| Doc. no. | J16/02-0027 = WG21 N1369 |
| Date: | 10 May 2002 |
| Project: | Programming Language C++ |
| Reply to: | Matt Austern <austern@research.att.com> |
Reference ISO/IEC IS 14882:1998(E)
Also see:
The purpose of this document is to record the status of issues which have come before the Library Working Group (LWG) of the ANSI (J16) and ISO (WG21) C++ Standards Committee. Issues represent potential defects in the ISO/IEC IS 14882:1998(E) document. Issues are not to be used to request new features or other extensions.
This document contains only library issues which are actively being considered by the Library Working Group. That is, issues which have a status of New, Open, Ready, and Review. See Library Defect Reports List for issues considered defects and Library Closed Issues List for issues considered closed.
The issues in these lists are not necessarily formal ISO Defect Reports (DR's). While some issues will eventually be elevated to official Defect Report status, other issues will be disposed of in other ways. See Issue Status.
This document is in an experimental format designed for both viewing via a world-wide web browser and hard-copy printing. It is available as an HTML file for browsing or PDF file for printing.
Prior to Revision 14, library issues lists existed in two slightly different versions; a Committee Version and a Public Version. Beginning with Revision 14 the two versions were combined into a single version.
This document includes [bracketed italicized notes] as a reminder to the LWG of current progress on issues. Such notes are strictly unofficial and should be read with caution as they may be incomplete or incorrect. Be aware that LWG support for a particular resolution can quickly change if new viewpoints or killer examples are presented in subsequent discussions.
For the most current official version of this document see http://www.dkuug.dk/jtc1/sc22/wg21. Requests for further information about this document should include the document number above, reference ISO/IEC 14882:1998(E), and be submitted to Information Technology Industry Council (ITI), 1250 Eye Street NW, Washington, DC 20005.
Public information as to how to obtain a copy of the C++ Standard, join the standards committee, submit an issue, or comment on an issue can be found in the C++ FAQ at http://www.research.att.com/~austern/csc/faq.html. Public discussion of C++ Standard related issues occurs on news:comp.std.c++.
For committee members, files available on the committee's private web site include the HTML version of the Standard itself. HTML hyperlinks from this issues list to those files will only work for committee members who have downloaded them into the same disk directory as the issues list files.
New - The issue has not yet been reviewed by the LWG. Any Proposed Resolution is purely a suggestion from the issue submitter, and should not be construed as the view of LWG.
Open - The LWG has discussed the issue but is not yet ready to move the issue forward. There are several possible reasons for open status:
A Proposed Resolution for an open issue is still not be construed as the view of LWG. Comments on the current state of discussions are often given at the end of open issues in an italic font. Such comments are for information only and should not be given undue importance.
Dup - The LWG has reached consensus that the issue is a duplicate of another issue, and will not be further dealt with. A Rationale identifies the duplicated issue's issue number.
NAD - The LWG has reached consensus that the issue is not a defect in the Standard, and the issue is ready to forward to the full committee as a proposed record of response. A Rationale discusses the LWG's reasoning.
Review - Exact wording of a Proposed Resolution is now available for review on an issue for which the LWG previously reached informal consensus.
Ready - The LWG has reached consensus that the issue is a defect in the Standard, the Proposed Resolution is correct, and the issue is ready to forward to the full committee for further action as a Defect Report (DR).
DR - (Defect Report) - The full J16 committee has voted to forward the issue to the Project Editor to be processed as a Potential Defect Report. The Project Editor reviews the issue, and then forwards it to the WG21 Convenor, who returns it to the full committee for final disposition. This issues list accords the status of DR to all these Defect Reports regardless of where they are in that process.
TC - (Technical Corrigenda) - The full WG21 committee has voted to accept the Defect Report's Proposed Resolution as a Technical Corrigenda. Action on this issue is thus complete and no further action is possible under ISO rules.
RR - (Record of Response) - The full WG21 committee has determined that this issue is not a defect in the Standard. Action on this issue is thus complete and no further action is possible under ISO rules.
Future - In addition to the regular status, the LWG believes that this issue should be revisited at the next revision of the standard. It is usually paired with NAD.
Issues are always given the status of New when they first appear on the issues list. They may progress to Open or Review while the LWG is actively working on them. When the LWG has reached consensus on the disposition of an issue, the status will then change to Dup, NAD, or Ready as appropriate. Once the full J16 committee votes to forward Ready issues to the Project Editor, they are given the status of Defect Report ( DR). These in turn may become the basis for Technical Corrigenda (TC), or are closed without action other than a Record of Response (RR ). The intent of this LWG process is that only issues which are truly defects in the Standard move to the formal ISO DR status.
Section: 22.2.2.1.2 [lib.facet.num.get.virtuals] Status: Open Submitter: Nathan Myers Date: 6 Aug 1998
The current description of numeric input does not account for the possibility of overflow. This is an implicit result of changing the description to rely on the definition of scanf() (which fails to report overflow), and conflicts with the documented behavior of traditional and current implementations.
Users expect, when reading a character sequence that results in a value unrepresentable in the specified type, to have an error reported. The standard as written does not permit this.
Further comments from Dietmar:
I don't feel comfortable with the proposed resolution to issue 23: It kind of simplifies the issue to much. Here is what is going on:
Currently, the behavior of numeric overflow is rather counter intuitive and hard to trace, so I will describe it briefly:
Further discussion from Redmond:
The basic problem is that we've defined our behavior, including our error-reporting behavior, in terms of C90. However, C90's method of reporting overflow in scanf is not technically an "input error". The strto_* functions are more precise.
There was general consensus that failbit should be set upon overflow. We considered three options based on this:
Straw poll: (1) 5; (2) 0; (3) 8.
PJP will provide wording.
Proposed resolution:
Section: 27 [lib.input.output] Status: Open Submitter: Nathan Myers Date: 6 Aug 1998
Many of the specifications for iostreams specify that character values or their int_type equivalents are compared using operators == or !=, though in other places traits::eq() or traits::eq_int_type is specified to be used throughout. This is an inconsistency; we should change uses of == and != to use the traits members instead.
Proposed resolution:
[Kona: Nathan to supply proposed wording]
[ Tokyo: the LWG reaffirmed that this is a defect, and requires careful review of clause 27 as the changes are context sensitive. ]
Section: 21.3.7.9 [lib.string.io] Status: Ready Submitter: Nico Josuttis Date: 29 Sep 1998
Operator >> and getline() for strings read until eof() in the input stream is true. However, this might never happen, if the stream can't read anymore without reaching EOF. So shouldn't it be changed into that it reads until !good() ?
Proposed resolution:
In 21.3.7.9 [lib.string.io], paragraph 1, replace:
Effects: Begins by constructing a sentry object k as if k were constructed by typename basic_istream<charT,traits>::sentry k( is). If bool( k) is true, it calls str.erase() and then extracts characters from is and appends them to str as if by calling str.append(1, c). If is.width() is greater than zero, the maximum number n of characters appended is is.width(); otherwise n is str.max_size(). Characters are extracted and appended until any of the following occurs:
with:
Effects: Behaves as a formatted input function (27.6.1.2.1 [lib.istream.formatted.reqmts]). After constructing a sentry object, if the sentry converts to true, calls str.erase() and then extracts characters from is and appends them to str as if by calling str.append(1,c). If is.width() is greater than zero, the maximum number n of characters appended is is.width(); otherwise n is str.max_size(). Characters are extracted and appended until any of the following occurs:
In 21.3.7.9 [lib.string.io], paragraph 6, replace
Effects: Begins by constructing a sentry object k as if by typename basic_istream<charT,traits>::sentry k( is, true). If bool( k) is true, it calls str.erase() and then extracts characters from is and appends them to str as if by calling str.append(1, c) until any of the following occurs:
with:
Effects: Behaves as an unformatted input function (27.6.1.3 [lib.istream.unformatted]), except that it does not affect the value returned by subsequent calls to basic_istream<>::gcount(). After constructing a sentry object, if the sentry converts to true, calls str.erase() and then extracts characters from is and appends them to str as if by calling str.append(1,c) until any of the following occurs:
[Redmond: Made changes in proposed resolution. operator>> should be a formatted input function, not an unformatted input function. getline should not be required to set gcount, since there is no mechanism for gcount to be set except by one of basic_istream's member functions.]
[Curaçao: Nico agrees with proposed resolution.]
Rationale:
The real issue here is whether or not these string input functions get their characters from a streambuf, rather than by calling an istream's member functions, a streambuf signals failure either by returning eof or by throwing an exception; there are no other possibilities. The proposed resolution makes it clear that these two functions do get characters from a streambuf.
Section: 25 [lib.algorithms] Status: Open Submitter: Nico Josuttis Date: 29 Sep 1998
The standard does not state, how often a function object is copied, called, or the order of calls inside an algorithm. This may lead to surprising/buggy behavior. Consider the following example:
class Nth { // function object that returns true for the nth element
private:
int nth; // element to return true for
int count; // element counter
public:
Nth (int n) : nth(n), count(0) {
}
bool operator() (int) {
return ++count == nth;
}
};
....
// remove third element
list<int>::iterator pos;
pos = remove_if(coll.begin(),coll.end(), // range
Nth(3)), // remove criterion
coll.erase(pos,coll.end());
This call, in fact removes the 3rd AND the 6th element. This happens because the usual implementation of the algorithm copies the function object internally:
template <class ForwIter, class Predicate>
ForwIter std::remove_if(ForwIter beg, ForwIter end, Predicate op)
{
beg = find_if(beg, end, op);
if (beg == end) {
return beg;
}
else {
ForwIter next = beg;
return remove_copy_if(++next, end, beg, op);
}
}
The algorithm uses find_if() to find the first element that should be removed. However, it then uses a copy of the passed function object to process the resulting elements (if any). Here, Nth is used again and removes also the sixth element. This behavior compromises the advantage of function objects being able to have a state. Without any cost it could be avoided (just implement it directly instead of calling find_if()).
Proposed resolution:
In [lib.function.objects] 20.3 Function objects add as new paragraph 6 (or insert after paragraph 1):
Option 1:
Predicates are functions or function objects that fulfill the following requirements:
- They return a Boolean value (bool or a value convertible to bool)
- It doesn't matter for the behavior of a predicate how often it is copied or assigned and how often it is called.
Option 2:
- if it's a function:
- All calls with the same argument values yield the same result.
- if it's a function object:
- In any sequence of calls to operator () without calling any non-constant member function, all calls with the same argument values yield the same result.
- After an assignment or copy both objects return the same result for the same values.
[Santa Cruz: The LWG believes that there may be more to this than meets the eye. It applies to all function objects, particularly predicates. Two questions: (1) must a function object be copyable? (2) how many times is a function object called? These are in effect questions about state. Function objects appear to require special copy semantics to make state work, and may fail if calling alters state and calling occurs an unexpected number of times.]
[Dublin: Pete Becker felt that this may not be a defect, but rather something that programmers need to be educated about. There was discussion of adding wording to the effect that the number and order of calls to function objects, including predicates, not affect the behavior of the function object.]
[Pre-Kona: Nico comments: It seems the problem is that we don't have a clear statement of "predicate" in the standard. People including me seemed to think "a function returning a Boolean value and being able to be called by an STL algorithm or be used as sorting criterion or ... is a predicate". But a predicate has more requirements: It should never change its behavior due to a call or being copied. IMHO we have to state this in the standard. If you like, see section 8.1.4 of my library book for a detailed discussion.]
[Kona: Nico will provide wording to the effect that "unless otherwise specified, the number of copies of and calls to function objects by algorithms is unspecified". Consider placing in 25 [lib.algorithms] after paragraph 9.]
[Pre-Tokyo: Angelika Langer comments: if the resolution is that algorithms are free to copy and pass around any function objects, then it is a valid question whether they are also allowed to change the type information from reference type to value type.]
[Tokyo: Nico will discuss this further with Matt as there are multiple problems beyond the underlying problem of no definition of "Predicate".]
[Post-Tokyo: Nico provided the above proposed resolutions.]
[Curaçao: Nico will provide wording to make options clearer: are the exclusive, or is one a superset of the other?]
Section: 23.2.5 [lib.vector.bool] Status: Open Submitter: AFNOR Date: 7 Oct 1998
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.
Proposed resolution:
[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.]
[Tokyo: Not discussed by the full LWG; no one claimed new insights and so time was more productively spent on other issues. In private discussions it was asserted that requirements for any solution include 1) Increasing the full committee's understanding of the problem, and 2) providing compiler vendors, authors, teachers, and of course users with specific suggestions as to how to apply the eventual solution.]
Section: 24.1.1 [lib.input.iterators] Status: Open Submitter: AFNOR Date: 7 Oct 1998
Table 72 in 24.1.1 [lib.input.iterators] specifies semantics for *r++ of:
{ T tmp = *r; ++r; return tmp; }
There are two problems with this. First, the return type is specified to be "T", as opposed to something like "convertible to T". This is too specific: we want to allow *r++ to return an lvalue.
Second, writing the semantics in terms of code misleadingly suggests that the effects *r++ should precisely replicate the behavior of this code, including side effects. (What if it's a user-defined type whose copy constructor has observable behavior?) We should replace the code with words, or else put some blanket statement in clause 17 saying that code samples aren't intended to specify exactly how many times a copy constructor is called, even if the copy constructor has observable behavior. (See issue 334 for a similar problem.)
[Issue still isn't clear. Matt will try to explain it more clearly at the next meeting.]
Proposed resolution:
Section: 17.4.3.1 [lib.reserved.names] Status: Open Submitter: Judy Ward Date: 15 Dec 1998
The original issue asked whether a library implementor could specialize standard library templates for built-in types. (This was an issue because users are permitted to explicitly instantiate standard library templates.)
Specializations are no longer a problem, because of the resolution to core issue 259. Under the proposed resolution, it will be legal for a translation unit to contain both a specialization and an explicit instantiation of the same template, provided that the specialization comes first. In such a case, the explicit instantiation will be ignored. Further discussion of library issue 120 assumes that the core 259 resolution will be adopted.
However, as noted in lib-7047, one piece of this issue still remains: what happens if a standard library implementor explicitly instantiates a standard library templates? It's illegal for a program to contain two different explicit instantiations of the same template for the same type in two different translation units (ODR violation), and the core working group doesn't believe it is practical to relax that restriction.
The issue, then, is: are users allowed to implicitly instantiate standard library templates for non-user defined types? The status quo answer is 'yes'. Changing it to 'no' would give library implementors more freedom.
This is an issue because, for performance reasons, library implementors often need to explicitly instantiate standard library templates. (for example, std::basic_string<char>) Does giving users freedom to explicitly instantiate standard library templates for non-user defined types make it impossible or painfully difficult for library implementors to do this?
John Spicer suggests, in lib-8957, that library implementors have a mechanism they can use for explicit instantiations that doesn't prevent users from performing their own explicit instantiations: put each explicit instantiation in its own object file. (Different solutions might be necessary for Unix DSOs or MS-Windows DLLs.) On some platforms, library implementors might not need to do anything special: the "undefined behavior" that results from having two different explicit instantiations might be harmless.
Proposed resolution:
Option 1.
Append to 17.4.3.1 [lib.reserved.names] paragraph 1:
A program may explicitly instantiate any templates in the standard library only if the declaration depends on a user-defined name of external linkage and the instantiation meets the standard library requirements for the original template.
Option 2.
In light of the resolution to core issue 259, no normative changes in the library clauses are necessary. Add the following non-normative note to the end of 17.4.3.1 [lib.reserved.names] paragraph 1:
[Note: A program may explicitly instantiate standard library templates, even when an explicit instantiation does not depend on a user-defined name. --end note]
[Copenhagen: LWG discussed three options. (1) Users may not explicitly instantiate standard library templates, except on user-defined types. Consequence: library implementors may freely specialize or instantiate templates. (2) Users may explicitly instantiate any standard library template. Consequence: if implementors specialize or instantiate library templates, they may need to take special steps to make sure users can do it too. (3) It is implementation defined whether users may explicitly instantiate standard library templates on non-user-defined types. Consequence: library implementors may freely specialize or instantiate templates, but may need to document some or all templates that have been explicitly instantiated. ]
[Straw poll (first number is favor, second is strongly oppose): 1 - 4, 0; 2 - 9, 1; 3 - 0, 9. (Proposed resolution 1 was the original proposed resolution.) Because there was no support for option 3, no wording is provided.]
[Redmond: discussed again; straw poll had results similar to those of Copenhagen (1 - 1, 3; 2 - 8, 4; 3 - 6, 2). Most people said they could live with any option. The only objection to option 2 is potential implementation difficulty. Steve Clamage volunteered do a survey to see if there are any popular platforms where option 2 would present a real problem for implementors. See his reflector message, c++std-lib-9002. ]
[Steve and Pete Becker will talk to Jonathan Caves. The Microsoft linker might present a problem if there are multiple copies, some of which have static data and some of which are in DLLs. There may be similar problems with the Apple linker; Matt will clarify that.]
Section: 26.3.5.4 [lib.slice.arr.fill], 26.3.7.4 [lib.gslice.array.fill], 26.3.8.4 [lib.mask.array.fill], 26.3.9.4 [lib.indirect.array.fill] Status: Ready Submitter: Judy Ward Date: 15 Dec 1998
One of the operator= in the valarray helper arrays is const and one is not. For example, look at slice_array. This operator= in Section 26.3.5.2 [lib.slice.arr.assign] is const:
void operator=(const valarray<T>&) const;
but this one in Section 26.3.5.4 [lib.slice.arr.fill] is not:
void operator=(const T&);
The description of the semantics for these two functions is similar.
Proposed resolution:
26.3.5 [lib.template.slice.array] Template class slice_array
In the class template definition for slice_array, replace the member function declaration
void operator=(const T&);with
void operator=(const T&) const;
26.3.5.4 [lib.slice.arr.fill] slice_array fill function
Change the function declaration
void operator=(const T&);to
void operator=(const T&) const;
26.3.7 [lib.template.gslice.array] Template class gslice_array
In the class template definition for gslice_array, replace the member function declaration
void operator=(const T&);with
void operator=(const T&) const;
26.3.7.4 [lib.gslice.array.fill] gslice_array fill function
Change the function declaration
void operator=(const T&);to
void operator=(const T&) const;
26.3.8 [lib.template.mask.array] Template class mask_array
In the class template definition for mask_array, replace the member function declaration
void operator=(const T&);with
void operator=(const T&) const;
26.3.8.4 [lib.mask.array.fill] mask_array fill function
Change the function declaration
void operator=(const T&);to
void operator=(const T&) const;
26.3.9 [lib.template.indirect.array] Template class indirect_array
In the class template definition for indirect_array, replace the member function declaration
void operator=(const T&);with
void operator=(const T&) const;
26.3.9.4 [lib.indirect.array.fill] indirect_array fill function
Change the function declaration
void operator=(const T&);to
void operator=(const T&) const;
[Redmond: Robert provided wording.]
Rationale:
There's no good reason for one version of operator= being const and another one not. Because of issue 253, this now matters: these functions are now callable in more circumstances. In many existing implementations, both versions are already const.
Section: 27.6.2.5.4 [lib.ostream.inserters.character] Status: Open Submitter: Dietmar Kühl Date: 20 Jul 1999
Paragraph 4 states that the length is determined using traits::length(s). Unfortunately, this function is not defined for example if the character type is wchar_t and the type of s is char const*. Similar problems exist if the character type is char and the type of s is either signed char const* or unsigned char const*.
Proposed resolution:
Change 27.6.2.5.4 [lib.ostream.inserters.character] paragraph 4 from:
Effects: Behaves like an formatted inserter (as described in lib.ostream.formatted.reqmts) of out. After a sentry object is constructed it inserts characters. The number of characters starting at s to be inserted is traits::length(s). Padding is determined as described in lib.facet.num.put.virtuals. The traits::length(s) characters starting at s are widened using out.widen (lib.basic.ios.members). The widened characters and any required padding are inserted into out. Calls width(0).
to:
Effects: Behaves like an formatted inserter (as described in lib.ostream.formatted.reqmts) of out. After a sentry object is constructed it inserts characters. The number len of characters starting at s to be inserted is
- traits::length((const char*)s) if the second argument is of type const charT*
- char_traits<char>::length(s) if the second argument is of type const char*, const signed char*, or const unsigned char* and and charT is not char.
Padding is determined as described in lib.facet.num.put.virtuals. The len characters starting at s are widened using out.widen (lib.basic.ios.members). The widened characters and any required padding are inserted into out. Calls width(0).
[Kona: It is clear to the LWG there is a defect here. Dietmar will supply specific wording.]
[Post-Tokyo: Dietmar supplied the above wording.]
[Toronto: The original proposed resolution involved char_traits<signed char> and char_traits<unsigned char>. There was strong opposition to requiring that library implementors provide those specializations of char_traits.]
[Copenhagen: This still isn't quite right: proposed resolution text got garbled when the signed char/unsigned char specializations were removed. Dietmar will provide revised wording.]
Section: 23.1 [lib.container.requirements] Status: Ready Submitter: Judy Ward Date: 2 Jul 1998
Currently the following will not compile on two well-known standard library implementations:
#include <set>
using namespace std;
void f(const set<int> &s)
{
set<int>::iterator i;
if (i==s.end()); // s.end() returns a const_iterator
}
The reason this doesn't compile is because operator== was implemented as a member function of the nested classes set:iterator and set::const_iterator, and there is no conversion from const_iterator to iterator. Surprisingly, (s.end() == i) does work, though, because of the conversion from iterator to const_iterator.
I don't see a requirement anywhere in the standard that this must work. Should there be one? If so, I think the requirement would need to be added to the tables in section 24.1.1. I'm not sure about the wording. If this requirement existed in the standard, I would think that implementors would have to make the comparison operators non-member functions.
This issues was also raised on comp.std.c++ by Darin Adler. The example given was:
bool check_equal(std::deque<int>::iterator i,
std::deque<int>::const_iterator ci)
{
return i == ci;
}
Comment from John Potter:
In case nobody has noticed, accepting it will break reverse_iterator.
The fix is to make the comparison operators templated on two types.
template <class Iterator1, class Iterator2> bool operator== (reverse_iterator<Iterator1> const& x, reverse_iterator<Iterator2> const& y);Obviously: return x.base() == y.base();
Currently, no reverse_iterator to const_reverse_iterator compares are valid.
BTW, I think the issue is in support of bad code. Compares should be between two iterators of the same type. All std::algorithms require the begin and end iterators to be of the same type.
Proposed resolution:
Insert this paragraph after 23.1 [lib.container.requirements] paragraph 7:
In the expressions
i == j i != j i < j i <= j i >= j i > j i - jWhere i and j denote objects of a container's iterator type, either or both may be replaced by an object of the container's const_iterator type referring to the same element with no change in semantics.
[post-Toronto: Judy supplied a proposed resolution saying that iterator and const_iterator could be freely mixed in iterator comparison and difference operations.]
[Redmond: Dave and Howard supplied a new proposed resolution which explicitly listed expressions; there was concern that the previous proposed resolution was too informal.]
Rationale:
The LWG believes it is clear that the above wording applies only to the nested types X::iterator and X::const_iterator, where X is a container. There is no requirement that X::reverse_iterator and X::const_reverse_iterator can be mixed. If mixing them is considered important, that's a separate issue. (Issue 280.)
Section: 25.2.2 [lib.alg.swap] Status: Open Submitter: Andrew Koenig Date: 14 Aug 1999
The description of iter_swap in 25.2.2 paragraph 7,says that it ``exchanges the values''
of the objects to which two iterators refer.
What it doesn't say is whether it does so using swap or using the assignment operator and copy constructor.
This question is an important one to answer, because swap is specialized to work efficiently for standard containers.
For example:
vector<int> v1, v2; iter_swap(&v1, &v2);
Is this call to iter_swap equivalent to calling swap(v1, v2)? Or is it equivalent to
{
vector<int> temp = v1;
v1 = v2;
v2 = temp;
}
The first alternative is O(1); the second is O(n).
A LWG member, Dave Abrahams, comments:
Not an objection necessarily, but I want to point out the cost of that requirement:
iter_swap(list<T>::iterator, list<T>::iterator)
can currently be specialized to be more efficient than iter_swap(T*,T*) for many T (by using splicing). Your proposal would make that optimization illegal.
[Kona: The LWG notes the original need for iter_swap was proxy iterators which are no longer permitted.]
Proposed resolution:
Change the effect clause of iter_swap in 25.2.2 paragraph 7 from:
Exchanges the values pointed to by the two iterators a and b.
to
swap(*a, *b).
[post-Toronto: The LWG is concerned about possible overspecification: there may be cases, such as Dave Abrahams's example above, and such as vector<bool>'s iterators, where it makes more sense for iter_swap to do something other than swap. If performance is a concern, it may be better to have explicit complexity requirements than to say how iter_swap should be implemented.]
[Redmond: Discussed, with no consensus. There was very little support for the proposed resolution. Some people favored closing this issue as NAD. Others favored a more complicated specification of iter_swap, which might distinguish between ordinary iterators and proxies. A possible new issue: how do we know that the iterators passed to iter_swap have Assignable value types? (If this new issue is real, it extends far beyond just iter_swap.)]
Section: 20.1.5 [lib.allocator.requirements], 23.1 [lib.container.requirements] Status: Open Submitter: Andy Sawyer Date: 21 Oct 1999
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:
Change 20.1.5 [lib.allocator.requirements] table 32 max_size() wording from:
the largest value that can meaningfully be
passed to X::allocate
to:
the value of the largest constant expression
(5.19 [expr.const]) that could ever meaningfully be passed to X::allocate
Change 23.1 [lib.container.requirements] table 65 max_size() wording from:
size() of the largest possible container.
to:
the value of the largest constant expression
(5.19 [expr.const]) that could ever meaningfully be returned by X::size().
[Kona: The LWG informally discussed this and asked Andy Sawyer to submit an issue.]
[Tokyo: The LWG believes (1) above is the intended meaning.]
[Post-Tokyo: Beman Dawes supplied the above resolution at the request of the LWG. 21.3.3 [lib.string.capacity] was not changed because it references max_size() in 23.1. The term "compile-time" was avoided because it is not defined anywhere in the standard (even though it is used several places in the library clauses).]
[Copenhagen: Exactly what max_size means is still unclear. It may have a different meaning as a container member function than as an allocator member function. For the latter, it is probably best thought of as an architectural limit. Nathan will provide new wording.]
Section: 24.1.3 [lib.forward.iterators] Status: Ready Submitter: Matt Austern Date: 19 Nov 1999
In table 74, the return type of the expression *a is given as T&, where T is the iterator's value type. For constant iterators, however, this is wrong. ("Value type" is never defined very precisely, but it is clear that the value type of, say, std::list<int>::const_iterator is supposed to be int, not const int.)
Proposed resolution:
In table 74, in the *a and *r++ rows, change the return type from "T&" to "T& if X is mutable, otherwise const T&". In the a->m row, change the return type from "U&" to "U& if X is mutable, otherwise const U&".
[Tokyo: The LWG believes this is the tip of a larger iceberg; there are multiple const problems with the STL portion of the library and that these should be addressed as a single package. Note that issue 180 has already been declared NAD Future for that very reason.]
[Redmond: the LWG thinks this is separable from other constness issues. This issue is just cleanup; it clarifies language that was written before we had iterator_traits. Proposed resolution was modified: the original version only discussed *a. It was pointed out that we also need to worry about *r++ and a->m.]
Section: 18.2.1 [lib.limits] Status: Open Submitter: Stephen Cleary Date: 21 Dec 1999
In some places in this section, the terms "fundamental types" and "scalar types" are used when the term "arithmetic types" is intended. The current usage is incorrect because void is a fundamental type and pointers are scalar types, neither of which should have specializations of numeric_limits.
Proposed resolution:
Change 18.2 [lib.support.limits] para 1 from:
The headers <limits>, <climits>, and <cfloat> supply characteristics of implementation-dependent fundamental types (3.9.1).
to:
The headers <limits>, <climits>, and <cfloat> supply characteristics of implementation-dependent arithmetic types (3.9.1).
Change 18.2.1 [lib.limits] para 1 from:
The numeric_limits component provides a C++ program with information about various properties of the implementation's representation of the fundamental types.
to:
The numeric_limits component provides a C++ program with information about various properties of the implementation's representation of the arithmetic types.
Change 18.2.1 [lib.limits] para 2 from:
Specializations shall be provided for each fundamental type. . .
to:
Specializations shall be provided for each arithmetic type. . .
Change 18.2.1 [lib.limits] para 4 from:
Non-fundamental standard types. . .
to:
Non-arithmetic standard types. . .
Change 18.2.1.1 [lib.numeric.limits] para 1 from:
The member is_specialized makes it possible to distinguish between fundamental types, which have specializations, and non-scalar types, which do not.
to:
The member is_specialized makes it possible to distinguish between arithmetic types, which have specializations, and non-arithmetic types, which do not.
[post-Toronto: The opinion of the LWG is that the wording in the standard, as well as the wording of the proposed resolution, is flawed. The term "arithmetic types" is well defined in C and C++, and it is not clear that the term is being used correctly. It is also not clear that the term "implementation dependent" has any useful meaning in this context. The biggest problem is that numeric_limits seems to be intended both for built-in types and for user-defined types, and the standard doesn't make it clear how numeric_limits applies to each of those cases. A wholesale review of numeric_limits is needed. A paper would be welcome.]
Section: 25.2.8 [lib.alg.unique] Status: Ready Submitter: Andrew Koenig Date: 13 Jan 2000
What should unique() do if you give it a predicate that is not an equivalence relation? There are at least two plausible answers:
1. You can't, because 25.2.8 says that it it "eliminates all but the first element from every consecutive group of equal elements..." and it wouldn't make sense to interpret "equal" as meaning anything but an equivalence relation. [It also doesn't make sense to interpret "equal" as meaning ==, because then there would never be any sense in giving a predicate as an argument at all.]
2. The word "equal" should be interpreted to mean whatever the predicate says, even if it is not an equivalence relation (and in particular, even if it is not transitive).
The example that raised this question is from Usenet:
int f[] = { 1, 3, 7, 1, 2 };
int* z = unique(f, f+5, greater<int>());
If one blindly applies the definition using the predicate greater<int>, and ignore the word "equal", you get:
Eliminates all but the first element from every consecutive group of elements referred to by the iterator i in the range [first, last) for which *i > *(i - 1).
The first surprise is the order of the comparison. If we wanted to
allow for the predicate not being an equivalence relation, then we
should surely compare elements the other way: pred(*(i - 1), *i). If
we do that, then the description would seem to say: "Break the
sequence into subsequences whose elements are in strictly increasing
order, and keep only the first element of each subsequence". So the
result would be 1, 1, 2. If we take the description at its word, it
would seem to call for strictly DEcreasing order, in which case the
result should be 1, 3, 7, 2.
In fact, the SGI implementation of unique() does neither: It yields 1,
3, 7.
Proposed resolution:
Change 25.2.8 [lib.alg.unique] paragraph 1 to:
For a nonempty range, eliminates all but the first element from every consecutive group of equivalent elements referred to by the iterator i in the range [first+1, last) for which the following conditions hold: *(i-1) == *i or pred(*(i-1), *i) != false.
Also insert a new paragraph, paragraph 2a, that reads: "Requires: The comparison function must be an equivalence relation."
[Redmond: discussed arguments for and against requiring the comparison function to be an equivalence relation. Straw poll: 14-2-5. First number is to require that it be an equivalence relation, second number is to explicitly not require that it be an equivalence relation, third number is people who believe they need more time to consider the issue. A separate issue: Andy Sawyer pointed out that "i-1" is incorrect, since "i" can refer to the first iterator in the range. Matt provided wording to address this problem.]
[Curaçao: The LWG changed "... the range (first, last)..." to "... the range [first+1, last)..." for clarity. They considered this change close enough to editorial to not require another round of review.]
Rationale:
The LWG also considered an alternative resolution: change 25.2.8 [lib.alg.unique] paragraph 1 to:
For a nonempty range, eliminates all but 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 or pred(*(i-1), *i) != false.
Also insert a new paragraph, paragraph 1a, that reads: "Notes: The comparison function need not be an equivalence relation."
Informally: the proposed resolution imposes an explicit requirement that the comparison function be an equivalence relation. The alternative resolution does not, and it gives enough information so that the behavior of unique() for a non-equivalence relation is specified. Both resolutions are consistent with the behavior of existing implementations.
Section: 17.4.4.3 [lib.global.functions] Status: Open Submitter: Dave Abrahams Date: 01 Apr 2000
Are algorithms in std:: allowed to use other algorithms without qualification, so functions in user namespaces might be found through Koenig lookup?
For example, a popular standard library implementation includes this implementation of std::unique:
namespace std {
template <class _ForwardIter>
_ForwardIter unique(_ForwardIter __first, _ForwardIter __last) {
__first = adjacent_find(__first, __last);
return unique_copy(__first, __last, __first);
}
}
Imagine two users on opposite sides of town, each using unique on his own sequences bounded by my_iterators . User1 looks at his standard library implementation and says, "I know how to implement a more efficient unique_copy for my_iterators", and writes:
namespace user1 {
class my_iterator;
// faster version for my_iterator
my_iterator unique_copy(my_iterator, my_iterator, my_iterator);
}
user1::unique_copy() is selected by Koenig lookup, as he intended.
User2 has other needs, and writes:
namespace user2 {
class my_iterator;
// Returns true iff *c is a unique copy of *a and *b.
bool unique_copy(my_iterator a, my_iterator b, my_iterator c);
}
User2 is shocked to find later that his fully-qualified use of std::unique(user2::my_iterator, user2::my_iterator, user2::my_iterator) fails to compile (if he's lucky). Looking in the standard, he sees the following Effects clause for unique():
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
The standard gives user2 absolutely no reason to think he can interfere with std::unique by defining names in namespace user2. His standard library has been built with the template export feature, so he is unable to inspect the implementation. User1 eventually compiles his code with another compiler, and his version of unique_copy silently stops being called. Eventually, he realizes that he was depending on an implementation detail of his library and had no right to expect his unique_copy() to be called portably.
On the face of it, and given above scenario, it may seem obvious that the implementation of unique() shown is non-conforming because it uses unique_copy() rather than ::std::unique_copy(). Most standard library implementations, however, seem to disagree with this notion.
[Tokyo: Steve Adamczyk from the core working group indicates that "std::" is sufficient; leading "::" qualification is not required because any namespace qualification is sufficient to suppress Koenig lookup.]
Proposed resolution:
Add a paragraph and a note at the end of 17.4.4.3 [lib.global.functions]:
Unless otherwise specified, no global or non-member function in the standard library shall use a function from another namespace which is found through argument-dependent name lookup (3.4.2 [basic.lookup.koenig]).
[Note: the phrase "unless otherwise specified" is intended to allow Koenig lookup in cases like that of ostream_iterators:
Effects:*out_stream << value;
if(delim != 0) *out_stream << delim;
return (*this);--end note]
[Tokyo: The LWG agrees that this is a defect in the standard, but is as yet unsure if the proposed resolution is the best solution. Furthermore, the LWG believes that the same problem of unqualified library names applies to wording in the standard itself, and has opened issue 229 accordingly. Any resolution of issue 225 should be coordinated with the resolution of issue 229.]
[Toronto: The LWG is not sure if this is a defect in the standard. Most LWG members believe that an implementation of std::unique like the one quoted in this issue is already illegal, since, under certain circumstances, its semantics are not those specified in the standard. The standard's description of unique does not say that overloading adjacent_find should have any effect.]
[Curaçao: An LWG-subgroup spent an afternoon working on issues 225, 226, and 229. Their conclusion was that the issues should be separated into an LWG portion (Howard will write a proposal), and a EWG portion (Dave will write a proposal). The LWG and EWG had (separate) discussions of this plan the next day.]
Section: 17.4.3.1 [lib.reserved.names] Status: Open Submitter: Dave Abrahams Date: 01 Apr 2000
The issues are:
1. How can a 3rd party library implementor (lib1) write a version of a standard algorithm which is specialized to work with his own class template?
2. How can another library implementor (lib2) write a generic algorithm which will take advantage of the specialized algorithm in lib1?
This appears to be the only viable answer under current language rules:
namespace lib1 { // arbitrary-precision numbers using T as a basic unit template <class T> class big_num { //... };// defining this in namespace std is illegal (it would be an // overload), so we hope users will rely on Koenig lookup template <class T> void swap(big_int<T>&, big_int<T>&); }#include <algorithm> namespace lib2 { template <class T> void generic_sort(T* start, T* end) { ... // using-declaration required so we can work on built-in types using std::swap; // use Koenig lookup to find specialized algorithm if available swap(*x, *y); } }
This answer has some drawbacks. First of all, it makes writing lib2 difficult and somewhat slippery. The implementor needs to remember to write the using-declaration, or generic_sort will fail to compile when T is a built-in type. The second drawback is that the use of this style in lib2 effectively "reserves" names in any namespace which defines types which may eventually be used with lib2. This may seem innocuous at first when applied to names like swap, but consider more ambiguous names like unique_copy() instead. It is easy to imagine the user wanting to define these names differently in his own namespace. A definition with semantics incompatible with the standard library could cause serious problems (see issue 225).
Why, you may ask, can't we just partially specialize std::swap()? It's because the language doesn't allow for partial specialization of function templates. If you write:
namespace std
{
template <class T>
void swap(lib1::big_int<T>&, lib1::big_int<T>&);
}
You have just overloaded std::swap, which is illegal under the current language rules. On the other hand, the following full specialization is legal:
namespace std
{
template <>
void swap(lib1::other_type&, lib1::other_type&);
}
This issue reflects concerns raised by the "Namespace issue with specialized swap" thread on comp.lang.c++.moderated. A similar set of concerns was earlier raised on the boost.org mailing list and the ACCU-general mailing list. Also see library reflector message c++std-lib-7354.
J. C. van Winkel points out (in c++std-lib-9565) another unexpected fact: it's impossible to output a container of std::pair's using copy and an ostream_iterator, as long as both pair-members are built-in or std:: types. That's because a user-defined operator<< for (for example) std::pair<const std::string, int> will not be found: lookup for operator<< will be performed only in namespace std. Opinions differed on whether or not this was a defect, and, if so, whether the defect is that something is wrong with user-defined functionality and std, or whether it's that the standard library does not provide an operator<< for std::pair<>.
Proposed resolution:
[Tokyo: Summary, "There is no conforming way to extend std::swap for user defined templates." The LWG agrees that there is a problem. Would like more information before proceeding. This may be a core issue. Core issue 229 has been opened to discuss the core aspects of this problem. It was also noted that submissions regarding this issue have been received from several sources, but too late to be integrated into the issues list. ]
[Post-Tokyo: A paper with several proposed resolutions, J16/00-0029==WG21/N1252, "Shades of namespace std functions " by Alan Griffiths, is in the Post-Tokyo mailing. It should be considered a part of this issue.]
[Toronto: Dave Abrahams and Peter Dimov have proposed a resolution that involves core changes: it would add partial specialization of function template. The Core Working Group is reluctant to add partial specialization of function templates. It is viewed as a large change, CWG believes that proposal presented leaves some syntactic issues unanswered; if the CWG does add partial specialization of function templates, it wishes to develop its own proposal. The LWG continues to believe that there is a serious problem: there is no good way for users to force the library to use user specializations of generic standard library functions, and in certain cases (e.g. transcendental functions called by valarray and complex) this is important. Koenig lookup isn't adequate, since names within the library must be qualified with std (see issue 225), specialization doesn't work (we don't have partial specialization of function templates), and users aren't permitted to add overloads within namespace std. ]
[Copenhagen: Discussed at length, with no consensus. Relevant papers in the pre-Copenhagen mailing: N1289, N1295, N1296. Discussion focused on four options. (1) Relax restrictions on overloads within namespace std. (2) Mandate that the standard library use unqualified calls for swap and possibly other functions. (3) Introduce helper class templates for swap and possibly other functions. (4) Introduce partial specialization of function templates. Every option had both support and opposition. Straw poll (first number is support, second is strongly opposed): (1) 6, 4; (2) 6, 7; (3) 3, 8; (4) 4, 4.]
[Redmond: Discussed, again no consensus. Herb presented an argument that a user who is defining a type T with an associated swap should not be expected to put that swap in namespace std, either by overloading or by partial specialization. The argument is that swap is part of T's interface, and thus should to in the same namespace as T and only in that namespace. If we accept this argument, the consequence is that standard library functions should use unqualified call of swap. (And which other functions? Any?) A small group (Nathan, Howard, Jeremy, Dave, Matt, Walter, Marc) will try to put together a proposal before the next meeting.]
[Curaçao: An LWG-subgroup spent an afternoon working on issues 225, 226, and 229. Their conclusion was that the issues should be separated into an LWG portion (Howard will write a proposal), and a EWG portion (Dave will write a proposal). The LWG and EWG had (separate) discussions of this plan the next day.]
Section: 17.4.1.1 [lib.contents] Status: Open Submitter: Steve Clamage Date: 19 Apr 2000
Throughout the library chapters, the descriptions of library entities refer to other library entities without necessarily qualifying the names.
For example, section 25.2.2 "Swap" describes the effect of swap_ranges in terms of the unqualified name "swap". This section could reasonably be interpreted to mean that the library must be implemented so as to do a lookup of the unqualified name "swap", allowing users to override any ::std::swap function when Koenig lookup applies.
Although it would have been best to use explicit qualification with "::std::" throughout, too many lines in the standard would have to be adjusted to make that change in a Technical Corrigendum.
Issue 182, which addresses qualification of size_t, is a special case of this.
Proposed resolution:
To section 17.4.1.1 "Library contents" Add the following paragraph:
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.
[Post-Tokyo: Steve Clamage submitted this issue at the request of the LWG to solve a problem in the standard itself similar to the problem within implementations of library identified by issue 225. Any resolution of issue 225 should be coordinated with the resolution of this issue.]
[post-Toronto: Howard is undecided about whether it is appropriate for all standard library function names referred to in other standard library functions to be explicitly qualified by std: it is common advice that users should define global functions that operate on their class in the same namespace as the class, and this requires argument-dependent lookup if those functions are intended to be called by library code. Several LWG members are concerned that valarray appears to require argument-dependent lookup, but that the wording may not be clear enough to fall under "unless explicitly described otherwise".]
[Curaçao: An LWG-subgroup spent an afternoon working on issues 225, 226, and 229. Their conclusion was that the issues should be separated into an LWG portion (Howard will write a proposal), and a EWG portion (Dave will write a proposal). The LWG and EWG had (separate) discussions of this plan the next day.]
Section: 22.2.2.2.2 [lib.facet.num.put.virtuals] Status: Open Submitter: James Kanze, Stephen Clamage Date: 25 Apr 2000
What is the following program supposed to output?
#include <iostream>
int
main()
{
std::cout.setf( std::ios::scientific , std::ios::floatfield ) ;
std::cout.precision( 0 ) ;
std::cout << 1.00 << '\n' ;
return 0 ;
}
From my C experience, I would expect "1e+00"; this is what printf("%.0e" , 1.00 ); does. G++ outputs "1.000000e+00".
The only indication I can find in the standard is 22.2.2.2.2/11, where it says "For conversion from a floating-point type, if (flags & fixed) != 0 or if str.precision() > 0, then str.precision() is specified in the conversion specification." This is an obvious error, however, fixed is not a mask for a field, but a value that a multi-bit field may take -- the results of and'ing fmtflags with ios::fixed are not defined, at least not if ios::scientific has been set. G++'s behavior corresponds to what might happen if you do use (flags & fixed) != 0 with a typical implementation (floatfield == 3 << something, fixed == 1 << something, and scientific == 2 << something).
Presumably, the intent is either (flags & floatfield) != 0, or (flags & floatfield) == fixed; the first gives something more or less like the effect of precision in a printf floating point conversion. Only more or less, of course. In order to implement printf formatting correctly, you must know whether the precision was explicitly set or not. Say by initializing it to -1, instead of 6, and stating that for floating point conversions, if precision < -1, 6 will be used, for fixed point, if precision < -1, 1 will be used, etc. Plus, of course, if precision == 0 and flags & floatfield == 0, 1 should be = used. But it probably isn't necessary to emulate all of the anomalies of printf:-).
Proposed resolution:
In 22.2.2.2.2 [lib.facet.num.put.virtuals], paragraph 11, change "if (flags & fixed) != 0" to "if (flags & floatfield) == fixed || (flags & floatfield) == scientific"
Rationale:
The floatfield determines whether numbers are formatted as if with %f, %e, or %g. If the fixed bit is set, it's %f, if scientific it's %e, and if both bits are set, or neither, it's %e.
Turning to the C standard, a precision of 0 is meaningful for %f and %e, but not for %g: for %g, precision 0 is taken to be the same as precision 1.
The proposed resolution has the effect that the output of the above program will be "1e+00".
[Curaçao: Howard will send Matt improved wording dealing with case not covered by current PR.]
Section: 23.1.2 [lib.associative.reqmts] Status: Open Submitter: Andrew Koenig Date: 30 Apr 2000
If mm is a multimap and p is an iterator into the multimap, then mm.insert(p, x) inserts x into mm with p as a hint as to where it should go. Table 69 claims that the execution time is amortized constant if the insert winds up taking place adjacent to p, but does not say when, if ever, this is guaranteed to happen. All it says it that p is a hint as to where to insert.
The question is whether there is any guarantee about the relationship between p and the insertion point, and, if so, what it is.
I believe the present state is that there is no guarantee: The user can supply p, and the implementation is allowed to disregard it entirely.
Additional comments from Nathan:
The vote [in Redmond] was on whether to elaborately specify the use of
the hint, or to require behavior only if the value could be inserted
adjacent to the hint. I would like to ensure that we have a chance to
vote for a deterministic treatment: "before, if possible, otherwise
after, otherwise anywhere appropriate", as an alternative to the
proposed "before or after, if possible, otherwise [...]".
Proposed resolution:
In table 69 "Associative Container Requirements" in 23.1.2 [lib.associative.reqmts], in the row for a.insert(p, t), change
iterator p is a hint pointing to where the insert should start to search.
to
insertion adjacent to iterator p is preferred if more than one insertion point is valid.
and change
logarithmic in general, but amortized constant if t is inserted right after p.
to
logarithmic in general, but amortized constant if t is inserted adjacent to iterator p.
[Toronto: there was general agreement that this is a real defect: when inserting an element x into a multiset that already contains several copies of x, there is no way to know whether the hint will be used. The proposed resolution was that the new element should always be inserted as close to the hint as possible. So, for example, if there is a subsequence of equivalent values, then providing a.begin() as the hint means that the new element should be inserted before the subsequence even if a.begin() is far away. JC van Winkel supplied precise wording for this proposed resolution, and also for an alternative resolution in which hints are only used when they are adjacent to the insertion point.]
[Copenhagen: the LWG agreed to the original proposed resolution, in which an insertion hint would be used even when it is far from the insertion point. This was contingent on seeing a reference implementation showing that it is possible to implement this requirement without loss of efficiency. John Potter provided such a reference implementation.]
[Redmond: The LWG was reluctant to adopt the proposal that emerged from Copenhagen: it seemed excessively complicated, and went beyond fixing the defect that we identified in Toronto. PJP provided the new wording described in this issue. Nathan agrees that we shouldn't adopt the more detailed semantics, and notes: "we know that you can do it efficiently enough with a red-black tree, but there are other (perhaps better) balanced tree techniques that might differ enough to make the detailed semantics hard to satisfy."]
[Curaçao: Nathan should give us the alternative wording he suggests so the LWG can decide between the two options.]
Section: 25.2.8 [lib.alg.unique] Status: Ready Submitter: Angelika Langer Date: May 15 2000
Some popular implementations of unique_copy() create temporary copies of values in the input sequence, at least if the input iterator is a pointer. Such an implementation is built on the assumption that the value type is CopyConstructible and Assignable.
It is common practice in the standard that algorithms explicitly specify any additional requirements that they impose on any of the types used by the algorithm. An example of an algorithm that creates temporary copies and correctly specifies the additional requirements is accumulate(), 26.4.1 [lib.accumulate].
Since the specifications of unique() and unique_copy() do not require CopyConstructible and Assignable of the InputIterator's value type the above mentioned implementations are not standard-compliant. I cannot judge whether this is a defect in the standard or a defect in the implementations.
Proposed resolution:
In 25.2.8 change:
-4- Requires: The ranges [first, last) and [result, result+(last-first)) shall not overlap.
to:
-4- Requires: The ranges [first, last) and [result, result+(last-first)) shall not overlap. The expression *result = *first must be valid. If neither InputIterator nor OutputIterator meets the requirements of forward iterator then the value type of InputIterator must be copy constructible. Otherwise copy constructible is not required.
[Redmond: the original proposed resolution didn't impose an explicit requirement that the iterator's value type must be copy constructible, on the grounds that an input iterator's value type must always be copy constructible. Not everyone in the LWG thought that this requirement was clear from table 72. It has been suggested that it might be possible to implement unique_copy without requiring assignability, although current implementations do impose that requirement. Howard provided new wording.]
[ Curaçao: The LWG changed the PR editorially to specify "neither...nor...meet..." as clearer than "both...and...do not meet...". Change believed to be so minor as not to require re-review. ]
Section: 23.2.4.3 [lib.vector.modifiers] Status: Open Submitter: Lisa Lippincott Date: 06 June 2000
Paragraph 2 of 23.2.4.3 [lib.vector.modifiers] describes the complexity of vector::insert:
Complexity: If first and last are forward iterators, bidirectional iterators, or random access iterators, the complexity is linear in the number of elements in the range [first, last) plus the distance to the end of the vector. If they are input iterators, the complexity is proportional to the number of elements in the range [first, last) times the distance to the end of the vector.
First, this fails to address the non-iterator forms of insert.
Second, the complexity for input iterators misses an edge case -- it requires that an arbitrary number of elements can be added at the end of a vector in constant time.
At the risk of strengthening the requirement, I suggest simply
Complexity: The complexity is linear in the number of elements inserted plus the distance to the end of the vector.
For input iterators, one may achieve this complexity by first inserting at the end of the vector, and then using rotate.
I looked to see if deque had a similar problem, and was surprised to find that deque places no requirement on the complexity of inserting multiple elements (23.2.1.3 [lib.deque.modifiers], paragraph 3):
Complexity: In the worst case, inserting a single element into a deque takes time linear in the minimum of the distance from the insertion point to the beginning of the deque and the distance from the insertion point to the end of the deque. Inserting a single element either at the beginning or end of a deque always takes constant time and causes a single call to the copy constructor of T.
I suggest:
Complexity: The complexity is linear in the number of elements inserted plus the shorter of the distances to the beginning and end of the deque. Inserting a single element at either the beginning or the end of a deque causes a single call to the copy constructor of T.
Proposed resolution:
[Toronto: It's agreed that there is a defect in complexity of multi-element insert for vector and deque. For vector, the complexity should probably be something along the lines of c1 * N + c2 * distance(i, end()). However, there is some concern about whether it is reasonable to amortize away the copies that we get from a reallocation whenever we exceed the vector's capacity. For deque, the situation is somewhat less clear. Deque is notoriously complicated, and we may not want to impose complexity requirements that would imply any implementation technique more complicated than a while loop whose body is a single-element insert.]
Section: 26.3.2.1 [lib.valarray.cons], 26.3.2.2 [lib.valarray.assign] Status: Ready Submitter: Robert Klarer Date: 31 Jul 2000
This discussion is adapted from message c++std-lib-7056 posted November 11, 1999. I don't think that anyone can reasonably claim that the problem described below is NAD.
These valarray constructors can never be called:
template <class T>
valarray<T>::valarray(const slice_array<T> &);
template <class T>
valarray<T>::valarray(const gslice_array<T> &);
template <class T>
valarray<T>::valarray(const mask_array<T> &);
template <class T>
valarray<T>::valarray(const indirect_array<T> &);
Similarly, these valarray assignment operators cannot be called:
template <class T>
valarray<T> valarray<T>::operator=(const slice_array<T> &);
template <class T>
valarray<T> valarray<T>::operator=(const gslice_array<T> &);
template <class T>
valarray<T> valarray<T>::operator=(const mask_array<T> &);
template <class T>
valarray<T> valarray<T>::operator=(const indirect_array<T> &);
Please consider the following example:
#include <valarray>
using namespace std;
int main()
{
valarray<double> va1(12);
valarray<double> va2(va1[slice(1,4,3)]); // line 1
}
Since the valarray va1 is non-const, the result of the sub-expression va1[slice(1,4,3)] at line 1 is an rvalue of type const std::slice_array<double>. This slice_array rvalue is then used to construct va2. The constructor that is used to construct va2 is declared like this:
template <class T>
valarray<T>::valarray(const slice_array<T> &);
Notice the constructor's const reference parameter. When the constructor is called, a slice_array must be bound to this reference. The rules for binding an rvalue to a const reference are in 8.5.3, paragraph 5 (see also 13.3.3.1.4). Specifically, paragraph 5 indicates that a second slice_array rvalue is constructed (in this case copy-constructed) from the first one; it is this second rvalue that is bound to the reference parameter. Paragraph 5 also requires that the constructor that is used for this purpose be callable, regardless of whether the second rvalue is elided. The copy-constructor in this case is not callable, however, because it is private. Therefore, the compiler should report an error.
Since slice_arrays are always rvalues, the valarray constructor that has a parameter of type const slice_array<T> & can never be called. The same reasoning applies to the three other constructors and the four assignment operators that are listed at the beginning of this post. Furthermore, since these functions cannot be called, the valarray helper classes are almost entirely useless.
Proposed resolution:
slice_array:
gslice_array:
mask_array:
indirect_array:
[This wording is taken from Robert Klarer's reflector message, c++std-lib-7827. Gabriel Dos Reis agrees that this general solution is correct.]
Rationale:
Keeping the valarray constructors private is untenable. Merely making valarray a friend of the helper classes isn't good enough, because access to the copy constructor is checked in the user's environment.
Making the assignment operator public is not strictly necessary to solve this problem. A majority of the LWG (straw poll: 13-4) believed we should make the assignment operators public, in addition to the copy constructors, for reasons of symmetry and user expectation.
Section: 19.1 [lib.std.exceptions] Status: Open Submitter: Dave Abrahams Date: 01 Aug 2000
Many of the standard exception types which implementations are required to throw are constructed with a const std::string& parameter. For example:
19.1.5 Class out_of_range [lib.out.of.range]
namespace std {
class out_of_range : public logic_error {
public:
explicit out_of_range(const string& what_arg);
};
}
1 The class out_of_range defines the type of objects thrown as excep-
tions to report an argument value not in its expected range.
out_of_range(const string& what_arg);
Effects:
Constructs an object of class out_of_range.
Postcondition:
strcmp(what(), what_arg.c_str()) == 0.
There are at least two problems with this:
There may be no cure for (1) other than changing the interface to out_of_range, though one could reasonably argue that (1) is not a defect. Personally I don't care that much if out-of-memory is reported when I only have 20 bytes left, in the case when out_of_range would have been reported. People who use exception-specifications might care a lot, though.
There is a cure for (2), but it isn't completely obvious. I think a note for implementors should be made in the standard. Avoiding possible termination in this case shouldn't be left up to chance. The cure is to use a reference-counted "string" implementation in the exception object. I am not necessarily referring to a std::string here; any simple reference-counting scheme for a NTBS would do.
Further discussion, in email:
...I'm not so concerned about (1). After all, a library implementation can add const char* constructors as an extension, and users don't need to avail themselves of the standard exceptions, though this is a lame position to be forced into. FWIW, std::exception and std::bad_alloc don't require a temporary basic_string.
...I don't think the fixed-size buffer is a solution to the problem,
strictly speaking, because you can't satisfy the postcondition
strcmp(what(), what_arg.c_str()) == 0
For all values of what_arg (i.e. very long values). That means that
the only truly conforming solution requires a dynamic allocation.
Further discussion, from Redmond:
The most important progress we made at the Redmond meeting was realizing that there are two separable issues here: the const string& constructor, and the copy constructor. If a user writes something like throw std::out_of_range("foo"), the const string& constructor is invoked before anything gets thrown. The copy constructor is potentially invoked during stack unwinding.
The copy constructor is a more serious problem, becuase failure during stack unwinding invokes terminate. The copy constructor must be nothrow. Curaçao: Howard thinks this requirement is already present.
The fundamental problem is that it's difficult to get the nothrow requirement to work well with the requirement that the exception objects store a string of unbounded size, particularly if you also try to make the const string& constructor nothrow. Options discussed include:
(Not all of these options are mutually exclusive.)
Proposed resolution:
[Toronto: some LWG members thought this was merely a QoI issue, but most believed that it was at least a borderline defect. There was more support for nonnormative advice to implementors than for a normative change.]
[Redmond: discussed, without definite conclusion. Most LWG members thought there was a real defect lurking here. A small group (Herb, Kevlin, Howard, Martin, Dave) will try to make a recommendation.]
[Curaçao: Howard will nag the others to work on a recommendation.]
Section: 20.1.5 [lib.allocator.requirements] Status: Open Submitter: Matt Austern Date: 22 Aug 2000
From lib-7752:
I've been assuming (and probably everyone else has been assuming) that allocator instances have a particular property, and I don't think that property can be deduced from anything in Table 32.
I think we have to assume that allocator type conversion is a homomorphism. That is, if x1 and x2 are of type X, where X::value_type is T, and if type Y is X::template rebind<U>::other, then Y(x1) == Y(x2) if and only if x1 == x2.
Further discussion: Howard Hinnant writes, in lib-7757:
I think I can prove that this is not provable by Table 32. And I agree it needs to be true except for the "and only if". If x1 != x2, I see no reason why it can't be true that Y(x1) == Y(x2). Admittedly I can't think of a practical instance where this would happen, or be valuable. But I also don't see a need to add that extra restriction. I think we only need:
if (x1 == x2) then Y(x1) == Y(x2)
If we decide that == on allocators is transitive, then I think I can prove the above. But I don't think == is necessarily transitive on allocators. That is:
Given x1 == x2 and x2 == x3, this does not mean x1 == x3.
Example:
x1 can deallocate pointers from: x1, x2, x3
x2 can deallocate pointers from: x1, x2, x4
x3 can deallocate pointers from: x1, x3
x4 can deallocate pointers from: x2, x4x1 == x2, and x2 == x4, but x1 != x4
Proposed resolution:
[Toronto: LWG members offered multiple opinions. One opinion is that it should not be required that x1 == x2 implies Y(x1) == Y(x2), and that it should not even be required that X(x1) == x1. Another opinion is that the second line from the bottom in table 32 already implies the desired property. This issue should be considered in light of other issues related to allocator instances.]
Section: 23.2.2.4 [lib.list.ops] Status: Open Submitter: P.J. Plauger Date: 27 Nov 2000
Section 23.2.2.4 [lib.list.ops] states that
void splice(iterator position, list<T, Allocator>& x);
invalidates all iterators and references to list x.
But what does the C++ Standard mean by "invalidate"? You can still dereference the iterator to a spliced list element, but you'd better not use it to delimit a range within the original list. For the latter operation, it has definitely lost some of its validity.
If we accept the proposed resolution to issue 250, then we'd better clarify that a "valid" iterator need no longer designate an element within the same container as it once did. We then have to clarify what we mean by invalidating a past-the-end iterator, as when a vector or string grows by reallocation. Clearly, such an iterator has a different kind of validity. Perhaps we should introduce separate terms for the two kinds of "validity."
Proposed resolution:
Add the following text to the end of section 24.1 [lib.iterator.requirements], after paragraph 5:
An invalid iterator is an iterator that may be singular. [Footnote: This definition applies to pointers, since pointers are iterators. The effect of dereferencing an iterator that has been invalidated is undefined.]
[post-Copenhagen: Matt provided wording.]
[Redmond: General agreement with the intent, some objections to the wording. Dave provided new wording.]
[Curaçao: The definition of "singular" is contentious. The 278 resolution must be made consistent with issue 208 and 24.1/5. Furthermore, a Rationale paragraph is required.]
Section: 24.4.1 [lib.reverse.iterators] Status: Open Submitter: Steve Cleary Date: 27 Nov 2000
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 believed it should be a separate issue. There was also some reservations about whether this was a worthwhile problem to fix.
Steve said: "Fixing reverse_iterator. std::reverse_iterator can (and should) be changed to preserve these additional requirements." He also said in email that it can be done without breaking user's code: "If you take a look at my suggested solution, reverse_iterator doesn't have to take two parameters; there is no danger of breaking existing code, except someone taking the address of one of the reverse_iterator global operator functions, and I have to doubt if anyone has ever done that. . . But, just in case they have, you can leave the old global functions in as well -- they won't interfere with the two-template-argument functions. With that, I don't see how any user code could break."
Proposed resolution:
Section: 24.4.1.1 [lib.reverse.iterator] add/change the following declarations:
A) Add a templated assignment operator, after the same manner
as the templated copy constructor, i.e.:
template < class U >
reverse_iterator < Iterator >& operator=(const reverse_iterator< U >& u);
B) Make all global functions (except the operator+) have
two template parameters instead of one, that is, for
operator ==, !=, <, >, <=, >=, - replace:
template < class Iterator >
typename reverse_iterator< Iterator >::difference_type operator-(
const reverse_iterator< Iterator >& x,
const reverse_iterator< Iterator >& y);
with:
template < class Iterator1, class Iterator2 >
typename reverse_iterator < Iterator1 >::difference_type operator-(
const reverse_iterator < Iterator1 > & x,
const reverse_iterator < Iterator2 > & y);
Also make the addition/changes for these signatures in 24.4.1.3 [lib.reverse.iter.ops].
[ Copenhagen: The LWG is concerned that the proposed resolution introduces new overloads. Experience shows that introducing overloads is always risky, and that it would be inappropriate to make this change without implementation experience. It may be desirable to provide this feature in a different way. ]
Section: 22.2.2.2.2 [lib.facet.num.put.virtuals] Status: Open Submitter: Howard Hinnant Date: 5 Dec 2000
Paragraph 16 mistakenly singles out integral types for inserting thousands_sep() characters. This conflicts with the syntax for floating point numbers described under 22.2.3.1/2.
Proposed resolution:
Change paragraph 16 from:
For integral types, punct.thousands_sep() characters are inserted into the sequence as determined by the value returned by punct.do_grouping() using the method described in 22.2.3.1.2 [lib.facet.numpunct.virtuals].
To:
For arithmetic types, punct.thousands_sep() characters are inserted into the sequence as determined by the value returned by punct.do_grouping() using the method described in 22.2.3.1.2 [lib.facet.numpunct.virtuals].
[ Copenhagen: Opinions were divided about whether this is actually an inconsistency, but at best it seems to have been unintentional. This is only an issue for floating-point output: The standard is unambiguous that implementations must parse thousands_sep characters when performing floating-point. The standard is also unambiguous that this requirement does not apply to the "C" locale. ]
[ A survey of existing practice is needed; it is believed that some implementations do insert thousands_sep characters for floating-point output and others fail to insert thousands_sep characters for floating-point input even though this is unambiguously required by the standard. ]
[Curaçao: Howard will email Bill and other implementors to try to move the issue forward.]
Section: 25.2.4 [lib.alg.replace] Status: Open Submitter: Martin Sebor Date: 15 Dec 2000
The requirements in 25.2.4 [lib.alg.replace], p1 that T to be Assignable (23.1 [lib.container.requirements]) is not necessary or sufficient for either of the algorithms. The algorithms require that std::iterator_traits<ForwardIterator>::value_type be Assignable and that both std::iterator_traits<ForwardIterator>::value_type and be EqualityComparable (20.1.1 [lib.equalitycomparable]) with respect to one another.
Further discussion, from Jeremy:
There are a number of problems with the requires clauses for the algorithms in 25.1 [lib.alg.nonmodifying] and 25.2 [lib.alg.modifying.operations]. The requires clause of each algorithm should describe the necessary and sufficient requirements on the inputs to the algorithm such that the algorithm compiles and runs properly. Many of the requires clauses fail to do this. Here is a summary of the kinds of mistakes:
Here is the list of algorithms that contain mistakes:
Also, in the requirements for EqualityComparable, the requirement that the operator be defined for const objects is lacking.
Proposed resolution:
20.1.1 [lib.equalitycomparable] Change p1 from
In Table 28, T is a type to be supplied by a C++ program instantiating a template, a, b, and c are values of type T.
to
In Table 28, T is a type to be supplied by a C++ program instantiating a template, a, b, and c are values of type const T.
25.1.2 [lib.alg.find] Change p1 from
Requires: Type T is EqualityComparable (20.1.1).
to
Requires: There must be a equality operator defined that accepts type std::iterator_traits<InputIterator>::reference for the left operand and const T for the right operand.
25.1.3 [lib.alg.find.end] Add the following requires clause
Requires: There must be an equality operator defined that accepts type const std::iterator_traits<ForwardIterator1>::value_type for the left operand and const std::iterator_traits<ForwardIterator2>::value_type for the right operand.
25.1.4 [lib.alg.find.first.of] Add the following requires clause
Requires: There must be an equality operator defined that accepts type const std::iterator_traits<ForwardIterator1>::value_type for the left operand and const std::iterator_traits<ForwardIterator2>::value_type for the right operand.
25.1.5 [lib.alg.adjacent.find] Add the following requires clause
Requires: T must be EqualityComparable (20.1.1).
25.1.6 [lib.alg.count] Change p1 from
Requires: Type T is EqualityComparable (20.1.1).
to
Requires: There must be a equality operator defined that accepts type std::iterator_traits<InputIterator>::reference for the left operand and const T for the right operand.
25.1.7 [lib.mismatch] Add the following requires clause
Requires: There must be an equality operator defined that accepts type std::iterator_traits<InputIterator1>::reference for the left operand and std::iterator_traits<InputIterator2>::reference for the right operand.
25.1.8 [lib.alg.equal] Add the following requires clause
Requires: There must be an equality operator defined that accepts type std::iterator_traits<InputIterator1>::reference for the left operand and std::iterator_traits<InputIterator2>::reference for the right operand.
25.1.9 [lib.alg.search] Add the following requires clause
Requires: There must be an equality operator defined that accepts type const std::iterator_traits<ForwardIterator1>::value_type for the left operand and const std::iterator_traits<ForwardIterator2>::value_type for the right operand.
Change change p4 from
Requires: Type T is EqualityComparable (20.1.1), type Size is convertible to integral type (4.7.12.3).
to
Requires: There must be an equality operator defined that accepts const std::iterator_traits<ForwardIterator>::value_type for the left operand and const T for the right operand. The type Size is convertible to integ