

Addresses JP-32

Representations of reference link are not unified. Most reference links to clause (table) number, say X, are in the form "Clause X" ("Table X") capitalized, and subsection Y.Y.Y is referenced with its number only in the form "Y.Y.Y". Whether they are parenthesized or not depends on the context. However there are some notations "(Z)" consisting of only a number Z in parentheses to confer Clause or Table number Z.

Proposed resolution:

Change "(10)" to "(Clause 10)".


1396(i). regex should support allocators

Section: 32.7 [re.regex] Status: NAD Submitter: INCITS Opened: 2010-08-25 Last modified: 2020-11-09

Priority: Not Prioritized

View all other issues in [re.regex].

View all issues with NAD status.

Duplicate of: 1451

Discussion:

Addresses US-104, US-141

std::basic_regex should have an allocator for all the reasons that a std::string does. For example, I can use boost::interprocess to put a string or vector in shared memory, but not a regex.

[ Resolution proposed by ballot comment ]

Add allocators to regexes

[ 2010-10-24 Daniel adds: ]

Accepting n3171 would solve this issue.

[2011-03-22 Madrid]

Close 1396 as NAD Future.

[LEWG Kona 2017]

Recommend Open: Covered in LEWG9, P0269, which is in wording review.

[2020-07-17; status changed to Tentatively NAD in issue processing telecon]

P0269R0 has been superseded by P1294R0. Requested LEWG to confirm which version they forwarded to LWG. In any case, this is a feature request, not a defect, and will be dealt with as a proposal not an issue.

[2020-11-09 Status changed: Tentatively NAD → NAD.]

Rationale:

No consensus for a change at this time

Proposed resolution:


1398(i). Users should be able to specialize functors without depending on whole <functional> header

Section: 22.10 [function.objects] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [function.objects].

View all issues with NAD status.

Discussion:

Addresses GB-96

The function templates hash, less and equal_to are important customization points for user-defined types to be supported by several standard containers. These are accessed through the <functional> header which has grown significantly larger in C++0x, exposing many more facilities than a user is likely to need through there own header, simply to declare the necessary specialization. There should be a smaller header available for users to make the necessary customization.

[ Resolution proposed by ballot comment ]

Provide a tiny forwarding header for important functor types in the <functional> header that a user may want to specialize. This should contain the template declaration for equal_to, hash and less.

[ Rapperswill summary ]

Alisdair: Would recommend NAD unless someone takes the issue.

Daniel: Volunteers to write a paper for this.

[ 2010-11-07 Daniel provides a paper available on the Batavia document list ]

[ 2010 Batavia: ]

Closed as NAD - the consensus was that forwarding headers such as <iosfwd> do not bring the expected benefits, and are not widely used (to the surprise of some active users in the room!). Without real experience reporting a benefit, there is no further interest in pursuing this issue as an extension - hence NAD rather than NAD Future.

Rationale:

No consensus to make a change

Proposed resolution:

See paper "Forwarding <functional> functor templates" on the Batavia LWG document list


1406(i). Support hashing smart-pointers based on owner

Section: 20.3.2.2 [util.smartptr.shared] Status: NAD Submitter: Japan Opened: 2010-08-25 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [util.smartptr.shared].

View all issues with NAD status.

Discussion:

Addresses JP-5

Hash support based on ownership sharing should be supplied for shared_ptr and weak_ptr. For two shared_ptr objects p and q, two distinct equivalence relations can be defined. One is based on equivalence of pointer values, which is derived from the expression p.get() == q.get() (hereafter called address based equivalence relation), the other is based on equivalence of ownership sharing, which is derived from the expression !p.owner_before(q) && !q.owner_before(p) (hereafter called ownership-based equivalence relation). These two equivalence relations are independent in general. For example, a shared_ptr object created by the constructor of the signature shared_ptr(shared_ptr<U> const &, T *) could reveal a difference between these two relations. Therefore, hash support based on each equivalence relation should be supplied for shared_ptr. However, while the standard library provides the hash support for address-based one (20.9.11.6 paragraph 2), it lacks the hash support for ownership-based one. In addition, associative containers work well in combination with the shared_ptr's ownership-based comparison but unordered associative containers don't. This is inconsistent.

For the case of weak_ptr, hash support for the ownership based equivalence relation can be safely defined on weak_ptrs, and even on expired ones. The absence of hash support for the ownership-based equivalence relation is fatal, especially for expired weak_ptrs. And the absence of such hash support precludes some quite effective use-cases, e.g. erasing the unordered_map entry of an expired weak_ptr key from a customized deleter supplied to shared_ptrs.

Hash support for the ownership-based equivalence relation cannot be provided by any user-defined manner because information about ownership sharing is not available to users at all. Therefore, the only way to provide ownership-based hash support is to offer it intrusively by the standard library.

As far as we know, such hash support is implementable. Typical implementation of such hash function could return the hash value of the pointer of the counter object that is internally managed by shared_ptr and weak_ptr.

[2010 Rapperswil:]

No consensus to make this change at this time.

[LEWG Kona 2017]

Recommend NAD: Needs a paper. Feature. Exposing an implementation detail (indirectly observable via hash).

Proposed resolution:

Add the following non-static member functions to shared_ptr and weak_ptr class template;

Update [util.smartptr.shared], 20.9.11.2 paragraph 1

namespace std{
template<class T> class shared_ptr {
public:
...
  size_t owner_hash() const;
...
};
}

Update [util.smartptr.weak], 20.9.11.3 paragraph 1

namespace std{
template<class T> class weak_ptr {
public:
...
  size_t owner_hash() const;
...
};
}

These functions satisfy the following requirements. Let p and q be objects of either shared_ptr or weak_ptr, H be a hypothetical function object type that satisfies the hash requirements ([hash.requirements], 20.2.4) and h be an object of the type H. The expression p.owner_hash() behaves as if it were equivalent to the expression h(p). In addition, h(p) == h(q) must become true if p and q share ownership.


1411(i). Add a compile-time flag to detect monotonic_clock

Section: 99 [time.clock.monotonic] Status: Dup Submitter: DIN Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [time.clock.monotonic].

View all issues with Dup status.

Duplicate of: 1410

Discussion:

Addresses DE-20

The library component monotonic_clock is conditionally supported, but no compile-time flag exists that allows user-code to query its existence. Further-on there exist no portable means to simulate such a query. (To do so, user code would be required to add types to namespace std::chrono.)

Proposed resolution:

Provide a compile-time flag (preferably a macro) that can be used to query the existence of monotonic_clock.


1413(i). Specify whether high_resolution_clock is a distinct type or a typedef

Section: 29.7.8 [time.clock.hires] Status: NAD Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses US-112

What it means for high_resolution_clock to be a synonym is undefined. If it may or may not be a typedef, then certain classes of programs become unportable.

[ Resolution proposed in ballot comment ]

Require that it be a distinct class type.

[ 2010 Batavia ]

This is not a defect. Threre are a number of places in the standard where we allow implentations to choose their preferred technique, the most obvious example being the iterator/const_iterator types of set.

Typically, this means it is not portable to declare function overloads that differ only in their use of these types.

Proposed resolution:


1415(i). Iterator stability bans the short-string optimization

Section: 24.2.2.1 [container.requirements.general] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [container.requirements.general].

View all other issues in [container.requirements.general].

View all issues with NAD Editorial status.

Discussion:

Requirements on iterators swapping allegiance would disallow the small-string optimization.

[ Resolved in Rapperswil by paper N3108. ]

Proposed resolution:

Add an exclusion for basic_string to the sentence beginning “Every iterator referring to an element...”. Add a sentence to 21.4.6.8/2 saying that iterators and references to string elements remain valid, but it is not specified whether they refer to the same string or the other string.


1419(i). forward_list::erase_after should return an iterator

Section: 24.3.9 [forward.list] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2023-02-07

Priority: Not Prioritized

View all other issues in [forward.list].

View all issues with NAD Editorial status.

Discussion:

Addresses US-117

forward_list::erase_after should return an iterator.

[ Resolved in Rapperswil by a motion to directly apply the words from the ballot comment in N3102. ]

Proposed resolution:

See Appendix 1 - Additional Details


1433(i). random_shuffle and shuffle should have consistent signatures

Section: 27.7.13 [alg.random.shuffle] Status: Dup Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.random.shuffle].

View all issues with Dup status.

Duplicate of: 1432

Discussion:

Addresses GB-119

The functions random_shuffle and shuffle both take arguments providing a source of randomness, but one take its argument by rvalue reference, and the other requires an lvalue reference. The technical merits of which form of argument passing should be settled for this specific case, and a single preferred form used consistently.

Proposed resolution:

[DEPENDS ON WHETHER RVALUE OR LVALUE REFERENCE IS THE PREFERRED FORM]


1434(i). For min/max functions replace variadic arguments by initializer_list argument

Section: 27.8.9 [alg.min.max] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [alg.min.max].

View all issues with NAD Editorial status.

Discussion:

Addresses US-122

It was the LWG's intent in Pittsburgh that N2772 be applied to the WP.

[ Resolved in Rapperswil by paper N3106. ]

Proposed resolution:

Apply N2772 to the WP.


1442(i). "happens-before" should be "synchronizes-with"

Section: 33 [thread] Status: NAD Editorial Submitter: Canada Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread].

View all issues with NAD Editorial status.

Duplicate of: 1443

Discussion:

Addresses CA-9, GB-122

[CA-9:]

Imposed happens-before edges should be in synchronizes-with
Each use of the words "happens-before" should be replaced with the words "synchronizes-with" in the following sentences:
27.2.3p2
30.3.1.2p6
30.3.1.5p7
30.6.4p7
30.6.9p5
30.6.10.1p23
Rationale: Happens-before is defined in 1.10p11 in a way that (deliberately) does not make it explicitly transitively closed. Adding edges to happens-before directly, as in 27.2.3p2 etc., does not provide transitivity with sequenced-before or any other existing happens-before edge. This lack of transitivity seems to be unintentional.

[GB-122]

At various points in the standard new edges are added to happens-before, for example 27.2.3:2 adds happens-before edges between writes and reads from a stream:

If one thread makes a library call a that writes a value to a stream and, as a result, another thread reads this value from the stream through a library call b such that this does not result in a data race, then a happens before b.

Happens-before is defined in 1.10:11 in a deliberate way that makes it not explicitly transitively closed. Adding edges to happens-before directly, as in 27.2.3:2, does not provide transitivity with sequenced-before or any other existing happens-before edge. This lack of transitivity seems to be unintentional. In order to achieve transitivity we suggest each edge be added to inter-thread-happens-before as a synchronises-with edge (as per conversation with Hans Boehm). In the standard, each use of the words "happens-before" should be replaced with the words "synchronizes-with" in the following sentences:

27.2.3:2, 30.3.1.2:6, 30.3.1.5:7, 30.6.4:7, 30.6.9:5, 30.6.10.1:23

Proposed resolution:

[Beman provided specific wording for the proposed resolution.]

Change 27.2.3 Thread Safety [iostreams.threadsafety] paragraph 2:

If one thread makes a library call a that writes a value to a stream and, as a result, another thread reads this value from the stream through a library call b such that this does not result in a data race, then a happens before synchronizes with b.

Change 30.3.1.2 thread constructors [thread.thread.constr] paragraph 6:

Synchronization: The invocation of the constructor happens before synchronizes with the invocation of the copy of f.

Change 30.3.1.5 thread members [thread.thread.member] paragraph 7:

Synchronization: The completion of the thread represented by *this happens before synchronizes with (1.10) join() returns returning. [ Note: Operations on *this are not synchronized. --end note ]

Change 30.6.4 Associated asynchronous state [futures.state] paragraph 7:

Calls to functions that successfully set the stored result of an associated asynchronous state synchronize with (1.10) calls to functions successfully detecting the ready state resulting from that setting. The storage of the result (whether normal or exceptional) into the associated asynchronous state happens before synchronizes with (1.10) that state is being set to ready.

Change 30.6.9 Function template async [futures.async] paragraph 5:

Synchronization: the invocation of async happens before synchronizes with (1.10) the invocation of f. [ Note: this statement applies even when the corresponding future object is moved to another thread. — end note ] If the invocation is not deferred, a call to a waiting function on an asynchronous return object that shares the associated asynchronous state created by this async call shall block until the associated thread has completed. If the invocation is not deferred, the join() on the created thread happens before synchronizes with (1.10) the first function that successfully detects the ready status of the associated asynchronous state returns or before the function that gives up the last reference to the associated asynchronous state returns, whichever happens first. If the invocation is deferred, the completion of the invocation of the deferred function happens before synchronizes with the calls to the waiting functions return.

Change 30.6.10.1 packaged_task member functions [futures.task.members] paragraph 23:

Synchronization: a successful call to operator() synchronizes with (1.10) a call to any member function of a future, shared_future, or atomic_future object that shares the associated asynchronous state of *this. The completion of the invocation of the stored task and the storage of the result (whether normal or exceptional) into the associated asynchronous state happens before synchronizes with (1.10) the state is being set to ready. [ Note: operator() synchronizes and serializes with other functions through the associated asynchronous state. —end note ]


1443(i). Imposed happens-before edges are not made transitive

Section: 33 [thread] Status: Dup Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread].

View all issues with Dup status.

Duplicate of: 1442

Discussion:

Addresses GB-122

At various points in the standard new edges are added to happens-before, for example 27.2.3:2 adds happens-before edges between writes and reads from a stream:

If one thread makes a library call a that writes a value to a stream and, as a result, another thread reads this value from the stream through a library call b such that this does not result in a data race, then a happens before b.

Happens-before is defined in 1.10:11 in a deliberate way that makes it not explicitly transitively closed. Adding edges to happens-before directly, as in 27.2.3:2, does not provide transitivity with sequenced-before or any other existing happens-before edge. This lack of transitivity seems to be unintentional. In order to achieve transitivity we suggest each edge be added to inter-thread-happens-before as a synchronises-with edge (as per conversation with Hans Boehm). In the standard, each use of the words "happens-before" should be replaced with the words "synchronizes-with" in the following sentences:

27.2.3:2, 30.3.1.2:6, 30.3.1.5:7, 30.6.4:7, 30.6.9:5, 30.6.10.1:23

Proposed resolution:

Request the concurrency working group to determine if changes are needed


1444(i). OFF_T is not defined

Section: 31.5.3.2 [fpos.operations] Status: Dup Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [fpos.operations].

View all issues with Dup status.

Duplicate of: 1414

Discussion:

Addresses GB-123

Several rows in table 124 specify a Return type of 'OFF_T', which does not appear to be a type defined in this standard.

Proposed resolution:

Resolve outstanding references to the removed type 'OFF_T'.


1446(i). Move and swap for I/O streams

Section: 31.7 [iostream.format] Status: NAD Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iostream.format].

View all issues with NAD status.

Discussion:

Addresses US-138

For istreams and ostreams, the move-constructor does not move-construct, the move-assignment operator does not move-assign, and the swap function does not swap because these operations do not manage the rdbuf() pointer. Useful applications of these operations are prevented both by their incorrect semantics and because they are protected.

[ Resolution proposed by ballot comment: ]

In short: reverse the resolution of issue 900, then change the semantics to move and swap the rdbuf() pointer. Add a new protected constructor that takes an rvalue reference to a stream and a pointer to a streambuf, a new protected assign() operator that takes the same arguments, and a new protected partial_swap() function that doesn't swap rdbuf(). See Appendix 1 - Additional Details

[ 2010-10-24 Daniel adds: ]

Accepting n3179 would solve this issue.

[ 2010-11 Batavia ]

Closed as NAD.

The Library Working Group reviewed n3179 and concluded that this change alone was not sufficient, as it would require changes to some of the derived stream types in the library. The preference is to not make such a broad fix, and retain the current semantics. This is closed as NAD rather than NAD future as it will be difficult to rename the new functions introduced in the C++0x revision of the standard at a later date.

Proposed resolution:


1451(i). regex should support allocators

Section: 32.7 [re.regex] Status: Dup Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [re.regex].

View all issues with Dup status.

Duplicate of: 1396

Discussion:

Addresses US-141

std::basic_regex should have an allocator for all the reasons that a std::string does. For example, I can use boost::interprocess to put a string or vector in shared memory, but not a regex.

Proposed resolution:

Add allocators to regexes; see paper N3171 in the pre-Batavia mailing.


1452(i). "target sequence" is not defined

Section: 32.9.5 [re.results.acc] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [re.results.acc].

View all issues with NAD status.

Discussion:

Addresses GB-125

The term "target sequence" is not defined (32.9.5 [re.results.acc] p. 2).

[ Resolution proposed by ballot comment: ]

Replace "target sequence" with "string being searched/matched"

[ 2010-11-01 Daniel comments: ]

The proposed resolution looks incomplete to me, there are more normative usages of the term target sequence in clause 28, e.g. 32.11.2 [re.tokiter] p. 7.

[2011-03-22 Madrid meeting: Moved to NAD]

Rationale:

Standard is correct as written

Proposed resolution:

Wording changes are against N3126. They are intended not to conflict with the wording changes suggested by n3158.

Change 32.9.5 [re.results.acc] p. 2 as indicated:

difference_type position(size_type sub = 0) const;

2 Returns: The distance from the start of the target sequencestring being matched to (*this)[sub].first.


1454(i). Ensure C compatibility for atomics

Section: 33.5 [atomics] Status: Dup Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics].

View all issues with Dup status.

Duplicate of: 1455

Discussion:

Addresses GB-128

WG14 has made some late changes to their specification of atomics, and care should be taken to ensure that we retain a common subset of language/library syntax to declare headers that are portable to both languages. Ideally, such headers would not require users to define their own macros, especially not macros that map to keywords (which remains undefined behaviour)

[ Resolution proposed by ballot comment ]

Depends on result of the review of WG14 work, which is expected to be out to ballot during the time wg21 is resolving its own ballot comments. Liaison may also want to file comments in WG14 to ensure compatibity from both sides.

Proposed resolution:


1456(i). Missing fixed-size atomic_ typedefs

Section: 33.5 [atomics] Status: NAD Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics].

View all issues with NAD status.

Discussion:

Addresses GB-129

Table 143 lists the typedefs for various atomic types corresponding to the various standard integer typedefs, such as atomic_int_least8_t for int_least8_t, and atomic_uint_fast64_t for uint_fast64_t. However, there are no atomic typedefs corresponding to the fixed-size standard typedefs int8_t, int16_t, and so forth.

[ 2010-10-24 Daniel adds: ]

Accepting n3164 would solve this issue.

[ 2011-02-15 Anthony corrects numbering/naming for N3225, Howard suggests improvement for the position of '(optional)', Daniel reorders rows in harmony to remaining entries and suggests specific optionality comments: ]

[2011-02-16 Reflector discussion]

Moved to Tentatively Ready after 5 votes.

[2011-03-16: Hans reopenes and comments]

WG14 briefly discussed LWG 1456. It turns out that they had previously made a conscious decision not to make a similar change. If C++ would deviate, this would introduce a C divergence.

We should reopen the issue and, in my opinion, probably resolve it as NAD instead. This is in a part of the standard that is there mostly for C compatibility, so introducing divergence here seems to make no sense.

[2011-03-24 Madrid]

Rationale:

WG14 does not require these typedefs, and we see no reason to be gratuitously different.

Proposed resolution:

Add the following entries to table 143:

Table 146 — atomic <inttypes.h> typedefs
Atomic typedef <inttypes.h> type
... ...
atomic_intmax_t intmax_t
atomic_uintmax_t uintmax_t
atomic_int8_t // iff int8_t is provided int8_t
atomic_uint8_t // iff uint8_t is provided uint8_t
atomic_int16_t // iff int16_t is provided int16_t
atomic_uint16_t // iff uint16_t is provided uint16_t
atomic_int32_t // iff int32_t is provided int32_t
atomic_uint32_t // iff uint32_t is provided uint32_t
atomic_int64_t // iff int64_t is provided int64_t
atomic_uint64_t // iff uint64_t is provided uint64_t

1458(i). Overlapping evaluations are allowed

Section: 33.5.4 [atomics.order] Status: Dup Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [atomics.order].

View all other issues in [atomics.order].

View all issues with Dup status.

Duplicate of: 1459

Discussion:

Addresses GB-131

33.5.5 [atomics.lockfree] p.8 states:

An atomic store shall only store a value that has been computed from constants and program input values by a finite sequence of program evaluations, such that each evaluation observes the values of variables as computed by the last prior assignment in the sequence.

... but 6.9.1 [intro.execution] p.13 states:

If A is not sequenced before B and B is not sequenced before A, then A and B are unsequenced. [ Note: The execution of unsequenced evaluations can overlap. — end note ]

Overlapping executions can make it impossible to construct the sequence described in 33.5.5 [atomics.lockfree] p.8. We are not sure of the intention here and do not offer a suggestion for change, but note that 33.5.5 [atomics.lockfree] p.8 is the condition that prevents out-of-thin-air reads.

Proposed resolution:

Request the concurrency working group to determine if changes are needed. Consider changing the use of "sequence" in 33.5.5 [atomics.lockfree]


1461(i). Rename all ATOMIC_* macros as STD_ATOMIC_*

Section: 33.5 [atomics] Status: NAD Submitter: Canada Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics].

View all issues with NAD status.

Discussion:

Addresses CA-1

All ATOMIC_... macros should be prefixed with STD_ as in STD_ATOMIC_... to indicate they are STD macros as other standard macros. The rationale that they all seem too long seems weak.

[2011-03-06: Daniel adapts suggested wording to N3242 and comments]

I suggest to declare this issue as NAD. Reason for this suggestion is, that C1x is currently going to suggest exactly the same macros as additions to header <stdatomic.h>, therefore C++0x should not define a whole new set. I'm making this suggestion with the understanding that C1x is intending to keep in sync in this regard. For example, the most recent draft of C1x does contain the macro ATOMIC_ADDRESS_LOCK_FREE which has recently been removed from the C++ working draft.

[2011-03-24]

Rationale:

C is not going to change the name of these macros, and it is important they have the same name for compatibility

Proposed resolution:

  1. Change sub-clause 33.5.2 [atomics.syn] as indicated:

    [..]
    // [atomics.lockfree], lock-free property
    #define STD_ATOMIC_CHAR_LOCK_FREE unspecified
    #define STD_ATOMIC_CHAR16_T_LOCK_FREE unspecified
    #define STD_ATOMIC_CHAR32_T_LOCK_FREE unspecified
    #define STD_ATOMIC_WCHAR_T_LOCK_FREE unspecified
    #define STD_ATOMIC_SHORT_LOCK_FREE unspecified
    #define STD_ATOMIC_INT_LOCK_FREE unspecified
    #define STD_ATOMIC_LONG_LOCK_FREE unspecified
    #define STD_ATOMIC_LLONG_LOCK_FREE unspecified
    
    // [atomics.types.operations.req], operations on atomic types
    #define STD_ATOMIC_VAR_INIT(value) see below
    [..]
    
  2. Change 33.5.5 [atomics.lockfree] p. 1 as indicated:

    #define STD_ATOMIC_CHAR_LOCK_FREE implementation-defined
    #define STD_ATOMIC_CHAR16_T_LOCK_FREE implementation-defined
    #define STD_ATOMIC_CHAR32_T_LOCK_FREE implementation-defined
    #define STD_ATOMIC_WCHAR_T_LOCK_FREE implementation-defined
    #define STD_ATOMIC_SHORT_LOCK_FREE implementation-defined
    #define STD_ATOMIC_INT_LOCK_FREE implementation-defined
    #define STD_ATOMIC_LONG_LOCK_FREE implementation-defined
    #define STD_ATOMIC_LLONG_LOCK_FREE implementation-defined
    

    1 The STD_ATOMIC_..._LOCK_FREE macros indicate the lock-free property of the corresponding atomic types, [..]

  3. Change 99 [atomics.types.operations.req] p. 6 as indicated:

    #define STD_ATOMIC_VAR_INIT(value) see below
    

    5 Remarks: The macro expands to a token sequence suitable for constant initialization an atomic variable of static storage duration of a type that is initialization-compatible with value. [ Note: This operation may need to initialize locks. — end note ] Concurrent access to the variable being initialized, even via an atomic operation, constitutes a data race. [ Example:

    atomic<int> v = STD_ATOMIC_VAR_INIT(5);
    

    end example ]

  4. Change 33.5.10 [atomics.flag] p. 1+4 as indicated:

    namespace std {
      [..]
      #define STD_ATOMIC_FLAG_INIT see below
    }
    

    [..] 4 The macro STD_ATOMIC_FLAG_INIT shall be defined in such a way that it can be used to initialize an object of type atomic_flag to the clear state. For a static-duration object, that initialization shall be static. It is unspecified whether an unitialized atomic_flag object has an initial state of set or clear. [ Example:

    atomic_flag guard = STD_ATOMIC_FLAG_INIT;
    

    end example ]


1463(i). Inconsistent value assignment for atomic_bool

Section: 99 [atomics.types.integral] Status: Dup Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.integral].

View all issues with Dup status.

Duplicate of: 1462

Discussion:

Addresses US-157

atomic_bool has a volatile assignment operator but not a non-volatile operator. The other integral types have both.

Proposed resolution:

Add a non-volatile assignment operator to atomic_bool.


1470(i). "Same-ness" curiosities

Section: 33.5.8.2 [atomics.types.operations] Status: Dup Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.operations].

View all issues with Dup status.

Duplicate of: 1474

Discussion:

Addresses US-165

According to 33.5.8.2 [atomics.types.operations] p. 23:

“is the same that same as that of” is not grammatical (and is not clear)

Proposed resolution:


1471(i). Default constructor of atomics needs specification

Section: 33.5.8.2 [atomics.types.operations] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.operations].

View all issues with NAD Editorial status.

Discussion:

Addresses US-168

33.5.8.2 [atomics.types.operations] around p. 4: The definition of the default constructor needs exposition.

Proposed resolution:

Insert a new general prototype description following the current 33.5.8.2 [atomics.types.operations] p. 3 as indicated:

3 [Note: Many operations are volatile-qualified. The “volatile as device register” semantics have not changed in the standard. This qualification means that volatility is preserved when applying these operations to volatile objects. It does not mean that operations on non-volatile objects become volatile. Thus, volatile qualified operations on non-volatile objects may be merged under some conditions. -- end note]

A::A() = default;

? Effects: Leaves the atomic object in an uninitialized state. [Note: These semantics ensure compatiblity with C. -- end note]

constexpr A::A(C desired);
[..]

1472(i). Incorrect semantics of atomic_init

Section: 33.5.8.2 [atomics.types.operations] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.operations].

View all issues with NAD Editorial status.

Discussion:

Addresses US-171

As of 33.5.8.2 [atomics.types.operations] p. 7:

The atomic_init definition "Non-atomically assigns the value" is not quite correct, as the atomic_init purpose is initialization.

Proposed resolution:

Change 33.5.8.2 [atomics.types.operations] p. 7 as indicated:

void atomic_init(volatile A *object, C desired);
void atomic_init(A *object, C desired);

7 Effects: Non-atomically assigns the value desired to *objectInitializes *object with value desired. Concurrent access from another thread, even via an atomic operation, constitutes a data race. [Note: This function should only be applied to objects that have been default constructed. These semantics ensure compatibility with C. — end note]


1473(i). Incomplete memory order specifications

Section: 33.5.8.2 [atomics.types.operations] Status: NAD Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.operations].

View all issues with NAD status.

Discussion:

Addresses US-172

As of 33.5.8.2 [atomics.types.operations] p. 9, 13, 17, 20:

The order specifications are incomplete because the non-_explicit functions do not have such parameters.

Add a new sentence: "If the program does not specify an order, it shall be memory_order_seq_cst." Or perhaps: "The non-_explicit non-member functions shall affect memory as though they were _explicit with memory_order_seq_cst."

[ 2010 Batavia ]

The Concurrency subgroup reviewed this, and deemed it NAD according to 33.5.8.2 [atomics.types.operations] paragraph 2, bullet 4.

Rationale:

The working paper is correct as written.

Proposed resolution:

  1. Change 33.5.8.2 [atomics.types.operations] p. 9 as indicated:
    void atomic_store(volatile A* object, C desired);
    void atomic_store(A* object, C desired);
    void atomic_store_explicit(volatile A *object, C desired, memory_order order);
    void atomic_store_explicit(A* object, C desired, memory_order order);
    void A::store(C desired, memory_order order = memory_order_seq_cst) volatile;
    void A::store(C desired, memory_order order = memory_order_seq_cst);
    

    8 Requires: The order argument shall not be memory_order_consume, memory_order_acquire, nor memory_order_acq_rel.

    9 Effects: Atomically replaces the value pointed to by object or by this with the value of desired. Memory is affected according to the value of order. If the program does not specify an order, it shall be memory_order_seq_cst.

  2. Change 33.5.8.2 [atomics.types.operations] p. 13 as indicated:
    C atomic_load(const volatile A* object);
    C atomic_load(const A* object);
    C atomic_load_explicit(const volatile A* object, memory_order);
    C atomic_load_explicit(const A* object, memory_order);
    C A::load(memory_order order = memory_order_seq_cst) const volatile;
    C A::load(memory_order order = memory_order_seq_cst) const;
    

    12 Requires: The order argument shall not be memory_order_release nor memory_order_acq_rel.

    13 Effects: Memory is affected according to the value of order. If the program does not specify an order, it shall be memory_order_seq_cst.

    14 Returns: Atomically returns the value pointed to by object or by this.

  3. Change 33.5.8.2 [atomics.types.operations] p. 17 as indicated:
    C atomic_exchange(volatile A* object, C desired);
    C atomic_exchange(A* object, C desired);
    C atomic_exchange_explicit(volatile A* object, C desired, memory_order);
    C atomic_exchange_explicit(A* object, C desired, memory_order);
    C A::exchange(C desired, memory_order order = memory_order_seq_cst) volatile;
    C A::exchange(C desired, memory_order order = memory_order_seq_cst);
    

    17 Effects: Atomically replaces the value pointed to by object or by this with desired. Memory is affected according to the value of order. These operations are atomic read-modify-write operations (1.10). If the program does not specify an order, it shall be memory_order_seq_cst.

    18 Returns: Atomically returns the value pointed to by object or by this immediately before the effects.

  4. Change 33.5.8.2 [atomics.types.operations] p. 20 as indicated:
    bool atomic_compare_exchange_weak(volatile A* object, C * expected, C desired);
    bool atomic_compare_exchange_weak(A* object, C * expected, C desired);
    bool atomic_compare_exchange_strong(volatile A* object, C * expected, C desired);
    bool atomic_compare_exchange_strong(A* object, C * expected, C desired);
    bool atomic_compare_exchange_weak_explicit(volatile A* object, C * expected, C desired,
      memory_order success, memory_order failure);
    bool atomic_compare_exchange_weak_explicit(A* object, C * expected, C desired,
      memory_order success, memory_order failure);
    bool atomic_compare_exchange_strong_explicit(volatile A* object, C * expected, C desired,
      memory_order success, memory_order failure);
    bool atomic_compare_exchange_strong_explicit(A* object, C * expected, C desired,
      memory_order success, memory_order failure);
    bool A::compare_exchange_weak(C & expected, C desired,
      memory_order success, memory_order failure) volatile;
    bool A::compare_exchange_weak(C & expected, C desired,
      memory_order success, memory_order failure);
    bool A::compare_exchange_strong(C & expected, C desired,
      memory_order success, memory_order failure) volatile;
    bool A::compare_exchange_strong(C & expected, C desired,
      memory_order success, memory_order failure);
    bool A::compare_exchange_weak(C & expected, C desired,
      memory_order order = memory_order_seq_cst) volatile;
    bool A::compare_exchange_weak(C & expected, C desired,
      memory_order order = memory_order_seq_cst);
    bool A::compare_exchange_strong(C & expected, C desired,
      memory_order order = memory_order_seq_cst) volatile;
    bool A::compare_exchange_strong(C & expected, C desired,
      memory_order order = memory_order_seq_cst);
    

    19 Requires: The failure argument shall not be memory_order_release nor memory_order_acq_rel. The failure argument shall be no stronger than the success argument.

    20 Effects: Atomically, compares the contents of the memory pointed to by object or by this for equality with that in expected, and if true, replaces the contents of the memory pointed to by object or by this with that in desired, and if false, updates the contents of the memory in expected with the contents of the memory pointed to by object or by this. Further, if the comparison is true, memory is affected according to the value of success, and if the comparison is false, memory is affected according to the value of failure. When only one memory_order argument is supplied, the value of success is order, and the value of failure is order except that a value of memory_order_acq_rel shall be replaced by the value memory_order_acquire and a value of memory_order_release shall be replaced by the value memory_order_relaxed. If the program does not specify an order, it shall be memory_order_seq_cst. If the operation returns true, these operations are atomic read-modify-write operations (1.10). Otherwise, these operations are atomic load operations.

    [..]


1475(i). weak compare-and-exchange confusion II

Section: 33.5.8.2 [atomics.types.operations] Status: Dup Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.operations].

View all issues with Dup status.

Duplicate of: 1474

Discussion:

Addresses CH-23

33.5.8.2 [atomics.types.operations] p. 23: The first sentence has non-English syntax.

[ Resolution proposed in ballot comment: ]

Change to "The weak compare-and-exchange operations may fail spuriously, that is, return false while leaving the contents of memory pointed to by expected unchanged."

[ Daniel translates NB comment in a proposed resolution ]

Change 33.5.8.2 [atomics.types.operations] p. 23 as indicated:

23 Remark: The weak compare-and-exchange operations may fail spuriously, that is, return false while leaving the contents of memory pointed to by expected before the operation is the same that same as that of the object and the same as that of expected after the operationThe weak compare-and-exchange operations may fail spuriously, that is, return false while leaving the contents of memory pointed to by expected unchanged.. [ Note: This spurious failure enables implementation of compare-and-exchange on a broader class of machines, e.g., loadlocked store-conditional machines. A consequence of spurious failure is that nearly all uses of weak compare-and-exchange will be in a loop.

When a compare-and-exchange is in a loop, the weak version will yield better performance on some platforms. When a weak compare-and-exchange would require a loop and a strong one would not, the strong one is preferable. — end note ]

[ See 1474 for the proposed resolution ]

Proposed resolution:


1476(i). Meaningless specification of spurious failure

Section: 33.5.8.2 [atomics.types.operations] Status: Dup Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.operations].

View all issues with Dup status.

Duplicate of: 1474

Discussion:

Addresses US-177

The first sentence of this paragraph doesn't make sense.

[ Resolution proposed in ballot comment ]

Figure out what it's supposed to say, and say it.

Proposed resolution:


1477(i). weak compare-and-exchange confusion III

Section: 33.5.8.2 [atomics.types.operations] Status: Dup Submitter: BSI Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [atomics.types.operations].

View all issues with Dup status.

Duplicate of: 1474

Discussion:

Addresses GB-135

The first sentence of 33.5.8.2 [atomics.types.operations] p.23 was changed by n2992 but now makes no sense: "that is, return false while leaving the contents of memory pointed to by expected before the operation is the same that same as that of the object and the same as that of expected after the operation." There's a minor editorial difference between n2992 ("is that same as that" vs "is the same that same as that") but neither version makes sense. Also, the remark talks about "object" which should probably be "object or this" to cover the member functions which have no object parameter.

[ Resolution proposed in ballot comment: ]

Fix the Remark to say whatever was intended.

Proposed resolution:


1483(i). __STDCPP_THREADS spelling

Section: 33.4 [thread.threads] Status: NAD Editorial Submitter: DIN Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

Addresses DE-23

Predefined macros usually start and end with two underscores, see 16.8 and FDIS 29124 = WG21 N3060 clause 7. __STDCPP_THREADS should blend in.

[ Resolved in Rapperswil by a motion to directly apply the words from the ballot comment in N3102. ]

Proposed resolution:

Change the macro name to __STDCPP_THREADS__.


1485(i). Unclear thread::id specification

Section: 33.4.3.2 [thread.thread.id] Status: NAD Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.thread.id].

View all issues with NAD status.

Discussion:

Addresses US-184

It is unclear when a thread::id ceases to be meaningful. The sentence "The library may reuse the value of a thread::id of a terminated thread that can no longer be joined." implies that some terminated threads can be joined. It says nothing about detached threads.

[ Resolution proposed by ballot comment: ]

Require a unique thread::id for every thread that is (1) detached and not terminated or (2) has an associated std::thread object.

[ 2010-11-22 Howard Hinnant observes ]

A thread can either be running or terminated. Additionally a thread can be joined, detached, or neither. These combine into the five possible states shown in this table:

RunningTerminated
Neither joined nor detachedshall not reuse idshall not reuse id
detachedshall not reuse idmay reuse id
joinedimpossible statemay reuse id

Only if a thread is neither joined nor detached can it be joined. Or said differently, if a thread has already been joined or detached, then it can not be joined. The sentence:

The library may reuse the value of a thread::id of a terminated thread that can no longer be joined.

precisely defines the two states shown in the above table where a thread::id may be reused.

The following program illustrates all of the possibilities:

#include <mutex>
#include <thread>
#include <iostream>
#include <chrono>

std::mutex mut;

void f()
{
   std::lock_guard<std::mutex> _(mut);
   std::cout << "f id = " << std::this_thread::get_id() << " terminating\n";
}

void g()
{
   std::lock_guard<std::mutex> _(mut);
   std::cout << "g id = " << std::this_thread::get_id() << " terminating\n";
}

int main()
{
   std::cout << "main id = " << std::this_thread::get_id() << "\n";
   std::thread t1(f);
   std::thread(g).detach();
   std::this_thread::sleep_for(std::chrono::seconds(1));
   std::cout << "g's thread::id can be reused here because g has terminated and is detached.\n";
   std::cout << "f's thread::id can't be reused here because f has terminated but is still joinable.\n";
   std::cout << "f id = " << t1.get_id() << "\n";
   t1.join();
   std::cout << "f's thread::id can be reused here because f has terminated and is joined.\n";
   std::cout << "f id = " << t1.get_id() << "\n";
}

main id = 0x7fff71197ca0
f id = 0x100381000 terminating
g id = 0x100581000 terminating
g's thread::id can be reused here because g has terminated and is detached.
f's thread::id can't be reused here because f has terminated but is still joinable.
f id = 0x100381000
f's thread::id can be reused here because f has terminated and is joined.
f id = 0x0

[2011-02-11 Reflector discussion]

Moved to Tentatively NAD after 5 votes.

Proposed resolution:


1486(i). Value of this_thread::get_id() underspecified for detached thread

Section: 33.4.5 [thread.thread.this] Status: NAD Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.thread.this].

View all issues with NAD status.

Discussion:

Addresses CH-24

What would be the value this_thread::get_id() when called from a detached thread?

[ Resolution proposed by ballot comment: ]

Add some text to clarify that get_id() still returns the same value even after detaching.

[ 2010-11-22 Howard Hinnant observes ]

33.4.5 [thread.thread.this]/1 contains the following sentence describing this_thread::get_id():

... No other thread of execution shall have this id and this thread of execution shall always have this id.

I don't object to adding "even if detached" to this sentence, but it seems unnecessary to me. "Always" means always.

[2011-02-11 Reflector discussion]

Moved to Tentatively NAD after 5 votes.

Proposed resolution:


1489(i). unlock functions and unlock mutex requirements are inconsistent

Section: 33.6 [thread.mutex] Status: NAD Editorial Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.mutex].

View all issues with NAD Editorial status.

Discussion:

Addresses CH-26

Specifications of unlock member functions and unlock mutex requirements are inconsistent wrt to exceptions and pre- and postconditions.

[ Resolution proposed by ballot comment: ]

unlock should specifiy the precondition that the current thread "owns the lock", this will make calls without holding the locks "undefined behavior". unlock in [mutex.requirements] should either be noexcept(true) or be allowed to throw system_error like unique_lock::unlock, or the latter should be nothrow(true) and have the precondition owns == true. Furthermore unique_lock's postcondition is wrong in the case of a recursive mutex where owns might stay true, when it is not the last unlock needed to be called.

Proposed resolution:


1495(i). Condition variable wait_for return value insufficient

Section: 33.7 [thread.condition] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.condition].

View all issues with NAD Editorial status.

Discussion:

Addresses US-191

The condition variable wait_for returning cv_status is insufficient.

[ Resolution proposed by ballot comment: ]

Return a duration of timeout remaining instead. See Appendix 1 of n3141 - Additional Details, p. 211

Proposed resolution:


1496(i). condition_variable not implementable

Section: 33.7.4 [thread.condition.condvar] Status: NAD Editorial Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.condition.condvar].

View all issues with NAD Editorial status.

Discussion:

Addresses CH-28

Requiring wait_until makes it impossible to implement condition_variable correctly using respective objects provided by the operating system (i.e. implementing the native_handle() function) on many platforms (e.g. POSIX, Windows, MacOS X) or using the same object as for the condition variable proposed for C.

[ Resolution proposed by ballot comment: ]

Remove the wait_until functions or make them at least conditionally supported.

Proposed resolution:


1499(i). Condition variables preclude wakeup optimization

Section: 33.7 [thread.condition] Status: NAD Submitter: INCITS Opened: 2010-08-25 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [thread.condition].

View all issues with NAD status.

Discussion:

Addresses US-193

Condition variables preclude a wakeup optimization.

[ Resolution proposed by ballot comment: ]

Change condition_variable to allow such optimization. See Appendix 1 - Additional Details

[ 2010 Batavia ]

The Concurrency subgroup reviewed the issue, and deemed it an extension to be handled after C++0x.

Rationale:

The LWG does not wish to make the change at this time.

[LEWG Kona 2017]

Recommend NAD: (N4618 numbering) 30.5.1[thread.condition.condvar] p10.3 allows spurious wakeups. This issue is out of date.

[2017-03-01, Kona]

SG1 recommends: Close as NAD

The approach suggested there raises all sorts of issues about thread_locals, etc. It's probably way too late to change this anyway, but this would have required a careful paper.

Proposed resolution:


1500(i). Consider removal of native_handle()

Section: 33.7.5 [thread.condition.condvarany] Status: NAD Editorial Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.condition.condvarany].

View all issues with NAD Editorial status.

Discussion:

Addresses CH-32

Given that the lock type can be something the underlying doesn't know 'native_handle()' is probably unimplementable on essentially all platforms.

[ Resolved in Rapperswil by a motion to directly apply the words from the ballot comment in N3102. ]

Proposed resolution:

Consider the removal of 'native_handle()'.


1503(i). "associated asynchronous state" must go

Section: 33.10.5 [futures.state] Status: NAD Editorial Submitter: Switzerland Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [futures.state].

View all issues with NAD Editorial status.

Discussion:

Addresses CH-35

The term "associated asynchronous state" is long, ugly and misleading terminology. When introduced we agreed upon that we should come up with a better name. Here it is: "liaison state". Since the state is hidden and provides synchronization of a future with its corresponding promise, we believe "liaison state" is a much better and shorter name (liaison ~ (typically hidden) relationship)

[ Resolution proposed by ballot comment: ]

Change all occurrences of "associated asynchronous state" to "liaison state".

Proposed resolution:

The project editor may supply a more appopriate term, or use "liaison state", at his own discretion.


1506(i). set_exception with a null pointer

Section: 33.10.6 [futures.promise] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [futures.promise].

View all other issues in [futures.promise].

View all issues with NAD Editorial status.

Discussion:

Addresses US-198

promise::set_exception can be called with a null pointer, but none of the descriptions of the get() functions for the three types of futures say what happens for this case.

[ Resolved in Rapperswil by a motion to directly apply the words from the ballot comment in N3102. ]

Proposed resolution:

Add the following sentence to the end of 30.6.5/22: The behavior of a program that calls set_exception with a null pointer is undefined.


1509(i). No restriction on calling future::get more than once

Section: 99 [futures.atomic_future] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [futures.atomic_future].

View all issues with NAD Editorial status.

Discussion:

Addresses US-202

The note in this paragraph says "unlike future, calling get more than once on the same atomic_future object is well defined and produces the result again." There is nothing in future that says anything negative about calling get more than once.

[ Resolution proposed by ballot comment: ]

Remove this note, or add words to the requirements for future that reflect what this note says.

Proposed resolution:


1510(i). Should be undefined behaviour to call atomic_future operations unless valid()

Section: 99 [futures.atomic_future] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [futures.atomic_future].

View all issues with NAD Editorial status.

Discussion:

Addresses US-203

Both future and shared_future specify that calling most member functions on an object for which valid() == false produces undefined behavior. There is no such statement for atomic_future.

[ Resolution proposed by ballot comment: ]

Add a new paragraph after 99 [futures.atomic_future]/2 with the same words as [futures.shared_future]/3.

[ 2010-11-02 Daniel translates proposed changes into specific deltas and comments: ]

While applying the wording, I notice that [futures.shared_future]/3 does speak of the move-assignment operator, and not of the copy-assignment operator. atomic_future obviously needs this to be true for the copy-assignment operator, but I strongly assume that shared_future needs to mention both special member assignment operators in this paragraph. To keep this consistent, the following P/R also provides wording to fix the corresponding location for shared_future.

Proposed resolution:

  1. Change [futures.shared_future]/3 as indicated:

    3 The effect of calling any member function other than the destructor, the copy-assignment operator, the move-assignment operator, or valid() on a shared_future object for which valid() == false is undefined.

  2. Following 99 [futures.atomic_future]/2, add a new paragraph:

    ? The effect of calling any member function other than the destructor, the copy-assignment operator, or valid() on a atomic_future object for which valid() == false is undefined.


1511(i). Synchronize the move-constructor for atomic_future

Section: 99 [futures.atomic_future] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [futures.atomic_future].

View all issues with NAD Editorial status.

Discussion:

Addresses US-204

According to the definition of atomic_future, all members of atomic_future are synchronizing except constructors. However, it would probably be appropriate for a move constructor to be synchronizing on the source object. If not, the postconditions on paragraphs 7-8, might not be satisfied. This may be applicable if a collection of futures are being doled out to a set of threads that process their value.

[ Resolution proposed by ballot comment: ]

Make the move constructor for atomic future lock the source

Proposed resolution:


1512(i). Conflict in specification: block or join?

Section: 33.10.9 [futures.async] Status: NAD Editorial Submitter: INCITS Opened: 2010-08-25 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [futures.async].

View all other issues in [futures.async].

View all issues with NAD Editorial status.

Discussion:

Addresses US-205

33.10.9 [futures.async] p. 3: The third sentence says "If the invocation is not deferred, a call to a waiting function on an asynchronous return object that shares the associated asynchronous state created by this async call shall block until the associated thread has completed." The next sentence says "If the invocation is not deferred, the join() on the created thread..." Blocking until a thread completes is not necessarily a join.

[ Resolution proposed by ballot comment: ]

Decide whether the requirement is to block until finished or to call join, and rewrite to match.

Proposed resolution:


2006(i). emplace broken for associative containers

Section: 24.2.8 [unord.req] Status: NAD Submitter: Pablo Halpern Opened: 2010-10-18 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with NAD status.

Discussion:

The current definition of emplace(args) for associative containers as described in Table 99 is:

Requires: T shall be constructible from args.

Effects: Inserts a T object t constructed with std::forward<Args>(args)... if and only if there is no element in the container with key equivalent to the key of t. The bool component of the returned pair is true if and only if the insertion takes place, and the iterator component of the pair points to the element with key equivalent to the key of t.

There is similar language in Table 100 for unordered associative containers.

The first issue is editorial: T should be value_type throughout both tables.

The major issue is that, if the container is map, multimap, unordered_map, or unordered_multimap, then the only way to construct an object of value_type is to supply exactly two arguments for Key and Value, a pair<Key,Value>, or a piecewise_construct_t followed by two tuples. The original emplace() proposal would have allowed you to specify a Key value followed by any number of constructor arguments for Value. When we removed the variadic constructor to pair, this ability went away. I don't think that was deliberate.

Fixing this is non-trivial, I think. I think that emplace() for map and multimap need several overloads: one for each overloaded constructor in pair<Key,Value>, and one for the emplace(Key, valueargs...) case. And it probably needs some SFINAE meta-programming to ensure that the last case doesn't override any of the other ones. Alternatively, one could say that there are exactly two cases: emplace(args) where pair<Key,Value> is constructible from args, and emplace(args) where Key is constructible form the first arg and Value is constructible from the rest.

Alternatively, the status quo is to use piecewise_construct_t if you want to construct an object.

[ 2010 Batavia: ]

N3178 was looked at in session and moved to NAD.

Proposed resolution:


2026(i). hash should be std qualified for unordered container

Section: 24.5 [unord] Status: NAD Submitter: Pete Becker Opened: 2011-02-07 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [unord].

View all issues with NAD status.

Discussion:

Tom Plum pointed out to me that there's an apparent inconsistency in the std:: qualification of template names in the unordered containers:

 template <class Key,
           class T,
           class Hash = hash<Key>,
           class Pred = std::equal_to<Key>,
           class Alloc = std::allocator<std::pair<const Key, T> > >
   class unordered_map;

Is there a reason that hash is not qualified with std::? TR1 also does not use std:: here.

[ 2011-02-07 Chris Jefferson adds: ]

I assumed (I might be wrong) it is because hash is designed to be a customisation point, like swap.

[ 2011-02-07 Howard Hinnant adds: ]

I think this is incorrect. We mean std::hash, though clients are free to specialize std::hash on user-defined types. With the possible exception of begin/end (which I'm not sure if we've settled that), swap is the only intended customization point (look up a function by ADL) in the std:: lib.

[ 2011-02-24 Chris Jefferson adds: ]

I recommend NAD, due to 16.4.2.2 [contents] p3:

Whenever a name x defined in the standard library is mentioned, the name x is assumed to be fully qualified as ::std::x, unless explicitly described otherwise. For example, if the Effects section for library function F is described as calling library function G, the function ::std::G is meant.

[2011-02-25 Reflector discussion]

Moved to Tentatively NAD after 5 votes.

Proposed resolution:


2036(i). istream >> char and eofbit

Section: 31.7.5.2 [istream] Status: NAD Submitter: Howard Hinnant Opened: 2011-02-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [istream].

View all issues with NAD status.

Discussion:

The question is: When a single character is extracted from an istream using operator>>, does eofbit get set if this is the last character extracted from the stream? The current standard is at best ambiguous on the subject. 31.7.5.2 [istream]/p3 describes all extraction operations with:

3 If rdbuf()->sbumpc() or rdbuf()->sgetc() returns traits::eof(), then the input function, except as explicitly noted otherwise, completes its actions and does setstate(eofbit), which may throw ios_base::failure (31.5.4.4 [iostate.flags]), before returning.

And [istream::extractors]/p12 in describing operator>>(basic_istream<charT,traits>& in, charT& c); offers no further clarification:

12 Effects: Behaves like a formatted input member (as described in [istream.formatted.reqmts]) of in. After a sentry object is constructed a character is extracted from in, if one is available, and stored in c. Otherwise, the function calls in.setstate(failbit).

I coded it one way in libc++, and g++ coded it another way. Chris Jefferson noted that some boost code was sensitive to the difference and fails for libc++. Therefore I believe that it is very important that we specify this extraction operator in enough detail that both vendors and clients know what behavior is required and expected.

Here is a brief code example demonstrating the issue:

#include <sstream>
#include <cassert>

int main()
{
  std::istringstream ss("1");
  char t;
  ss >> t;
  assert(!ss.eof());
};

For every type capable of reading this istringstream but char, ss.eof() will be true after the extraction (bool, int, double, etc.). So for consistency's sake we might want to have char behave the same way as other built-in types.

However Jean-Marc Bourguet offers this counter example code using an interactive stream. He argues that setting eof inhibits reading the next line:

#include <iostream>

int main()
{
 char c;
 std::cin >> std::noskipws;
 std::cout << "First line: ";
 while (std::cin >> c) {
    if (c == '\n') {
       std::cout << "Next line: ";
    }
 }
}

As these two code examples demonstrate, whether or not eofbit gets set is an observable difference and it is impacting real-world code. I feel it is critical that we clearly and unambiguously choose one behavior or the other. I am proposing wording for both behaviors and ask the LWG to choose one (and only one!).

Wording for setting eof bit:

Modify [istream::extractors]/p12 as follows:

12 Effects: Behaves like a formatted input member (as described in [istream.formatted.reqmts]) of in. After a sentry object is constructed a character is extracted from in, if one is available, and stored in c. Otherwise, the function calls in.setstate(failbit). If a character is extracted and it is the last character in the pending sequence, the function calls in.setstate(eofbit). If a character is not extracted the function calls in.setstate(failbit | eofbit).

Wording for not setting eof bit:

12 Effects: Behaves like a formatted input member (as described in [istream.formatted.reqmts]) of in. After a sentry object is constructed a character is extracted from in, if one is available, and stored in c. Otherwise, the function calls in.setstate(failbit). with in.rdbuf()->sbumpc(). If traits::eof() is returned, the function calls in.setstate(failbit | eofbit). Otherwise the return value is converted to type charT and stored in c.

[2011-02-27: Jean-Marc Bourguet comments]

Just for completeness: it [the counter example] doesn't inhibit to read the next line, it inhibits the prompt to be put at the appropriate time.

More information to take into account when deciding:

[2011-02-28: Martin Sebor comments]

[Responds to bullet 1 of Jean-Marc's list]

Yes, this matches the stdcxx test suite for num_get and time_get but not money_get when the currency symbol is last. I don't see where in the locale.money.get.virtuals section we specify whether eofbit is or isn't set and when.

IMO, if we try to fix the char extractor to be consistent we should also fix all the others extractors and manipulators that aren't consistent (including std::get_money and std::get_time).

[2011-03-24 Madrid meeting]

Dietmar convinced Howard, that the standard does already say the right words

Rationale:

Reading the last character does not set eofbit and the standard says so already

Proposed resolution:


2043(i). std{in,out,err} should be usable as field names

Section: 31.13 [c.files] Status: NAD Submitter: Jeffrey Yasskin Opened: 2011-03-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [c.files].

View all issues with NAD status.

Discussion:

People often define structs and classes with fields named stdin, stdout, or stderr. According to 31.13 [c.files], though, these are macros.

glibc defines them to themselves, allowing their non-portable use as field names, while the Mac OS X libc defines them to either __stdoutp or (&__sF[1]), etc depending on __DARWIN_UNIX03. It's possible to allow their use while, as far as I can see, only requiring minor changes to various libc's, so C++1x should allow it.

[ 2011 Bloomington ]

Closed as NAD. This is an extension request that has been an issue for over 20 years. Supporting the extension would place a burden on the underlying C library that we may not be in a position to influence.

Proposed resolution:

This wording is relative to the FDIS.

  1. In 31.13 [c.files] add "stderr", "stdin", and "stdout" to a new Values section in Table 134 — Header <cstdio> synopsis:

    134 — Header <cstdio> synopsis
    Type Name(s)
    Macros:
    BUFSIZ FOPEN_MAX SEEK_CUR TMP_MAX _IONBF stdout
    EOF L_tmpnam SEEK_END _IOFBF stderr  
    FILENAME_MAX NULL <cstdio> SEEK_SET _IOLBF stdin  
    Types: FILE fpos_t size_t <cstdio>    
    Functions:
    Values:
    stderr stdin stdout      
  2. Add a new paragraph after paragraph 2 as indicated:

    2 Calls to the function tmpnam with an argument of NULL may introduce a data race (17.6.5.9) with other calls to tmpnam with an argument of NULL.
    See also: ISO C 7.9, Amendment 1 4.6.2.

    ? The macros stderr, stdin, and stdout shall expand to stderr, stdin, and stdout, respectively. [Note: This allows uses of #ifdef to detect their presence, while allowing code in other scopes to use them as identifiers. — end note]

  3. In C.7 [diff.library] add "stderr", "stdin", and "stdout" to Table 150 — Standard values:

    Table 150 — Standard values
    CHAR_BIT FLT_DIG INT_MIN MB_LEN_MAX
    SHRT_MIN
    stderr
    stdin
    stdout
    UCHAR_MAX

2046(i). shared_future(future<R>&&) should be allowed to throw

Section: 33.10.8 [futures.shared.future] Status: NAD Submitter: Peter Sommerlad Opened: 2011-04-04 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [futures.shared.future].

View all issues with NAD status.

Discussion:

Requiring the constructor shared_future(future<R>&& rhs) not to throw is a pessimisation of the case where a future is returned from a call to async(function,launch::deferred) and possible other cases.

Such a future not dealing with multiple threads only needs to keep (a copy of) the function to be called it later. However, creating a shared_future from that future will require more infrastructure, like space for the value of type R, an exception_ptr, and a synchronized reference counter for the shared_future's instances.

Enforcing the constructor shared_future(future<R>&& rhs) not to throw, implies that any implementation of future will need to pre-allocate space for shared_future's infrastructure, that also requires an operating system resource for synchronization, regardless if is ever needed.

All this came up when discussing D/N3267 and Concurrency Working Group decided that the constructor shared_future(future<R>&& rhs) should be allowed to throw.

[ 2011 Bloomington ]

Closed as NAD. Rationale to follow by email...

Proposed resolution:

Apply the proposed resolution of n3269


2060(i). unique_ptr<T[]>(nullptr_t) missing noexcept

Section: 20.3.1.4 [unique.ptr.runtime] Status: NAD Editorial Submitter: Howard Hinnant, Paolo Carlini Opened: 2011-05-27 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [unique.ptr.runtime].

View all issues with NAD Editorial status.

Discussion:

The synopsis in 20.3.1.3 [unique.ptr.single] specifies:

constexpr unique_ptr(nullptr_t) noexcept

which looks correct to me. However the corresponding constructor in 20.3.1.4 [unique.ptr.runtime] is missing noexcept:

constexpr unique_ptr(nullptr_t) : unique_ptr() { }

[Bloomington, 2011]

Closed as NAD Editorial.

Proposed resolution:

This wording is relative to the FDIS.

  1. Modify the synopsis in 20.3.1.4 [unique.ptr.runtime]:

    namespace std {
      template <class T, class D> class unique_ptr<T[], D> {
      public:
        typedef see below pointer;
        typedef T element_type;
        typedef D deleter_type;
    
        // 20.7.1.3.1, constructors
        constexpr unique_ptr() noexcept;
        […]
        constexpr unique_ptr(nullptr_t) noexcept : unique_ptr() { }
    	
        […]
      };
    }
    

2068(i). std::pair not C++03-compatible with defaulted copy c'tor

Section: 22.3.2 [pairs.pair] Status: NAD Submitter: Daniel Krügler Opened: 2011-06-18 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [pairs.pair].

View all other issues in [pairs.pair].

View all issues with NAD status.

Discussion:

The specification of the copy semantics of the C++03 version of std::pair is defined by the class synopsis in [lib.pairs]:

template <class T1, class T2>
struct pair {
  typedef T1 first_type;
  typedef T2 second_type;

  T1 first;
  T2 second;
  pair();
  pair(const T1& x, const T2& y);
  template<class U, class V> pair(const pair<U, V> &p);
};

The effect of this specification is, that the copy constructor is compiler-declared with the proper form depending on the contained member types. In particular, the instantiation of pair is well-formed with an element type that has a copy constructor with non-const first parameter type like specialzations of auto_ptr or any user-defined type like the following one:

struct A {
  A(A&){}
};

In contrast to container types which require CopyConstructible value types, the C++03 pair does support these, albeit unusual, element types.

The FDIS version of the std::pair specification does specify the same semantics by defaulting the copy and move constructor in 22.3.2 [pairs.pair]:

template <class T1, class T2>
struct pair {
  typedef T1 first_type;
  typedef T2 second_type;

  T1 first;
  T2 second;
  pair(const pair&) = default;
  pair(pair&&) = default;
  pair();
  […]
};

But according to the current core rules this makes the instantiation of e.g. std::pair<A, int> ill-formed, because of the const mismatch of the compiler-declared form of the copy constructor with that of the defaulted declaration.

Unfortunately there seems to be no simple library solution for this problem. If the defaulted declarations were removed, both copy c'tor and move c'tor would be deleted, because there exist user-declared copy assignment and move assignment operators in the FDIS. But these operations need to be user-defined to realize the wanted semantics of these operations for element types that are reference types. If core rules would not be changed to fix that, I see the following options:

  1. Intentionally decide to break the support for element types with non-const copy c'tors in pair.
  2. User-declare both copy and move ctor to at least support the instantiation of the pair specializations, but this would still not allow to copy them by the copy constructor.
  3. User-declare both the const and non-const copy ctors, the move ctor, and additionally the non-const copy assignment operator to support the instantiation of the pair specializations and of these members. This would support all element types as it did in C++03, but all copy/move members would be non-trivial.
  4. Intentionally decide to give up support for element types that are references for pair, but still keep the allocator support with the effect of removing all declarations of the special copy/move members. User code that needs to use tuple instead. But this would be a rather drastic step requiring further corrections of the draft, e.g. a change of the signature of the algorithm minmax (not the overload with the initializer_list) with a different return type.

This problem does not extend as backward-compatibility problem to tuple, because the TR1 specification did explicitly declare copy constructor and copy assignment operator via the "normal" form:

tuple(const tuple&);
tuple& operator=(const tuple&);

[Bloomington, 2011]

Closed as NAD.

This is an unfortunate change of behavior between C++03 and C++11, but is consistent with tuple. There is no desire to go to lengths supporting types like auto_ptr now that rvalue references are in the language.

There may be an issue for Core/EWG to look at, so that some simple =default syntax could be used that would do the right thing. If such a facility became availabile, LWG might revisit this issue.

Proposed resolution:


2073(i). Library exceptions that take string arguments

Section: 19.2 [std.exceptions], 19.5.8 [syserr.syserr], 31.5.2.2.1 [ios.failure] Status: NAD Submitter: Eelis van der Weegen Opened: 2011-08-16 Last modified: 2023-02-07

Priority: Not Prioritized

View all other issues in [std.exceptions].

View all issues with NAD status.

Discussion:

This is an extension issue for LWG to add constructor overloads that take a string by an rvalue reference in order to move the string into the exception.

[2012, Kona]

Move to NAD.

This was discussed during C++11 standardization, and deemed (at the time) to be a conforming extension that vendors are free to add, but there seemed no need to call it out in the standard. Since then it has been noted that the rvalue-reference overloads do not give you the move-semantic guarantee the proposer is thought to be looking for, as in order to meet the requirements that copy constructors do not throw (for standard exceptions) the exceptions that store strings must actually store a reference-counted immutable string, rather than an std::string internally. Therefore, an rvalue-reference overload is going to have to allocate memory in exactly the same way as copying from a const string& argument.

Proposed resolution:


2079(i). Required pow() overloads

Section: 28.7 [c.math] Status: NAD Submitter: Steve Clamage Opened: 2011-08-29 Last modified: 2016-01-28

Priority: 3

View all other issues in [c.math].

View all issues with NAD status.

Discussion:

LWG issue 550 removed the functions:

float       pow(float, int);
double      pow(double, int);
long double pow(long double, int);

from header <cmath>. This change does not seem to be mentioned in Annex C, C.2.14.

Howard:

N3290 28.7 [c.math]/p11 says:

Moreover, there shall be additional overloads sufficient to ensure:

  1. If any argument corresponding to a double parameter has type long double, then all arguments corresponding to double parameters are effectively cast to long double.
  2. Otherwise, if any argument corresponding to a double parameter has type double or an integer type, then all arguments corresponding to double parameters are effectively cast to double.
  3. Otherwise, all arguments corresponding to double parameters are effectively cast to float.

From C99 7.12.7.4 we have:

double pow(double, double);

28.7 [c.math]/p11/b2 says that if the client calls pow(2.0f, 2), then the int for second argument causes the following effective call to be made:

pow(static_cast<double>(2.0f), static_cast<double>(2)) -> double

The first sentence of p11 implies that this is done by supplying the following additional overload:

double pow(float, int);

If the client calls pow(2.0, 2), then the same reasoning (b2 again) implies the following additional overload:

double pow(double, int);

If the client calls pow(2.0l, 2), then b1 implies the following additional overload:

long double pow(long double, int);

In all, p11 implies hundreds (perhaps thousands?) of extra overloads. All but one of which is a superset of the overloads required by C++98/03 (that one being pow(float, int) which had its return type changed from float to double).

In practice, at least some vendors implement p11 by using templated overloads as opposed to ordinary overloads.

Steve Clamage:

Thanks. I didn't see that those extra overloads were actually implied by p11, despite the first sentence. Without examples, the point is a bit subtle (at least for me).

[2015-05-05 Lenexa: Move to NAD]

Billy: I believe this is NAD.

STL: Oh, Steve himself agrees.

Wakely: The issue marked as NAD will be sufficient.

STL: Yes, we should get rid of this.

Billy: I don't see any minutes from Issaquah.

Marshall: Since Steve agrees, does anyone object to marking as NAD?

Nope.

Proposed resolution:


2082(i). Misleading complexity requirements in <algorithm>

Section: 27 [algorithms] Status: NAD Submitter: Nicolai Josuttis Opened: 2011-09-02 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [algorithms].

View all other issues in [algorithms].

View all issues with NAD status.

Discussion:

The partition_point() algorithm is specified with:

Complexity: O(log(last - first)) applications of pred.

While this is correct, it gives the impression that this is a logarithmic algorithm. But unless random access iterators are used it is not logarithmic because for advancing the iterator we have last-first steps, which means that the complexity becomes linear here.

Shouldn't we clarify the complexity here to something like:

Complexity: logarithmic for random-access iterators and linear otherwise (in any case O(log(last - first)) applications of pred).

Or should we even require O(log(last - first) for random-access iterators only because for other iterators just iterating over all elements, while calling pred for each element, might often be faster.

Daniel Krügler:

I agree that especially this description is a bit misleading. I'm not convinced that this is a real defect, because the whole bunch of templates within <algorithm> document the complexity solely of applications* of predicates, assignment, or swaps, but never the complexity of traversal operations (e.g. increment or iterator equality tests). This means, the standard is consistent for this function template, even though it could say a bit more.

I would like to see a wording improvement, but I would rather think that the complexity of the predicate should be mentioned first (as in other algorithms) and that a non-normative note could be added for specifically this algorithm to point out that this does not imply a logarithmic traversal complexity. The note could give more details, by explicity pointing out the linear traversal complexity for non-random-Access iterators.

Howard Hinnant:

If we are going to make such improvements, they should be made across the board in <algorithm>, not to just partition_point. For example all 4 algorithms in 27.8.4 [alg.binary.search] have the same issue, and have since C++98.

stable_partition and inplace_merge should be inspected as well.

Perhaps a new paragraph in 27.1 [algorithms.general], similar to p12 would be a better place to address this issue.

[2012, Kona]

Move to NAD.

There was some concern that the issue, if it were to be addressed, is much larger than the couple of algorithms called out here, and applies across the whole library. There is no interest in looking at this or similar issues without a paper addressing the whole library. In fairness to anyone considering writing such a paper, it should be noted that there was not much interest in such a paper in the group, although no strong opposition either.

Proposed resolution:


2084(i). basic_string use of charT*

Section: 23.4.3 [basic.string] Status: NAD Submitter: Howard Hinnant Opened: 2011-09-11 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [basic.string].

View all other issues in [basic.string].

View all issues with NAD status.

Discussion:

For C++11 we gave all of the containers, including basic_string new generalized pointer types:

typedef typename allocator_traits<Allocator>::pointer       pointer:
typedef typename allocator_traits<Allocator>::const_pointer const_pointer;

However, the constructors, assignment, and member functions still traffic exclusively in terms of const charT*, for example:

basic_string(const charT* s, const Allocator& a = Allocator());

Was this an oversight? Did we mean instead:

basic_string(const_pointer s, const Allocator& a = Allocator());

Rationale:

It's intentional. char_traits assumes that all elements of a string can be accessed indirect on plain pointers. So basic_string doesn't support allocators with fancy pointers or references. And we meant to do that.

Let's take the constructor example you called out:

basic_string(const charT* s, const Allocator& a = Allocator());

This constructor allows us to create a basic_string object from a string literal. If we were to change the pointer type, that would no longer be possible.

There is no issue here, as the implementation of the constructor must make a copy of the string pointed-to by the pointer 's' rather than adopt ownership of that buffer. It is that internal copy that must make use of the allocator::pointer type.

Now what about the return value of 'c_str()', should that return an allocator::pointer?

Again, the answer (I believe) is 'no' because this is the function that allows us to pass the string's contents to a legacy/OS 'C' API. It is deliberately returning a raw pointer for a reason.

There was an issue where vector::data was changed to return an allocator::pointer to the internal buffer, and this was changed back exactly because this was intended to support passing to external APIs.

Do we have a use-case where the pointer type of internal data structures of our containers (notably basic_string and vector) need to be exposed through a public API? All my current use-cases for allocator::pointer are specific to the implementation of containers themselves.

Proposed resolution:


2090(i). Minor Overconstraint in Mutex Types

Section: 33.6.4.2 [thread.mutex.requirements.mutex] Status: NAD Submitter: Pete Becker Opened: 2011-10-17 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.mutex.requirements.mutex].

View all issues with NAD status.

Discussion:

33.6.4.2 [thread.mutex.requirements.mutex]/6, fourth bullet requires the return type of m.lock() to be void.

This is over-constrained. The true requirement is that the standard library ignores any value that the function returns. Yes, allowing non-void return types means that users can't store a pointer to this member function. No, that's not the least bit important.

[See also the discussion following c++std-lib-31318]

[2012, Kona]

This does not specify a concept; it specifies requirements on the concrete mutex types.

Proposed resolution:


2107(i). Some iterator category should guarantee the lifetime of references

Section: 25.3 [iterator.requirements] Status: NAD Submitter: Jeffrey Yasskin Opened: 2011-11-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [iterator.requirements].

View all issues with NAD status.

Discussion:

Many iterators guarantee that references and pointers returned from their methods will outlive the iterator itself. Other useful iterators can't guarantee this, leading to the rule in 25.3 [iterator.requirements] p9 that "Destruction of an iterator may invalidate pointers and references previously obtained from that iterator."

Some algorithms can take advantage of long-lived references by returning them, while they can adapt to short-lived references by returning by value instead. However, there doesn't seem to be a way in the standard to distinguish between these two types of iterators.

The ForwardIterator requirements come close by saying "If a and b are both dereferenceable, then a == b if and only if *a and *b are bound to the same object." (25.3.5.5 [forward.iterators] p6) However, there are some subtle ways to satisfy this rule and still return a short-lived reference, meaning algorithms can't be guaranteed that forward_iterator_tag will imply long-lived references.

On the other hand, defect 198, which added the invalidation wording to iterator.requirements.general, refers to iterators with short-lived references being used as arguments to reverse_iterator, which requires BidirectionalIterators. If ForwardIterator required long-lived references, this would be impossible.

Either ForwardIterator should be clarified to require long-lived references, or a new category should be added that does.

See also the discussion around c++std-lib-31477.

Daniel: Related to this issue is that when applying N3066 we unintentionally lost some forward iterator requirements from C++03, where we had the post-conditions a == X(a) of X(a), and u == a of any copy operation from a to u. This wording must be restored as well.

[2012, Kona]

Move to NAD.

This issue affects only Input Iterators, as all other categories are required to return a native reference, and are not (currently) allowed to return proxies. The issue with Input Iterators is known, and has been present since the original standard. Any change in this regard would be an extension requiring a more substantial paper than treatment as a simple issue.

Proposed resolution:


2113(i). Do library implementers have the freedom to add final to non-polymorphic components?

Section: 16.4.6 [conforming] Status: NAD Submitter: Daniel Krügler Opened: 2011-11-30 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [conforming].

View all issues with NAD status.

Discussion:

Related to LWG 2112 the question has been raised whether a library implementation may declare non-polymorphic library components, such as class template std::vector or std::basic_string, as final class types.

This issue is not suggesting to enforce that libraries are required to do that, it is asking whether libraries should have the freedom to do so.

The existing wording in 16.4.6.12 [derivation] does not give a clear permission to do so. In my opinion this position should be clarified in either direction.

Giving implementations this freedom would have both advantages and disadvantages. Several opponents where worried about breakage of code of existing user implementations. On the other hand such types where not designed to be used as base classes. Allowing implementations to mark these components as final could allow them to provide compile-modes that are intentionally restrictive to the advantage of user code that want to be alterted about that. Any implementation that would be concerned about user complaints would not take advantage of this feature anyway.

If agreement exists that such implementation freedom would be useful, wording like

An implementation may declare additional non-virtual member function signatures within a class as final.

or

An implementation may declare additional class without virtual member function signatures as final.

should be added to 16.4.6 [conforming] with corresponding exceptions of these rules (e.g. iterator, unary_function, or pair).

If such freedom should not exist, it seems better to clarify this as well, e.g. by adding around 16.4.6.12 [derivation]:

An implementation shall not declare any class or any member function signature as final.

[2012, Kona]

Move to NAD.

Unless the library uses the keyword final in a specification, the user clearly has freedom to derive from such a class, and so equally clearly, the library vendor does not have freedom to add a final overrider or class attribute. Howard observed there may be some wiggle-room with 'unspecified types' such as those returned from bind expressions, or iterators, but we did not see a need to further clarify the issue. Note that, for example, a vector::iterator may be implemented as a raw pointer, so users cannot generally assume the ability to derive from unspecified library types.

Proposed resolution:


2121(i). app for string streams

Section: 31.8.5.2 [stringstream.cons] Status: NAD Submitter: Nicolai Josuttis Opened: 2012-01-15 Last modified: 2018-06-07

Priority: 3

View all issues with NAD status.

Discussion:

This issue was raised while discussing issue 1448.

Note the following program:

string s("s1: 123456789");
ostringstream s1(s, ios_base::out|ios_base::app);
s1 << "hello";
cout << s1.str() << endl;

With g++4.x it prints:

s1: 123456789hello

With VisualC++10 it prints:

hello23456789

From my intuitive understanding the flag "app" should result in the output of g++4.x. I also would read that from [ios::openmode] claiming:

app   seek to end before each write

However in issue 1448 P.J.Plauger comments:

I think we should say nothing special about app at construction time (thus leaving the write pointer at the beginning of the buffer). Leave implementers wiggle room to ensure subsequent append writes as they see fit, but don't change existing rules for initial seek position.

Note that the flag ate on both platforms appends "hello" to s.

[2018-06; Rapperswil Wednesday evening]

MC: shouldn't this be Open?
BO: I think we changed it to do the same thing as gcc
JM: NAD - gcc is right
ACTION close as NAD, noting that gcc is right

Proposed resolution:


2124(i). Seed sequence over-specified

Section: 28.5.3.2 [rand.req.seedseq] Status: NAD Submitter: Alberto Ganesh Barbati Opened: 2012-01-16 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.req.seedseq].

View all issues with NAD status.

Discussion:

The seed sequence requirements described in 28.5.3.2 [rand.req.seedseq] appear to be over-specified. All seed sequence types are required to have a result_type nested type, a specific set of constructors, function members size() and param(), which are never used by the library. In fact, the only library components that actively use seed sequences are the random engines and all the engines need is the generate() member function. In particular, library components never attempts to construct seed sequence objects. These extraneous requirements are clearly written to describe the library provided type seed_seq type; while it's good that seed_seq has all those constructors and members, it's not a compelling reason to require a user-provided seed sequence type to implement all of them.

Suppose I want to write my own seed sequence class, this should do fine (and actually works as expected with libc++):

class my_seed_seq
{
  /* internals */
public:
  my_seed_seq(/* my own parameters */);

  template <class It>
  void generate(It first, It last);
};

my_seed_seq s(/* params */);
std::default_random_engine e(s);

The only reason to have these extra members would be to provide some support for generic serializability/persistence of seed sequence objects. I believe that would be out of the scope of the random library, so I doubt we will ever need those requirements in the future.

I therefore propose to remove all requirements from 28.5.3.2 [rand.req.seedseq] except for the presence of the generate() function.

[2012, Kona]

Move to Tenatively NAD. (Tentative as issue was not in pre-meeting mailing)

The 'overspecification', as such, was a deliberate intent to provide guarantees consumers of the whole random number framework may rely upon, especially in generic code. While the standard engines may be built without relying on these guarantees, this specification is part of a commitment to a broader framework, and Walter indicated future proposals in preparation for parallel generation of random numbers that may depend more inimately on these existing requirements.

Alisdair noted that the result_type typedef was a call-back to how we used to specify adaptable functors before TR1 result_of and the addition of std::bind and is probably not something we should be actively promoting in future libraries. However, it is too late to remove this requirement from seed sequences unless we are doing further surgery, as recommended by this issue.

Walter notes that the result_type protocol has not been formally deprecated by the standard. Alisdair replies that was the intent of deprecating the bind_1st/ unary_function set of templates in C++11, although we did not say anything about result_type in general.

Proposed resolution:

This wording is relative to the FDIS.

  1. Edit 28.5.3.2 [rand.req.seedseq] p2 as indicated:

    A class S satisfies the requirements of a seed sequence if the expressions shown in Table 115 are valid and have the indicated semantics, and if S also satisfies all other requirements of this section 28.5.3.2 [rand.req.seedseq]. In that Table and throughout this section:

    1. T is the type named by S's associated result_type;
    2. q is a value of S and r is a possibly const value of S; and
    3. ib and ie are input iterators with an unsigned integer value_type of at least 32 bits;
    4. rb and re are mutable random access iterators with an unsigned integer value_type of at least 32 bits;
    5. ob is an output iterator; and
    6. il is a value of initializer_list<T>.
  2. Ditto, in Table 115, remove all rows except the one describing q.generate(rb, re):

    Table 115 — Seed sequence requirements
    Expression Return type Pre/Post-condition Complexity
    S::result_type T T is an unsigned integer
    type (6.8.2 [basic.fundamental]) of at least 32 bits.
    compile-time
    S()   Creates a seed sequence with
    the same initial state as all
    other default-constructed seed
    sequences of type S.
    constant
    S(ib,ie)   Creates a seed sequence having
    internal state that depends on
    some or all of the bits of the
    supplied sequence [ib, ie).
    𝒪(ie - ib)
    S(il)   Same as S(il.begin(),
    il.end())
    .
    same as
    S(il.begin(),
    il.end())
    q.generate(rb,re) void Does nothing if rb == re.
    Otherwise, fills the supplied
    sequence [rb, re) with 32-bit
    quantities that depend on the
    sequence supplied to the
    constructor and possibly also
    depend on the history of
    generate's previous
    invocations.
    𝒪(re - rb)
    r.size() size_t The number of 32-bit units that
    would be copied by a call to
    r.param.
    constant
    r.param(ob) void Copies to the given destination a sequence of 32-bit units that
    can be provided to the
    constructor of a second object
    of type S, and that would
    reproduce in that second object
    a state indistinguishable from
    the state of the first object.
    𝒪(r.size())

2125(i). TimedMutex specification problem

Section: 33.6.4.3 [thread.timedmutex.requirements], 33.6.4.3.2 [thread.timedmutex.class] Status: Pending NAD Editorial Submitter: Vicente J. Botet Escriba Opened: 2012-01-01 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.timedmutex.requirements].

View all issues with Pending NAD Editorial status.

Discussion:

33.6.4.3.2 [thread.timedmutex.class] says:

The class timed_mutex shall satisfy all of the TimedMutex requirements (33.6.4.3 [thread.timedmutex.requirements]). It shall be a standardlayout class (Clause 11 [class]).

Problem here is that 33.6.4.3 [thread.timedmutex.requirements] does not define a requirement set named "TimedMutex", it only refers to "timed mutex types"

[See also issue 2126]

[2012, Portland: move to Tentatively NAD Editorial]

We have timed mutex type, but it is labeled timed mutex requirements

We can make a suggestion, but will send to the editor as it seems purely editorial. There is a typo, and we don't have the timed mutex but 33.6.4.3 [thread.timedmutex.requirements] already says timed mutex type, and we need to reuse that term down in the class to fulfil the mutex requirement.

[To Editor:]

Replace this one with timed mutex type.

Proposed resolution:


2126(i). Several specification problems in regard to mutex requirements

Section: 33.6.4 [thread.mutex.requirements], 33.6.4.2.2 [thread.mutex.class], 33.6.4.2 [thread.mutex.requirements.mutex], 33.6.4.2.3 [thread.mutex.recursive], 33.6.4.3 [thread.timedmutex.requirements], 33.6.4.3.2 [thread.timedmutex.class], 33.6.4.3.3 [thread.timedmutex.recursive] Status: Pending NAD Editorial Submitter: Pete Becker Opened: 2012-01-16 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [thread.mutex.requirements].

View all other issues in [thread.mutex.requirements].

View all issues with Pending NAD Editorial status.

Discussion:

33.6.4.2.2 [thread.mutex.class]/3 says that the class mutex "shall satisfy all the Mutex requirements (33.6.4 [thread.mutex.requirements])". 33.6.4.2.2 [thread.mutex.class] is part of 33.6.4 [thread.mutex.requirements], so at the very least, this requirement is recursive. But worse, there is nothing that says what "the Mutex requirements" refers to. For example, the "Lockable requirements" section starts with "A type L meets the Lockable requirements if …". There is no such statement for "the Mutex requirements".

Organizationally, paragraphs 1-26 in 33.6.4.2 [thread.mutex.requirements.mutex] should probably be in a subclause with a name. (This is actually an ISO requirement, to avoid exactly this kind of ambiguous referencing) Then the first sentence of 33.6.4.2.2 [thread.mutex.class]/3 can become a note: "The class mutex meets the requirements of (whatever)", since that subclause already says that the mutex types "shall meet the requirements set out in this section."

And similarly for 33.6.4.2.3 [thread.mutex.recursive]/2 (recursive_mutex).

33.6.4.3 [thread.timedmutex.requirements], Timed mutex types, also needs the same rearrangement: its introductory requirements should be moved into a subclause, and the first sentences of 33.6.4.3.2 [thread.timedmutex.class]/2 and 33.6.4.3.3 [thread.timedmutex.recursive]/2 should be turned into notes that refer to this new subclause and to the new subclause in 33.6.4.2 [thread.mutex.requirements.mutex].

[See also issue 2125]

[2012, Portland: move to Tentatively NAD Editorial]

Seems no real ambiguity. May need some reorg of text rather then changing the wording.

Is there much that needs to be changed? But Pete's suggestion of putting requirement in separate sub section is good. Should be the direction to editor.

Suggest this is an editorial change. Happy with Pete's comments.

Proposed resolution:


2131(i). Member function getline taking a string as parameter

Section: 31.7.5.4 [istream.unformatted] Status: NAD Submitter: Loïc Joly Opened: 2012-03-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [istream.unformatted].

View all issues with NAD status.

Discussion:

I think the following code should be legal:

void f(std::istream& is)
{
  std::string s;
  is.getline(s); // Would be equivalent to std::getline(is, s)
}

[2013-04-20, Bristol]

Unanimous that this is a new feature request and not a issue.

Resolution: Tentatively NAD

Proposed resolution:

This wording is relative to N3376.

  1. Change the class template basic_istream synopsis, 31.7.5.2 [istream], as indicated

    namespace std {
      template <class charT, class traits = char_traits<charT> >
      class basic_istream : virtual public basic_ios<charT,traits> {
      public:
        […]
        // 27.7.2.3 Unformatted input:
        […]
        basic_istream<charT,traits>& getline(char_type* s, streamsize n);
        basic_istream<charT,traits>& getline(char_type* s, streamsize n,
          char_type delim);
        template<class Allocator>
        basic_istream<charT,traits>& getline(basic_string<charT,traits,Allocator>& str);
        template<class Allocator>
        basic_istream<charT,traits>& getline(basic_string<charT,traits,Allocator>& str,
          char_type delim);
        […]
      };
    }
    
  2. Insert the following two new prototype descriptions after 31.7.5.4 [istream.unformatted] paragraph 24:

    basic_istream<charT,traits>& getline(char_type* s, streamsize n);
    

    -24- Returns: getline(s,n,widen('\n'))

    template<class Allocator>
    basic_istream<charT,traits>& getline(basic_string<charT,traits,Allocator>& str);
    

    -??- Returns: std::getline(*this, str)

    template<class Allocator>
    basic_istream<charT,traits>& getline(basic_string<charT,traits,Allocator>& str, char_type delim);
    

    -??- Returns: std::getline(*this, str, delim)


2134(i). Redundant Mutex requirement?

Section: 33.6.4.2 [thread.mutex.requirements.mutex] Status: Pending NAD Editorial Submitter: Pete Becker Opened: 2012-03-05 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [thread.mutex.requirements.mutex].

View all issues with Pending NAD Editorial status.

Discussion:

33.6.4.2 [thread.mutex.requirements.mutex]/11 says that prior unlock operations synchronize with m.lock().

33.6.4.2 [thread.mutex.requirements.mutex]/19 says that if m.try_lock() succeeds, prior unlock operations synchronize with the operation.

33.6.4.2 [thread.mutex.requirements.mutex]/25 says that m.unlock() synchronizes with subsequent successful lock operations.

Does the third requirement add anything to the first two? If not, it should probably be a non-normative note.

[2012, Portland: move to Tentatively NAD Editorial]

Agree that third note should be non-normative and adds nothing.

Seems An Editorial change, but does changing a normative to non-normative wording makes it a non-editorial change?

Ask the editor. If not editorial, then we will agree on the fix as removal of the third point, then we will put it in ready state for Bristol.

Proposed resolution:


2161(i). const equivalence of std::map

Section: 24.4 [associative], 24.5 [unord] Status: NAD Submitter: Bjarne Stroustrup Opened: 2012-06-18 Last modified: 2016-08-02

Priority: 2

View all other issues in [associative].

View all issues with NAD status.

Discussion:

As described in the reflector discussion c++std-core-21860 consider the following example:

map<const int, int> mci{};
map<int, int> mi = mci; // ??
mci[1] = 2;
mi[1] = 2;

Should it be required that the marked initialization is well-formed? As a possible solution this could be realized by an alias template:

template <class K, class T>
struct OriginalMap { […] };

template <class K, class T>
using ImprovedMap = OriginalMap<const K, T>;

[2016-08 Chicago]

Recommend NAD. No other container supports conversion from different container specializations, so adding support directly might be more surprising than the omission.

We would welcome papers on the convertibility of containers of different element types directed to the LEWG in the future.

Proposed resolution:


2167(i). Copy assignment requirements of Containers

Section: 24.2.2.1 [container.requirements.general] Status: NAD Submitter: Dean Michael Berris Opened: 2012-07-13 Last modified: 2016-01-28

Priority: Not Prioritized

View other active issues in [container.requirements.general].

View all other issues in [container.requirements.general].

View all issues with NAD status.

Discussion:

Table 96 defines the general requirement for copy assignment (row 23, page 704) as:

Table 96 — Container requirements
Expression Return type Operational semantics Assertion/note pre-/post-condition Complexity
r = a X& post: r == a. linear

However there is no requirement that T is CopyInsertable into X.

[2012, Portland: Move to Tentatively NAD]

Howard notes that this may be a difficult requirement for std::array

We already have this requirement for allocator aware containers, and std::array already adds the appropriate extra requirement.

We say the necessary things in the necessary places, but the container requirements continue to cause confusion in where we sometimes say things. Consensus is that this issue remains NAD though.

Proposed resolution:

This wording is relative to N3376.

  1. Change Table 96 — "Container requirements" in 24.2.2.1 [container.requirements.general]:

    Table 96 — Container requirements
    Expression Return type Operational semantics Assertion/note pre-/post-condition Complexity
    r = a X& Requires: T is CopyInsertable into X.
    post: r == a.
    linear

2171(i). "swappable" undefined for swapping lvalue and rvalue

Section: 16.4.4.3 [swappable.requirements] Status: NAD Submitter: Robert Shearer Opened: 2012-07-24 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [swappable.requirements].

View all issues with NAD status.

Discussion:

Paragraph 16.4.4.3 [swappable.requirements] p4 states:

An rvalue or lvalue t is swappable if and only if t is swappable with any rvalue or lvalue, respectively, of type T.

This paragraph seems to establish two disjoint definitions of "swappable" — one for lvalues and one for rvalues — with neither definition including the case of swapping an rvalue with an lvalue.

Resolution proposal:

Delete the word "respectively".

[ 2012-10 Portland: Close as NAD ]

The current wording does intentionally specify two families of 'swappable' behaviors, for lvalues and for rvalues, and not for mixed behavior. The need to support rvalues is for types like vector<bool>::reference. Likewise, library types like string provide a swap for values, but not a mixed-mode swap between lvalues and rvalues, which were deliberately removed from C++11 after initally being part of the standard.

Accepting this resolution would break the library specification, as no current library type would meet the new requirements.

Proposed resolution:

This wording is relative to N3376.

Change 16.4.4.3 [swappable.requirements] p4 as indicated:

An rvalue or lvalue t is swappable if and only if t is swappable with any rvalue or lvalue , respectively, of type T.


2178(i). Allocator requirement changes not mentioned Annex C

Section: 16.4.4.6 [allocator.requirements], C.7 [diff.library] Status: Pending NAD Editorial Submitter: Nevin Liber Opened: 2012-08-14 Last modified: 2016-11-21

Priority: 3

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with Pending NAD Editorial status.

Discussion:

Given that a number of things were removed from the allocator requirements (reference, const_reference, address() in 16.4.4.6 [allocator.requirements]), it seems that these incompatible changes should be mentioned in Annex C.7 [diff.library], more specifically in [diff.cpp03].

[ 2012-10 Portland: Move to Open ]

It was clearly pointed out by Bill during the C++11 process that our change to allocator requirements potentially broke 3rd party user containers written to expect C++03 allocators, or rather, an allocator written to the minimal requirements of C++11 might not be guaranteed to work with a container written to the previous rules. This was a trade-off in making allocaters easier to write by use of the allocator_traits framework.

This probably does merit a write-up in Annex C, and we look forward to seeing wording. Until then, the best we can do is move the issue to Open.

[2016-11-12, Issaquah]

Sat PM: JW to provide Annex C wording Status to "Pending NAD Editorial".

Proposed resolution:


2199(i). unordered containers are required to have an initial max load factor of 1.0

Section: 24.2.8 [unord.req] Status: NAD Submitter: Alisdair Meredith Opened: 2012-10-09 Last modified: 2016-11-21

Priority: 3

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with NAD status.

Discussion:

The default constructor, allocator-aware constructor, and range-based constructors for the unordered containers do not offer a means to control the initial max_load_factor, so the standard mandates the value 1.0. This seems overly restrictive, as there is plenty of research suggesting a value between 0.5 and 1.0 is more often optimal for unique-key containers, and perhaps a slightly higher value might be appropriate for multi-containers.

Rather than guess at the appropriate max_load_factor, it seems reasonable that the standard should allow vendors to pick a value at their discretion, with perhaps a note of advice. It is less clear whether the default value should be implementation-defined or unspecified, given the ease of a user determining this by querying this attribute immediately after construction.

[2013-03-15 Issues Teleconference]

Moved to Open.

Alisdair to provide wording.

Marshall: It seems to me that what you really want is to be able to pass a max load factor in the constructor, but that's a different issue.

Alisdair agrees in principle, but concerned with adding yet more constructors to these classes.

[2016-11-12, Issaquah]

NAD - if someone wants to come back with a paper exploring design alternatives, we can re-open.

Proposed resolution:


2201(i). Missing macro entries from C standard library

Section: C.7 [diff.library] Status: NAD Editorial Submitter: Kevin McCarty Opened: 2012-02-03 Last modified: 2016-11-28

Priority: 2

View all other issues in [diff.library].

View all issues with NAD Editorial status.

Discussion:

It seems that in C.7 [diff.library], Table 150 the following macros from [c.limits], Table 31 are missing:

LLONG_MIN 
LLONG_MAX
ULLONG_MAX

In addition in C.7 [diff.library], Table 150 the following macros from [c.limits], Table 32 are missing:

DECIMAL_DIG 
FLT_EVAL_METHOD

Furtheron it seems that in C.7 [diff.library], Table 149/150 further macros are missing as well, e.g. HUGE_VALF, INFINITY, etc.

[2014-02 Issaquah:]

This is an issue, all of C has not been updated for C99, C99 functions are missing, whole section needs to be overhauled.

The issue needs to be updated for functions and other missing items and when that happens the issue title is wrong and needs to be adapted.

[2016-08 Chicago: Zhihao Yuan comments]

We suggest to either:

  1. Having a paper to rewrite C.5 C standard library by striking the content shared by C and C++ and leave only the differences;

  2. Having a paper, or a complete proposed wording to fix the missing names added between C89 and C99.

For 1), some C11 headers being excluded from C++ (such as thread.h) need to be added to this section.

[2016-11-12, Issaquah]

Sat PM: Thomas will attempt to fix this editorially If he cannot, we will reexamine.

[2016-11-16]

Resolved as NAD Editorial

Proposed resolution:


2204(i). reverse_iterator should not require a second copy of the base iterator

Section: 25.5.1.6 [reverse.iter.elem] Status: NAD Submitter: David Abrahams Opened: 2012-10-30 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [reverse.iter.elem].

View all issues with NAD status.

Discussion:

This note in [reverse.iter.op.star]/2:

[ Note: This operation must use an auxiliary member variable rather than a temporary variable to avoid returning a reference that persists beyond the lifetime of its associated iterator. (See 24.2.) —end note ]

is incorrect because such iterator implementations are ruled out by 25.3.5.5 [forward.iterators]/6, where it says:

If a and b are both dereferenceable, then a == b if and only if *a and *b are bound to the same object.

[2013-04-20, Bristol]

Alisdair: sugested to strike the "exposition only" member.

Daniel: we must check that it wouldn't conflict with a previous solution to another issue.

Dietmar: This is an issue but the proposing word is not correct. When we have proxies inside the sequence.

Solution: NAD thanks to a contrieved example by Dietmar.

Proposed resolution:

Strike the note, [reverse.iter.op.star]/2:

[ Note: This operation must use an auxiliary member variable rather than a temporary variable to avoid returning a reference that persists beyond the lifetime of its associated iterator. (See 24.2.) —end note ]

2226(i). wstring_convert methods do not take allocator instance

Section: D.27.2 [depr.conversions.string] Status: NAD Submitter: Glen Fernandes Opened: 2012-12-14 Last modified: 2019-02-26

Priority: Not Prioritized

View other active issues in [depr.conversions.string].

View all other issues in [depr.conversions.string].

View all issues with NAD status.

Discussion:

The wstring_convert class template, described in D.27.2 [depr.conversions.string], does not support custom stateful allocators. It only supports custom stateless allocators.

The to_bytes member function returns basic_string<char, char_traits<char>, Byte_alloc> but it does not take an instance of Byte_alloc to pass to the constructor of the basic_string.

Similarly the from_bytes member function returns basic_string<Elem, char_traits<Elem>, Wide_alloc> but it does not take an instance of Wide_alloc to pass to the constructor of the basic_string.

This makes these two member functions and the wstring_convert class template not usable when Wide_alloc or Byte_alloc are stateful allocators.

[2013-01-22, Glen provides wording]

[2013-03-15 Issues Teleconference]

Moved to NAD Future.

This is clearly an extension that the LEWG may want to take a look at, once we have more experience with appropriate use of allocators with the C++11 model.

[LEWG Kona 2017]

Recommend NAD: Does this follow the pattern? Should be discussed as a group. Do we have the experience with the C++11 allocator model to know that this is the addition to make?

Should to_string() also take an allocator? substr()? Any function that returns a string?

This suggests a larger change.

[Kona 2019]

Jonathan points out: The wstring_convert type is deprecated now.

Proposed resolution:

This wording is relative to N3485.

  1. In D.27.2 [depr.conversions.string]/2 and /6 "Class template wstring_convert synopsis" change the overloads of the member function from_bytes() so that all four overloads take an additional parameter which is an instance of Wide_alloc:

    wide_string from_bytes(char byte, const Wide_alloc& alloc = Wide_alloc());
    wide_string from_bytes(const char *ptr, const Wide_alloc& alloc = Wide_alloc());
    wide_string from_bytes(const byte_string& str, const Wide_alloc& alloc = Wide_alloc());
    wide_string from_bytes(const char *first, const char *last, const Wide_alloc& alloc = Wide_alloc());
    
  2. In D.27.2 [depr.conversions.string] /8 specify that this Wide_alloc allocator parameter is used to construct the wide_string object returned from the function:

    -7- Effects: The first member function shall convert the single-element sequence byte to a wide string. The second member function shall convert the null-terminated sequence beginning at ptr to a wide string. The third member function shall convert the sequence stored in str to a wide string. The fourth member function shall convert the sequence defined by the range [first, last) to a wide string.

    -8- In all cases:

  3. In D.27.2 [depr.conversions.string]/2 and /12 "Class template wstring_convert synopsis" change the overloads of the member function to_bytes() so that all four overloads take an additional parameter which is an instance of Byte_alloc:

    byte_string to_bytes(Elem wchar, const Byte_alloc& alloc = Byte_alloc());
    byte_string to_bytes(const Elem *wptr, const Byte_alloc& alloc = Byte_alloc());
    byte_string to_bytes(const wide_string& wstr, const Byte_alloc& alloc = Byte_alloc());
    byte_string to_bytes(const Elem *first, const Elem *last, const Byte_alloc& alloc = Byte_alloc());
    
  4. In D.27.2 [depr.conversions.string] /13 specify that this Byte_alloc allocator parameter is used to construct the byte_string object returned from the function:

    -12- Effects: The first member function shall convert the single-element sequence wchar to a byte string. The second member function shall convert the null-terminated sequence beginning at wptr to a byte string. The third member function shall convert the sequence stored in wstr to a byte string. The fourth member function shall convert the sequence defined by the range [first, last) to a byte string.

    -13- In all cases:


2242(i). [uninitialized_]copy_n() defect

Section: 27.7.1 [alg.copy], 27.11.5 [uninitialized.copy] Status: NAD Submitter: Sean Parent Opened: 2013-02-14 Last modified: 2017-06-05

Priority: 2

View other active issues in [alg.copy].

View all other issues in [alg.copy].

View all issues with NAD status.

Discussion:

copy_n() and uninitialized_copy_n() only return the output iterator, and not the input iterator. Likely the interface was simply copied from the original STL. Unfortunately the interface in the original STL contains a bug.

copy_n() and uninitialized_copy_n() must return the resulting input iterator as well as the output iterator (I would suggest returning a pair). Without this, there is no way to continue reading from an actual input iterator — and if it is really a forward iterator, it will cost n increments to get back to where you were.

[2016-08 Chicago]

Tues PM: refer to LEWG

[2017-02 in Kona, LEWG recommends NAD]

Potentially breaks lots of things. Cannot overload on return type. Pair, tuple, or struct return? New name? Needs paper with a new algorithm (copy_n_works)? Not an issue for istream iterators (can get stream and get a new iterator). Don’t use this function, use ranges.

[2017-06-02 Issues Telecon]

This change has been made in the Ranges TS.

Resolve as NAD

Proposed resolution:


2251(i). C++ library should define ssize_t

Section: 17.2 [support.types] Status: NAD Submitter: Matt Austern Opened: 2013-04-19 Last modified: 2016-01-28

Priority: 3

View all other issues in [support.types].

View all issues with NAD status.

Discussion:

The C++ standard library defines size_t, a typedef for an implementation defined unsigned integer type that can represent the sizes of objects. The POSIX standard augments this with ssize_t, a typedef for a signed integer type that corresponds to size_t.

The ssize_t typedef is useful — useful enough that the C++ standard even refers to it. (In a non-normative footnote in 31.2.2 [stream.types].) Also, lots of OS vendors add it to their headers anyway, even though it isn't part of the C or C++ standards, because those vendors are trying to define headers that conform to multiple standards at once. We should make users' and implementers' lives easier by adding ssize_t to 17.2 [support.types].

[2013-09-29, Suggested wording from Jayson Oldfather]

I decided to use the phrase to describe ssize_t below because of the text describing it in the POSIX standard. In it, it describes ssize_t with the value range of [-1,{SSIZE_MAX}]. SSIZE_MAX is specified in the POSIX standard as a minimum value of _POSIX_SSIZE_MAX. This macro is referenced in the wording below.

[Lenexa 2015-05-05: NAD - no consensus for a change]

Billy : ssize_t that was promised to be signed, was based on rsize_t from safe secure C

NM : ssize_t s ptrdif_t

Z : ptrdiff_t is full range, ssize_t has only -1 as negative value

Billy : motivations for ptrdiff_T, ssize_t and rsize_T all fuzzy. - Reads rsize max -

NM : ptrdiff_T not big enough to rep difference of pointers anymore

STL : description incorporates posixisms

Billy : Don't need it

NM : rather remove it from footnote

Z : Name has precise meaning

STL : everyone understands ptrdiff_t is signed counterpart to size_t

Billy : Not in all implementations anymore

DK : footnote says something different from ...

Z/NM : off_t historically tainted

STL : we have a type trait to make signed version of size_t. we should just use that

MC : NAD; is feature request

TP : It's not cstdsef

Proposed resolution:

  1. Ammend 17.2 [support.types], Table 30 as indicated:

    Table 30 — Header <cstddef> synopsis
    Type Name(s)
    Macros: NULL offset_t
    Types: ptrdiff_t ssize_t size_t max_align_t nullptr_t

    Add the following paragraph to describe ssize_t

    -?- The type ssize_t is an implementation-defined signed integer type that shall contain the minimum range [-1, {SSIZE_MAX}] where SSIZE_MAX is specified at a minimum of _POSIX_SSIZE_MAX.

    Ammend p7 as follows:

    -7- [Note: It is recommended that implementations choose types for ptrdiff_t, ssize_t, and size_t whose integer conversion ranks …


2253(i). [arrays.ts] dynarray should state which container requirements aren't met

Section: 99 [arrays.ts::dynarray.overview] Status: NAD Arrays Submitter: Jonathan Wakely Opened: 2013-04-23 Last modified: 2016-10-31

Priority: 0

View all issues with NAD Arrays status.

Discussion:

Addresses: arrays.ts

99 [arrays.ts::dynarray.overview] p2 says:

"Unless otherwise specified, all dynarray operations have the same requirements and semantics as specified in 99 [arrays.ts::container.requirements]."

Some differences from 99 [arrays.ts::container.requirements] are not explicitly specified, including at least the lack of a default constructor, copy assignment and swap member.

The wording could be similar to 24.3.7.1 [array.overview] which says "An array satisfies all of the requirements of a container and of a reversible container (99 [arrays.ts::container.requirements]), except that a default constructed array object is not empty and that swap does not have constant complexity."

[2013-09 Chicago:]

Move to Deferred. This feature will ship after C++14 and should be revisited then.

[2014-06-06 pre-Rapperswil]

This issue has been reopened as arrays-ts.

[2014-06-16 Rapperswil]

Move to Ready

[2014/11 Urbana]

Held at Ready status, pending clarification of Arrays TS

Proposed resolution:

  1. Add to 99 [arrays.ts::dynarray.overview] p2:

    -2- A dynarray satisfies all of the requirements of a container and of a reversible container (99 [arrays.ts::container.requirements]), except for default construction, assignment and swap. Unless otherwise specified, all dynarray operations have the same requirements and semantics as specified in 99 [arrays.ts::container.requirements].


2254(i). [arrays.ts] Is dynarray an allocator-aware container?

Section: 99 [arrays.ts::container.requirements.general] Status: NAD Arrays Submitter: Jonathan Wakely Opened: 2013-04-23 Last modified: 2016-10-31

Priority: 3

View all issues with NAD Arrays status.

Discussion:

Addresses: arrays.ts

99 [arrays.ts::container.requirements.general] p3 says:

"All of the containers defined in this Clause and in (21.4) except array meet the additional requirements of an allocator-aware container, as described in Table 99."

Is this true of dynarray? I believe the answer must be no because dynarray has no allocator_type, and morally should be no, so that operations are defined in terms of std::allocator<T>, which p13 says doesn't actually need to be used (which allows the elements to be default-initialized as is intended, rather than "default-inserted into the container" using an allocator.)

The requirement that "each element is constructed with uses-allocator construction" provides roughly equivalent behaviour to the "CopyInsertable into X" requirements for allocator-aware containers, allowing an allocator to control construction of the dynarray elements.

[2013-09 Chicago]

Move to Deferred. This feature will ship after C++14 and should be revisited then.

[2014-06-06 pre-Rapperswill]

This issue has been reopened as arrays-ts.

Proposed resolution:

  1. Change to 99 [arrays.ts::container.requirements.general] p13:

    -13- All of the containers defined in this Clause and in (21.4) except array and dynarray meet the additional requirements of an allocator-aware container, as described in Table 99.


2255(i). [arrays.ts] dynarray constructor ambiguity

Section: 99 [arrays.ts::dynarray.cons] Status: NAD Arrays Submitter: Jonathan Wakely Opened: 2013-04-23 Last modified: 2016-03-08

Priority: 0

View all issues with NAD Arrays status.

Discussion:

Addresses: arrays.ts

These constructors can interact badly::

template<class Alloc>
  dynarray(size_type c, const Alloc& alloc);
dynarray(size_type c, const T& v);

Unless the second argument is a value of exactly the type T you will get the first constructor, i.e. all of these will fail to compile:

dynarray<long> dlong(1, 1);   // 1 is not long
dynarray<float> dflt(1, 1.0);  // 1.0 is not float
dynarray<int*> dptr(1, nullptr);  // nullptr is not int*
dynarray<void*> doh(1, 0);  // 0 is not void*

The nullptr case is particularly annoying, a user trying to do the right thing by saying nullptr instead of NULL still gets the wrong result.

The constructor taking an allocator requires that "Alloc shall meet the requirements for an Allocator" but doesn't actually say "shall not participate in overload resolution unless ..."

I believe we have no precedent for using SFINAE to check "the requirements for an Allocator" because it's a pretty complicated set of requirements. We could say it shall not participate in overload resolution if Alloc is implicitly convertible to value_type.

Alternatively, we could follow the same approach used by other types that can be constructed with an unconstrained allocator type and use std::allocator_arg_t as the first argument instead of adding an allocator after the other arguments.

[2013-09 Chicago:]

Move to Deferred. This feature will ship after C++14 and should be revisited then.

[2014-06-06 pre-Rapperswil]

This issue has been reopened as arrays-ts.

[2014-06-16 Rapperswil]

Move to Ready for alternative A

Previous resolution [SUPERSEDED]:

  1. Either use the correct way to unambiguously call a constructor taking any type of allocator, i.e. change the constructors to take dynarray(std::allocator_arg_t, const Alloc&, ...) by modifying both the synopsis 99 [arrays.ts::dynarray.overview] p2 and 99 [arrays.ts::dynarray.cons] before p9 like so:

    template <class Alloc>
      dynarray(allocator_arg_t, const Alloc& a, size_type c, const Alloc& alloc);
    template <class Alloc>
      dynarray(allocator_arg_t, const Alloc& a, size_type c, const T& v, const Alloc& alloc);
    template <class Alloc>
      dynarray(allocator_arg_t, const Alloc& a, const dynarray& d, const Alloc& alloc);
    template <class Alloc>
      dynarray(allocator_arg_t, const Alloc& a, initializer_list<T>, const Alloc& alloc);
    
  2. or constrain the problematic constructor by adding a new paragraph to 99 [arrays.ts::dynarray.cons]:

    template <class Alloc>
      dynarray(size_type c, const Alloc& alloc);
    template <class Alloc>
      dynarray(size_type c, const T& v, const Alloc& alloc);
    template <class Alloc>
      dynarray(const dynarray& d, const Alloc& alloc);
    template <class Alloc>
      dynarray(initializer_list<T>, const Alloc& alloc);
    

    -9- Requires: Alloc shall meet the requirements for an Allocator (16.4.4.6 [allocator.requirements]).

    -10- Effects: Equivalent to the preceding constructors except that each element is constructed with uses-allocator construction (20.2.8.2 [allocator.uses.construction]).

    -?- Remarks: The first constructor shall not participate in overload resolution unless Alloc is not implicitly convertible to T.

[2014/11 Urbana]

Held at Ready status, pending clarification of Arrays TS

Proposed resolution:

  1. Use the correct way to unambiguously call a constructor taking any type of allocator, i.e. change the constructors to take dynarray(std::allocator_arg_t, const Alloc&, ...) by modifying both the synopsis 99 [arrays.ts::dynarray.overview] p2 and 99 [arrays.ts::dynarray.cons] before p9 like so:

    template <class Alloc>
      dynarray(allocator_arg_t, const Alloc& a, size_type c, const Alloc& alloc);
    template <class Alloc>
      dynarray(allocator_arg_t, const Alloc& a, size_type c, const T& v, const Alloc& alloc);
    template <class Alloc>
      dynarray(allocator_arg_t, const Alloc& a, const dynarray& d, const Alloc& alloc);
    template <class Alloc>
      dynarray(allocator_arg_t, const Alloc& a, initializer_list<T>, const Alloc& alloc);
    

2256(i). On vector iterator invalidation

Section: 24.3.11.5 [vector.modifiers] Status: NAD Submitter: Howard Hinnant Opened: 2013-04-29 Last modified: 2020-05-08

Priority: 3

View all other issues in [vector.modifiers].

View all issues with NAD status.

Discussion:

24.3.11.5 [vector.modifiers]/p3 says:

iterator erase(const_iterator position);
iterator erase(const_iterator first, const_iterator last);

Effects: Invalidates iterators and references at or after the point of the erase.

Consider this example:

#include <vector>
#include <cassert>

int main()
{
  typedef std::vector<int> C;
  C c = {1, 2, 3, 4};
  C::iterator i = c.begin() + 1;
  C::iterator j = c.end() - 1;
  assert(*i == 2);
  assert(*j == 4);
  c.erase(c.begin());
  assert(*i == 3); // Why is this not perfectly fine?!
}

Why has the iterator i been invalidated? It still refers to a perfectly reasonable, fully constructed object. If vector::iterator were to be implemented as a pointer (which is legal), it is not possible for that last line to do anything but run fine.

The iterator j on the other hand now points at end, and any iterators that may now point beyond end(), into uninitialized memory, are clearly invalid.

But why do we say that an iterator that must point to a valid object is invalid? This looks to me like we simply got sloppy in our specification.

[2016-05 Issues Telecon]

This is related to 2698

[2017-03-04, Kona]

NAD. "Works as designed" Also, the example does not work in today's world of launder.

Proposed resolution:


2264(i). [arrays.ts] std::dynarray defines its initializer-list constructor in terms of a non-existent constructor

Section: 99 [arrays.ts::dynarray], 99 [arrays.ts::container.requirements] Status: NAD Arrays Submitter: Povilas Kanapickas Opened: 2013-05-22 Last modified: 2016-03-08

Priority: 1

View all issues with NAD Arrays status.

Discussion:

Addresses: arrays.ts

std::dynarray member listing at 99 [arrays.ts::dynarray.overview] includes this constructor:

dynarray(initializer_list<T>);

Also, 99 [arrays.ts::dynarray.overview] p. 2 says:

Unless otherwise specified, all dynarray operations have the same requirements and semantics as specified in 23.2.

The constructor in question isn't mentioned in 99 [arrays.ts::dynarray.cons] or anywhere else. This means requirements from 99 [arrays.ts::container.requirements] apply. However, Table 100 in 24.2.4 [sequence.reqmts] says:

X(il)               Equivalent to X(il.begin(), il.end())

std::dynarray does not provide this constructor.

The proposed resolution below adds the missing constructor and a complementary constructor with an allocator parameter. The new constructors, differently from the rest of containers, accept iterators that have forward iterator category. This is needed because the size requirements must be known in order to allocate appropriately-sized storage.

An alternative resolution could be to properly specify the initializer-list constructor.

[2013-09 Chicago:]

Move to Deferred. This feature will ship after C++14 and should be revisited then.

[2014-06-06 pre-Rapperswill]

This issue has been reopened as arrays-ts.

Proposed resolution:

  1. Add the following to the std::dynarray synopsis at 99 [arrays.ts::dynarray.overview]:

    namespace std {
      template <class T>
      class dynarray {
        […]
        // 23.3.4.2 construct/copy/destroy:
        […]
        template <class ForwardIterator>
        dynarray(ForwardIterator first, ForwardIterator last);
        template <class ForwardIterator, class Alloc>
        dynarray(ForwardIterator first, ForwardIterator last, const Alloc& alloc);
        […]
      };
    }
    
  2. Add the following to 99 [arrays.ts::dynarray.cons] after p. 8:

    template <class ForwardIterator>
    dynarray(ForwardIterator first, ForwardIterator last);
    

    -?- Requires: T shall meet the CopyConstructible requirements.

    -?- Effects: Allocates storage for distance(first, last) elements. The distance(first, last) elements of the dynarray are direct-initialized (9.4 [dcl.init]) with the corresponding elements from the range [first,last). May or may not invoke the global operator new.

    -?- Complexity: distance(first, last).

    -?- Throws: std::bad_array_length when the size requested is larger than implementable, std::bad_alloc when there is insufficient memory.

  3. Add the following to the list of constructors at 99 [arrays.ts::dynarray.cons] before p. 9:

    template <class Alloc>
    dynarray(size_type c, const Alloc& alloc);
    template <class Alloc>
    dynarray(size_type c, const T& v, const Alloc& alloc);
    template <class Alloc>
    dynarray(const dynarray& d, const Alloc& alloc);
    template <class Alloc>
    dynarray(initializer_list<T>, const Alloc& alloc);
    template <class ForwardIterator, class Alloc>
    dynarray(ForwardIterator first, ForwardIterator last, const Alloc& alloc);
    

2270(i). Inconsistent to_string overloads

Section: 23.4.5 [string.conversions] Status: NAD Submitter: Raf Schietekat Opened: 2013-07-02 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [string.conversions].

View all issues with NAD status.

Discussion:

For internal consistency, to_string() should either list all relevant types (including bool, char, etc.), or only those that are the destination types of integral or floating-point promotion (float not being among them).

A defensible reason for having (or rather keeping) the float overloads anyway could be to exactly mirror the adjacent sets of stoX() function overloads (even without round-trip fidelity for floating-point numbers).

Unfortunately, that reveals a bigger issue than redundant overloads: the glaring and indefensible omission of an overloaded function stoui(). Adding that is not as trivial as removing redundant overloads, of course, because it requires everybody to take action. Still, it is the preferable remedy for the present situation.

As far as I can tell from easily accessible information, C++ has already created the precedent with stoi(), which is not the equivalent of a pair of functions strtoi()/wcstoi() in C, but it would be if such functions existed. The function atoi() may look similar, but it does not qualify because it is as different from a hypothetical strtoi() as atol() currently is from strtol(), with the latter two both Standard C. It is only logical to act on this one-sided precedent by completing the set. Whether or not Standard C leads the way (or follows suit) is immaterial, but an invitation could be extended.

[2013-09 Chicago]

These overloads were very carefully and experimentally determined to be the minimal set, when all (known) promotion and conversion scenarios were considered. Removing superfluous-looking overloads is likely to result in ambiguities.

Proposed resolution:

This wording is relative to N3691.

  1. Modify 23.4 [string.classes], header <string> synopsis, as indicated:

    #include <initializer_list>
    
    namespace std {
      […]
      string to_string(int val);
      string to_string(unsigned val);
      string to_string(long val);
      string to_string(unsigned long val);
      string to_string(long long val);
      string to_string(unsigned long long val);
      string to_string(float val);
      string to_string(double val);
      string to_string(long double val);
      […]
    
      […]
      wstring to_wstring(int val);
      wstring to_wstring(unsigned val);
      wstring to_wstring(long val);
      wstring to_wstring(unsigned long val);
      wstring to_wstring(long long val);
      wstring to_wstring(unsigned long long val);
      wstring to_wstring(float val);
      wstring to_wstring(double val);
      wstring to_wstring(long double val);
      […]
    }
    
  2. Modify 23.4.5 [string.conversions] p7+14 as indicated:

    string to_string(int val);
    string to_string(unsigned val);
    string to_string(long val);
    string to_string(unsigned long val);
    string to_string(long long val);
    string to_string(unsigned long long val);
    string to_string(float val);
    string to_string(double val);
    string to_string(long double val);
    

    -7- Returns: Each function returns a string object holding the character representation of the value of its argument that would be generated by calling sprintf(buf, fmt, val) with a format specifier of "%d", "%u", "%ld", "%lu", "%lld", "%llu", "%f", "%f", or "%Lf", respectively, where buf designates an internal character buffer of sufficient size.

    […]

    wstring to_wstring(int val);
    wstring to_wstring(unsigned val);
    wstring to_wstring(long val);
    wstring to_wstring(unsigned long val);
    wstring to_wstring(long long val);
    wstring to_wstring(unsigned long long val);
    wstring to_wstring(float val);
    wstring to_wstring(double val);
    wstring to_wstring(long double val);
    

    -14- Returns: Each function returns a wstring object holding the character representation of the value of its argument that would be generated by calling swprintf(buf, buffsz, fmt, val) with a format specifier of L"%d", L"%u", L"%ld", L"%lu", L"%lld", L"%llu", L"%f", L"%f", or L"%Lf", respectively, where buf designates an internal character buffer of sufficient size buffsz.


2277(i). [arrays.ts] <dynarray> is missing in 24.7/1

Section: 99 [arrays.ts::iterator.range] Status: NAD Arrays Submitter: Cassio Neri Opened: 2013-07-31 Last modified: 2016-03-08

Priority: 3

View all issues with NAD Arrays status.

Discussion:

Addresses: arrays.ts

Section 99 [arrays.ts::iterator.range] p1 specifies header files that, in addition to <iterator>, make available the function templates in 24.7 (begin, end, etc.) but it fails to mention <dynarray>. This seems to be just an oversight.

[2013-09 Chicago:]

Move to Deferred. This feature will ship after C++14 and should be revisited then.

[2014-06-06 pre-Rapperswill]

This issue has been reopened as arrays-ts.

Proposed resolution:

This wording is relative to N3691.

  1. Modify 99 [arrays.ts::iterator.range] p1 as indicated:

    -1- In addition to being available via inclusion of the <iterator> header, the function templates in 24.7 are available when any of the following headers are included: <array>, <deque>, <dynarray>, <forward_list>, <list>, <map>, <regex>, <set>, <string>, <unordered_map>, <unordered_set>, and <vector>.


2279(i). Carefully state effects of list::splice function

Section: 24.3.10.5 [list.ops], 24.3.9.6 [forward.list.ops] Status: NAD Submitter: Arseny Klimovsky Opened: 2013-08-15 Last modified: 2023-02-07

Priority: Not Prioritized

View all other issues in [list.ops].

View all issues with NAD status.

Discussion:

I think that the effects of list::splice function should be stated more carefully.

Function transferring a single element is described now in the following way (24.3.10.5 [list.ops] p7):

void splice(const_iterator position, list& x, const_iterator i);
void splice(const_iterator position, list&& x, const_iterator i);

Effects: Inserts an element pointed to by i from list x before position and removes the element from x. The result is unchanged if position == i or position == ++i. Pointers and references to *i continue to refer to this same element but as a member of *this. Iterators to *i (including i itself) continue to refer to the same element, but now behave as iterators into *this, not into x.

But it is incorrect to talk about operator== for iterators that are not from the same container (after acceptance of N3066, 25.3.5.5 [forward.iterators] p2). So, the text operates with an undefined behaviour.

One is formally allowed to have list implementation where two iterators from different lists return true to operator==. For example, this can only happen to non-dereferenceable iterators, and position and ++i can be non-dereferenceable. So, literally according to the standard, it is not allowed in this implementation to transfer such elements with splice function.

[2013-09 Chicago (late night issues)]

Moved to NAD.

The condition under which the list is unchanged is not program code, so there is no undefined behavior to protect against. Rather, the precondition that the evaluation can be performed is implicit if determining when the condition applies.

Proposed resolution:

This wording is relative to N3691.

  1. Modify [forwardlist.ops] p6 as indicated:

    void splice_after(const_iterator position, forward_list& x, const_iterator i);
    void splice_after(const_iterator position, forward_list&& x, const_iterator i);
    

    […]

    -6- Effects: Inserts the element following i into *this, following position, and removes it from x. The result is unchanged if &x == this and the following condition is satisfied: position == i or position == ++i. Pointers and references to *++i continue to refer to the same element but as a member of *this. Iterators to *++i continue to refer to the same element, but now behave as iterators into *this, not into x.

  2. Modify 24.3.10.5 [list.ops] p7 as indicated:

    void splice(const_iterator position, list& x, const_iterator i);
    void splice(const_iterator position, list&& x, const_iterator i);
    

    -7- Effects: Inserts an element pointed to by i from list x before position and removes the element from x. The result is unchanged if &x == this and the following condition is satisfied: position == i or position == ++i. Pointers and references to *i continue to refer to this same element but as a member of *this. Iterators to *i (including i itself) continue to refer to the same element, but now behave as iterators into *this, not into x.


2281(i). C99 cross-reference typo in [using.linkage]

Section: 16.4.3.3 [using.linkage] Status: NAD Editorial Submitter: Chris Sharpe Opened: 2013-08-23 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [using.linkage].

View all issues with NAD Editorial status.

Discussion:

There is a footnote at section 16.4.3.3 [using.linkage]/2 that reads:

"The only reliable way to declare an object or function signature from the Standard C library is by including the header that declares it, notwithstanding the latitude granted in 7.1.7 of the C Standard."

There is no section 7.1.7 in the C99 Standard (or C11 final draft). I think the relevant section is:

"Provided that a library function can be declared without reference to any type defined in a header, it is also permissible to declare the function and use it without including its associated header."

at 7.1.4/2 from C99.

[2013-09 Chicago]

Moved to NAD Editorial.

Proposed resolution:

This wording is relative to N3691.

  1. Edit footnote 182, 16.4.3.3 [using.linkage] as indicated:

    The only reliable way to declare an object or function signature from the Standard C library is by including the header that declares it, notwithstanding the latitude granted in 7.1.77.1.4 of the C Standard.


2297(i). [CD] Missing type requirements for std::exchange

Section: 22.2.3 [utility.exchange] Status: NAD Submitter: Alisdair Meredith Opened: 2013-09-22 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [utility.exchange].

View all issues with NAD status.

Discussion:

Addresses GB 5

The wording describes example code including the call of a move constructor, but there is no requirement stated that T be move constructible.

We would like to add a new Para 1 before existing paragraph:

Requires: Type T shall be MoveConstructible (Table 20) and MoveAssignable (Table 22).

However the MoveAssignable concept currently does not cover cases where the source and destination types may differ.

[2013-09 Chicago]

The requirements are implicit according to 16.3.2.4 [structure.specifications]p4. There is no desire to redundantly repeat a set of requirements.

Proposed resolution:


2302(i). Passing null pointer to placement new

Section: 17.6.3.4 [new.delete.placement] Status: Pending NAD Submitter: Marc Glisse Opened: 2013-09-12 Last modified: 2017-09-07

Priority: 2

View all other issues in [new.delete.placement].

View all issues with Pending NAD status.

Discussion:

Based on this discussion and as discussed in c++std-core-23998 and c++std-lib-34442, calling placement new currently forces the compiler to check if the pointer is null before initializing the object (a non-negligible cost). It seems many people were not aware of this and they consider it a user error to pass a null pointer to it.

Proposed resolution: for operator new and operator new[], add:

Requires: ptr shall not be a null pointer.

[2014-02-15 post-Issaquah session : move to Tentatively NAD]

AJM to supply the rationale...

Proposed resolution:

This wording is relative to N3691.

  1. Change 17.6.3.4 [new.delete.placement] as indicated:

    void* operator new(std::size_t size, void* ptr) noexcept;
    

    -?- Requires: ptr shall not be a null pointer.

    -2- Returns: ptr.

    -3- Remarks: Intentionally performs no other action.

    -4- [Example: This can be useful for constructing an object at a known address:

    void* place = operator new(sizeof(Something));
    Something* p = new (place) Something();
    

    end example]

    void* operator new[](std::size_t size, void* ptr) noexcept;
    

    -?- Requires: ptr shall not be a null pointer.

    -5- Returns: ptr.

    -6- Remarks: Intentionally performs no other action.


2305(i). [fund.ts] optional forwarding construction/assignment

Section: 5.3.1 [fund.ts::optional.object.ctor] Status: NAD Submitter: Cassio Neri Opened: 2013-09-23 Last modified: 2015-10-26

Priority: 4

View all issues with NAD status.

Discussion:

Addresses: fund.ts

Consider:

struct foo {
  foo(std::initializer_list<int>&);        // 1
  foo(const std::initializer_list<int>&);  // 2
  foo(std::initializer_list<int>&&);       // 3
  foo(const std::initializer_list<int>&&); // 4
};

std::initializer_list<int> il{0, 1, 2};

foo foo_0{1, 2, 3};                                 // calls 3
foo foo_1{il};                                      // calls 1
foo foo_2((const std::initializer_list<int>&) il);  // calls 2
foo foo_3{(std::initializer_list<int>&&) il};       // calls 3
foo foo_4((const std::initializer_list<int>&&) il); // calls 4

Although the constructors of foo are unusual (initializer_lists are normally passed by value) users of optional could naturally expect perfect forwarding of initializer_lists. However, all lines below end up calling 1.

optional<foo> opt0{in_place, {1, 2, 3}};
optional<foo> opt1{in_place, il};                    
optional<foo> opt3{in_place, (const std::initializer_list<int>&) il};
optional<foo> opt2{in_place, (std::initializer_list<int>&&) il};
optional<foo> opt4{in_place, (const std::initializer_list<int>&&) il};

opt0.emplace({1, 2, 3});
opt0.emplace(il);
opt0.emplace((const std::initializer_list<int>&) il);
opt0.emplace((std::initializer_list<int>&&) il);
opt0.emplace((const std::initializer_list<int>&&) il);

The constructor

template <class... Args> constexpr explicit optional(in_place_t, Args&&... args);

can handle all constructor calls above, except the one taking {1, 2, 3}. Hence, a simple modification of

template <class U, class... Args>
constexpr explicit optional(in_place_t, initializer_list<U>&& il, Args&&... args);

allows perfect forwarding of std::initializer_list<U>s to be complete.

[2014-06-06 pre-Rapperswil]

This issue has been reopened as fundamentals-ts.

[2014-06-17, Rapperswil]

Move to NAD

Proposed resolution:

This wording is relative to N3691.

  1. Change 5.3.1 [fund.ts::optional.object.ctor] as indicated:

    template <class U, class... Args> 
    constexpr explicit optional(in_place_t, initializer_list<U>&& il, Args&&... args);
    

    -27- Requires: is_constructible<T, initializer_list<U>&&, Args&&...>::value is true.

    -28- Effects: Initializes the contained value as if constructing an object of type T with the arguments ilstd::move(il), std::forward<Args>(args)....

    […]

    -31- Remarks: The function shall not participate in overload resolution unless is_constructible<T, initializer_list<U>&, Args&&...>::value is true.

  2. Change 5.3.3 [fund.ts::optional.object.assign] as indicated:

    template <class U, class... Args>
    void optional<T>::emplace(initializer_list<U>&& il, Args&&... args);
    

    -27- Requires: is_constructible<T, initializer_list<U>&&, Args&&...>::value is true.

    -28- Effects: Calls *this = nullopt. Then initializes the contained value as if constructing an object of type T with the arguments ilstd::move(il), std::forward<Args>(args)....

    […]

    -32- Remarks: This function shall not participate in overload resolution unless is_constructible<T, initializer_list<U>&, Args&&...>::value is true.


2311(i). Allocator requirements should be further minimized

Section: 16.4.4.6 [allocator.requirements] Status: NAD Submitter: Stephan T. Lavavej Opened: 2013-09-21 Last modified: 2016-01-28

Priority: 2

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with NAD status.

Discussion:

C++11's minimized allocator requirements are great, but they're still requiring more things from users than absolutely necessary.

[2014-02-14 Issaquah: Close as NAD]

Different vendors rely on each of the different elements suggested to be removed.

While value_type my be deduced as suggested, far too much wording relies on it being available, and the standard churn is likely to be much harder than presented here.

Proposed resolution:

This wording is relative to N3691.

  1. Change in 16.4.4.6 [allocator.requirements], Table 28 — "Allocator requirements" as indicated:

    Table 28 — Allocator requirements (continued)
    Expression Return type Assertion/note pre-/post-condition Default
    X::value_type Identical to T   See Note B, below.
    a1 == a2 bool returns true only if storage
    allocated from each can be
    deallocated via the other.
    operator== shall be reflexive,
    symmetric, and transitive, and
    shall not exit via an exception.
    true
    a1 != a2 bool same as !(a1 == a2)  
    a == b bool same as a ==
    Y::rebind<T>::other(b)
     
    a != b bool same as !(a == b)  
    X a(b);   Shall not exit via an exception.
    post: Y(a) == b, a == X(b)
     
    X a(move(b));   Shall not exit via an exception.
    post: a equals the prior value of X(b).
     
  2. After 16.4.4.6 [allocator.requirements] p3, add a new paragraph:

    Note B: If Allocator is a class template instantiation of the form SomeAllocator<T, Args>, where Args is zero or more type arguments, and Allocator does not supply a nested type named value_type, the standard allocator_traits template uses T in place of Allocator::value_type by default. For allocator types that are not template instantiations of the above form, no default is provided.

  3. In the example provided in 16.4.4.6 [allocator.requirements]/5, delete as indicated:

    template <class Tp>
    struct SimpleAllocator {
      typedef Tp value_type;
      SimpleAllocator(ctor args);
      template <class T> SimpleAllocator(const SimpleAllocator<T>& other);
      Tp *allocate(std::size_t n);
      void deallocate(Tp *p, std::size_t n);
    };
    
    template <class T, class U>
    bool operator==(const SimpleAllocator<T>&, const SimpleAllocator<U>&);
    template <class T, class U>
    bool operator!=(const SimpleAllocator<T>&, const SimpleAllocator<U>&);
    
  4. Edit 20.2.9 [allocator.traits]p1, class template allocator_traits synopsis, as indicated:

    namespace std {
      template <class Alloc> struct allocator_traits {
        typedef Alloc allocator_type;
    
        typedef typename Alloc::value_typesee below value_type;
    
        […]
    
        static Alloc select_on_container_copy_construction(const Alloc& rhs);
    
        static bool equal(const Alloc& a1, const Alloc& a2) noexcept;
      };
    }
    
  5. At the beginning of 20.2.9.2 [allocator.traits.types], add a new paragraph:

    typedef see below value_type;
    

    Type: Alloc::value_type if such a type exists; otherwise, T if Alloc is a class template instantiation of the form Alloc<T, Args>, where Args is zero or more type arguments; otherwise, the program is ill-formed.

  6. At the end of 20.2.9.3 [allocator.traits.members], add a new paragraph:

    static bool equal(const Alloc& a1, const Alloc& a2) noexcept;
    

    -?- Returns: a1 == a2 if that expression is well-formed; otherwise, true.


2319(i). basic_string's move constructor should not be noexcept

Section: 23.4.3.3 [string.cons] Status: NAD Submitter: Stephan T. Lavavej Opened: 2013-09-21 Last modified: 2017-03-21

Priority: 1

View all other issues in [string.cons].

View all issues with NAD status.

Discussion:

In debugging implementations, containers (including basic_string) may need to own dynamically allocated helper objects at all times, including in their default-constructed and moved-from states. This means that their default constructors and move constructors may throw exceptions. Therefore, the Standard should not mark them as noexcept. (Other implementations will still be permitted to add noexcept.)

[2014-02, Issaquah : move to Ready]

The issue discussion was highly controversial: The arguments in favour was that implementations exist that always need to allocate memory even for the move operations (similar as for some other containers) and that this cleans up an inconsistency between std::string and other container types. Counter arguments were that potentially throwing move operations reduce much of the advantages of move-support, e.g. in vector<string>.

straw poll: accept wording in the issue
SF 4 WF 4 N 0 WA 1 SA 1

straw poll: 14 or 17?
C++14: 4 C++17: 4

Move to Ready for C++17, as too close to 14 DIS without strong consensus.

It was suggested to introduce a special library vocabulary that specifies a "normative encouragement to not throw exceptions" for functions like these.

NJ: I offer to write a proposal to add encouragement for not throwing ... "noexcept in italics means should not throw"

[2014/11 Urbana]

Resolved by paper N4258

[2015-10]

Richard Smith pointed out that N4258 explicitly decided not to change this call; leaving it as noexcept. In response, I am changing the resolution of this issue from 'Resolved' to 'NAD'

Proposed resolution:

This wording is relative to N3691.

  1. In 23.4.3 [basic.string]/5, class template basic_string synopsis, and 23.4.3.3 [string.cons]/2 change as indicated:

    basic_string(basic_string&& str) noexcept;
    
  2. Edit 23.4.3.3 [string.cons]/17 as indicated:

    basic_string(const basic_string& str, const Allocator& alloc);
    basic_string(basic_string&& str, const Allocator& alloc);
    

    […]

    -17- Throws: The second form throws nothing if alloc == str.get_allocator()..


2326(i). uniform_int_distribution<unsigned char> should be permitted

Section: 28.5.3.1 [rand.req.genl] Status: NAD Submitter: Stephan T. Lavavej Opened: 2013-09-21 Last modified: 2017-02-02

Priority: 2

View all issues with NAD status.

Discussion:

28.5.3.1 [rand.req.genl]/1 says: "Throughout this subclause 26.5, the effect of instantiating a template [...] that has a template type parameter named IntType is undefined unless the corresponding template argument is cv-unqualified and is one of short, int, long, long long, unsigned short, unsigned int, unsigned long, or unsigned long long." 28.5.9.2.1 [rand.dist.uni.int] specifies template<class IntType = int> class uniform_int_distribution, so this forbids uniform_int_distribution<char/signed char/unsigned char>.

I am not aware of anything in <random> that works with 16-bit integers but fails with 8-bit integers, so I suspect that IntType and UIntType could simply be extended to permit the char family. Alternatively, this change could be limited to uniform_int_distribution alone, where it is definitely safe. A <random> expert should decide which change is best.

[2015-04-04 Geoffrey provides wording]

I think it's time to call the question; it's just silly that we have a random number library with no natural way to generate random bytes. However, I don't think it's sufficient to fix only uniform_int_distribution, or even all of IntType. At a bare minimum we need to also fix independent_bits_engine (arguably the cleanest way of generating a random byte) and that's specified in terms of UIntType.

The wording provided below is equivalent to adding unsigned char to item "f" and adding signed char and unsigned char to item "e". That means it still excludes char, but I'm OK with that. If you want to generate a 1-byte number, you should probably pick a signedess, and if you want to generate a raw byte, the "true" raw byte type is unsigned char. This also excludes extended integral types and wide char types, which seem like nice-to-haves at best. I have no objection to supporting any of those types; I just picked this to simplify the wording and hopefully maximize consensus. Note that if we want to broaden IntType to permit any integral type, we'll need to decide if we want to exclude bool.

For reference, IntType is used as a parameter of the following templates:

uniform_int_distribution
binomial_distribution
geometric_distribution
negative_binomial_distribution
poisson_distribution
discrete_distribution

and UIntType is used as a parameter of the following templates:

linear_congruential_engine
mersenne_twister_engine
subtract_with_carry_engine
independent_bits_engine

[2015-9-11, Telecon]

Walter feels very strongly that this is not a defect, but a feature request.

In a previous telecon, Aaron offered to write a paper proposing this.

Jonathan offered to help.

Closing as NAD

Proposed resolution:

This wording is relative to N4296.

  1. Change in 28.5.3.1 [rand.req.genl] p1 as indicated:

    -1- Throughout this subclause 26.5, the effect of instantiating a template:

    1. […]

    2. that has a template type parameter named IntType is undefined unless the corresponding template argument is cv-unqualified and is a standard integer type (6.8.2 [basic.fundamental])one of short, int, long, long long, unsigned short, unsigned int, unsigned long, or unsigned long long.

    3. that has a template type parameter named UIntType is undefined unless the corresponding template argument is cv-unqualified and is a standard unsigned integer type (6.8.2 [basic.fundamental])one of unsigned short, unsigned int, unsigned long, or unsigned long long.


2327(i). Non-power-of-two URNGs should be forbidden

Section: 28.5.3.3 [rand.req.urng] Status: NAD Submitter: Stephan T. Lavavej Opened: 2013-09-21 Last modified: 2016-01-28

Priority: Not Prioritized

View all other issues in [rand.req.urng].

View all issues with NAD status.

Discussion:

28.5.3.3 [rand.req.urng] allows URNGs with non-power-of-two (NPOT) ranges, like [0, 1729]. This is unnecessarily permissive (I cannot imagine a realistic source of randomness that would generate such a range) and has real costs for implementers, as uniform_int_distribution must be prepared to accept such URNGs. The most efficient way to accumulate randomness is to concatenate random bits, so NPOT randomness is not just useless, it is actively harmful (to avoid bias, if a URNG generates a random number outside of a power-of-two range, the number must be discarded).

Forbidding NPOT URNGs wouldn't affect users, and would simplify Standard Library implementations. It would be nice to require min() to be 0, but this is not necessary; it is simple for implementations to say g() - G::min() and this will optimize away if min() is 0. (It is vaguely plausible for a URNG to have a nonzero minimum; I can imagine something that simply masks off low-order bits without shifting the rest downwards.) What is important is for the entire range to have a power-of-two width; [1729, 1984] is acceptable as its size is 256.

[2013-10-12: Howard presents a counterexample]

Consider:

#include <random>
#include <string>
#include <iostream>

template <class Int>
bool is_power_2m1(Int i)
{
  return (i & (i + 1)) == 0;
}

template <class URNG>
void test(const std::string& urng)
{
  using namespace std;
  typename URNG::result_type rng = URNG::max() - URNG::min();
  if (!is_power_2m1(rng))
  {
    cout << hex;
    cout << urng << " : min = " << URNG::min() << ", max = " << URNG::max()
         << ", max-min = " << rng << '\n';
  }
};

int main()
{
    using namespace std;
    test<minstd_rand0>("minstd_rand0");
    test<minstd_rand>("minstd_rand");
    test<mt19937>("mt19937");
    test<mt19937_64>("mt19937_64");
    test<ranlux24_base>("ranlux24_base");
    test<ranlux48_base>("ranlux48_base");
    test<ranlux24>("ranlux24");
    test<ranlux48>("ranlux48");
    test<knuth_b>("knuth_b");
}

Which for me outputs:

minstd_rand0 : min = 1, max = 7ffffffe, max-min = 7ffffffd
minstd_rand : min = 1, max = 7ffffffe, max-min = 7ffffffd
knuth_b : min = 1, max = 7ffffffe, max-min = 7ffffffd

We do not want to outlaw these three URNG's, and the proposed wording would do that.

[Issaquah 2014-02-10: Moved to NAD]

STL withdraws the issue, non-power-of-2 URNGs are used in the field, it is too late to consider removing them.

Proposed resolution:

This wording is relative to N3691.

  1. Add a new paragraph at the end of 28.5.3.3 [rand.req.urng] as indicated:

    -3- The following relation shall hold: G::min() < G::max().

    -?- G::max() - G::min() shall be 2n - 1 for some n > 0.


2335(i). array<array<int, 3>, 4> should be layout-compatible with int[4][3]

Section: 24.3.7 [array] Status: NAD Submitter: Jeffrey Yasskin Opened: 2013-10-04 Last modified: 2020-11-09

Priority: 3

View all other issues in [array].

View all issues with NAD status.

Discussion:

In order to replace some uses of C arrays with std::array, we need it to be possible to cast from a std::array<> to an equivalent C array. Core wording doesn't appear to be in quite the right state to allow casting, but if we specify that appropriate types are layout-compatible, we can at least write:

union {
  array<array<array<int, 2>, 3>, 4> arr;
  int carr[4][3][2];
};

to view memory as the other type: C++14 CD [class.mem]p18.

I believe it's sufficient to add "array<T, N> shall be layout-compatible (6.8 [basic.types]) with T[N]." to 24.3.7.1 [array.overview], but we might also need some extension to 11.4 [class.mem] to address the possibility of layout-compatibility between struct and array types.

I checked that libc++ on MacOS already implements this, although it would be good for someone else to double-check; I haven't checked any other standard libraries.

[2020-02-14, Prague]

LWG discussions and decision for NAD.

[2020-11-09 Status changed: Tentatively NAD → NAD.]

Rationale:

The desire to use std::array like this seems like an "XY problem". The goal should be "replace C arrays" not "replace C arrays with std::array", because std::array is not suitable here. There are superior solutions being proposed, and will be available in a future version of C++ (e.g. using mdspan as a multi-dimensional view on an array).

Proposed resolution:


2337(i). shared_ptr operator*() should not be noexcept

Section: 20.3.2.2.6 [util.smartptr.shared.obs] Status: NAD Submitter: Stephan T. Lavavej Opened: 2013-10-05 Last modified: 2017-07-17

Priority: 2

View all other issues in [util.smartptr.shared.obs].

View all issues with NAD status.

Discussion:

20.3.1.3.5 [unique.ptr.single.observers]/3: "pointer operator->() const noexcept; Requires: get() != nullptr."

20.3.2.2.6 [util.smartptr.shared.obs]/2: "T& operator*() const noexcept; Requires: get() != 0."

20.3.2.2.6 [util.smartptr.shared.obs]/5: "T* operator->() const noexcept; Requires: get() != 0."

Narrow-contract functions should not be noexcept.

[2014-02-15 Issaquah]

Issue is contentious, raise to P2.

[2015-02 Cologne]

AM: This ship has sailed. JM: What's the issue? AM: operator-> has narrow contract and should never have had noexcept. DK: Not quite. We explicitly called out that for shared_ptr this is fine. You said so in your "narrow contract" paper. GR: This would be a fairly major regression in the design of {unique,shared}_ptr over raw pointers; raw pointer dereferencing is noexcept. It's not a performance regression but a usability regression. AM: Do we expect users to query noexpect on dereference expressions? Room: Yes. VV: We don't just expect it, we have seen it. JM: Yes, users may be querying something like noexcept(x->y) and expect to be checking y, but silently end up checking x->.

Close as NAD, with explanation from GR.

Previous resolution [SUPERSEDED]:

This wording is relative to N3691.

  1. In 20.3.1.3 [unique.ptr.single]/1, class template unique_ptr synopsis for single objects, change as indicated:

    pointer operator->() const noexcept;
    
  2. In 20.3.1.3.5 [unique.ptr.single.observers] change as indicated:

    pointer operator->() const noexcept;
    

    -3- Requires: get() != nullptr.

    -4- Returns: get().

    -?- Throws: Nothing.

    -5- Note: use typically requires that T be a complete type.

  3. In 20.3.2.2 [util.smartptr.shared]/1, class template shared_ptr synopsis, change as indicated:

    T& operator*() const noexcept;
    T* operator->() const noexcept;
    
  4. In 20.3.2.2.6 [util.smartptr.shared.obs] change as indicated:

    T& operator*() const noexcept;
    

    -2- Requires: get() != 0.

    -3- Returns: *get().

    -?- Throws: Nothing.

    -4- Remarks: When T is void, it is unspecified whether this member function is declared. If it is declared, it is unspecified what its return type is, except that the declaration (although not necessarily the definition) of the function shall be well formed.

    T* operator->() const noexcept;
    

    -5- Requires: get() != 0.

    -6- Returns: get().

    -?- Throws: Nothing.

[2015-03-03, Geoffrey provides rationale]

Rationale:

It is by design that these members are noexcept, and changing that now would be a substantial regression in functionality. These classes were designed to substitute for plain pointers as transparently as possible, so since those operations are effectively noexcept on plain pointers, they should be noexcept on unique_ptr and shared_ptr as well. This matters in practice because we expect these members to be used fairly often inside the noexcept operator, and such code could be broken by this change. These design considerations override our general policy against noexcept for narrow-contract functions.

It is notable that N3279, which proposed this policy, did not propose striking noexcept from these operations. It's not clear if the omission of operator* and operator-> was an oversight, or an intentional reflection of the above considerations. N3279 was based on N3248 by the same authors, which states that:

"Most applications of noexcept for unique_ptr and shared_ptr are on functions with wide contracts. However, there are preconditions on the atomic access functions, so these should lose the specification."

Proposed resolution:


2345(i). integer_sequence should have a self-typedef ::type

Section: 21.2.2 [intseq.intseq] Status: NAD Submitter: Stephan T. Lavavej Opened: 2013-11-01 Last modified: 2016-01-28

Priority: 2

View all issues with NAD status.

Discussion:

21.3.4 [meta.help] says that integral_constant<T, v> provides ::value_type (for T) and ::type (for itself).

21.2.2 [intseq.intseq] says that integer_sequence<T, I...> provides ::value_type (for T), but nothing for itself.

Self-typedefs can be useful when users create chains of derived classes, then want to get the Standard base type. This is especially relevant to integer_sequence, as variadic templates encourage recursive inheritance.

[2014-02-13 Issaquah: Close as NAD]

AJM: My own implementation used a different alias for types representing parameter packs, and specifically did not define type. I tried it both ways, and found bugs more quickly when type was not defined.

Proposed resolution:

This wording is relative to N3797.

  1. Edit 21.2.2 [intseq.intseq] as indicated:

    namespace std {
      template<class T, T... I>
      struct integer_sequence {
        typedef T value_type;
        typedef integer_sequence<T, I...> type;
        static constexpr size_t size() noexcept { return sizeof...(I); }
      };
    }
    

2347(i). reverse_iterator::operator[] calls const version of current[]

Section: 25.5.1.6 [reverse.iter.elem] Status: NAD Submitter: Timo Bingmann Opened: 2013-11-11 Last modified: 2021-06-06

Priority: 2

View all other issues in [reverse.iter.elem].

View all issues with NAD status.

Discussion:

Currently reverse_iterator::operator[]() returns "current[-n-1]" and has an "unspecified" return type.

When Iterator is a mutable random access iterator, the expression "current[-n-1]" calls "Iterator::operator[] const", which returns a const reference. This const reference cannot be converted back to a mutable reference.

This issue is related to the "unspecified" return value of reverse_iterator::operator[], see defect 386.

The -1 is due to "current" pointing one item beyond the reverse_iterator's real current value.

The current libstdc++ implementation reads "*(current + n)" for reverse_iterator::operator[].

This copied current, advances (backwards) via operator+ and dereferences. It bypasses the issues due to reverse_iterator::operator[] being const by copying the iterator.

[2014-02-13 Issaquah : close as NAD]

Proposed resolution:

This wording is relative to N3797.

  1. Edit 25.5.1.2 [reverse.iterator], class template reverse_iterator synopsis, as indicated:

    namespace std {
      template <class Iterator>
      class reverse_iterator : public
      iterator<typename iterator_traits<Iterator>::iterator_category,
        typename iterator_traits<Iterator>::value_type,
        typename iterator_traits<Iterator>::difference_type,
        typename iterator_traits<Iterator>::pointer,
        typename iterator_traits<Iterator>::reference> {
      public:
        […]
        unspecifiedreference operator[](difference_type n) const;
        […]
      };
      […]
    }
    
  2. Edit [reverse.iter.opindex] as indicated:

    unspecifiedreference operator[](
      typename reverse_iterator<Iterator>::difference_type n) const;
    

    -1- Returns: current[-n-1]*(current + n).


2351(i). Does .seed() completely reset state of engine?

Section: 28.5.4 [rand.eng] Status: NAD Submitter: Thomas Plum Opened: 2013-12-02 Last modified: 2016-01-28

Priority: 2

View all other issues in [rand.eng].

View all issues with NAD status.

Discussion:

With regard to Random number engine class templates 28.5.4 [rand.eng], the Standard can be read in two different ways: when the member function

.seed(result_type s = default_seed)

is invoked, is all associated state (such as carry) reset to the same state that would have been created by the constructor

explicit engine-type(result_type s = default_seed)

or is the exact state unspecified?

Implementations differ.

[2014-02-13, Issaquah]

Walter Brown says that Table 117 makes this very clear, and that the answer is "Yes"

Suggested resolution: NAD

[2015-05-05 Lenexa: Move to NAD]

Proposed resolution:

Suggested resolution: NAD


2352(i). Is a default-constructed std::seed_seq intended to produce a predictable .generate()?

Section: 28.5.8.1 [rand.util.seedseq] Status: NAD Submitter: Thomas Plum Opened: 2013-12-02 Last modified: 2022-08-24

Priority: 2

View all other issues in [rand.util.seedseq].

View all issues with NAD status.

Discussion:

With respect to class seed_seq 28.5.8.1 [rand.util.seedseq], is a default-constructed std::seed_seq intended to produce a predictable .generate() sequence?

Implementations differ.

[2014-02-10]

Priority set to 2

[2022-08-24 Status changed: New → NAD.]

Discussed in LWG telecon. The default constructor effects are clear. Implementations agree.

Proposed resolution:


2355(i). "s" UDL suffix should be reserved for a compile-time string library type

Section: 23.4.7 [basic.string.literals] Status: NAD Submitter: Michael Price Opened: 2014-01-18 Last modified: 2016-01-28

Priority: 1

View all issues with NAD status.

Discussion:

The current draft uses the "s" UDL suffix as a UDL for basic_string<charT> (23.4.7 [basic.string.literals]). In light of EWG active issue 66 (concerning N3599), the "s" suffix (when applied to character string literals) should be reserved for a compile-time string library type.

[Issaquah 2014-10-12: Move to NAD]

We discussed leaving the s UDL suffix for string_view in Portland, and voted strongly in favor of using it for std::string. string is also an extremely widely used type, and the difference is observable in type deduction cases. In addition, a compile-time string is likely to cost significant compile time, which we don't want to make the default with s.

Mark 2355 as NAD?
SFFNASA
8 4020

Proposed resolution:


2372(i). Assignment from int to std::string

Section: 23.4.3 [basic.string] Status: NAD Submitter: Andrzej Krzemieński Opened: 2014-03-13 Last modified: 2018-06-23

Priority: 4

View other active issues in [basic.string].

View all other issues in [basic.string].

View all issues with NAD status.

Discussion:

The following code works in C++:

int i = 300;
std::string threeHundred;
threeHundred = i;

"Works" == "Compiles and doesn't have an undefined behavior". But it may not be obvious and in fact misleading what it does. This assignment converts an int to char and then uses string's assignment from char. While the assignment from char can be considered a feature, being able to assign from an int looks like a safety gap. Someone may believe C++ works like "dynamically typed" languages and expect a lexical conversion to take place.

Ideally the assignment from char could be deprecated and later removed, but as a less intrusive alternative one could consider adding a SFINAEd deleted function template:

template <typename IntT> // enable if is_integral<IntT>::value
basic_string& operator=(IntT) = delete;

[Lenexa 2015-06-06: Move to LEWG]

RS: std::string x('0' + n); broken by this.

MC: This is an extension, move to LEWG.

Move to LEWG, consensus.

Previous resolution [SUPERSEDED]:

This wording is relative to N3936.

  1. To 23.4.3 [basic.string], class template basic_string synopsis, add as indicated:

    basic_string& operator=(const basic_string& str);
    basic_string& operator=(basic_string&& str) noexcept;
    basic_string& operator=(const charT* s);
    basic_string& operator=(charT c);
    template <class IntT> basic_string& operator=(IntT i) = delete;
    basic_string& operator=(initializer_list<charT>);
    
  2. Add after 23.4.3.3 [string.cons] p26 as indicated:

    basic_string& operator=(charT c);
    

    -26- Returns: *this = basic_string(1,c).

    template <class IntT> basic_string& operator=(IntT i) = delete;
    

    -?- Remarks: This signature shall not participate in overload resolution unless is_integral<T>::value is true.

[LEWG: 2016-03, Jacksonville]

is_integral<T>::valueis_arithmetic<tmpl-arg>::value

This needs a paper; close the issue

We don't think the breakage is acceptable.

Guidance to author: Look for a way to encourage a warning; discomfort with calling that "deprecation".

Consider += and push_back.

Proposed resolution:

This should be addressed by a paper addressed to LEWG.


2373(i). Make new entities and names in namespace std conforming extensions

Section: 16.4.6 [conforming] Status: NAD Submitter: Chandler Carruth Opened: 2014-03-22 Last modified: 2015-05-22

Priority: 3

View all other issues in [conforming].

View all issues with NAD status.

Discussion:

Technically, right now, it is not a conforming extension to add a new function to namespace std. Doing so could cause unqualified lookup on the name of that function in the presence of a using directive to find a different function. This seems an unreasonable restriction on library vendors providing conforming extensions, as such a using directive seems inherently risky in unqualified name lookup.

16.4.6.5 [member.functions] implies that adding overloads to a method is a conforming extension, and within some limits the same is true for global functions due to 16.4.6.4 [global.functions].

It would likely be useful to specify that other new entities are valid conforming extensions, or preclude them where they pose serious compatibility problems.

[Lenexa 2015-05-06: Move to NAD]

JY: It's a design question, move to LEWG?

AM: NAD: extensions led to us being unable to use the names hash_map, leading to unordered_map etc. Will result in collisions between members.

MC: Agrees, implementations that extend std:: must use __ugly_names for this reason.

JY: I would not oppose NAD.

Move to NAD, consensus.

Proposed resolution:


2379(i). Obtaining native handle of the current thread

Section: 33.2.3 [thread.req.native], 33.4.5 [thread.thread.this] Status: NAD Submitter: Matt Austern Opened: 2014-03-31 Last modified: 2015-10-21

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Class thread contains an implementation-defined type thread::native_handle_type, and an implementation-defined function thread::native_handle() that returns a value of that type. The presence and semantics of those members is implementation-defined; the intention is that they can be used for interoperability with libraries that rely on operating system specific features. (Posix libraries that accept arguments of type pthread_t, for example.)

Unfortunately, there appears to be no native handle support for the equivalent of pthread_self(). We can use this_thread::get_id() to obtain the thread::id of the current thread, but there is no mechanism for converting a thread::id to a thread::native_handle.

Proposed wording:

In 33.4.5 [thread.thread.this] add:

thread::native_handle_type native_handle(); // See 30.2.3

to the this_thread namespace synopsis.

Rationale:

Informally, we could address this issue either by adding a new function in this_thread or by providing a mechanism for converting between thread::id and thread::native_handle. I propose the former because it seems more localized, and doesn't involve saying anything more about implementation defined native functionality than we currently do.

It's intentional that the proposed resolution adds a declaration of native_handle() without adding a paragraph explaining what it does. This is because everything about native_handle() is implementation-defined. The standard does the same thing in 33.4.3.6 [thread.thread.member].

[2015-02 Cologne]

Handed over to SG1.

[2015-05 Lenexa, SG1 response]

Strong "don't care" reaction from SG1, after pointing out that this only matters in non-portable code, which can call pthread_self() or the like anyway, but the change also doesn't add any non-trivial implementation requirements. This defused initial strong opinions on both sides. Since this is essentially a feature request, we did not have sufficient consensus to proceed at this point. There was a feeling we should reconsider after making more sense out of the much more general TLS issues we have been discussing.

Proposed resolution:

This wording is relative to N3936.

  1. Change 33.2.3 [thread.req.native] p1 as indicated:

    -1- Several classes described in this Clause have membersThis Clause includes several members named native_handle_type and native_handle. The presence of these members and their semantics is implementation-defined. […]

  2. In 33.4 [thread.threads], header <thread> synopsis, add:

    namespace std {
      […]
      namespace this_thread {
        thread::id get_id() noexcept;
        thread::native_handle_type native_handle();
        […]
      }
    }
    
  3. In 33.4.5 [thread.thread.this] add:

    namespace std {
      namespace this_thread {
        thread::id get_id() noexcept;
        thread::native_handle_type native_handle(); // See 33.2.3 [thread.req.native]
        […]
      }
    }
    

2382(i). Unclear order of container update versus object destruction on removing an object

Section: 16.4.6.9 [reentrancy] Status: Pending NAD Submitter: Peter Kasting Opened: 2014-05-06 Last modified: 2014-11-04

Priority: 2

View all other issues in [reentrancy].

View all issues with Pending NAD status.

Discussion:

The standard does not seem to discuss reentrant access to a container during removal of an element, leaving it unclear whether a removed object is destroyed before or after it is removed from the container. For example, the behavior of the following code seems to be unspecified:

#include <iostream>
#include <map>
#include <memory>

struct T;
typedef std::map<int, std::shared_ptr<T>> TMap;

struct T {
  T(TMap* t_map, int index) : t_map(t_map), index(index) {}
  ~T() {
    std::cout << "Object " << index << " is ";
    if (t_map->count(index))
      std::cout << "destroyed before being removed from the map" << std::endl;
    else
      std::cout << "removed from the map before being destroyed" << std::endl;
  }

  static void AddToMap(TMap* map, int index) {
    (*map)[index] = std::make_shared<T>(map, index);
  }

  TMap* t_map;
  int index;
};

int main()
{
  TMap t_map;
  T::AddToMap(&t_map, 0);
  T::AddToMap(&t_map, 1);
  t_map.erase(1);
  t_map.erase(0);
}

The output of this program in Visual Studio 2013 is:

Object 1 is removed from the map before being destroyed
Object 0 is destroyed before being removed from the map

The core issue here is whether an object removed from a container should be destroyed before or after it is removed from the container. The current standard seems to be silent on this issue. The above output demonstrates that the behavior is actually inconsistent. (It's difficult to fully describe Visual Studio's behavior; for example, changing main() in the above example to the following:)

int main()
{
  TMap t_map;
  T::AddToMap(&t_map, 0);
  T::AddToMap(&t_map, 1);
  T::AddToMap(&t_map, 2);
  T::AddToMap(&t_map, 3);
  t_map.erase(3);
  t_map.clear();
}

(...gives this output:)

Object 3 is removed from the map before being destroyed
Object 2 is destroyed before being removed from the map
Object 1 is destroyed before being removed from the map
Object 0 is removed from the map before being destroyed

In my opinion, the standard should explicitly describe when objects are destroyed as part of removal from a container. To me, it makes the most sense to say that objects should be removed from the container before they are destroyed.

[2014-05-07, Jeffrey Yasskin comments]

I think there are two main points here beyond this writeup:

  1. We can't make recursive use of a standard library container valid in all cases.

  2. If recursion through especially erase() is undefined behavior, that's pretty scary for existing large applications with code in destructors. Of course, "scary" doesn't mean we have to define the behavior.

I'll add a third: The language in 16.4.6.9 [reentrancy] nearly makes this undefined behavior already. I think any fix is probably going to live there, and extend the current "implementation-defined" on recursive reentrancy for individual functions to recursive reentrancy on class instances. I'm not sure exactly how to word that.

[2014-06 Rapperswil]

STL: We need more wording about how container methods can be reentrency.

Jeffrey: The title for this issue is confusing, what we really want is "reentrancy for objects".

Alisdair: Should we then close 2382 as NAD with a link to the new issue?

Proposed resolution:


2386(i). function::operator= handles allocators incorrectly

Section: 22.10.17.3.2 [func.wrap.func.con] Status: NAD Submitter: Pablo Halpern Opened: 2014-05-23 Last modified: 2015-05-05

Priority: 1

View all other issues in [func.wrap.func.con].

View all issues with NAD status.

Discussion:

The Effects clauses for the assignment operator for class template function are written as code that constructs a temporary function and then swaps it with *this. The intention appears to be that assignment should have the strong exception guarantee, i.e., *this is not modified if an exception is thrown. However, the current description is incorrect when *this was constructed using an allocator.

Part of the problem is the under-specification of swap, which does not state the allocator requirements or allocator postconditions. If swap behaves like the rest of the standard library, swapping function objects constructed with different allocators would be undefined behavior. Alternatively swap could exchange the allocators, though I would argue against this specification.

For either specification of swap, the current Effects clauses for operator= are incorrect. If swap does not exchange the allocators, then operator= would have undefined behavior, which is clearly not desired. If swap does exchange the allocators, then operator= would always leave the left-hand side (lhs) of the assignment with a default allocator. The latter would be surprising behavior, as the allocator instance is normally unchanged for the lifetime of an object (for good reason), and is certainly not reset to default arbitrarily.

The desired behavior is that assignment would leave the allocator of the lhs unchanged. The way to achieve this behavior is to construct the temporary function using the original allocator. Unfortunately, we cannot describe the desired behavior in pure code, because there is no way to name the type-erased value of the allocator. (N3916 would improve this situation for the Library Fundamentals TS, but even with those changes, there is no way to recover the original type of the allocator.) The PR below, therefore, uses pseudo-code, inventing a fictitious ALLOCATOR_OF(f) expression that evaluates to the actual allocator type, even if that allocator was type erased. I have implemented this PR successfully.

Previous resolution [SUPERSEDED]:

This wording is relative to N3936.

  1. Change 22.10.17.3.2 [func.wrap.func.con] as indicated:

    In the following descriptions, ALLOCATOR_OF(f) is a copy of the allocator specified in the construction of function f, or allocator<char>() if no allocator was specified.

    function& operator=(const function& f);
    

    -12- Effects: function(allocator_arg, ALLOCATOR_OF(*this), f).swap(*this);

    -13- Returns: *this

    function& operator=(function&& f);
    

    -14- Effects: Replaces the target of *this with the target of f.function(allocator_arg, ALLOCATOR_OF(*this), std::move(f)).swap(*this);

    -15- Returns: *this

    function& operator=(nullptr_t);
    

    -16- Effects: If *this != nullptr, destroys the target of this.

    -17- Postconditions: !(*this). The allocator is unchanged.

    -18- Returns: *this

    -?- Throws: Nothing.

    template<class F> function& operator=(F&& f);
    

    -19- Effects: function(allocator_arg, ALLOCATOR_OF(*this), std::forward<F>(f)).swap(*this);

    -20- Returns: *this

    -21- Remarks: This assignment operator shall not participate in overload resolution unless declval<typename decay<F>::type&>() is Callable (20.9.11.2) for argument types ArgTypes... and return type R.

    template<class F> function& operator=(reference_wrapper<F> f);
    

    -22- Effects: function(allocator_arg, ALLOCATOR_OF(*this), f).swap(*this);

    -23- Returns: *this

[2015-05, Lenexa]

STL: think this is NAD, don't think this is implementable or even should be.
STL: think this issue should be dealt with the same as 2370, don't think this should be done ever.
STL: NAD because there is nothing broken here.
STL: already fixed operator= noexcept so Throws nothing is not needed
STL: nothing to salvage here
MC: consensus for NAD

Proposed resolution:

There was consensus by the committee that the issue does not constitute as defect.


2388(i). Handling self-assignment in the proposed library function std::exchange

Section: 22.2.3 [utility.exchange] Status: NAD Submitter: Nick Calus Opened: 2014-05-09 Last modified: 2015-05-05

Priority: 2

View all other issues in [utility.exchange].

View all issues with NAD status.

Discussion:

In paper N3668, the addition of a template function std::exchange had been proposed. In the rationale provided by the paper, we find the following:

I chose the name for symmetry with atomic_exchange, since they behave the same except for this function not being atomic.

and:

Atomic objects provide an atomic_exchange function ([atomics.types.operations.req]p18) that assigns a new value to the object and returns the old value. This operation is also useful on non-atomic objects, and this paper proposes adding it to the library.

But the specified semantics of std::exchange is defined as follows:

template <class T, class U=T> T exchange(T& obj, U&& new_val);

Effects: Equivalent to:

T old_val = std::move(obj);
obj = std::forward<U>(new_val);
return old_val;

When looking at the post-condition of the std::exchange function, one would expect the return value to be the old value of obj and also that obj now contains the value of new_value. This post-condition is violated when obj is a reference to the same object as new_value and type T has move semantics.

Given it's specification, it is clear that std::exchange is meant to be used with types that have move semantics. Therefore, the post-condition is violated for self-assignments.

Suppose the following situation:

You have a vector of objects. The objects implement move semantics and are emptied when moved from. You provide a function that allows you to replace an object at a specific index by a new object (provided by reference as an argument to your function). When replacing an object, your function calls a member-function do_something_fancy on the old object.

void your_function(int i, X& new_val) {
  std::exchange(vec[i], new_val).do_something_fancy();
}

Your function gets called with a given index and the corresponding element of said vector. (by coincidence or by purpose, it doesn't really matter)

your_function(5, vec[5]);

This will cause the object at vec[5] to be in an empty state. If this object would not implement move semantics, assignment performance is potentially worse, but at least it is not in an empty (to my business logic, invalid) state.

So to me, the current reference implementation of std::exchange does not have the behavior it is expected to have.

[2014-12-18 Telecon]

MC: does this resolution solve the problem?

JW: and is the cost of the extra construction and move acceptable?

AM: not all moves are cheap

VV: seems like a design change

JW: maybe this should be rolled into my unwritten paper on self-swap, so we deal with them consistently

VV: we should update the issue saying something like that and maybe NAD Future

MC: instead, add Requires clause saying the arguments are not the same.

JW: interesting, that can even be checked in a debug mode assertion

MC: ACTION: send alternative P/R that we can consider

Previous resolution [SUPERSEDED]:

This wording is relative to N3936.

  1. Change 22.2.3 [utility.exchange] as indicated:

    template <class T, class U=T> T exchange(T& obj, U&& new_val);
    

    -1- Effects: Equivalent to:

    T tmp = std::forward<U>(new_val);
    T old_val = std::move(obj);
    obj = std::forward<U>(new_val)std::move(tmp);
    return old_val;
    

Previous resolution from Marshall [SUPERSEDED]:

This wording is relative to N4296.

  1. Change 22.2.3 [utility.exchange] as indicated:

    template <class T, class U=T> T exchange(T& obj, U&& new_val);
    

    -?- Requires: obj and new_val shall not refer to the same object.

    -1- Effects: Equivalent to:

    T old_val = std::move(obj);
    obj = std::forward<U>(new_val);
    return old_val;
    

[2015-03-30, Marshall provides alternative wording]

[2015-05, Lenexa]

MC: self exchange does not work
MC: PR is just to add requires
STL: what if the new thing is a subobject, isn't that just as bad, any aliasing
STL: don't know that we need to do anything here if we aren't changing the implementation
NM: should remove the requires
MC: so NAD
STL: could add note
NM: remove requires
DK: requires isn't already there
RL: no note?
STL: no note, NAD, burden for next person that asks about aliasing
DK: what do we do for swap?
STL: self swap has always been noop, exchange could do something bad because it clears out
MC: alright, NAD it is

Proposed resolution:

The current specification is clear about the implications described by the issue example and is as intended.

2402(i). basic_string(const basic_string& str, size_type pos, size_type n = npos) shouldn't use Allocator()

Section: 23.4.3.3 [string.cons] Status: NAD Submitter: Stephan T. Lavavej Opened: 2014-06-14 Last modified: 2016-03-07

Priority: 3

View all other issues in [string.cons].

View all issues with NAD status.

Discussion:

23.4.3.3 [string.cons] p3 specifies:

basic_string(const basic_string& str, size_type pos, size_type n = npos, const Allocator& a = Allocator());

But this implies that basic_string(str, pos) and basic_string(str, pos, n) use Allocator() instead of getting an allocator from str.

23.4.3.2 [string.require] p3 says "The Allocator object used shall be obtained as described in 23.2.1." 24.2.2.1 [container.requirements.general] p8 says "Copy constructors for these container types obtain an allocator by calling allocator_traits<allocator_type>::select_on_container_copy_construction on the allocator belonging to the container being copied.", but this isn't exactly a copy constructor. Then it talks about move constructors (which this definitely isn't), and finally says that "All other constructors for these container types take a const allocator_type& argument. […] A copy of this allocator is used for any memory allocation performed".

[2015-05-06 Lenexa: move to Open]

STL: there an allocator right there in str, why default-construct one

STL: my fix, which may not be right, splits out functions with and without allocators

JW: there are other ways to propagate the allocator from str to the new object

PJP: hard to get motivated about this one

JW: I think this is not a copy operation, this is init'ing a string from a range of characters which happens to originate in a string. It makes it inconsistent with the similar ctor taking a const char pointer, and if we had a std::string_view we wouldn't even have this ctor, and it wouldn't be possible to propagate the allocator.

STL: but people with stateful allocators want it to propagate

JW: I think the people using stateful allocators will alter the default behaviour of select_on_container_copy_construction so that it doesn't propagate, but will return a default-constructed one (to ensure a stateful allocator referring to a stack buffer doesn't leak to a region where the stack buffer has gone). So for those people, your proposed change does nothing, it changes one default-constructed allocator to a call to select_on_container_copy_construction which returns a default-constructed allocator. For other people who have different stateful allocators they can still provide the right allocator (whatever that may be) by passing it in.

STL: OK, that's convincing.

PJP: I agree with Jonathan

JW: would like to run both our arguments by Pablo in case I'm totally misrepresenting the expected users of allocator-traits stuff

[2015-10, Kona Saturday afternoon]

Everyone thinks this seems right.

STL: It'd be really weird if you copy from a string with a stateful allocator and you'd just default-construct a new allocator.

EF: We definitely need this for polymorphic allocators.

TK: Whether you think this is kind of copy-constructor or a constructor from raw string data, the new form in the PR is more flexible. You can still get the default-constructed allocator if you want, but conversely, getting the select_on_container_copy is really hard to type in the old form.

JW has objections (written in the issue) but won't block "Review" status.

Move to Review, hopefully to be made ready at a telecon.

[2015-11-22, Pablo comments]

I like the direction of the PR, but it is incomplete. Consider the following (assuming the PR):

typedef basic_string<char, char_traits<char>, A<char>> stringA;
vector<stringA, scoped_allocator_adaptor<A<stringA>>> vs;
stringA s;

vs.emplace_back(s, 2);  // Ill-formed

The problem is that uses-allocator construction requires that we be able to pass an allocator to the constructor stringA(s, 2, allocator), but no such constructor exists. I think this defect already exists, but we should fix it a the same time that we fix 2402. So, I would say we need a third constructor:

basic_string(const basic_string& str, size_type pos, const Allocator& a);

[2016-01-05, Pablo comments]

I've reconsidered and I think that the issue as stated, is NAD. I do not like the PR at all. In fact, I think it reverses a previous fix and it could break existing code.

There are two patterns that are at work here:

  1. Every constructor needs a version with and without an allocator argument (possibly through the use of default arguments).

  2. Every constructor except the copy constructor for which an allocator is not provided uses a default-constructed allocator.

The constructors in question are not copy constructors. I do not think it is compelling that the allocator should come from its argument any more than it should come from any other object that happens to supply characters for a string constructor.

[2016-03 Jacksonville]

Closed as NAD, noting that 2583 is a related issue.

Previous resolution [SUPERSEDED]:

This wording is relative to N3936.

  1. Change 23.4.3 [basic.string] p5, class template basic_string synopsis, as indicated:

    […]
    // 21.4.2, construct/copy/destroy:
    […]
    basic_string(basic_string&& str) noexcept;
    basic_string(const basic_string& str, size_type pos, size_type n = npos);
    basic_string(const basic_string& str, size_type pos, size_type n = npos,
                const Allocator& a = Allocator());
    […]
    
  2. Change 23.4.3.3 [string.cons] around p3 as indicated:

    basic_string(const basic_string& str, 
                 size_type pos, size_type n = npos);
    basic_string(const basic_string& str, 
                 size_type pos, size_type n = npos,
                 const Allocator& a = Allocator());
    

    […]

    -5- Effects: Constructs an object of class basic_string and determines the effective length rlen of the initial string value as the smaller of n and str.size() - pos, as indicated in Table 65. The first constructor obtains an allocator by calling allocator_traits<allocator_type>::select_on_container_copy_construction on the allocator belonging to str.

    Table 65 — basic_string(const basic_string&, size_type, size_type) and basic_string(const basic_string&, size_type, size_type, const Allocator&) effects

Proposed resolution:

The existing wording is intended.


2405(i). rotate()'s return value is incorrect when middle == first

Section: 27.7.11 [alg.rotate] Status: NAD Submitter: Stephan T. Lavavej Opened: 2014-06-14 Last modified: 2014-06-17

Priority: Not Prioritized

View all other issues in [alg.rotate].

View all issues with NAD status.

Discussion:

When LWG 488 was resolved, the intention was to return "where the subrange [first, middle) starts after the rotate is performed". However, this wasn't achieved in one case.

When middle == last, rotate() does nothing and returns first. This is good.

When middle == first, rotate() does nothing and returns last. This is bad. In addition to being inconsistent with the other do-nothing case, it prevents rotate() from providing the useful guarantee that LWG 488 wanted: when [first, last) is non-empty, rotate()'s return value should always be dereferenceable to get the originally-first element.

Howard Hinnant:

As the author of LWG 488 I can assure everyone that the edge cases the proposed resolution specifies were not specified by accident. rotate(i, i, j) should return j and rotate(i, j, j) should return i. Doing otherwise will break working code. These return values were motivated by the uses of rotate in the implementation of algorithms such as stable_partition and inplace_merge. The results of these edge cases were not chosen lightly.

Also a good read:

notes-on-programming-2006-10-13

Summary: NAD.

[2014-06-16 Rapperswill]

Closed as NAD.

Proposed resolution:

This wording is relative to N3936.

  1. Change 27.7.11 [alg.rotate] p2 as indicated:

    -2- Returns: If middle == first, returns first. Otherwise, returns first + (last - middle).


2417(i). [fund.ts.v2] std::experimental::optional::operator< and LessThanComparable requirement

Section: 5.7 [fund.ts.v2::optional.relops], 5.9 [fund.ts.v2::optional.comp_with_t] Status: NAD Submitter: Daniel Krügler Opened: 2014-06-20 Last modified: 2021-06-06

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: fund.ts.v2

Currently, std::experimental::optional::operator== imposes the EqualityComparable requirement which provides two guarantees: It ensures that operator!= can rely on the equivalence-relation property and more importantly, that the BooleanTestable requirements suggested by issue 2114 are automatically implied.

std::experimental::optional::operator< doesn't provide a LessThanComparable requirement, but there was quite an historic set of changes involved with that family of types: As of N3527 this operator was defined in terms of operator< of the contained type T and imposed the LessThanComparable requirement. In the final acceptance step of optional by the committee, the definition was expressed in terms of std::less and the LessThanComparable requirement had been removed.

The inconsistency between operator== and operator< should be removed. One possible course of action would be to add the LessThanComparable to std::experimental::optional::operator<. The EqualityComparable requirement of operator== could also be removed, but in that case both operators would at least need to require the BooleanTestable requirements (see 2114) for the result type of T's operator== and operator<.

Arguably, corresponding operators for pair and tuple do not impose LessThanComparable (nor EqualityComparable), albeit the definition of the "derived" relation functions depend on properties ensured by LessThanComparable. According to the SGI definition, the intention was to imposed both EqualityComparable and LessThanComparable. If this is not intended, the standard should clarify this position.

[2015-02 Cologne]

VV, DK, JY discuss why and when LessThanComparable was removed. AM: Move to LEWG. Please tell LWG when you look at it.

[2016-11-08, Issaquah]

Not adopted during NB comment resolution

[LEWG Kona 2017]

Recommend NAD: We've done a lot of work getting the C++17 semantics we want in this area; we're not going to change them 3 days from DIS or change the TSv2 behavior to be different from '17.

Proposed resolution:


2429(i). std::basic_ostringstream is missing an allocator-extended constructor

Section: 31.8.4 [ostringstream] Status: NAD Submitter: Markus Kemp Opened: 2014-09-03 Last modified: 2017-03-21

Priority: Not Prioritized

View all other issues in [ostringstream].

View all issues with NAD status.

Discussion:

I initially brought this issue up on Stack Overflow, where I was then told to make a topic about this problem on std-discussion, where I was then in turn asked to report the issue.

The problem: The std::basic_ostringstream class template can be instantiated with an allocator type, but none of the constructors provided accept an allocator argument, which means it's impossible to use std::basic_ostringstream with stateful allocators. The C++ Standard Library Defect Report List seems to already mention a similar issue (2210).

[2014-11 Urbana]

Closed as NAD

This is not a rejection of the suggestion, but an observation that simply adding an allocator-aware constructor is only part of the problem. stringstream returns the string assembled in its buffer by value, as the result of a call to str, and typically this will not use the same allocator as would be supplied at construction.

The appropriate way to make progress on this issue, if motivated, is to submit a paper to LEWG addressing the larger design concerns in addition to 'just' adding an (optional) allocator to the constructors.

Proposed resolution:


2430(i). Heterogeneous container lookup should be enabled using meta-function instead of nested type

Section: 24.2.7 [associative.reqmts] Status: NAD Submitter: Tomasz Kamiński Opened: 2014-07-14 Last modified: 2018-06-23

Priority: Not Prioritized

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with NAD status.

Discussion:

Currently the heterogeneous lookup in associative container are enabled by presence of is_transparent nested type in the comparator type (24.2.7 [associative.reqmts]). This complicates the definition of call wrapper types that want to define is_transparent if they wrap a callable type that defines is_transparent, and requires the target to be a complete type in cases where an incomplete type would otherwise be ok.

Another problem is that users cannot add the is_transparent member to a third-party comparison type that they do not control, even if they know it supports heterogeneous comparisons.

If the associative containers used a trait instead of checking for an is_transparent member type then it would avoid the requirement for complete types, and would allow customization of the trait without modifying the comparator type. This would also be consistent with the traits is_placeholder and is_bind_expression.

For backward compatibility with the existing design, the default implementation of the is_transparent trait could depend on the presence of the is_transparent nested type.

[2014-11 Urbana]

Move to LEWG

Request for a new metafunction should first be responded to by LEWG.

[2017-02 in Kona, LEWG recommends NAD]

In all discussions of heterogeneous lookup, the entire set of those discussions has been different integer types or string_view vs string. There actually hasn't been an example, that we are aware of, besides those two. -- Could always wrap the third-party type with the comparator? -- Could benefit from a motivating example. -- is_transparent<>, what does it mean?

[2017-06-02 Issues Telecon]

The advantages of changing the API now are not sufficiently clear. We invite a paper presenting motivating examples and more details of the proposed change, but at present consider this Not A Defect.

Resolve as NAD

Proposed resolution:


2432(i). initializer_list assignability

Section: 17.10 [support.initlist] Status: NAD Submitter: David Krauss Opened: 2014-09-30 Last modified: 2022-11-30

Priority: 2

View other active issues in [support.initlist].

View all other issues in [support.initlist].

View all issues with NAD status.

Discussion:

std::initializer_list::operator= 17.10 [support.initlist] is horribly broken and it needs deprecation:

std::initializer_list<foo> a = {{1}, {2}, {3}};
a = {{4}, {5}, {6}};
// New sequence is already destroyed.

Assignability of initializer_list isn't explicitly specified, but most implementations supply a default assignment operator. I'm not sure what 16.3 [description] says, but it probably doesn't matter.

[Lenexa 2015-05-05: Send to EWG as discussed in Telecon]

[2022-08-24; Reflector poll]

Set status to Tentatively NAD after reflector poll in October 2021.

"If somebody wants to revisit it, they'll need to write a paper to demonstrate what the change would break, whether that would be a problem in practice, and convince the evolution groups to make a change. But it's not an LWG issue."

[2022-11-25; see EWG 1369]

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

  1. Edit 17.10 [support.initlist] p1, class template initializer_list synopsis, as indicated:

    namespace std {
      template<class E> class initializer_list {
      public:
        […]
        constexpr initializer_list() noexcept;
      
        initializer_list(const initializer_list&) = default;
        initializer_list(initializer_list&&) = default;
        initializer_list& operator=(const initializer_list&) = delete;
        initializer_list& operator=(initializer_list&&) = delete;
        
        constexpr size_t size() const noexcept;
        […]
      };
      […]
    }
    

2446(i). Unspecialized std::tuple_size should be defined

Section: 22.4.1 [tuple.general] Status: NAD Submitter: Nevin Liber Opened: 2014-10-10 Last modified: 2018-06-23

Priority: Not Prioritized

View all other issues in [tuple.general].

View all issues with NAD status.

Discussion:

In 22.4.1 [tuple.general] paragraph 2, the unspecialized std::tuple_size is undefined. It would be a lot more useful with SFINAE if it were defined as an empty struct; that way, it can be used with enable_if for determining whether or not it is valid to use tuple_size, tuple_element and get on the corresponding data structure.

[2014-11 Urbana]

Moved to LEWG 42.

This request goes beyond simply making an API respond well to SFINAE, but coupling that with an implication for other tuple APIs. The proper place for such design discussions is LEWG.

[2017-02 in Kona, LEWG recommends NAD]

We believe there are other ways to achieve the same results (at least for all of the use cases we could see). We invite Nevin to provide more details / motivations if this is necessary, not just helpful in a limited number of cases. If the dr is raised, variant_size should be considered for the same change as well.

[2017-06-02 Issues Telecon]

It's unclear what benefit this change would have, but making it complete would have very undesirable interactions with structured bindings. The current Core wording for structured bindings depends on tuple_size being an incomplete type.

Resolve as NAD

Proposed resolution:

This wording is relative to N3936.

  1. Change 22.4.1 [tuple.general] p2, header <tuple> synopsis, as indicated

    […]
    // 20.4.2.5, tuple helper classes:
    template <class T> class tuple_size; // undefined
    […]
    
  2. Change 22.4.7 [tuple.helper] as indicated

    […]
    template <class T> struct tuple_size { };
    […]
    

2449(i). vector::insert invalidates end()?

Section: 24.3.11.5 [vector.modifiers] Status: NAD Submitter: Marc Glisse Opened: 2014-10-21 Last modified: 2017-03-14

Priority: 3

View all other issues in [vector.modifiers].

View all issues with NAD status.

Discussion:

this issue is based on the discussion here.

24.3.11.5 [vector.modifiers] says about vector::insert: "If no reallocation happens, all the iterators and references before the insertion point remain valid." This doesn't seem to guarantee anything about the iterator at the point of insertion.

The question comes from people asking if the following is valid, assuming a sufficient call to reserve() was done first:

v.insert(v.end(), v.begin(), v.end());

It could fail for an implementation using a sentinel for the end of the vector, but I don't know of any (it would be quite inconvenient). And for any implementation using a simple position as iterator (pointer (possibly in a wrapper), or base+offset), this is needlessly restrictive. The fact that this alternative:

v.insert(v.end(), &v[0], &v[0]+v.size())

is arguably valid (again assuming a large enough reserve()) makes it a bit confusing that the first version isn't (24.2.4 [sequence.reqmts] has a precondition that iterator arguments to insert() do not point into the sequence, but vector::insert is more refined and seems to give enough guarantees that it cannot fail).

Then we might as well say that vector iterators act as positions, and that after a reallocation-free operation an iterator points to the same position, whatever may be there now…

[2017-03-04, Kona]

NAD. The problem description is incorrect; it is a violation of table entry 87 a.insert(p, i, j) - "i and j are not iterators into a".

Proposed resolution:


2474(i). <cmath> functions unfriendly to integral_constant arguments

Section: 28.7 [c.math] Status: NAD Submitter: Matheus Izvekov Opened: 2015-02-13 Last modified: 2015-12-17

Priority: 4

View all other issues in [c.math].

View all issues with NAD status.

Discussion:

Using numeric wrappers with <cmath> functions is currently troublesome. Code such as: "std::exp2(std::integral_constant<int, 5>{})" will fail to compile because of ambiguity.

Arguments which are implicitly convertible to an arithmetic type should be accepted whenever the latter would be accepted.

David Krauss pointed out that some issue may present itself with numeric libraries which provide <cmath> equivalents in their own namespace which are not more specialized than the ones provided in <cmath>. If the changes proposed are implemented, then cases where the user brings those into scope through ADL might become ambiguous.

It's hard to assess how much breakage this would cause in the wild, and how much work it would take to fix. Should this be determined to be a problem, it's possible to make the new behaviour optional, and disabled by default for all user-defined types.

[2015-10, Kona Saturday afternoon]

STL: This should be NAD, NAD-future and NAD-go-away. Users can already solve this using Walter's call syntax.

VV: I don't personally have this problem, but the proposed resolution seems frightening to me.

VV: There's a related issue, 2294, and also 2192.

STL: 2086 is about user-defined types in <cmath>, fixed in C++14. The PR for 2474 wants to undo the fix.

Link to 2086 and NAD.

Proposed resolution:

This wording is relative to N4296.

  1. Change 28.7 [c.math] p11 b2 as indicated:

    -11- Moreover, there shall be additional overloads sufficient to ensure:

    1. […]

    2. Otherwise, if any arithmetic argument corresponding to a double parameter has type double or an integer typea type which is not floating-point, but which is implicitly convertible to double, then all arithmetic arguments corresponding to double parameters are effectively cast to double.

    3. […]


2535(i). Inconsistency between ostream::write and ostream::operator<<

Section: 23.4.4.4 [string.io], 31.7.6.4 [ostream.unformatted] Status: NAD Submitter: Marshall Clow Opened: 2015-09-10 Last modified: 2016-07-13

Priority: 2

View all other issues in [string.io].

View all issues with NAD status.

Discussion:

Consider the following program:

#include <iostream>
#include <ostream>
#include <string>

template <class CharT>
class testbuf : public std::basic_streambuf<CharT> 
{
public:
  testbuf() {}

protected:
  virtual std::streamsize xsputn(const CharT *s, std::streamsize n)
  {
    std::cout << "xsputn('" << s << "', " << n << ")\n";
    return n;
  }
};

int main () 
{
  testbuf<char> sb;
  std::ostream os (&sb);
  
  std::string s1{"abc"};
  os.write(s1.data(), s1.size());
  
  os.write(s1.data(), 0);
  
  std::string s2{"def"};
  os << s2;
  
  std::string s3{""};
  os << s3;
  
  os << "";
}

What should it print?

libc++:

xsputn('abc', 3)
xsputn('def', 3)

libstdc++:

xsputn('abc', 3)
xsputn('abc', 0)
xsputn('def', 3)
xsputn('', 0)
xsputn('', 0)

VS:

xsputn('abc', 3)
xsputn('def', 3)
xsputn('', 0)
xsputn('', 0)

23.4.4.4 [string.io]/5 seems to say that an implementation is required to call sputn (which calls xsputn) even if there's nothing to output (in the case of ostream::operator<<(basic_string)).

31.7.6.4 [ostream.unformatted]/5.1 implies that it's OK to not call sputn if there's nothing to output (in the case of ostream::write)

Backstory: A user has a ostream with a subclass of basic_streambuf. it creates an output file on first write. Sometimes, he calls ostream::write(p, 0), and expects this to create the file. This doesn't work in libc++, and then he pointed out the inconsistency between operator<< and write.

For reference to a bug report see here.

There are two obvious possible resolutions:

  1. a) require all output functions to call sputn, even if there are no characters to output. In practice, this reduces to "string-like" things which are empty (string, string_view, char*, etc), and write(ptr, len).

  2. b) remove the requirement that ostream::operator<< call sputn when there are no characters to output.

[06-2016 Oulu, Saturday morning]

MC: Problem is this program produces different outputs on different platforms. The issue is what happens when you write 0 bytes: do you have to call xsputn? This affects a real customer.

NJ: Why is this a problem? Why not QOI?

DKu: I don’t think it’s a problem.

MC: Making your own streambuf is an explicit customization point.

DKu: But you should expect different numbers of calls.

NJ: They may even split the input, and call xsputn more than once for an input.

MC: Do we actually say that anywhere?

DKu: I think so.

MC: If you can find that, I’d be OK with NAD. The other thing that bothers me is that in one case it says you’re required to call sputn even if there’s no input, but the other wording doesn’t contain that requirement.

DKu: The first wording says “as if by”, which may give wiggle room to not call it.

DKu: In the second wording, sputc will never call sputn; it puts character into buffer, and calls overflow if the buffer is full. sputn is strictly an optimization.

MC: OK, I’m convinced this could be NAD, that the standard simply gives no guarantees about this. Are we OK with this lack of precision and implementation variance, or does the spec need to be more precise?

DKu: If you look at [ostream]/p2, it deliberately doesn’t specify how the functions are called. Even if sputn is called, no guarantee that xsputn is called at all: if there’s space in the buffer, the implementation may just put the characters in the buffer. This flexibility makes user implementations nicer, so I think this is definitely NAD

BD: This stuff is incredibly chewed-over. There used to always be a group working on this stuff; it’s hard to believe there’s anything in here that’s not deliberate, so you can’t change it with this small group; you need to talk to all the implementers.

MC: Any objections to NAD?

no objections.

Closing as NAD.

Proposed resolution:


2538(i). [parallel.ts] Requirements on data race behavior of iterators and swap should be clarified

Section: 99 [parallel.ts::parallel.alg.general.exec] Status: NAD Submitter: Robert Geva Opened: 2015-09-22 Last modified: 2015-10-21

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: parallel.ts

Parallel algorithms as of N4352 in general need to assume that iterator operations and swap have the expected data race behavior. For example, sort would not work if the exchange operation were thread-unsafe and, for example, always updated a single unprotected global counter, independent of the objects being swapped. A parallel sort has to be able to assume that disjoint pairs of objects can be swapped concurrently, i.e. that elemental functions have the same sort of thread-safety behavior that we generally promise for the standard library versions. I don't see that assumption stated anywhere.

We should then probably also be clearer that the 99 [parallel.ts::parallel.alg.general.exec] ordering rules are further constrained by algorithm correctness requirements, and by the requirement not to introduce data races when elemental functions satisfy their thread-safety expectations.

It's tempting to generally add basic thread-safety requirements to various library requirements clauses. But that adds backwards compatibility issues for single-threaded code.

Most probably this applies to other operations as well. We do state in [parallel.alg.general.user]:

Function objects passed into parallel algorithms as objects of type BinaryPredicate, Compare, and BinaryOperation shall not directly or indirectly modify objects via their arguments.

But that only seems to cover the easy cases.

Proposed resolution:


2552(i). priority_queue doesn't work with move-only types

Section: 24.6.7 [priority.queue] Status: NAD Submitter: Matt Austern Opened: 2015-10-27 Last modified: 2016-02-07

Priority: 3

View all other issues in [priority.queue].

View all issues with NAD status.

Discussion:

Suppose we want to remove and process the topmost element of a priority_queue<T>. For a copyable type we might write

auto tmp = q.top();
q.pop();

but of course that doesn't work if T is move-only. Nothing of that sort can work, since moving out of top() would make the subsequent call to pop() fail.

Currently, pop() is defined to perform

pop_heap(c.begin(), c.end(), comp);
c.pop_back();

so the removed value continues to exist between the first and second lines but there isn't any access to it. The sort of primitive that would allow consuming and removing the topmost element would be some variation on this, e.g.

pop_heap(c.begin(), c.end(), comp);
auto tmp = move(c.back());
c.pop_back();
return tmp;

[2016-02, Issues Telecon]

This should be addressed by a paper addressed to LEWG.

Proposed resolution:


2553(i). [fund.ts.v2] basic_string_view substring constructor

Section: 7.3 [fund.ts.v2::string.view.cons] Status: NAD Submitter: Evan Teran Opened: 2015-10-29 Last modified: 2018-06-23

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: fund.ts.v2

string_view can be tremendously useful for dealing with sub-strings without copying. However, the current proposal for basic_string_view, has no constructor which provides a direct way of creating a view of a sub-string of a basic_string. Instead, we construct a view of the whole basic_string, and then as a second step create a sub-string, for example using substr. To simplify what I believe to be a common use case, I suggest adding an additional constructor.

The proposed wording for this is as follows:

template <class Allocator>
basic_string_view(const basic_string<charT, traits, Allocator>& str, size_type pos, size_type count = npos);

Throws: out_of_range if pos >= str.size().

Effects: Determines the effective length rlen of the string to reference as the smaller of count and size() - pos.

Postcondition:

data_ = str.data() + pos size_ = rlen

In other words, the result is as if constructed via: basic_string_view(basic_string_view(str).substr(pos, count));

An example implementation could look like this:

template <class Allocator>
basic_string_view(const basic_string<charT, traits, Allocator>& str, size_type pos, size_type count = npos) 
  : data_(nullptr), size_(0) 
{
  basic_string_view(str).substr(pos, count).swap(*this);
}

Note that while we have a default parameter for count, pos does not. I believe that it is best to have this as a separate overload, as opposed to default parameters on the current constructor for two reasons:

  1. The current constructor taking a basic_string does not throw, this overload can throw if pos >= str.size().

  2. This constructor performs slightly more work, it is not necessary to impose this extra work on the basic case of constructing a view of a whole string.

This has been briefly discussed in the isocpp forums. There were no obvious objections to this small improvement. Additionally, another reason to consider this addition is to provide a more consistent interface. With raw strings, we have the ability to construct a basic_string_view which is a sub-string. For example:

const char* s = "hello world";
auto v = string_view(s + 6);

But there is no constructor which easily does the same when starting with a basic_string.

Finally, As a example, consider the following (trivial) code:

void print_string(string_view v) {
  std::cout << v << '\n';
}

int main() {
  std::string s = "hello world"; // for example, we want to print the sub-string "world", without copies

  // current method:
  print_substring(string_view(s).substr(6));

  // suggested method:
  print_substring(string_view(s, 6);
}

Previous resolution [SUPERSEDED]:

This wording is relative to N4529.

  1. Insert between 7.3 [fund.ts.v2::string.view.cons] p5 and p6 the following sequence of paragraphs:

    template <class Allocator>
    basic_string_view(const basic_string<charT, traits, Allocator>& str, size_type pos, size_type count = npos);
    

    -?- Throws: out_of_range if pos >= str.size().

    -?- Effects: Determines the effective length rlen of the string to reference as the smaller of count and size() - pos.

    -?- Postcondition: Constructs a basic_string_view, with the postconditions in Table ?

    Table ? — basic_string_view(const basic_string<charT, traits, Allocator>&, size_type, size_type) effects
    Element Value
    data_ str.data() + pos
    size_ rlen

[2016-03, Jacksonville]

Change status to "LEWG"

LEWG: Do we want this constructor?

SF F N A SA

0 3 2 13 1

Proposed resolution:

Not a defect. The LWG believes this missing feature is not sufficiently serious to constitute a defect.


2563(i). LWG 2259 relaxes requirements, perhaps unintentionally

Section: 16.4.6.5 [member.functions] Status: NAD Submitter: Ville Voutilainen Opened: 2015-11-29 Last modified: 2018-11-12

Priority: 2

View all other issues in [member.functions].

View all issues with NAD status.

Discussion:

The combination of 16.4.6.5 [member.functions], paragraphs 2 and 3 that LWG 2259 does seems to drop a requirement that any call behaves as if no overloads were added. Paragraph 3 used to say "A call to a member function signature described in the C ++ standard library behaves as if the implementation declares no additional member function signatures." whereas the new wording says "provided that any call to the member function that would select an overload from the set of declarations described in this standard behaves as if that overload were selected."

This can be read as meaning that if there's no default constructor specified, like for instance for std::ostream, an implementation is free to add it. It can also be read as meaning that an implementation is free to add any overloads that wouldn't change the overload resolution result of any call expression that would select a specified overload. That's vastly different from allowing extensions that add new functions rather than new overloads.

Was this relaxation intentional?

[2016-04, Issues Telecon]

Ville provides a motivating example.

#include <iostream>

class my_stream : std::ostream
{
};

int main()
{
    my_stream ms;
}

This example is accepted by libstdc++, msvc rejects it, and clang+libc++ segfault on melpon.org/wandbox o_O. An earlier clang+libc++ just accepts it. I don't think the implementation divergence is caused by the acceptance of the referred-to 2259, but it certainly seems to increasingly bless the implementation divergence.

[2016-05 Issues Telecon]

This is related to issue 2695.

[2018-08 Batavia Monday issue discussion]

Ville to provide wording.

[2018-08 Batavia Monday issue discussion]

Ville recommends NAD; because closing this would outlaw conforming extensions.

[2018-11 San Diego Thursday night issue processing]

Status to NAD

Proposed resolution:


2580(i). Who is definitive: operator= or assign?

Section: 23.4.3.3 [string.cons], 23.4.3.7.3 [string.assign], 32.7.3 [re.regex.assign] Status: NAD Submitter: Marshall Clow Opened: 2016-01-05 Last modified: 2016-11-12

Priority: 4

View all other issues in [string.cons].

View all issues with NAD status.

Discussion:

There are two "containers" in the standard who have member functions named assign that take parameters of the type of the container (as opposed to iterators, pointers, what have you).

In string's case, we define assign in terms of operator=. In regex's case, we define operator= in terms of assign.

We should pick a style and use use it.

In 23.4.3.3 [string.cons], we have:

basic_string& operator=(const basic_string& str);

-17- Effects: If *this and str are not the same object, modifies *this as shown in Table 70.

-18- If *this and str are the same object, the member has no effect.

-19- Returns: *this

basic_string& operator=(basic_string&& str)
  noexcept(allocator_traits<Allocator>::propagate_on_container_move_assignment::value || 
           allocator_traits<Allocator>::is_always_equal::value);

-20- Effects: Move assigns as a sequence container (23.2), except that iterators, pointers and references may be invalidated.

-21- Returns: *this

In 23.4.3.7.3 [string.assign], we have:

basic_string& assign(const basic_string& str);

-1- Effects: Equivalent to assign(str, 0, npos).

-2- Returns: *this.

basic_string& assign(basic_string&& str) 
  noexcept(allocator_traits<Allocator>::propagate_on_container_move_assignment::value ||
           allocator_traits<Allocator>::is_always_equal::value);

-2- Effects: Equivalent to *this = std::move(str).

-3- Returns: *this.

Marshall says: There is another issue 2579 here, to change /1 to be similar to /2.

In 32.7.3 [re.regex.assign], we have:

basic_regex& operator=(const basic_regex& e);

-1- Effects: returns assign(e).

basic_regex& operator=(basic_regex&& e) noexcept;

-2- Effects: returns assign(std::move(e)).

and

basic_regex& assign(const basic_regex& that);

-7- Effects: copies that into *this and returns *this.

-8- Postconditions: flags() and mark_count() return that.flags() and that.mark_count(), respectively.

basic_regex& assign(basic_regex&& that) noexcept;

-9- Effects: move assigns from that into *this and returns *this.

-10- Postconditions: flags() and mark_count() return the values that that.flags() and that.mark_count(), respectively, had before assignment. that is in a valid state with unspecified value.

[2016-02, Issues Telecon]

Marshall to see if this can be dealt with editorially. Change Regex so that assign is in terms of op=

[2016-02]

Changed basic_regex to match string as an editorial change. Closing as NAD

Proposed resolution:


2600(i). ios_base must store inaccessible iostate flags

Section: 31.5.2.6 [ios.base.storage] Status: NAD Submitter: David Krauss Opened: 2016-03-14 Last modified: 2019-02-26

Priority: Not Prioritized

View all other issues in [ios.base.storage].

View all issues with NAD status.

Discussion:

DR 41, "Ios_base needs clear(), exceptions()" stopped short of providing the interface suggested in its title, but it did require the underlying state to be stored in ios_base. Because rdstate() is also missing, ios_base manipulators relying on iword and pword cannot detect failure. The only safe alternative is to manipulate a derived class, which must be a template.

libc++ already provides the interface as a nonconforming extension. libstdc++ implements the internal state but leaves it frustratingly inaccessible, as specified. Any conforming implementation should be able to provide the interface without ABI problems.

[2016-04, Issues Telecon]

This is really a request for an (feature) API. Passing to LEWG.

[LEWG Kona 2017]

Recommend NAD: iostreams aren't used in this way enough to spend committee time on it. However, a paper could change our minds.

Proposed resolution:


2604(i). [filesys.ts] [PDTS] Need definition of dot and dot-dot

Section: 4.14 [filesys.ts::fs.def.parent] Status: NAD Submitter: CH-3 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

The concept of a parent directory for dot or dotdot exists, but the definition doesn't apply.

Suggested action:

Remove the paragraph. This concept does not apply to dot and dot-dot. Add a definition for dot and dot-dot.

[2014-02-07, Beman Dawes comments]

Suggest it is editorial and should be passed to the project editor.

[ 2014-02-11 Issaquah: Beman to provide wording for review next meeting. Also see related issue 5. ]

[22 May 2014 Beman Dawes comments:]

I've now reviewed this issue carefully and believe it is NAD. "parent" is used in four places in the WP, and so deserves a definition. The current definition is copied word-for-word and in its entirety from the POSIX definition. I believe strongly that the File System TS needs to stay in alignment with POSIX on this matter, and that the best way to do that is simply to use the POSIX wording.

[17 Jun 2014 Rapperswil LWG closes as NAD. No concensus for change.]

Proposed resolution:


2610(i). [filesys.ts] [PDTS] Apparently inconsistent return types from several functions

Section: 6 [filesys.ts::fs.filesystem.synopsis] Status: NAD Submitter: FI-4 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.filesystem.synopsis].

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

It is unclear why copy, copy_file and copy_symlink have different return types, and why the attribute-version of create_directory has a different return type than the create_directory that takes no attributes. The status/error reporting in these functions seems inconsistent.

Resolution:

Make the status/error reporting consistent or add a note explaining why it's different.

[ 2014-02-11 Issaquah: NAD. LWG/SG-3 reviewed each function and return type, and found that since they have different functionality different return types are warranted. create_directory has an inconsistent return type between the synopsis and the description. This has subsequently been corrected editorially. ]

Proposed resolution:


2611(i). [filesys.ts] [PDTS] Lack of relative() operation function

Section: 6 [filesys.ts::fs.filesystem.synopsis], 15 [filesys.ts::fs.op.funcs] Status: NAD Future Submitter: GB-1 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.filesystem.synopsis].

View all issues with NAD Future status.

Discussion:

Addresses: filesys.ts

There is no relative() operation, to complement both absolute() and canonical()

The TS introduces relative paths.

However there is no way to create a relative path as a path relative to another. Methods are provided to create absolute and canonical paths.

In section 15.1 Absolute [fs.op.absolute]:

path absolute(const path& p, const path& base=current_path());

and in section 15.2 Canonical [fs.op.canonical]

path canonical(const path& p, const path& base = current_path());

path canonical(const path& p, error_code& ec);

path canonical(const path& p, const path& base, error_code& ec);

By providing a operations to achieve absolute and canonical paths there is no impediment to providing a similar operation relative() that attempts to return a new path relative to some base path.

For example:

path relative(const path& p, const path& to = current_path());

path relative(const path& p, error_code& ec);

path relative(const path& p, const path& to, error_code& ec);

This would return a path, if possible, that is relative to to. The implementation can make use of absolute() and canonical() to determine the relative path, if it exists.

The File System TS is based on the boost::filesystem library and it too suffers from this anomaly. There are open tickets for this in Boost Trac:

and it is the subject of several posts on StackOverflow for example:

Other languages typically provide a similar function. For example python provides:

os.path.relpath(path[, start])

Return a relative filepath to path either from the current directory or from an optional start directory. This is a path computation: the filesystem is not accessed to confirm the existence or nature of path or start. start defaults to os.curdir.

[2014-02-07, Beman Dawes comments]

A relative() function is useful and much requested. I've seen such a function provided by users and have written it myself in app code. It is one of those things I've been meaning to do for years, and have just never gotten around to.

That said, my mild preference is to treat this as "NAD, Future" for File System TS1, but treat it as a priority for TS2.

[ 2014-02-11 Issaquah ]

The LWG/SG-3 voted strongly in favor of adding this functionality, and doing so in this TS. That implies quite a bit of work before the next meeting to validate that the proposed interface works as desired for various platforms. There was general agreement not to hold FS STS1 if this functionality isn't ready when the rest of the TS is ready.

[2014-05-19 Beman Dawes supplied wording.]

The design benefited from discussions with Jamie Allsop, who was the source of the original NB comment. Thanks to Bjorn Reese for corrections and suggestions. Although there was also discussion and experimentation with additional relative functions that took into account symlinks and normalization, these are not proposed here since even the proponents of such functions were unsure of appropriate semantics.

[2014-06-17 Rapperswil LWG closes as NAD, Future.]

Although there is strong concensus for eventually providing both lexical and existence based flavors of relative() functionality, discussion of the many possible design choices led to the conclusion that more research and actual user experience is necessary before moving forward. Interested parties should submit papers.

Original proposed resolution:

  1. Modify header <filesystem> synopsis, 6 [fs.filesystem.synopsis], by adding the operational functions after canonical:

    path relative(const path& p, const path& to = current_path());
    path relative(const path& p, error_code& ec);
    path relative(const path& p, const path& to, error_code& ec);
    
  2. Insert the section:

    15.3 Relative [fs.op.relative]

    path relative(const path& p, const path& to = current_path());
    path relative(const path& p, error_code& ec);
    path relative(const path& p, const path& to, error_code& ec);

    Overview: Return a relative path of p to the current directory or from an optional to path.

    Returns: A relative path such that canonical(to)/relative(p,to) == canonical(p), otherwise path(). If canonical(to) == canonical(p) the path path(".") is returned. For the overload without a to argument, to is current_path(). Signatures with argument ec return path() if an error occurs.

    Throws: As specified in Error reporting.

    Remarks: !exists(p) or !exists(to) or !is_directory(to) is an error.

    and bump all following sections up by 0.1. Update the contents and any cross-references accordingly.

Question: Should Returns be specified in terms of equivalence? For example: equivalent( canonical(to)/relative(p,to), canonical(p) )

Question: Should canonical(to) == canonical(p) return path(".") or path()? Why?

Question: Should to be spelled start?

Proposed resolution:

To 6 Header <experimental/filesystem> synopsis [fs.filesystem.synopsis], add:

path lexically_relative(const path& p, const path& base);

At the end of 8.6 path non-member functions [path.non-member], add

8.6.3 path lexically_relative function [path.lexically.relative]

path lexically_relative(const path& p, const path& base);

Creates a path from the trailing elements of p that are lexically relative to base, which must be a prefix of p.

Effects: If the number of elements in [ p.begin(), p.end() ) is less than or equal to the number of elements in [ base.begin(), base.end() ), or if any element in [base.begin(), base.end()) is not equal to the corresponding element in [p.begin(), p.end()), throw an exception of type filesystem_error.

Remarks: Equality or inequality are determined by path::operator== or path::operator!= respectively.

Returns: An object of class path containing the first element of p that does not have a corresponding element in base, followed by the subsequent elements of p appended as if by path::operator/=.

Throws: filesystem_error.

[Note: Behavior is determined by the lexical value of the elements of p and base - the external file system is not accessed. The case where an element of base is not equal to corresponding element of p is treated as an error to avoid returning an incorrect result in the event of symlinks.  --end note]

A possible implementation would be:

        
auto mm = std::mismatch( p.begin(), p.end(), base.begin(), base.end());
if (mm.first == p.end() || mm.second != base.end())
{
throw filesystem_error(
"p does not begin with base, so can not be made relative to base",
p, base,
error_code(errc::invalid_argument, generic_category()));
}
path tmp(*mm.first++);
for (; mm.first != p.end(); ++mm.first)
tmp /= *mm.first;
return tmp;

2612(i). [filesys.ts] [PDTS] uintmax_t too small for large file sizes

Section: 6 [filesys.ts::fs.filesystem.synopsis], 15.14 [filesys.ts::fs.op.file_size] Status: NAD Future Submitter: CH-8 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.filesystem.synopsis].

View all issues with NAD Future status.

Discussion:

Addresses: filesys.ts

uintmax_t is specified to hold at least 64 bit. This is not enough for sizes beyond 4 [sic] exabytes.

Specify whether an implementation must provide a uintmax_t that can hold the maximum possible space and file size values.

[2014-02-06: Jeffery Yasskin points out 64-bits unsigned actually has a maximum value of "16 exabytes, not 4"]

[2014-02-07: Beman Dawes suggests: This should be NAD. Such ultra-large files are the province of enterprise-wide filesystems such as requested by IBM and others for a follow-on SG3 TS. That would be the best vehicle to address this concern IMO.]

[ 2014-02-11 Issaquah: NAD Future. ]

Proposed resolution:


2613(i). [filesys.ts] [PDTS] Missing actual error conditions thrown

Section: 7 [filesys.ts::fs.err.report] Status: NAD Submitter: CH 9 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.err.report].

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

The specification of the actual error conditions for the functions that specify Throws: As specified in Error reporting. is missing.

Add those specifications.

[2014-02-07, Beman Dawes comments]

The actual error codes, and thus the error conditions, are determined by the operating system, and thus operating system dependent.

[2014-02-11 Issaquah]

There is no consensus for a change. LWG/SG3 requested a separate issue be opened to clarify 7 [fs.err.report]. See issue 55.

Proposed resolution:


2617(i). [filesys.ts] [PDTS] path member swap() unnecessary

Section: 8.4.5 [filesys.ts::path.modifiers] Status: NAD Submitter: CH-12 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

As we have move semantics, member swap functions shouldn't be necessary any more.

[2014-02-12 LWG/SG-3 Issaquah ]

No consensus for change. STL pointed out that swap should be noexcept and will submit a separate issue.

Proposed resolution:

Remove swap().


2623(i). [filesys.ts] [PDTS] Request for create_regular_file() and/or touch()

Section: 15 [filesys.ts::fs.op.funcs] Status: NAD Submitter: CH-14 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.op.funcs].

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

Since create_symlink(), create_hardlink(), and create_directory() exist, there's no reason not to have a create_regular_file() function.

Consider adding a function create_regular_file() with the behaviour of the POSIX touch command.

[Beman comments]

create_regular_file() and touch() should be different functions since their behavior would differ if the file already exists; touch() updates the last write date.

I have often wanted create_regular_file() but never got around to adding it.

While touch is quite useful from the command line, I'm less sure of its usefulness as a library function. It is trivial for a user to implement.

Whether or not it is appropriate to add operational functions this late in the PDTS process is questionable.

[2014-02-13 LWG/SG-3 Issaquah: No consensus for change at this time.]

Proposed resolution:


2626(i). [filesys.ts] [PDTS] Equivalence is a volatile property

Section: 15.13 [filesys.ts::fs.op.equivalent] Status: NAD Submitter: CH-16 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

Equivalence is a volatile property.

Consider adding a note that equivalence cannot be determined race-free.

[2014-02-13 LWG/SG-3 Issaquah: No consensus for change. Section 2.1 description of races is sufficient.]

Proposed resolution:


2628(i). [filesys.ts] [PDTS] Possible last_write_time() postcondition?

Section: 15.25 [filesys.ts::fs.op.last_write_time] Status: NAD Submitter: GB-15 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.op.last_write_time].

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

The constraint on last_write_time is too weak: It is noted that the postcondition of last_write_time(p) == new_time is not specified since it might not hold for file systems with coarse time granularity.

However, might it be possible to have a postcondition that last_write_time(p) <= new_time ?

Add postcondition: last_write_time(p) <= new_time

[2014-02-09, Beman Dawes comments:]

That assumes the file system rounds down. We don't know which direction a file system rounds. Nice try, but this one looks NAD to me.

[2014-02-13 LWG/SG-3 Issaquah: No consensus for change.]

Proposed resolution:


2630(i). [filesys.ts] [PDTS] remove() must avoid race

Section: 15.28 [filesys.ts::fs.op.remove] Status: NAD Submitter: CH-17 Opened: 2014-01-20 Last modified: 2016-08-10

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

The specification can be read to require the existence test. As this introduces a race, the existence test must not happen.

Change to: "Effects: p is removed as if by POSIX remove()."

[2014-02-13 LWG/SG-3 Issaquah: Insufficient consensus for change. Vote for NAD: 9 0 0 2 1.]

Proposed resolution:

This wording is relative to SG3 working draft.

  1. Change 15.28 [filesys.ts::fs.op.remove] as indicated:

    Effects: If exists(symlink_status(p,ec)), itp is removed as if by POSIX remove().


2631(i). [filesys.ts] [PDTS] POSIX guarantees atomicity for rename()

Section: 15.30 [filesys.ts::fs.op.rename] Status: NAD Submitter: CH-18 Opened: 2014-01-20 Last modified: 2016-08-12

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

POSIX guarantees some kind of atomicity for rename().

Clarify that POSIX' rename() guarantee "If the rename() function fails for any reason other than [EIO], any file named by new shall be unaffected." holds for C++ as well.

[2014-02-10 Beman Dawes]

Section 2.1, POSIX conformance, [fs.conform.9945] specifies the POSIX conformance requirements for TS implementations in carefully crafted and specific detail. Repeating a portion of the POSIX standard's specification for a particular TS function would do great harm as it would bring into question all of the portions of the POSIX specification for the function that were not repeated.

Furthermore, all the caveats and other details of the 2.1 specification would have to be analyzed and possibly appended; it ties the hands of implementors if they are not given latitude to deviate as needed when working with non-POSIX operating systems.

I strongly recommend NAD for this issue.

[2014-02-13 LWG/SG-3 Issaquah: No consensus for change.]

Proposed resolution:


2638(i). [filesys.ts] [PDTS] Make certain functions noexcept and drop error_code version

Section: 12.3 [filesys.ts::directory_entry.obs], 15.12 [filesys.ts::fs.op.exists], 15.16 [filesys.ts::fs.op.is_block_file], 15.17 [filesys.ts::fs.op.is_char_file], 15.18 [filesys.ts::fs.op.is_directory], 15.19 [filesys.ts::fs.op.is_empty], 15.20 [filesys.ts::fs.op.is_fifo], 15.21 [filesys.ts::fs.op.is_other], 15.22 [filesys.ts::fs.op.is_regular_file], 15.23 [filesys.ts::fs.op.is_socket], 15.24 [filesys.ts::fs.op.is_symlink], 15.33 [filesys.ts::fs.op.status], 15.35 [filesys.ts::fs.op.symlink_status], 99 [filesys.ts::fs.op.unique_path] Status: NAD Submitter: P.J. Plauger Opened: 2014-01-30 Last modified: 2016-08-12

Priority: Not Prioritized

View all other issues in [filesys.ts::directory_entry.obs].

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

exists(const path&) should be noexcept (drop error_code version).
is_*(const path&) should be noexcept (drop error_code version).
status(const path&) should be noexcept (drop error_code version).
symlink_status(const path&) should be noexcept (drop error_code version).
file_status::status() should be noexcept (drop error_code version).
file_status::symlink_status() should be noexcept (drop error_code version).
unique_path(const path&) should be noexcept (drop error_code version).

[2014-02-08: Daniel comments]

unique_path(const path&) cannot be declared as noexcept, because it returns an object during whose construction a memory allocation request may fail, see the rationale provided in 2637.

exists(const path&) and the is_*(const path&) functions cannot be noexcept, because they are specified in terms of status(const path&), which again may throw an exception, which is explicitly described in the Effects (filesystem_error), because the non-throwing function (status(const path&, error_code&)) may fail to satisfy the request.

symlink_status(const path&) may throw an exception for similar reasons that status(const path&) could fail.

The reference to file_status::status()/symlink_status() looks like a typo to me (there are no such functions in file_status), presumably directory_entry::status()/symlink_status() was meant. In this case I see no reason how these could be marked as noexcept, because these functions all may fail and may throw an exception.

Based on this interpretation of the issue discussion I recommend to resolve this issue as NAD.

[Beman Dawes 2014-02-27]

Issues 2637, 2638, 2641, and 2649 are concerned with signatures which should or should not be noexcept. I will provide unified proposed wording for these issues, possibly in a separate paper.

[17 Jun 2014 Rapperswil LWG closes as NAD. Working paper correct as written.]

Proposed resolution:


2639(i). [filesys.ts] [PDTS] permissions() is missing from synopsis

Section: 6 [filesys.ts::fs.filesystem.synopsis] Status: NAD Editorial Submitter: P.J. Plauger Opened: 2014-01-30 Last modified: 2016-08-12

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.filesystem.synopsis].

View all issues with NAD Editorial status.

Discussion:

Addresses: filesys.ts

permissions function is missing from the summary list.

Proposed resolution:

[2014-02-07: Beman Dawes comments: Fixed as Editorial.]


2642(i). [filesys.ts] [PDTS] class path should have defaulted constructors/destructor/assignments.

Section: 8 [filesys.ts::class.path] Status: NAD Submitter: P.J. Plauger Opened: 2014-01-30 Last modified: 2016-08-12

Priority: Not Prioritized

View all other issues in [filesys.ts::class.path].

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

class path should have defaulted constructors/destructor/assignments.

[2014-02-26 Beman Dawes comments]

Suggest NAD. Earlier versions did have defaulted constructors/destructor/assignments, but they were removed at the request of Alberto Ganesh Barbati (c++std-filesys January 8, 2013):

Speaking of =default, we have to be careful not over-constrain the specification. I mean, if we just specify the intended meaning of those function with proper wording, the implementation is allowed to use =default in case it provides an equivalent behaviour, but if we put =default in the specification, the implementation is required to use it. However, we have to remember that "for exposition only" data members may not be an exhaustive list and that 17.5.2.3/2 allows implementations to provide an equivalent behaviour using different members for which "default" construction/copy/assignment may not be appropriate.

=default is what we want for tuple, atomics, etc. It might be appropriate for simple types like file_status, but, for more complex types like path itself, I'd remove it and add proper wording.

[17 Jun 2014 Rapperswil LWG closes as NAD. Ganesh's analysis is correct. WP correct as written.]

Proposed resolution:


2643(i). [filesys.ts] [PDTS] path::compare(const string&) should be path::compare(const string_type&)

Section: 8 [filesys.ts::class.path] Status: Dup Submitter: P.J. Plauger Opened: 2014-01-30 Last modified: 2016-08-12

Priority: Not Prioritized

View all other issues in [filesys.ts::class.path].

View all issues with Dup status.

Discussion:

Addresses: filesys.ts

path::compare(const string&) should be path::compare(const string_type&).

[2014-02-08 Daniel comments]

This issue is a duplicate of 2650. The suggested wording of that issue would resolve this issue here as well.

[2014-02-13 LWG/SG-3 Issaquah: Agrees with Daniel.]

Proposed resolution:


2646(i). [filesys.ts] [PDTS] Do we really need generic*?

Section: 8.4.7 [filesys.ts::path.generic.obs] Status: NAD Submitter: P.J. Plauger Opened: 2014-01-30 Last modified: 2016-08-12

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

Do we really need generic*?

[2014-02-08 Daniel comments]

These functions should exist for more than one reason.

First, the generic pathname format is a well-defined concept for the path specification and defining just a single way into a path but not out of it looks like an incomplete design.

More importantly, the existence of these functions have demonstrated to be quite useful in practice, because the generic pathname format concept is popular for many tools such as maven or the some configuration files for the Eclipse IDE do understand this syntax uniformly on all platforms and it allows to generate or modify such files using C++ code based on filesystem::path. The practical problem of the non-portable root-names doesn't matter in such contexts, because the serialized path names are in general relative names or depend on initial parts that are determined by properties or environment variables.

In other words: I'm recommending NAD for this issue.

[2014-02-13 LWG/SG-3 Issaquah: Withdrawn by submitter.]

Proposed resolution:


2651(i). [filesys.ts] [PDTS] directory_iterator, recursive_directory_iterator, pointer/reference typedefs wrong

Section: 13 [filesys.ts::class.directory_iterator], 14 [filesys.ts::class.rec.dir.itr] Status: Dup Submitter: Stephan T. Lavavej Opened: 2014-02-03 Last modified: 2016-08-11

Priority: Not Prioritized

View all other issues in [filesys.ts::class.directory_iterator].

View all issues with Dup status.

Discussion:

Addresses: filesys.ts

directory_iterator and recursive_directory_iterator are constant iterators, but their pointer/reference typedefs are wrong (std::iterator defaults to providing modifiable ones).

[2014-02-08 Daniel comments]

I noticed the same problem when trying to resolve 2652. The currently suggested wording for that issue should fix the here mentioned problem as well.

I recommend to solve this issue as "Resolved by the proposed wording for 2652".

[2014-02-13 LWG/SG-3 Issaquah: Daniel's resolution accepted.]

Proposed resolution:


2654(i). [filesys.ts] [PDTS] Concerns with security and testability

Section: 1 [filesys.ts::fs.scope] Status: NAD Future Submitter: Google Opened: 2014-01-20 Last modified: 2016-08-11

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.scope].

View all issues with NAD Future status.

Discussion:

Addresses: filesys.ts

We have two primary concerns with the interface as specified:

(a) its interface repeats the mistake of V7 Unix in 1979 by exposing access checking (and similarly file creation) independently from opening and mutating the file, and

(b) it provides no realistic means of testing a software library which uses the standard interface for accessing the filesystem under fault scenarios.

Due to the extent of (a), TOCTTOU [1] security vulnerabilities are guaranteed, if not during access checking[2], during other common operations such as temporary file creation[3].

Due to (b) it is impossible to portably test libraries using the proposed interface against critical correctness and security edge cases.

[1]: TOCTTOU: Time-of-check-to-time-of-use.  Operating system integrity in OS/VS2

[2]: Fixing Races for Fun and Profit: How to use access(2)

[3]: Checking for Race Conditions in File Accesses

[Beman Dawes: 10 Feb 2014: Suggested response: NAD, Future]

We share your concerns and look forward to receiving specific proposals to address them. Whether they will addressed by a revision of TS 18822 or a new TS will be decided as proposals progress through the committee process. See How To Submit a Proposal.

[17 Jun 2014 Rapperswil LWG agrees NAD, Future with rationale as stated above.]

Proposed resolution:


2659(i). [filesys.ts] [PDTS] Invalid expressions for bitmask types

Section: 10 [filesys.ts::fs.enum] Status: NAD Editorial Submitter: Daniel Krügler Opened: 2014-03-01 Last modified: 2016-08-11

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

Addresses: filesys.ts

copy_options is declared as enum class type that is a bismask type, but the specification repeatedly uses expressions that are invalid for scoped enums, such as:

!(options)

because there is no contextual conversion to bool, not even the || operator in:

((options & copy_options::recursive) || !(options))

Affected are basically all formulations in the form:

"if options & copy_options::create_symlinks [..]"

because all rely on contextual conversion to bool. The only other specifically mention scoped enumeration in the standard that is also a bit mask type is the launch enum and the wording there always uses forms such as:

"if policy & launch::deferred is non-zero"

which better acknowledges the fact that the obtained values does not necessarily undergo an implicit conversion.

I think the current wording in the file system spec. must be changed, especially for invalid expressions of the form:

((options & copy_options::recursive) || !(options))

A similar problem arises in the usage of the bitmask type perms for the expression:

((prms & add_perms) && (prms & remove_perms))

The only way how to describe this with a scoped enum is the lengthier form

((prms & perms::add_perms) != perms::none && (prms & perms::remove_perms) != perms::none)

thus fixing several problems:

[20 May 2014 Beman Dawes provides proposed wording. Fixing invalid C++ is editorial, but treating this as an issue ensures more people review the proposed changes.]

[17 Jun 2014 Rapperswil LWG requests issue be handled as editorial.]

Proposed resolution:

Change 15.3 Copy [fs.op.copy]:

Before the first use of f and t:

Report an error as specified in Error reporting (7) if:

If is_symlink(f), then:

Otherwise if is_regular_file(f), then:

Otherwise if is_directory(f) && ((options & copy_options::recursive) != copy_options::none || !(options == copy_options::none)) then:

Change 15.4 Copy file [fs.op.copy_file]:

If  exists(to) && !(options & (copy_options::skip_existing | copy_options::overwrite_existing | copy_options::update_existing)) == copy_options::none report a file already exists error as specified in Error reporting (7).

If !exists(to) || (options & copy_options::overwrite_existing) != copy_options::none || ((options & copy_options::update_existing) != copy_options::none && last_write_time(from) > last_write_time(to)) || !(options & (copy_options::skip_existing | copy_options::overwrite_existing | copy_options::update_existing)) == copy_options::none copy the contents and attributes of the file from resolves to the file to resolves to.

Change 15.26 Permissions [fs.op.permissions]:

Requires: !((prms & perms::add_perms) != perms::none && (prms & perms::remove_perms) != perms::none).


2661(i). [filesys.ts] Surprising equivalent() behavior if neither file exists

Section: 6 [filesys.ts::fs.filesystem.synopsis] Status: NAD Submitter: Beman Dawes Opened: 2014-04-12 Last modified: 2016-08-11

Priority: Not Prioritized

View all other issues in [filesys.ts::fs.filesystem.synopsis].

View all issues with NAD status.

Discussion:

Addresses: filesys.ts

bool equivalent(const path& p1, const path& p2); has always thrown a exception if neither file exists, with rationale that if they don't exist, it isn't possible to tell if two paths are equivalent. Dave Abrahams has reported that this is counter-intuitive and hard to teach.

An alternative if neither path exists would be to return true if they are lexically equal (operator==), otherwise return false.

This was not a national body comment, and Dave is the only one I can recall ever complaining about the current behavior. On the other hand, any complaint from Dave deserves serious consideration.

[17 Jun 2014 Rapperswil LWG considers this NAD. Mixing lexical and existence based behavior is not desirable.]

Proposed resolution:


2666(i). Bitmask operations should use bitmask terms

Section: 31.12 [filesystems] Status: NAD Editorial Submitter: Jonathan Wakely Opened: 2014-06-30 Last modified: 2017-06-15

Priority: 3

View all other issues in [filesystems].

View all issues with NAD Editorial status.

Discussion:

C++14 17.5.2.1.3 Bitmask types [bitmask.types] paragraph 4 specifies:

The following terms apply to objects and values of bitmask types:

The TS should use these forms where applicable throughout the document.

[2016-02, Jacksonville]

This is an editoral change.

[Apr 2016 Issue updated to address the C++ Working Paper. Previously addressed File System TS]

Proposed resolution:


2668(i). path::operator+= is defined, but not operator+

Section: 31.12.6 [fs.class.path] Status: NAD Submitter: Jonathan Wakely Opened: 2014-07-03 Last modified: 2017-03-19

Priority: 3

View all other issues in [fs.class.path].

View all issues with NAD status.

Discussion:

This doesn't seem to be in Boost.Filesystem, so maybe it isn't needed, but since path += path2 works then it seems reasonable to expect path1 + path2 to work as well.

[04 Jul 2014 Beman Dawes comments:]

The 12 overloads required by basic_string operator+ scared me off, and I never came back to the issue.

[22 Nov 2015 Beman supplies proposed resolution wording.]

[Apr 2016 Issue updated to address the C++ Working Paper. Previously addressed File System TS]

[2016-11-12, Issaquah]

Sat PM: JW is no longer convinced that this is a good idea.

Proposed resolution:

NAD, Future.

It is not necessary to provide every possible string operation for class path because it is always possible to convert to a string, perform the operation, and convert back to a path. The most commonly needed string operations are provided for class path as a convenience, but every added function comes at the cost of increased interface complexity. In this case, the cost is judged to outweigh the convenience.

Future changes to the language, such as concepts, and changes to the library, such as basic_string_view, may allow reduction of the complexity of the class path interface. The LWG may wish to reconsider this issue at that time.


2692(i). Overspecification of lvalueness of bitmask elements

Section: 16.3.3.3.3 [bitmask.types] Status: NAD Submitter: Hubert Tong Opened: 2016-04-14 Last modified: 2017-07-17

Priority: 3

View all other issues in [bitmask.types].

View all issues with NAD status.

Discussion:

The usual pattern now used for identifying where bitmask elements are declared, namely, as variables, preclude declaring them as enumerators.

Compare: ctype_base::space in C++03 subclause 22.2.1 [lib.category.ctype] versus the same in N4582 subclause 22.4.1 [category.ctype].

It is unclear whether this is intentional. Further it is unclear if odr-use of bitmask elements is intended to be allowed.

[2016-05 Issues Telecon]

Jonathan believes that this was intentional, and was done by N3110. Jonathan will provide more precise references.

Proposed resolution:


2700(i). resource_adaptor went missing

Section: 20.4 [mem.res] Status: NAD Submitter: Jonathan Wakely Opened: 2016-05-04 Last modified: 2016-11-26

Priority: 1

View all other issues in [mem.res].

View all issues with NAD status.

Discussion:

The alias template std::experimental::resource_adaptor is an important part of the polymorphic memory resources framework, but was not adopted into the WP along with the rest of LFTSv1. This seems to be an accidental omission. It was present in P0220R0, but absent in P0220R1. There was no discussion of leaving it out in Jacksonville, it is one of the "core features" discussed and intended for adoption.

[Oulu, 2016-06]

Tuesday night: This was not an accidental omission.

std::experimental::resource_adaptor will remain in the LFTS and we shall continue to gain implementation experience.

Proposed resolution:

  1. Add section 8.7 from N4562 (LFTSv2 DTS) into the WP.


2701(i). Unclear requirement in [memory.resource.private]

Section: 20.4.2.3 [mem.res.private] Status: NAD Editorial Submitter: Jonathan Wakely Opened: 2016-05-04 Last modified: 2017-06-15

Priority: 3

View all other issues in [mem.res.private].

View all issues with NAD Editorial status.

Discussion:

[memory.resource.private] says:

virtual void* do_allocate(size_t bytes, size_t alignment) = 0;

-1- Requires: Alignment shall be a power of two.

Is that supposed to be a requirement on the alignment specified by the alignment parameter? If so the word "Alignment" should not be capitalized and in text font.

[2016-05 Issues Telecon]

The resolution should reference [basic.align]. Jonathan to update the wording.

[2016-08 - Chicago]

Thurs AM: Moved to NAD Editorial. Will change "Alignment" to alignment and add a reference to [basic.align]

Proposed resolution:

  1. Modify [memory.resource.private] paragraph 1 as shown:

    virtual void* do_allocate(size_t bytes, size_t alignment) = 0;
    

    -1- Requires: AlignmentThe alignment specified by alignment shall be a power of two.


2717(i). scoped_allocator_adaptor uses forward to do move's job

Section: 20.5.4 [allocator.adaptor.members] Status: NAD Submitter: Billy Robert O'Neal III Opened: 2016-05-24 Last modified: 2017-07-17

Priority: Not Prioritized

View all other issues in [allocator.adaptor.members].

View all issues with NAD status.

Discussion:

scoped_allocator_adaptor is specified to use forward when what it is really doing is moving elements. It should use move.

Previous resolution [SUPERSEDED]:

This wording is relative to N4582.

  1. Edit 20.5.4 [allocator.adaptor.members] p15 as indicated:

    template <class T1, class T2, class U, class V>
      void construct(pair<T1, T2>* p, pair<U, V>&& x);
    

    Effects: Equivalent to this->construct(p, piecewise_construct, forward_as_tuple(std::forwardmove<U>(x.first)), forward_as_tuple(std::forwardmove<V>(x.second))).

Proposed resolution:

Withdrawn by the submitter, since the prerequisites were incorrect.


2761(i). directory_entry comparisons are members

Section: 31.12.10.4 [fs.dir.entry.obs] Status: NAD Submitter: Ville Voutilainen Opened: 2016-08-03 Last modified: 2017-03-19

Priority: 2

View all other issues in [fs.dir.entry.obs].

View all issues with NAD status.

Discussion:

The comparison operators for directory_entry are member functions (31.12.10.4 [fs.dir.entry.obs]).

That means that a directory_entry can be compared with a path, but a path cannot be compared with a directory_entry, because the left-hand side is not subject to conversions. This seems unfortunate.

Note that the comparisons for path are non-members (31.12.6.8 [fs.path.nonmember]).

The meta-p/r is to turn the directory_entry comparisons into non-member functions.

[2016-08 Chicago]

Wed PM: Priority 2

[2016-10-05 Ville provides concrete wording]

[2016-10 Telecon]

This (comparing a path with a directory_entry) works today w/o any changes. Closing as NAD.

Proposed resolution:

This wording is relative to N4606.

  1. In [fs.class.directory_entry] synopsis, edit as follows:

    namespace std::filesystem {
    class directory_entry {
    public:
      […]
      bool operator< (const directory_entry& rhs) const noexcept;
      bool operator==(const directory_entry& rhs) const noexcept;
      bool operator!=(const directory_entry& rhs) const noexcept;
      bool operator<=(const directory_entry& rhs) const noexcept;
      bool operator> (const directory_entry& rhs) const noexcept;
      bool operator>=(const directory_entry& rhs) const noexcept;
    
    private:
      path pathobject; // exposition only
    };
    
    bool operator< (const directory_entry& lhs, const directory_entry& rhs) noexcept;
    bool operator==(const directory_entry& lhs, const directory_entry& rhs) noexcept;
    bool operator!=(const directory_entry& lhs, const directory_entry& rhs) noexcept;
    bool operator<=(const directory_entry& lhs, const directory_entry& rhs) noexcept;
    bool operator> (const directory_entry& lhs, const directory_entry& rhs) noexcept;
    bool operator>=(const directory_entry& lhs, const directory_entry& rhs) noexcept;
    
    }
    
  2. In 31.12.10.4 [fs.dir.entry.obs]/6-11, edit as follows:

    bool operator==(const directory_entry& lhs, const directory_entry& rhs) const noexcept;
    

    -6- Returns: pathobjectlhs.path() == rhs.pathobjectpath().

    bool operator!=(const directory_entry& lhs, const directory_entry& rhs) const noexcept;
    

    -7- Returns:: pathobjectlhs.path() != rhs.pathobjectpath().

    bool operator< (const directory_entry& lhs, const directory_entry& rhs) const noexcept;
    

    -8- Returns:: pathobjectlhs.path() < rhs.pathobjectpath().

    bool operator<<=(const directory_entry& lhs, const directory_entry& rhs) const noexcept;
    

    -9- Returns:: pathobjectlhs.path() <= rhs.pathobjectpath().

    bool operator> (const directory_entry& lhs, const directory_entry& rhs) const noexcept;
    

    -10- Returns:: pathobjectlhs.path() > rhs.pathobjectpath().

    bool operator>=(const directory_entry& lhs, const directory_entry& rhs) const noexcept;
    

    -11- Returns:: pathobjectlhs.path() >= rhs.pathobjectpath().


2764(i). Are <cstddint> macros optional?

Section: 17.4.1 [cstdint.syn] Status: Dup Submitter: Thomas Koeppe Opened: 2016-08-10 Last modified: 2017-03-21

Priority: 3

View all other issues in [cstdint.syn].

View all issues with Dup status.

Discussion:

Are the macros INT[8, 16, 32, 64]_MAX etc. optional?

The entire <cstddint> header is specified to have all types and macros "defined the same as in C". But C is also unclear about this: the fixed-width types like int32_t are optional in C and in C++. The corresponding macro INT32_MAX is defined in terms of an expression of the same type as the "corresponding type converted according to the integral promotions". But if the "corresponding type" does not exist, then surely the macro too cannot exist? It seems that the macros should also be optional.

Suggested resolution: See e.g. here, or equivalent wording to the effect that the macros INT*_MAX etc are defined if and only if the corresponding integer type is defined.

(Note that the types intptr_t and uintptr_t are also optional.)

[2016-08-11, Richard comments]

C allows other values for N in addition to 8, 16, 32, 64, whereas it appears that C++ does not. Is the difference intentional?

[2016-09-09 Issues Resolution Telecon]

We need to answer Richard's question before making this ready

[2017-03-04, Kona]

Close as a duplicate of 2820 per Thomas' request.

Proposed resolution:


2772(i). Inconsistency in the insert(node) interface

Section: 24.2.7 [associative.reqmts], 24.2.8 [unord.req], 24.4.2 [associative.map.syn], 24.5.2 [unord.map.syn] Status: NAD Submitter: Tomasz Kamiński Opened: 2016-09-06 Last modified: 2018-01-28

Priority: 2

View other active issues in [associative.reqmts].

View all other issues in [associative.reqmts].

View all issues with NAD status.

Discussion:

In C++17 the interface of the unique map was extended to include following function:

pair<iterator, bool> try_emplace(const key_type& k, Args&&... args); //and move version
iterator try_emplace(const_iterator hint, const key_type& k, Args&&... args); //and move version
iterator insert(const_iterator hint, node_type&& nh)
insert_return_type insert(node_type&& nh);

All of the functions share a common property, that they are performing basically no-operation in case when element with given key (part of node) is already stored in the map. However there is major difference in their interface. The first three functions:

pair<iterator, bool> try_emplace(key_type&& k, Args&&... args); //and copy version
iterator try_emplace(const_iterator hint, key_type&& k, Args&&... args); //and copy version
iterator insert(const_iterator hint, node_type&& nh)

are guaranteeing that the value of the arguments (k, nh, args...) will not be changed if the map already contains a key with given value, so the programmer is free to reuse it for their own purpose.

However, the interface of the fourth function is a bit different:

insert_return_type insert(node_type&& nh);

The insert_return_type is an unspecified type that contains:

bool inserted;
X::iterator position;
X::node_type node;

As we can see, the insert function is returning a node. This difference is actually misleading, as the programmer may start to wonder, why the function returns a node handle, instead of being guaranteed that the argument will not be modified (as other functions do). Most reasonable explanation is that, this function actually return a handle to a different node, that one passed as the argument, i.e. this function replaces an existing node with the nh argument and returns the handle to the old node. However, this function actually has the same semantics as the other insert function and returns a node that was passed as argument.

In addition, this design makes the interface of the insert function for the map inconsistent. Value inserting functions are returning pair<iterator, bool> while node inserting function is returning an unspecified type with guaranteed set of members.

The only potential benefit of this signature is that it could potentially allow programmer to use decomposition declaration, so instead of:

auto nh = node_provider();
if (map.insert(std::move(nh)).second)
  handle_node_in_other_way(std::move(nh));

The user would be able to write:

if (auto [it, ins, nh] = map.insert(node_provider); ins)
   handle_node_in_other_way(std::move(nh));

However, the insert_return_type is not currently required to work with decomposition declaration, so this is only "potential" benefit that could be added in future.

Furthermore, this change is preventing a user to use structured binding with combination with insert in generic code:

template<typename UniqMap, typename Elem>
void log_duplicate_insertion(UniqMap& map, Elem&& elem)
{
  if (auto [it, ins] = map.insert(std::forward<Elem>(elem)); !ins)
    std::cout << "attempt to insert duplicate for " << *it;
}

Currently, log_duplicate_insertion will not work with node_handle_type.

So, I am proposing to change the interface of the insert(node_handle) function for associative containers with unique keys, to be consistent with the other insert operation and try_emplace function. I.e. change the signature to:

std::pair<iterator, bool> insert(node_type&& nh);

and provide the guarantee that nh will be unchanged if an element was not inserted.

[2016-09-06, Howard comments]

This is related to LWG 839.

[2018-1-26 issues processing telecon]

Status to 'NAD'; this is "specified as designed". This was considered during the design, but not accepted.

Proposed resolution:


2775(i). reverse_iterator is does not compile for fancy pointers

Section: 25.5.1.6 [reverse.iter.elem] Status: Dup Submitter: Billy Robert O'Neal III Opened: 2016-09-23 Last modified: 2021-06-06

Priority: Not Prioritized

View all other issues in [reverse.iter.elem].

View all issues with Dup status.

Duplicate of: 1052

Discussion:

std::reverse_iterator<Iterator>::operator->() is specified to call addressof(operator*()), but it returns iterator_traits<Iterator>::pointer, which may be a fancy pointer type from which no conversion from T* is possible. It should instead delegate to the underlying iterator's operator->().

[Issues processing Telecon 2016-10-7]

Set status to Duplicate

Proposed resolution:

This wording is relative to N4606.

  1. Modify [reverse.iter.opref] as indicated:

    constexpr pointer operator->() const;
    

    -1- Returns: addressof(operator*()).Effects: If Iterator is a pointer type, as if by:

    Iterator tmp = current;
    return --tmp;
    

    Otherwise, as if by:

    Iterator tmp = current;
    --tmp;
    return tmp.operator->();
    

2822(i). Resolution for LWG 2742 introduces ambiguities

Section: 23.4.3.3 [string.cons] Status: NAD Submitter: Jonathan Wakely Opened: 2016-11-15 Last modified: 2017-01-30

Priority: Not Prioritized

View all other issues in [string.cons].

View all issues with NAD status.

Discussion:

The new constructor added for LWG 2742 causes an ambiguity when attempting to construct from a (non-const) char*

char s[] = "whoops"; 
std::string str(s, 2, 4);

This is the same problem discussed in 2758, and the fix is the same: disable that new constructor if T is convertible to const_pointer, so that the old constructor is used instead.

Previous resolution [SUPERSEDED]:

This wording is relative to N4606.

  1. Modify the new basic_string(const T& t, size_type pos, size_type n, const Allocator& a = Allocator()); constructor in 23.4.3.3 [string.cons] as shown:

    template<class T>
    basic_string(const T& t, size_type pos, size_type n, const Allocator& a = Allocator());
    

    -?- Effects: Creates a variable, sv, as if by basic_string_view<charT, traits> sv = t; and then behaves the same as:

    basic_string(sv.substr(pos, n), a)
    

    -?- Remarks: This constructor shall not participate in overload resolution unless is_convertible_v<const T&, basic_string_view<charT, traits>> is true and is_convertible_v<const T&, const charT*> is false.

[2016-12-21, Jonathan comments and reommends NAD]

Tim Song pointed out that this constructor isn't ambiguous, and I can no longer reproduce the errors I was getting, so I think this issue should be resolved NAD. The code snippet in the issue discussion compiles OK without the proposed change.

[2017-01-27 Telecon]

NAD

Proposed resolution:


2828(i). Clarify <cstdalign> (following adoption of P0063r3)

Section: 21.3.5.4 [meta.unary.prop] Status: NAD Editorial Submitter: Thomas Koeppe Opened: 2016-11-20 Last modified: 2020-09-06

Priority: Not Prioritized

View other active issues in [meta.unary.prop].

View all other issues in [meta.unary.prop].

View all issues with NAD Editorial status.

Discussion:

In Oulu, we adopted P0063r3 to base C++ on C11. C11 is the first version of C that defines a header <stdalign.h>. However, C++11 (!) already seems to have defined headers <cstdalign> and <stdalign.h>, even though C99 does not have <stdalign.h>.

This appears to be the consequence of the partial application of N3093 from 2010, which proposes both to add <cstdalign> and to rebase C++ onto C11. However, the latter part didn't happen, leaving the header dangling without a sensible specification.

I suggest two actions:

  1. 1) Recognize that C++11 and C++14 are defective in mentioning <cstdalign> and <stdalign.h>.
  2. 2a) Consequently, P0063r3 is adding <cstdalign> as a new feature; I would like to revert that decision and not add this header. Instead, P0063r3 should only add <stdalign.h> (which should be done in Annex D) for compatibility with C.
  3. 2b) If we do not want to un-add <cstdalign>, we should at least put the specification of <cstdalign> into Annex D.

Moreover, I would like to ask LWG for editorial dispensation to move the specifications of <ccomplex>, <ctgmath> and <cstdbool> wholly into Annex D, too, since these are now deprecated features that are not otherwise referred to or needed by the main text. An example change is available on github, which incorporates solution 2b. Wording for solution 2a can be produced on request.

[2017-01-27 Telecon]

NAD Editorial; Thomas will make the editorial changes (option 2B)

Proposed resolution:


2852(i). Specifications of operator== for std::basic_strings and std::basic_string_views are difficult to conform to

Section: 23.4.4.2 [string.cmp], 23.4.3.8.4 [string.compare] Status: NAD Submitter: Ahti Leppänen Opened: 2017-01-09 Last modified: 2021-06-06

Priority: 2

View all issues with NAD status.

Discussion:

Currently (N4618, 2016-11-28) the specification of operator== for std::basic_string and std::basic_string_view objects is clearly defined, but when interpreted as written, it may lead to comparison of strings of different sizes being a 𝒪(n) operation instead of a simple size check. Actual implementations in standard libraries vary so that in practice the programmers can't rely neither on having the literal version of the standard specification nor reasonable performance characteristics.

The definition for basic_string operator== in N4618 is as follows:

[string.operator==]

bool operator==(const basic_string<charT, traits, Allocator>& lhs,
                const basic_string<charT, traits, Allocator>& rhs) noexcept;

-1- Returns: lhs.compare(rhs) == 0.

23.4.3.8.4 [string.compare]

int compare(const basic_string& str) const noexcept;

-6- Effects: Equivalent to: return compare(basic_string_view<charT, traits>(str));

23.4.3.8.4 [string.compare]

int compare(basic_string_view<charT, traits> sv) const noexcept;

-1- Effects: Determines the effective length rlen of the strings to compare as the smaller of size() and sv.size(). The function then compares the two strings by calling traits::compare(data(), sv.data(), rlen).

-2- Returns: The nonzero result if the result of the comparison is nonzero. Otherwise, returns a value as indicated in Table 63.

Table 63 — compare() results
Condition Return Value
size() < sv.size() < 0
size() == sv.size() 0
size() > sv.size() > 0

From these it seems that compare() of strings of different sizes can't return zero and operator== will return false. However some implementations do not seem to call traits::compare() for basic_strings of different sizes even when the traits and it's compare() are user-defined. And those that call, make the operator== a worst case 𝒪(n) operation even for strings of different sizes.

This defect report does not propose a wording, but on a general level the wording should allow standard library implementers to write a standard conforming operator== for basic_string and basic_string_view (others?) in such a way that it's performance characteristics are reasonable and the programmers can rely on having a consistent behaviour across implementations. Perhaps the key issue here is that operator== is defined through compare() == 0: while it returns the intended result, for some inputs it does computations that are not needed by operator==. There are also related specifications that may need to be revised, for example operator!= for basic_string_views is defined in 23.3.4 [string.view.comparison] as

Returns: lhs.compare(rhs) != 0

[2017-01-26, Jonathan Wakely comments and provides proposed resolution]

As mentioned above, some implementations do not make a call to Traits::compare if the string lengths are not equal, even though in general this is an observable side effect. Some implementations only perform that optimisation for std::string and std::wstring, where we know that calls to std::char_traits<char>::compare and std::char_traits<wchar_t>::compare are not observable.

My reading is that the Returns: element describes the value that must be returned, not the precise steps that must be taken to calculate that value. If we intended to specify the precise steps that must be taken then we could say that using "Effects: Equivalent to […]", but we don't do that.

I would prefer this issue to be closed NAD with the rationale that my reading is correct and comparing the lengths to avoid calling Traits::compare is already permitted. But if my reading is wrong we need to permit this obvious optimisation.

[2017-01-27 Telecon]

Priority 2

[2017-02-04, Ahti Leppänen comments and recommends NAD]

While there seems to be varying interpretations of the standards wording, given the comments in this defect report and definitions in [structure.specification] (N4618):

Effects: the actions performed by the function

Returns: a description of the value(s) returned by the function

I fail to see that the specification of operator== "Returns: lhs.compare(rhs) == 0" would require call to compare() and no longer consider the report valid.

[2016-07, Toronto Saturday afternoon issues processing]

Status to NAD; we accept Jonathan's reasoning. Note that several implementations do this today.

Proposed resolution:

This wording is relative to N4618.

  1. Preferred: NAD

  2. Alternative:

    1. Modify [string.operator==] p1 as shown:

      template<class charT, class traits, class Allocator>
        bool operator==(const basic_string<charT, traits, Allocator>& lhs,
                        const basic_string<charT, traits, Allocator>& rhs) noexcept;
      

      -1- Returns: lhs.size() == rhs.size() && lhs.compare(rhs) == 0.

    2. Modify 23.3.4 [string.view.comparison] as shown:

      [Example: A sample conforming implementation for operator== would be:

      template<class T> using __identity = decay_t<T>;
      template<class charT, class traits>
        constexpr bool operator==(basic_string_view<charT, traits> lhs,
                                  basic_string_view<charT, traits> rhs) noexcept {
          return lhs.size() == rhs.size() && lhs.compare(rhs) == 0;
        }
      template<class charT, class traits>
        constexpr bool operator==(basic_string_view<charT, traits> lhs,
                                  __identity<basic_string_view<charT, traits>> rhs) noexcept {
          return lhs.size() == rhs.size() && lhs.compare(rhs) == 0;
        }
      template<class charT, class traits>
        constexpr bool operator==(__identity<basic_string_view<charT, traits>> lhs,
                                  basic_string_view<charT, traits> rhs) noexcept {
          return lhs.size() == rhs.size() && lhs.compare(rhs) == 0;
        }
      

      end example]

      template<class charT, class traits>
        constexpr bool operator==(basic_string_view<charT, traits> lhs,
                                  basic_string_view<charT, traits> rhs) noexcept;
      

      -2- Returns: lhs.size() == rhs.size() && lhs.compare(rhs) == 0.

      template<class charT, class traits>
        constexpr bool operator!=(basic_string_view<charT, traits> lhs,
                                  basic_string_view<charT, traits> rhs) noexcept;
      

      -3- Returns: lhs.size() != rhs.size() || lhs.compare(rhs) != 0.


2854(i). wstring_convert provides no indication of incomplete input or output

Section: D.27.2 [depr.conversions.string] Status: NAD Submitter: PowerGamer Opened: 2017-01-08 Last modified: 2017-06-05

Priority: 3

View other active issues in [depr.conversions.string].

View all other issues in [depr.conversions.string].

View all issues with NAD status.

Discussion:

Example:

// Input UTF-16 string is incomplete - only first half of
// UTF-16 surrogate pair L"\xD843\xDEF9":
wchar_t in_utf16[] = L"\xD843";

std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>> cvt;
auto out_utf8 = cvt.to_bytes(in_utf16); // No error.

There is no indication that input was incomplete (the value returned by cvt.state() is not documented and so cannot be examined by user for that purpose). As such the user will not know that more input data should be provided in additional call to cvt.to_bytes().

The output can be incomplete too: MSVC2017 implementation (which as far as I can tell is standard conforming) produces "\xF0" in out_utf8. Again, no indication of incomplete output produced is provided by std::wstring_convert.

IMO it makes std::wstring_convert in its current state completely useless (it cannot be relied upon to either produce complete and valid UTF sequence or throw an error in all situations).

Imagine a file has UTF16 encoded text. You want to read all the data from a file at once and convert it into UTF8 using std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>>.

Now, if a file contains completely invalid UTF16 (for example, forbidden or incorrectly encoded Unicode code points) you will get an exception from std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>>.

But if a file contains incomplete (but in all other regards valid) UTF16 (for ex. file ends with only the first half of a valid surrogate pair) you will neither get an error exception from std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>> nor any indication that the input provided to std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>> was incomplete.

[2017-01-27 Telecon]

Priority 3; send to LEWG

[2017-02 in Kona, LEWG recommends NAD]

[2017-06-02 Issues Telecon]

This facility has a number of known problems, including poor error handling. The feature has been deprecated, and the plan is to replace it with better facilities with a better API.

Resolve as NAD

Proposed resolution:


2860(i). launder and base class subobjects

Section: 17.6.5 [ptr.launder] Status: NAD Submitter: Hubert Tong Opened: 2017-01-31 Last modified: 2020-09-06

Priority: 2

View all other issues in [ptr.launder].

View all issues with NAD status.

Discussion:

There is an apparent oversight in the wording for launder that allows it to return base class subobjects which differ in their polymorphic behaviour between calls to launder.

This can be fixed by restricting launder from returning pointers to base class subobjects:

  1. always, or

  2. only for polymorphic class types.

[2017-03-04, Kona]

Set priority to 2. This was discussed in EWG (via a paper). Assign this (and 2859) to Core.

[2017-08-14, CWG telecon note]

Core recommends NAD after discussion in EWG.

Hubert Tong summarizes the outcome of that discussion as follows:

launder does not provide positive confirmation of the dynamic type of the object; it is intended that launder can be used to induce a devirtualization barrier even when the static type of the (sub)object to which the returned pointer refers is apparently consistent with prior accesses related to the source pointer.

[2020-02 Status to NAD on Thursday night in Prague.]

Proposed resolution:

This wording is relative to N4618.

  1. Option 1:

    1. Modify 17.6.5 [ptr.launder] as indicated:

      template <class T> constexpr T* launder(T* p) noexcept;
      

      -1- Requires: p represents the address A of a byte in memory. An object X that is within its lifetime (6.7.3 [basic.life]) and whose type is similar (7.3.6 [conv.qual]) to T is located at the address A. X shall either be a most derived object, or pointer-interconvertible with a most derived object that is within its lifetime. All bytes of storage that would be reachable through the result are reachable through p (see below).

  2. Option 2:

    1. Modify 17.6.5 [ptr.launder] as indicated:

      template <class T> constexpr T* launder(T* p) noexcept;
      

      -1- Requires: p represents the address A of a byte in memory. An object X that is within its lifetime (6.7.3 [basic.life]) and whose type is similar (7.3.6 [conv.qual]) to T is located at the address A. If T is a polymorphic class type, then X shall be a most derived object. All bytes of storage that would be reachable through the result are reachable through p (see below).


2865(i). Resolve all open Library issues for C++17

Section: 16 [library] Status: NAD Submitter: Finland Opened: 2017-02-03 Last modified: 2017-03-12

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Addresses FI 2

All open Library Issues should be resolved.

Proposed change:

As LWG sees fit.

[2017-03-12, post-Kona]

Closing as NAD; this issue was opened by an over-zealous LWG chair.

Proposed resolution:


2871(i). User specializations of type traits should be ill-formed

Section: 21.3.3 [meta.type.synop] Status: NAD Submitter: United States Opened: 2017-02-03 Last modified: 2020-09-06

Priority: Not Prioritized

View other active issues in [meta.type.synop].

View all other issues in [meta.type.synop].

View all issues with NAD status.

Discussion:

Addresses US 10

A user specialization of any type trait should produce an ill-formed program, not merely one whose behavior is unspecified. See also the related comment re 22.12.3 [execpol.type].

Proposed change:

Reword the paragraph as follows:

Unless otherwise specified, a program that adds specializations for any of the templates defined in this subclause is ill-formed; no diagnostic required.

See also 2909

[2017-07 Toronto Thurs Issue Prioritization]

NAD; what the standard says is that this is "undefined", not "unspecified".

Proposed resolution:

This wording is relative to N4618.

  1. Modify 21.3.3 [meta.type.synop] as indicated:

    -1- The behavior of a program that adds specializations for any of the templates defined in this subclause is undefined unless otherwise specifiedUnless otherwise specified, a program that adds specializations for any of the templates defined in this subclause is ill-formed; no diagnostic required.


2886(i). Keep the empty() functions in any

Section: 22.7.4.5 [any.observers] Status: NAD Submitter: Finland Opened: 2017-02-03 Last modified: 2020-09-06

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses FI 7

The proposal p0032 has multiple problems: 1) it turns member function .empty() into .has_value(), negating the logic. Refactoring e.g. existing uses of std::experimental::any to use std::any thus involve non-trivial refactorings that are error- prone and can't be done via simple search-and-replace if there are containers in the same source files for which .empty() is used (based on the implementation experience of making the change in libstdc++ and refactoring the testsuite). Whilst any is not a container, the library is failing to go towards a direction where there would be a generic way to query for emptiness. 2) The use of function references for tag types makes the interface hard to use. The tag types do not have value semantics like every other tag type has, the tag types are hard to construct, and present surprises for certain kinds of overload sets. Furthermore, any attempts to decay the tag types produces a really surprising effect — as opposed to what the other tag types do, which is that the result of decaying them is the tag type itself, decaying these new tag types results in a pointer to function.

Proposed change: Keep the .empty() functions (and introduce them to all the types that are supposed to have a homogeneous interface), and make the tag types be regular tag types that are not references to functions.

[2017-02-03, Marshall notes]

P0504R0 (adopted in Issaquah) removed use of function references for tag types

[ 2017-06-26 Moved to Tentatively NAD after 5 positive votes on c++std-lib. ]

Proposed resolution:


2891(i). Relax library requirements on volatile types

Section: 16.4.6 [conforming] Status: NAD Submitter: Great Britain Opened: 2017-02-03 Last modified: 2017-07-15

Priority: Not Prioritized

View all other issues in [conforming].

View all issues with NAD status.

Discussion:

Addresses GB 35

Most implementations have poor testing and support for instantiating standard library templates with volatile-qualified types. We should grant a library-freedom to conforming implementations so that support for volatile (and const volatile) qualified types in standard library templates is not required unless explicitly specified — and mandate such support for all templates in the <type_traits> header. Additional support is already specified in most places we would be interested (e.g., tuple API). We may want to additionally guarantee support through forwarding references.

Proposed change: add a new 17.6.5.x Volatile Qualified Types [res.on.volatile.type] describing the intended level of support for volatile qualifiers.

[2017-07 Toronto Thurs Issue Prioritization]

This is not a defect, but a feature request (and a large one). Papers welcome, but closing as NAD

Proposed resolution:


2892(i). Relax the prohibition on libraries adding constexpr

Section: 16.4.6.7 [constexpr.functions] Status: NAD Submitter: Great Britain Opened: 2017-02-03 Last modified: 2020-09-06

Priority: 1

View all other issues in [constexpr.functions].

View all issues with NAD status.

Discussion:

Addresses GB 38

Relax the prohibition on libraries adding constexpr; this was a constraint requested by library implementers when constexpr was new, and those same implementers now feel unduly constrained.

Proposed change: Rewrite the whole sub-clause to support libraries adding constexpr in a compatible manner, much like the freedom to add a noexcept specification.

[2017-02-20, Marshall adds wording]

The simplest change would be to strike the sentence "An implementation shall not […] explicitly required". However, people seem to want a definite permission here, so I have provided one.

[2017-07 Toronto Thurs Issue Prioritization]

Priority 1

[2018-06 Rapperswil Thursday issues processing]

This is a feature request; not a defect. We've had two cross-group meetings over this, and there is no consensus for changing this. Closing as NAD.

Proposed resolution:

This wording is relative to N4640.

  1. Modify 16.4.6.7 [constexpr.functions] as indicated:

    This International Standard explicitly requires that certain standard library functions are constexpr (7.1.5). An implementation may declare additional standard library function signatures as constexprAn implementation shall not declare any standard library function signature as constexpr except for those where it is explicitly required. Within any header that provides any non-defining declarations of constexpr functions or constructors an implementation shall provide corresponding definitions.


2893(i). Parsing Hexadecimally in P0067R4

Section: 22 [utilities] Status: NAD Submitter: Great Britain Opened: 2017-02-03 Last modified: 2020-09-06

Priority: Not Prioritized

View all other issues in [utilities].

View all issues with NAD status.

Discussion:

Addresses GB 45

If P0067R4 is applied consider how to parse hexadecimally:

to_chars(beg, end, 42, 16); 16 for hex

to_chars(beg, end, 4.2, true); true means hex

to_chars(beg, end, 4.2, chars_format::hex);

to_chars(beg, end, 4.2, chars_format::hex, 2);

That is: We have 3 different formats to specify hex depending on value types and whether to use precision. Which application programmer should remember this?

May be even worse (I am not sure):

to_chars(beg, end, 4.2, 16);

would silently convert 4.2 to 4 and

to_chars(beg, end, 4, chars_format::hex);

would silently convert 4 to 4.000000.

Proposed change: The various options should be harmonized, possibly by use of an extended enum approach, having the values: dec, hex, scientific, fixed, general with dec (new!) as default for integral values and general for floats

[2017-02-23, Jens Maurer comments]

This issue is NAD; the facility was redesigned with these considerations in mind before it went into the working draft.

[ 2017-06-26 Moved to Tentatively NAD after 5 positive votes on c++std-lib. ]

Proposed resolution:


2896(i). The contents of <codecvt> are underspecified

Section: D.26 [depr.locale.stdcvt] Status: Dup Submitter: Great Britain Opened: 2017-02-03 Last modified: 2017-04-22

Priority: Not Prioritized

View all other issues in [depr.locale.stdcvt].

View all issues with Dup status.

Discussion:

Addresses GB 57

The contents of <codecvt> are underspecified, and will take a reasonable amount of work to identify and correct all of the issues. There appears to be a general feeling that this is not the best way to address unicode transcoding in the first place, and this library component should be retired to Annex D, along side <strstream>, until a suitable replacement is standardized

Proposed change: Deprecate and move the whole of clause [locale.stdcvt] to Annex D.

[2017-02-03, Daniel comments]

This issue unintentionally became added a second time, it should be considered as a pure duplicate of LWG 2869. By a wit of fortuity, both issues can also be transformed into each other by transposing the last two digits of the issue numbers.

Proposed resolution:

This issue is a duplicate of LWG 2869.


2898(i). Prefer not to use member typedefs as constructor parameters

Section: 16 [library] Status: NAD Submitter: United States Opened: 2017-02-03 Last modified: 2017-06-02

Priority: Not Prioritized

View other active issues in [library].

View all other issues in [library].

View all issues with NAD status.

Discussion:

Addresses US 106

Review the whole library for constructors using member typedefs to name constructor parameters rather than template type parameters, as this inhibits class template deduction. e.g., the unique_lock explicit constructor taking the mutex_type typedef would be better served naming Mutex directly, to preserve support for deduction.

Proposed change: Review each constructor of each library class template, and revise specification of parameter types as needed.

[2017-03-03, Kona Friday morning]

This is no longer true; core has fixed this

Setting to NAD

Proposed resolution:


2902(i). variant should only support complete types

Section: 22.6.3 [variant.variant] Status: NAD Submitter: United States Opened: 2017-02-03 Last modified: 2020-09-06

Priority: 0

View all other issues in [variant.variant].

View all issues with NAD status.

Discussion:

Addresses US 114

variant needs to know the size of an object in order to compute the size of its internal buffer, so require that any cv-qualified object type in Types... be a complete type.

Proposed change: Add 'complete' in p2

[2017-07-01, Daniel comments]

I recommend to close this issue as NAD: According to 16.4.5.8 [res.on.functions] p2 plus bullet (2.5):

-2- In particular, the effects are undefined in the following cases:

  1. — […]

  2. (2.5) — if an incomplete type (6.9) is used as a template argument when instantiating a template component, unless specifically allowed for that component.

In other words: The standard currently already requires type completeness for all types in Types... of std::variant<Types...>, we therefore should not repeat that again (Note that other templates — such as shared_ptrgrant type incompleteness allowance for that reason).

The criteria for the alternative resolution form Resolved do not apply for this case, because the quoted constraint do exist since C++98 (see 17.4.3.6 [lib.res.on.functions] p2 b5), so the issue was not even an issue at the point of the submission.

Previous resolution [SUPERSEDED]:

This wording is relative to N4618.

  1. Edit 22.6.3 [variant.variant]/2 as indicated:

    All types in Types... shall be (possibly cv-qualified) complete object types that are not arrays.

[2017-07-06 Moved to Tentatively NAD after 5 positive votes on c++std-lib.]

Proposed resolution:

The lack of the explicit constraint in 22.6.3 [variant.variant] is not a defect, because the general requirements in 16.4.5.8 [res.on.functions] p2 already impose the requested type completeness constraints.


2907(i). Semantics for destroying the deleter and the control-block of a shared_ptr are unclear

Section: 20.3.2.2.3 [util.smartptr.shared.dest] Status: NAD Submitter: United States Opened: 2017-02-03 Last modified: 2017-07-15

Priority: Not Prioritized

View all other issues in [util.smartptr.shared.dest].

View all issues with NAD status.

Discussion:

Addresses US 134

The semantics for destroying the deleter and the control-block are unclear. In particular, it is not clear that we guarantee a lack of race conditions destroying the control-block and deleter. Possible race-free implementations might destroy the deleter after running d(p), and before giving up the weak reference held by this shared_ptr; running the destructor for d only when the last weak_ptr is destroyed, potentially at a much later date, but ensuring that d(p) completes before the shared_ptr gives up its weak reference; making a copy of d in the destructor before manipulating the weak count, and then using this copy to run d(p), even while the control-block could be concurrently reclaimed with an expiring weak_ptr in another thread. Note that this may be related to LWG 2751. (Also, see the note in 20.11.2.2.10 p1 [util.smartptr.getdeleter])

Proposed change: Clarify that the shared_ptr weak ownership of the control block is released at the end of the destructor, and not as the destructor begins. Otherwise, the deleter might be destroyed even before the destructor gets to move a copy to call safely.

[2017-07 Toronto Thurs Issue Prioritization]

NAD; Alisdair to provide rationale

Proposed resolution:


2909(i). User specializations of is_execution_policy should be ill-formed

Section: 22.12.3 [execpol.type] Status: NAD Submitter: United States Opened: 2017-02-03 Last modified: 2017-07-16

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Addresses US 14

A user specialization of any is_execution_policy should produce an ill-formed program, not merely one whose behavior is unspecified. See also the related comment re [meta.type.synop] (20.15.2). (US 10)

Proposed change: Reword the paragraph as follows:

Unless otherwise specified, a program that adds specializations for is_execution_policy is ill-formed; no diagnostic required.

See also LWG 2871.

[2017-03-03, Kona Friday AM]

Jonathan: This breaks all known implementations of parallel algorithms, since they use custom execution policies.

Alisdair to check with SG1 to see if that's what they meant.

[2017-07 Toronto Thurs Issue Prioritization]

NAD; what the standard says is that this is "undefined", not "unspecified".

Proposed resolution:


2910(i). Template deduction and integral_constant

Section: 21.3.3 [meta.type.synop] Status: Dup Submitter: United States Opened: 2017-02-03 Last modified: 2020-09-06

Priority: Not Prioritized

View other active issues in [meta.type.synop].

View all other issues in [meta.type.synop].

View all issues with Dup status.

Discussion:

Addresses US 142

An alias template using the new template template auto deduction would make integral_constant slightly easier to use.

[2017-07 Toronto Thurs Issue Prioritization]

Duplicate of 2922.

Proposed resolution:

This wording is relative to N4618.

  1. Add to the synopsis of <type_traits>:

    template <auto N>
    using integer_constant = integral_constant<decltype(N), N>;
    

2916(i). Insert iterators should each have an instantiation guide to initialize from a container

Section: 25.5.2 [insert.iterators] Status: NAD Submitter: United States Opened: 2017-02-03 Last modified: 2020-09-06

Priority: Not Prioritized

View all other issues in [insert.iterators].

View all issues with NAD status.

Discussion:

Addresses US 151

The three insert iterators should each have an instantiation guide to initialize from a container.

Proposed change:

Add to the <iterator> header synopsis:

template <class Container>
back_insert_iterator(Container&) -> back_insert_iterator<Container>;
template <class Container>
front_insert_iterator(Container&) -> back_insert_iterator<Container>;
template <class Container>
insert_iterator(Container&, typename Container::iterator) -> insert_iterator<Container>;

[2017-07 Toronto Thurs Issue Prioritization]

After the core changes in Kona, these guides are no longer necessary; the implicit ones are sufficient.

Proposed resolution:

This wording is relative to N4618.

  1. Add to the synopsis of <iterator>:

    
    template <class Container> back_insert_iterator(Container&) -> back_insert_iterator<Container>;
    template <class Container> front_insert_iterator(Container&) -> back_insert_iterator<Container>;
    template <class Container> insert_iterator(Container&, typename Container::iterator) -> insert_iterator<Container>;
    

2930(i). Are implementations allowed to split non-member functions into several overloads?

Section: 16.4.6.4 [global.functions] Status: NAD Submitter: Morwenn Opened: 2017-02-04 Last modified: 2017-03-14

Priority: Not Prioritized

View all other issues in [global.functions].

View all issues with NAD status.

Discussion:

Section 16.4.6.5 [member.functions] makes it clear that standard library implementers are allowed to "replace a member function with default arguments with two or more member functions with equivalent behavior". However, it is unclear when reading 16.4.6.4 [global.functions] whether implementers are allowed to perform the same transformation for non-member functions. 16.4.6.4 [global.functions] p3 mentions that "an implementation shall not declare a non-member function signature with additional default arguments", but does not explicitly forbid to replace a function with default arguments with more functions with equivalent behavior.

The standard should make it clear whether such a transformation is conforming or not, probably through a note [global.functions].

Link to the Standard Discussion forum.

[2017-03-04, Kona]

We believe that [member.functions]p2 is clear.

Proposed resolution:


2967(i). std::equal on empty ranges

Section: 27.6.13 [alg.equal] Status: NAD Submitter: Gennaro Prota Opened: 2017-05-26 Last modified: 2020-09-06

Priority: Not Prioritized

View all other issues in [alg.equal].

View all issues with NAD status.

Discussion:

The description of the std::equal() algorithm in the standard doesn't make clear what the result of it is on empty ranges:

std::equal(first, first, second) ; // what does this return?

It should IMHO return true (two empty ranges are always equal).

[2017-07 Toronto Monday issue prioritization]

Closing as NAD; the existing wording covers empty sequences

Proposed resolution:

This wording is relative to N4659.

  1. Edit 27.6.13 [alg.equal] as indicated:

    [Drafting note: The current wording presented below uses two times the unusual phrase "[…] return […]" instead of "[…] returns […]". The project editor is kindly asked to consider to replace these unusual wording forms by the usual one. — end drafting note]

    template<class InputIterator1, class InputIterator2>
      bool equal(InputIterator1 first1, InputIterator1 last1,
                 InputIterator2 first2);
    […]
    

    -1- Remarks: If last2 was not given in the argument list, it denotes first2 + (last1 - first1) below.

    -2- Returns: If [first1, last1) and [first2, last2) are both empty, returns true. If last1 - first1 != last2 - first2, return false. Otherwise return true if for every iterator i in the range [first1, last1) the following corresponding conditions hold: *i == *(first2 + (i - first1)), pred(*i, *(first2 + (i - first1))) != false. Otherwise, returns false.

    […]


2971(i). variant should require Destructible types

Section: 22.6.3 [variant.variant] Status: NAD Submitter: Agustín K-ballo Bergé Opened: 2017-05-31 Last modified: 2020-09-06

Priority: Not Prioritized

View all other issues in [variant.variant].

View all issues with NAD status.

Discussion:

The state of a variant is not specified if an exception is thrown during the destruction of the active alternative. Exceptions should not be allowed to escape during destruction of an object used as a variant alternative. Otherwise, it should be specified that the variant object might not hold a value.

[2017-07 Toronto Monday issue prioritization]

NAD; [res.on.functions]/2 already gives us this guarantee.

Proposed resolution:

This wording is relative to N4659.

  1. Edit 22.6.3 [variant.variant] as indicated:

    -2- All types in Types... shall be (possibly cv-qualified) object types that are not arrays, and shall satisfy the requirements of Destructible (Table 27).


2992(i). system_category() and error_code::error_code() should be constexpr

Section: 19.5.3.5 [syserr.errcat.objects], 19.5.4.2 [syserr.errcode.constructors] Status: NAD Submitter: Peter Dimov Opened: 2017-06-27 Last modified: 2017-07-12

Priority: Not Prioritized

View all other issues in [syserr.errcat.objects].

View all issues with NAD status.

Discussion:

The default constructor of error_code should be constexpr to enable constant initialization; as a practical matter, there are reports that it regularly shows up in profiles because clearing error codes is so frequent.

Suggested resolution:

There was an objection that system_category() can't be made constexpr because it needs to "immortalize" the object so that it's not destroyed at process shutdown or module unload, in order for the error_code facility to remain usable. However, the following proof of concept shows how to achieve this and still make the function constexpr:

#include <new>

template<class _Ty>
  union _Immortalizer
  { // constructs _Ty, never destroys
  constexpr _Immortalizer(): __ty()
  {
  }

  ~_Immortalizer() noexcept {}
  _Immortalizer(const _Immortalizer&) = delete;
  _Immortalizer& operator=(const _Immortalizer&) = delete;

  _Ty __ty;
};

struct error_category
{
  virtual ~error_category() = default;
};

struct system_category_impl : public error_category
{
};

[[clang::require_constant_initialization]] static const _Immortalizer<system_category_impl> _System_category;

constexpr error_category const& system_category() noexcept
{
  return _System_category.__ty;
}

struct error_code
{
  int val_;
  const error_category* cat_;

  constexpr error_code() noexcept : val_(0), cat_(&system_category()) {}

  constexpr int value() const noexcept { return val_; }
  constexpr error_category const& category() const noexcept { return *cat_; }
};

constexpr error_code s_code;

static_assert(s_code.value() == 0);
static_assert(&s_code.category() == &system_category());

[2017-07 Toronto Tuesday PM issue prioritization]

NAD; This is a feature request; needs a paper.

Proposed resolution:

This wording is relative to N4659.

  1. Edit 19.5.3.1 [syserr.errcat.overview], class error_category synopsis, as indicated:

    class error_category {
    public:
      constexpr error_category() noexcept;
      virtual ~error_category();
      error_category(const error_category&) = delete;
      error_category& operator=(const error_category&) = delete;
      virtual const char* name() const noexcept = 0;
      virtual error_condition default_error_condition(int ev) const noexcept;
      virtual bool equivalent(int code, const error_condition& condition) const noexcept;
      virtual bool equivalent(const error_code& code, int condition) const noexcept;
      virtual string message(int ev) const = 0;
    
      bool operator==(const error_category& rhs) const noexcept;
      bool operator!=(const error_category& rhs) const noexcept;
      bool operator<(const error_category& rhs) const noexcept;
    };
    
    constexpr const error_category& generic_category() noexcept;
    constexpr const error_category& system_category() noexcept;
    
  2. Edit 19.5.3.5 [syserr.errcat.objects] as indicated:

    constexpr const error_category& generic_category() noexcept;
    

    […]

    constexpr const error_category& system_category() noexcept;
    

    […]

  3. Edit 19.5.4.1 [syserr.errcode.overview], class error_code synopsis, as indicated:

    class error_code {
    public:
      // 19.5.4.2 [syserr.errcode.constructors], constructors
      constexpr error_code() noexcept;
      constexpr error_code(int val, const error_category& cat) noexcept;
      template <class ErrorCodeEnum>
      error_code(ErrorCodeEnum e) noexcept;
      
      // 19.5.4.3 [syserr.errcode.modifiers], modifiers
      constexpr void assign(int val, const error_category& cat) noexcept;
      template <class ErrorCodeEnum>
      error_code& operator=(ErrorCodeEnum e) noexcept;
      constexpr void clear() noexcept;
      
      // 19.5.4.4 [syserr.errcode.observers], observers
      constexpr int value() const noexcept;
      constexpr const error_category& category() const noexcept;
      error_condition default_error_condition() const noexcept;
      string message() const;
      explicit operator bool() const noexcept;
      
    private:
      int val_; // exposition only
      const error_category* cat_; // exposition only
    };
  4. Edit 19.5.4.2 [syserr.errcode.constructors] as indicated:

    constexpr error_code() noexcept;
    

    […]

    constexpr error_code(int val, const error_category& cat) noexcept;
    

    […]

  5. Edit 19.5.4.3 [syserr.errcode.modifiers] as indicated:

    constexpr void assign(int val, const error_category& cat) noexcept;
    

    […]

    constexpr void clear() noexcept;
    

    […]

  6. Edit 19.5.4.4 [syserr.errcode.observers] as indicated:

    constexpr int value() const noexcept;
    

    […]

    constexpr const error_category& category() const noexcept;
    

    […]


3016(i). optional and over-aligned types

Section: 22.5.3 [optional.optional] Status: NAD Submitter: Tim Song Opened: 2017-09-04 Last modified: 2020-09-06

Priority: 3

View all other issues in [optional.optional].

View all issues with NAD status.

Discussion:

LWG issue 2555 added "It is implementation-defined whether over-aligned types are supported (C++14 §3.11)." to the specification of std::experimental::optional in LFTS, however that issue wasn't moved until optional had already been merged to the IS working paper, so it isn't present in the specification of std::optional. Should the same rule be added for std::optional as well?

[2017-11 Albuquerque Wednesday night issues processing]

Priority set to 3; Casey to provide rationale for closing as NAD.

2018-11 Closed as NAD with the adoption of P0899R1

Proposed resolution:

This wording is relative to N4687.

  1. Edit 22.5.3 [optional.optional] p1 as indicated:

    […] The contained value shall be allocated in a region of the optional<T> storage suitably aligned for the type T. It is implementation-defined whether over-aligned types are supported (6.7.6 [basic.align]). When an object of type optional<T> is contextually converted to bool, the conversion returns true if the object contains a value; otherwise the conversion returns false.


3033(i). basic_string move ctor is underspecified

Section: 23.4.3.3 [string.cons] Status: NAD Editorial Submitter: Peter Sommerlad Opened: 2017-11-09 Last modified: 2017-11-29

Priority: Not Prioritized

View all other issues in [string.cons].

View all issues with NAD Editorial status.

Discussion:

The standard has a hole in the specification of basic_string's move constructor that seems not only be editorial. In 23.4.3.3 [string.cons] p2:

basic_string(const basic_string& str);
basic_string(basic_string&& str) noexcept;

Effects: Constructs an object of class basic_string as indicated in Table 56. In the second form, str is left in a valid state with an unspecified value.

refers to Table 56, but that table only speaks about copy-ctor. According to Pablo, move and copy differ in what they do. Move ctor of basic_string needs separate table (entry) to describe what happens.

[28-Nov-2017 After discussion on the ML, closing as NAD Editorial]

Proposed resolution:


3068(i). Forbid assigning an rvalue basic_string to basic_string_view

Section: 23.3.3 [string.view.template] Status: NAD Submitter: Antony Polukhin Opened: 2018-02-19 Last modified: 2022-08-24

Priority: 2

View all other issues in [string.view.template].

View all issues with NAD status.

Discussion:

It is known that we cannot disable move construction of basic_string_view from rvalues of basic_string, because it breaks a valid use case:

string foo();
void bar(string_view );
bar(foo());

Though it is still possible to disable an absolutely wrong case of assigning an rvalue basic_string to basic_string_view:

string_view sw = "Hello";
sw = foo();

Some tests that make sure that other use cases are not affected are available here

[2018-06-18 after reflector discussion]

Priority set to 2; status to LEWG

[2020-05-28; LEWG issue reviewing]

LEWG issue processing voted to reject 3068 as NAD. Status change to Open.

SF F N A SA
15 5 1 0 0

[2022-08-24 Status changed: Open → NAD.]

LWG telecon: close based on LEWG direction. "This makes string_view depend on basic_string." "Request a paper with implementation if someone really wants this."

Proposed resolution:

This wording is relative to N4727.

  1. Change 23.3.3 [string.view.template], class template basic_string_view synopsis, as indicated:

    […]
    // 23.3.3.2 [string.view.cons], construction and assignment
    constexpr basic_string_view() noexcept;
    constexpr basic_string_view(const basic_string_view&) noexcept = default;
    constexpr basic_string_view& operator=(const basic_string_view&) noexcept = default;
    template <class A>
    basic_string_view& operator=(const basic_string<charT, traits, A>&&) = delete;
    constexpr basic_string_view(const charT* str);
    constexpr basic_string_view(const charT* str, size_type len);
    […]
    

3106(i). nothrow should be inline constexpr rather that extern const

Section: 17.6.2 [new.syn] Status: NAD Submitter: Antony Polukhin Opened: 2018-04-25 Last modified: 2020-09-06

Priority: 2

View all issues with NAD status.

Discussion:

std::nothrow in [new.syn] is declared as 'extern const nothrow_t nothrow;'. Unfortunately this declaration requires linking with Standard Library (which does not work well on GCC with -nostdlib flag). Moreover the declaration differs from other Standard Library tags that are 'inline constexpr'.

[2018-06-18 after reflector discussion]

Priority set to 2

[2018-08-23 Batavia Issues processing]

Status to Tentatively NAD - if we had a status 'Wont Fix' that would be correct.

This is an ABI break for most implementations, and people can define their own variable of type std::nothrow_t if they want a constexpr one.

Proposed resolution:

This wording is relative to N4741.

  1. Edit 17.6.2 [new.syn] as indicated:

    struct nothrow_t { explicit nothrow_t() = default; };
    extern constinline constexpr nothrow_t nothrow;
    

3138(i). There is no such thing as assertion-level

Section: 99 [support.contract.cviol] Status: NAD Submitter: Tim Song Opened: 2018-07-10 Last modified: 2019-08-05

Priority: 2

View all other issues in [support.contract.cviol].

View all issues with NAD status.

Discussion:

contract_violation has an assertion_level member function that is specified to return text "describing the assertion-level of the violated contract". There is, however, no assertion-level in [dcl.attr.contract], only contract-level.

By itself, this can probably be handled editorially as a typo, but we may want to reconsider the name of the function, which currently mirrors the (incorrect) name of the nonterminal.

[2018-11 Reflector prioritization]

Set Priority to 2

[Post Cologne Mailing]

Contracts were removed from the C++20 draft in Cologne; closing as NAD.

Proposed resolution:


3139(i). contract_violation's special member functions

Section: 99 [support.contract.cviol] Status: NAD Submitter: Tim Song Opened: 2018-07-10 Last modified: 2019-08-05

Priority: 1

View all other issues in [support.contract.cviol].

View all issues with NAD status.

Discussion:

contract_violation is currently defined as:

class contract_violation {
public:
  uint_least32_t line_number() const noexcept;
  string_view file_name() const noexcept;
  string_view function_name() const noexcept;
  string_view comment() const noexcept;
  string_view assertion_level() const noexcept;
};

which implies a full set of defaulted special member functions with entirely unclear semantics, and can arguably be read to imply that it is an aggregate.

It seems unlikely that we meant for users to default construct or copy contract_violation objects, but if we did want to do that, we need to at least specify the behavior of a default constructed contract_violation object and how copying and moving work with respect to the lifetime of the strings referenced by the string_views returned by its member functions.

[2018-11 Reflector prioritization]

Set Priority to 1

[2019 Cologne Wednesday night]

= delete for copy/move assignment; Status to 'Open'; Daniel to provide wording

[Post Cologne Mailing]

Contracts were removed from the C++20 draft in Cologne; closing as NAD.

Proposed resolution:


3163(i). [networking.ts] Buffer sequence iterator equivalency

Section: 16.2 [networking.ts::buffer.reqmts] Status: NAD Submitter: Vinnie Falco Opened: 2018-10-07 Last modified: 2020-09-06

Priority: Not Prioritized

View all other issues in [networking.ts::buffer.reqmts].

View all issues with NAD status.

Discussion:

Addresses: networking.ts

Buffer sequence iterators in [networking.ts] differ from the current standard iterator requirements in that the value returned upon dereferencing is not required to be a reference type. This change is vital for allowing authors to create useful, lazily-evaluated views which meet the requirements of buffer sequences. Unfortunately the updated language is incompatible with the standard iterator requirements for equality testing, because the addresses of value types returned by buffer sequence iterators may not be stable. The change proposed in this defect report defines a new metric for buffer sequence iterator equality.

11-2018 Status to NAD after discussion on the reflector.

Proposed resolution:

This wording is relative to the N4734.

  1. Modify 16.2.1 [networking.ts::buffer.reqmts.mutablebuffersequence], Table 12 "MutableBufferSequence requirements", as indicated:

    Table 12 — MutableBufferSequence requirements
    expression return type assertion/note
    pre/post-condition
    net::buffer_sequence_begin(x)
    net::buffer_sequence_end(x)
    An iterator type whose reference type
    is convertible to mutable_buffer and which
    satisfies all the requirements for bidirectional
    iterators (C++ 2014 [bidirectional.iterators]) except that:
    1. there is no requirement that operator-> is provided, and
    2. there is no requirement that reference
      be a reference type., and
    3. for iterators a and b
      there is no requirement that a == b if and
      only if *a and *b are bound to the
      same object, and
    4. a == b if and only if
      distance(buffer_sequence_begin(x), a) == distance(buffer_sequence_begin(x), b)
    […]
  2. Modify 16.2.2 [networking.ts::buffer.reqmts.constbuffersequence], Table 13 "ConstBufferSequence requirements", as indicated:

    Table 13 — ConstBufferSequence requirements
    expression return type assertion/note
    pre/post-condition
    net::buffer_sequence_begin(x)
    net::buffer_sequence_end(x)
    An iterator type whose reference type is convertible to
    const_buffer and which satisfies all the requirements for
    bidirectional iterators (C++ 2014 [bidirectional.iterators]) except that:
    1. there is no requirement that operator-> is provided, and
    2. there is no requirement that reference be a reference
      type., and
    3. for iterators a and b there is no
      requirement that a == b if and only if *a and
      *b are bound to the same object, and
    4. a == b if and only if
      distance(buffer_sequence_begin(x), a) == distance(buffer_sequence_begin(x), b)
    […]

3164(i). Unhelpful "shall not participate" constraints for unique_ptr with reference deleter

Section: 20.3.1.3.2 [unique.ptr.single.ctor] Status: NAD Submitter: Jonathan Wakely Opened: 2018-10-09 Last modified: 2018-11-27

Priority: Not Prioritized

View all other issues in [unique.ptr.single.ctor].

View all issues with NAD status.

Discussion:

[unique.ptr.single.ctor] defines these constructors:

unique_ptr(pointer p, const D& d) noexcept;
unique_ptr(pointer p, remove_reference_t<D>&& d) noexcept;

and p13 says:

Remarks: If D is a reference type, the second constructor is defined as deleted. These constructors shall not participate in overload resolution unless is_constructible_v<D, decltype(d)> is true.

The first sentence in the Remarks has no effect. If D is a reference type A&, then for the second constructor the condition is is_constructible<A&, A&&> which is false. So the second constructor never participates in overload resolution for reference deleters, and so it's irrelevant whether it's deleted or not.

We should either strike that sentence, or adjust the is_constructible constraint so that the deleted constructor participates in overload resolution for deleters of reference type. That way trying to initialize a deleter of reference type from an rvalue will resolve to the deleted function as intended.

I think we can just change the "shall not participate" condition to only apply to non-reference deleters. For reference deleters the condition is always true for the first constructor, because is_constructible<A&, A&> is true, and always false for the second constructor (but we want it to be true so the deleted constructor is used). So for both constructors, the "shall not participate" isn't useful when D is a reference type.

11-2018 Status to NAD after discussion on the reflector.

Proposed resolution:

This wording is relative to N4778.

  1. Change 20.3.1.3.2 [unique.ptr.single.ctor] as indicated:

    unique_ptr(pointer p, const D& d) noexcept;
    unique_ptr(pointer p, remove_reference_t<D>&& d) noexcept;
    

    […]

    -11- Remarks: If D is a reference type, the second constructor is defined as deleted. If D is not a reference type, tThese constructors shall not participate in overload resolution unless is_constructible_v<D, decltype(d)> is true.

    […]


3165(i). All starts_with() overloads should be called "begins_with"

Section: 23.4.3 [basic.string], 23.3.3 [string.view.template] Status: NAD Submitter: Gennaro Prota Opened: 2018-10-22 Last modified: 2020-08-21

Priority: 2

View other active issues in [basic.string].

View all other issues in [basic.string].

View all issues with NAD status.

Discussion:

Throughout the standard library, the opposite of the verb "to end" is "to begin", not to "to start". In this respect, all the overloads of std::basic_string::starts_with and std::basic_string_view::starts_with are named inconsistently. These calls were added in P0457R2, which refers to analogous functions in Java, Python, Qt, LLVM, WebKit, all of which use the term "start". But, in our opinion, the C++ standard library should first be consistent with itself.

[2018-11: Referred to LEWG after reflector discussion.]

[2020-05-28; LEWG issue reviewing]

LEWG issue processing voted to reject 3165 as NAD. Status change to Open.

Reject LWG3165 as NAD

SF F N A SA
21 1 0 1  0

[2020-08-21; Issue processing telecon: moved to NAD based on LEWG recommendation]

Proposed resolution:

This wording is relative to N4778.

  1. Change 23.4.3 [basic.string], class template basic_string synopsis, as indicated:

    bool beginstarts_with(basic_string_view<charT, traits> x) const noexcept;
    bool beginstarts_with(charT x) const noexcept;
    bool beginstarts_with(const charT* x) const;
    
  2. Change 23.4.3.8.5 [string.starts.with] as indicated

    bool beginstarts_with(basic_string_view<charT, traits> x) const noexcept;
    bool beginstarts_with(charT x) const noexcept;
    bool beginstarts_with(const charT* x) const;
    

    -1- Effects: Equivalent to:

    return basic_string_view<charT, traits>(data(), size()).beginstarts_with(x);
    
  3. Change 23.3.3 [string.view.template], class template basic_string_view synopsis, as indicated:

    constexpr bool beginstarts_with(basic_string_view x) const noexcept;
    constexpr bool beginstarts_with(charT x) const noexcept;
    constexpr bool beginstarts_with(const charT* x) const;
    
  4. Change 23.3.3.8 [string.view.ops] as indicated:

    constexpr bool beginstarts_with(basic_string_view x) const noexcept;
    

    -20- Effects: Equivalent to: return compare(0, npos, x) == 0;

    constexpr bool beginstarts_with(charT x) const noexcept;
    

    -21- Effects: Equivalent to: return beginstarts_with(basic_string_view(&x, 1));

    constexpr bool beginstarts_with(const charT* x) const;
    

    -22- Effects: Equivalent to: return beginstarts_with(basic_string_view(x));


3181(i). split_view::outer_iterator converting constructor is misconstrained

Section: 26.7.16.3 [range.lazy.split.outer] Status: NAD Submitter: Casey Carter Opened: 2019-01-05 Last modified: 2023-02-07

Priority: Not Prioritized

View other active issues in [range.lazy.split.outer].

View all other issues in [range.lazy.split.outer].

View all issues with NAD status.

Discussion:

The constraints on split_view::outer_iterator's converting constructor:

constexpr outer_iterator(outer_iterator<!Const> i)
  requires Const && ConvertibleTo<iterator_t<V>, iterator_t<const V>>;
are problematic. Since const V doesn't depend on a template parameter of outer_iterator, compilers want to expand iterator_t<const V> when outer_iterator is instantiated - regardless of the value of Const - which can result in a hard error when iterator_t<const V> is ill-formed. The constraint should instead use iterator_t<Base>, as do the similar converting constructors for the other range adaptor iterator and sentinel types.

[2019-01-20 Reflector prioritization]

Set status to NAD

Proposed resolution:

This wording is relative to N4791.

  1. Change [range.split.outer] as follows:

    namespace std::ranges {
      template<class V, class Pattern>
      template<bool Const>
      struct split_view<V, Pattern>::outer_iterator {
        […]
        constexpr outer_iterator(outer_iterator<!Const> i)
          requires Const && ConvertibleTo<iterator_t<V>, iterator_t<const VBase>>;
        […]
      };
    }
    

    […]

    constexpr outer_iterator(outer_iterator<!Const> i)
      requires Const && ConvertibleTo<iterator_t<V>, iterator_t<const VBase>>;
    

    -4- Effects: Initializes parent_ with i.parent_ and current_ with std::move(i.current_).

    […]


3207(i). N in ssize(const T (&)[N]) should be size_t

Section: 25.7 [iterator.range] Status: NAD Submitter: Nevin Liber Opened: 2019-05-23 Last modified: 2020-11-09

Priority: Not Prioritized

View other active issues in [iterator.range].

View all other issues in [iterator.range].

View all issues with NAD status.

Discussion:

The N in ssize(const T (&)[N]) is specified to be of type ptrdiff_t. It should be size_t to be consistent with the rest of the standard library, such as the array overloads for all other range access functions, the swap overload for arrays, and other function template overloads for arrays. (Note: the return type of this function should still be ptrdiff_t.)

[2019-06-12 Tentatively NAD after reflector discussion]

[2020-11-09 Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4810.

  1. Modify 25.7 [iterator.range] as indicated:

    template<class T, ptrdiffsize_t N> constexpr ptrdiff_t ssize(const T (&array)[N]) noexcept;
    

    -19- Returns: static_cast<ptrdiff_t>(N).


3271(i). Parsing functions should save and restore stream format state

Section: 29.13 [time.parse] Status: NAD Submitter: Howard Hinnant Opened: 2019-09-02 Last modified: 2020-09-06

Priority: 3

View all other issues in [time.parse].

View all issues with NAD status.

Discussion:

from_stream overloads may need to set various formatting state and flags of the basic_istream to implement the parsing functionality. Such settings must not persist beyond the parsing functions. Thus these functions need to save and restore any settings on the basic_istream they may need to change.

[2019-10 Priority set to 3 after reflector discussion]

[2020-02, Prague; Close as NAD]

LWG reviewed the issue and decided no change is needed. Unless specified to change the flags, it's assumed that the stream flags are unchanged by any input function.

Proposed resolution:

This wording is relative to N4830.

[Drafting note: The modification of 29.13 [time.parse] p1 is intended to be non-conflictingly mergeable with the change suggested by LWG 3269 at the same paragraph.]

  1. Modify 29.13 [time.parse] as indicated:

    -1- Each parse overload specified in this subclause calls from_stream unqualified, so as to enable argument dependent lookup (6.5.4 [basic.lookup.argdep]). In the following paragraphs, let is denote an object of type basic_istream<charT, traits>, where charT and traits are template parameters in that context.

    […]

    -10- All from_stream overloads behave as unformatted input functions, except that they have an unspecified effect on the value returned by subsequent calls to basic_istream<>::gcount(). Any changes made to is.fill(), is.width() or is.flags() are undone prior to an returning or exceptional execution ([thread.once.callonce]). Each overload takes a format string containing ordinary characters and flags which have special meaning. Each flag begins with a %. Some flags can be modified by E or O. During parsing each flag interprets characters as parts of date and time types according to Table [tab:time.parse.spec]. Some flags can be modified by a width parameter given as a positive decimal integer called out as N below which governs how many characters are parsed from the stream in interpreting the flag. All characters in the format string that are not represented in Table [tab:time.parse.spec], except for white space, are parsed unchanged from the stream. A white space character matches zero or more white space characters in the input stream.


3298(i). Range adaptors introduced by P1035 do not require viewable_range

Section: 26.7.11.2 [range.take.while.view], 26.7.12.2 [range.drop.view], 26.7.13.2 [range.drop.while.view] Status: NAD Submitter: Christopher Di Bella Opened: 2019-09-24 Last modified: 2020-09-06

Priority: Not Prioritized

View all other issues in [range.take.while.view].

View all issues with NAD status.

Discussion:

After adoption of P1035 range adaptor closure is described to take a viewable_range as input, and return a view as output. The deduction-guides for take_while_view, drop_view, and drop_while_view do not currently impose the viewable_range requirement.

[2020-02 Status to NAD on Thursday morning in Prague.]

Proposed resolution:

This wording is relative to N4830.

  1. Modify 26.7.11.2 [range.take.while.view], class template take_while_view synopsis, as indicated:

    namespace std::ranges {
      […]
      template<classinput_range R, class Pred>
        requires viewable_range<R> && 
            indirect_unary_predicate<const Pred, iterator_t<R>>
          take_while_view(R&&, Pred) -> take_while_view<all_view<R>, Pred>;
    }
    
  2. Modify 26.7.12.2 [range.drop.view], class template drop_view synopsis, as indicated:

    namespace std::ranges {
      […]
      template<classinput_range R>
        requires viewable_range<R>
          drop_view(R&&, range_difference_t<R>) -> drop_view<all_view<R>>;
    }
    
  3. Modify 26.7.13.2 [range.drop.while.view], class template drop_while_view synopsis, as indicated:

    namespace std::ranges {
      […]
      template<classinput_range R, class Pred>
        requires viewable_range<R> && 
            indirect_unary_predicate<const Pred, iterator_t<R>>
          drop_while_view(R&&, Pred) -> drop_while_view<all_view<R>, Pred>;
    }
    

3311(i). basic_string::operator=(charT c) should be constrained

Section: 23.4.3.3 [string.cons] Status: Dup Submitter: Russia Opened: 2019-11-04 Last modified: 2019-11-16

Priority: Not Prioritized

View all other issues in [string.cons].

View all issues with Dup status.

Discussion:

Addresses RU 013

Because of the implicit conversion of arithmetic types it is error prone to use the basic_string::operator=(charT c):

double d = 3.14;
std::string s;
s = d; // Compiles

Make sure that the program is ill-formed if an implicit conversion from arithmetic type happens while assigning to std::basic_string. Or at least make sure that the program is ill-formed if an implicit conversion from floating point type happens while assigning to std::basic_string.

[2019-11 Status to Duplicate during Tuesday morning issue processing in Belfast.]

Duplicate of 2372.

Proposed resolution:

This wording is relative to N4835.

  1. Modify 23.4.3 [basic.string] p3, class template basic_string synopsis, as indicated:

    template<class T>
    constexpr basic_string& operator=(charT c);
    
  2. Modify 23.4.3.3 [string.cons] as indicated:

    template<class T>
    constexpr basic_string& operator=(charT c);
    

    -?- Constraints: is_same_v<T, charT> is true.

    -30- Effects: Equivalent to:

    return *this = basic_string_view<charT, traits>(addressof(c), 1);
    


3312(i). polymorphic_allocator::allocate_object and new_object should be [[nodiscard]]

Section: 20.4.3 [mem.poly.allocator.class] Status: Dup Submitter: United States Opened: 2019-11-04 Last modified: 2019-11-16

Priority: Not Prioritized

View all other issues in [mem.poly.allocator.class].

View all issues with Dup status.

Discussion:

Addresses US 217

Add [[nodiscard]] in front of the return type for allocate_object and new_object in class declaration and in member-function description for polymorphic_allocator template.

[Daniel comments]

This issue is related to LWG 3304.

[2019-11 Status to Duplicate during Tuesday morning issue processing in Belfast.]

Duplicate of 3304.

Proposed resolution:

This wording is relative to N4835.

  1. Modify 20.4.3 [mem.poly.allocator.class] p2, class template polymorphic_allocator synopsis, as indicated:

    // 20.4.3.3 [mem.poly.allocator.mem], member functions
    [[nodiscard]] Tp* allocate(size_t n);
    void deallocate(Tp* p, size_t n);
    
    void* allocate_bytes(size_t nbytes, size_t alignment = alignof(max_align_t));
    void deallocate_bytes(void* p, size_t nbytes, size_t alignment = alignof(max_align_t));
    template<class T> [[nodiscard]] T* allocate_object(size_t n = 1);
    template<class T> void deallocate_object(T* p, size_t n = 1);
    template<class T, class... CtorArgs> [[nodiscard]] T* new_object(CtorArgs&&... ctor_args);
    template<class T> void delete_object(T* p);
    
  2. Modify 20.4.3.3 [mem.poly.allocator.mem] as indicated:

    template<class T>
      [[nodiscard]] T* allocate_object(size_t n = 1);
    

    -8- Effects: Allocates memory suitable for holding an array of n objects of type T, as follows:

    1. (8.1) — if SIZE_MAX / sizeof(T) < n, throws length_error,

    2. (8.2) — otherwise equivalent to:

      return static_cast<T*>(allocate_bytes(n*sizeof(T), alignof(T)));
      

    […]

    template<class T, class CtorArgs...>
      [[nodiscard]] T* new_object(CtorArgs&&... ctor_args);
    

    -11- Effects: Allocates and constructs an object of type T, as follows. Equivalent to:

    T* p = allocate_object<T>();
    try {
      construct(p, std::forward<CtorArgs>(ctor_args)...);
    } catch (...) {
      deallocate_object(p);
      throw;
    }
    return p;
    

    […]


3333(i). ranges::cbegin/ranges::cend, (and maybe ranges::crbegin/ranges::crend) are under-specified to allow rvalue-arrays

Section: 26.3.4 [range.access.cbegin], 26.3.5 [range.access.cend] Status: NAD Submitter: Christopher Di Bella Opened: 2019-11-06 Last modified: 2020-02-13

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

The following should be ill-formed, but is not, likely due to decay-to-pointer or reference collapsing.

double x[321];
ranges::cbegin(std::move(x)); // should be ill-formed
ranges::cend(std::move(x));   // should be ill-formed

This might be a problem for ranges::crbegin and ranges::crend, but I haven't implemented any of rbegin, rend, crbegin, and crend to confirm as much.

[2020-02 Status to NAD on Thursday morning in Prague.]

Proposed resolution:

This wording is relative to N4835.

  1. Change 26.3.4 [range.access.cbegin] as indicated:

    -1- The name ranges::cbegin denotes a customization point object (16.3.3.3.5 [customization.point.object]). If tThe expression ranges::cbegin(E) for some subexpression E of type T is well-formed, then the expression ranges::cbegin(E) is expression-equivalent to:

    1. (1.1) — ranges::begin(static_cast<const T&>(E)) if E is an lvalue.

    2. (1.2) — Otherwise, ranges::begin(static_cast<const T&&>(E)).

  2. Change 26.3.5 [range.access.cend] as indicated:

    -1- The name ranges::cend denotes a customization point object (16.3.3.3.5 [customization.point.object]). If tThe expression ranges::cend(E) for some subexpression E of type T is well-formed, then the expression ranges::cend(E) is expression-equivalent to:

    1. (1.1) — ranges::end(static_cast<const T&>(E)) if E is an lvalue.

    2. (1.2) — Otherwise, ranges::end(static_cast<const T&&>(E)).


3365(i). Rename ref-is-glvalue to deref-is-ref

Section: 26.7.14.3 [range.join.iterator] Status: NAD Submitter: Johel Ernesto Guerrero Peña Opened: 2020-01-07 Last modified: 2020-11-09

Priority: 0

View all other issues in [range.join.iterator].

View all issues with NAD status.

Discussion:

The name of join_view::iterator::ref-is-glvalue, defined as is_reference_v<range_reference_t<Base>>, doesn't take into account the fact that it may also be true because range_reference_t<Base> is an rvalue (e.g. for move_iterator<int>).

We suggest renaming it to deref-is-ref.

[2020-01-14 Status set to Tentatively NAD after five positive votes on the reflector. This issue was based on a misunderstanding by the submitter which had been cleared up in an related editorial issue.]

[2020-11-09 Status changed: Tentatively NAD → NAD.]

Proposed resolution:

Rename all occurrences referring to the symbol join_view::iterator::ref-is-glvalue to deref-is-ref.


3394(i). ranges::basic_istream_view::iterator has an empty iterator_traits

Section: 26.6.6.3 [range.istream.iterator] Status: NAD Submitter: Patrick Palka Opened: 2020-02-09 Last modified: 2020-11-09

Priority: 2

View all other issues in [range.istream.iterator].

View all issues with NAD status.

Discussion:

Every instantiation of ranges::basic_istream_view::iterator has an empty iterator_traits, i.e. the type

iterator_traits<ranges::basic_istream_view<Val, CharT, Traits>::iterator>

has no members.

This happens because basic_istream_view::iterator neither models cpp17-iterator (since this concept requires copyability, which this iterator is by design not) nor does it define all four of the member types difference_type, value_type, reference, and iterator_category (it is missing reference). Therefore by 25.3.2.3 [iterator.traits] p3, this iterator's specialization of iterator_traits will be empty if generated from the iterator_traits primary template.

Since basic_istream_view::iterator is indeed an iterator, its iterator_traits should certainly not be empty. The simplest solution here is to define the member type reference in the definition of basic_istream_view::iterator, which will enable its iterator_traits specialization to be appropriately populated from the primary template.

[2020-02-10; Jonathan comments]

Jonathan and Casey have concerns about the proposed resolution. Casey: The wording makes it look as if this iterator is supposed to be an cpp17-input-iterator.

See also LWG 3283 and 3289.

[2020-02 Prioritized as P2 Monday morning in Prague]

Previous resolution [SUPERSEDED]:

This wording is relative to N4849.

  1. Modify 26.6.6.3 [range.istream.iterator], class template basic_istream_view::iterator synopsis, as indicated:

    namespace std::ranges {
      template<class Val, class CharT, class Traits>
      class basic_istream_view<Val, CharT, Traits>::iterator { // exposition only
      public:
        using iterator_category = input_iterator_tag;
        using difference_type = ptrdiff_t;
        using value_type = Val;
        using reference = Val&;
    
        iterator() = default;
        […]
      };
    }
    

[2020-02-13, Prague]

LWG decided for NAD: The ranges::basic_istream_view::iterator is a move-only type and thus cannot meet the Cpp17 requirements (even output_iterator_tag), as such it should not specialize iterator_traits, to avoid misleading results when it is passed to new algorithms.

A related issue, LWG 3397, is supposed to take care for a problem with the definition of the iterator_category member type of this template.

[2020-11-09 Status changed: Tentatively NAD → NAD.]

Proposed resolution:


3399(i). basic_syncbuf::emit() + Qt's #define emit = Big Bada-Boom

Section: 31.11.2.1 [syncstream.syncbuf.overview] Status: NAD Submitter: Marc Mutz Opened: 2019-02-14 Last modified: 2020-11-09

Priority: Not Prioritized

View all other issues in [syncstream.syncbuf.overview].

View all issues with NAD status.

Discussion:

The current IS contains a function called emit() (in basic_syncbuf, added by P0053).

emit is a macro in pervasive use in every Qt program. #include'ing <osyncstream> after any Qt header would break, because emit was #define'd to nothing by the Qt headers.

It is understood that the committee cannot check every 3rd-party library out there that chooses (badly, as I’d readily concur) to #define a macro of all-lowercase letters, but the min/max issue in the Windows headers caused so much pain for our users, that we probably should avoid a breakage here, for the benefit of our users that have to work Qt.

It also doesn’t seem like emit() is a particularly mandatory name for the syncbuf function. Since it returns bool, it could just as easily be called try_emit() and the issue would be solved.

Suggested approach:

In basic_syncbuf, rename bool emit() to bool try_emit(). In basic_osyncstream, where the function doesn't return bool, but sets the stream's failbit, rename void emit() to void emit_or_fail().

[2020-02-14, Prague]

The issue was send to LEWG, who made the following poll:

We're open to renaming osyncstream::emit() (and related).

SF F N A SA
1  1 5 8 20

[2020-11-09 Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4849.

  1. Modify 31.7.6.5 [ostream.manip] as indicated:

    template<class charT, class traits>
      basic_ostream<charT, traits>& flush_emit(basic_ostream<charT, traits>& os);
    

    -12- Effects: Calls os.flush(). Then, if os.rdbuf() is a basic_syncbuf<charT, traits, Allocator>*, called buf for the purpose of exposition, calls buf->try_emit().

  2. Modify 31.11.2.1 [syncstream.syncbuf.overview] as indicated:

    […]
    // 31.11.2.4 [syncstream.syncbuf.members], member functions
    bool try_emit();
    streambuf_type* get_wrapped() const noexcept;
    […]
    

    -1- Class template basic_syncbuf stores character data written to it, known as the associated output, into internal buffers allocated using the object's allocator. The associated output is transferred to the wrapped stream buffer object *wrapped when try_emit() is called or when the basic_syncbuf object is destroyed. Such transfers are atomic with respect to transfers by other basic_syncbuf objects with the same wrapped stream buffer object.

  3. Modify 31.11.2.2 [syncstream.syncbuf.cons] as indicated:

    ~basic_syncbuf();
    

    -7- Effects: Calls try_emit().

    -8- Throws: Nothing. If an exception is thrown from try_emit(), the destructor catches and ignores that exception.

  4. Modify 31.11.2.3 [syncstream.syncbuf.assign] as indicated:

    basic_syncbuf& operator=(basic_syncbuf&& rhs) noexcept;
    

    -1- Effects: Calls try_emit() then move assigns from rhs. After the move assignment *this has the observable state it would have had if it had been move constructed from rhs (31.11.2.2 [syncstream.syncbuf.cons]).

  5. Replace in 31.11.2.4 [syncstream.syncbuf.members] all occurrences of emit() by try_emit() (five occurrences)

  6. Replace in 31.11.2.5 [syncstream.syncbuf.virtuals] all occurrences of emit() by try_emit() (three occurrences)

  7. Modify 31.11.3.1 [syncstream.osyncstream.overview] as indicated:

    […]
    // 31.11.3.3 [syncstream.osyncstream.members], member functions
    void emit_or_fail();
    streambuf_type* get_wrapped() const noexcept;
    […]
    
  8. Replace in 31.11.3.2 [syncstream.osyncstream.cons] all occurrences of emit() by emit_or_fail() (three occurrences)

  9. Replace in [syncstream.osyncstream.assign] all occurrences of emit() by emit_or_fail() (two occurrences)

  10. Replace in 31.11.3.3 [syncstream.osyncstream.members] all occurrences of emit() by emit_or_fail() (six occurrences)


3415(i). back_insert_iterator fails when a container is also its value type

Section: 25.5.2.2 [back.insert.iterator] Status: NAD Submitter: Billy O'Neal III Opened: 2020-03-03 Last modified: 2020-07-17

Priority: Not Prioritized

View all other issues in [back.insert.iterator].

View all issues with NAD status.

Discussion:

Consider the following:

#include <algorithm>
#include <iterator>
#include <vector>

struct example_container 
{
  using value_type = std::back_insert_iterator<example_container>;
  void push_back(const value_type&) {}
};

int main() 
{
  std::vector<std::back_insert_iterator<example_container>> v;
  example_container ex;
  std::copy(v.begin(), v.end(), std::back_inserter(ex));
}

This example is out-of-contract in the current standard because it creates back_insert_iterator<incomplete>, as per 16.4.5.8 [res.on.functions]/2. However, it might be something we are considering for future iterators and proxy reference types. In practice, the "Ill-formed, no diagnostic required" the user is likely to get is an ambiguity between what back_insert_iterator's copy assignment operator, and its "push back assigning operator". We could resolve this by changing the return type of operator* to a proxy in the same way istream_iterator does, though that might be ABI breaking for some implementations.

We should consider having a standing LWG/LEWG policy that iterators are not their own proxy operator* type if we intend to leave the door open to more incomplete type support in the standard library.

[2020-07-17; Status changed to NAD in telecon]

We reviewed the reflector discussion and were not motivated to support this. There were concerns that adding incomplete type support elsewhere in containers has caused us regrettable problems, and we're not sure we could get this right even if we wanted to support it.

Proposed resolution:


3440(i). Aggregate-paren-init breaks direct-initializing a tuple or optional from {aggregate-member-value}

Section: 22.4.4.1 [tuple.cnstr], 22.5.3.2 [optional.ctor] Status: NAD Submitter: Ville Voutilainen Opened: 2020-05-01 Last modified: 2020-11-09

Priority: 2

View other active issues in [tuple.cnstr].

View all other issues in [tuple.cnstr].

View all issues with NAD status.

Discussion:

For reference, see this gcc bug report.

Constructing a tuple or optional from an element value of an aggregate is broken in C++20. tuple<c> t({val}); and optional<c> t({val}); invoked a non-forwarding constructor before, but now the perfect-forwarding converting constructors are a match, because the element is constructible from {val}. But it's not convertible, so overload resolution chooses the explicit constructor, and the initialization fails.

Tim Song explains the overload resolution in this reflector discussion.

Now that we understand that C++17 called the non-forwarding conversion constructor, and C++20 tries to use the forwarding conversion constructor, we have the solution. SFINAE away the forwarding conversion constructor when it would convert an aggregate.

This also means that tuple<c> t(0); won't work, which is unfortunate because tuple<c>/optional<c> no longer mirrors what c can do. That's okay; in this LWG issue, we first restore feature parity with C++17, and later, as an extension, enable such initializations so that tuple/optional mirrors what c can do in C++20.

The proposed wording below has been implemented and tested.

[2020-05-09; Reflector prioritization]

Set priority to 2 after reflector discussions.

[2020-06-11; LWG Telecon: Status changed: New → LEWG.]

Ask LEWG if it's desirable to make ({val}) work again. Tomasz would prefer it to be explicit e.g. via std::in_place.

[2020-06-23; LEWG Telecon]

POLL: Make ({val}) work again, at the risk of non-transparency of tuple constructors and further complicating the tuple and optional overload set.

SF F N A SA
0  5 6 9 0

No consensus for change. Close as Not a defect.

[2020-11-09 Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4861.

  1. Modify 22.4.4.1 [tuple.cnstr] as indicated:

    template<class... UTypes> constexpr explicit(see below) tuple(UTypes&&... u);
    

    -11- Constraints: sizeof...(Types) equals sizeof...(UTypes) and sizeof...(Types) ≥ 1 and is_constructible_v<Ti, Ui> is true for all i and conjunction_v<is_aggregate<remove_reference_t<Ti>>, negation<is_same<remove_reference_t<Ti>, remove_reference_t<Ui>>>> is false for all i.

  2. Modify 22.5.3.2 [optional.ctor] as indicated:

    template<class U = T> constexpr explicit(see below) optional(U&& v);
    

    -22- Constraints: is_constructible_v<T, U> is true, is_same_v<remove_cvref_t<U>, in_place_t> is false, and is_same_v<remove_cvref_t<U>, optional> is false, and conjunction_v<is_aggregate<T>, negation<is_same<T, remove_reference_t<U>>>> is false.


3468(i). Transparent lookups in unordered containers are inconsistent

Section: 24.2.8 [unord.req] Status: NAD Submitter: Marshall Clow Opened: 2020-07-23 Last modified: 2020-08-21

Priority: Not Prioritized

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with NAD status.

Discussion:

In C++14, we added "transparent lookups" into the ordered associative containers. This was sold as an efficiency concern, as removing the need to create temporary objects just to compare against.

However, people found clever ways to use this. One of them, in fact, was in the original paper, and was the subject of a question on Stack Overflow.

This all works because the elements in the ordered associative containers are, well, ordered.

For C++20, we added this facility to the unordered containers.

Consider the following code:

#include <unordered_set>
#include <string>
#include <iostream>

struct DumbHash // put everything in the same bucket
{
  using is_transparent = void;

  template<typename T>
  size_t operator()(const T&) const { return 0; }
};

struct CompareEQ 
{
  using is_transparent = void;

  bool operator()(const std::string& lhs, const std::string& rhs) const
  { return lhs == rhs; }

  bool operator()(const std::string& lhs, char rhs) const
  { return !lhs.empty() && (lhs[0] == rhs); }

  bool operator()(char lhs, const std::string& rhs) const
  { return !rhs.empty() && (lhs == rhs[0]); }
};

int main () 
{
  const char* one[] = {"a", "b",  "c", "d",  "e", "bb"};
  const char* two[] = {"b", "e",  "d", "bb", "c", "a"};
  const char* thr[] = {"b", "bb", "a", "c",  "d", "e"};

  typedef std::unordered_set<std::string, DumbHash, CompareEQ> MS;
  MS m1{std::begin(one), std::end(one)};
  MS m2{std::begin(two), std::end(two)};
  MS m3{std::begin(thr), std::end(thr)};

  for (const auto& s: m1) 
    std::cout << s << ' '; 
  std::cout << std::endl;
  for (const auto& s: m2) 
    std::cout << s << ' '; 
  std::cout << std::endl;
  for (const auto& s: m3) 
    std::cout << s << ' '; 
  std::cout << std::endl;

  std::cout << "m1:" << m1.count('b') << ' ';
  std::cout << "m2:" << m2.count('b') << ' ';
  std::cout << "m3:" << m3.count('b');
}

When I run this program on my Mac, I get the following output:

bb e d c b a
a c bb d e b
e d c a bb b
m1:1 m2:1 m3:2

(This is using unreleased code, but I have confirmed this with VS2019's unordered_multiset.)

This is clearly bad; three containers, containing the same elements, doing the same lookups, giving different results. This also applies to the transparent versions of find, equal_range, and contains.

The problem is that the elements in the unordered containers are only partially ordered; i.e, all the elements that are equal (according to the non-transparent version of the comparison predicate) are adjacent in the container, but are unordered relative to the other elements in the container.

My recommendation is to declare this all UB.

Suggested resolution:

Add a precondition to all of the transparent lookup functions for the unordered containers forbidding stuff like this. This should probably go in 24.2.8 [unord.req], maybe at the end of Table [tab:container.hash.req].

Precondition: The value being searched matches at most one unique key in the container.

[2020-08-21 Issue processing telecon: NAD based on reflector discussion. Status changed: New → NAD.]

Proposed resolution:


3558(i). elements_view::sentinel's first operator- has wrong return type

Section: 26.7.22.4 [range.elements.sentinel] Status: NAD Editorial Submitter: Hewill Kang Opened: 2021-05-28 Last modified: 2021-06-07

Priority: Not Prioritized

View all issues with NAD Editorial status.

Discussion:

Because the iterator type of the first operator- is const iterator<OtherConst>&, its return type should be range_difference_t<maybe-const<OtherConst, V>>.

[2021-06-07 Fixed by editorial issue #4603. Status changed: New → NAD Editorial.]

Proposed resolution:

This wording is relative to N4885.

  1. Modify 26.7.22.4 [range.elements.sentinel] as indicated:

        […]
        template<bool OtherConst>
          requires sized_sentinel_for<sentinel_t<Base>, iterator_t<maybe-const<OtherConst, V>>>
        friend constexpr range_difference_t<Basemaybe-const<OtherConst, V>>
          operator-(const iterator<OtherConst>& x, const sentinel& y);
        
        template<bool OtherConst>
          requires sized_sentinel_for<sentinel_t<Base>, iterator_t<maybe-const<OtherConst, V>>>
        friend constexpr range_difference_t<maybe-const<OtherConst, V>>
          operator-(const sentinel& x, const iterator<OtherConst>& y);
      };
    }
    
    […]
    template<bool OtherConst>
      requires sized_sentinel_for<sentinel_t<Base>, iterator_t<maybe-const<OtherConst, V>>>
    friend constexpr range_difference_t<Basemaybe-const<OtherConst, V>>
      operator-(const iterator<OtherConst>& x, const sentinel& y);
    

    -5- Effects: Equivalent to: return x.current_ - y.end_;


3562(i). Superseding nullopt_t's requirement to not be DefaultConstructible

Section: 22.5.4 [optional.nullopt] Status: NAD Submitter: David Friberg Opened: 2021-06-04 Last modified: 2021-06-14

Priority: Not Prioritized

View all other issues in [optional.nullopt].

View all issues with NAD status.

Discussion:

22.5.4 [optional.nullopt] p2 requires that nullopt_t shall not have a default constructor ([…] or an initializer-list constructor, and shall not be an aggregate). However, after the final resolution of LWG issue 2510 and CWG 1518 it seems as if the following synopsis would suffice:

struct nullopt_t { explicit nullopt_t() = default; };

Which would also be consistent with other tag types.

22.5.4 [optional.nullopt] p2 in its current form was written prior to the final resolution of LWG issue 2510 and CWG 1518, and is arguably based on using an earlier proposed approach for tag types which was later superseded.

Details:

P0032R3 was part of introducing optional and contained a discussion around the nullopt_t type, tag types in general, and the DefaultConstructible requirement. Particularly, from section 'Not assignable from {}':

To re-enforce this design, there is an pending issue 2510-Tag types should not be DefaultConstructible Core issue 2510.

At the time of P0032R3, the resolution of LWG issue 2510 was indeed for all tag types to not be DefaultConstructible, but this resolution was later superseded as part of the resolution of CWG issue CWG 1518, which in turn reverted parts of the resolution of CWG issue CWG 1630 (which "went too far in allowing use of explicit constructors for default initialization").

The final resolution for CWG 1518 (P0398R0) lead to the requirement for aggregate classes to not have any explicit [user-declared] constructors, and LWG 2510 was resolved accordingly by making all tag types have explicit user-declared (constexpr) default constructors.

However, the spec of nullopt_t never changed after these events or as part of updating other tag types, and it is still explicitly required to not be DefaultConstructible (nor have a initialization-list constructor) and to not be an aggregate, whereas the implementation of it is otherwise unspecified.

We may note that we still see some compiler variance on the topic of CWG issues 1518 & 1630, see this godbolt link.

[2021-06-14 Reflector poll: Current wording prevents an implicit conversion sequence from {}. Status changed: New → NAD.]

Proposed resolution:

This wording is relative to N4885.

  1. Modify 22.5.2 [optional.syn], header <optional> synopsis, as indicated:

    […]
    
    // 22.5.4 [optional.nullopt], no-value state indicator
    struct nullopt_t{see below explicit nullopt_t() = default; };
    inline constexpr nullopt_t nullopt{}(unspecified);
    
    […]
    
  2. Modify 22.5.4 [optional.nullopt] as indicated:

    struct nullopt_t{see below explicit nullopt_t() = default; };
    inline constexpr nullopt_t nullopt{}(unspecified);
    

    -1- The struct nullopt_t is an empty class type used as a unique type to indicate the state of not containing a value for optional objects. In particular, optional<T> has a constructor with nullopt_t as a single argument; this indicates that an optional object not containing a value shall be constructed.

    -2- Type nullopt_t shall not have a default constructor or an initializer-list constructor, and shall not be an aggregate.


3579(i). Complexity guarantees for resize() and append() functions across the library

Section: 23.4.3.5 [string.capacity], 24.3.8.3 [deque.capacity], 24.3.10.3 [list.capacity], 24.3.11.3 [vector.capacity], 24.3.9.5 [forward.list.modifiers], 28.6.2.8 [valarray.members], 23.4.3.7.2 [string.append], 31.12.6.5.3 [fs.path.append] Status: NAD Submitter: Louis Dionne Opened: 2021-08-11 Last modified: 2022-08-23

Priority: 3

View all other issues in [string.capacity].

View all issues with NAD status.

Discussion:

This issue requests to clarify the complexity requirements for resize and append member functions across the library. Currently, none of them have complexity requirements associated to them, which means that implementations are free to use a geometric growth approach or not. A geometric growth approach (like what's required by push_back) implies that calling resize/append with successively larger sizes will have amortized 𝒪(N) complexity, whereas using a resize-exactly approach makes that pattern 𝒪(N2).

I believe the Standard should either specify that those member functions are required to have behavior that is consistent with push_back, or explicitly mention that implementations are allowed to use whatever growth strategy they want. If we do the former, users could start relying on this guarantee in a portable manner. If we do the latter, it would make it clear to users that they should not rely on the amortized 𝒪(N) guarantee if they want their code to be portable.

The following classes have a resize() member function and also a push_back() member function:

The following classes have a resize() member function but do not support push_back():

The following classes have an append() member function:

None of them specify a complexity requirement. I think we should update all of them to describe what is expected or permitted in an implementation.

[2021-08-20; Reflector poll]

Set priority to 3 and status to "LEWG" after reflector poll.

[2022-02-22 LEWG telecon; Status changed: LEWG → Tentatively NAD.]

A paper would be needed. Such a paper would need to include discussion on whether allocate_at_least (new for C++23) has an impact.

[2022-08-23 Status changed: Tentatively NAD → NAD.]

Proposed resolution:


3588(i). Strike out purposeless UB involving the deleter in members functions of unique_ptr

Section: 20.3.1.3.1 [unique.ptr.single.general] Status: NAD Submitter: Johel Ernesto Guerrero Peña Opened: 2021-09-07 Last modified: 2021-09-20

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

This originated from the editorial issues #4871 and #4872.

Several member functions of unique_ptr are noexcept, and yet, they have the precondition that an expression involving the deleter does not exit via an exception. There's nothing an implementation or user can take advantage of in presence of this UB. Since the behavior otherwise would be a call to std::terminate, these preconditions should be striked out.

Note that although ~unique_ptr() is not noexcept, 16.4.6.13 [res.on.exception.handling] p3 specifies it to behave as if it were.

[2021-09-20 Status changed: New → NAD.]

The current specification allows the compiler to omit noexcept-enforcement.

Proposed resolution:

This wording is relative to N4892.

  1. Modify 20.3.1.3.2 [unique.ptr.single.ctor] as indicated:

    constexpr unique_ptr() noexcept;
    constexpr unique_ptr(nullptr_t) noexcept;
    

    -1- […]

    -2- Preconditions: D meets the Cpp17DefaultConstructible requirements (Table 27), and that construction does not throw an exception.

    […]

    explicit unique_ptr(pointer p) noexcept;
    

    -5- Constraints: […]

    -6- Mandates: […]

    -7- Preconditions: D meets the Cpp17DefaultConstructible requirements (Table 27), and that construction does not throw an exception.

    […]

    unique_ptr(pointer p, const D& d) noexcept;
    unique_ptr(pointer p, remove_reference_t<D>&& d) noexcept;
    

    -10- Constraints: […]

    -11- Mandates: […]

    -12- Preconditions: For the first constructor, if D is not a reference type, D meets the Cpp17CopyConstructible requirements and such construction does not exit via an exception. For the second constructor, if D is not a reference type, D meets the Cpp17MoveConstructible requirements and such construction does not exit via an exception.

    […]

    unique_ptr(unique_ptr&& u) noexcept;
    

    -17- Constraints: […]

    -18- Preconditions: If D is not a reference type, D meets the Cpp17MoveConstructible requirements (Table 28). Construction of the deleter from an rvalue of type D does not throw an exception.

    […]

    template<class U, class E> unique_ptr(unique_ptr<U, E>&& u) noexcept;
    

    -21- Constraints: […]

    -22- Preconditions: If E is not a reference type, construction of the deleter from an rvalue of type E is well-formed and does not throw an exception. Otherwise, E is a reference type and construction of the deleter from an lvalue of type E is well-formed and does not throw an exception.

    […]

  2. Modify 20.3.1.3.3 [unique.ptr.single.dtor] as indicated:

    ~unique_ptr();
    

    -1- Preconditions: The expression get_deleter()(get()) is well-formed, and has well-defined behavior, and does not throw exceptions.

    […]

  3. Modify 20.3.1.3.4 [unique.ptr.single.asgn] as indicated:

    unique_ptr& operator=(unique_ptr&& u) noexcept;
    

    -1- Constraints: […]

    -2- Preconditions: If D is not a reference type, D meets the Cpp17MoveAssignable requirements (Table 30) and assignment of the deleter from an rvalue of type D does not throw an exception. Otherwise, D is a reference type; remove_reference_t<D> meets the Cpp17CopyAssignable requirements and assignment of the deleter from an lvalue of type D does not throw an exception.

    […]

    template<class U, class E> unique_ptr& operator=(unique_ptr<U, E>&& u) noexcept;
    

    -6- Constraints: […]

    -7- Preconditions: If E is not a reference type, assignment of the deleter from an rvalue of type E is well-formed and does not throw an exception. Otherwise, E is a reference type and assignment of the deleter from an lvalue of type E is well-formed and does not throw an exception.

    […]

  4. Modify 20.3.1.3.6 [unique.ptr.single.modifiers] as indicated:

    void reset(pointer p = pointer()) noexcept;
    

    -3- Preconditions: The expression get_deleter()(get()) is well-formed, and has well-defined behavior, and does not throw exceptions.

    […]

    void swap(unique_ptr& u) noexcept;
    

    -6- Preconditions: get_deleter() is swappable (16.4.4.3 [swappable.requirements]) and does not throw an exception under swap.

    […]


3596(i). ranges::starts_with and ranges::ends_with are underspecified

Section: 27.6.16 [alg.starts.with], 27.6.17 [alg.ends.with] Status: NAD Submitter: Michael Schellenberger Costa Opened: 2021-09-17 Last modified: 2021-09-24

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

When implementing C++23 new ranges::starts_with and ranges::ends_with for MSVC we came around some issues regarding infinite ranges.

In short, there is no possible answers to ends_with, if the second range is infinite. Similarly, there is no meaningful answer to starts_with if both ranges are infinite.

We should constraint both algorithms to forbid those corner cases (preferred resolution), or explicitly specify a given return value for those (The reasonable choice would be "return false;")

[2021-09-24 infinite ranges are invalid by [iterator.requirements.general] p10, so undefined by p12. Might be nice to relax this in future, but needs a paper. Status changed: New → NAD.]

Proposed resolution:


3611(i). Should compare_exchange be allowed to modify the expected value on success?

Section: 33.5.8.2 [atomics.types.operations] Status: NAD Submitter: Jonathan Wakely Opened: 2021-09-29 Last modified: 2022-08-23

Priority: 2

View all other issues in [atomics.types.operations].

View all issues with NAD status.

Discussion:

Currently a compare_exchange will only update the expected parameter if the compare_exchange fails. This precludes unconditionally clearing the padding bits of the expected object prior to doing the compare_exchange, which complicates the implementation by requiring additional work (e.g. making a copy of the expected value and clearing the copy's padding, then copying it back to expected only if the compare_exchange fails).

We should consider whether we want to allow modifications to expected in the success case, if such modifications only affect padding bits (i.e. do not alter the value). If we want to allow it, we need to say so explicitly. The current wording does not permit modifications in the success case, and any such modification could create a data race if another thread is reading the same memory location (if it knows a priori that a compare_exchange_strong would succeed and so not write to that location).

[2021-10-14; Reflector poll]

Set priority to 2 after reflector poll. Send to SG1.

This is 2426 again, but the new requirement to clear padding bits changes things.

[2022-07-06; Move to "Open" following SG1 feedback]

Allow compare_exchange to modify the expected value on success?

SFFNASA
00054

[2022-07-14; Move to "Tentatively NAD" following reflector poll]

[2022-08-23 Status changed: Tentatively NAD → NAD.]

Proposed resolution:


3652(i). Can we relax the preconditions of longjmp?

Section: 17.13.3 [csetjmp.syn] Status: NAD Submitter: Jiang An Opened: 2021-12-15 Last modified: 2022-01-30

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

IMO the UB in 17.13.3 [csetjmp.syn]/2 is because of the design that "every destructor call for an object shall be in effect when the lifetime of that object ends". A similar requirement is in 6.7.3 [basic.life]/5, but it's more relaxed than the requirement for setjmp/longjmp. It seems better to harmonize the requirements in 6.7.3 [basic.life]/5 and 17.13.3 [csetjmp.syn]/2, which allows longjmp to skip non-trivial but no-op destructor calls. Given that Microsoft UCRT's longjmp can sometimes (but not always) call destructors, IMO we should say that it's unspecified whether longjmp calls destructor for any object whose lifetime ends.

[2022-01-30 Duplicate of CWG 2361. Status changed: New → NAD.]

Proposed resolution:

This wording is relative to N4901.

  1. Modify 17.13.3 [csetjmp.syn] as indicated:

    -2- The function signature longjmp(jmp_buf jbuf, int val) has more restricted behavior in this document. A setjmp/longjmp call pair has undefined behavior iIf replacing the setjmp and longjmp in a setjmp/longjmp call pair by catch and throw would invoke any non-trivial destructors for any objects with automatic storage duration, it is unspecified whether the destructor is invoked for each such object. The behavior is undefined if the program depends on the side effects produced by any such unspecified destructor calls. A call to setjmp or longjmp has undefined behavior if invoked in a suspension context of a coroutine (7.6.2.4 [expr.await]).


3667(i). std::cout << &X::f prints 1

Section: 31.7.6.3 [ostream.formatted] Status: NAD Submitter: Peter Dimov Opened: 2022-01-31 Last modified: 2022-08-23

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

At present, the program

#include <iostream>

struct X
{
  void f() {}
};

int main()
{
  std::cout << &X::f;
}

prints 1. That's because member pointers implicitly convert to bool, and there's operator<< overload for bool in 31.7.6.3.2 [ostream.inserters.arithmetic].

This behavior is rarely useful. In C++20, we added deleted overloads to prevent a similar counter-intuitive output for the case in which e.g. L"str" is output to std::cout, which used to print the pointer value using the operator<< overload for const void*.

We should similarly consider adding a deleted overload for member pointers.

[2022-03-04; Reflector poll; Status changed: New → Tentatively NAD]

Needs a paper to LEWG if anything should change here.

[2022-08-23 Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4901.

  1. Modify 31.7.6.2.1 [ostream.general], class template basic_ostream synopsis, as indicated:

    namespace std {
      […]
      
      template<class traits>
        basic_ostream<wchar_t, traits>&
          operator<<(basic_ostream<wchar_t, traits>&, const char16_t*) = delete;
      template<class traits>
        basic_ostream<wchar_t, traits>&
          operator<<(basic_ostream<wchar_t, traits>&, const char32_t*) = delete;
          
      template<class charT, class traits, class T, class C>
        basic_ostream<charT, traits>& 
          operator<<(basic_ostream<charT, traits>&, T C::*) = delete;
    }
    

3695(i). The standard-layout property of char-like types serves for nothing

Section: 23.1 [strings.general] Status: NAD Submitter: Jiang An Opened: 2022-05-07 Last modified: 2022-08-23

Priority: 4

View all other issues in [strings.general].

View all issues with NAD status.

Discussion:

Currently char-like types are defined as non-array trivial and standard-layout types (23.1 [strings.general]). However, as far as I know, there is no actual operation specified in the standard or used in implementations requiring a char-like type to be standard-layout (unlike the triviality, which is related to charT(), etc.). Implementations are merely static_assert-ing the standard-layout property of element types.

As the standard library almost never requires a user-provided type to be standard-layout now (other suspicious exceptions are program-defined specializations of std::atomic<T*>), I think char-like types should just be non-array trivial types.

[2022-05-17; Reflector poll]

Set priority to 4 after reflector poll. Six votes for NAD. Changing this would require a proposal through LEWG.

[2022-08-23 Status changed: Tentatively NAD → NAD.]

Proposed resolution:


3726(i). reverse_iterator::operator-> is underconstrained for non-pointer iterators

Section: 25.5.1.6 [reverse.iter.elem] Status: NAD Submitter: Hewill Kang Opened: 2022-06-27 Last modified: 2022-11-30

Priority: Not Prioritized

View all other issues in [reverse.iter.elem].

View all issues with NAD status.

Discussion:

For non-pointer types, reverse_iterator::operator-> only requires the expression i.operator->() to be well-formed.

Since the return type of this function is explicitly specified as pointer, this will cause a hard error in the function body when the return type of i.operator->() cannot be converted to pointer.

We should add a return type constraint for this.

[2022-08-23; Reflector poll: NAD]

pointer is iterator_traits<Iterator>::pointer, which is required to name decltype(i.operator->()) (25.3.2.3 [iterator.traits]/1) so the postulated problem simply does not arise in valid code.

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4910.

  1. Modify 25.5.1.6 [reverse.iter.elem] as indicated:

    constexpr pointer operator->() const
      requires (is_pointer_v<Iterator> ||
                requires(const Iterator i) { { i.operator->() } -> convertible_to<pointer>; });
    

    -2- Effects:

    1. (2.1) — If Iterator is a pointer type, equivalent to: return prev(current);

    2. (2.2) — Otherwise, equivalent to: return prev(current).operator->();


3727(i). reverse_iterator/common_iterator's operator-> should not require the underlying iterator's operator-> to be a const member function

Section: 25.5.1.6 [reverse.iter.elem], 25.5.5.4 [common.iter.access] Status: NAD Submitter: Hewill Kang Opened: 2022-06-27 Last modified: 2022-11-30

Priority: Not Prioritized

View all other issues in [reverse.iter.elem].

View all issues with NAD status.

Discussion:

For non-pointer types, reverse_iterator::operator-> requires that the Iterator must have an operator->() with const-qualifier, whereas in the Effects clause, it always invokes the non-const object's operator->().

common_iterator::operator-> also requires that I::operator->() must be const-qualified, which seems reasonable since the return type of get<I>(v_) is const I&. However, LWG 3672 makes common_iterator::operator->() always return a value, which makes it unnecessary to detect the constness of I::operator->(), because it will be invoked with a non-const returned object anyway.

I think we should remove this constraint as I don't see the benefit of doing that. Constraining iterator's operator->() to be const and finally invoking non-const overload doesn't feel right to me either. In <ranges>, the exposition-only constraint has-arrow (26.5.2 [range.utility.helpers]) for operator->() does not require that the underlying iterator's operator->() to be const, we should make them consistent, and I believe this relaxation of constraints can bring some value.

Daniel:

This issue's second part of the resolution actually depends on 3672 being applied. But note that the reference wording below is still N4910.

[2022-08-23; Reflector poll: NAD]

Implicit variations apply to those requires-expressions, so calling as non-const (and rvalue) is fine. The PR actually loses that property and makes those overloads truly underconstrained. Motivation for relaxing it is vague. As for consistency, we should fix has-arrow instead.

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4910.

  1. Modify 25.5.1.6 [reverse.iter.elem] as indicated:

    constexpr pointer operator->() const
      requires (is_pointer_v<Iterator> ||
                requires(const Iterator i) { i.operator->(); });
    

    -2- Effects:

    1. (2.1) — If Iterator is a pointer type, equivalent to: return prev(current);

    2. (2.2) — Otherwise, equivalent to: return prev(current).operator->();

  2. Modify 25.5.5.4 [common.iter.access] as indicated:

    constexpr decltype(auto) operator->() const
      requires see below;
    

    -3- The expression in the requires-clause is equivalent to:

      indirectly_readable<const I> &&
      (requires(const I& i) { i.operator->(); } ||
      is_reference_v<iter_reference_t<I>> ||
      constructible_from<iter_value_t<I>, iter_reference_t<I>>)
      

    -4- Preconditions: holds_alternative<I>(v_) is true.

    -5- Effects:

    1. (5.1) — If I is a pointer type or if the expression get<I>(v_).operator->() is well-formedrequires(I i) { i.operator->(); } is true, equivalent to: return get<I>(v_);

    2. (5.2) — Otherwise, if iter_reference_t<I> is a reference type, equivalent to:

        auto&& tmp = *get<I>(v_);
        return addressof(tmp);
        
    3. (5.3) — Otherwise, equivalent to: return proxy(*get<I>(v_)); where proxy is the exposition-only class:

        class proxy {
          iter_value_t<I> keep_;
          constexpr proxy(iter_reference_t<I>&& x)
            : keep_(std::move(x)) {}
        public:
          constexpr const iter_value_t<I>* operator->() const noexcept {
            return addressof(keep_);
          }
        };
        

    […]


3735(i). views::adjacent<0> should be prohibited

Section: 26.7.26.1 [range.adjacent.overview], 26.7.27.1 [range.adjacent.transform.overview] Status: NAD Submitter: Hewill Kang Opened: 2022-07-13 Last modified: 2022-11-30

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

views::adjacent is very similar to views::slide, except that the window size N is given at compile time.

Since the case where N is 0 does not make sense for slide_view, LWG 3711 and LWG 3712 added preconditions to the constructor and removed the default constructor, respectively.

But for views::adjacent, we can still specify N to be 0. According to the description of 26.7.26.1 [range.adjacent.overview], it will return views::empty<tuple<>> as in the case of views::zip applied to an empty pack. And for views::adjacent_transform<0>(E, F), it will return views::zip_transform(F) and eventually return empty_view for some type.

This doesn't seem reasonable to me. The reason why views::zip can return views::empty<tuple<>> is that the parameter pack can indeed be empty, so this still makes some sense. However, there is no meaningful sense for the word "adjacent" when N is 0.

I don't see any observable value in allowing views::adjacent<0>, we should disable it for consistency with views::slide.

[2022-08-23; Reflector poll: NAD]

views::zip() is exactly as meaningful as views::adjacent<0>(E) - it's just the edge case.

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4910.

  1. Modify 26.7.26.1 [range.adjacent.overview] as indicated:

    -2- The name views::adjacent<N> denotes a range adaptor object (26.7.2 [range.adaptor.object]). Given a subexpression E and a constant expression N, the expression views::adjacent<N>(E) is expression-equivalent to:

    1. (2.1) — ((void)E, auto(views::empty<tuple<>>))Iif N is equal to 0, views::adjacent<N>(E) is ill-formed.

    2. (2.2) — Ootherwise, adjacent_view<views::all_t<decltype((E))>, N>(E).

  2. Modify 26.7.27.1 [range.adjacent.transform.overview] as indicated:

    -2- The name views::adjacent_transform<N> denotes a range adaptor object (26.7.2 [range.adaptor.object]). Given subexpressions E and F and a constant expression N:

    1. (2.1) — If N is equal to 0, views::adjacent_transform<N>(E, F) is ill-formedexpression-equivalent to ((void)E, views::zip_transform(F)), except that the evaluations of E and F are indeterminately sequenced.

    2. (2.2) — Otherwise, the expression views::adjacent_transform<N>(E, F) is expression-equivalent to adjacent_transform_view<views::all_t<decltype((E))>, decay_t<decltype((F))>, N>(E, F).


3739(i). chunk_view::size should preserve the signedness of the size of the underlying range

Section: 26.7.28.2 [range.chunk.view.input], 26.7.28.6 [range.chunk.view.fwd] Status: NAD Submitter: Hewill Kang Opened: 2022-07-15 Last modified: 2022-11-30

Priority: Not Prioritized

View all other issues in [range.chunk.view.input].

View all issues with NAD status.

Discussion:

Currently, the Effects of chunk_view::size simply returns to-unsigned-like(div-ceil(ranges::distance(base_), n_)), where div-ceil is defined in 26.7.28.2 [range.chunk.view.input] as:

template<class I>
constexpr I div-ceil(I num, I denom) { // exposition only
  I r = num / denom;
  if (num % denom)
    ++r;
  return r;
}

There are two problems here. First, for the const version of chunk_view::size, the types of ranges::distance(base_) and n_ are range_difference_t<const V> and range_difference_t<V> respectively, and the two parameters of div-ceil have the same type I. Given that the standard does not guarantee that V and const V must have the same difference_type, this makes the div-ceil's template deduction fail when the two are different.

Second, the standard does not guarantee that ranges::size must return an unsigned type, but here we use to-unsigned-like to unconditionally convert the return type of chunk_view::size to an unsigned type, which is inconsistent with the behavior of other range adaptors such as take_view and drop_view.

We should try to preserve the characteristics of the range_size_t of the underlying range as much as possible.

[2022-08-23; Reflector poll: NAD (would need a paper for LEWG)]

The "range_difference_t<const V> and range_difference_t<V> can be both valid but have different types" part is something I expect Mr. Carter's eventual paper to outlaw. The "preserve signedness" part is LEWG.

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4910.

  1. Modify 26.7.28.2 [range.chunk.view.input] as indicated:

    constexpr auto size() requires sized_range<V>;
    constexpr auto size() const requires sized_range<const V>;
    

    -5- Effects: Equivalent to:

    return to-unsigned-like(div-ceil(ranges::sizedistance(base_), static_cast<decltype(ranges::size(base_))>(n_));
    

  2. Modify 26.7.28.6 [range.chunk.view.fwd] as indicated:

    constexpr auto size() requires sized_range<V>;
    constexpr auto size() const requires sized_range<const V>;
    

    -3- Effects: Equivalent to:

    return to-unsigned-like(div-ceil(ranges::sizedistance(base_), static_cast<decltype(ranges::size(base_))>(n_));
    


3740(i). slide_view::size should preserve the signedness of underlying range's size

Section: 26.7.29.2 [range.slide.view] Status: NAD Submitter: Hewill Kang Opened: 2022-07-15 Last modified: 2022-11-30

Priority: Not Prioritized

View all other issues in [range.slide.view].

View all issues with NAD status.

Discussion:

Currently, slide_view::size const has the following Effects:

auto sz = ranges::distance(base_) - n_ + 1;
if (sz > 0) sz = 0;
return to-unsigned-like(sz);

There are two problems worth noting here. First, as described in LWG 3739, ranges::distance(base_) and n_ may have different types, which makes the actual type of sz not deterministic. Also, the return type is unconditionally converted to an unsigned type, even though the underlying range may have a signed size type.

Second, even if V has the same difference_type as const V, there may still be integer promotion issues mentioned by LWG 3730 since we add an integer 1 at the end here.

I think converting sz to the size type of the underlying range before returning is the appropriate thing to do.

[2022-08-23; Reflector poll: NAD]

Paper author: "I did consider promotion and decided not to care. The code compiles and conforms to all ranges requirements, and that seems entirely sufficient to me." "Even if we don't outlaw those being different types entirely, people playing those games will still get exactly one unsigned-integer-like type back. It's totally deterministic."

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4910.

  1. Modify 26.7.29.2 [range.slide.view] as indicated:

    constexpr auto size() requires sized_range<V>;
    constexpr auto size() const requires sized_range<const V>;
    

    -8- Effects: Equivalent to:

    auto sz = ranges::distance(base_) - n_ + 1;
    if (sz < 0) sz = 0;
    return static_cast<decltype(ranges::size(base_))>to-unsigned-like(sz);
    


3741(i). std::chrono::abs(duration) is ill-formed with non-reduced periods

Section: 29.5.10 [time.duration.alg] Status: NAD Submitter: Charlie Barto Opened: 2022-07-16 Last modified: 2022-11-30

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Currently 29.5.10 [time.duration.alg] specifies abs(duration) as:

Returns: if d >= d.zero(), return d, otherwise return -d.

Because unary minus on durations is defined to return common_type_t<duration>(-rep_), and common_type_t for durations is specified to reduce the period, this is ill-formed with durations such as duration<int, ratio<1000, 1000>>, or any other type where the numerator and denominator of the period are not coprime.

[2022-08-23; Reflector poll: NAD]

Not ill-formed, implementation should do a conversion. Changing it to return the reduced duration as an improvement would be for LEWG.

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4910.

  1. Modify 29.2 [time.syn], header <chrono> synopsis, as indicated:

    […]
    // 29.5.10 [time.duration.alg], specialized algorithms
    template<class Rep, class Period>
      constexpr common_type_t<duration<Rep, Period>> abs(duration<Rep, Period> d);
    […]
    
  2. Modify 29.5.10 [time.duration.alg] as indicated:

    [Drafting note: This will cause abs to reduce the period before returning it, much like the other arithmetic operators.

    This is not a breaking change, because code that was using abs with a non-reduced period before did not compile. ]

    template<class Rep, class Period>
      constexpr common_type_t<duration<Rep, Period>> abs(duration<Rep, Period> d);
    

    -1- Constraints: numeric_limits<Rep>::is_signed is true.

    -2- Returns: If d >= d.zero(), return +d, otherwise return -d.


3752(i). Should string::substr forward the allocator to the newly created string?

Section: 23.4.3.8.3 [string.substr] Status: NAD Submitter: Igor Zhukov Opened: 2022-08-10 Last modified: 2022-11-30

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Adrian Vogelsgesang noticed that libcxx's string::substr forward the allocator to the newly created string.

Nikolas Klauser found that MSVC STL does the same thing.

While Casey Carter and we all agree that this is a bug and we will fix it, we've "always" been nonconforming here and it's weird though that there was never an issue filed and two implementations have the exact same change.

So we want a clarification from LWG.

[2022-08-24; Reflector poll]

Set status to Tentatively NAD after reflector poll.

"If you want to provide a different allocator, you can use the substring constructor."

"This matches the constructor, whose choice of allocator has been already considered in 2402. Any changes here need a paper."

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:


3768(i). possibly-const-range is overconstrained

Section: 26.2 [ranges.syn] Status: NAD Submitter: Hewill Kang Opened: 2022-09-08 Last modified: 2022-11-30

Priority: Not Prioritized

View other active issues in [ranges.syn].

View all other issues in [ranges.syn].

View all issues with NAD status.

Discussion:

Due to the possibly-const-range constraint that the template parameter R must model input_range, this makes the ranges::cend that tries using it for meaningful const casting never be applied to an output_range, even though const_sentinel does not require the template parameter S to model input_iterator.

This is unnecessary, we should relax its constraint.

[2022-09-23; Reflector poll]

Tentatively NAD. "const output ranges don’t seem to be useful."

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4917.

  1. Modify 26.2 [ranges.syn] as indicated:

    #include <compare>              // see 17.11.1 [compare.syn]
    #include <initializer_list>     // see 17.10.2 [initializer.list.syn]
    #include <iterator>             // see 25.2 [iterator.synopsis]
    
    namespace std::ranges {
      […]
    
      // 26.7.21 [range.as.const], as const view
      template<input_range R>
        constexpr auto& possibly-const-range(R& r) {          // exposition only
          if constexpr (constant_range<const R> && !constant_range<R>) {
            return const_cast<const R&>(r);
          } else {
            return r;
          }
        } 
      
      […]
    }
    

3779(i). ranges::fold_* can unintentionally const_cast and reinterpret_cast

Section: 27.6.18 [alg.fold] Status: NAD Submitter: Nicole Mazzuca Opened: 2022-09-15 Last modified: 2022-11-30

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

In the Effects element of ranges::fold_right, we get the following code:

using U = decay_t<invoke_result_t<F&, iter_reference_t<I>, T>>;
if (first == last)
  return U(std::move(init)); // functional-style C cast
[…]

Given the following function object:

struct Second {
  static char* operator()(const char*, char* rhs) {
    return rhs;
  }
};

calling fold_right as:

char* p = fold_right(views::empty<char*>, "Hello", Second{});

initializes p with const_cast<char*>("Hello").

The same problem exists in fold_left_with_iter, and thus in fold_left.

One can get the reinterpret_cast behavior by replacing const char* with unsigned long long.

[2022-10-12; Reflector poll]

Set status to "Tentatively NAD" after reflector poll.

"The example doesn't compile. The accumulator should be be the second param, but with that fixed the constraints are not satisfied. The convertible_to constraint prevents the undesirable casting."

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4917.

  1. Modify 27.6.18 [alg.fold] as indicated:

    template<bidirectional_iterator I, sentinel_for<I> S, class T,
             indirectly-binary-right-foldable<T, I> F>
      constexpr auto ranges::fold_right(I first, S last, T init, F f);
    template<bidirectional_range R, class T,
             indirectly-binary-right-foldable<T, iterator_t<R>> F>
      constexpr auto ranges::fold_right(R&& r, T init, F f);
    

    -3- Effects: Equivalent to:

    using U = decay_t<invoke_result_t<F&, iter_reference_t<I>, T>>;
    if (first == last)
      return static_cast<U>(std::move(init));
    I tail = ranges::next(first, last);
    U accum = invoke(f, *--tail, std::move(init));
    while (first != tail)
      accum = invoke(f, *--tail, std::move(accum));
    return accum;
    
    […]
    template<input_iterator I, sentinel_for<I> S, class T,
             indirectly-binary-left-foldable<T, I> F>
      constexpr see below ranges::fold_left_with_iter(I first, S last, T init, F f);
    template<input_range R, class T, indirectly-binary-left-foldable<T, iterator_t<R>> F>
      constexpr see below ranges::fold_left_with_iter(R&& r, T init, F f);
    

    -6- Let U be decay_t<invoke_result_t<F&, T, iter_reference_t<I>>>.

    -7- Effects: Equivalent to:

    if (first == last)
      return {std::move(first), static_cast<U>(std::move(init))};
    U accum = invoke(f, std::move(init), *first);
    for (++first; first != last; ++first)
      accum = invoke(f, std::move(accum), *first);
    return {std::move(first), std::move(accum)};
    

3789(i). Precondition of (not replaced) operator delete[]

Section: 17.6.3.3 [new.delete.array] Status: NAD Submitter: blacktea hamburger Opened: 2022-09-25 Last modified: 2022-11-30

Priority: Not Prioritized

View all other issues in [new.delete.array].

View all issues with NAD status.

Discussion:

Consider (operator delete[](std::size_t) and operator new(std::size_t) is not replaced):

operator delete[](operator new(1));

(even not replaced) void* operator new(std::size_t) does not return void* operator new[](std::size_t). So the behavior is undefined according to 17.6.3.3 [new.delete.array] paragraph 9:

Preconditions: ptr is a null pointer or its value represents the address of a block of memory allocated by an earlier call to a (possibly replaced) operator new[](std::size_t) or operator new[](std::size_t, std::align_val_t) which has not been invalidated by an intervening call to operator delete[].

However, consider (operator delete(std::size_t) and operator new[](std::size_t) is not replaced):

operator delete(operator new[](1));

(not replaced) operator new[](std::size_t) simply returns operator new(std::size_t) according to 17.6.3.3 [new.delete.array] paragraph 4:

Default behavior: Returns operator new(size), or operator new(size, alignment), respectively.

So it is well-formed according to 17.6.3.2 [new.delete.single] paragraph 10:

Preconditions: ptr is a null pointer or its value represents the address of a block of memory allocated by an earlier call to a (possibly replaced) operator new(std::size_t) or operator new(std::size_t, std::align_val_t) which has not been invalidated by an intervening call to operator delete.

The behavior should be consistent.

[2022-10-10; Reflector poll]

Set status to "Tentatively NAD" after reflector poll.

"No reason to carve out an exception covering a case on something which can’t be observed by the program (whether the allocation operators are replaced). This just makes things more complicated for no good reason." "This would require changes to sanitizers and other dynamic analyzers, for zero practical benefit (except allowing bad code to go un-diagnosed)."

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4917.

  1. Modify 17.6.3.3 [new.delete.array] as indicated:

    void operator delete[](void* ptr) noexcept;
    void operator delete[](void* ptr, std::size_t size) noexcept;
    void operator delete[](void* ptr, std::align_val_t alignment) noexcept;
    void operator delete[](void* ptr, std::size_t size, std::align_val_t alignment) noexcept;
    

    -9- Preconditions: ptr is a null pointer or its value represents the address of a block of memory allocated by an earlier call to a (possibly replaced) operator new[](std::size_t), or operator new[](std::size_t, std::align_val_t), (not replaced) operator new(std::size_t), or operator new(std::size_t, std::align_val_t) which has not been invalidated by an intervening call to operator delete[].

    […]


3800(i). No deduction guide for std::match_results

Section: 32.9.1 [re.results.general] Status: NAD Submitter: Jonathan Wakely Opened: 2022-10-25 Last modified: 2022-11-30

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

std::match_results meets some of the requirements for an allocator-aware container, which means that its allocator_type must have the same value_type as the container. This means that class template argument deduction should work for:

std::match_results mr(alloc);

The allocator's value_type will be the sub_match<Iter> type stored in the match_results object, and so the first template argument for the match_results type will be value_type::iterator.

P0433R2 added a deduction guide for std::basic_regex, but I see no rationale for not adding one for std::match_results. This seems like a defect, because all other allocator-aware containers support deduction from an allocator argument.

[2022-11-01; reflector poll]

Status changed to Tentatively NAD. The issue is wrong: other containers do not support deduction from an allocator type.

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

Proposed resolution:

This wording is relative to N4917.

  1. Modify 32.9.1 [re.results.general], class template match_results synopsis, as indicated:

    namespace std {
      template<class BidirectionalIterator,
               class Allocator = allocator<sub_match<BidirectionalIterator>>>
        class match_results {
          […]
          void swap(match_results& that);
        };
        
      template<class Allocator>
        match_results(Allocator) -> match_results<typename Allocator::value_type::iterator, Allocator>;
    }
    

3806(i). Should concept formattable<T, charT> default to char?

Section: 22.14.6.2 [format.formattable] Status: NAD Submitter: Jonathan Wakely Opened: 2022-10-28 Last modified: 2023-01-24

Priority: 2

View other active issues in [format.formattable].

View all other issues in [format.formattable].

View all issues with NAD status.

Discussion:

For many uses of <format> there's no need to explicitly mention the output type as char. There are either typedefs for char specializations (format_context, format_parse_context, format_args, etc.) or a default template argument (formatter, range_formatter). But for the new formattable concept you always need to specify the character type:

static_assert( std::formattable<int> ); // ill-formed
static_assert( std::formattable<int, char> ); // OK

Should the concept have a default template argument for the second parameter, to make it easier to check whether something is formattable as char?

[2022-11-01; Reflector poll]

Set priority to 2 after reflector poll. Two votes for NAD, the convenience makes it easier to misuse.

[2022-11-30; Reflector poll]

Set status to "Tentatively NAD" after ten votes in reflector poll.

[2022-11-30 LWG telecon. Status changed: Tentatively NAD → NAD.]

[2023-01-24 LEWG electronic poll; weak consensus to reject the propsed change.]

Proposed resolution:

This wording is relative to N4917.

  1. Modify 22.14.1 [format.syn] as indicated:

    […]
    // 22.14.6 [format.formatter], formatter
    template<class T, class charT = char> struct formatter;
    
    // 22.14.6.2 [format.formattable], concept formattable
    template<class T, class charT = char>
      concept formattable = see below;
    […]
    
  2. Modify 22.14.6.2 [format.formattable] as indicated:

    [Drafting note: This repeats the default template argument already shown in the synopsis, which would not be valid in C++ code. That is consistent with our presentation style though, as this is not C++ code, it's a specification. See e.g. indirect_strict_weak_order and subrange.]

    -1- Let fmt-iter-for<charT> be an unspecified type that models output_iterator<const charT&> (25.3.4.10 [iterator.concept.output]).

    template<class T, class charT = char>
      concept formattable =
        semiregular<formatter<remove_cvref_t<T>, charT>> &&
        requires(formatter<remove_cvref_t<T>, charT> f,
                 const formatter<remove_cvref_t<T>, charT> cf,
                 T t,
                 basic_format_context<fmt-iter-for<charT>, charT> fc,
                 basic_format_parse_context<charT> pc) {
          { f.parse(pc) } -> same_as<basic_format_parse_context<charT>::iterator>;
          { cf.format(t, fc) } -> same_as<fmt-iter-for<charT>>;
        };
    

3808(i). Inconsistent feature test macros for ranges algorithms

Section: 17.3.2 [version.syn] Status: NAD Submitter: Daniel Marshall Opened: 2022-11-02 Last modified: 2022-11-12

Priority: Not Prioritized

View other active issues in [version.syn].

View all other issues in [version.syn].

View all issues with NAD status.

Discussion:

ranges::shift is seemingly the only ranges algorithm that incremented the value of an existing feature test macro.

From what I can tell the only other one that could have done that but didn't is __cpp_lib_starts_ends_with.

Should there have been a __cpp_lib_ranges_shift or should have __cpp_lib_starts_ends_with incremented its value when the ranges algorithms were added?

[Kona 2022-11-12 Status changed: New → NAD.]

Proposed resolution:


3858(i). basic_const_iterator is too strict to provide iterator_category

Section: 25.5.3.4 [const.iterators.types] Status: NAD Submitter: Hewill Kang Opened: 2023-01-22 Last modified: 2023-02-06

Priority: Not Prioritized

View all issues with NAD status.

Discussion:

Currently, basic_const_iterator provides iterator_category member only when the underlying iterator models forward_iterator, which means that it will not provide input_iterator_tag when applied to any C++17 legacy iterator.

This seems undesirable, and it should be more appropriate to provide iterator_category only by detecting whether the iterator_traits specialization of the underlying iterator has a valid iterator_category, as is common practice for other iterator adaptors in <iterator>.

[2023-02-06 Status changed: New → NAD.]

Set status to NAD after reflector poll.

Proposed resolution:

This wording is relative to N4928.

  1. Modify 25.5.3.4 [const.iterators.types] as indicated:

    -2- The member typedef-name iterator_category is defined if and only if the qualified-id iterator_traits<Iterator>::iterator_category is valid and denotes a typeIterator models forward_iterator. In that case, basic_const_iterator<Iterator>::iterator_category denotes the type iterator_traits<Iterator>::iterator_category.


3874(i). Rename __cpp_lib_ranges_to_container to __cpp_lib_ranges_to

Section: 17.3.2 [version.syn] Status: NAD Submitter: Hewill Kang Opened: 2023-02-05 Last modified: 2023-02-08

Priority: Not Prioritized

View other active issues in [version.syn].

View all other issues in [version.syn].

View all issues with NAD status.

Discussion:

LWG 3785 enables ranges::to to construct a non-range object from a range, confirming that this is the intended part of the original design.

However, the feature test macro for ranges::to is still named __cpp_lib_ranges_to_container, where the "container" part seems confusing to users.

[Issaquah 2023-02-08 Status changed: New → NAD.]

No consensus to change anything. The macro name is just a unique identifier, it doesn't have to be a precise description of the feature.

Proposed resolution:

This wording is relative to N4928.

  1. Modify 17.3.2 [version.syn], header <version> synopsis, as indicated:

    […]
    #define __cpp_lib_ranges_stride                     202207L // also in <ranges>
    #define __cpp_lib_ranges_to_container               202202L // also in <ranges>
    #define __cpp_lib_ranges_zip                        202110L // also in <ranges>, <tuple>, <utility>
    […]