Doc. no. N3438=12-0128
Date: 2012-09-24
Project: Programming Language C++
Reply to: Alisdair Meredith <lwgchair@gmail.com>

C++ Standard Library Active Issues List (Revision R79)

Revised 2012-09-24 at 14:09:21 UTC

Reference ISO/IEC IS 14882:2011(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 INCITS PL22.16 and ISO WG21 C++ Standards Committee. Issues represent potential defects in the ISO/IEC IS 14882:2011(E) document.

This document contains only library issues which are actively being considered by the Library Working Group, i.e., issues which have a status of New, Open, Ready, or 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.

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.open-std.org/jtc1/sc22/wg21/. Requests for further information about this document should include the document number above, reference ISO/IEC 14882:2011(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 comp.std.c++ FAQ.

How to submit an issue

  1. Mail your issue to the author of this list.
  2. Specify a short descriptive title. If you fail to do so, the subject line of your mail will be used as the issue title.
  3. If the "From" on your email is not the name you wish to appear as issue submitter, then specify issue submitter.
  4. Provide a brief discussion of the problem you wish to correct. Refer to the latest working draft or standard using [section.tag] and paragraph numbers where appropriate.
  5. Provide proposed wording. This should indicate exactly how you want the standard to be changed. General solution statements belong in the discussion area. This area contains very clear and specific directions on how to modify the current draft. If you are not sure how to word a solution, you may omit this part. But your chances of a successful issue greatly increase if you attempt wording.
  6. It is not necessary for you to use html markup. However, if you want to, you can <ins>insert text like this</ins> and <del>delete text like this</del>. The only strict requirement is to communicate clearly to the list maintainer exactly how you want your issue to look.
  7. It is not necessary for you to specify other html font/formatting mark-up, but if you do the list maintainer will attempt to respect your formatting wishes (as described by html markup, or other common idioms).
  8. It is not necessary for you to specify open date or last modified date (the date of your mail will be used).
  9. It is not necessary for you to cross reference other issues, but you can if you like. You do not need to form the hyperlinks when you do, the list maintainer will take care of that.
  10. One issue per email is best.
  11. Between the time you submit the issue, and the next mailing deadline (date at the top of the Revision History), you own this issue. You control the content, the stuff that is right, the stuff that is wrong, the format, the misspellings, etc. You can even make the issue disappear if you want. Just let the list maintainer know how you want it to look, and he will try his best to accommodate you. After the issue appears in an official mailing, you no longer enjoy exclusive ownership of it.

Revision History

Issue Status

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.

Deferred - The LWG has discussed the issue, is not yet ready to move the issue forward, but neither does it deem the issue significant enough to delay publishing a standard or Technical Report. A typical deferred issue would be seeking to clarify wording that might be technically correct, but easily mis-read.

A Proposed Resolution for a deferred 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.

NAD Editorial - The LWG has reached consensus that the issue can either be handled editorially, or is handled by a paper (usually linked to in the rationale).

NAD Concepts - The LWG has reached consensus that the issue is NAD for now, but represents a real issue when the library is done with language-supported concepts.

NAD Future - In addition to the regular status, the LWG believes that this issue should be revisited at the next revision of the standard.

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).

Resolved - The LWG has reached consensus that the issue is a defect in the Standard, but the resolution adopted to resolve the issue came via some other mechanism than this issue in the list - typically by applying a formal paper, occasionally as a side effect of consolidating several interacting issue resolutions into a single issue.

DR - (Defect Report) - The full WG21/PL22.16 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.

TC1 - (Technical Corrigenda 1) - The full WG21/PL22.16 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.

CD1 - (Committee Draft 2008) - The full WG21/PL22.16 committee has voted to accept the Defect Report's Proposed Resolution into the Fall 2008 Committee Draft.

TRDec - (Decimal TR defect) - The LWG has voted to accept the Defect Report's Proposed Resolution into the Decimal TR. Action on this issue is thus complete and no further action is expected.

WP - (Working Paper) - The proposed resolution has not been accepted as a Technical Corrigendum, but the full WG21/PL22.16 committee has voted to apply the Defect Report's Proposed Resolution to the working paper.

Tentatively - This is a status qualifier. The issue has been reviewed online, or at an unofficial meeting, but not in an official meeting, and some support has been formed for the qualified status. Tentatively qualified issues may be moved to the unqualified status and forwarded to full committee (if Ready) within the same meeting. Unlike Ready issues, Tentatively Ready issues will be reviewed in subcommittee prior to forwarding to full committee. When a status is qualified with Tentatively, the issue is still considered active.

Pending - This is a status qualifier. When prepended to a status this indicates the issue has been processed by the committee, and a decision has been made to move the issue to the associated unqualified status. However for logistical reasons the indicated outcome of the issue has not yet appeared in the latest working paper.

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 (TC1), or are closed without action other than a Record of Response (Resolved ). The intent of this LWG process is that only issues which are truly defects in the Standard move to the formal ISO DR status.

Active Issues


1169. num_get not fully compatible with strto*

Section: 22.4.2.1.2 [facet.num.get.virtuals] Status: Open Submitter: Cosmin Truta Opened: 2009-07-04 Last modified: 2012-09-24

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

View all issues with Open status.

Discussion:

As specified in the latest draft, N2914, num_get is still not fully compatible with the following C functions: strtoul, strtoull, strtof and strtod.

In C, when conversion of a string to an unsigned integer type falls outside the representable range, strtoul and strtoull return ULONG_MAX and ULLONG_MAX, respectively, regardless whether the input field represents a positive or a negative value. On the other hand, the result of num_get conversion of negative values to unsigned integer types is zero. This raises a compatibility issue.

Moreover, in C, when conversion of a string to a floating-point type falls outside the representable range, strtof, strtod and strtold return ±HUGE_VALF, ±HUGE_VAL and ±HUGE_VALL, respectively. On the other hand, the result of num_get conversion of such out-of-range floating-point values results in the most positive/negative representable value. Although many C library implementations do implement HUGE_VAL (etc.) as the highest representable (which is, usually, the infinity), this isn't required by the C standard. The C library specification makes no statement regarding the value of HUGE_VAL and friends, which potentially raises the same compatibility issue as in the above case of unsigned integers. In addition, neither C nor C++ define symbolic constants for the maximum representable floating-point values (they only do so only for the maximum representable finite floating-point values), which raises a usability issue (it would be hard for the programmer to check the result of num_get against overflow).

As such, we propose to adjust the specification of num_get to closely follow the behavior of all of its underlying C functions.

[ 2010 Rapperswil: ]

Some concern that this is changing the specification for an existing C++03 function, but it was pointed out that this was underspecified as resolved by issue 23. This is clean-up for that issue in turn. Some concern that we are trying to solve the same problem in both clause 22 and 27.

Bill: There's a change here as to whether val is stored to in an error case.

Pablo: Don't think this changes whether val is stored to or not, but changes the value that is stored.

Bill: Remembers having skirmishes with customers and testers as to whether val is stored to, and the resolution was not to store in error cases.

Howard: Believes since C++03 we made a change to always store in overflow.

Everyone took some time to review the issue.

Pablo: C++98 definitely did not store any value during an error condition.

Dietmar: Depends on the question of what is considered an error, and whether overflow is an error or not, which was the crux of LWG 23.

Pablo: Yes, but given the "zero, if the conversion function fails to convert the entire field", we are requiring every error condition to store.

Bill: When did this happen?

Alisdair: One of the last two or three meetings.

Dietmar: To store a value in case of failure is a very bad idea.

Move to Open, needs more study.

[2011-03-24 Madrid meeting]

Move to deferred

[ 2011 Bloomington ]

The proposed wording looks good, no-one sure why this was held back before. Move to Review.

[2012,Kona]

Move to Open.

THe issues is what to do with -1. Should it match 'C' or do the "sane" thing. A fix here changes behavior, but is probably what we want.

Pablo to provide wording, with help from Howard.

Proposed resolution:

Change 22.4.2.1.2 [facet.num.get.virtuals] as follows:

Stage 3: The sequence of chars accumulated in stage 2 (the field) is converted to a numeric value by the rules of one of the functions declared in the header <cstdlib>:

The numeric value to be stored can be one of:

The resultant numeric value is stored in val. If the conversion function fails to convert the entire field, or if the field represents a value outside the range of representable values, ios_base::failbit is assigned to err.


1175. unordered complexity

Section: 23.2.5 [unord.req] Status: Review Submitter: Pablo Halpern Opened: 2009-07-17 Last modified: 2012-09-24

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with Review status.

Discussion:

When I look at the unordered_* constructors, I think the complexity is poorly described and does not follow the style of the rest of the standard.

The complexity for the default constructor is specified as constant. Actually, it is proportional to n, but there are no invocations of value_type constructors or other value_type operations.

For the iterator-based constructor the complexity should be:

Complexity: exactly n calls to construct value_type from InputIterator::value_type (where n = distance(f,l)). The number of calls to key_equal::operator() is proportional to n in the average case and n*n in the worst case.

[ 2010 Rapperswil: ]

Concern that the current wording may require O(1) where that cannot be delivered. We need to look at both the clause 23 requirements tables and the constructor description of each unordered container to be sure.

Howard suggests NAD Editorial as we updated the container requirement tables since this issue was written.

Daniel offers to look deeper, and hopefully produce wording addressing any outstanding concerns at the next meeting.

Move to Open.

[2011-02-26: Daniel provides wording]

I strongly suggest to clean-up the differences between requirement tables and individual specifications. In the usual way, the most specific specifications wins, which is in this case the wrong one. In regard to the concern expressed about missing DefaultConstructible requirements of the value type I disagree: The function argument n is no size-control parameter, but only some effective capacity parameter: No elements will be value-initialized by these constructors. The necessary requirement for the value type, EmplaceConstructible into *this, is already listed in Table 103 — Unordered associative container requirements. Another part of the proposed resolution is the fact that there is an inconsistency of the complexity counting when both a range and a bucket count is involved compared to constructions where only bucket counts are provided: E.g. the construction X a(n); has a complexity of n bucket allocations, but this part of the work is omitted for X a(i, j, n);, even though it is considerable larger (in the average case) for n ≫ distance(i, j).

[2011-03-24 Madrid meeting]

Move to deferred

[ 2011 Bloomington ]

The proposed wording looks good. Move to Review.

[2012, Kona]

Fix up some presentation issues with the wording, combining the big-O expressions into single expressions rather than the sum of two separate big-Os.

Strike "constant or linear", prefer "linear in the number of buckets". This allows for number of buckets being larger than requested n as well.

Default n to "unspecified" rather than "implementation-defined". It seems an un-necessary burden asking vendors to document a quantity that is easily determined through the public API of these classes.

Replace distance(f,l) with "number of elements in the range [f,l)"

Retain in Review with the updated wording

Proposed resolution:

  1. Modify the following rows in Table 103 — Unordered associative container requirements to add the explicit bucket allocation overhead of some constructions. As editorial recommendation it is suggested not to shorten the sum 𝒪(n) + 𝒪(N) to 𝒪(n + N), because two different work units are involved.

    Table 103 — Unordered associative container requirements (in addition to container)
    Expression Return type Assertion/note pre-/post-condition Complexity
    X(i, j, n, hf, eq)
    X a(i, j, n, hf, eq)
    X
    Effects: Constructs an empty container with at least n
    buckets, using hf as the hash function and eq as the key
    equality predicate, and inserts elements from [i, j) into it.
    Average case 𝒪(n + N) (N is distance(i, j)),
    worst case 𝒪(n) + 𝒪(N2)
    X(i, j, n, hf)
    X a(i, j, n, hf)
    X
    Effects: Constructs an empty container with at least n
    buckets, using hf as the hash function and key_equal() as the key
    equality predicate, and inserts elements from [i, j) into it.
    Average case 𝒪(n + N) (N is distance(i, j)),
    worst case 𝒪(n + N2)
    X(i, j, n)
    X a(i, j, n)
    X
    Effects: Constructs an empty container with at least n
    buckets, using hasher() as the hash function and key_equal() as the key
    equality predicate, and inserts elements from [i, j) into it.
    Average case 𝒪(n + N) (N is distance(i, j)),
    worst case 𝒪(n + N2)
  2. Modify 23.5.4.2 [unord.map.cnstr] p. 1-4 as indicated (The edits of p. 1 and p. 3 attempt to fix some editorial oversight.):

    explicit unordered_map(size_type n = see below,
                           const hasher& hf = hasher(),
                           const key_equal& eql = key_equal(),
                           const allocator_type& a = allocator_type());
    

    1 Effects: Constructs an empty unordered_map using the specified hash function, key equality function, and allocator, and using at least n buckets. If n is not provided, the number of buckets is unspeciedimpldefdefault number of buckets in unordered_map. max_load_factor() returns 1.0.

    2 Complexity: ConstantLinear in the number of buckets.

    template <class InputIterator>
    unordered_map(InputIterator f, InputIterator l,
                  size_type n = see below,
                  const hasher& hf = hasher(),
                  const key_equal& eql = key_equal(),
                  const allocator_type& a = allocator_type());
    

    3 Effects: Constructs an empty unordered_map using the specified hash function, key equality function, and allocator, and using at least n buckets. If n is not provided, the number of buckets is unspecifiedimpldefdefault number of buckets in unordered_map. Then inserts elements from the range [f, l). max_load_factor() returns 1.0.

    4 Complexity: Average case linear, worst case quadraticLinear in the number of buckets. In the average case linear in N and in the worst case quadratic in N to insert the elements, where N is equal to number of elements in the range [f,l).

  3. Modify 23.5.5.2 [unord.multimap.cnstr] p. 1-4 as indicated (The edits of p. 1 and p. 3 attempt to fix some editorial oversight.):

    explicit unordered_multimap(size_type n = see below,
                                const hasher& hf = hasher(),
                                const key_equal& eql = key_equal(),
                                const allocator_type& a = allocator_type());
    

    1 Effects: Constructs an empty unordered_multimap using the specified hash function, key equality function, and allocator, and using at least n buckets. If n is not provided, the number of buckets is unspecifiedimpldefdefault number of buckets in unordered_multimap. max_load_factor() returns 1.0.

    2 Complexity: ConstantLinear in the number of buckets.

    template <class InputIterator>
    unordered_multimap(InputIterator f, InputIterator l,
                       size_type n = see below,
                       const hasher& hf = hasher(),
                       const key_equal& eql = key_equal(),
                       const allocator_type& a = allocator_type());
    

    3 Effects: Constructs an empty unordered_multimap using the specified hash function, key equality function, and allocator, and using at least n buckets. If n is not provided, the number of buckets is unspecifiedimpldefdefault number of buckets in unordered_multimap. Then inserts elements from the range [f, l). max_load_factor() returns 1.0.

    4 Complexity: Average case linear, worst case quadraticLinear in the number of buckets. In the average case linear in N and in the worst case quadratic in N to insert the elements, where N is equal to number of elements in the range [f,l).

  4. Modify 23.5.6.2 [unord.set.cnstr] p. 1-4 as indicated (The edits of p. 1 and p. 3 attempt to fix some editorial oversight.):

    explicit unordered_set(size_type n = see below,
                           const hasher& hf = hasher(),
                           const key_equal& eql = key_equal(),
                           const allocator_type& a = allocator_type());
    

    1 Effects: Constructs an empty unordered_set using the specified hash function, key equality function, and allocator, and using at least n buckets. If n is not provided, the number of buckets is unspecifiedimpldefdefault number of buckets in unordered_set. max_load_factor() returns 1.0.

    2 Complexity: ConstantLinear in the number of buckets.

    template <class InputIterator>
    unordered_set(InputIterator f, InputIterator l,
                  size_type n = see below,
                  const hasher& hf = hasher(),
                  const key_equal& eql = key_equal(),
                  const allocator_type& a = allocator_type());
    

    3 Effects: Constructs an empty unordered_set using the specified hash function, key equality function, and allocator, and using at least n buckets. If n is not provided, the number of buckets is unspecifiedimpldefdefault number of buckets in unordered_set. Then inserts elements from the range [f, l). max_load_factor() returns 1.0.

    4 Complexity: Average case linear, worst case quadraticLinear in the number of buckets. In the average case linear in N and in the worst case quadratic in N to insert the elements, where N is equal to number of elements in the range [f,l).

  5. Modify 23.5.7.2 [unord.multiset.cnstr] p. 1-4 as indicated (The edits of p. 1 and p. 3 attempt to fix some editorial oversight.):

    explicit unordered_multiset(size_type n = see below,
                                const hasher& hf = hasher(),
                                const key_equal& eql = key_equal(),
                                const allocator_type& a = allocator_type());
    

    1 Effects: Constructs an empty unordered_multiset using the specified hash function, key equality function, and allocator, and using at least n buckets. If n is not provided, the number of buckets is unspecifiedimpldefdefault number of buckets in unordered_multiset. max_load_factor() returns 1.0.

    2 Complexity: ConstantLinear in the number of buckets.

    template <class InputIterator>
    unordered_multiset(InputIterator f, InputIterator l,
                       size_type n = see below,
                       const hasher& hf = hasher(),
                       const key_equal& eql = key_equal(),
                       const allocator_type& a = allocator_type());
    

    3 Effects: Constructs an empty unordered_multiset using the specified hash function, key equality function, and allocator, and using at least n buckets. If n is not provided, the number of buckets is unspecifiedimpldefdefault number of buckets in unordered_multiset. Then inserts elements from the range [f, l). max_load_factor() returns 1.0.

    4 Complexity: Average case linear, worst case quadraticLinear in the number of buckets. In the average case linear in N and in the worst case quadratic in N to insert the elements, where N is equal to number of elements in the range [f,l).


1213. Meaning of valid and singular iterator underspecified

Section: 24.2 [iterator.requirements] Status: Deferred Submitter: Daniel Krügler Opened: 2009-09-19 Last modified: 2012-09-24

View all other issues in [iterator.requirements].

Discussion:

The terms valid iterator and singular aren't properly defined. The fuzziness of those terms became even worse after the resolution of 208 (including further updates by 278). In 24.2 [iterator.requirements] as of N2723 the standard says now:

5 - 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. [...] 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. [...] Dereferenceable values are always non-singular.

10 - An invalid iterator is an iterator that may be singular.

First, issue 208 intentionally removed the earlier constraint that past-the-end values are always non-singular. The reason for this was to support null pointers as past-the-end iterators of e.g. empty sequences. But there seem to exist different views on what a singular (iterator) value is. E.g. according to the SGI definition a null pointer is not a singular value:

Dereferenceable iterators are always nonsingular, but the converse is not true. For example, a null pointer is nonsingular (there are well defined operations involving null pointers) even thought it is not dereferenceable.

and proceeds:

An iterator is valid if it is dereferenceable or past-the-end.

Even if the standard prefers a different meaning of singular here, the change was incomplete, because by restricting feasible expressions of singular iterators to destruction and assignment isn't sufficient for a past-the-end iterator: Of-course it must still be equality-comparable and in general be a readable value.

Second, the standard doesn't clearly say whether a past-the-end value is a valid iterator or not. E.g. 20.6.12 [specialized.algorithms]/1 says:

In all of the following algorithms, the formal template parameter ForwardIterator is required to satisfy the requirements of a forward iterator (24.1.3) [..], and is required to have the property that no exceptions are thrown from [..], or dereference of valid iterators.

The standard should make better clear what "singular pointer" and "valid iterator" means. The fact that the meaning of a valid value has a core language meaning doesn't imply that for an iterator concept the term "valid iterator" has the same meaning.

Let me add a final example: In X [allocator.concepts.members] of N2914 we find:

pointer X::allocate(size_type n);

11 Returns: a pointer to the allocated memory. [Note: if n == 0, the return value is unspecified. —end note]

[..]

void X::deallocate(pointer p, size_type n);

Preconditions: p shall be a non-singular pointer value obtained from a call to allocate() on this allocator or one that compares equal to it.

If singular pointer value would include null pointers this make the preconditions unclear if the pointer value is a result of allocate(0): Since the return value is unspecified, it could be a null pointer. Does that mean that programmers need to check the pointer value for a null value before calling deallocate?

[ 2010-11-09 Daniel comments: ]

A later paper is in preparation.

[ 2010 Batavia: ]

Doesn't need to be resolved for Ox

Proposed resolution:

Consider to await the paper.


1450. [FCD] Contradiction in regex_constants

Section: 28.5.2 [re.matchflag] Status: Open Submitter: BSI Opened: 2010-08-25 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

Addresses GB-127

The Bitmask Type requirements in 17.5.2.1.3 [bitmask.types] p.3 say that all elements on a bitmask type have distinct values, but 28.5.2 [re.matchflag] defines regex_constants::match_default and regex_constants::format_default as elements of the bitmask type regex_constants::match_flag_type, both with value 0. This is a contradiction.

[ Resolution proposed by ballot comment: ]

One of the bitmask elements should be removed from the declaration and should be defined separately, in the same manner as ios_base::adjustfield, ios_base::basefield and ios_base::floatfield are defined by 27.5.3.1.2 [ios::fmtflags] p.2 and Table 120. These are constants of a bitmask type, but are not distinct elements, they have more than one value set in the bitmask. regex_constants::format_default should be specified as a constant with the same value as regex_constants::match_default.

[ 2010-10-31 Daniel comments: ]

Strictly speaking, a bitmask type cannot have any element of value 0 at all, because any such value would contradict the requirement expressed in 17.5.2.1.3 [bitmask.types] p. 3:

for any pair Ci and Cj, Ci & Ci is nonzero

So, actually both regex_constants::match_default and regex_constants::format_default are only constants of the type regex_constants::match_flag_type, and no bitmask elements.

[ 2010-11-03 Daniel comments and provides a proposed resolution: ]

The proposed resolution is written against N3126 and considered as a further improvement of the fixes suggested by n3110.

Add the following sentence to 28.5.2 [re.matchflag] paragraph 1:

1 The type regex_constants::match_flag_type is an implementation-defined bitmask type (17.5.2.1.3). Matching a regular expression against a sequence of characters [first,last) proceeds according to the rules of the grammar specified for the regular expression object, modified according to the effects listed in Table 136 for any bitmask elements set. Type regex_constants::match_flag_type also defines the constants regex_constants::match_default and regex_constants::format_default.

[ 2011 Bloomington ]

It appears the key problem is the phrasing of the bitmask requirements. Jeremiah supplies updated wording.

Pete Becker has also provided an alternative resolution.

Ammend 17.5.2.1.3 [bitmask.types]:

Change the list of values for "enum bit mask" in p2 from

V0 = 1 << 0, V1 = 1 << 1, V2 = 1 << 2, V3 = 1 << 3, ....

to

V0 = 0, V1 = 1 << 0, V2 = 1 << 1, V3 = 1 << 2, ....

Here, the names C0, C1, etc. represent bitmask elements for this particular bitmask type. All such non-zero elements have distinct values such that, for any pair Ci and Cj where i != j, Ci & Ci is nonzero and Ci & Cj is zero.

Change bullet 3 of paragraph 4:

TheA non-zero value Y is set in the object X if the expression X & Y is nonzero.

Proposed resolution:

Ammend 17.5.2.1.3 [bitmask.types] p3:

Here, the names C0, C1, etc. represent bitmask elements for this particular bitmask type. All such elements have distinct, non-zero values such that, for any pair Ci and Cj where i != j, Ci & Ci is nonzero and Ci & Cj is zero. Additionally, the value 0 is used to represent an empty bitmask, in which no bitmask elements are set.

Add the following sentence to 28.5.2 [re.matchflag] paragraph 1:

1 The type regex_constants::match_flag_type is an implementation-defined bitmask type (17.5.2.1.3). The constants of that type, except for match_default and format_default, are bitmask elements. The match_default and format_default constants are empty bitmasks. Matching a regular expression against a sequence of characters [first,last) proceeds according to the rules of the grammar specified for the regular expression object, modified according to the effects listed in Table 136 for any bitmask elements set.


1526. [FCD] C++ should not impose thread safety requirements on C99 library implementations

Section: 17.6.5.9 [res.on.data.races] Status: Open Submitter: BSI Opened: 2011-03-24 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

Addresses GB-111

Section 17.6.5.9 [res.on.data.races], Data Race Avoidance, requires the C++ Standard Library to avoid data races that might otherwise result from two threads making calls to C++ Standard Library functions on distinct objects. The C standard library is part of the C++ Standard Library and some C++ Standary library functions (parts of the Localization library, as well as Numeric Conversions in 21.5), are specified to make use of the C standard library. Therefore, the C++ standard indirectly imposes a requirement on the thread safety of the C standard library. However, since the C standard does not address the concept of thread safety conforming C implementations exist that do no provide such guarantees. This conflict needs to be reconciled.

Suggested resolution by national body comment:

remove the requirement to make use of strtol() and sprintf() since these functions depend on the global C locale and thus cannot be made thread safe.

[2011-03-24 Madrid meeting]

Deferred

[ 2011 Bloomington ]

Alisdair: PJ, does this cause a problem in C?

PJ: Every implementation know of is thread safe.

Pete: There a couple of effects that are specified on strtol() and sprintf() which is a problem.

PJ: When C++ talks about C calls it should be "as if" calling the function.

Pete: Culprit is to string stuff. My fault.

PJ: Not your fault. You did what you were told. Distinct resolution to change wording.

Dietmar: What would we break if we change it back?

Pete: Nothing. If implemented on top of thread safe C library you are just fine.

Alisdair: Anyone want to clean up wording and put it back to what Pete gave us?

Alisdair: No volunteers. Do we want to mark as NAD? We could leave it as deferred.

Stefanus: Did original submitter care about this?

Lawrence: There is some work to make local calls thread safe. The resolution would be to call those thread safe version.

Pete: "As if called under single threaded C program"

Action Item (Alisdair): Write wording for this issue.

[2012, Kona]

Re-opened at the request of the concurrency subgroup, who feel there is an issue that needs clarifying for the (planned) 2017 standard.

Rationale:

No consensus to make a change at this time

Proposed resolution:


2003. String exception inconsistency in erase.

Section: 21.4.1 [string.require] Status: Open Submitter: José Daniel García Sánchez Opened: 2010-10-21 Last modified: 2012-09-24

View other active issues in [string.require].

View all other issues in [string.require].

View all issues with Open status.

Discussion:

Clause 21.4.1 [string.require]p3 states:

No erase() or pop_back() member function shall throw any exceptions.

However in 21.4.6.5 [string::erase] p2 the first version of erase has

Throws: out_of_range if pos > size().

[2011-03-24 Madrid meeting]

Beman: Don't want to just change this, can we just say "unless otherwise specified"?

Alisdair: Leave open, but update proposed resolution to say something like "unless otherwise specified".

General agreement that it should be corrected but not a stop-ship.

Action: Update proposed wording for issue 2003 as above, but leave Open.

Proposed resolution:

Update [string.require]p/3:

3 No erase() or pop_back() member function shall throw any exceptions.


2005. unordered_map::insert(T&&) protection should apply to map too

Section: 23.4.4.4 [map.modifiers], 23.4.5.3 [multimap.modifiers], X [unord.map.modifiers], X [unord.multimap.modifiers] Status: Tentatively Ready Submitter: P.J. Plauger Opened: 2010-10-14 Last modified: 2012-09-24

View all issues with Tentatively Ready status.

Discussion:

In [unord.map.modifiers], the signature:

template <class P>
    pair<iterator, bool> insert(P&& obj);

now has an added Remarks paragraph:

Remarks: This signature shall not participate in overload resolution unless P is implicitly convertible to value_type.

The same is true for unordered_multimap.

But neither map nor multimap have this constraint, even though it is a Good Thing(TM) in those cases as well.

[ The submitter suggests: Add the same Remarks clause to [map.modifiers] and [multimap.modifiers]. ]

[ 2010-10-29 Daniel comments: ]

I believe both paragraphs need more cleanup: First, the current Requires element conflict with the Remark; second, it seems to me that the whole single Requires element is intended to be split into a Requires and an Effects element; third, the reference to tuple is incorrect (noticed by Paolo Carlini); fourth, it refers to some non-existing InputIterator parameter relevant for a completely different overload; sixth, the return type of the overload with hint is wrong. The following proposed resolution tries to solve these issues as well and uses similar wording as for the corresponding unordered containers. Unfortunately it has some redundancy over Table 99, but I did not remove the specification because of the more general template parameter P - the Table 99 requirements apply only for an argument identical to value_type.

Daniel's Proposed resolution (not current):

  1. Change 23.4.4.4 [map.modifiers] around p. 1 as indicated:
    template <class P> pair<iterator, bool> insert(P&& x);
    template <class P> pair<iterator, bool> insert(const_iterator position, P&& x);
    

    1 Requires: P shall be convertible to value_type is constructible from std::forward<P>(x)..

    If P is instantiated as a reference type, then the argument x is copied from. Otherwise x is considered to be an rvalue as it is converted to value_type and inserted into the map. Specifically, in such cases CopyConstructible is not required of key_type or mapped_type unless the conversion from P specifically requires it (e.g., if P is a tuple<const key_type, mapped_type>, then key_type must be CopyConstructible). The signature taking InputIterator parameters does not require CopyConstructible of either key_type or mapped_type if the dereferenced InputIterator returns a non-const rvalue pair<key_type,mapped_type>. Otherwise CopyConstructible is required for both key_type and mapped_type.
    ? Effects: Inserts x converted to value_type if and only if there is no element in the container with key equivalent to the key of value_type(x). For the second form, the iterator position is a hint pointing to where the search should start.

    ? Returns: For the first form, the bool component of the returned pair object indicates whether the insertion took place and the iterator component - or for the second form the returned iterator - points to the element with key equivalent to the key of value_type(x).

    ? Complexity: Logarithmic in general, but amortized constant if x is inserted right before position.

    ? Remarks: These signatures shall not participate in overload resolution unless P is implicitly convertible to value_type.

  2. Change 23.4.5.3 [multimap.modifiers] around p. 1 as indicated:
    template <class P> iterator insert(P&& x);
    template <class P> iterator insert(const_iterator position, P&& x);
    

    1 Requires: P shall be convertible to value_type is constructible from std::forward<P>(x).

    If P is instantiated as a reference type, then the argument x is copied from. Otherwise x is considered to be an rvalue as it is converted to value_type and inserted into the map. Specifically, in such cases CopyConstructible is not required of key_type or mapped_type unless the conversion from P specifically requires it (e.g., if P is a tuple<const key_type, mapped_type>, then key_type must be CopyConstructible). The signature taking InputIterator parameters does not require CopyConstructible of either key_type or mapped_type if the dereferenced InputIterator returns a non-const rvalue pair<key_type, mapped_type>. Otherwise CopyConstructible is required for both key_type and mapped_type.
    ? Effects: Inserts x converted to value_type. For the second form, the iterator position is a hint pointing to where the search should start.

    ? Returns: An iterator that points to the element with key equivalent to the key of value_type(x).

    ? Complexity: Logarithmic in general, but amortized constant if x is inserted right before position.

    ? Remarks: These signatures shall not participate in overload resolution unless P is implicitly convertible to value_type.

[ 2010 Batavia: ]

We need is_convertible, not is_constructible, both in ordered and unordered containers.

[ 2011 Bloomington ]

The effects of these inserts can be concisely stated in terms of emplace(). Also, the correct term is "EmplaceConstructible", not "constructible".

New wording by Pablo, eliminating duplicate requirements already implied by the effects clause. Move to Review.

[ 2011-10-02 Daniel comments and refines the proposed wording ]

Unfortunately the template constraints expressed as "P is implicitly convertible to value_type" reject the intended effect to support move-only key types, which was the original intention when the library became move-enabled through the rvalue-reference proposals by Howard (This can clearly be deduced from existing carefully selected wording that emphasizes that CopyConstructible is only required for special situations involving lvalues or const rvalues as arguments). The root of the problem is based on current core rules, where an "implicitly converted" value has copy-initialization semantics. Consider a move-only key type KM, some mapped type T, and a source value p of type P equal to std::pair<KM, T>, this is equivalent to:

std::pair<const KM, T> dest = std::move(p);

Now 8.5 [dcl.init] p16 b6 sb2 says that the effects of this heterogeneous copy-initialization (p has a different type than dest) are as-if a temporary of the target type std::pair<const KM, T> is produced from the rvalue p of type P (which is fine), and this temporary is used to initialize dest. This second step cannot succeed, because we cannot move from const KM to const KM. This means that std::is_convertible<P, std::pair<const KM, T>>::value is false.

But the actual code that is required (with the default allocator) is simply a direct-initialization from P to value_type, so imposing an implicit conversion is more than necessary. Therefore I strongly recommend to reduce the "overload participation" constraint to std::is_constructible<std::pair<const KM, T>, P>::value instead. This change is the only change that has been performed to the previous proposed wording from Pablo shown below.

[2012, Kona]

Moved to Tentatively Ready by the post-Kona issues processing subgroup, after much discussion on Daniel's analysis of Copy Initialization and move semantics, which we ultimately agreed with.

Proposed resolution:

  1. Change 23.4.4.4 [map.modifiers] around p. 1 as indicated:
    template <class P> pair<iterator, bool> insert(P&& x);
    template <class P> pair<iterator, bool> insert(const_iterator position, P&& x);
    
    1 Requires: P shall be convertible to value_type.

    If P is instantiated as a reference type, then the argument x is copied from. Otherwise x is considered to be an rvalue as it is converted to value_type and inserted into the map. Specifically, in such cases CopyConstructible is not required of key_type or mapped_type unless the conversion from P specifically requires it (e.g., if P is a tuple<const key_type, mapped_type>, then key_type must be CopyConstructible). The signature taking InputIterator parameters does not require CopyConstructible of either key_type or mapped_type if the dereferenced InputIterator returns a non-const rvalue pair<key_type,mapped_type>. Otherwise CopyConstructible is required for both key_type and mapped_type.
    ? Effects: The first form is equivalent to return emplace(std::forward<P>(x)). The second form is equivalent to return emplace_hint(position, std::forward<P>(x)).

    ? Remarks: These signatures shall not participate in overload resolution unless std::is_constructible<value_type, P&&>::value is true.

  2. Change 23.4.5.3 [multimap.modifiers] around p. 1 as indicated:
    template <class P> iterator insert(P&& x);
    template <class P> iterator insert(const_iterator position, P&& x);
    
    1 Requires: P shall be convertible to value_type.

    If P is instantiated as a reference type, then the argument x is copied from. Otherwise x is considered to be an rvalue as it is converted to value_type and inserted into the map. Specifically, in such cases CopyConstructible is not required of key_type or mapped_type unless the conversion from P specifically requires it (e.g., if P is a tuple<const key_type, mapped_type>, then key_type must be CopyConstructible). The signature taking InputIterator parameters does not require CopyConstructible of either key_type or mapped_type if the dereferenced InputIterator returns a non-const rvalue pair<key_type, mapped_type>. Otherwise CopyConstructible is required for both key_type and mapped_type.
    ? Effects: The first form is equivalent to return emplace(std::forward<P>(x)). The second form is equivalent to return emplace_hint(position, std::forward<P>(x)).

    ? Remarks: These signatures shall not participate in overload resolution unless std::is_constructible<value_type, P&&>::value is true.

  3. Change 23.5.4.4 [unord.map.modifers] around p. 1 as indicated:
    template <class P>
    pair<iterator, bool> insert(P&& obj);
    
    1 Requires: value_type is constructible from std::forward<P>(obj).

    2 Effects: equivalent to return emplace(std::forward<P>(obj)). Inserts obj converted to value_type if and only if there is no element in the container with key equivalent to the key of value_type(obj).

    3 Returns: The bool component of the returned pair object indicates whether the insertion took place and the iterator component points to the element with key equivalent to the key of value_type(obj).

    4 Complexity: Average case O(1), worst case O(size()).

    53 Remarks: This signature shall not participate in overload resolution unless P is implicitly convertible to value_typestd::is_constructible<value_type, P&&>::value is true.

    template <class P>
    iterator insert(const_iterator hint, P&& obj);
    
    6 Requires: value_type is constructible from std::forward<P>(obj).

    7? Effects: equivalent to return emplace_hint(hint, std::forward<P>(obj)). Inserts obj converted to value_type if and only if there is no element in the container with key equivalent to the key of value_type(obj). The iterator hint is a hint pointing to where the search should start.

    8 Returns: An iterator that points to the element with key equivalent to the key of value_type(obj).

    9 Complexity: Average case O(1), worst case O(size()).

    10? Remarks: This signature shall not participate in overload resolution unless P is implicitly convertible to value_typestd::is_constructible<value_type, P&&>::value is true.

  4. Change 23.5.5.3 [unord.multimap.modifers] around p. 1 as indicated:
    template <class P>
    iterator insert(P&& obj);
    
    1 Requires: value_type is constructible from std::forward<P>(obj).

    2 Effects: equivalent to return emplace(std::forward<P>(obj)). Inserts obj converted to value_type.

    3 Returns: An iterator that points to the element with key equivalent to the key of value_type(obj).

    4 Complexity: Average case O(1), worst case O(size()).

    53 Remarks: This signature shall not participate in overload resolution unless P is implicitly convertible to value_typestd::is_constructible<value_type, P&&>::value is true.

    template <class P>
    iterator insert(const_iterator hint, P&& obj);
    
    6 Requires: value_type is constructible from std::forward<P>(obj).

    7? Effects: equivalent to return emplace_hint(hint, std::forward<P>(obj)). Inserts obj converted to value_type. The iterator hint is a hint pointing to where the search should start.

    8 Returns: An iterator that points to the element with key equivalent to the key of value_type(obj).

    9 Complexity: Average case O(1), worst case O(size()).

    10? Remarks: This signature shall not participate in overload resolution unless P is implicitly convertible to value_typestd::is_constructible<value_type, P&&>::value is true.


2011. Unexpected output required of strings

Section: 21.4.8.9 [string.io] Status: Tentatively Ready Submitter: James Kanze Opened: 2010-07-23 Last modified: 2012-09-24

View all other issues in [string.io].

View all issues with Tentatively Ready status.

Discussion:

What should the following code output?

#include <string>
#include <iostream>
#include <iomanip>

int main() 
{ 
   std::string test("0X1Y2Z"); 
   std::cout.fill('*'); 
   std::cout.setf(std::ios::internal, std::ios::adjustfield); 
   std::cout << std::setw(8) << test << std::endl; 
} 

I would expect "**0X1Y2Z", and this is what the compilers I have access to (VC++, g++ and Sun CC) do. But according to the standard, it should be "0X**1Y2Z":

21.4.8.9 [string.io]/5:

template<class charT, class traits, class Allocator>
  basic_ostream<charT, traits>&
    operator<<(basic_ostream<charT, traits>& os, const basic_string<charT,traits,Allocator>& str);

Effects: Behaves as a formatted output function (27.7.3.6.1 [ostream.formatted.reqmts]). After constructing a sentry object, if this object returns true when converted to a value of type bool, determines padding as described in 22.4.2.2.2 [facet.num.put.virtuals], then inserts the resulting sequence of characters seq as if by calling os.rdbuf()->sputn(seq, n), where n is the larger of os.width() and str.size(); then calls os.width(0).

22.4.2.2.2 [facet.num.put.virtuals]/5:

[…]

Stage 3: A local variable is initialized as

fmtflags adjustfield= (flags & (ios_base::adjustfield));

The location of any padding is determined according to Table 88.

If str.width() is nonzero and the number of charT's in the sequence after stage 2 is less than str.width(), then enough fill characters are added to the sequence at the position indicated for padding to bring the length of the sequence to str.width(). str.width(0) is called.

Table 88 — Fill padding
State Location
adjustfield == ios_base::left pad after
adjustfield == ios_base::right pad before
adjustfield == internal and a sign occurs in the representation pad after the sign
adjustfield == internal and representation after stage 1 began with 0x or 0X pad after x or X
otherwise pad before

Although it's not 100% clear what "the sequence after stage 2" should mean here, when there is no stage 2, the only reasonable assumption is that it is the contents of the string being output. In the above code, the string being output is "0X1Y2Z", which starts with "0X", so the padding should be inserted "after x or X", and not before the string. I believe that this is a defect in the standard, and not in the three compilers I tried.

[ 2010 Batavia (post meeting session) ]

Consensus that all known implementations are consistent, and disagree with the standard. Preference is to fix the standard before implementations start trying to conform to the current spec, as the current implementations have the preferred form. Howard volunteered to drught for Madrid, move to Open.

[2011-03-24 Madrid meeting]

Daniel Krügler volunteered to provide wording, interacting with Dietmar and Bill.

[2011-06-24 Daniel comments and provides wording]

The same problem applies to the output provided by const char* and similar character sequences as of 27.7.3.6.4 [ostream.inserters.character] p. 5. and even for single character output (!) as described in 27.7.3.6.4 [ostream.inserters.character] p. 1, just consider the character value '-' where '-' is the sign character. In this case Table 91 — "Fill padding" requires to pad after the sign, i.e. the output for the program

#include <iostream>
#include <iomanip>

int main() 
{ 
   char c = '-'; 
   std::cout.fill('*'); 
   std::cout.setf(std::ios::internal, std::ios::adjustfield); 
   std::cout << std::setw(2) << c << std::endl; 
} 

According to the current wording this program should output "-*", but all tested implementations output "*-" instead.

I suggest to replace the reference to 22.4.2.2.2 [facet.num.put.virtuals] in all three places. It is not very complicated to describe the padding rules for simple character sequences "inline". A similar approach is used as for the money_put functions.

[ 2011 Bloomington ]

Move to Review, the resolution seems correct but it would be nice if some factoring of the common words were proposed.

[2012, Kona]

Moved to Tentatively Ready by the post-Kona issues processing subgroup.

While better factoring of the common words is desirable, it is also editorial and should not hold up the progress of this issue. As the edits impact two distinct clauses, it is not entirely clear what a better factoring should look like.

Proposed resolution:

The new wording refers to the FDIS numbering.

  1. Change 21.4.8.9 [string.io]/5 as indicated:

    template<class charT, class traits, class Allocator>
      basic_ostream<charT, traits>&
        operator<<(basic_ostream<charT, traits>& os,
                   const basic_string<charT,traits,Allocator>& str);
    

    -5- Effects: Behaves as a formatted output function ([ostream.formatted.reqmts]). After constructing a sentry object, if this object returns true when converted to a value of type bool, determines padding as described in [facet.num.put.virtuals],follows: A charT character sequence is produced, initially consisting of the elements defined by the range [str.begin(), str.end()). If str.size() is less than os.width(), then enough copies of os.fill() are added to this sequence as necessary to pad to a width of os.width() characters. If (os.flags() & ios_base::adjustfield) == ios_base::left is true, the fill characters are placed after the character sequence; otherwise, they are placed before the character sequence. Tthen inserts the resulting sequence of characters seq as if by calling os.rdbuf()->sputn(seq, n), where n is the larger of os.width() and str.size(); then calls os.width(0).

  2. Change 27.7.3.6.4 [ostream.inserters.character]/1 as indicated (An additional editorial fix is suggested for the first prototype declaration):

    template<class charT, class traits>
      basic_ostream<charT,traits>& operator<<(basic_ostream<charT,traits>& out,
                                              charT c});
    template<class charT, class traits>
      basic_ostream<charT,traits>& operator<<(basic_ostream<charT,traits>& out,
                                              char c);
    // specialization
    template<class traits>
      basic_ostream<char,traits>& operator<<(basic_ostream<char,traits>& out,
                                             char c);
    // signed and unsigned
    template<class traits>
      basic_ostream<char,traits>& operator<<(basic_ostream<char,traits>& out,
                                             signed char c);
    template<class traits>
      basic_ostream<char,traits>& operator<<(basic_ostream<char,traits>& out,
                                             unsigned char c);
    

    -1- Effects: Behaves like a formatted inserter (as described in [ostream.formatted.reqmts]) of out. After a sentry object is constructed it inserts characters. In case c has type char and the character type of the stream is not char, then the character to be inserted is out.widen(c); otherwise the character is c. Padding is determined as described in [facet.num.put.virtuals]follows: A character sequence is produced, initially consisting of the insertion character. If out.width() is greater than one, then enough copies of out.fill() are added to this sequence as necessary to pad to a width of out.width() characters. If (out.flags() & ios_base::adjustfield) == ios_base::left is true, the fill characters are placed after the insertion character; otherwise, they are placed before the insertion character. width(0) is called. The insertion character and any required padding are inserted into out; then calls os.width(0).

  3. Change 27.7.3.6.4 [ostream.inserters.character]/5 as indicated:

    template<class charT, class traits>
      basic_ostream<charT,traits>& operator<<(basic_ostream<charT,traits>& out,
                                              const charT* s);
    template<class charT, class traits>
      basic_ostream<charT,traits>& operator<<(basic_ostream<charT,traits>& out,
                                              const char* s);
    template<class traits>
      basic_ostream<char,traits>& operator<<(basic_ostream<char,traits>& out,
                                             const char* s);
    template<class traits>
      basic_ostream<char,traits>& operator<<(basic_ostream<char,traits>& out,
                                             const signed char* s);
    template<class traits>
      basic_ostream<char,traits>& operator<<(basic_ostream<char,traits>& out,
                                             const unsigned char* s);
    

    […]

    -5- Padding is determined as described in [facet.num.put.virtuals]. The n characters starting at s are widened using out.widen ([basic.ios.members])follows: A character sequence is produced, initially consisting of the elements defined by the n characters starting at s widened using out.widen ([basic.ios.members]). If n is less than out.width(), then enough copies of out.fill() are added to this sequence as necessary to pad to a width of out.width() characters. If (out.flags() & ios_base::adjustfield) == ios_base::left is true, the fill characters are placed after the character sequence; otherwise, they are placed before the character sequence. The widened characters and any required padding are inserted into out. Calls width(0).


2012. Associative maps should insert pair, not tuple

Section: 23.4 [associative] Status: Open Submitter: Paolo Carlini Opened: 2010-10-29 Last modified: 2012-09-24

View other active issues in [associative].

View all other issues in [associative].

View all issues with Open status.

Discussion:

I'm seeing something strange in the paragraphs 23.4.4.4 [map.modifiers] and 23.4.5.3 [multimap.modifiers]: they both talk about tuple<const key_type, mapped_type> but I think they should be talking about pair<const key_type, mapped_type> because, among other reasons, a tuple is not convertible to a pair. If I replace tuple with pair everything makes sense to me.

The proposed resolution is obvious.

[ 2010-11-07 Daniel comments ]

This is by far not the only necessary fix within both sub-clauses. For details see the 2010-10-29 comment in 2005.

[2011-03-24 Madrid meeting]

Paolo: Don't think we can do it now.

Daniel K: Agrees.

[ 2011 Bloomington ]

Consensus that this issue will be resolved by 2005, but held open until that issue is resolved.

Proposed resolution:

Apply the resolution proposed by the 2010-10-29 comment in 2005.


2013. Do library implementers have the freedom to add constexpr?

Section: 17.6.5.6 [constexpr.functions] Status: Review Submitter: Matt Austern Opened: 2010-11-12 Last modified: 2012-09-24

View all issues with Review status.

Discussion:

Suppose that a particular function is not tagged as constexpr in the standard, but that, in some particular implementation, it is possible to write it within the constexpr constraints. If an implementer tags such a function as constexpr, is that a violation of the standard or is it a conforming extension?

There are two questions to consider. First, is this allowed under the as-if rule? Second, if it does not fall under as-if, is there (and should there be) any special license granted to implementers to do this anyway, sort of the way we allow elision of copy constructors even though it is detectable by users?

I believe that this does not fall under "as-if", so implementers probably don't have that freedom today. I suggest changing the WP to grant it. Even if we decide otherwise, however, I suggest that we make it explicit.

[ 2011 Bloomington ]

General surprise this was not already in 'Ready' status, and so moved.

[ 2012 Kona ]

Some concern expressed when presented to full committee for the vote to WP status that this issue had been resolved without sufficient thought of the consequences for diverging library implementations, as users may use SFINAE to observe different behavior from otherwise identical code. Issue moved back to Review status, and will be discussed again in Portland with a larger group. Note for Portland: John Spicer has agreed to represent Core's concerns during any such discussion within LWG.

Proposed resolution:

In 17.6.4.6 [constexpr.functions], change paragraph 1 to:

This standard explicitly requires that certain standard library functions are constexpr [dcl.constexpr]. Additionally, an implementation may declare any function to be constexpr if that function's definition satisfies the necessary constraints. Within any header that provides any non-defining declarations of constexpr functions or constructors an implementation shall provide corresponding definitions.


2016. Allocators must be no-throw swappable

Section: 17.6.3.5 [allocator.requirements] Status: Open Submitter: Daniel Krügler Opened: 2010-11-17 Last modified: 2012-09-24

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with Open status.

Discussion:

During the Batavia meeting it turned out that there is a definition hole for types satisfying the Allocators requirements: The problem became obvious when it was discussed whether all swap functions of Containers with internal data handles can be safely tagged with noexcept or not. While it is correct that the implicit swap function of an allocator is required to be a no-throw operation (because move/copy-constructors and assignment operators are required to be no-throw functions), there are no such requirements for specialized swap overloads for a particular allocator.

But this requirement is essential because the Containers are required to support swappable Allocators, when the value allocator_traits<>::propagate_on_container_swap evaluates to true.

[2011-02-10 Alberto, Daniel, and Pablo collaborated on the proposed wording]

The proposed resolution (based on N3225) attempts to solve the following problems:

  1. Table 44 — Allocator requirements, expression rows X::propagate_on_container_copy_assignment, X::propagate_on_container_move_assignment, and X::propagate_on_container_swap only describe operations, but no requirements. In fact, if and only if these compile-time predicates evaluate to true, the additional requirements CopyAssignable, no-throw MoveAssignable, and no-throw lvalue Swappable, respectively, are imposed on the allocator types.
  2. 23.2.1 [container.requirements.general] p. 9 misses to refer to the correct swap conditions: The current wording does not relate to 17.6.3.2 [swappable.requirements] as it should and omits to mention that lvalues shall be swapped. Additional there is one situation described twice in p. 8 and p. 9 (undefined behaviour unless a.get_allocator() == b.get_allocator() or allocator_traits<allocator_type>::propagate_on_container_swap::value == true), which should be cleaned up.

[2011-04-08 Pablo comments]

I'm implementing a version of list now and I actually do find it impossible to write an exception-safe assignment operator unless I can assume that allocator assignment does not throw. (The problem is that I use a sentinel node and I need to allocate a new sentinel using the new allocator without destroying the old one -- then swap the allocator and sentinel pointer in atomically, without risk of an exception leaving one inconsistent with the other.

Please update the proposed resolution to add the nothrow requirement to copy-assignment.

Proposed resolution:

  1. Adapt the following three rows from Table 44 — Allocator requirements:

    Table 44 — Allocator requirements
    Expression Return type Assertion/note
    pre-/post-condition
    Default
    X::propagate_on_container_copy_assignment Identical to or derived from true_type
    or false_type
    true_type only if an allocator of type X should be copied
    when the client container is copy-assigned. See Note B, below.
    false_type
    X::propagate_on_container_move_assignment Identical to or derived from true_type
    or false_type
    true_type only if an allocator of type X should be moved
    when the client container is move-assigned. See Note B, below.
    false_type
    X::propagate_on_container_swap Identical to or derived from true_type
    or false_type
    true_type only if an allocator of type X should be swapped
    when the client container is swapped. See Note B, below.
    false_type
  2. Following 17.6.3.5 [allocator.requirements] p. 3 insert a new normative paragraph:

    Note B: If X::propagate_on_container_copy_assignment::value is true, X shall satisfy the CopyAssignable requirements (Table 39 [copyassignable]) and the copy operation shall not throw exceptions. If X::propagate_on_container_move_assignment::value is true, X shall satisfy the MoveAssignable requirements (Table 38 [moveassignable]) and the move operation shall not throw exceptions. If X::propagate_on_container_swap::value is true, lvalues of X shall be swappable (17.6.3.2 [swappable.requirements]) and the swap operation shall not throw exceptions.

  3. Modify 23.2.1 [container.requirements.general] p. 8 and p. 9 as indicated:

    8 - [..] The allocator may be replaced only via assignment or swap(). Allocator replacement is performed by copy assignment, move assignment, or swapping of the allocator only if allocator_traits<allocator_type>::propagate_on_container_copy_assignment::value, allocator_traits<allocator_type>::propagate_on_container_move_assignment::value, or allocator_traits<allocator_type>::propagate_on_container_swap::value is true within the implementation of the corresponding container operation. The behavior of a call to a container's swap function is undefined unless the objects being swapped have allocators that compare equal or allocator_traits<allocator_type>::propagate_on_container_swap::value is true. In all container types defined in this Clause, the member get_allocator() returns a copy of the allocator used to construct the container or, if that allocator has been replaced, a copy of the most recent replacement.

    9 - The expression a.swap(b), for containers a and b of a standard container type other than array, shall exchange the values of a and b without invoking any move, copy, or swap operations on the individual container elements. Lvalues of aAny Compare, Pred, or Hash objects belonging to a and b shall be swappable and shall be exchanged by unqualified calls to non-member calling swap as described in 17.6.3.2 [swappable.requirements]. If allocator_traits<allocator_type>::propagate_on_container_swap::value is true, then lvalues of allocator_type shall be swappable and the allocators of a and b shall also be exchanged using a an unqualified call to non-member swap call as described in 17.6.3.2 [swappable.requirements]. Otherwise, theythe allocators shall not be swapped, and the behavior is undefined unless a.get_allocator() == b.get_allocator(). Every iterator referring to an element in one container before the swap shall refer to the same element in the other container after the swap. It is unspecified whether an iterator with value a.end() before the swap will have value b.end() after the swap.


2018. regex_traits::isctype Returns clause is wrong

Section: 28.7 [re.traits] Status: Open Submitter: Jonathan Wakely Opened: 2010-11-16 Last modified: 2012-09-24

View all other issues in [re.traits].

View all issues with Open status.

Discussion:

28.7 [re.traits] p. 12 says:

returns true if f bitwise or'ed with the result of calling lookup_classname with an iterator pair that designates the character sequence "w" is not equal to 0 and c == '_'

If the bitmask value corresponding to "w" has a non-zero value (which it must do) then the bitwise or with any value is also non-zero, and so isctype('_', f) returns true for any f. Obviously this is wrong, since '_' is not in every ctype category.

There's a similar problem with the following phrases discussing the "blank" char class.

[2011-05-06: Jonathan Wakely comments and provides suggested wording]

DR 2019 added isblank support to <locale> which simplifies the definition of regex_traits::isctype by removing the special case for the "blank" class.

My suggestion for 2018 is to add a new table replacing the lists of recognized names in the Remarks clause of regex_traits::lookup_classname. I then refer to that table in the Returns clause of regex_traits::isctype to expand on the "in an unspecified manner" wording which is too vague. The conversion can now be described using the "is set" term defined by 17.5.2.1.3 [bitmask.types] and the new table to convey the intented relationship between e.g. [[:digit:]] and ctype_base::digit, which is not actually stated in the FDIS.

The effects of isctype can then most easily be described in code, given an "exposition only" function prototype to do the not-quite-so-unspecified conversion from char_class_type to ctype_base::mask.

The core of LWG 2018 is the "bitwise or'ed" wording which gives the wrong result, always evaluating to true for all values of f. That is replaced by the condition (f&x) == x where x is the result of calling lookup_classname with "w". I believe that's necessary, because the "w" class could be implemented by an internal "underscore" class i.e. x = _Alnum|_Underscore in which case (f&x) != 0 would give the wrong result when f==_Alnum.

The proposed resolution also makes use of ctype::widen which addresses the problem that the current wording only talks about "w" and '_' which assumes charT is char. There's still room for improvement here: the regex grammar in 28.13 [re.grammar] says that the class names in the table should always be recognized, implying that e.g. U"digit" should be recognized by regex_traits<char32_t>, but the specification of regex_traits::lookup_classname doesn't cover that, only mentioning char and wchar_t. Maybe the table should not distinguish narrow and wide strings, but should just have one column and add wording to say that regex_traits widens the name as if by using use_facet<ctype<charT>>::widen().

Another possible improvement would be to allow additional implementation-defined extensions in isctype. An implementation is allowed to support additional class names in lookup_classname, e.g. [[:octdigit:]] for [0-7] or [[:bindigit:]] for [01], but the current definition of isctype provides no way to use them unless ctype_base::mask also supports them.

[2011-05-10: Alberto and Daniel perform minor fixes in the P/R]

[ 2011 Bloomington ]

Consensus that this looks to be a correct solution, and the presentation as a table is a big improvement.

Concern that the middle section wording is a little muddled and confusing, Stefanus volunteered to reword.

Proposed resolution:

This wording is relative to the FDIS.

  1. Modify 28.7 [re.traits] p. 10 as indicated:

    template <class ForwardIterator>
      char_class_type lookup_classname(
        ForwardIterator first, ForwardIterator last, bool icase = false) const;
    

    -9- Returns: an unspecified value that represents the character classification named by the character sequence designated by the iterator range [first,last). If the parameter icase is true then the returned mask identifies the character classification without regard to the case of the characters being matched, otherwise it does honor the case of the characters being matched.(footnote 335) The value returned shall be independent of the case of the characters in the character sequence. If the name is not recognized then returns a value that compares equal to 0.

    -10- Remarks: For regex_traits<char>, at least the names "d", "w", "s", "alnum", "alpha", "blank", "cntrl", "digit", "graph", "lower", "print", "punct", "space", "upper" and "xdigit"narrow character names in Table X shall be recognized. For regex_traits<wchar_t>, at least the names L"d", L"w", L"s", L"alnum", L"alpha", L"blank", L"cntrl", L"digit", L"graph", L"lower", L"print", L"punct", L"space", L"upper" and L"xdigit"wide character names in Table X shall be recognized.

  2. Modify 28.7 [re.traits] p. 12 as indicated:

    bool isctype(charT c, char_class_type f) const;
    

    -11- Effects: Determines if the character c is a member of the character classification represented by f.

    -12- Returns: Converts f into a value m of type std::ctype_base::mask in an unspecified manner, and returns true if use_facet<ctype<charT> >(getloc()).is(m, c) is true. Otherwise returns true if f bitwise or'ed with the result of calling lookup_classname with an iterator pair that designates the character sequence "w" is not equal to 0 and c == '_', or if f bitwise or'ed with the result of calling lookup_classname with an iterator pair that designates the character sequence "blank" is not equal to 0 and c is one of an implementation-defined subset of the characters for which isspace(c, getloc()) returns true, otherwise returns false.except that when f represents membership of a character class named in Table X, the corresponding ctype_base::mask value shall be set in m. Given the function prototype

    template<class C>
       ctype_base::mask convert(typename regex_traits<C>::char_class_type);
    

    the result is determined as if by

    
    ctype_base::mask m = convert<charT>(f);
    const ctype<charT>& ct = use_facet<ctype<charT> >(getloc());
    if (ct.is(m, c))
      return true;
    charT w[1] = { ct.widen('w') };
    char_class_type x = lookup_classname(w, w+1);
    if ((f&x) == x && c == ct.widen('_'))
      return true;
    return false;
    

    [Example:

    
    regex_traits<char> t;
    string d("d");
    string u("upper");
    regex_traits<char>::char_class_type f;
    f = t.lookup_classname(d.begin(), d.end());
    f |= t.lookup_classname(u.begin(), u.end());
    ctype_base::mask m = convert<char>(f); // m == ctype_base::digit|ctype_base::upper
    

    end example]

    [Example:

    
    regex_traits<char> t;
    string w("w");
    regex_traits<char>::char_class_type f;
    f = t.lookup_classname(w.begin(), w.end());
    t.isctype('A', f); // returns true
    t.isctype('_', f); // returns true
    t.isctype(' ', f); // returns false
    

    end example]

  3. At the end of [re.traits] add a new Table X — Character class names and corresponding ctype masks:

    Table X — Character class names and corresponding ctype masks
    Narrow character name Wide character name Corresponding ctype_base::mask value
    "alnum" L"alnum" ctype_base::alnum
    "alpha" L"alpha" ctype_base::alpha
    "blank" L"blank" ctype_base::blank
    "cntrl" L"cntrl" ctype_base::cntrl
    "digit" L"digit" ctype_base::digit
    "d" L"d" ctype_base::digit
    "graph" L"graph" ctype_base::graph
    "lower" L"lower" ctype_base::lower
    "print" L"print" ctype_base::print
    "punct" L"punct" ctype_base::punct
    "space" L"space" ctype_base::space
    "s" L"s" ctype_base::space
    "upper" L"upper" ctype_base::upper
    "w" L"w" ctype_base::alnum
    "xdigit" L"xdigit" ctype_base::xdigit

2035. Output iterator requirements are broken

Section: 24.2.4 [output.iterators] Status: Open Submitter: Daniel Krügler Opened: 2011-02-27 Last modified: 2012-09-24

View other active issues in [output.iterators].

View all other issues in [output.iterators].

View all issues with Open status.

Discussion:

During the Pittsburgh meeting the proposal N3066 became accepted because it fixed several severe issues related to the iterator specification. But the current working draft (N3225) does not reflect all these changes. Since I'm unaware whether every correction can be done editorial, this issue is submitted to take care of that. To give one example: All expressions of Table 108 — "Output iterator requirements" have a post-condition that the iterator is incrementable. This is impossible, because it would exclude any finite sequence that is accessed by an output iterator, such as a pointer to a C array. The N3066 wording changes did not have these effects.

[2011-03-01: Daniel comments:]

This issue has some overlap with the issue 2038 and I would prefer if we could solve both at one location. I suggest the following approach:

  1. The terms dereferencable and incrementable could be defined in a more general way not restricted to iterators (similar to the concepts HasDereference and HasPreincrement from working draft N2914). But on the other hand, all current usages of dereferencable and incrementable are involved with types that satisfy iterator requirements. Thus, I believe that it is sufficient for C++0x to add corresponding definitions to 24.2.1 [iterator.requirements.general] and to let all previous usages of these terms refer to this sub-clause. Since the same problem occurs with the past-the-end iterator, this proposal suggest providing similar references to usages that precede its definition as well.

  2. We also need to ensure that all iterator expressions get either an operational semantics in terms of others or we need to add missing pre- and post-conditions. E.g. we have the following ones without semantics:

    *r++ = o // output iterator
    *r--     // bidirectional iterator
    

    According to the SGI specification these correspond to

    { *r = o; ++r; }                         // output iterator
    { reference tmp = *r; --r; return tmp; } // bidirectional iterator
    

    respectively. Please note especially the latter expression for bidirectional iterator. It fixes a problem that we have for forward iterator as well: Both these iterator categories provide stronger guarantees than input iterator, because the result of the dereference operation is reference, and not only convertible to the value type (The exact form from the SGI documentation does not correctly refer to reference).

[2011-03-14: Daniel comments and updates the suggested wording]

In addition to the before mentioned necessary changes there is another one need, which became obvious due to issue 2042: forward_list<>::before_begin() returns an iterator value which is not dereferencable, but obviously the intention is that it should be incrementable. This leads to the conclusion that imposing dereferencable as a requirement for the expressions ++r is wrong: We only need the iterator to be incrementable. A similar conclusion applies to the expression --r of bidirectional iterators.

[ 2011 Bloomington ]

Consensus this is the correct direction, but there are (potentially) missing incrementable preconditions on some table rows, and the Remarks on when an output iterator becomes dereferencable are probably better handled outside the table, in a manner similar to the way we word for input iterators.

There was some concern about redundant pre-conditions when the operational semantic is defined in terms of operations that have preconditions, and a similar level of concern over dropping such redundancies vs. applying a consistent level of redundant specification in all the iterator tables. Wording clean-up in either direction would be welcome.

[2011-08-18: Daniel adapts the proposed resolution to honor the Bloomington request]

There is only a small number of further changes suggested to get rid of superfluous requirements and essentially non-normative assertions. Operations should not have extra pre-conditions, if defined by "in-terms-of" semantics, see e.g. a != b or a->m for Table 107. Further, some remarks, that do not impose anything or say nothing new have been removed, because I could not find anything helpful they provide. E.g. consider the remarks for Table 108 for the operations dereference-assignment and preincrement: They don't provide additional information say nothing surprising. With the new pre-conditions and post-conditions it is implied what the remarks intend to say.

[ 2011-11-03: Some observations from Alexander Stepanov via c++std-lib-31405 ]

The following sentence is dropped from the standard section on OutputIterators:

"In particular, the following two conditions should hold: first, any iterator value should be assigned through before it is incremented (this is, for an output iterator i, i++; i++; is not a valid code sequence); second, any value of an output iterator may have at most one active copy at any given time (for example, i = j; *++i = a; *j = b; is not a valid code sequence)."

[ 2011-11-04: Daniel comments and improves the wording ]

In regard to the first part of the comment, the intention of the newly proposed wording was to make clear that for the expression

*r = o

we have the precondition dereferenceable and the post-condition incrementable. And for the expression

++r

we have the precondition incrementable and the post-condition dereferenceable or past-the-end. This should not allow for a sequence like i++; i++; but I agree that it doesn't exactly say that.

In regard to the second point: To make this point clearer, I suggest to add a similar additional wording as we already have for input iterator to the "Assertion/note" column of the expression ++r:

"Post: any copies of the previous value of r are no longer required to be dereferenceable or incrementable."

The proposed has been updated to honor the observations of Alexander Stepanov.

Proposed resolution:

  1. Add a reference to 24.2.1 [iterator.requirements.general] to the following parts of the library preceding Clause 24 Iterators library: (I stopped from 23.2.5 [unord.req] on, because the remaining references are the concrete containers)

    1. 17.6.3.2 [swappable.requirements] p5:

      -5- A type X satisfying any of the iterator requirements (24.2) is ValueSwappable if, for any dereferenceable (24.2.1 [iterator.requirements.general]) object x of type X, *x is swappable.

    2. 17.6.3.5 [allocator.requirements], Table 27 — "Descriptive variable definitions", row with the expression c:

      a dereferenceable (24.2.1 [iterator.requirements.general]) pointer of type C*

    3. 20.6.3.2 [pointer.traits.functions]:

      Returns: The first template function returns a dereferenceable (24.2.1 [iterator.requirements.general]) pointer to r obtained by calling Ptr::pointer_to(r); […]

    4. 21.4.3 [string.iterators] p. 2:

      Returns: An iterator which is the past-the-end value (24.2.1 [iterator.requirements.general]).

    5. 22.4.5.1.2 [locale.time.get.virtuals] p. 11:

      iter_type do_get(iter_type s, iter_type end, ios_base& f,
        ios_base::iostate& err, tm *t, char format, char modifier) const;
      

      Requires: t shall be dereferenceable (24.2.1 [iterator.requirements.general]).

    6. 23.2.1 [container.requirements.general] p. 6:

      […] end() returns an iterator which is the past-the-end (24.2.1 [iterator.requirements.general]) value for the container. […]

    7. 23.2.3 [sequence.reqmts] p. 3:

      […] q denotes a valid dereferenceable (24.2.1 [iterator.requirements.general]) const iterator to a, […]

    8. 23.2.4 [associative.reqmts] p. 8 (I omit intentionally one further reference in the same sub-clause):

      […] q denotes a valid dereferenceable (24.2.1 [iterator.requirements.general]) const iterator to a, […]

    9. 23.2.5 [unord.req] p. 10 (I omit intentionally one further reference in the same sub-clause):

      […] q and q1 are valid dereferenceable (24.2.1 [iterator.requirements.general]) const iterators to a, […]

  2. Edit 24.2.1 [iterator.requirements.general] p. 5 as indicated (The intent is to properly define incrementable and to ensure some further library guarantee related to past-the-end iterator values):

    -5- 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 sequence. These values are called past-the-end values. Values of an iterator i for which the expression *i is defined are called dereferenceable. Values of an iterator i for which the expression ++i is defined are called incrementable. The library never assumes that past-the-end values are dereferenceable or incrementable. Iterators can also have singular values that are not associated with any sequence. […]

  3. Modify the column contents of Table 106 — "Iterator requirements", 24.2.2 [iterator.iterators], as indicated:

    Table 106 — Iterator requirements
    Expression Return type Operational semantics Assertion/note
    pre-/post-condition
    *r reference   pre: r is dereferenceable.
    ++r X&   pre: r is incrementable.
  4. Modify the column contents of Table 107 — "Input iterator requirements", 24.2.3 [input.iterators], as indicated [Rationale: The wording changes attempt to define a minimal "independent" set of operations, namely *a and ++r, and to specify the semantics of the remaining ones. This approach seems to be in agreement with the original SGI specificationend rationale]:

    Table 107 — Input iterator requirements (in addition to Iterator)
    Expression Return type Operational semantics Assertion/note
    pre-/post-condition
    a != b contextually
    convertible to bool
    !(a == b) pre: (a, b) is in the domain
    of ==.
    *a convertible to T   pre: a is dereferenceable.
    The expression
    (void)*a, *a is equivalent
    to *a.
    If a == b and (a,b) is in
    the domain of == then *a is
    equivalent to *b.
    a->m   (*a).m pre: a is dereferenceable.
    ++r X&   pre: r is dereferenceableincrementable.
    post: r is dereferenceable or
    r is past-the-end.
    post: any copies of the
    previous value of r are no
    longer required either to be
    dereferenceable, incrementable,
    or to be in the domain of ==.
    (void)r++   (void)++r equivalent to (void)++r
    *r++ convertible to T { T tmp = *r;
    ++r;
    return tmp; }
     
  5. Modify the column contents of Table 108 — "Output iterator requirements", 24.2.4 [output.iterators], as indicated [Rationale: The wording changes attempt to define a minimal "independent" set of operations, namely *r = o and ++r, and to specify the semantics of the remaining ones. This approach seems to be in agreement with the original SGI specificationend rationale]:

    Table 108 — Output iterator requirements (in addition to Iterator)
    Expression Return type Operational semantics Assertion/note
    pre-/post-condition
    *r = o result is not used   pre: r is dereferenceable.
    Remark: After this operation
    r is not required to be
    dereferenceable and any copies of
    the previous value of r are no
    longer required to be dereferenceable
    or incrementable.

    post: r is incrementable.
    ++r X&   pre: r is incrementable.
    &r == &++r.
    Remark: After this operation
    r is not required to be
    dereferenceable.
    Remark: After this operation
    r is not required to be
    incrementable and any copies of
    the previous value of r are no
    longer required to be dereferenceable
    or incrementable.

    post: r is dereferenceable
    or r is past-the-end
    incrementable.
    r++ convertible to const X& { X tmp = r;
    ++r;
    return tmp; }
    Remark: After this operation
    r is not required to be
    dereferenceable.
    post: r is incrementable.
    *r++ = o result is not used { *r = o; ++r; } Remark: After this operation
    r is not required to be
    dereferenceable.
    post: r is incrementable.
  6. Modify the column contents of Table 109 — "Forward iterator requirements", 24.2.5 [forward.iterators], as indicated [Rationale: Since the return type of the expression *r++ is now guaranteed to be type reference, the implied operational semantics from input iterator based on value copies is wrong — end rationale]

    Table 109 — Forward iterator requirements (in addition to input iterator)
    Expression Return type Operational semantics Assertion/note
    pre-/post-condition
    r++ convertible to const X& { X tmp = r;
    ++r;
    return tmp; }
     
    *r++ reference { reference tmp = *r;
    ++r;
    return tmp; }
     
  7. Modify the column contents of Table 110 — "Bidirectional iterator requirements", 24.2.6 [bidirectional.iterators], as indicated:

    Table 110 — Bidirectional iterator requirements (in addition to forward iterator)
    Expression Return type Operational semantics Assertion/note
    pre-/post-condition
    --r X&   pre: there exists s such that
    r == ++s.
    post: r is dereferenceableincrementable.
    --(++r) == r.
    --r == --s implies r == s.
    &r == &--r.
    r-- convertible to const X& { X tmp = r;
    --r;
    return tmp; }
     
    *r-- reference { reference tmp = *r;
    --r;
    return tmp; }
     

2038. Missing definition for incrementable iterator

Section: 24.2.4 [output.iterators] Status: Open Submitter: Pete Becker Opened: 2011-02-27 Last modified: 2012-09-24

View other active issues in [output.iterators].

View all other issues in [output.iterators].

View all issues with Open status.

Discussion:

In comp.lang.c++, Vicente Botet raises the following questions:

"In "24.2.4 Output iterators" there are 3 uses of incrementable. I've not found the definition. Could some one point me where it is defined?

Something similar occurs with dereferenceable. While the definition is given in "24.2.1 In general" it is used several times before.

Shouldn't these definitions be moved to some previous section?"

He's right: both terms are used without being properly defined.

There is no definition of "incrementable".

While there is a definition of "dereferenceable", it is, in fact, a definition of "dereferenceable iterator". "dereferenceable" is used throughout Clause 23 (Containers) before its definition in Clause 24. In almost all cases it's referring to iterators, but in 17.6.3.2 [swappable.requirements] there is a mention of "dereferenceable object"; in 17.6.3.5 [allocator.requirements] the table of Descriptive variable definitions refers to a "dereferenceable pointer"; 20.6.3.2 [pointer.traits.functions] refers to a "dereferenceable pointer"; in 22.4.5.1.2 [locale.time.get.virtuals]/11 (do_get) there is a requirement that a pointer "shall be dereferenceable". In those specific cases it is not defined.

[2011-03-02: Daniel comments:]

I believe that the currently proposed resolution of issue 2035 solves this issue as well.

[ 2011 Bloomington ]

Agree with Daniel, this will be handled by the resolution of 2035.

Proposed resolution:


2048. Unnecessary mem_fn overloads

Section: 20.8 [function.objects], 20.8.10 [func.memfn] Status: Tentatively Ready Submitter: Jonathan Wakely Opened: 2011-04-18 Last modified: 2012-09-24

View other active issues in [function.objects].

View all other issues in [function.objects].

View all issues with Tentatively Ready status.

Discussion:

The mem_fn overloads for member functions are redundant and misleading and should be removed from the post-C++11 WP.

I believe the history of the overloads is as follows:

In TR1 and in C++0x prior to the N2798 draft, mem_fn was specified by a single signature:

template<class R, class T> unspecified mem_fn(R T::* pm);

and was accompanied by the remark "Implementations may implement mem_fn as a set of overloaded function templates." This remark predates variadic templates and was presumably to allow implementations to provide overloads for a limited number of function parameters, to meet the implementation-defined limit on numbers of template parameters.

N2770 "Concepts for the C++0x Standard Library: Utilities" added separate overloads for pointers to member functions, apparently so that function parameters would require the CopyConstructible concept (those overloads first appeared in N2322.) The overloads failed to account for varargs member functions (i.e. those declared with an ellipsis in the parameter-declaration-clause) e.g.

struct S {
 int f(int, ...);
};

Syntactically such a function would be handled by the original mem_fn(R T::* pm) signature, the only minor drawback being that there would be no CopyConstructible requirement on the parameter list. (Core DR 547 clarifies that partial specializations can be written to match cv-qualified and ref-qualified functions to support the case where R T::* matches a pointer to member function type.)

LWG issue 920 pointed out that additional overloads were missing for member functions with ref-qualifiers. These were not strictly necessary, because such functions are covered by the mem_fn(R T::* pm) signature.

Concepts were removed from the draft and N3000 restored the original single signature and accompanying remark.

LWG 1230 was opened to strike the remark again and to add an overload for member functions (this overload was unnecessary for syntactic reasons and insufficient as it didn't handle member functions with cv-qualifiers and/or ref-qualifiers.)

920 (and 1230) were resolved by restoring a full set of (non-concept-enabled) overloads for member functions with cv-qualifiers and ref-qualifiers, but as in the concept-enabled draft there were no overloads for member functions with an ellipsis in the parameter-declaration-clause. This is what is present in the FDIS.

Following the thread beginning with message c++std-lib-30675, it is my understanding that all the mem_fn overloads for member functions are unnecessary and were only ever added to allow concept requirements. I'm not aware of any reason implementations cannot implement mem_fn as a single function template. Without concepts the overloads are redundant, and the absence of overloads for varargs functions can be interpreted to imply that varargs functions are not intended to work with mem_fn. Clarifying the intent by adding overloads for varargs functions would expand the list of 12 redundant overloads to 24, it would be much simpler to remove the 12 redundant overloads entirely.

[Bloomington, 2011]

Move to Review.

The issue and resolution appear to be correct, but there is some concern that the wording of INVOKE may be different depending on whether you pass a pointer-to-member-data or pointer-to-member-function. That might make the current wording necessary after all, and then we might need to add the missing elipsis overloads.

There was some concern that the Remark confirming implementors had freedom to implement this as a set of overloaded functions may need to be restored if we delete the specification for these overloads.

[2012, Kona]

Moved to Tentatively Ready by the post-Kona issues processing subgroup.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change the <functional> synopsis 20.8 [function.objects] p. 2 as follows:

    namespace std {
      […]
      // [func.memfn], member function adaptors:
      template<class R, class T> unspecified mem_fn(R T::*);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...));
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) const);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) volatile);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) const volatile);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) &);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) const &);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) volatile &);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) const volatile &);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) &&);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) const &&);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) volatile &&);
      template<class R, class T, class... Args>
      unspecified mem_fn(R (T::*)(Args...) const volatile &&);
    
      […]
    }
    
  2. Change 20.8.10 [func.memfn] as follows:

    template<class R, class T> unspecified mem_fn(R T::*);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...));
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) const);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) volatile);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) const volatile);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) &);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) const &);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) volatile &);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) const volatile &);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) &&);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) const &&);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) volatile &&);
    template<class R, class T, class... Args>
    unspecified mem_fn(R (T::*)(Args...) const volatile &&);
    

2049. is_destructible is underspecified

Section: 20.9.4.3 [meta.unary.prop] Status: Tentatively Ready Submitter: Daniel Krügler Opened: 2011-04-18 Last modified: 2012-09-24

View other active issues in [meta.unary.prop].

View all other issues in [meta.unary.prop].

View all issues with Tentatively Ready status.

Discussion:

The conditions for the type trait is_destructible to be true are described in Table 49 — Type property predicates:

For a complete type T and given
template <class U> struct test { U u; };,
test<T>::~test() is not deleted.

This specification does not say what the result would be for function types or for abstract types:

  1. For an abstract type X the instantiation test<X> is already ill-formed, so we cannot say anything about whether the destructor would be deleted or not.
  2. In regard to function types, there exists a special rule in the core language, 14.3.1 [temp.arg.type] p. 3, which excludes member functions to be declared via the type of the template parameter:

    If a declaration acquires a function type through a type dependent on a template-parameter and this causes a declaration that does not use the syntactic form of a function declarator to have function type, the program is ill-formed.

    [ Example:

    template<class T> struct A {
      static T t;
    };
    typedef int function();
    A<function> a; // ill-formed: would declare A<function>::t
                   // as a static member function
    

    end example ]

    which has the same consequence as for abstract types, namely that the corresponding instantiation of test is already ill-formed and we cannot say anything about the destructor.

To solve this problem, I suggest to specify function types as trivially and nothrowing destructible, because above mentioned rule is very special for templates. For non-templates, a typedef can be used to introduce a member as member function as clarified in 8.3.5 [dcl.fct] p. 10.

For abstract types, two different suggestions have been brought to my attention: Either declare them as unconditionally non-destructible or check whether the expression

std::declval<T&>().~T()

is well-formed in an unevaluated context. The first solution is very easy to specify, but the second version has the advantage for providing more information to user-code. This information could be quite useful, if generic code is supposed to invoke the destructor of a reference to a base class indirectly via a delete expression, as suggested by Howard Hinnant:

template <class T>
my_pointer<T>::~my_pointer() noexcept(is_nothrow_destructible<T>::value)
{
   delete ptr_;
}

Additional to the is_destructible traits, its derived forms is_trivially_destructible and is_nothrow_destructible are similarly affected, because their wording refers to "the indicated destructor" and probably need to be adapted as well.

[ 2011 Bloomington ]

After discussion about to to handle the exceptional cases of reference types, function types (available by defererencing a function pointer) and void types, Howard supplied proposed wording.

[ 2011-08-20 Daniel comments and provides alternatives wording ]

The currently proposed wording would have the consequence that every array type is not destructible, because the pseudo-destructor requires a scalar type with the effect that the expression

std::declval<T&>().~T()

is not well-formed for e.g. T equal to int[3]. The intuitive solution to fix this problem would be to adapt the object type case to refer to the expression

std::declval<U&>().~U()

with U equal to remove_all_extents<T>::type, but that would have the effect that arrays of unknown bounds would be destructible, if the element type is destructible, which was not the case before (This was intentionally covered by the special "For a complete type T" rule in the FDIS).

Suggestion: Use the following definition instead:

Let U be remove_all_extents<T>::type.
For incomplete types and function types, is_destructible<T>::value is false.
For object types, if the expression std::declval<U&>().~U() is well-formed
when treated as an unevaluated operand (Clause 5), then is_destructible<T>::value
is true, otherwise it is false.
For reference types, is_destructible<T>::value is true.

This wording also harmonizes with the "unevaluated operand" phrase used in other places, there does not exist the definition of an "unevaluated context"

Note: In the actually proposed wording this wording has been slightly reordered with the same effects.

Howard's (old) proposed resolution:

Update 20.9.4.3 [meta.unary.prop], table 49:

template <class T> struct is_destructible; For a complete type T and given template <class U> struct test { U u; };, test<T>::~test() is not deleted.
For object types, if the expression: std::declval<T&>().~T() is well-formed in an unevaluated context then is_destructible<T>::value is true, otherwise it is false.
For void types, is_destructible<T>::value is false.
For reference types, is_destructible<T>::value is true.
For function types, is_destructible<T>::value is false.
T shall be a complete type, (possibly cv-qualified) void, or an array of unknown bound.

[2012, Kona]

Moved to Tentatively Ready by the post-Kona issues processing subgroup.

Proposed resolution:

Update 20.9.4.3 [meta.unary.prop], table 49:

template <class T> struct is_destructible; For a complete type T and given template <class U> struct test { U u; };, test<T>::~test() is not deleted.
For reference types, is_destructible<T>::value is true.
For incomplete types and function types, is_destructible<T>::value is false.
For object types and given U equal to remove_all_extents<T>::type,
if the expression std::declval<U&>().~U() is well-formed when treated as an
unevaluated operand (Clause 5 [expr]), then is_destructible<T>::value is true,
otherwise it is false.
T shall be a complete type, (possibly cv-qualified) void, or an array of unknown bound.

2052. Mixup between mapped_type and value_type for associative containers

Section: 23.2.4 [associative.reqmts] Status: Open Submitter: Marc Glisse Opened: 2011-05-04 Last modified: 2012-09-24

View all other issues in [associative.reqmts].

View all issues with Open status.

Discussion:

(this is basically reopening the first part of issue 2006, as discussed in the thread starting at c++std-lib-30698 )

Section 23.2.4 [associative.reqmts]

In Table 102, several uses of T (which means mapped_type here) should be value_type instead. This is almost editorial. For instance:

a_uniq.emplace(args)

Requires: T shall be EmplaceConstructible into X 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.

[ 2011 Bloomington ]

Not even an exhaustive list of problem locations. No reason to doubt issue.

Pablo agrees to provide wording.

[ 2011-09-04 Pablo Halpern provides improved wording ]

Proposed resolution:

In both section 23.2.4 [associative.reqmts] Table 102 and 23.2.5 [unord.req], Table 103, make the following text replacements:

Original text, in FDIS Replacement text
T is CopyInsertable into X and CopyAssignable. value_type is CopyInsertable into X, key_type is CopyAssignable, and mapped_type is CopyAssignable (for containers having a mapped_type)
T is CopyInsertable value_type is CopyInsertable
T shall be CopyInsertable value_type shall be CopyInsertable
T shall be MoveInsertable value_type shall be MoveInsertable
T shall be EmplaceConstructible value_type shall be EmplaceConstructible
T object value_type object

[ Notes to the editor: The above are carefully selected phrases that can be used for global search-and-replace within the specified sections without accidentally making changes to correct uses T. ]


2054. time_point constructors need to be constexpr

Section: 20.11.6 [time.point] Status: Open Submitter: Anthony Williams Opened: 2011-05-13 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

In 20.11.6 [time.point], time_point::min() and time_point::max() are listed as constexpr. However, time_point has no constexpr constructors, so is not a literal type, and so these functions cannot be constexpr without adding a constexpr constructor for implementation purposes.

Proposed resolution: Add constexpr to the constructors of time_point. The effects of the constructor template basically imply that the member function time_since_epoch() is intended to be constexpr as well.

Proposed resolution:

This wording is relative to the FDIS.

  1. Alter the class template definition in 20.11.6 [time.point] as follows:

    template <class Clock, class Duration = typename Clock::duration>
    class time_point {
      […]
    public:
      // 20.11.6.1, construct:
      constexpr time_point(); // has value epoch
      constexpr explicit time_point(const duration& d); // same as time_point() + d
      template <class Duration2>
        constexpr time_point(const time_point<clock, Duration2>& t);
    
      // 20.11.6.2, observer:
      constexpr duration time_since_epoch() const;
      […]
    };
    
  2. Alter the declarations in 20.11.6.1 [time.point.cons]:

    constexpr time_point();
    

    -1- Effects: Constructs an object of type time_point, initializing d_ with duration::zero(). Such a time_point object represents the epoch.

    constexpr explicit time_point(const duration& d);
    

    -2- Effects: Constructs an object of type time_point, initializing d_ with d. Such a time_point object represents the epoch + d.

    template <class Duration2>
      constexpr time_point(const time_point<clock, Duration2>& t);
    

    -3- Remarks: This constructor shall not participate in overload resolution unless Duration2 is implicitly convertible to duration.

    -4- Effects: Constructs an object of type time_point, initializing d_ with t.time_since_epoch().

  3. Alter the declaration in 20.11.6.2 [time.point.observer]:

    constexpr duration time_since_epoch() const;
    

    -1- Returns: d_.


2056. future_errc enums start with value 0 (invalid value for broken_promise)

Section: 30.6.1 [futures.overview] Status: Tentatively Ready Submitter: Nicolai Josuttis Opened: 2011-05-18 Last modified: 2012-09-24

View all other issues in [futures.overview].

View all issues with Tentatively Ready status.

Discussion:

In 30.6.1 [futures.overview] enum class future_errc is defined as follows:

enum class future_errc {
  broken_promise,
  future_already_retrieved,
  promise_already_satisfied,
  no_state
};

With this declaration broken_promise has value 0, which means that for a future_error f with this code

f.code().operator bool()

yields false, which makes no sense. 0 has to be reserved for "no error". So, the enums defined here have to start with 1.

Howard, Anthony, and Jonathan have no objections.

[Discussion in Bloomington 2011-08-16]

Previous resolution:

This wording is relative to the FDIS.

  1. In 30.6.1 [futures.overview], header <future> synopsis, fix the declaration of future_errc as follows:

    namespace std {
      enum class future_errc {
        broken_promise,
        future_already_retrieved = 1,
        promise_already_satisfied,
        no_state,
        broken_promise
      };
      […]
    }
    

Is this resolution overspecified? These seem to be all implementation-defined. How do users add new values and not conflict with established error codes?

PJP proxy says: over-specified. boo.

Other error codes: look for is_error_code_enum specializations. Only one exists io_errc

Peter: I don't see any other parts of the standard that specify error codes where we have to do something similar.

Suggest that for every place where we add an error code, the following:

  1. no zero values
  2. all implementation defined values, so future_already_retrieved = implementation_defined
  3. values are distinct

[2012, Kona]

Moved to Tentatively Ready by the post-Kona issues processing subgroup.

Proposed resolution:

This wording is relative to the FDIS.

In 30.6.1 [futures.overview], header <future> synopsis, fix the declaration of future_errc as follows:

namespace std {
  enum class future_errc {
    broken_promise = implementation defined,
    future_already_retrieved = implementation defined,
    promise_already_satisfied = implementation defined,
    no_state = implementation defined
  };
  […]
}

In 30.6.1 [futures.overview], header <future> synopsis, add a paragraph after paragraph 2 as follows:

The enum values of future_errc are distinct and not zero.

2057. time_point + duration semantics should be made constexpr conforming

Section: 20.11.6.5 [time.point.nonmember] Status: Open Submitter: Daniel Krügler Opened: 2011-05-21 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

It has been observed by LWG 2054 that the specification of some time_point member functions already imply that time_point needs to be a literal type and suggests to specify the constructors and the member function time_since_epoch() as constexpr functions at the minimum necessary. Adding further constexpr specifier to other operations should clearly be allowed and should probably be done as well. But to allow for further constexpr functions in the future requires that their semantics is compatible to operations allowed in constexpr functions. This is already fine for all operations, except this binary plus operator:

template <class Clock, class Duration1, class Rep2, class Period2>
time_point<Clock, typename common_type<Duration1, duration<Rep2, Period2>>::type>
operator+(const time_point<Clock, Duration1>& lhs, const duration<Rep2, Period2>& rhs);

-1- Returns: CT(lhs) += rhs, where CT is the type of the return value.

for similar reasons as those mentioned in 2020. The semantics should be fixed to allow for making them constexpr. This issue should also be considered as a placeholder for a request to make the remaining time_point operations similarly constexpr as had been done for duration.

Proposed resolution:

This wording is relative to the FDIS.

  1. In 20.11.6.5 [time.point.nonmember], p.1 change the Returns element semantics as indicated:

    template <class Clock, class Duration1, class Rep2, class Period2>
    time_point<Clock, typename common_type<Duration1, duration<Rep2, Period2>>::type>
    operator+(const time_point<Clock, Duration1>& lhs, const duration<Rep2, Period2>& rhs);
    

    -1- Returns: CT(lhs) += rhsCT(lhs.time_since_epoch() + rhs), where CT is the type of the return value.


2058. valarray and begin/end

Section: 26.6 [numarray] Status: Tentatively Ready Submitter: Gabriel Dos Reis Opened: 2011-05-17 Last modified: 2012-09-24

View all other issues in [numarray].

View all issues with Tentatively Ready status.

Discussion:

It was just brought to my attention that the pair of functions begin/end were added to valarray component. Those additions strike me as counter to the long standing agreement that valarray<T> is not yet another container. Valarray values are in general supposed to be treated as a whole, and as such has a loose specification allowing expression template techniques.

The addition of these functions sound to me as making it much harder (or close to impossible) to effectively use expression templates as implementation techniques, for no clear benefits.

My recommendation would be to drop begin/end - or at least for the const valarray<T>& version. I strongly believe those are defects.

[This issue was discussed on the library reflector starting from c++std-lib-30761. Some of the key conclusions of this discussion were:]

  1. The begin/end members were added to allow valarray to participate in the new range-based for-loop by n2930 and not to make them container-like.
  2. It is currently underspecified when the iterator values returned from begin/end become invalidated. To fix this, these invalidation rules need at least to reflect the invalidation rules of the references returned by the operator[] overloads of valarray (26.6.2.4 [valarray.access]).
  3. A further problem is that the requirements expressed in 26.6.1 [valarray.syn] p.3-5 enforce an implementation to provide further overloads of begin/end, if the replacement type technique is used (which was clearly part of the design of valarray). Providing such additional overloads would also lead to life-time problems in examples like begin(x + y) where x and y are expressions involving valarray objects. To fix this, the begin/end overloads could be explicitly excluded from the general statements of 26.6.1 [valarray.syn] p.3-5. This would make it unspecified whether the expression begin(x + y) would be well-formed, portable code would need to write this as begin(std::valarray<T>(x + y)).

[ 2011 Bloomington ]

The intent of these overloads is entirely to support the new for syntax, and not to create new containers.

Stefanus provides suggested wording.

[2012, Kona]

Moved to Tenatively Ready by post-meeting issues processing group, after confirmation from Gaby.

Proposed resolution:

In 26.6.1 [valarray.syn]/4, make the following insertion:

4 Implementations introducing such replacement types shall provide additional functions and operators as follows:

In 26.6.10 [valarray.range], make the following insertion:

1 In the begin and end function templates that follow, unspecified1 is a type that meets the requirements of a mutable random access iterator (24.2.7) whose value_type is the template parameter T and whose reference type is T&. unspecified2 is a type that meets the requirements of a constant random access iterator (24.2.7) whose value_type is the template parameter T and whose reference type is const T&.

2 The iterators returned by begin and end for an array are guaranteed to be valid until the member function resize(size_t, T) (26.6.2.8 [valarray.members]) is called for that array or until the lifetime of that array ends, whichever happens first.


2059. C++0x ambiguity problem with map::erase

Section: 23.4.4 [map] Status: Open Submitter: Christopher Jefferson Opened: 2011-05-18 Last modified: 2012-09-24

View all other issues in [map].

View all issues with Open status.

Discussion:

map::erase (and several related methods) took an iterator in C++03, but take a const_iterator in C++0x. This breaks code where the map's key_type has a constructor which accepts an iterator (for example a template constructor), as the compiler cannot choose between erase(const key_type&) and erase(const_iterator).

#include <map>

struct X
{
  template<typename T>
  X(T&) {}
};

bool operator<(const X&, const X&) { return false; }

void erasor(std::map<X,int>& s, X x)
{
  std::map<X,int>::iterator it = s.find(x);
  if (it != s.end())
    s.erase(it);
}

[ 2011 Bloomington ]

This issue affects only associative container erase calls, and is not more general, as these are the only functions that are also overloaded on another single arguement that might cause confusion - the erase by key method. The complete resolution should simply restore the iterator overload in addition to the const_iterator overload for all eight associative containers.

Proposed wording supplied by Alan Talbot, and moved to Review.

[2012, Kona]

Moved back to Open by post-meeting issues processing group.

Pablo very unhappy about case of breaking code with ambiguous conversion between both iterator types.

Alisdair strongly in favor of proposed resolution, this change from C++11 bit Chris in real code, and it took a while to track down the cause.

Move to open, bring in front of a larger group

Proposed wording from Jeremiah: erase(key) shall not participate in overload resolution if iterator is convertible to key. Note that this means making erase(key) a template-method

Poll Chris to find out if he already fixed his code, or fixed his library

Jeremiah - allow both overloads, but enable_if the const_iterator form as a template, requiring is_same to match only const_iterator.

Poll PJ to see if he has already applied this fix?

Proposed resolution:

Editorial note: The following things are different between 23.2.4 [associative.reqmts] p.8 and 23.2.5 [unord.req] p.10. These should probably be reconciled.

  1. First uses the convention "denotes"; second uses the convention "is".
  2. First redundantly says: "If no such element exists, returns a.end()." in erase table entry, second does not.

23.2.4 [associative.reqmts] Associative containers

8 In Table 102, X denotes an associative container class, a denotes a value of X, a_uniq denotes a value of X when X supports unique keys, a_eq denotes a value of X when X supports multiple keys, u denotes an identifier, i and j satisfy input iterator requirements and refer to elements implicitly convertible to value_type, [i,j) denotes a valid range, p denotes a valid const iterator to a, q denotes a valid dereferenceable const iterator to a, r denotes a valid dereferenceable iterator to a, [q1, q2) denotes a valid range of const iterators in a, il designates an object of type initializer_list<value_type>, t denotes a value of X::value_type, k denotes a value of X::key_type and c denotes a value of type X::key_compare. A denotes the storage allocator used by X, if any, or std::allocator<X::value_type> otherwise, and m denotes an allocator of a type convertible to A.

23.2.4 [associative.reqmts] Associative containers Table 102

Add row:

a.erase(r) iterator erases the element pointed to by r. Returns an iterator pointing to the element immediately following r prior to the element being erased. If no such element exists, returns a.end(). amortized constant

23.2.5 [unord.req] Unordered associative containers

10 In table 103: X is an unordered associative container class, a is an object of type X, b is a possibly const object of type X, a_uniq is an object of type X when X supports unique keys, a_eq is an object of type X when X supports equivalent keys, i and j are input iterators that refer to value_type, [i, j) is a valid range, p and q2 are valid const iterators to a, q and q1 are valid dereferenceable const iterators to a, r is a valid dereferenceable iterator to a, [q1,q2) is a valid range in a, il designates an object of type initializer_list<value_type>, t is a value of type X::value_type, k is a value of type key_type, hf is a possibly const value of type hasher, eq is a possibly const value of type key_equal, n is a value of type size_type, and z is a value of type float.

23.2.5 [unord.req] Unordered associative containers Table 103

Add row:

a.erase(r) iterator Erases the element pointed to by r. Returns the iterator immediately following r prior to the erasure. Average case O(1), worst case O(a.size()).

23.4.4.1 [map.overview] Class template map overview p. 2

iterator erase(iterator position);
iterator erase(const_iterator position);
size_type erase(const key_type& x);
iterator erase(const_iterator first, const_iterator last);

23.4.5.1 [multimap.overview] Class template multimap overview p. 2

iterator erase(iterator position);
iterator erase(const_iterator position);
size_type erase(const key_type& x);
iterator erase(const_iterator first, const_iterator last);

23.4.6.1 [set.overview] Class template set overview p. 2

iterator erase(iterator position);
iterator erase(const_iterator position);
size_type erase(const key_type& x);
iterator erase(const_iterator first, const_iterator last);

23.4.7.1 [multiset.overview] Class template multiset overview

iterator erase(iterator position);
iterator erase(const_iterator position);
size_type erase(const key_type& x);
iterator erase(const_iterator first, const_iterator last);

23.5.4.1 [unord.map.overview] Class template unordered_map overview p. 3

iterator erase(iterator position);
iterator erase(const_iterator position);
size_type erase(const key_type& x);
iterator erase(const_iterator first, const_iterator last);

23.5.5.1 [unord.multimap.overview] Class template unordered_multimap overview p. 3

iterator erase(iterator position);
iterator erase(const_iterator position);
size_type erase(const key_type& x);
iterator erase(const_iterator first, const_iterator last);

23.5.6.1 [unord.set.overview] Class template unordered_set overview p. 3

iterator erase(iterator position);
iterator erase(const_iterator position);
size_type erase(const key_type& x);
iterator erase(const_iterator first, const_iterator last);

23.5.7.1 [unord.multiset.overview] Class template unordered_multiset overview p. 3

iterator erase(iterator position);
iterator erase(const_iterator position);
size_type erase(const key_type& x);
iterator erase(const_iterator first, const_iterator last);

[diff.cpp03.containers] C.2.12 Clause 23: containers library

23.2.3, 23.2.4

Change: Signature changes: from iterator to const_iterator parameters

Rationale: Overspecification. Effects: The signatures of the following member functions changed from taking an iterator to taking a const_iterator:

Valid C++ 2003 code that uses these functions may fail to compile with this International Standard.


2062. Effect contradictions w/o no-throw guarantee of std::function swaps

Section: 20.8.11.2 [func.wrap.func], 20.8.11.2.2 [func.wrap.func.mod] Status: Open Submitter: Daniel Krügler Opened: 2011-05-28 Last modified: 2012-09-24

View all other issues in [func.wrap.func].

View all issues with Open status.

Discussion:

Howard Hinnant observed in reflector message c++std-lib-30841 that 20.8.11.2 [func.wrap.func] makes the member swap noexcept, even though the non-member swap is not noexcept.

The latter was an outcome of the discussions during the Batavia meeting and the Madrid meeting involving LWG 1349, which seems to indicate that the remaining noexcept specifier at the member swap is incorrect and should be removed.

But if we allow for a potentially throwing member swap of std::function, this causes another conflict with the exception specification for the following member function:

template<class F> function& operator=(reference_wrapper<F> f) noexcept;

Effects: function(f).swap(*this);

Note that in this example the sub-expression function(f) does not cause any problems, because of the nothrow-guarantee given in 20.8.11.2.1 [func.wrap.func.con] p. 10. The problem is located in the usage of the swap which could potentially throw given the general latitude.

So, either the Madrid meeting decision need to be revised (and both member and free swap of std::function should be noexcept), or this function needs to be adapted as well, e.g. by taking the exception-specification away or by changing the semantics.

One argument for "swap-may-throw" would be to allow for small-object optimization techniques where the copy of the target may throw. But given the fact that the swap function has been guaranteed to be "Throws: Nothing" from TR1 on, it seems to me that that there would still be opportunities to perform small-object optimizations just restricted to the set of target copies that cannot throw.

In my opinion member swap of std::function has always been intended to be no-throw, because otherwise there would be no good technical reason to specify the effects of several member functions in terms of the "construct-swap" idiom (There are three functions that are defined this way), which provides the strong exception safety in this case. I suggest to enforce that both member swap and non-member swap of std::function are nothrow functions as it had been guaranteed since TR1 on.

[ 2011 Bloomington ]

Dietmar: May not be swappable in the first place.

Alisdair: This is wide contact. Then we should be taking noexcept off instead of putting it on. This is preferred resolution.

Pablo: This is bigger issue. Specification of assignment in terms of swap is suspect to begin with. It is over specification. How this was applied to string is a better example to work from.

Pablo: Two problems: inconsistency that should be fixed (neither should have noexcept), the other issues is that assignment should not be specified in terms of swap. There are cases where assignment should succeed where swap would fail. This is easier with string as it should follow container rules.

Action Item (Alisdair): There are a few more issues found to file.

Dave: This is because of allocators? The allocator makes this not work.

Howard: There is a type erased allocator in shared_ptr. There is a noexcept allocator in shared_ptr.

Pablo: shared_ptr is a different case. There are shared semantics and the allocator does move around. A function does not have shared semantics.

Alisdair: Function objects think they have unique ownership.

Howard: In function we specify semantics with copy construction and swap.

Action Item (Pablo): Write this up better (why assignment should not be defined in terms of swap)

Howard: Not having trouble making function constructor no throw.

Dietmar: Function must allocate memory.

Howard: Does not put stuff that will throw on copy or swap in small object optimization. Put those on heap. Storing allocator, but has to be no throw copy constructable.

Pablo: Are you allowed to or required to swap or move allocators in case or swap or move.

Dave: An allocator that is type erased should be different...

Pablo: it is

Dave: Do you need to know something about allocator types? But only at construction time.

Pablo: You could have allocators that are different types.

Dave: Swap is two ended operation.

Pablo: Opinion is that both have to say propagate on swap for them to swap.

John: It is not arbitrary. If one person says no. No is no.

Howard: Find noexcept swap to be very useful. Would like to move in that direction and bring container design along.

Dave: If you have something were allocator must not propagate you can detect that at construction time.

...

Pablo: Need to leave this open and discuss in smaller group.

Alisdair: Tried to add boost::any as TR2 proposal and ran into this issue. Only the first place where we run into issues with type erased allocators. Suggest we move it to open.

Action Item: Move to open.

Action Item (Pablo works with Howard and Daniel): Address the more fundamental issue (which may be multiple issues) and write up findings.

[ Original resolution: ]

This wording is relative to the FDIS.

  1. Modify the header <functional> synopsis in 20.8 [function.objects] as indicated:

    namespace std {
      […]
    
      template<class R, class... ArgTypes>
      void swap(function<R(ArgTypes...)>&, function<R(ArgTypes...)>&) noexcept;
    
      […]
    }
    
  2. Modify the class template function synopsis in 20.8.11.2 [func.wrap.func] as indicated:

    namespace std {
      […]
    
      // [func.wrap.func.alg], specialized algorithms:
      template<class R, class... ArgTypes>
      void swap(function<R(ArgTypes...)>&, function<R(ArgTypes...)>&) noexcept;
    
      […]
    }
    
  3. Modify 20.8.11.2.7 [func.wrap.func.alg] as indicated:

    template<class R, class... ArgTypes>
    void swap(function<R(ArgTypes...)>& f1, function<R(ArgTypes...)>& f2) noexcept;
    

    -1- Effects: f1.swap(f2);

Proposed resolution:


2063. Contradictory requirements for string move assignment

Section: 21.4 [basic.string] Status: Open Submitter: Howard Hinnant Opened: 2011-05-29 Last modified: 2012-09-24

View all other issues in [basic.string].

View all issues with Open status.

Discussion:

21.4.1 [string.require]/p4 says that basic_string is an "allocator-aware" container and behaves as described in 23.2.1 [container.requirements.general].

23.2.1 [container.requirements.general] describes move assignment in p7 and Table 99.

If allocator_traits<allocator_type>::propagate_on_container_move_assignment::value is false, and if the allocators stored in the lhs and rhs sides are not equal, then move assigning a string has the same semantics as copy assigning a string as far as resources are concerned (resources can not be transferred). And in this event, the lhs may have to acquire resources to gain sufficient capacity to store a copy of the rhs.

However 21.4.2 [string.cons]/p22 says:

basic_string<charT,traits,Allocator>&
operator=(basic_string<charT,traits,Allocator>&& str) noexcept;

Effects: If *this and str are not the same object, modifies *this as shown in Table 71. [Note: A valid implementation is swap(str). — end note ]

These two specifications for basic_string::operator=(basic_string&&) are in conflict with each other. It is not possible to implement a basic_string which satisfies both requirements.

Additionally assign from an rvalue basic_string is defined as:

basic_string& assign(basic_string&& str) noexcept;

Effects: The function replaces the string controlled by *this with a string of length str.size() whose elements are a copy of the string controlled by str. [ Note: A valid implementation is swap(str). — end note ]

It seems contradictory that this member can be sensitive to propagate_on_container_swap instead of propagate_on_container_move_assignment. Indeed, there is a very subtle chance for undefined behavior here: If the implementation implements this in terms of swap, and if propagate_on_container_swap is false, and if the two allocators are unequal, the behavior is undefined, and will likely lead to memory corruption. That's a lot to go wrong under a member named "assign".

[ 2011 Bloomington ]

Alisdair: Can this be conditional noexcept?

Pablo: We said we were not going to put in many conditional noexcepts. Problem is not allocator, but non-normative definition. It says swap is a valid operation which it is not.

Dave: Move assignment is not a critical method.

Alisdair: Was confusing assignment and construction.

Dave: Move construction is critical for efficiency.

Kyle: Is it possible to test for noexcept.

Alisdair: Yes, query the noexcept operator.

Alisdair: Agreed there is a problem that we cannot unconditionally mark these operations as noexcept.

Pablo: How come swap is not defined in alloc

Alisdair: It is in utility.

Pablo: Swap has a conditional noexcept. Is no throw move constructable, is no throw move assignable.

Pablo: Not critical for strings or containers.

Kyle: Why?

Pablo: They do not use the default swap.

Dave: Important for deduction in other types.

Alisdair: Would change the policy we adopted during FDIS mode.

Pablo: Keep it simple and get some vendor experience.

Alisdair: Is this wording correct? Concerned with bullet 2.

Pablo: Where does it reference containers section.

Alisdair: String is a container.

Alisdair: We should not remove redundancy piecemeal.

Pablo: I agree. This is a deviation from rest of string. Missing forward reference to containers section.

Pablo: To fix section 2. Only the note needs to be removed. The rest needs to be a forward reference to containers.

Alisdair: That is a new issue.

Pablo: Not really. Talking about adding one sentence, saying that basic string is a container.

Dave: That is not just a forward reference, it is a semantic change.

PJ: We intended to make it look like a container, but it did not satisfy all the requirements.

Pablo: Clause 1 is correct. Clause 2 is removing note and noexcept (do not remove the rest). Clause 3 is correct.

Alisdair: Not sure data() is correct (in clause 2).

Conclusion: Move to open, Alisdair and Pablo volunteered to provide wording

[ originally proposed wording: ]

This wording is relative to the FDIS.

  1. Modify the class template basic_string synopsis in 21.4 [basic.string]:

    namespace std {
      template<class charT, class traits = char_traits<charT>,
        class Allocator = allocator<charT> >
      class basic_string {
      public:
        […]
        basic_string& operator=(basic_string&& str) noexcept;
        […]
        basic_string& assign(basic_string&& str) noexcept;
        […]
      };
    }
    
  2. Remove the definition of the basic_string move assignment operator from 21.4.2 [string.cons] entirely, including Table 71 — operator=(const basic_string<charT, traits, Allocator>&&). This is consistent with how we define move assignment for the containers in Clause 23:

    basic_string<charT,traits,Allocator>&
    operator=(basic_string<charT,traits,Allocator>&& str) noexcept;
    

    -22- Effects: If *this and str are not the same object, modifies *this as shown in Table 71. [ Note: A valid implementation is swap(str). — end note ]

    -23- If *this and str are the same object, the member has no effect.

    -24- Returns: *this

    Table 71 — operator=(const basic_string<charT, traits, Allocator>&&)
    Element Value
    data() points at the array whose first element was pointed at by str.data()
    size() previous value of str.size()
    capacity() a value at least as large as size()
  3. Modify the paragraphs prior to 21.4.6.3 [string::assign] p.3 as indicated (The first insertion recommends a separate paragraph number for the indicated paragraph):

    basic_string& assign(basic_string&& str) noexcept;
    

    -?- Effects: Equivalent to *this = std::move(str). The function replaces the string controlled by *this with a string of length str.size() whose elements are a copy of the string controlled by str. [ Note: A valid implementation is swap(str). — end note ]

    -3- Returns: *this

[ 2012-08-11 Joe Gottman observes: ]

One of the effects of basic_string's move-assignment operator (21.4.2 [string.cons], Table 71) is

Element Value
data() points at the array whose first element was pointed at by str.data()

If a string implementation uses the small-string optimization and the input string str is small enough to make use of it, this effect is impossible to achieve. To use the small string optimization, a string has to be implemented using something like

union
{
   char buffer[SMALL_STRING_SIZE];
   char *pdata;
};

When the string is small enough to fit inside buffer, the data() member function returns static_cast<const char *>(buffer), and since buffer is an array variable, there is no way to implement move so that the moved-to string's buffer member variable is equal to this->buffer.

Resolution proposal:

Change Table 71 to read:

Element Value
data() points at the array whose first element was pointed at by str.data() that contains the same characters in the same order as str.data() contained before operator=() was called

Proposed resolution:


2070. allocate_shared should use allocator_traits<A>::construct

Section: 20.7.2.2.6 [util.smartptr.shared.create] Status: Open Submitter: Jonathan Wakely Opened: 2011-07-11 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

20.7.2.2.6 [util.smartptr.shared.create] says:

-2- Effects: Allocates memory suitable for an object of type T and constructs an object in that memory via the placement new expression ::new (pv) T(std::forward<Args>(args)...). The template allocate_shared uses a copy of a to allocate memory. If an exception is thrown, the functions have no effect.

This explicitly requires placement new rather than using allocator_traits<A>::construct(a, (T*)pv, std::forward<Args>(args)...) In most cases that would result in the same placement new expression, but would allow more control over how the object is constructed e.g. using scoped_allocator_adaptor to do uses-allocator construction, or using an allocator declared as a friend to construct objects with no public constructors.

[ 2011-08-16 Bloomington: ]

Agreed to fix in principle, but believe that make_shared and allocate_shared have now diverged enough that their descriptions should be separated. Pablo and Stefanus to provide revised wording.

Daniel's (old) proposed resolution:

This wording is relative to the FDIS.

  1. Change the following paragraphs of 20.7.2.2.6 [util.smartptr.shared.create] as indicated (The suggested removal of the last sentence of p1 is not strictly required to resolve this issue, but is still recommended, because it does not say anything new but may give the impression that it says something new):

    template<class T, class... Args> shared_ptr<T> make_shared(Args&&... args);
    template<class T, class A, class... Args>
      shared_ptr<T> allocate_shared(const A& a, Args&&... args);
    

    -1- Requires: For the template make_shared, tThe expression ::new (pv) T(std::forward<Args>(args)...), where pv has type void* and points to storage suitable to hold an object of type T, shall be well formed. For the template allocate_shared, the expression allocator_traits<A>::construct(a, pt, std::forward<Args>(args)...), where pt has type T* and points to storage suitable to hold an object of type T, shall be well formed. A shall be an allocator ([allocator.requirements]). The copy constructor and destructor of A shall not throw exceptions.

    -2- Effects: Allocates memory suitable for an object of type T and constructs an object in that memory. The template make_shared constructs the object via the placement new expression ::new (pv) T(std::forward<Args>(args)...). The template allocate_shared uses a copy of a to allocate memory and constructs the object by calling allocator_traits<A>::construct(a, pt, std::forward<Args>(args)...). If an exception is thrown, the functions have no effect.

    -3- Returns: A shared_ptr instance that stores and owns the address of the newly constructed object of type T.

    -4- Postconditions: get() != 0 && use_count() == 1

    -5- Throws: bad_alloc, or, for the template make_shared, an exception thrown from the constructor of T, or, for the template allocate_shared, an exception thrown from A::allocate or from allocator_traits<A>::constructfrom the constructor of T.

    -6- Remarks: Implementations are encouraged, but not required, to perform no more than one memory allocation. [ Note: This provides efficiency equivalent to an intrusive smart pointer. — end note ]

    -7- [ Note: These functions will typically allocate more memory than sizeof(T) to allow for internal bookkeeping structures such as the reference counts. — end note ]

[2011-12-04: Jonathan and Daniel improve wording]

See also c++std-lib-31796

Proposed resolution:

This wording is relative to the FDIS.

  1. Change the following paragraphs of 20.7.2.2.6 [util.smartptr.shared.create] as indicated:

    template<class T, class... Args> shared_ptr<T> make_shared(Args&&... args);
    template<class T, class A, class... Args>
      shared_ptr<T> allocate_shared(const A& a, Args&&... args);
    

    -1- Requires: The expression ::new (pv) T(std::forward<Args>(args)...), where pv has type void* and points to storage suitable to hold an object of type T, shall be well formed. A shall be an allocator (17.6.3.5 [allocator.requirements]). The copy constructor and destructor of A shall not throw exceptions.

    -2- Effects: Equivalent to

     
    return allocate_shared<T>(allocator<T>(), std::forward<Args>(args)...);
    

    Allocates memory suitable for an object of type T and constructs an object in that memory via the placement new expression ::new (pv) T(std::forward<Args>(args)...). The template allocate_shared uses a copy of a to allocate memory. If an exception is thrown, the functions have no effect.

    -?- Remarks: An implementation may meet the effects (and the implied guarantees) without creating the allocator object [Note: That is, user-provided specializations of std::allocator may not be instantiated, the expressions ::new (pv) T(std::forward<Args>(args)...) and pv->~T() may be evaluated directly — end note].

    -3- Returns: A shared_ptr instance that stores and owns the address of the newly constructed object of type T.

    -4- Postconditions: get() != 0 && use_count() == 1

    -5- Throws: bad_alloc, or an exception thrown from A::allocate or from the constructor of T.

    -6- Remarks: Implementations are encouraged, but not required, to perform no more than one memory allocation. [Note: This provides efficiency equivalent to an intrusive smart pointer. — end note]

    -7- [Note: These functions will typically allocate more memory than sizeof(T) to allow for internal bookkeeping structures such as the reference counts. — end note]

  2. Add the following set of new paragraphs immediately following the previous paragraph 7 of 20.7.2.2.6 [util.smartptr.shared.create]:

    template<class T, class A, class... Args>
      shared_ptr<T> allocate_shared(const A& a, Args&&... args);
    

    -?- Requires: The expressions allocator_traits<A>::construct(b, pt, std::forward<Args>(args)...) and allocator_traits<A>::destroy(b, pt) shall be well-formed and well-defined, where b has type A and is a copy of a and where pt has type T* and points to storage suitable to hold an object of type T. A shall meet the allocator requirements (17.6.3.5 [allocator.requirements]).

    -?- Effects: Uses an object a2 of type allocator_traits<A>::rebind_alloc<unspecified> that compares equal to a to allocate memory suitable for an object of type T. Uses a copy b of type A from a to construct an object of type T in that memory by calling allocator_traits<A>::construct(b, pt, std::forward<Args>(args)...). If an exception is thrown, the function has no effect.

    -?- Returns: A shared_ptr instance that stores and owns the address of the newly constructed object of type T. When ownership is given up, the effects are as follows: Uses a copy b2 of type A from a to destruct an object of type T by calling allocator_traits<A>::destroy(b2, pt2) where pt2 has type T* and refers to the newly constructed object. Then uses an object of type allocator_traits<A>::rebind_alloc<unspecified> that compares equal to a to deallocate the allocated memory.

    -?- Postconditions: get() != 0 && use_count() == 1

    -?- Throws: Nothing unless memory allocation or allocator_traits<A>::construct throws an exception.

    -?- Remarks: Implementations are encouraged, but not required, to perform no more than one memory allocation. [Note: Such an implementation provides efficiency equivalent to an intrusive smart pointer. — end note]

    -?- [Note: This function will typically allocate more memory than sizeof(T) to allow for internal bookkeeping structures such as the reference counts. — end note]


2071. std::valarray move-assignment

Section: 26.6.2.3 [valarray.assign] Status: Ready Submitter: Paolo Carlini Opened: 2011-05-05 Last modified: 2012-09-24

View all other issues in [valarray.assign].

View all issues with Ready status.

Discussion:

Yesterday I noticed that the language we have in the FDIS about std::valarray move assignment is inconsistent with the resolution of LWG 675. Indeed, we guarantee constant complexity (vs linear complexity). We also want it to be noexcept, that is more subtle, but again it's at variance with all the containers.

Also, even if we suppose that LWG 675 applies only to the containers proper, I don't think the current "as if by calling resize(v.size())" is internally consistent with the noexcept requirement.

So, what do we really want for std::valarray? Shall we maybe just strike or fix the as-if, consider it some sort of pasto from the copy-assignment text, thus keep the noexcept and constant complexity requirements (essentially the whole operation would boild down to a swap of POD data members). Or LWG 675 should be explicitly extended to std::valarray too? In that case both noexcept and constant complexity would go, I think, and the operation would boil down to the moral equivalent of clear() (which doesn't really exist in this case) + swap?

Howard: I agree the current wording is incorrect. The complexity should be linear in size() (not v.size()) because the first thing this operator needs to do is resize(0) (or clear() as you put it).

I think we can keep the noexcept.

As for proper wording, here's a first suggestion:

Effects: *this obtains the value of v. The value of v after the assignment is not specified.

Complexity: linear.

See also reflector discussion starting with c++std-lib-30690.

[2012, Kona]

Move to Ready.

Some discussion on the types supported by valarray concludes that the wording is trying to say something similar to the core wording for trivial types, but significantly predates it, and does allow for types with non-trivial destructors. Howard notes that the only reason for linear complexity, rather than constant, is to support types with non-trivial destructors.

AJM suggests replacing the word 'value' with 'state', but straw poll prefers moving forward with the current wording, 5 to 2.

Proposed resolution:

This wording is relative to the FDIS.

In 26.6.2.3 [valarray.assign] update as follows:

valarray<T>& operator=(valarray<T>&& v) noexcept;

3 Effects: *this obtains the value of v. If the length of v is not equal to the length of *this, resizes *this to make the two arrays the same length, as if by calling resize(v.size()), before performing the assignment.The value of v after the assignment is not specified.

4 Complexity: ConstantLinear.


2072. Unclear wording about capacity of temporary buffers

Section: 20.6.11 [temporary.buffer] Status: Open Submitter: Kazutoshi Satoda Opened: 2011-08-10 Last modified: 2012-09-24

View all other issues in [temporary.buffer].

View all issues with Open status.

Discussion:

According to 20.6.11 [temporary.buffer] p1+2:

template <class T>
pair<T*, ptrdiff_t> get_temporary_buffer(ptrdiff_t n) noexcept;

-1- Effects: Obtains a pointer to storage sufficient to store up to n adjacent T objects. It is implementation-defined whether over-aligned types are supported (3.11).

-2- Returns: A pair containing the buffer's address and capacity (in the units of sizeof(T)), or a pair of 0 values if no storage can be obtained or if n <= 0.

I read this as prohibiting to return a buffer of which capacity is less than n, because such a buffer is not sufficient to store n objects.

The corresponding description in SGI STL is clear on this point, but I think it is a bit too verbose:

(for the return value, a pair P) [...] the buffer pointed to by P.first is large enough to hold P.second objects of type T. P.second is greater than or equal to 0, and less than or equal to len.

There seems to be two different targets of the "up to n" modification: The capacity of obtained buffer, and the actual number that the caller will store into the buffer.

First I read as the latter, and got surprised seeing that libstdc++ implementation can return a smaller buffer. I started searching about get_temporary_buffer(). After reading a quote from TC++PL at stackoverflow, I realized that the former is intended.

Such misinterpretation seems common:

Proposed resolution:


2074. Off by one error in std::reverse_copy

Section: 25.3.10 [alg.reverse] Status: Ready Submitter: Peter Miller Opened: 2011-08-17 Last modified: 2012-09-24

View all other issues in [alg.reverse].

View all issues with Ready status.

Discussion:

The output of the program below should be:

"three two one null \n"

But when std::reverse_copy is implemented as described in N3291 25.3.10 [alg.reverse] it's:

"null three two one \n"

because there's an off by one error in 25.3.10 [alg.reverse]/4; the definition should read:

*(result + (last - first) - 1 - i) = *(first + i)

Test program:

#include <algorithm>
#include <iostream>

template <typename BiIterator, typename OutIterator>
auto
reverse_copy_as_described_in_N3291(
  BiIterator first, BiIterator last, OutIterator result )
-> OutIterator
{
  // 25.3.10/4 [alg.reverse]:
  // "...such that for any non-negative integer i < (last - first)..."
  for ( unsigned i = 0; i < ( last - first ); ++i )
    // "...the following assignment takes place:"
    *(result + (last - first) - i) = *(first + i);

  // 25.3.10/6
  return result + (last - first);
}

int main()
{
  using std::begin;
  using std::end;
  using std::cout;

  static const char*const in[3]  { "one", "two", "three" };
  const char*             out[4] { "null", "null", "null", "null" };

  reverse_copy_as_described_in_N3291( begin( in ), end( in ), out );

  for ( auto s : out )
    cout << s << ' ';

  cout < std::endl;

  return 0;
}

[2012, Kona]

Move to Ready.

Proposed resolution:

This wording is relative to the FDIS.

Change 25.3.10 [alg.reverse] p4 as follows:

template<class BidirectionalIterator, class OutputIterator>
  OutputIterator
    reverse_copy(BidirectionalIterator first,
                 BidirectionalIterator last, OutputIterator result);

-4- Effects: Copies the range [first,last) to the range [result,result+(last-first)) such that for any non-negative integer i < (last - first) the following assignment takes place: *(result + (last - first) - 1 - i) = *(first + i).

-5- Requires: The ranges [first,last) and [result,result+(last-first)) shall not overlap.

-6- Returns: result + (last - first).

-7- Complexity: Exactly last - first assignments.


2075. Progress guarantees, lock-free property, and scheduling assumptions

Section: 1.10 [intro.multithread], 29.4 [atomics.lockfree], 29.6.5 [atomics.types.operations.req] Status: New Submitter: Torvald Riegel Opened: 2011-08-18 Last modified: 2012-09-24

View all issues with New status.

Discussion:

According to 1.10 [intro.multithread] p2:

"Implementations should ensure that all unblocked threads eventually make progress."

29.4 [atomics.lockfree] p2 declares the lock-free property for a particular object. However, "lock-free" is never defined, and in discussions that I had with committee members it seemed as if the standard's lock-free would be different from what lock-free means in other communities (eg, research, text books on concurrent programming, etc.).

Following 29.6.5 [atomics.types.operations.req] p7 is_lock_free() returns "true if the object is lock-free". What is returned if the object is only sometimes lock-free?

Basically, I would like to see clarifications for the progress guarantees so that users know what they can expect from implementations (and what they cannot expect!), and to give implementors a clearer understanding of which user expectations they have to implement.

  1. Elaborate on the intentions of the progress guarantee in 1.10 [intro.multithread] p2. As I don't know about your intentions, it's hard to suggest a resolution.

  2. Define the lock-free property. The definition should probably include the following points:

  3. Add a note explaining that compare-exchange-weak is not necessarily lock-free (but is nonblocking)? Or is it indeed intended to be lock-free (only allowed to fail spuriously but guaranteed to not fail eventually)? Implementing the latter might be a challenge on LL-SC machines or lead to space overheads I suppose, see the cacheline sharing example above.

[2011-12-01: Hans comments]

1.10 [intro.multithread] p2 was an intentional compromise, and it was understood at the time that it was not a precise statement. The wording was introduced by N3209, which discusses some of the issues. There were additional reflector discussions.

This is somewhat separable from the question of what lock-free means, which is probably a more promising question to focus on.

[2012, Kona]

General direction: lock-free means obstruction-free. Leave the current "should" recommendation for progress. It would take a lot of effort to try to do better.

Proposed resolution:


2076. Bad CopyConstructible requirement in set constructors

Section: 23.4.6.2 [set.cons] Status: New Submitter: Jens Maurer Opened: 2011-08-20 Last modified: 2012-09-24

View all issues with New status.

Discussion:

23.4.6.2 [set.cons] paragraph 4 says:

Requires: If the iterator’s dereference operator returns an lvalue or a non-const rvalue, then Key shall be CopyConstructible.

I'm confused why a "non-const rvalue" for the return value of the iterator would require CopyConstructible; isn't that exactly the situation when you'd want to apply the move constructor?

The corresponding requirement for multimap seems better in that regard ([multimap.cons] paragraph 3):

Requires: If the iterator’s dereference operator returns an lvalue or a const rvalue pair<key_type, mapped_type>, then both key_type and mapped_type shall be CopyConstructible.

Obviously, if I have a const rvalue, I can't apply the move constructor (which will likely attempt modify its argument).

Dave Abrahams:

I think you are right. Proposed resolution: drop "non-" from 23.4.6.2 [set.cons] paragraph 3.

[2012, Kona]

The wording is in this area will be affected by Pablo's paper being adopted at this meeting. Wait for that paper to be applied before visiting this issue - deliberately leave in New status until the next meeting.

Proposed resolution:

This wording is relative to the FDIS.

Change 23.4.6.2 [set.cons] p3 as follows:

template <class InputIterator>
  set(InputIterator first, InputIterator last,
    const Compare& comp = Compare(), const Allocator& = Allocator());

-3- Effects: Constructs an empty set using the specified comparison object and allocator, and inserts elements from the range [first,last).

-4- Requires: If the iterator’s dereference operator returns an lvalue or a non-const rvalue, then Key shall be CopyConstructible.

-5- Complexity: Linear in N if the range [first,last) is already sorted using comp and otherwise N logN, where N is last - first.


2077. Further incomplete constraints for type traits

Section: 20.9.4.3 [meta.unary.prop] Status: Open Submitter: Daniel Krügler Opened: 2011-08-20 Last modified: 2012-09-24

View other active issues in [meta.unary.prop].

View all other issues in [meta.unary.prop].

View all issues with Open status.

Discussion:

The currently agreed on proposed wording for 2015 using remove_all_extents<T>::type instead of the "an array of unknown bound" terminology in the precondition should be extended to some further entries especially in Table 49, notably the is_*constructible, is_*assignable, and is_*destructible entries. To prevent ODR violations, incomplete element types of arrays must be excluded for value-initialization and destruction for example. Construction and assignment has to be honored, when we have array-to-pointer conversions or pointer conversions of incomplete pointees in effect.

[2012, Kona]

The issue is that in three type traits, we are accidentally saying that in certain circumstances the type must give a specified answer when given an incomplete type. (Specifically: an array of unknown bound of incomplete type.) The issue asserts that there's an ODR violation, since the trait returns false in that case but might return a different version when the trait is completed.

Howard argues: no, there is no risk of an ODR violation. is_constructible<A[]> must return false regardless of whether A is complete, so there's no reason to forbid an array of unknown bound of incomplete types. Same argument applies to is_assignable. General agreement with Howard's reasoning.

There may be a real issue for is_destructible. None of us are sure what is_destructible is supposed to mean for an array of unknown bound (regardless of whether its type is complete), and the standard doesn't make it clear. The middle column doesn't say what it's supposed to do for incomplete types.

In at least one implementation, is_destructible<A[]> does return true if A is complete, which would result in ODR violation unless we forbid it for incomplete types.

Move to open. We believe there is no issue for is_constructible or is_assignable, but that there is a real issue for is_destructible.

Proposed resolution:


2078. Throw specification of async() incomplete

Section: 30.6.8 [futures.async] Status: New Submitter: Nicolai Josuttis Opened: 2011-08-29 Last modified: 2012-09-24

View other active issues in [futures.async].

View all other issues in [futures.async].

View all issues with New status.

Discussion:

The current throw specification of async() does state:

-6- Throws: system_error if policy is launch::async and the implementation is unable to start a new thread.

First it seems not clear whether this only applies if policy equals launch::async of if the async launch mode flag is set (if policy|launch::async!=0)

In the discussion Lawrence Crowl also wrote:

More generally, I think what we want to say is that if the implementation cannot successfully execute on one of the policies allowed, then it must choose another. The principle would apply to implementation-defined policies as well.

Peter Sommerlad:

Should not throw. That was the intent. "is async" meat exactly.

Proposed resolution:


2079. Required pow() overloads

Section: 26.8 [c.math] Status: New Submitter: Steve Clamage Opened: 2011-08-29 Last modified: 2012-09-24

View other active issues in [c.math].

View all other issues in [c.math].

View all issues with New 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 26.8 [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);

26.8 [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).

Proposed resolution:


2080. Specify when once_flag becomes invalid

Section: 30.4.4 [thread.once] Status: Review Submitter: Nicolai Josuttis Opened: 2011-08-30 Last modified: 2012-09-24

View all issues with Review status.

Discussion:

In function call_once 30.4.4.2 [thread.once.callonce] paragraph 4 and 5 specify for call_once():

Throws: system_error when an exception is required (30.2.2 [thread.req.exception]), or any exception thrown by func.

Error conditions:

However, nowhere in 30.4.4 [thread.once] is specified, when a once-flag becomes invalid.

As far as I know this happens if the flag is used for different functions. So we either have to have to insert a sentence/paragraph in

30.4.4.2 Function call_once [thread.once.callonce]

or

30.4.4 Call once [thread.once]

explaining when a once_flag becomes invalidated or we should state as error condition something like:

Anthony Williams:

A once_flag is invalidated if you destroy it (e.g. it is an automatic object, or heap allocated and deleted, etc.)

If the library can detect that this is the case then it will throw this exception. If it cannot detect such a case then it will never be thrown.

Jonathan Wakely:

I have also wondered how that error can happen in C++, where the type system will reject a non-callable type being passed to call_once() and should prevent a once_flag being used after its destructor runs.

If a once_flag is used after its destructor runs then it is indeed undefined behaviour, so implementations are already free to throw any exception (or set fire to a printer) without the standard saying so.

My assumption was that it's an artefact of basing the API on pthreads, which says:

The pthread_once() function may fail if:

[EINVAL] If either once_control or init_routine is invalid.

Pete Becker:

Yes, probably. We had to clean up several UNIXisms that were in the original design.

[2012, Kona]

Remove error conditions, move to Review.

Proposed resolution:

This wording is relative to N3337.

  1. Change 30.4.4.2 [thread.once.callonce] as indicated:

    template<class Callable, class ...Args>
    void call_once(once_flag& flag, Callable&& func, Args&&... args);
    

    […]

    -4- Throws: system_error when an exception is required (30.2.2), or any exception thrown by func.

    -5- Error conditions:

    • invalid_argument — if the once_flag object is no longer valid.

2081. Allocator requirements should include CopyConstructible

Section: 17.6.3.5 [allocator.requirements] Status: Ready Submitter: Jonathan Wakely Opened: 2011-08-30 Last modified: 2012-09-24

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with Ready status.

Discussion:

As discussed in c++std-lib-31054 and c++std-lib-31059, the Allocator requirements implicitly require CopyConstructible because a.select_on_container_copy_construction() and container.get_allocator() both return a copy by value, but the requirement is not stated explicitly anywhere.

In order to clarify that allocators cannot have 'explicit' copy constructors, the requirements should include CopyConstructible.

[2012, Kona]

Move to Ready.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change Table 28 — Allocator requirements in 17.6.3.5 [allocator.requirements]:

    Table 28 — Allocator requirements
    Expression Return type Assertion/note pre-/post-condition Default
    X a1(a);
    X a1 = a;
    Shall not exit via an exception.
    post: a1 == a
    X a1(move(a));
    X a1 = move(a);
    Shall not exit via an exception.
    post: a1 equals the prior value
    of a.
  2. Change 17.6.3.5 [allocator.requirements] paragraph 4:

    An allocator type X shall satisfy the requirements of CopyConstructible (17.6.3.1 [utility.arg.requirements]). The X::pointer, X::const_pointer, X::void_pointer, and X::const_void_pointer types shall satisfy the requirements of NullablePointer (17.6.3.3 [nullablepointer.requirements]). No constructor, comparison operator, copy operation, move operation, or swap operation on these types shall exit via an exception. X::pointer and X::const_pointer shall also satisfy the requirements for a random access iterator (24.2 [iterator.requirements]).


2083. const-qualification on weak_ptr::owner_before

Section: 20.7.2.3 [util.smartptr.weak], 20.7.2.3.5 [util.smartptr.weak.obs] Status: Ready Submitter: Ai Azuma Opened: 2011-09-06 Last modified: 2012-09-24

View all other issues in [util.smartptr.weak].

View all issues with Ready status.

Discussion:

Is there any reason why weak_ptr::owner_before member function templates are not const-qualified?

Daniel Krügler:

I don't think so. To the contrary, without these to be const member function templates, the semantics of the specializations owner_less<weak_ptr<T>> and owner_less<shared_ptr<T>> described in 20.7.2.3.7 [util.smartptr.ownerless] is unclear.

It is amusing to note that this miss has remain undetected from the accepted paper n2637 on. For the suggested wording changes see below.

[2012, Kona]

Move to Ready.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change the class template weak_ptr synopsis in 20.7.2.3 [util.smartptr.weak] as indicated:

    namespace std {
      template<class T> class weak_ptr {
      public:
        typedef T element_type;
        […]
        template<class U> bool owner_before(shared_ptr<U> const& b) const;
        template<class U> bool owner_before(weak_ptr<U> const& b) const;
      };
      […]
    }
    
  2. Change the prototypes in 20.7.2.3.5 [util.smartptr.weak.obs] before p6 as indicated:

    template<class U> bool owner_before(shared_ptr<U> const& b) const;
    template<class U> bool owner_before(weak_ptr<U> const& b) const;
    

2085. Wrong description of effect 1 of basic_istream::ignore

Section: 27.7.2.3 [istream.unformatted] Status: New Submitter: Krzysztof Zelechowski Opened: 2011-09-11 Last modified: 2012-09-24

View other active issues in [istream.unformatted].

View all other issues in [istream.unformatted].

View all issues with New status.

Discussion:

27.7.2.3 [istream.unformatted] in N3242 currently has the following to say about the semantics of basic_istream::ignore:

[..]. Characters are extracted until any of the following occurs:

This statement, apart from being slightly ungrammatical, indicates that if (n == numeric_limits<streamsize>::max()), the method returns without extracting any characters.

The description intends to describe the observable behaviour of an implementation in terms of logical assertions. Logical assertions are not "bullets" that can be "entered" but need not; they are predicates that can evaluate to true or false.

The description contains two predicates, either of them causes extraction to terminate. In the incriminated case, the first predicate is evaluates to true because its premise is false, therefore no characters will be extracted.

The intended semantics would be described by the following statement:

[..]. Characters are extracted until any of the following occurs:

Proposed resolution:

This wording is relative to the FDIS.

Change 27.7.2.3 [istream.unformatted] p25 as indicated:

basic_istream<charT,traits>&
  ignore(streamsize n = 1, int_type delim = traits::eof());

-25- Effects: Behaves as an unformatted input function (as described in 27.7.2.3 [istream.unformatted], paragraph 1). After constructing a sentry object, extracts characters and discards them. Characters are extracted until any of the following occurs:


2086. Overly generic type support for math functions

Section: 26.8 [c.math] Status: Ready Submitter: Daniel Krügler Opened: 2011-09-22 Last modified: 2012-09-24

View other active issues in [c.math].

View all other issues in [c.math].

View all issues with Ready status.

Discussion:

26.8 [c.math] ends with a description of a rule set for "sufficient overloads" in p11:

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.

My impression is that this rule set is probably more generic as intended, my assumption is that it is written to mimic the C99/C1x rule set in 7.25 p2+3 in the "C++" way:

-2- Of the <math.h> and <complex.h> functions without an f (float) or l (long double) suffix, several have one or more parameters whose corresponding real type is double. For each such function, except modf, there is a corresponding type-generic macro. (footnote 313) The parameters whose corresponding real type is double in the function synopsis are generic parameters. Use of the macro invokes a function whose corresponding real type and type domain are determined by the arguments for the generic parameters. (footnote 314)

-3- Use of the macro invokes a function whose generic parameters have the corresponding real type determined as follows:

where footnote 314 clarifies the intent:

If the type of the argument is not compatible with the type of the parameter for the selected function, the behavior is undefined.

The combination of the usage of the unspecific term "cast" with otherwise no further constraints (note that C constraints the valid set to types that C++ describes as arithmetic types, but see below for one important difference) has the effect that it requires the following examples to be well-formed and well-defined:

#include <cmath>

enum class Ec { };

struct S { explicit operator long double(); };

void test(Ec e, S s) {
 std::sqrt(e); // OK, behaves like std::sqrt((float) e);
 std::sqrt(s); // OK, behaves like std::sqrt((float) s);
}

GCC 4.7 does not accept any of these examples.

I found another example where the C++ rule differs from the C set, but in this case I'm not so sure, which direction C++ should follow. The difference is located in the fact, that in C enumerated types are integer types as described in 6.2.5 p17 (see e.g. n1569 or n1256):

"The type char, the signed and unsigned integer types, and the enumerated types are collectively called integer types. The integer and real floating types are collectively called real types."

This indicates that in C the following code

#include <math.h>

enum E { e };

void test(void) {
  sqrt(e); // OK, behaves like sqrt((double) e);
}

seems to be well-defined and e is cast to double, but in C++ referring to

#include <cmath>

enum E { e };

void test() {
  std::sqrt(e); // OK, behaves like sqrt((float) e);
}

is also well-defined (because of our lack of constraints) but we must skip bullet 2 (because E is not an integer type) and effectively cast e to float. Accepting this, we would introduce a silent, but observable runtime difference for C and C++.

GCC 4.7 does not accept this example, but causes an ambiguity error among the three floating point overloads of sqrt.

My current suggestion to fix these problems would be to constrain the valid argument types of these functions to arithmetic types.

Howard provided wording to solve the issue.

[2012, Kona]

Moved to Ready. The proposed wording reflects both original intent from TR1, and current implementations.

Proposed resolution:

This wording is relative to the FDIS.

Change 26.8 [c.math] p11 as indicated:

Moreover, there shall be additional overloads sufficient to ensure:

  1. If any arithmetic argument corresponding to a double parameter has type long double, then all arithmetic arguments corresponding to double parameters are effectively cast to long double.
  2. Otherwise, if any arithmetic argument corresponding to a double parameter has type double or an integer type, then all arithmetic arguments corresponding to double parameters are effectively cast to double.
  3. Otherwise, all arithmetic arguments corresponding to double parameters are effectively cast tohave type float.

2087. iostream_category() and noexcept

Section: 27.5 [iostreams.base] Status: New Submitter: Nicolai Josuttis Opened: 2011-09-22 Last modified: 2012-09-24

View all other issues in [iostreams.base].

View all issues with New status.

Discussion:

In <system_error> we have:

const error_category& generic_category() noexcept;
const error_category& system_category() noexcept;

In <future> we have:

const error_category& future_category() noexcept;

But in <ios> we have:

const error_category& iostream_category();

Is there any reason that iostream_category() is not declared with noexcept or is this an oversight?

Daniel:

This looks like an oversight to me. We made the above mentioned changes as part of noexcept-ifying the thread library but iostream_category() was skipped, so it seems to be forgotten. There should be no reason, why it cannot be noexcept. When doing so, we should also make these functions noexcept (similar to corresponding overloads):

error_code make_error_code(io_errc e);
error_condition make_error_condition(io_errc e);

Suggested wording provided by Daniel.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change 27.5.1 [iostreams.base.overview], header <ios> synopsis as indicated:

    #include <iosfwd>
    namespace std {
      […]
      error_code make_error_code(io_errc e) noexcept;
      error_condition make_error_condition(io_errc e) noexcept;
      const error_category& iostream_category() noexcept;
    }
    
  2. Change the prototype declarations in 27.5.6.5 [error.reporting] as indicated:

    error_code make_error_code(io_errc e) noexcept;
    

    -1- Returns: error_code(static_cast<int>(e), iostream_category()).

    error_condition make_error_condition(io_errc e) noexcept;
    

    -2- Returns: error_condition(static_cast<int>(e), iostream_category()).

    const error_category& iostream_category() noexcept;
    

    -3- Returns: A reference to an object of a type derived from class error_category.

    -4- 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 "iostream".


2088. std::terminate problem

Section: 18.8.3 [exception.terminate] Status: Open Submitter: Daniel Krügler Opened: 2011-09-25 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

Andrzej Krzemienski reported the following on comp.std.c++:

In N3290, which is to become the official standard, in 18.8.3.4 [terminate], paragraph 1 reads

Remarks: Called by the implementation when exception handling must be abandoned for any of several reasons (15.5.1), in effect immediately after evaluating the throw-expression (18.8.3.1). May also be called directly by the program.

It is not clear what is "in effect". It was clear in previous drafts where paragraphs 1 and 2 read:

Called by the implementation when exception handling must be abandoned for any of several reasons (15.5.1). May also be called directly by the program.

Effects: Calls the terminate_handler function in effect immediately after evaluating the throw-expression (18.8.3.1), if called by the implementation, or calls the current terminate_handler function, if called by the program.

It was changed by N3189. The same applies to function unexpected (D. 11.4, paragraph 1).

Assuming the previous wording is still intended, the wording can be read "unless std::terminate is called by the program, we will use the handler that was in effect immediately after evaluating the throw-expression".

This assumes that there is some throw-expression connected to every situation that triggers the call to std::terminate. But this is not the case:

Which one is referred to?

In case std::nested_exception::rethrow_nested is called for an object that has captured no exception, there is no throw-expression involved directly (and may no throw be involved even indirectly).

Next, 18.8.3.1 [terminate.handler], paragraph 2 says

Required behavior: A terminate_handler shall terminate execution of the program without returning to the caller.

This seems to allow that the function may exit by throwing an exception (because word "return" implies a normal return).

One could argue that words "terminate execution of the program" are sufficient, but then why "without returning to the caller" would be mentioned. In case such handler throws, noexcept specification in function std::terminate is violated, and std::terminate would be called recursively - should std::abort not be called in case of recursive std::terminate call? On the other hand some controlled recursion could be useful, like in the following technique.

The here mentioned wording changes by N3189 in regard to 18.8.3.4 [terminate] p1 were done for a better separation of effects (Effects element) and additional normative wording explanations (Remarks element), there was no meaning change intended. Further, there was already a defect existing in the previous wording, which was not updated when further situations where defined, when std::terminate where supposed to be called by the implementation.

The part

"in effect immediately after evaluating the throw-expression"

should be removed and the quoted reference to 18.8.3.1 [terminate.handler] need to be part of the effects element where it refers to the current terminate_handler function, so should be moved just after

"Effects: Calls the current terminate_handler function."

It seems ok to allow a termination handler to exit via an exception, but the suggested idiom should better be replaced by a more simpler one based on evaluating the current exception pointer in the terminate handler, e.g.

void our_terminate (void) {
  std::exception_ptr p = std::current_exception();
  if (p) {
    ... // OK to rethrow and to determine it's nature
  } else {
    ... // Do something else
  }
}

[2011-12-09: Daniel comments]

A related issue is 2111.

[2012, Kona]

Move to Open.

There is an interaction with Core issues in this area that Jens is already supplying wording for. Review this issue again once Jens wording is available.

Alisdair to review clause 15.5 (per Jens suggestion) and recommend any changes, then integrate Jens wording into this issue.

Proposed resolution:


2089. std::allocator::construct should use uniform initialization

Section: 20.6.9.1 [allocator.members] Status: Open Submitter: David Krauss Opened: 2011-10-07 Last modified: 2012-09-24

View all other issues in [allocator.members].

View all issues with Open status.

Discussion:

When the EmplaceConstructible (23.2.1 [container.requirements.general]/13) requirement is used to initialize an object, direct-initialization occurs. Initializing an aggregate or using a std::initializer_list constructor with emplace requires naming the initialized type and moving a temporary. This is a result of std::allocator::construct using direct-initialization, not list-initialization (sometimes called "uniform initialization") syntax.

Altering std::allocator<T>::construct to use list-initialization would, among other things, give preference to std::initializer_list constructor overloads, breaking valid code in an unintuitive and unfixable way — there would be no way for emplace_back to access a constructor preempted by std::initializer_list without essentially reimplementing push_back.

std::vector<std::vector<int>> v;
v.emplace_back(3, 4); // v[0] == {4, 4, 4}, not {3, 4} as in list-initialization

The proposed compromise is to use SFINAE with std::is_constructible, which tests whether direct-initialization is well formed. If is_constructible is false, then an alternative std::allocator::construct overload is chosen which uses list-initialization. Since list-initialization always falls back on direct-initialization, the user will see diagnostic messages as if list-initialization (uniform-initialization) were always being used, because the direct-initialization overload cannot fail.

I can see two corner cases that expose gaps in this scheme. One occurs when arguments intended for std::initializer_list satisfy a constructor, such as trying to emplace-insert a value of {3, 4} in the above example. The workaround is to explicitly specify the std::initializer_list type, as in v.emplace_back(std::initializer_list<int>(3, 4)). Since this matches the semantics as if std::initializer_list were deduced, there seems to be no real problem here.

The other case is when arguments intended for aggregate initialization satisfy a constructor. Since aggregates cannot have user-defined constructors, this requires that the first nonstatic data member of the aggregate be implicitly convertible from the aggregate type, and that the initializer list have one element. The workaround is to supply an initializer for the second member. It remains impossible to in-place construct an aggregate with only one nonstatic data member by conversion from a type convertible to the aggregate's own type. This seems like an acceptably small hole.

The change is quite small because EmplaceConstructible is defined in terms of whatever allocator is specified, and there is no need to explicitly mention SFINAE in the normative text.

[2012, Kona]

Move to Open.

There appears to be a real concern with initializing aggregates, that can be performed only using brace-initialization. There is little interest in the rest of the issue, given the existence of 'emplace' methods in C++11.

Move to Open, to find an acceptable solution for intializing aggregates. There is the potential that EWG may have an interest in this area of language consistency as well.

Proposed resolution:

This wording is relative to the FDIS.

Change 20.6.9.1 [allocator.members] p12 as indicated:

template <class U, class... Args>
  void construct(U* p, Args&&... args);

12 Effects: ::new((void *)p) U(std::forward<Args>(args)...) if is_constructible<U, Args...>::value is true, else ::new((void *)p) U{std::forward<Args>(args)...}


2091. Misplaced effect in m.try_lock_for()

Section: 30.4.1.3 [thread.timedmutex.requirements] Status: Review Submitter: Pete Becker Opened: 2011-10-18 Last modified: 2012-09-24

View other active issues in [thread.timedmutex.requirements].

View all other issues in [thread.timedmutex.requirements].

View all issues with Review status.

Discussion:

30.4.1.3 [thread.timedmutex.requirements]/4 says, in part,

"Requires: If the tick period of [the argument] is not exactly convertible … [it] shall be rounded up …"

This doesn't belong in the requires clause. It's an effect. It belongs in paragraph 5. Nitpickingly, this would be a technical change: as written it imposes an obligation on the caller, while moving it imposes an obligation on the callee. Although that's certainly not what was intended.

Peter Dimov comments:

Not to mention that it should round down, not up. :-)

Incidentally, I see that the wrong try_lock requirement that the caller shall not own the mutex has entered the standard, after all. Oh well. Let's hope that the real world continues to ignore it.

[2012, Kona]

Remove the offending sentence from the requirements clause. Do not add it back anywhere else. The implementation already must have license to wake up late, so the rounding is invisible.

Move to Review.

Proposed resolution:

This wording is relative to N3337.

  1. Change 30.4.1.3 [thread.timedmutex.requirements]/4 as indicated:

    -3- The expression m.try_lock_for(rel_time) shall be well-formed and have the following semantics:

    -4- Requires: If the tick period of rel_time is not exactly convertible to the native tick period, the duration shall be rounded up to the nearest native tick period. If m is of type std::timed_mutex, the calling thread does not own the mutex.


2092. Vague Wording for condition_variable_any

Section: 30.5.2 [thread.condition.condvarany] Status: Review Submitter: Pete Becker Opened: 2011-10-20 Last modified: 2012-09-24

View all other issues in [thread.condition.condvarany].

View all issues with Review status.

Discussion:

30.5.2 [thread.condition.condvarany]/4 says, in part, that condition_variable_any() throws an exception "if any native handle type manipulated is not available".

I don't know what that means. Is this intended to say something different from the analogous words for condition_variable() [30.5.1 [thread.condition.condvar]/4], "if some non-memory resource limitation prevents initialization"? If not, it should be worded the same way.

[2012, Kona]

Copy the corresponding wording from the condition_variable constructor in 30.5.1 [thread.condition.condvar] p4.

Move to Review.

Proposed resolution:

This wording is relative to N3337.

  1. Change 30.4.1.3 [thread.timedmutex.requirements]/4 as indicated:

    condition_variable_any();
    

    […]

    -4- Error conditions:

    • resource_unavailable_try_againif any native handle type manipulated is not available if some non-memory resource limitation prevents initialization.
    • operation_not_permitted — if the thread does not have the privilege to perform the operation.

2093. Throws clause of condition_variable::wait with predicate

Section: 30.5.1 [thread.condition.condvar] Status: Review Submitter: Alberto Ganesh Barbati Opened: 2011-10-27 Last modified: 2012-09-24

View other active issues in [thread.condition.condvar].

View all other issues in [thread.condition.condvar].

View all issues with Review status.

Discussion:

the Throws: clause of condition_variable::wait/wait_xxx functions that take a predicate argument is:

Throws: system_error when an exception is required (30.2.2 [thread.req.exception]).

If executing the predicate throws an exception, I would expect such exception to propagate unchanged to the caller, but the throws clause seems to indicate that it gets mutated into a system_error. T hat's because of 17.5.1.4 [structure.specifications]/4:

"If F’s semantics contains a Throws:, Postconditions:, or Complexity: element, then that supersedes any occurrences of that element in the code sequence."

Is my interpretation correct? Does it match the intent?

Daniel comments:

I don't think that this interpretation is entirely correct, the wording does not say that std::system_error or a derived class must be thrown, it simply is underspecified in this regard (The extreme interpretation is that the behaviour would be undefined, but that would be too far reaching I think). We have better wording for this in 30.4.4.2 [thread.once.callonce] p4, where it says:

"Throws: system_error when an exception is required (30.2.2 [thread.req.exception]), or any exception thrown by func."

or in 30.3.2 [thread.thread.this] p6/p9:

"Throws: Nothing if Clock satisfies the TrivialClock requirements (20.11.3 [time.clock.req]) and operations of Duration do not throw exceptions. [ Note: instantiations of time point types and clocks supplied by the implementation as specified in 20.11.7 [time.clock] do not throw exceptions. — end note ]"

So, the here discussed Throws elements should add lines along the lines of

"Any exception thrown by operations of pred."

and similar wording for time-related operations:

"Any exception thrown by operations of Duration",

"Any exception thrown by operations of chrono::duration<Rep, Period>"

[2011-11-28: Ganesh comments and suggests wording]

As for the discussion about the exception thrown by the manipulation of time-related objects, I believe the argument applies to all functions declared in 30 [thread]. Therefore, instead of adding wording to each member, I would simply move those requirements from 30.3.2 [thread.thread.this] p6/p9 to a new paragraph in 30.2.4 [thread.req.timing].

As for 30.5.2 [thread.condition.condvarany], the member functions wait() and wait_until() are described only in terms of the Effects: clause (so strictly speaking, they need no changes), however, wait_for() is described with a full set of clauses including Throws: and Error conditions:. Either we should add those clauses to wait/wait_until with changes similar to the one above, or remove paragraphs 29 to 34 entirely. By the way, even paragraph 26 could be removed IMHO.

[2012, Kona]

We like the idea behind the proposed resolution.

Modify the first addition to read instead: "Functions that specify a timeout, will throw if an operation on a clock, time point, or time duration throws an exception."

In the note near the bottom change "even if the timeout has already expired" to "or if the timeout has already expired". (This is independent, but the original doesn't seem to make sense.)

Move to Review.

Proposed resolution:

This wording is relative to N3337.

  1. Add a new paragraph at the end of 30.2.4 [thread.req.timing]:

    […]

    -6- 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.

    -7- Implementation-provided clocks that are used for these functions shall meet the TrivialClock requirements (20.11.3 [time.clock.req]).

    -?- Functions that specify a timeout, will throw if an operation on a clock, time point, or time duration throws an exception. [ Note: instantiations of clock, time point and duration types supplied by the implementation as specified in 20.11.7 [time.clock] do not throw exceptions. — end note]

  2. Change 30.3.2 [thread.thread.this] as indicated:

    template <class Clock, class Duration>
      void sleep_until(const chrono::time_point<Clock, Duration>& abs_time);;
    

    -4- Effects: Blocks the calling thread for the absolute timeout (30.2.4 [thread.req.timing]) specified by abs_time.

    -5- Synchronization: None.

    -6- Throws: timeout-related exceptions (30.2.4 [thread.req.timing]).Nothing if Clock satisfies the TrivialClock requirements (20.11.3 [time.clock.req]) and operations of Duration do not throw exceptions. [ Note: instantiations of time point types and clocks supplied by the implementation as specified in 20.11.7 [time.clock] do not throw exceptions. — end note]

    template <class Rep, class Period>
      void sleep_for(const chrono::duration<Rep, Period>& rel_time);;
    

    -7- Effects: Blocks the calling thread for the relative timeout (30.2.4 [thread.req.timing]) specified by rel_time.

    -8- Synchronization: None.

    -9- Throws: timeout-related exceptions (30.2.4 [thread.req.timing]).Nothing if operations of chrono::duration<Rep, Period> do not throw exceptions. [ Note: instantiations of time point types and clocks supplied by the implementation as specified in 20.11.7 [time.clock] do not throw exceptions. — end note]

  3. Change 30.4.1.3 [thread.timedmutex.requirements] as indicated:

    -3- The expression m.try_lock_for(rel_time) shall be well-formed and have the following semantics:

    […]

    -5- Effects: The function attempts to obtain ownership of the mutex within the relative timeout (30.2.4 [thread.req.timing]) specified by rel_time. If the time specified by rel_time is less than or equal to rel_time.zero(), the function attempts to obtain ownership without blocking (as if by calling try_lock()). The function shall return within the timeout specified by rel_time only if it has obtained ownership of the mutex object. [Note: As with try_lock(), there is no guarantee that ownership will be obtained if the lock is available, but implementations are expected to make a strong effort to do so. — end note]

    […]

    -8- Synchronization: If try_lock_for() returns true, prior unlock() operations on the same object synchronize with (1.10 [intro.multithread]) this operation.

    -9- Throws: timeout-related exceptions (30.2.4 [thread.req.timing]).Nothing.

    -10- The expression m.try_lock_until(abs_time) shall be well-formed and have the following semantics:

    […]

    -12- Effects: The function attempts to obtain ownership of the mutex. If abs_time has already passed, the function attempts to obtain ownership without blocking (as if by calling try_lock()). The function shall return before the absolute timeout (30.2.4 [thread.req.timing]) specified by abs_time only if it has obtained ownership of the mutex object. [Note: As with try_lock(), there is no guarantee that ownership will be obtained if the lock is available, but implementations are expected to make a strong effort to do so. — end note]

    […]

    -15- Synchronization: If try_lock_until() returns true, prior unlock() operations on the same object synchronize with (1.10 [intro.multithread]) this operation.

    -16- Throws: timeout-related exceptions (30.2.4 [thread.req.timing]).Nothing.

  4. Change 30.5.1 [thread.condition.condvar] as indicated:

    template <class Predicate>
      void wait(unique_lock<mutex>& lock, Predicate pred);
    

    […]

    -15- Effects: Equivalent to:

    while (!pred())
      wait(lock);
    

    […]

    -17- Throws: std::system_error when an exception is required (30.2.2 [thread.req.exception]), timeout-related exceptions (30.2.4 [thread.req.timing]), or any exception thrown by pred.

    […]

    template <class Clock, class Duration>
      cv_status wait_until(unique_lock<mutex>& lock,
                           const chrono::time_point<Clock, Duration>& abs_time);
    

    […]

    -23- Throws: system_error when an exception is required (30.2.2 [thread.req.exception]) or timeout-related exceptions (30.2.4 [thread.req.timing]).

    […]

    template <class Rep, class Period>
      cv_status wait_for(unique_lock<mutex>& lock,
                         const chrono::duration<Rep, Period>& rel_time);
    

    […]

    -26- Effects: as ifEquivalent to:

    return wait_until(lock, chrono::steady_clock::now() + rel_time);
    

    […]

    -29- Throws: system_error when an exception is required (30.2.2 [thread.req.exception]) or timeout-related exceptions (30.2.4 [thread.req.timing]).

    […]

    template <class Clock, class Duration, class Predicate>
      bool wait_until(unique_lock<mutex>& lock,
                      const chrono::time_point<Clock, Duration>& abs_time,
                      Predicate pred);
    

    […]

    -32- Effects: Equivalent to:

    while (!pred())
      if (wait_until(lock, abs_time) == cv_status::timeout)
        return pred();
    return true;
    

    -33- Returns: pred()

    […]

    -36- Throws: std::system_error when an exception is required (30.2.2 [thread.req.exception]), timeout-related exceptions (30.2.4 [thread.req.timing]), or any exception thrown by pred.

    […]

    template <class Rep, class Period, class Predicate>
      bool wait_for(unique_lock<mutex>& lock,
                    const chrono::duration<Rep, Period>& rel_time,
                    Predicate pred);
    

    […]

    -39- Effects: as ifEquivalent to:

    return wait_until(lock, chrono::steady_clock::now() + rel_time, std::move(pred));
    

    […]

    -42- Returns: pred()

    […]

    -44- Throws: system_error when an exception is required (30.2.2 [thread.req.exception]), timeout-related exceptions (30.2.4 [thread.req.timing]), or any exception thrown by pred.

    […]

  5. Change 30.5.2 [thread.condition.condvarany] as indicated:

    template <class Lock, class Predicate>
      void wait(Lock& lock, Predicate pred);
    

    -14- Effects: Equivalent to:

    while (!pred())
      wait(lock);
    
    template <class Lock, class Clock, class Duration>
      cv_status wait_until(Lock& lock, const chrono::time_point<Clock, Duration>& abs_time);
    

    […]

    -18- Throws: system_error when an exception is required (30.2.2 [thread.req.exception]) or any timeout-related exceptions (30.2.4 [thread.req.timing]).

    […]

    template <class Lock, class Rep, class Period>
      cv_status wait_for(Lock& lock, const chrono::duration<Rep, Period>& rel_time);
    

    […]

    -20- Effects: as ifEquivalent to:

    return wait_until(lock, chrono::steady_clock::now() + rel_time);
    

    […]

    -23- Throws: system_error when an exception is required (30.2.2 [thread.req.exception]) or any timeout-related exceptions (30.2.4 [thread.req.timing]).

    […]

    template <class Lock, class Clock, class Duration, class Predicate>
      bool wait_until(Lock& lock, const chrono::time_point<Clock, Duration>& abs_time, Predicate pred);
    

    -25- Effects: Equivalent to:

    while (!pred())
      if (wait_until(lock, abs_time) == cv_status::timeout)
        return pred();
    return true;
    

    -26- Returns: pred()[Note: There is no blocking if pred() is initially true, or if the timeout has already expired. — end note]

    -27- [Note: The returned value indicates whether the predicate evaluates to true regardless of whether the timeout was triggered. end note]

    template <class Lock, class Rep, class Period, class Predicate>
      bool wait_for(Lock& lock, const chrono::duration<Rep, Period>& rel_time, Predicate pred);
    

    -28- Effects: as ifEquivalent to:

    return wait_until(lock, chrono::steady_clock::now() + rel_time, std::move(pred));
    

    -29- [Note: There is no blocking if pred() is initially true, even if the timeout has already expired. — end note] -30- Postcondition: lock is locked by the calling thread.

    -31- Returns: pred()

    -32- [Note: The returned value indicates whether the predicate evaluates to true regardless of whether the timeout was triggered. — end note]

    -33- Throws: system_error when an exception is required (30.2.2 [thread.req.exception]).

    -34- Error conditions:

    • equivalent error condition from lock.lock() or lock.unlock().

2094. duration conversion overflow shouldn't participate in overload resolution

Section: 20.11.5.1 [time.duration.cons] Status: Review Submitter: Vicente J. Botet Escriba Opened: 2011-10-31 Last modified: 2012-09-24

View all other issues in [time.duration.cons].

View all issues with Review status.

Discussion:

20.11.5.1 [time.duration.cons] says:

template <class Rep2, class Period2>
  constexpr duration(const duration<Rep2, Period2>& d);

Remarks: This constructor shall not participate in overload resolution unless treat_as_floating_point<rep>::value is true or both ratio_divide<Period2, period>::den is 1 and treat_as_floating_point<Rep2>::value is false.

The evaluation of ratio_divide<Period2, period>::den could make ratio_divide<Period2, period>::num overflow.

This occur for example when we try to create a millisecond (period=ratio<1,1000>) from an exa-second (Period2=ratio<1018>).

ratio_divide<ratio<1018>, ratio<1,1000>>::num is 1021 which overflows which makes the compiler error.

If the function f is overloaded with milliseconds and seconds

void f(milliseconds);
void f(seconds);

The following fails to compile.

duration<int,exa> r(1);
f(r);

While the conversion to seconds work, the conversion to milliseconds make the program fail at compile time. In my opinion, this program should be well formed and the constructor from duration<int,exa> to milliseconds shouldn't participate in overload resolution as the result can not be represented.

I think the wording of the standard can be improved so no misinterpretations are possible by adding that "no overflow is induced by the conversion".

[2012, Kona]

Move to Review.

Pete: The wording is not right.

Howard: Will implement this to be sure it works.

Jeffrey: If ratio needs a new hook, should it be exposed to users for their own uses?

Pete: No.

Move to Review, Howard to implement in a way that mere mortals can understand.

Proposed resolution:

This wording is relative to the FDIS.

Change the following paragraphs of 20.11.5.1 [time.duration.cons] p4 indicated:

template <class Rep2, class Period2>
  constexpr duration(const duration<Rep2, Period2>& d);

Remarks: This constructor shall not participate in overload resolution unless no overflow is induced in the conversion and treat_as_floating_point<rep>::value is true or both ratio_divide<Period2, period>::den is 1 and treat_as_floating_point<Rep2>::value is false. [ Note: This requirement prevents implicit truncation error when converting between integral-based duration types. Such a construction could easily lead to confusion about the value of the duration. — end note ]


2095. promise and packaged_task missing constructors needed for uses-allocator construction

Section: 30.6.5 [futures.promise], 30.6.9 [futures.task] Status: New Submitter: Jonathan Wakely Opened: 2011-11-01 Last modified: 2012-09-24

View other active issues in [futures.promise].

View all other issues in [futures.promise].

View all issues with New status.

Discussion:

This example is ill-formed according to C++11 because uses_allocator<promise<R>, A>::value is true, but is_constructible<promise<R>, A, promise<R>&&>::value is false. Similarly for packaged_task.

#include <future>
#include <memory>
#include <tuple>

using namespace std;

typedef packaged_task<void()> task;
typedef promise<void> prom;
allocator<task> a;

tuple<task, prom> t1{ allocator_arg, a };
tuple<task, prom> t2{ allocator_arg, a, task{}, prom{} };

Proposed resolution:

This wording is relative to the FDIS.

  1. Add to 30.6.5 [futures.promise], class template promise synopsis, as indicated:

    namespace std {
      template <class R>
      class promise {
      public:
        promise();
        template <class Allocator>
        promise(allocator_arg_t, const Allocator& a);
        template <class Allocator>
        promise(allocator_arg_t, const Allocator& a, promise&& rhs) noexcept;
        promise(promise&& rhs) noexcept;
        promise(const promise& rhs) = delete;
        ~promise();	
        […]
      };
      […]
    }
    
  2. Change 30.6.5 [futures.promise] as indicated:

    promise(promise&& rhs) noexcept;
    template <class Allocator>
    promise(allocator_arg_t, const Allocator& a, promise&& rhs) noexcept;
    

    -5- Effects: constructs a new promise object and transfers ownership of the shared state of rhs (if any) to the newly-constructed object.

    -6- Postcondition: rhs has no shared state.

    -?- [Note: a is not used — end note]

  3. Add to 30.6.9 [futures.task], class template packaged_task synopsis, as indicated:

    namespace std {
      template<class> class packaged_task; // undefined
    
      template<class R, class... ArgTypes>
      class packaged_task<R(ArgTypes...)> {
      public:
        // construction and destruction
        packaged_task() noexcept;
        template <class F>
        explicit packaged_task(F&& f);
        template <class Allocator>
        explicit packaged_task(allocator_arg_t, const Allocator& a) noexcept;
        template <class Allocator>
        explicit packaged_task(allocator_arg_t, const Allocator& a, packaged_task&& rhs) noexcept;
        template<class Allocator>
        explicit packaged_task(allocator_arg_t, const Allocator& a, const packaged_task&) = delete;
        template <class F, class Allocator>
        explicit packaged_task(allocator_arg_t, const Allocator& a, F&& f);
        ~packaged_task();
        […]
      };
      […]
    }
    
  4. Change 30.6.9.1 [futures.task.members] as indicated:

    packaged_task() noexcept;
    template <class Allocator>
      explicit packaged_task(allocator_arg_t, const Allocator& a) noexcept;
    

    -1- Effects: constructs a packaged_task object with no shared state and no stored task.

    -?- [Note: a is not used — end note]

    […]

    packaged_task(packaged_task&& rhs) noexcept;
    template <class Allocator>
      explicit packaged_task(allocator_arg_t, const Allocator& a, packaged_task&& rhs) noexcept;
    

    -5- Effects: constructs a new packaged_task object and transfers ownership of rhs’s shared state to *this, leaving rhs with no shared state. Moves the stored task from rhs to *this.

    -6- Postcondition: rhs has no shared state.

    -?- [Note: a is not used — end note]


2097. packaged_task constructors should be constrained

Section: 30.6.9.1 [futures.task.members] Status: New Submitter: Jonathan Wakely Opened: 2011-11-02 Last modified: 2012-09-24

View other active issues in [futures.task.members].

View all other issues in [futures.task.members].

View all issues with New status.

Discussion:

With the proposed resolution of 2067, this no longer selects the copy constructor:

std::packaged_task<void()> p1;
std::packaged_task<void()> p2(p1);

Instead this constructor is a better match:

template <class F>
 explicit packaged_task(F&& f);

This attempts to package a packaged_task, which internally tries to copy p2, which fails because the copy constructor is deleted. For at least one implementation the resulting error message is much less helpful than the expected "cannot call deleted function" because it happens after instantiating several more templates rather than in the context where the constructor is called.

I believe the solution is to constrain to the template constructors so the template argument F cannot be deduced as (possibly cv) packaged_task& or packaged_task. It could be argued this constraint is already implied because packaged_task is not copyable and the template constructors require that "invoking a copy of f shall behave the same as invoking f".

Daniel points out that the variadic constructor of std::thread described in 30.3.1.2 [thread.thread.constr] has a similar problem and suggests a similar wording change, which has been integrated below.

An alternative is to declare thread(thread&) and packaged_task(packaged_task&) as deleted.

Proposed resolution:

This wording is relative to the FDIS.

  1. Insert a new Remarks element to 30.3.1.2 [thread.thread.constr] around p3 as indicated:

    template <class F, class ...Args> explicit thread(F&& f, Args&&... args);
    

    -3- Requires: F and each Ti in Args shall satisfy the MoveConstructible requirements. INVOKE(DECAY_COPY ( std::forward<F>(f)), DECAY_COPY (std::forward<Args>(args))...) (20.8.2) shall be a valid expression.

    -?- Remarks: This constructor shall not participate in overload resolution if decay<F>::type is the same type as std::thread.

  2. Insert a new Remarks element to 30.6.9.1 [futures.task.members] around p2 as indicated:

    template <class F>
      packaged_task(F&& f);
    template <class F, class Allocator>
      explicit packaged_task(allocator_arg_t, const Allocator& a, F&& f);
    

    -2- Requires: INVOKE(f, t1, t2, ..., tN, R), where t1, t2, ..., tN are values of the corresponding types in ArgTypes..., shall be a valid expression. Invoking a copy of f shall behave the same as invoking f.

    -?- Remarks: These constructors shall not participate in overload resolution if decay<F>::type is the same type as std::packaged_task<R(ArgTypes...)>.


2098. Minor Inconsistency between promise::set_value and promise::set_value_at_thread_exit

Section: 30.6.5 [futures.promise] Status: New Submitter: Pete Becker Opened: 2011-11-14 Last modified: 2012-09-24

View other active issues in [futures.promise].

View all other issues in [futures.promise].

View all issues with New status.

Discussion:

30.6.5 [futures.promise]/16 says that promise::set_value(const R&) throws any exceptions thrown by R's copy constructor, and that promise_set_value(R&&) throws any exceptions thrown by R's move constructor.

30.6.5 [futures.promise]/22 is the Throws: clause for promise::set_value_at_thread_exit. It has no corresponding requirements, only that these functions throw "future_error if an error condition occurs."

Daniel suggests wording to fix this: The approach is a bit more ambitious and also attempts to fix wording glitches of 30.6.5 [futures.promise]/16, because it would be beyond acceptable efforts of implementations to determine whether a constructor call of a user-defined type will indeed call a copy constructor or move constructor (in the first case it might be a template constructor, in the second case it might also be a copy-constructor, if the type has no move constructor).

Proposed resolution:

This wording is relative to the FDIS.

  1. Change 30.6.5 [futures.promise]/16 as indicated:

    void promise::set_value(const R& r);
    void promise::set_value(R&& r);
    void promise<R&>::set_value(R& r);
    void promise<void>::set_value();
    

    […]

    -16- Throws:

    • future_error if its shared state already has a stored value or exception, or
    • for the first version, any exception thrown by the copy constructor ofconstructor selected to copy an object of R, or
    • for the second version, any exception thrown by the move constructor ofconstructor selected to move an object of R.
  2. Change 30.6.5 [futures.promise]/22 as indicated:

    void promise::set_value_at_thread_exit(const R& r);
    void promise::set_value_at_thread_exit(R&& r);
    void promise<R&>::set_value_at_thread_exit(R& r);
    void promise<void>::set_value_at_thread_exit();
    

    […]

    -16- Throws: future_error if an error condition occurs.

    • future_error if its shared state already has a stored value or exception, or
    • for the first version, any exception thrown by the constructor selected to copy an object of R, or
    • for the second version, any exception thrown by the constructor selected to move an object of R.

2099. Unnecessary constraints of va_start() usage

Section: 18.10 [support.runtime] Status: Ready Submitter: Daniel Krügler Opened: 2011-11-12 Last modified: 2012-09-24

View other active issues in [support.runtime].

View all other issues in [support.runtime].

View all issues with Ready status.

Discussion:

In 18.10 [support.runtime] p3 we find (emphasis mine):

The restrictions that ISO C places on the second parameter to the va_start() macro in header <stdarg.h> are different in this International Standard. The parameter parmN is the identifier of the rightmost parameter in the variable parameter list of the function definition (the one just before the ...).227 If the parameter parmN is declared with a function, array, or reference type, or with a type that is not compatible with the type that results when passing an argument for which there is no parameter, the behavior is undefined.

It seems astonishing that the constraints on function types and array types imposes these on the declared parameter parmN, not to the adjusted one (which would not require this extra wording, because that is implicit). This seems to say that a function definition of the form (Thanks to Johannes Schaub for this example)

#include <stdarg.h>

void f(char const paramN[], ...) {
  va_list ap;
  va_start(ap, paramN);
  va_end(ap);
}

would produce undefined behaviour when used.

Similar wording exists in C99 and in the most recent C11 draft in 7.16.1.4 p4

In my opinion the constraints in regard to array types and function types are unnecessary and should be relaxed. Are there really implementations out in the wild that would (according to my understanding incorrectly) provide the declared and not the adjusted type of paramN as deduced type to va_start()?

[2012, Kona]

Move to Ready.

Proposed resolution:

This wording is relative to the FDIS.

Change 18.10 [support.runtime] p3 as indicated:

The restrictions that ISO C places on the second parameter to the va_start() macro in header <stdarg.h> are different in this International Standard. The parameter parmN is the identifier of the rightmost parameter in the variable parameter list of the function definition (the one just before the ...).227 If the parameter parmN is declared withof a function, array, or reference type, or withof a type that is not compatible with the type that results when passing an argument for which there is no parameter, the behavior is undefined.


2100. timed waiting functions cannot timeout if launch::async policy used

Section: 30.6.8 [futures.async] Status: New Submitter: Jonathan Wakely Opened: 2011-11-14 Last modified: 2012-09-24

View other active issues in [futures.async].

View all other issues in [futures.async].

View all issues with New status.

Discussion:

30.6.8 [futures.async] p5 says

If the implementation chooses the launch::async policy,

That should say a non-timed waiting function, otherwise, calling a timed waiting function can block indefinitely waiting for the associated thread to complete, rather than timing out after the specified time.

Since std::thread does not provide a timed_join() function (nor does Pthreads, making it impossible on many platforms) there is no way for a timed waiting function to try to join but return early due to timeout, therefore timed waiting functions either cannot guarantee to timeout or cannot be used to meet the requirement to block until the thread is joined. In order to allow timed waiting functions to timeout the requirement should only apply to non-timed waiting functions.

Proposed resolution:

This wording is relative to the FDIS.

Change 30.6.8 [futures.async] p5 as indicated:

If the implementation chooses the launch::async policy,


2101. Some transformation types can produce impossible types

Section: 20.9.7 [meta.trans] Status: Open Submitter: Daniel Krügler Opened: 2011-11-18 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

Table 53 — "Reference modifications" says in regard to the type trait add_lvalue_reference (emphasize mine)

If T names an object or function type then the member typedef type shall name T&;

The problem with this specification is that function types with cv-qualifier or ref-qualifier, like

void() const
void() &

are also affected by the first part of the rule, but this would essentially mean, that instantiating add_lvalue_reference with such a type would attempt to form a type that is not defined in the C++ type system, namely

void(&)() const
void(&)() &

The general policy for TransformationTraits is to define always some meaningful mapping type, but this does not hold for add_lvalue_reference, add_rvalue_reference, and in addition to these two for add_pointer as well. The latter one would attempt to form the invalid types

void(*)() const
void(*)() &

A possible reason why those traits were specified in this way is that in C++03 (and that means for TR1), cv-qualifier were underspecified in the core language and several compilers just ignored them during template instantiations. This situation became fixed by adopting CWG issues 295 and 547.

While there is possibly some core language clarification needed (see reflector messages starting from c++std-core-20740), it seems also clear that the library should fix the specification. The suggested resolution follows the style of the specification of the support concepts PointeeType and ReferentType defined in N2914.

[21012, Kona]

Move to NAD.

These cv- and ref-qualified function types are abberations in the type system, and do not represent any actual entity defined by the language. The notion of cv- and ref- qualification applies only to the implicit *this reference in a member function.

However, these types can be produced by quirks of template metaprogramming, the question remains what the library should do about it. For exmaple, add_reference returns the original type if passed a reference type, or a void type. Conversely, add_pointer will refurn a pointer to the referenced type when passed a reference.

It is most likely that the 'right' answer in any case will depend on the context that the question is being asked, in terms of forming these obscure types. The best the LWG can do is allow an error to propogate back to the user, so they can provide their own meaningful answer in their context - with additional metaprogramming on their part. The consensus is that if anyone is dangerous enough with templates to get themselves into this problem, they will also have the skills to resolve the problem themselves. This is not going to trip up the non-expert developer.

Lastly, it was noted that this problem arises only because the language is inconsistent in providing us these nonesense types that do no really represent anything in the language. There may be some way Core or Evolution could give us a more consistent type system so that the LWG does not need to invent an answer at all, should this question need resolving. This is another reason to not specify anything at the LWG trait level at this time, leaving the other working groups free to produce the 'right' answer that we can then follow without changing the meaning of exisitng, well-defined programs.

[21012, post-Kona]

Move back to Open. Daniel is concerned that this is not an issue we can simply ignore, further details to follow.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change Table 53 — "Reference modifications" in 20.9.7.2 [meta.trans.ref] as indicated:

    Table 53 — Reference modifications
    Template Comments
    template <class T>
    struct
    add_lvalue_reference;
    If T names an object type or if T names a function type that does not have
    cv-qualifiers or a ref-qualifier
    then the member typedef type
    shall name T&; otherwise, if T names a type “rvalue reference to T1” then
    the member typedef type shall name T1&; otherwise, type shall name T.
    template <class T>
    struct
    add_rvalue_reference;
    If T names an object type or if T names a function type that does not have
    cv-qualifiers or a ref-qualifier
    then the member typedef type
    shall name T&&; otherwise, type shall name T. [ Note: This rule reflects
    the semantics of reference collapsing (8.3.2 [dcl.ref]). For example, when a type T
    names a type T1&, the type add_rvalue_reference<T>::type is not an
    rvalue reference. — end note ]
  2. Change Table 56 — "Pointer modifications" in 20.9.7.5 [meta.trans.ptr] as indicated:

    Table 56 — Pointer modifications
    Template Comments
    template <class T>
    struct add_pointer;
    The member typedef type shall name the same type as
    If T names a function type that has cv-qualifiers or a ref-qualifier
    then the member typedef type shall name T; otherwise, it
    shall name the same type as
    remove_reference<T>::type*.

2103. std::allocator_traits<std::allocator<T>>::propagate_on_container_move_assignment

Section: 20.6.9 [default.allocator] Status: Ready Submitter: Ai Azuma Opened: 2011-11-08 Last modified: 2012-09-24

View all other issues in [default.allocator].

View all issues with Ready status.

Discussion:

"std::allocator_traits<std::allocator<T>>::propagate_on_container_move_assignment::value" is specified as "false", according to (20.6.9 [default.allocator]) and (20.6.8.1 [allocator.traits.types]). However, according to (23.2.1 [container.requirements.general]), this specification leads to the unneeded requirements (MoveInsertable and MoveAssignable of the value type) on the move assignment operator of containers with the default allocator.

Proposed resolution:

Either of the following two changes;

  1. adding the nested typedef like "typedef std::true_type propagate_on_container_move_assignment;" in the definition of std::allocator class template,
  2. adding the explicit partial specialization of "std::allocator_traits" class template for "std::allocator" class template, in which "propagate_on_container_move_assignment" nested typedef is specified as "std::true_type".

Pablo prefers the first resolution.

[2011-12-02: Pablo comments]

This issue has potentially some overlap with 2108. Should the trait always_compare_equal been added, this issue's resolution should be improved based on that.

[2012, Kona]

Move to Ready.

Proposed resolution:

This wording is relative to the FDIS.

Change 20.6.9 [default.allocator], the class template allocator synopsis as indicated:

namespace std {
  template <class T> class allocator;

  // specialize for void:
  template <> class allocator<void> {
  public:
    typedef void* pointer;
    typedef const void* const_pointer;
    // reference-to-void members are impossible.
    typedef void value_type;
    template <class U> struct rebind { typedef allocator<U> other; };
  };

  template <class T> class allocator {
  public:
    typedef size_t size_type;
    typedef ptrdiff_t difference_type;
    typedef T* pointer;
    typedef const T* const_pointer;
    typedef T& reference;
    typedef const T& const_reference;
    typedef T value_type;
    template <class U> struct rebind { typedef allocator<U> other; };
    typedef true_type propagate_on_container_move_assignment;

    […]
  };
}

2104. unique_lock move-assignment should not be noexcept

Section: 30.4.2.2 [thread.lock.unique] Status: New Submitter: Anthony Williams Opened: 2011-11-27 Last modified: 2012-09-24

View all issues with New status.

Discussion:

I just noticed that the unique_lock move-assignment operator is declared noexcept. This function may call unlock() on the wrapped mutex, which may throw.

Suggested change: remove the noexcept specification from unique_lock::operator=(unique_lock&&) in 30.4.2.2 [thread.lock.unique] and 30.4.2.2.1 [thread.lock.unique.cons].

Daniel:

I think the situation is actually a bit more complex as it initially looks.

First, the effects of the move-assignment operator are (emphasize mine):

Effects: If owns calls pm->unlock().

Now according to the BasicLockable requirements:

m.unlock()

3 Requires: The current execution agent shall hold a lock on m.

4 Effects: Releases a lock on m held by the current execution agent.

Throws: Nothing.

This shows that unlock itself is a function with narrow contract and for this reasons no unlock function of a mutex or lock itself does have a noexcept specifier according to our mental model.

Now the move-assignment operator attempts to satisfy these requirement of the function and calls it only when it assumes that the conditions are ok, so from the view-point of the caller of the move-assignment operator it looks as if the move-assignment operator would in total a function with a wide contract.

The problem with this analysis so far is, that it depends on the assumed correctness of the state "owns".

Looking at the construction or state-changing functions, there do exist several ones that depend on caller-code satisfying the requirements and there is one guy, who looks most suspicious:

unique_lock(mutex_type& m, adopt_lock_t);

11 Requires: The calling thread own the mutex.
[…]
13 Postconditions: pm == &m and owns == true.

because this function does not even call lock() (which may, but is not required to throw an exception if the calling thread does already own the mutex). So we have in fact still a move-assignment operator that might throw an exception, if the mutex was either constructed or used (call of lock) incorrectly.

The correct fix seems to me to also add a "Throws: Nothing" element to the move-assignment operator, because using it correctly shall now throw an exception.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change 30.4.2.2 [thread.lock.unique], class template unique_lock synopsis as indicated:

    namespace std {
      template <class Mutex>
      class unique_lock {
      public:
        typedef Mutex mutex_type;
        […]
        unique_lock(unique_lock&& u) noexcept;
        unique_lock& operator=(unique_lock&& u) noexcept;
        […]
      };
    }
    
  2. Change 30.4.2.2.1 [thread.lock.unique.cons] around p22 as indicated:

    unique_lock& operator=(unique_lock&& u) noexcept;
    

    -22- Effects: If owns calls pm->unlock().

    -23- Postconditions: pm == u_p.pm and owns == u_p.owns (where u_p is the state of u just prior to this construction), u.pm == 0 and u.owns == false.

    -24- [Note: With a recursive mutex it is possible for both *this and u to own the same mutex before the assignment. In this case, *this will own the mutex after the assignment and u will not. — end note]

    -??- Throws: Nothing.


2105. Inconsistent requirements on const_iterator's value_type

Section: 23.2.1 [container.requirements.general] Status: Ready Submitter: Jeffrey Yasskin Opened: 2011-11-28 Last modified: 2012-09-24

View other active issues in [container.requirements.general].

View all other issues in [container.requirements.general].

View all issues with Ready status.

Discussion:

In the FDIS, Table 96 specifies X::const_iterator as a "constant iterator type whose value type is T". However, Table 97 specifies X::const_reverse_iterator as an "iterator type whose value type is const T" and which is defined as reverse_iterator<const_iterator>. But reverse_iterator::value_type is just "typename iterator_traits<Iterator>::value_type" 24.5.1.1 [reverse.iterator], so const_iterator and const_reverse_iterator must have the same value_type.

The resolution to issue 322 implies that const_reverse_iterator should change.

[2012, Kona]

Move to Ready.

Proposed resolution:

This wording is relative to the FDIS.

Change Table 97 — "Reversible container requirements" as indicated

Table 97 — Reversible container requirements
Expression Return type Assertion/note pre-/post-condition Complexity
X::reverse_-
iterator
iterator type whose value type
is T
reverse_iterator<iterator> compile time
X::const_-
reverse_-
iterator
constant iterator type whose value type
is const T
reverse_iterator<const_iterator> compile time

2106. move_iterator wrapping iterators returning prvalues

Section: 24.5.3 [move.iterators] Status: Review Submitter: Dave Abrahams Opened: 2011-11-30 Last modified: 2012-09-24

View all other issues in [move.iterators].

View all issues with Review status.

Discussion:

Shouldn't move_iterator be specialized so that if the iterator it wraps returns a prvalue when dereferenced, the move_iterator also returns by value? Otherwise, it creates a dangling reference.

Howard: I believe just changing move_iterator<I>::reference would do. A direction might be testing on is_reference<iterator_traits<I>::reference>, or is_reference<decltype(*declval<I>())>.

Daniel: I would prefer to use a consistent style among the iterator adaptors, so I suggest to keep with the iterator_traits typedefs if possible.

using reference = typename conditional<
  is_reference<typename iterator_traits<Iterator>::reference>::value,
  value_type&&,
  value_type
>::type;

We might also want to ensure that if Iterator's reference type is a reference, the referent is equal to value_type (after removal of cv-qualifiers). In general we have no such guarantee.

Marc: In the default case where we don't return value_type&&, should we use value_type or should we keep the reference type of the wrapped iterator?

Daniel: This suggestion looks appealing at first, but the problem here is that using this typedef can make it impossible for move_iterator to satisfy its contract, which means returning an rvalue of the value type (Currently it says rvalue-reference, but this must be fixed as of this issue anyway). I think that user-code can reasonably expect that when it has constructed an object m of move_iterator<It>, where It is a valid mutable iterator type, the expression

It::value_type&& rv = *m;

is well-formed.

Let's set R equal to iterator_traits<Iterator>::reference in the following. We can discuss the following situations:

  1. R is a reference type: We can only return the corresponding xvalue of R, if value_type is reference-related to the referent type, else this is presumably no forward iterator and we cannot say much about it, except that it must be convertible to value_type, so it better should return a prvalue.
  2. R is not a reference type: In this case we can rely on a conversion to value_type again, but not much more. Assume we would return R directly, this might turn out to have a conversion to an lvalue-reference type of the value type (for example). If that is the case, this would indirectly violate the contract of move_iterator.

In regard to the first scenario I suggest that implementations are simply required to check that V2 = remove_cv<remove_reference<R>::type>::type is equal to the value type V1 as a criterion to return this reference as an xvalue, in all other cases it should return the value type directly as prvalue.

The additional advantage of this strategy is, that we always ensure that reference has the correct cv-qualification, if R is a real reference.

It is possible to improve this a bit by indeed supporting reference-related types, this would require to test is_same<V1, V2>::value || is_base_of<V1, V2>::value instead. I'm unsure whether (a) this additional effort is worth it and (b) a strict reading of the forward iterator requirements seems not to allow to return a reference-related type (Whether this is a defect or not is another question).

[2011-12-05: Marc Glisse comments and splits into two resolution alternatives]

I guess I am looking at the speed of:

value_type x;
x = *m;

(copy construction would likely trigger copy elision and thus be neutral) instead of the validity of:

value_type&& x = *m;

In this sense, Daniels earlier proposition that ignored value_type and just did switch_lvalue_ref_to_rvalue_ref<reference> was easier to understand (and it didn't require thinking about reference related types).

The currently proposed resolution has been split into two alternatives.

[2012, Kona]

Move to Review.

Alisdair: This only applies to input iterators, so keep that in mind when thinking about this.

STL: I see what B is doing, but not A.

Howard: I agree.

Alisdair: Should we use add_rvalue_reference?

STL: No, we do not want reference collapsing.

STL: Re A, messing with the CV qualification scares me.

Alisdair: Agree. That would break my intent.

STL: Actually I don't think it's actually wrong, but I still don't see what it's doing.

Alisdair: A is picking the value type, B is picking the proxy type.

Howard: I like returning the proxy type.

STL: Returning a reference (B) seems right, because the requirements say "reference". I suspect that B works correctly if you have a move iterator wrapping a move iterator wrapping a thing. I think that A would mess up the type in the middle.

Considerable discussion about which version is correct, checking various examples.

STL: Still think B is right. Still don't understand A. In A we are losing the proxyness.

Howard: Agree 100%. We don't want to lose the proxy. If it's const, so be it.

STL: B is also understandable by mortals.

Howard: Remove to review, keep A but move it out of the proposed resolution area (but keep it for rational).

Alisdair: Adding an explanatory note might be a good idea, if someone wants to write one.

Walter: Concerned about losing the word "reference" in p.1.

Howard: move_iterator will return an xvalue or a prvalue, both of which are rvalues.

[Proposed resolution A, rejected in preference to the currently proposed resolution (B)

  1. Change 24.5.3 [move.iterators] p1 as indicated:

    Class template move_iterator is an iterator adaptor with the same behavior as the underlying iterator except that its dereference operator implicitly converts the value returned by the underlying iterator’s dereference operator to an rvalue referenceof the value type. Some generic algorithms can be called with move iterators to replace copying with moving.

  2. Change 24.5.3.1 [move.iterator], class template move_iterator synopsis, as indicated:

    namespace std {
      template <class Iterator>
      class move_iterator {
      public:
        typedef Iterator iterator_type;
        typedef typename iterator_traits<Iterator>::difference_type difference_type;
        typedef Iterator pointer;
        typedef typename iterator_traits<Iterator>::value_type value_type;
        typedef typename iterator_traits<Iterator>::iterator_category iterator_category;
        typedef value_type&&see below reference;
        […]
      };
    }
    
  3. Immediately following the class template move_iterator synopsis in 24.5.3.1 [move.iterator] insert a new paragraph as indicated:

    -?- Let R be iterator_traits<Iterator>::reference and let V be iterator_traits<Iterator>::value_type. If is_reference<R>::value is true and if remove_cv<remove_reference<R>::type>::type is the same type as V, the template instantiation move_iterator<Iterator> shall define the nested type named reference as a synonym for remove_reference<R>::type&&, otherwise as a synonym for V.

]

Proposed resolution:

This wording is relative to the FDIS.

This section shows two mutually exclusive resolutions — only one can be picked!

  1. Change 24.5.3 [move.iterators] p1 as indicated:

    Class template move_iterator is an iterator adaptor with the same behavior as the underlying iterator except that its dereference operator implicitly converts the value returned by the underlying iterator’s dereference operator to an rvalue reference. Some generic algorithms can be called with move iterators to replace copying with moving.

  2. Change 24.5.3.1 [move.iterator], class template move_iterator synopsis, as indicated:

    namespace std {
      template <class Iterator>
      class move_iterator {
      public:
        typedef Iterator iterator_type;
        typedef typename iterator_traits<Iterator>::difference_type difference_type;
        typedef Iterator pointer;
        typedef typename iterator_traits<Iterator>::value_type value_type;
        typedef typename iterator_traits<Iterator>::iterator_category iterator_category;
        typedef value_type&&see below reference;
        […]
      };
    }
    
  3. Immediately following the class template move_iterator synopsis in 24.5.3.1 [move.iterator] insert a new paragraph as indicated:

    -?- Let R be iterator_traits<Iterator>::reference. If is_reference<R>::value is true, the template instantiation move_iterator<Iterator> shall define the nested type named reference as a synonym for remove_reference<R>::type&&, otherwise as a synonym for R.


2108. No way to identify allocator types that always compare equal

Section: 17.6.3.5 [allocator.requirements] Status: Open Submitter: Jonathan Wakely Opened: 2011-12-01 Last modified: 2012-09-24

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with Open status.

Discussion:

Whether two allocator objects compare equal affects the complexity of container copy and move assignments and also the possibility of an exception being thrown by container move assignments. The latter point means container move assignment cannot be noexcept when propagate_on_container_move_assignment (POCMA) is false for the allocator because there is no way to detect at compile-time if two allocators will compare equal. LWG 2013 means this affects all containers using std::allocator, but even if that is resolved, this affects all stateless allocators which do not explicitly define POCMA to true_type.

One solution would be to add an "always_compare_equal" trait to allocator_traits, but that would be duplicating information that is already defined by the type's equality operator if that operator always returns true. Requiring users to write operator== that simply returns true and also explicitly override a trait to repeat the same information would be unfortunate and risk user errors that allow the trait and actual operator== to disagree.

Dave Abrahams suggested a better solution in message c++std-lib-31532, namely to allow operator== to return true_type, which is convertible to bool but also detectable at compile-time. Adopting this as the recommended way to identify allocator types that always compare equal only requires a slight relaxation of the allocator requirements so that operator== is not required to return bool exactly.

The allocator requirements do not make it clear that it is well-defined to compare non-const values, that should be corrected too.

In message c++std-lib-31615 Pablo Halpern suggested an always_compare_equal trait that could still be defined, but with a sensible default value rather than requiring users to override it, and using that to set sensible values for other allocator traits:

Do we still need always_compare_equal if we can have an operator== that returns true_type? What would its default value be? is_empty<A> || is_convertible<decltype(a == a), true_type>::value, perhaps? One benefit I see to such a definition is that stateless C++03 allocators that don't use the true_type idiom will still benefit from the new trait.

[…]

One point that I want to ensure doesn't get lost is that if we adopt some sort of always_compare_equal-like trait, then propagate_on_container_swap and propagate_on_container_move_assignment should default to always_compare_equal. Doing this will eliminate unnecessary requirements on the container element type, as per [LWG 2103].

Optionally, operator== for std::allocator could be made to return true_type, however if LWG 2103 is adopted that is less important.

Alberto Ganesh Barbati: Suggest either always_compare_equal, all_objects_(are_)equivalent, or all_objects_compare_equal.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change Table 27 — "Descriptive variable definitions" in 17.6.3.5 [allocator.requirements]:

    Table 27 — Descriptive variable definitions
    Variable Definition
    a3, a4 an rvalue ofvalues of (possibly const) type X
    b a value of (possibly const) type Y
  2. Change Table 28 — "Allocator requirements" in 17.6.3.5 [allocator.requirements]:

    Table 28 — Allocator requirements
    Expression Return type Assertion/note pre-/post-condition Default
    a1 == a2a3 == a4 convertible to 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.
    a1 != a2a3 != a4 convertible to bool same as !(a1 == a2)!(a3 == a4)
    a3 == b convertible to bool same as a3 ==
    Y::rebind<T>::other(b)
    a3 != b convertible to bool same as !(a3 == b)
    […]
    a.select_on_-
    container_copy_-
    construction()
    X Typically returns either a or
    X()
    return a;
    X::always_compares_equal Identical to or derived
    from true_type or
    false_type
    true_type if the expression x1 == x2 is
    guaranteed to be true for any two (possibly
    const) values x1, x2 of type X, when
    implicitly converted to bool. See Note B, below.
    true_type, if is_empty<X>::value is true or if
    decltype(declval<const X&>() == declval<const X&>())
    is convertible to true_type, otherwise false_type.
    […]

    Note A: […]

    Note B: If X::always_compares_equal::value or XX::always_compares_equal::value evaluate to true and an expression equivalent to x1 == x2 or x1 != x2 for any two values x1, x2 of type X evaluates to false or true, respectively, the behaviour is undefined.

  3. Change class template allocator_traits synopsis, 20.6.8 [allocator.traits] as indicated:

    namespace std {
      template <class Alloc> struct allocator_traits {
        typedef Alloc allocator_type;
        […]
        typedef see below always_compares_equal;
        typedef see below propagate_on_container_copy_assignment;
        […]
      };
    }
    
  4. Insert the following between 20.6.8.1 [allocator.traits.types] p6 and p7 as indicated:

    typedef see below always_compares_equal;
    

    -?- Type: Alloc::always_compares_equal if such a type exists; otherwise, true_type if is_empty<Alloc>::value is true or if decltype(declval<const Alloc&>() == declval<const Alloc&>()) is convertible to true_type; otherwise, false_type .

    typedef see below propagate_on_container_copy_assignment;
    

    -7- Type: Alloc::propagate_on_container_copy_assignment if such a type exits, otherwise false_type.

  5. Change class template allocator synopsis, 20.6.9 [default.allocator] as indicated:

    namespace std {
      template <class T> class allocator;
    
      // specialize for void:
      template <> class allocator<void> {
      public:
        typedef void* pointer;
        typedef const void* const_pointer;
        // reference-to-void members are impossible.
        typedef void value_type;
        template <class U> struct rebind { typedef allocator<U> other; };
      };
    
      template <class T> class allocator {
      public:
        typedef size_t size_type;
        typedef ptrdiff_t difference_type;
        typedef T* pointer;
        typedef const T* const_pointer;
        typedef T& reference;
        typedef const T& const_reference;
        typedef T value_type;
        template <class U> struct rebind { typedef allocator<U> other; };
        typedef true_type always_compares_equal;
    
        […]
      };
    }
    

2109. Incorrect requirements for hash specializations

Section: 19.5.5 [syserr.hash], 20.7.2.6 [util.smartptr.hash], 20.8.12 [unord.hash], 20.13.1 [type.index.synopsis], 21.6 [basic.string.hash], 23.3.7 [vector.bool], 30.3.1.1 [thread.thread.id] Status: Review Submitter: Daniel Krügler Opened: 2011-12-04 Last modified: 2012-09-24

View all issues with Review status.

Discussion:

20.7.2.6 [util.smartptr.hash] p2 is specified as follows:

Requires: the template specializations shall meet the requirements of class template hash (20.8.12).

The problem here is the usage of a Requires element, which is actually a pre-condition that a user of a component has to satisfy. But the intent of this wording is actually to be a requirement on implementations. The Requires element should be removed here and the wording should be improved to say what it was intended for.

We have similar situations in basically all other places where the specification of library-provided hash specializations is defined. Usually, the Requires element is incorrect. In the special case of hash<unique_ptr<T, D>> the implementation depends on the behaviour of hash specializations, that could be user-provided. In this case the specification needs to separate the requirements on these specializations and those that are imposed on the implementation.

[2012, Kona]

Update wording and move to Review.

Believe a simpler formulation is to simply string the term Requires: and leave the current wording intact, rather than strike the whole clause and replace it.

[Originally proposed wording for reference

  1. Change 19.5.5 [syserr.hash] as indicated:

    template <> struct hash<error_code>;
    

    -1- Requires: the template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <system_error> provides a definition for a specialization of the template hash<error_code>. The requirements for the members of this specialization are given in sub-clause 20.8.12 [unord.hash].

  2. Change 20.5.3 [bitset.hash] as indicated:

    template <size_t N> struct hash<bitset<N> >;
    

    -1- Requires: the template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <bitset> provides a definition for a partial specialization of the hash class template for specializations of class template bitset<N>. The requirements for the members of instantiations of this specialization are given in sub-clause 20.8.12 [unord.hash].

  3. Change 20.7.2.6 [util.smartptr.hash] as indicated:

    template <class T, class D> struct hash<unique_ptr<T, D> >;
    

    -1- Requires: the template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <memory> provides a definition for a partial specialization of the hash class template for specializations of class template unique_ptr<T, D>. The requirements for the members of instantiations of this specialization are given in sub-clause 20.8.12 [unord.hash]. For an object p of type UP, where UP is unique_ptr<T, D>, hash<UP>()(p) shall evaluate to the same value as hash<typename UP::pointer>()(p.get()). The specialization hash<typename UP::pointer> shall be well-formed.

    -?- Requires: The specialization hash<typename UP::pointer> shall be well-formed and well-defined [Note: the general requirements of class template hash (20.8.12 [unord.hash]) are implied — end note].

    template <class T> struct hash<shared_ptr<T> >;
    

    -2- Requires: the template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <memory> provides a definition for a partial specialization of the hash class template for specializations of class template shared_ptr<T>. The requirements for the members of instantiations of this specialization are given in sub-clause 20.8.12 [unord.hash]. For an object p of type shared_ptr<T>, hash<shared_ptr<T> >()(p) shall evaluate to the same value as hash<T*>()(p.get()).

  4. Change 20.8.12 [unord.hash] p2 as indicated: [Comment: For unknown reasons the extended integer types are not mentioned here, which looks like an oversight to me and makes also the wording more complicated. See 2119 for this part of the problem. — end comment]

    template <> struct hash<bool>;
    template <> struct hash<char>;
    […]
    template <> struct hash<long double>;
    template <class T> struct hash<T*>;
    

    -2- Requires: the template specializations shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <functional> provides definitions for specializations of the hash class template for each cv-unqualified arithmetic type except for the extended integer types. This header also provides a definition for a partial specialization of the hash class template for any pointer type. The requirements for the members of these specializations are given in sub-clause 20.8.12 [unord.hash].

  5. Change 20.13.4 [type.index.hash] p1 as indicated:

    template <> struct hash<type_index>;
    

    -1- Requires: the template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <typeindex> provides a definition for a specialization of the template hash<type_index>. The requirements for the members of this specialization are given in sub-clause 20.8.12 [unord.hash]. For an object index of type type_index, hash<type_index>()(index) shall evaluate to the same result as index.hash_code().

  6. Change 21.6 [basic.string.hash] p1 as indicated:

    template <> struct hash<string>;
    template <> struct hash<u16string>;
    template <> struct hash<u32string>;
    template <> struct hash<wstring>;
    

    -1- Requires: the template specializations shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <string> provides definitions for specializations of the hash class template for the types string, u16string, u32string, and wstring. The requirements for the members of these specializations are given in sub-clause 20.8.12 [unord.hash].

  7. Change 23.3.7 [vector.bool] p7 as indicated:

    template <class Allocator> struct hash<vector<bool, Allocator> >;
    

    -7- Requires: the template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <vector> provides a definition for a partial specialization of the hash class template for specializations of class template vector<bool, Allocator>. The requirements for the members of instantiations of this specialization are given in sub-clause 20.8.12 [unord.hash].

  8. Change 30.3.1.1 [thread.thread.id] p14 as indicated:

    template <> struct hash<thread::id>;
    

    -14- Requires: the template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <thread> provides a definition for a specialization of the template hash<thread::id>. The requirements for the members of this specialization are given in sub-clause 20.8.12 [unord.hash].

]

Proposed resolution:

  1. Change 19.5.5 [syserr.hash] as indicated:

    template <> struct hash<error_code>;
    

    -1- Requires: tThe template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash].

  2. Change 20.5.3 [bitset.hash] as indicated:

    template <size_t N> struct hash<bitset<N> >;
    

    -1- Requires: tThe template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash]).

  3. Change 20.7.2.6 [util.smartptr.hash] as indicated:

    template <class T, class D> struct hash<unique_ptr<T, D> >;
    

    -1- Requires: tThe template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash]). For an object p of type UP, where UP is unique_ptr<T, D>, hash<UP>()(p) shall evaluate to the same value as hash<typename UP::pointer>()(p.get()). The specialization hash<typename UP::pointer> shall be well-formed.

    -?- Requires: The specialization hash<typename UP::pointer> shall be well-formed and well-defined, and shall meet the requirements of class template hash (20.8.12 [unord.hash]).

    template <class T> struct hash<shared_ptr<T> >;
    

    -2- Requires: tThe template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash]). For an object p of type shared_ptr<T>, hash<shared_ptr<T> >()(p) shall evaluate to the same value as hash<T*>()(p.get()).

  4. Change 20.8.12 [unord.hash] p2 as indicated: [Comment: For unknown reasons the extended integer types are not mentioned here, which looks like an oversight to me and makes also the wording more complicated. See 2119 for this part of the problem. — end comment]

    template <> struct hash<bool>;
    template <> struct hash<char>;
    […]
    template <> struct hash<long double>;
    template <class T> struct hash<T*>;
    

    -2- Requires: tThe template specializations shall meet the requirements of class template hash (20.8.12 [unord.hash]).

  5. Change 20.13.4 [type.index.hash] p1 as indicated:

    template <> struct hash<type_index>;
    

    -1- Requires: tThe template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash]). For an object index of type type_index, hash<type_index>()(index) shall evaluate to the same result as index.hash_code().

  6. Change 21.6 [basic.string.hash] p1 as indicated:

    template <> struct hash<string>;
    template <> struct hash<u16string>;
    template <> struct hash<u32string>;
    template <> struct hash<wstring>;
    

    -1- Requires: tThe template specializations shall meet the requirements of class template hash (20.8.12 [unord.hash]).

  7. Change 23.3.7 [vector.bool] p7 as indicated:

    template <class Allocator> struct hash<vector<bool, Allocator> >;
    

    -7- Requires: tThe template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash]).

  8. Change 30.3.1.1 [thread.thread.id] p14 as indicated:

    template <> struct hash<thread::id>;
    

    -14- Requires: tThe template specialization shall meet the requirements of class template hash (20.8.12 [unord.hash]).


2110. remove can't swap but note says it might

Section: 25.3.8 [alg.remove] Status: Ready Submitter: Howard Hinnant Opened: 2011-12-07 Last modified: 2012-09-24

View all other issues in [alg.remove].

View all issues with Ready status.

Discussion:

25.3.8 [alg.remove]/p1 says:

1 Requires: The type of *first shall satisfy the MoveAssignable requirements (Table 22).

This means that remove/remove_if can only use move assignment to permute the sequence. But then 25.3.8 [alg.remove]/p6 (non-normatively) contradicts p1:

6 Note: each element in the range [ret,last), where ret is the returned value, has a valid but unspecified state, because the algorithms can eliminate elements by swapping with or moving from elements that were originally in that range.

[2012, Kona]

Move to Ready.

Alisdair notes we could extend permission to use swap if it is available, but there is no interest. Accept the proposed resolution as written.

Proposed resolution:

This wording is relative to the FDIS.

Change 25.3.8 [alg.remove] as indicated:

template<class ForwardIterator, class T>
  ForwardIterator remove(ForwardIterator first, ForwardIterator last,
                         const T& value);

template<class ForwardIterator, class Predicate>
  ForwardIterator remove_if(ForwardIterator first, ForwardIterator last,
                            Predicate pred);

[…]

-6-Note: each element in the range [ret,last), where ret is the returned value, has a valid but unspecified state, because the algorithms can eliminate elements by swapping with or moving from elements that were originally in that range.


2111. Which unexpected/terminate handler is called from the exception handling runtime?

Section: 18.8.3.4 [terminate], D.13.3 [unexpected] Status: Open Submitter: Howard Hinnant Opened: 2011-12-06 Last modified: 2012-09-24

View all other issues in [terminate].

View all issues with Open status.

Discussion:

Prior to N3242, modified by N3189, we said this about unexpected():

Effects: Calls the unexpected_handler function in effect immediately after evaluating the throw-expression (D.13.1), if called by the implementation, or calls the current unexpected_handler, if called by the program.

and this about terminate():

Effects: Calls the terminate_handler function in effect immediately after evaluating the throw-expression (18.8.3.1), if called by the implementation, or calls the current terminate_handler function, if called by the program.

But now in both places we say:

Calls the current unexpected_handler function.

and:

Calls the current terminate function.

The difference is that in C++98/03 if a destructor reset a handler during stack unwinding, that new handler was not called if the unwinding later led to unexpected() or terminate() being called. But these new words say that this new handler is called. This is an ABI-breaking change in the way exceptions are handled. Was this change intentional?

N3189 was mainly about introducing exception safety and getters for the handlers. I don't recall the issue of which handler gets called being part of the discussion.

I propose that we revert to the C++98/03 behavior in this regard, lest ABI's such as the Itanium ABI are invalidated. A mechanical way to do this is to revert bullets 9 and 12 of N3189.

[2011-12-09: Daniel comments]

There was no such semantic change intended. It was an unfortunate side effect when trying to better separate different responsibilities in the previous wording.

A related issue is 2088.

[2012-01-30: Howard comments]

The C++98/03 wording is somewhat ambiguous:

Calls the terminate_handler function in effect immediately after evaluating the throw-expression...

There are potentially two throw-expressions being referred to here, and it is not clear if this sentence is referring to just the first or both:

  1. throw assignment-expression;
  2. throw;

There is ample evidence in current implementations that it is understood that only 1. was meant. But clearly both 1 and 2 could have been meant. We need a clarification. Does an execution of a rethrow (throw;) update which handlers can potentially be called?

  1. throw; // update handlers to get_xxx()?

My opinion: Go with existing practice, and clarify what that practice is, if surveys find that everyone does the same thing. Gcc 4.2 and Apple do 1. only, and do not reset the handlers to the current handlers on throw;.

If current practice is not unanimously one way or the other, I have no strong opinion. I have not found a motivating use case for the use of any particular handler. Most applications set the handlers once at the beginning of the program and then do not change them, and so will not be impacted by whatever decision is made here.

Proposed resolution:


2112. User-defined classes that cannot be derived from

Section: 17.6.5 [conforming], 20.6.8 [allocator.traits], 20.12.1 [allocator.adaptor.syn] Status: Open Submitter: Daniel Krügler Opened: 2011-11-30 Last modified: 2012-09-24

View all other issues in [conforming].

View all issues with Open status.

Discussion:

It is a very established technique for implementations to derive internally from user-defined class types that are used to customize some library component, e.g. deleters and allocators are typical candidates. The advantage of this approach is to possibly take advantage of the empty-base-class optimization (EBCO).

Whether or whether not libraries did take advantage of such a detail didn't much matter in C++03. Even though there did exist a portable idiom to prevent that a class type could be derived from, this idiom has never reached great popularity: The technique required to introduce a virtual base class and it did not really prevent the derivation, but only any construction of such a type. Further, such types are not empty as defined by the std::is_empty trait, so could easily be detected by implementations from TR1 on.

With the new C++11 feature of final classes and final member functions it is now very easy to define an empty, but not derivable from class type. From the point of the user it is quite natural to use this feature for types that he or she did not foresee to be derivable from.

On the other hand, most library implementations (including third-party libraries) often take advantage of EBCO applied to user-defined types used to instantiate library templates internally. As the time of submitting this issue the following program failed to compile on all tested library implementations:

#include <memory>

struct Noop final {
 template<class Ptr>
 void operator()(Ptr) const {}
};

std::unique_ptr<int, Noop> up;

In addition, many std::tuple implementations with empty, final classes as element types failed as well, due to a popular inheritance-based implementation technique. EBCO has also a long tradition to be used in library containers to efficiently store potentially stateless, empty allocators.

It seems that both user and library did the best they could: None of the affected types did impose explicit requirements on the corresponding user-defined types to be derivable from (This capability was not part of the required operations), and libraries did apply EBCO whereever possible to the convenience of the customer.

Nonetheless given the existence of non-derivable-from class types in C++11, libraries have to cope with failing derivations. How should that problem be solved?

It would certainly be possible to add weazel wording to the allocator requirements similar to what we had in C++03, but restricted to derivation-from requirements. I consider this as the bad solution, because it would add new requirements that never had existed before in this explicit form onto types like allocators.

Existing libraries presumably will need internal traits like __is_final or __is_derivable to make EBCO possible in the current form but excluding non-derivable class types. As of this writing this seems to happen already. Problem is that without a std::is_derivable trait, third-party libraries have no portable means to do the same thing as standard library implementations. This should be a good reason to make such a trait public available soon, but seems not essential to have now. Further, this issue should also be considered as a chance to recognice that EBCO has always been a very special corner case (There exist parallels to the previously existing odd core language rule that did make the interplay between std::auto_ptr and std::auto_ptr_ref possible) and that it would be better to provide explicit means for space-efficient storage, not necessarily restricted to inheritance relations, e.g. by marking data members with a special attribute.

At least two descriptions in the current standard should be fixed now for better clarification:

  1. As mentioned by Ganesh, 20.6.8 [allocator.traits] p1 currently contains a (non-normative) note "Thus, it is always possible to create a derived class from an allocator." which should be removed.

  2. As pointed out by Howard, the specification of scoped_allocator_adaptor as of 20.12.1 [allocator.adaptor.syn] already requires derivation from OuterAlloc, but only implies indirectly the same for the inner allocators due to the exposition-only description of member inner. This indirect implication should be normatively required for all participating allocators.

[2012, Kona]

What we really need is a type trait to indicate if a type can be derived from. Howard reports Clang and libc++ have had success with this approach.

Howard to provide wording, and AJM to alert Core that we may be wanting to add a new trait that requires compiler support.

Proposed resolution:


2114. Incorrect "contextually convertible to bool" requirements

Section: 17.6.3.3 [nullablepointer.requirements], 24.2.3 [input.iterators], 24.2.7 [random.access.iterators], 25.1 [algorithms.general], 25.4 [alg.sorting], 30.2.1 [thread.req.paramname] Status: Open Submitter: Daniel Krügler Opened: 2011-12-09 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

As of 17.6.3.1 [utility.arg.requirements] Table 17/18, the return types of the expressions

a == b

or

a < b

for types satisfying the EqualityComparable or LessThanComparable types, respectively, are required to be "convertible to bool" which corresponds to a copy-initialization context. But several newer parts of the library that refer to such contexts have lowered the requirements taking advantage of the new terminology of "contextually convertible to bool" instead, which corresponds to a direct-initialization context (In addition to "normal" direct-initialization constructions, operands of logical operations as well as if or switch conditions also belong to this special context).

One example for these new requirements are input iterators which satisfy EqualityComparable but also specify that the expression

a != b

shall be just "contextually convertible to bool". The same discrepancy exists for requirement set NullablePointer in regard to several equality-related expressions.

For random access iterators we have

a < b contextually convertible to bool

as well as for all derived comparison functions, so strictly speaking we could have a random access iterator that does not satisfy the LessThanComparable requirements, which looks like an artifact to me.

Even if we keep with the existing requirements based on LessThanComparable or EqualityComparable we still would have the problem that some current specifications are actually based on the assumption of implicit convertibility instead of "explicit convertibility", e.g. 20.7.1.4 [unique.ptr.special] p3:

template <class T1, class D1, class T2, class D2>
bool operator!=(const unique_ptr<T1, D1>& x, const unique_ptr<T2, D2>& y);

-3- Returns: x.get() != y.get().

Similar examples exist in 20.7.1.2.2 [unique.ptr.single.dtor] p2, 20.7.1.2.3 [unique.ptr.single.asgn] p9, 20.7.1.2.4 [unique.ptr.single.observers] p1+3+8, etc.

In all these places the expressions involving comparison functions (but not those of the conversion of a NullablePointer to bool!) assume to be "convertible to bool". I think this is a very natural assumption and all delegations of the comparison functions of some type X to some other API type Y in third-party code does so assuming that copy-initialization semantics will just work.

The actual reason for using the newer terminology can be rooted back to LWG 556. My hypotheses is that the resolution of that issue also needs a slight correction. Why so?

The reason for opening that issue were worries based on the previous "convertible to bool" wording. An expressions like "!pred(a, b)" might not be well-formed in those situations, because operator! might not be accessible or might have an unusual semantics (and similarly for other logical operations). This can indeed happen with unusual proxy return types, so the idea was that the evaluation of Predicate, BinaryPredicate (25.1 [algorithms.general] p8+9), and Compare (25.4 [alg.sorting] p2) should be defined based on contextual conversion to bool. Unfortunately this alone is not sufficient: In addition, I think, we also want the predicates to be (implicitly) convertible to bool! Without this wording, several conditions are plain wrong, e.g. 25.2.5 [alg.find] p2, which talks about "pred(*i) != false" (find_if) and "pred(*i) == false" (find_if_not). These expressions are not within a boolean context!

While we could simply fix all these places by proper wording to be considered in a "contextual conversion to bool", I think that this is not the correct solution: Many third-party libraries already refer to the previous C++03 Predicate definition — it actually predates C++98 and is as old as the SGI specification. It seems to be a high price to pay to switch to direct initialization here instead of fixing a completely different specification problem.

A final observation is that we have another definition for a Predicate in 30.2.1 [thread.req.paramname] p2:

If a parameter is Predicate, operator() applied to the actual template argument shall return a value that is convertible to bool.

The problem here is not that we have two different definitions of Predicate in the standard — this is confusing, but this fact alone is not a defect. The first (minor) problem is that this definition does not properly apply to function objects that are function pointers, because operator() is not defined in a strict sense. But the actually worse second problem is that this wording has the very same problem that has originally lead to LWG 556! We only need to look at 30.5.1 [thread.condition.condvar] p15 to recognice this:

while (!pred())
  wait(lock);

The negation expression here looks very familiar to the example provided in LWG 556 and is sensitive to the same "unusual proxy" problem. Changing the 30.2.1 [thread.req.paramname] wording to a corresponding "contextual conversion to bool" wouldn't work either, because existing specifications rely on "convertible to bool", e.g. 30.5.1 [thread.condition.condvar] p32+33+42 or 30.5.2 [thread.condition.condvarany] p25+26+32+33.

To summarize: I believe that LWG 556 was not completely resolved. A pessimistic interpretation is, that even with the current wording based on "contextually convertible to bool" the actual problem of that issue has not been fixed. What actually needs to be required here is some normative wording that basically expresses something along the lines of:

The semantics of any contextual conversion to bool shall be equivalent to the semantics of any implicit conversion to bool.

This is still not complete without having concepts, but it seems to be a better approximation. Another way of solving this issue would be to define a minimum requirements table with equivalent semantics. The proposed wording is a bit simpler but attempts to express the same thing.

[2012, Kona]

Agree with Daniel that we potentially broke some C++03 user code, accept the changes striking "contextually" from tables. Stefan to provide revised wording for section 25, and figure out changes to section 30.

Move to open, and then to Review when updated wording from Stefan is available.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change Table 25 — "NullablePointer requirements" in 17.6.3.3 [nullablepointer.requirements] as indicated:

    Table 25 — NullablePointer requirements
    Expression Return type Operational semantics
    […]
    a != b contextually convertible to bool !(a == b)
    a == np
    np == a
    contextually convertible to bool a == P()
    a != np
    np != a
    contextually convertible to bool !(a == np)
  2. Change Table 107 — "Input iterator requirements" in 24.2.3 [input.iterators] as indicated:

    Table 107 — Input iterator requirements (in addition to Iterator)
    Expression Return type Operational semantics Assertion/note
    pre-/post-condition
    a != b contextually convertible to bool !(a == b) pre: (a, b) is in the domain of ==.
    […]
  3. Change Table 111 — "Random access iterator requirements" in 24.2.7 [random.access.iterators] as indicated:

    Table 111 — Random access iterator requirements (in addition to bidirectional iterator)
    Expression Return type Operational semantics Assertion/note
    pre-/post-condition
    […]
    a < b contextually convertible to bool b - a > 0 < is a total ordering relation
    a > b contextually convertible to bool b < a > is a total ordering relation opposite to <.
    a >= b contextually convertible to bool !(a < b)
    a <= b contextually convertible to bool !(a > b)
  4. Change 25.1 [algorithms.general] p8+9 as indicated:

    -8- The Predicate parameter is used whenever an algorithm expects a function object (20.8 [function.objects]) that, when applied to the result of dereferencing the corresponding iterator, returns a value testable as true. In other words, if an algorithm takes Predicate pred as its argument and first as its iterator argument, it should work correctly in the construct pred(*first) implicitly or contextually converted to bool (Clause 4 [conv]). The function object pred shall not apply any non-constant function through the dereferenced iterator.

    -9- 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. In other words, 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 binary_pred(*first1, *first2) implicitly or contextually converted to bool (Clause 4 [conv]). BinaryPredicate always takes the first iterator's value_type as its first argument, that is, in those cases when T value is part of the signature, it should work correctly in the construct binary_pred(*first1, value) implicitly or contextually converted to bool (Clause 4 [conv]). binary_pred shall not apply any non-constant function through the dereferenced iterators.

  5. Change 25.4 [alg.sorting] p2 as indicated:

    -2- Compare is a function object type (20.8 [function.objects]). The return value of the function call operation applied to an object of type Compare, when implicitly or contextually converted to bool (4 [conv]), yields true if the first argument of the call is less than the second, and false otherwise. Compare comp is used throughout for algorithms assuming an ordering relation. It is assumed that comp will not apply any non-constant function through the dereferenced iterator.

  6. Change 30.2.1 [thread.req.paramname] p2 as indicated:

    -2- If a parameter is Predicate, operator() applied to the actual template argument shall return a value that is convertible to boolPredicate is a function object type (20.8 [function.objects]). The return value of the function call operation applied to an object of type Predicate, when implicitly or contextually converted to bool (4 [conv]), yields true if the corresponding test condition is satisfied, and false otherwise.


2115. Undefined behaviour for valarray assignments with mask_array index?

Section: 26.6.8 [template.mask.array] Status: Open Submitter: Thomas Plum Opened: 2011-12-10 Last modified: 2012-09-24

View all issues with Open status.

Discussion:

Recently I received a Service Request (SR) alleging that one of our testcases causes an undefined behavior. The complaint is that 26.6.8 [template.mask.array] in C++11 (and the corresponding subclause in C++03) are interpreted by some people to require that in an assignment "a[mask] = b", the subscript mask and the rhs b must have the same number of elements.

IMHO, if that is the intended requirement, it should be stated explicitly.

In any event, there is a tiny editorial cleanup that could be made:

In C++11, 26.6.8.1 [template.mask.array.overview] para 2 mentions

"the expression a[mask] = b;"

but the semicolon cannot be part of an expression. The correction could omit the semicolon, or change the word "expression" to "assignment" or "statement".

Here is the text of the SR, slightly modified for publication:

Subject: SR01174 LVS _26322Y31 has undefined behavior [open]

[Client:]
The test case t263.dir/_26322Y31.cpp seems to be illegal as it has an undefined behaviour. I searched into the SRs but found SRs were not related to the topic explained in this mail (SR00324, SR00595, SR00838).

const char vl[] = {"abcdefghijklmnopqrstuvwxyz"};
const char vu[] = {"ABCDEFGHIJKLMNOPQRSTUVWXYZ"};
const std::valarray<char> v0(vl, 27), vm5(vu, 5), vm6(vu, 6);
std::valarray<char> x = v0;
[…]
const bool vb[] = {false, false, true, true, false, true};
const std::valarray<bool> vmask(vb, 6);
x = v0;
x[vmask] = vm5;      // ***** HERE....
steq(&x[0], "abABeCghijklmnopqrstuvwxyz");
x2 = x[vmask];       // ***** ....AND HERE
[…]

This problem has already been discussed between [experts]: See thread http://gcc.gnu.org/ml/libstdc++/2009-11/threads.html#00051 Conclusion http://gcc.gnu.org/ml/libstdc++/2009-11/msg00099.html

[Plum Hall:]
Before I log this as an SR, I need to check one detail with you.

I did read the email thread you mentioned, and I did find a citation (see INCITS ISO/IEC 14882-2003 Section 26.3.2.6 on valarray computed assignments):

Quote: "If the array and the argument array do not have the same length, the behavior is undefined",

But this applies to computed assignment (*=, +=, etc), not to simple assignment. Here is the C++03 citation re simple assignment:

26.3.2.2 valarray assignment [lib.valarray.assign]

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

1 Each element of the *this array is assigned the value of the corresponding element of the argument array. The resulting behavior is undefined if the length of the argument array is not equal to the length of the *this array.

In the new C++11 (N3291), we find ...

26.6.2.3 valarray assignment [valarray.assign]

valarray<T>& operator=(const valarray<T>& v);

1 Each element of the *this array is assigned the value of the corresponding element of the argument array. If the length of v is not equal to the length of *this, resizes *this to make the two arrays the same length, as if by calling resize(v.size()), before performing the assignment.

So it looks like the testcase might be valid for C++11 but not for C++03; what do you think?

[Client:]
I quite agree with you but the two problems I mentioned:

x[vmask] = vm5;      // ***** HERE....
[…]
x2 = x[vmask];       // ***** ....AND HERE

refer to mask_array assignment hence target the C++03 26.3.8 paragraph. Correct?

[Plum Hall:]
I mentioned the contrast between C++03 26.3.2.2 para 1 versus C++11 26.6.2.3 para 1.

But in C++03 26.3.8, I don't find any corresponding restriction. Could you quote the specific requirement you're writing about?

[Client:]
I do notice the difference between c++03 26.3.2.2 and c++11 26.6.2.3 about assignments between different sized valarray and I perfectly agree with you.

But, as already stated, this is not a simple valarray assignment but a mask_array assignment (c++03 26.3.8 / c++11 26.6.8). See c++11 quote below:

26.6.8 Class template mask_array
26.6.8.1 Class template mask_array overview
[....]

  1. This template is a helper template used by the mask subscript operator: mask_array<T> valarray<T>::operator[](const valarray<bool>&).

  2. It has reference semantics to a subset of an array specified by a boolean mask. Thus, the expression a[mask] = b; has the effect of assigning the elements of b to the masked elements in a (those for which the corresponding element in mask is true.)

26.6.8.2 mask_array assignment

void operator=(const valarray<T>&) const;
const mask_array& operator=(const mask_array&) const;

1 These assignment operators have reference semantics, assigning the values of the argument array elements to selected elements of the valarray<T> object to which it refers.

In particular, [one of the WG21 experts] insisted on the piece "the elements of b".

That is why I reported the test t263.dir/_26322Y31.cpp having an undefined behaviour.

[Plum Hall:]
OK, I can see that I will have to ask WG21; I will file an appropriate issue with the Library subgroup. In the meantime, I will mark this testcase as "DISPUTED" so that it is not required for conformance testing, until we get a definitive opinion.

[2012, Kona]

Moved to Open.

There appears to be a real need for clarification in the standard, and implementations differ in their current interpretation. This will need some research by implementers and a proposed resolution before further discussion is likely to be fruitful.

Proposed resolution:


2116. std::swap noexcept(what?)

Section: 20.9.4.3 [meta.unary.prop] Status: Open Submitter: Dave Abrahams Opened: 2011-12-09 Last modified: 2012-09-24

View other active issues in [meta.unary.prop].

View all other issues in [meta.unary.prop].

View all issues with Open status.

Discussion:

IMO if we specified is_[nothrow_]constructible in terms of a variable declaration whose validity requires destructibility, it is clearly a bug in our specification and a failure to realize the actual original intent. The specification should have been in terms of placement-new.

Daniel:
At the time of the specification this was intended and the solution is not done by removing the destruction semantics of is_constructible.

The design of is_constructible was also impacted by the previous Constructible concept that explicitly contained destruction semantics, because during conceptification of the library it turned out to simplify the constraints in the library because you did not need to add Destructible all the time. It often was implied but never spoken out in C++03.

Pure construction semantics was considered as useful as well, so HasConstructor did also exist and would surely be useful as trait as well.

Another example that is often overlooked: This also affects wrapper types like pair, tuple, array that contain potentially more than one type: This is easy to understand if you think of T1 having a deleted destructor and T2 having a constructor that may throw: Obviously the compiler has potentially need to use the destructor of T1 in the constructor of std::pair<T1, T2> to ensure that the core language requirements are satisfied (All previous fully constructed sub-objects must be destructed).

The core language also honors this fact in 12.8 [class.copy] p11:

A defaulted copy/move constructor for a class X is defined as deleted (8.4.3 [dcl.fct.def.delete]) if X has:
[…]
— any direct or virtual base class or non-static data member of a type with a destructor that is deleted or inaccessible from the defaulted constructor,
[…]

Dave:
This is about is_nothrow_constructible in particular. The fact that it is foiled by not having a noexcept dtor is a defect.

[2012, Kona]

Move to Open.

is_nothrow_constructible is defined in terms of is_constructible, which is defined by looking at a hypothetical variable and asking whether the variable definition is known not to throw exceptions. The issue claims that this also examines the type's destructor, given the context, and thus will return false if the destructor can potentially throw. At least one implementation (Howard's) does return false if the constructor is noexcept(true) and the destructor is noexcept(false). So that's not a strained interpretation. The issue is asking for this to be defined in terms of placement new, instead of in terms of a temporary object, to make it clearer that is_nothrow_constructible looks at the noexcept status of only the constructor, and not the destructor.

Sketch of what the wording would look like:

require is_constructible, and then also require that a placement new operation does not throw. (Remembering the title of this issue... What does this imply for swap?

If we accept this resolution, do we need any changes to swap?

STL argues: no, because you are already forbidden from passing anything with a throwing desturctor to swap.

Dietmar argues: no, not true. Maybe statically the destructor can conceivably throw for some values, but maybe there are some values known not to throw. In that case, it's correct to pass those values to swap.

Proposed resolution:


2117. ios_base manipulators should have showgrouping/noshowgrouping

Section: 22.4.2.2.2 [facet.num.put.virtuals], 27.5.3.1.2 [ios::fmtflags], 27.5.6.1 [fmtflags.manip] Status: Open Submitter: Benjamin Kosnik Opened: 2011-12-15 Last modified: 2012-09-24

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

View all issues with Open status.

Discussion:

Iostreams should include a manipulator to toggle grouping on/off for locales that support grouped digits. This has come up repeatedly and been deferred. See LWG 826 for the previous attempt.

If one is using a locale that supports grouped digits, then output will always include the generated grouping characters. However, very plausible scenarios exist where one might want to output the number, un-grouped. This is similar to existing manipulators that toggle on/off the decimal point, numeric base, or positive sign.

See some user commentary here.

[21012, Kona]

Move to Open.

This is a feature request.

Walter is slightly uncomfortable with processing feature requests through the issues lists.

Alisdair says this is far from the first feature request that has come in from the issues list.

STL: The fact that you can turn off grouping on hex output is compelling.

Marshall: if we add this flag, we'll need to update tables 87-91 as well.

STL: If it has been implemented somewhere, and it works, we'd be glad to add it.

Howard: We need to say what the default is.

Alisdair sumarizes:

(1) We want clear wording that says what the effect is of turning the flag off;

(2) what the default values are, and

(3) how this fits into tables 87-90. (and 128)

Proposed resolution:

This wording is relative to the FDIS.

  1. Insert in 22.4.2.2.2 [facet.num.put.virtuals] paragraph 5:

    Stage 1: The first action of stage 1 is to determine a conversion specifier. The tables that describe this determination use the following local variables

    fmtflags flags = str.flags() ;
    fmtflags basefield = (flags & (ios_base::basefield));
    fmtflags uppercase = (flags & (ios_base::uppercase));
    fmtflags floatfield = (flags & (ios_base::floatfield));
    fmtflags showpos = (flags & (ios_base::showpos));
    fmtflags showbase = (flags & (ios_base::showbase));
    fmtflags showgrouping = (flags & (ios_base::showgrouping));
    
  2. Change header <ios> synopsis, 27.5.1 [iostreams.base.overview] as indicated:

    #include <iosfwd>
    
    namespace std {
      […]
      // 27.5.6, manipulators:
      […]
      ios_base& showpoint     (ios_base& str);
      ios_base& noshowpoint   (ios_base& str);
      ios_base& showgrouping  (ios_base& str);
      ios_base& noshowgrouping(ios_base& str);
      ios_base& showpos       (ios_base& str);
      ios_base& noshowpos     (ios_base& str);
      […]
    }
    
  3. Change class ios_base synopsis, 27.5.3 [ios.base] as indicated:

    namespace std {
      class ios_base {
      public:
      class failure;
        // 27.5.3.1.2 fmtflags
        typedef T1 fmtflags;
        […]
        static constexpr fmtflags showpoint = unspecified ;
        static constexpr fmtflags showgrouping = unspecified ;
        static constexpr fmtflags showpos = unspecified ;
        […]
      };
    }
    
  4. Add a new entry to Table 122 — "fmtflags effects" as indicated:

    Table 122 — fmtflags effects
    Element Effect(s) if set
    […]
    showpoint generates a decimal-point character unconditionally in generated floatingpoint output
    showgrouping generates grouping characters unconditionally in generated output
    […]
  5. After 27.5.3.1.2 [ios::fmtflags] p12 insert the following:

    ios_base& showgrouping(ios_base& str);
    

    -?- Effects: Calls str.setf(ios_base::showgrouping).

    -?- Returns: str.

    ios_base& noshowgrouping(ios_base& str);
    

    -?- Effects: Calls str.unsetf(ios_base::showgrouping).

    -?- Returns: str.


2118. unique_ptr for array does not support cv qualification conversion of actual argument

Section: 20.7.1.3 [unique.ptr.runtime] Status: Open Submitter: Alf P. Steinbach Opened: 2011-12-16 Last modified: 2012-09-24

View all other issues in [unique.ptr.runtime].

View all issues with Open status.

Discussion:

N3290 20.7.1.3.1 [unique.ptr.runtime.ctor] "unique_ptr constructors":

These constructors behave the same as in the primary template except that they do not accept pointer types which are convertible to pointer. [Note: One implementation technique is to create private templated overloads of these members. — end note]

This language excludes even pointer itself as type for the actual argument.

But of more practical concern is that both Visual C++ 10.0 and MinGW g++ 4.1.1 reject the code below, where only an implicit cv qualification is needed, which cv qualification is supported by the non-array version:

#include <memory>
using namespace std;

struct T {};

T* foo() { return new T; }
T const* bar() { return foo(); }

int main()
{
   unique_ptr< T const >       p1( bar() );        // OK
   unique_ptr< T const [] >    a1( bar() );        // OK

   unique_ptr< T const >       p2( foo() );        // OK
   unique_ptr< T const [] >    a2( foo() );        // ? this is line #15
}

The intent seems to be clearly specified in 20.7.1.3 [unique.ptr.runtime]/1 second bullet:

— Pointers to types derived from T are rejected by the constructors, and by reset.

But the following language in 20.7.1.3.1 [unique.ptr.runtime.ctor] then rejects far too much...

Proposed new wording of N3290 20.7.1.3.1 [unique.ptr.runtime.ctor] "unique_ptr constructors":

These constructors behave the same as in the primary template except that actual argument pointers p to types derived from T are rejected by the constructors. [Note: One implementation technique is to create private templated overloads of these members. — end note]

This will possibly capture the intent better, and avoid the inconsistency between the non-array and array versions of unique_ptr, by using nearly the exact same phrasing as for the paragraph explaining the intent.

[2012-08-25 Geoffrey Romer comments in c++std-lib-32978]

  1. The current P/R seems to intend to support at least two different implementation techniques — additional unusable templates that catch forbidden arguments or replacing existing constructors by templates that ensure ill-formed code inside the template body, when the requirements are not met. It seems unclear whether the current wording allows the second approach, though. It should be considered to allow both strategies or if that is not possible the note should be clearer.

  2. The very same problem exists for the reset member function, but even worse, because the current specification is more than clear that the deleted reset function will catch all cases not equal to pointer. It seems confusing at best to have different policies for the constructor and for the reset function. In this case, the question in regard to implementation freedom mentioned above is even more important.

  3. It's awkward to refer to "the constructors" twice in the same sentence; I suggest revising the sentence as "...except that they do not accept argument pointers p to types derived from T"

Proposed resolution:

This wording is relative to the FDIS.

Change 20.7.1.3.1 [unique.ptr.runtime.ctor] as indicated:

explicit unique_ptr(pointer p) noexcept;
unique_ptr(pointer p, see below d) noexcept;
unique_ptr(pointer p, see below d) noexcept;

These constructors behave the same as in the primary template except that they do not accept pointer types which are convertible to pointerargument pointers p to types derived from T are rejected by the constructors. [Note: One implementation technique is to create private templated overloads of these members. — end note]


2119. Missing hash specializations for extended integer types

Section: 20.8.12 [unord.hash] Status: Open Submitter: Daniel Krügler Opened: 2011-12-16 Last modified: 2012-09-24

View other active issues in [unord.hash].

View all other issues in [unord.hash].

View all issues with Open status.

Discussion:

According to the header <functional> synopsis 20.8 [function.objects] and to the explicit description in 20.8.12 [unord.hash] class template hash specializations shall be provided for all arithmetic types that are not extended integer types. This is not explicitly mentioned, but neither the list nor any normative wording does include them, so it follows by implication.

What are the reasons that extended integer types are excluded? E.g. for numeric_limits corresponding specializations are required. I would expect that an unordered_map with key type std::uintmax_t would just work, but that depends now on whether this type is an extended integer type or not.

This issue is not asking for also providing specializations for the cv-qualified arithmetic types. While this is surely a nice-to-have feature, I consider that restriction as a more secondary problem in practice.

The proposed resolution also fixes a problem mentioned in 2109 in regard to confusing requirements on user-defined types and those on implementations.

[2012, Kona]

Move to Open.

Agreed that it's a real issue and that the proposed wording fixes it. However, the wording change is not minimal and isn't consistent with the way we fixed hash wording elsewhere.

Alisdair will provide updated wording.

Proposed resolution:

This wording is relative to the FDIS.

Change 20.8.12 [unord.hash] p2 as indicated:

template <> struct hash<bool>;
template <> struct hash<char>;
[…]
template <> struct hash<long double>;
template <class T> struct hash<T*>;

-2- Requires: the template specializations shall meet the requirements of class template hash (20.8.12 [unord.hash])The header <functional> provides definitions for specializations of the hash class template for each cv-unqualified arithmetic type. This header also provides a definition for a partial specialization of the hash class template for any pointer type. The requirements for the members of these specializations are given in sub-clause 20.8.12 [unord.hash].


2120. What should async do if neither 'async' nor 'deferred' is set in policy?

Section: 30.6.8 [futures.async] Status: New Submitter: Jonathan Wakely Opened: 2012-01-01 Last modified: 2012-09-24

View other active issues in [futures.async].

View all other issues in [futures.async].

View all issues with New status.

Discussion:

Implementations already disagree, one returns an invalid future with no shared state, one chooses policy == async and one chooses policy == deferred, see c++std-lib-30839, c++std-lib-30840 and c++std-lib-30844. It's not clear if returning an invalid future is allowed by the current wording.

If the intention is to allow an empty future to be returned, then 30.6.8 [futures.async] p3 and p4 should be adjusted to clarify that a shared state might not be created and an invalid future might be returned.

If the intention is that a valid future is always returned, p3 should say something about the case where none of the conditions applies.

Proposed resolution:


2121. app for string streams

Section: 27.8.6 [stringstream.cons] Status: New Submitter: Nicolai Josuttis Opened: 2012-01-15 Last modified: 2012-09-24

View all issues with New 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 27.5.3.1.4 [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.

Proposed resolution:


2122. merge() stability for lists versus forward lists

Section: 23.3.5.5 [list.ops], 23.3.4.6 [forwardlist.ops] Status: Open Submitter: Nicolai Josuttis Opened: 2012-01-15 Last modified: 2012-09-24

View all other issues in [list.ops].

View all issues with Open status.

Discussion:

forward_list::merge() is specified in 23.3.4.6 [forwardlist.ops], p19 as follows:

This operation shall be stable: for equivalent elements in the two lists, the elements from *this shall always precede the elements from x.

But list::merge() is only specified in 23.3.5.5 [list.ops], p24 as follows:

Remarks: Stable.

Note that in general we define "stable" only for algorithms (see [defns.stable] and 17.6.5.7 [algorithm.stable]) so for member function we should explain it everywhere we use it.

Thus for lists we have to add:

Stable: for equivalent elements in the two lists, the elements from the list always precede the elements from the argument list.

This, BTW, was the specification we had with C++03.

In addition, I wonder whether we also have some guarantees regarding stability saying that the order of equivalent elements of each list merged remains stable (which would be my interpretation of just saying "stable", BTW).

Thus, I'd expect that for equivalent elements we guarantee that

[2012, Kona]

Move to Open.

STL says we need to fix up 17.6.5.7 to be stronger, and then the remarks for merge should just say "Remarks: Stable (see 17.6.5.7)"

Assigned to STL for word-smithing.

Proposed resolution:

This wording is relative to the FDIS.

  1. Change 23.3.5.5 [list.ops] as indicated:

    void                          merge(list<T,Allocator>& x);
    void                          merge(list<T,Allocator>&& x);
    template <class Compare> void merge(list<T,Allocator>& x, Compare comp);
    template <class Compare> void merge(list<T,Allocator>&& x, Compare comp);

    […]

    -24- Remarks: StableThis operation shall be stable: for equivalent elements in the two lists, the elements from *this shall always precede the elements from x and the order of equivalent elements of *this and x remains stable. If (&x != this) the range [x.begin(), x.end()) is empty after the merge. No elements are copied by this operation. The behavior is undefined if this->get_allocator() != x.get_allocator().

  2. Change 23.3.4.6 [forwardlist.ops] as indicated:

    void merge(forward_list<T,Allocator>& x);
    void merge(forward_list<T,Allocator>&& x);
    template <class Compare> void merge(forward_list<T,Allocator>& x, Compare comp);
    template <class Compare> void merge(forward_list<T,Allocator>&& x, Compare comp);

    […]

    -19- Effects: Merges x into *this. This operation shall be stable: for equivalent elements in the two lists, the elements from *this shall always precede the elements from x and the order of equivalent elements of *this and x remains stable. x is empty after the merge. If an exception is thrown other than by a comparison there are no effects. Pointers and references to the moved elements of x now refer to those same elements but as members of *this. Iterators referring to the moved elements will continue to refer to their elements, but they now behave as iterators into *this, not into x.


2123. merge() allocator requirements for lists versus forward lists

Section: 23.3.4.6 [forwardlist.ops] Status: Ready Submitter: Nicolai Josuttis Opened: 2012-01-15 Last modified: 2012-09-24

View all other issues in [forwardlist.ops].

View all issues with Ready status.

Discussion:

Sub-clause 23.3.5.5 [list.ops], p24 states for lists:

The behavior is undefined if this->get_allocator() != x.get_allocator().

But there is nothing like that for forward lists in 23.3.4.6 [forwardlist.ops], although I would expect the same undefined behavior there.

[2012, Kona]

Move to Ready.

Proposed resolution:

This wording is relative to the FDIS.

  1. Add a new paragraph after 23.3.4.6 [forwardlist.ops] p19 as indicated:

    void merge(forward_list<T,Allocator>& x);
    void merge(forward_list<T,Allocator>&& x);
    template <class Compare> void merge(forward_list<T,Allocator>& x, Compare comp);
    template <class Compare> void merge(forward_list<T,Allocator>&& x, Compare comp);

    […]

    -19- Effects: […]

    -?- Remarks: The behavior is undefined if this->get_allocator() != x.get_allocator().


2124. Seed sequence over-specified

Section: 26.5.1.2 [rand.req.seedseq] Status: Tentatively NAD Submitter: Alberto Ganesh Barbati Opened: 2012-01-16 Last modified: 2012-09-24

View other active issues in [rand.req.seedseq].

View all other issues in [rand.req.seedseq].

Discussion:

The seed sequence requirements described in 26.5.1.2 [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 26.5.1.2 [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 26.5.1.2 [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 26.5.1.2 [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 (3.9.1 [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. TimedMutex specification problem

Section: 30.4.1.3 [thread.timedmutex.requirements], 30.4.1.3.1 [thread.timedmutex.class] Status: New Submitter: Vicente J. Botet Escriba Opened: 2012-01-01 Last modified: 2012-09-24

View other active issues in [thread.timedmutex.requirements].

View all other issues in [thread.timedmutex.requirements].

View all issues with New status.

Discussion:

30.4.1.3.1 [thread.timedmutex.class] says:

The class timed_mutex shall satisfy all of the TimedMutex requirements (30.4.1.3 [thread.timedmutex.requirements]). It shall be a standardlayout class (Clause 9 [class]).

Problem here is that 30.4.1.3 [thread.timedmutex.requirements] does not define a requirement set named "TimedMutex", it only refers to "timed mutex types"

[See also issue 2126]

Proposed resolution:


2126. Several specification problems in regard to mutex requirements

Section: 30.4.1 [thread.mutex.requirements], 30.4.1.2.1 [thread.mutex.class], 30.4.1.2 [thread.mutex.requirements.mutex], 30.4.1.2.2 [thread.mutex.recursive], 30.4.1.3 [thread.timedmutex.requirements], 30.4.1.3.1 [thread.timedmutex.class], 30.4.1.3.2 [thread.timedmutex.recursive] Status: New Submitter: Pete Becker Opened: 2012-01-16 Last modified: 2012-09-24

View all other issues in [thread.mutex.requirements].

View all issues with New status.

Discussion:

30.4.1.2.1 [thread.mutex.class]/3 says that the class mutex "shall satisfy all the Mutex requirements (30.4.1 [thread.mutex.requirements])". 30.4.1.2.1 [thread.mutex.class] is part of 30.4.1 [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 30.4.1.2 [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 30.4.1.2.1 [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 30.4.1.2.2 [thread.mutex.recursive]/2 (recursive_mutex).

30.4.1.3 [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 30.4.1.3.1 [thread.timedmutex.class]/2 and 30.4.1.3.2 [thread.timedmutex.recursive]/2 should be turned into notes that refer to this new subclause and to the new subclause in 30.4.1.2 [thread.mutex.requirements.mutex].

[See also issue 2125]

Proposed resolution:


2127. Move-construction with raw_storage_iterator

Section: 20.6.10 [storage.iterator] Status: Open Submitter: Jonathan Wakely Opened: 2012-01-23 Last modified: 2012-09-24

View all other issues in [storage.iterator].

View all issues with Open status.

Discussion:

Aliaksandr Valialkin pointed out that raw_storage_iterator only supports constructing a new object from lvalues so cannot be used to construct move-only types:

template <typename InputIterator, typename T>
void move_to_raw_buffer(InputIterator first, InputIterator last, T *raw_buffer)
{
  std::move(first, last, std::raw_storage_iterator<T *, T>(raw_buffer));
}

This could easily be solved by overloading operator= for rvalues.

Dave Abrahams:

raw_storage_iterator causes exception-safety problems when used with any generic algorithm. I suggest leaving it alone and not encouraging its use.

Proposed resolution:

This wording is relative to N3337.

  1. Add a new signature to the synopsis in 20.6.10 [storage.iterator] p1:

    namespace std {
      template <class OutputIterator, class 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=(T&& element);
        raw_storage_iterator<OutputIterator,T>& operator++();
        raw_storage_iterator<OutputIterator,T> operator++(int);
    };
    }
    
  2. Insert the new signature and a new paragraph before p4:

    raw_storage_iterator<OutputIterator,T>& operator=(const T& element);
    raw_storage_iterator<OutputIterator,T>& operator=(T&& element);
    

    -?- Requires: For the first signature T shall be CopyConstructible. For the second signature T shall be MoveConstructible.

    -4- Effects: Constructs a value from element at the location to which the iterator points.

    -5- Returns: A reference to the iterator.


2128. Absence of global functions cbegin/cend

Section: 24.3 [iterator.synopsis], 24.6.5 [iterator.range] Status: Open Submitter: Dmitry Polukhin Opened: 2012-01-23 Last modified: 2012-09-24

View all other issues in [iterator.synopsis].

View all issues with Open status.

Discussion:

All standard containers support cbegin/cend member functions but corresponding global functions are missing. Proposed resolution it to add global cbegin/cend functions by analogy with global begin/end functions. This addition will unify things for users.

[2012, Kona]

STL: Range-based for loops do not use global begin/end (anymore).

Alisdair: We will have to make sure these will be available through many headers.

STL: Do this, including r and cr. This won't add any additional work.

Matt: Users will find it strange if these are not all available.

Alisdair: Should we have these available everywhere begin/end are available?

Marshall: Yes. Not any extra work.

Howard: Adding all of these means we need all of <iterator>.

STL: We already need it all.

Matt: We have to be careful what we are requiring if we include the r versions.

Jeffrey: If we include r, should they adapt if the container does not define reverse iteration?

STL: No. No special behavior. Should fail to compile. Up to user to add the reverse code--it's easy.

Howard: Anyway it will SFINAE out.

Alisdair: Error messages due to SFINAE are harder to understand than simple failure to compile.

STL: Agrees that SFINAE makes error messages much worse.

Action: STL to provide additional wording for the r variants. Move to Review once that wording is availalbe.

Proposed resolution:

This wording is relative to N3337.

  1. In 24.3 [iterator.synopsis], header iterator synopsis, add the following declarations::

    namespace std {
      […]
      // 24.6.5, range access:
      template <class C> auto begin(C& c) -> decltype(c.begin());
      template <class C> auto begin(const C& c) -> decltype(c.begin());
      template <class C> auto end(C& c) -> decltype(c.end());
      template <class C> auto end(const C& c) -> decltype(c.end());
      template <class C> auto cbegin(const C& c) -> decltype(c.cbegin());
      template <class C> auto cend(const C& c) -> decltype(c.cend());
      template <class T, size_t N> T* begin(T (&array)[N]);
      template <class T, size_t N> T* end(T (&array)[N]);
      template <class T, size_t N> const T* cbegin(T (&array)[N]);
      template <class T, size_t N> const T* cend(T (&array)[N]);
    }
    
  2. In 24.6.5 [iterator.range] after p5 add the following series of paragraphs:

    template <class C> auto cbegin(const C& c) -> decltype(c.cbegin());
    

    -?- Returns: c.cbegin().

    template <class C> auto cend(const C& c) -> decltype(c.cend());
    

    -?- Returns: c.cend().

    template <class T, size_t N> const T* cbegin(T (&array)[N]);
    

    -?- Returns: array.

    template <class T, size_t N> const T* cend(T (&array)[N]);
    

    -?- Returns: array + N.


2129. User specializations of std::initializer_list

Section: 17.6.4.2.1 [namespace.std], 18.9 [support.initlist] Status: Open Submitter: Richard Smith Opened: 2012-01-18 Last modified: 2012-09-24

View other active issues in [namespace.std].

View all other issues in [namespace.std].

View all issues with Open status.

Discussion:

Since the implementation is intended to magically synthesize instances of std::initializer_list (rather than by a constructor call, for instance), user specializations of this type can't generally be made to work. I can't find any wording which makes such specializations ill-formed, though, which leads me to suspect that they're technically legal under the provisions of 17.6.4.2.1 [namespace.std] p1.

[2012, Kona]

This sounds correct, but we need wording for a resultion.

Marshall Clow volunteers to produce wording.

Proposed resolution:


2130. Missing ordering constraints

Section: 29.3 [atomics.order] Status: New Submitter: Mark Batty Opened: 2012-02-22 Last modified: 2012-09-24

View all other issues in [atomics.order].

View all issues with New status.

Discussion:

C11 issue 407

It seems that both C11 and C++11 are missing the following two derivatives of this rule:

For atomic modifications A and B of an atomic object M, if there is a memory_order_seq_cst fence X such that A is sequenced before X, and X precedes B in S, then B occurs later than A in the modification order of M.

For atomic modifications A and B of an atomic object M, if there is a memory_order_seq_cst fence Y such that Y is sequenced before B, and A precedes Y in S, then B occurs later than A in the modification order of M.

Above wording has been suggested for the Technical Corrigendum of C11 via issue 407, details can be found here.

[2012-03-19: Daniel proposes a slightly condensed form to reduce wording duplications]

[2012-03-20: Hans comments]

The usage of the term atomic operations in 29.3 [atomics.order] p7 is actually incorrect and should better be replaced by atomic modifications as used in the C11 407 wording.

There seems to be a similar wording incorrectness used in 1.10 [intro.multithread] p17 which should be corrected as well.

Proposed resolution:

This wording is relative to N3376.

  1. [Drafting note: The project editor is kindly asked to consider to replace in 1.10 [intro.multithread] p17 the phrase "before an operation B on M" by "before a modification B of M".]

  2. Change 29.3 [atomics.order] paragraph 7 as indicated: [Drafting note: Note that the wording change intentionally does also replace the term atomic operation by atomic modification]

    -7- For atomic operations A and B on an atomic object 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 modification order of M. For atomic modifications A and B of an atomic object M, B occurs later than A in the modification order of M if:

    -8- [ Note: memory_order_seq_cst ensures sequential consistency only for a program that is free of data races and 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 ensure a total order only for the fences themselves. Fences cannot, in general, be used to restore sequential consistency for atomic operations with weaker ordering specifications. — end note ]


2131. Member function getline taking a string as parameter

Section: 27.7.2.3 [istream.unformatted] Status: New Submitter: Loïc Joly Opened: 2012-03-05 Last modified: 2012-09-24

View other active issues in [istream.unformatted].

View all other issues in [istream.unformatted].

View all issues with New 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)
}

Proposed resolution:

This wording is relative to N3376.

  1. Change the class template basic_istream synopsis, 27.7.2.1 [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 27.7.2.3 [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)


2132. std::function ambiguity

Section: 20.8.11.2.1 [func.wrap.func.con] Status: New Submitter: Ville Voutilainen Opened: 2012-02-28 Last modified: 2012-09-24

View all other issues in [func.wrap.func.con].

View all issues with New status.

Discussion:

Consider the following:

#include <functional>

void f(std::function<void()>) {}
void f(std::function<void(int)>) {}

int main() {
  f([]{});
  f([](int){});
}

The calls to f in main are ambiguous. Apparently because the conversion sequences to std::function from the lambdas are identical. The standard specifies that the function object given to std::function "shall be Callable (20.8.11.2) for argument types ArgTypes and return type R." It doesn't say that if this is not the case, the constructor isn't part of the overload set.

Daniel: During the preparation of N3123 it turned out that there are no longer reasons to refer to INVOKE as a conceptually entity alone, its real implementation as a function template invoke is possible but was deferred for a later point in time. Defining a type trait for the Callable requirement would also be possible, so there seem to be no technical reasons why the template constructor of std::function should not be constrained. The below suggested wording does this without introducing a special trait for this. This corresponds to the way that has been used to specify the result_of trait. Note that the definition of the Callable requirement is perfectly suitable for this, because it is a pure syntactically based requirement and can be directly transformed into a constrained template.

The suggested resolution also applies such wording to the "perfectly forwarding" assignment operator

template<class F> function& operator=(F&&);

The positive side-effect of this is that it automatically implements a solution to a problem similar to that mentioned in issue 1234.

It would be possible to apply similar constraints to the member signatures

template<class F> function& operator=(reference_wrapper<F>);

template<class F, class A> void assign(F&&, const A&);

as well. At this point there does not seem to be a pestering reason to do so.

Proposed resolution:

This wording is relative to N3376.

  1. Change the following paragraphs in 20.8.11.2.1 [func.wrap.func.con]: [Editorial comment: The removal of the seemingly additional no-throw requirements of copy constructor and destructor of A is recommended, because they are already part of the Allocator requirements. Similar clean-up has been suggested by 2070end comment]

    template<class F> function(F f);
    template<class F, class A> function(allocator_arg_t, const A& a, F f);
    

    -7- Requires: F shall be CopyConstructible. f shall be Callable (20.8.11.2 [func.wrap.func]) for argument types ArgTypes and return type R. The copy constructor and destructor of A shall not throw exceptions.

    -?- Remarks: These constructors shall not participate in overload resolution unless declval<F&>() is Callable (20.8.11.2 [func.wrap.func]) for argument types ArgTypes and return type R.

    […]

    template<class F> function& operator=(F&& f);
    

    -18- Effects: function(std::forward<F>(f)).swap(*this);

    -19- Returns: *this

    -?- Remarks: This assignment operator shall not participate in overload resolution unless declval<decay<F>::type&>() is Callable (20.8.11.2 [func.wrap.func]) for argument types ArgTypes and return type R.


2133. Attitude to overloaded comma for iterators

Section: 17.6.5.4 [global.functions] Status: New Submitter: Yakov Galka Opened: 2012-01-25 Last modified: 2012-09-24

View all other issues in [global.functions].

View all issues with New status.

Discussion:

17.6.5.4 [global.functions] says

Unless otherwise specified, global and non-member functions in the standard library shall not use functions from another namespace which are found through argument-dependent name lookup (3.4.2).

This sounds clear enough. There are just two problems:

  1. Both implementations I tested (VS2005 and GCC 3.4.3) do unqualified calls to the comma operator in some parts of the library with operands of user-defined types.

  2. The standard itself does this in the description of some algorithms. E.g. uninitialized_copy is defined as:

    Effects:

    for (; first != last; ++result, ++first)
      ::new (static_cast<void*>(&*result))
        typename iterator_traits<ForwardIterator>::value_type(*first);
    

If understood literally, it is required to call operator,(ForwardIterator, InputIterator).

For detailed discussion with code samples see here.

Proposal:

  1. Add an exception to the rule in 17.6.5.4 [global.functions] by permitting the implementation to call the comma operator as much as it wants to. I doubt we want this. or
  2. Fix the description of the said algorithms and perhaps add a note to 17.6.5.4 [global.functions] that brings attention of the implementers to avoid this pitfall.

Proposed resolution:


2134. Redundant Mutex requirement?

Section: 30.4.1.2 [thread.mutex.requirements.mutex] Status: New Submitter: Pete Becker Opened: 2012-03-05 Last modified: 2012-09-24

View all other issues in [thread.mutex.requirements.mutex].

View all issues with New status.

Discussion:

30.4.1.2 [thread.mutex.requirements.mutex]/11 says that prior unlock operations synchronize with m.lock().

30.4.1.2 [thread.mutex.requirements.mutex]/19 says that if m.try_lock() succeeds, prior unlock operations synchronize with the operation.

30.4.1.2 [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.

Proposed resolution:


2135. Unclear requirement for exceptions thrown in condition_variable::wait()

Section: 30.5.1 [thread.condition.condvar], 30.5.2 [thread.condition.condvarany] Status: New Submitter: Pete Becker Opened: 2012-03-06 Last modified: 2012-09-24

View other active issues in [thread.condition.condvar].

View all other issues in [thread.condition.condvar].

View all issues with New status.

Discussion:

condition_varible::wait() (and, presumably, condition_variable_any::wait(), although I haven't looked at it) says that it calls lock.unlock(), and if condition_variable::wait() exits by an exception it calls lock.lock() on the way out. But if the initial call to lock.unlock() threw an exception, does it make sense to call lock.lock()? We simply don't know the state of that lock object, and it's probably better not to touch it.

That aside, once the wait() call has been unblocked, it calls lock.lock(). If lock.lock() throws an exception, what happens? The requirement is:

If the function exits via an exception, lock.lock() shall be called prior to exiting the function scope.

That can be read in two different ways. One way is as if it said "lock.lock() shall have been called …", i.e. the original, failed, call to lock.lock() is all that's required. But a more natural reading is that wait has to call lock.lock() again, even though it already failed.

I think this wording suffers from being too general. There are two possible exception sources: the initial call to lock.unlock() and the final call to lock.lock(). Each one should have its own requirement. Lumping them together muddles things.

Proposed resolution:


2136. Postconditions vs. exceptions

Section: 17.5.1 [structure] Status: New Submitter: Jens Maurer Opened: 2012-03-08 Last modified: 2012-09-24

View all issues with New status.

Discussion:

The front matter in clause 17 should clarify that postconditions will not hold if a standard library function exits via an exception. Postconditions or guarantees that apply when an exception is thrown (beyond the basic guarantee) are described in an "Exception safety" section.

Proposed resolution:


2137. Misleadingly constrained post-condition in the presence of exceptions

Section: 28.8.3 [re.regex.assign] Status: New Submitter: Jonathan Wakely Opened: 2012-03-08 Last modified: 2012-09-24

View all other issues in [re.regex.assign].

View all issues with New status.

Discussion:

The post-conditions of basic_regex<>::assign 28.8.3 [re.regex.assign] p16 say:

If no exception is thrown, flags() returns f and mark_count() returns the number of marked sub-expressions within the expression.

The default expectation in the library is that post-conditions only hold, if there is no failure (see also 2136), therefore the initial condition should be removed to prevent any misunderstanding.

Proposed resolution:

This wording is relative to N3376.

template <class string_traits, class A>
  basic_regex& assign(const basic_string<charT, string_traits, A>& s,
    flag_type f = regex_constants::ECMAScript);

[…]

-15- Effects: Assigns the regular expression contained in the string s, interpreted according the flags specified in f. If an exception is thrown, *this is unchanged.

-16- Postconditions: If no exception is thrown, flags() returns f and mark_count() returns the number of marked sub-expressions within the expression.


2138. atomic_flag::clear should not accept memory_order_consume

Section: 29.7 [atomics.flag] Status: New Submitter: Ben Viglietta Opened: 2012-03-08 Last modified: 2012-09-24

View other active issues in [atomics.flag].

View all other issues in [atomics.flag].

View all issues with New status.

Discussion:

N3376 29.7 [atomics.flag]/7 says this about atomic_flag::clear:

Requires: The order argument shall not be memory_order_acquire or memory_order_acq_rel.

In addition, memory_order_consume should be disallowed, since it doesn't meaningfully apply to store operations. It's already disallowed on the analogous atomic<T>::store. The proposed updated text would be:

Requires: The order argument shall not be memory_order_consume, memory_order_acquire, or memory_order_acq_rel.

Proposed resolution:

This wording is relative to N3376.

void atomic_flag_clear(volatile atomic_flag *object) noexcept;
void atomic_flag_clear(atomic_flag *object) noexcept;
void atomic_flag_clear_explicit(volatile atomic_flag *object, memory_order order) noexcept;
void atomic_flag_clear_explicit(atomic_flag *object, memory_order order) noexcept;
void atomic_flag::clear(memory_order order = memory_order_seq_cst) volatile noexcept;
void atomic_flag::clear(memory_order order = memory_order_seq_cst) noexcept;

-7- Requires: The order argument shall not be memory_order_consume, memory_order_acquire, nor memory_order_acq_rel.

-8- Effects: Atomically sets the value pointed to by object or by this to false. Memory is affected according to the value of order.


2139. What is a user-defined type?

Section: 17.6.4.2.1 [namespace.std], 19.5 [syserr], 20.6.7.1 [allocator.uses.trait], 20.8.9.1.1 [func.bind.isbind], X [func.bind.isplace], 20.8.12 [unord.hash], 20.9.7.6 [meta.trans.other], 22.3.1 [locale], 22.4.1.4 [locale.codecvt], 28.12.1.4 [re.regiter.incr] Status: New Submitter: Loïc Joly Opened: 2012-03-08 Last modified: 2012-09-24

View other active issues in [namespace.std].

View all other issues in [namespace.std].

View all issues with New status.

Discussion:

The expression "user-defined type" is used in several places in the standard, but I'm not sure what it means. More specifically, is a type defined in the standard library a user-defined type?

From my understanding of English, it is not. From most of the uses of this term in the standard, it seem to be considered as user defined. In some places, I'm hesitant, e.g. 17.6.4.2.1 [namespace.std] p1:

A program may add a template specialization for any standard library template to namespace std only if the declaration depends on a user-defined type and the specialization meets the standard library requirements for the original template and is not explicitly prohibited.

Does it mean we are allowed to add in the namespace std a specialization for std::vector<std::pair<T, U>>, for instance?

Additional remarks from the reflector discussion: The traditional meaning of user-defined types refers to class types and enum types, but the library actually means here user-defined types that are not (purely) library-provided. Presumably a new term - like user-provided type - should be introduced and properly defined.

Proposed resolution:


2140. Meaning of notify_all_at_thread_exit synchronization requirement?

Section: 30.5 [thread.condition] Status: New Submitter: Pete Becker Opened: 2012-03-06 Last modified: 2012-09-24

View all other issues in [thread.condition].

View all issues with New status.

Discussion:

notify_all_at_thread_exit has the following synchronization requirement:

Synchronization: The call to notify_all_at_thread_exit and the completion of the destructors for all the current thread's variables of thread storage duration synchronize with (1.10 [intro.multithread]) calls to functions waiting on cond.

The functions waiting on cond have already been called, otherwise they wouldn't be waiting. So how can a subsequent call to notify_all_at_thread_exit synchronize with them?

Also, "synchronizes with" is a relationship between library calls (1.10 [intro.multithread]/8), so it's not meaningful for completion of destructors for non-library objects. Presumably the intention wasn't so make library destructors special here.

[2012-03-09 Jeffrey Yasskin comments:]

I think the text should say that "notify_all_at_thread_exit and destructor calls are sequenced before the lk.unlock()", and leave it at that, unless there's a funny implementation I haven't thought of.

[2012-03-19 Hans Boehm comments:]

I think the synchronization clause should just be replaced with (modulo wording tweaks):

"The implied lk.unlock() call is sequenced after the destruction of all objects with thread storage duration associated with the current thread."

as Jeffrey suggested.

To use this correctly, the notifying thread has to essentially acquire the lock, set a variable indicating it's done, call notify_all_at_thread_exit(), while the waiting thread acquires the lock, and repeatedly waits on the cv until the variable is set, and then releases the lock. That ensures that we have the proper synchronizes with relationship as a result of the lock.

Proposed resolution:

This wording is relative to N3376.

  1. Modify 30.5 [thread.condition] p8 as indicated:

    void notify_all_at_thread_exit(condition_variable& cond, unique_lock<mutex> lk);
    

    […]

    -8- Synchronization: The call to notify_all_at_thread_exit and the completion of the destructors for all the current thread's variables of thread storage duration synchronize with (1.10 [intro.multithread]) calls to functions waiting on cond The implied lk.unlock() call is sequenced after the destruction of all objects with thread storage duration associated with the current thread.


2141. common_type trait produces reference types

Section: 20.9.7.6 [meta.trans.other] Status: New Submitter: Doug Gregor Opened: 2012-03-11 Last modified: 2012-09-24

View all other issues in [meta.trans.other].

View all issues with New status.

Discussion:

The type computation of the common_type type trait is defined as

template <class T, class U>
 struct common_type<T, U> {
   typedef decltype(true ? declval<T>() : declval<U>()) type;
 };

This means that common_type<int, int>::type is int&&, because

Users of common_type do not expect to get a reference type as the result; the expectation is that common_type will return a non-reference type to which all of the types can be converted.

Daniel: In addition to that it should be noted that without such a fix the definition of std::unique_ptr's operator< in 20.7.1.4 [unique.ptr.special] (around p4) is also broken: In the most typical case (with default deleter), the determination of the common pointer type CT will instantiate std::less<CT> which can now be std::less<T*&&>, which will not be the specialization of pointer types that guarantess a total order.

Given the historic constext of common_type original specification, the proper resolution to me seems to be using std::decay instead of std::remove_reference:

template <class T, class U>
struct common_type<T, U> {
  typedef typename decay<decltype(true ? declval<T>() : declval<U>())>::type type;
};

At that time rvalues had no identity in this construct and rvalues of non-class types have no cv-qualification. With this change we would ensure that

common_type<int, int>::type == common_type<const int, const int>::type == int

Note that this harmonizes with the corresponding heterogenous case, which has already the exact same effect:

common_type<int, long>::type == common_type<const int, const long>::type == long

Proposed resolution:

This wording is relative to N3376.

  1. In 20.9.7.6 [meta.trans.other] p3, change the common_type definition to

    template <class T, class U>
    struct common_type<T, U> {
      typedef typename decay<decltype(true ? declval<T>() : declval<U>())>::type type;
    };
    

2142. packaged_task::operator() synchronization too broad?

Section: 30.6.9.1 [futures.task.members] Status: New Submitter: Pete Becker Opened: 2012-03-12 Last modified: 2012-09-24

View other active issues in [futures.task.members].

View all other issues in [futures.task.members].

View all issues with New status.

Discussion:

According to 30.6.9.1 [futures.task.members] p.18:

[A] successful call to [packaged_task::]operator() synchronizes with a call to any member function of a future or shared_future object that shares the shared state of *this.

This requires that the call to operator() synchronizes with calls to future::wait_for, future::wait_until, shared_future::wait_for, and shared_future::wait_until, even when these functions return because of a timeout.

Proposed resolution:


2143. ios_base::xalloc should be thread-safe

Section: 27.5.3 [ios.base] Status: New Submitter: Alberto Ganesh Barbati Opened: 2012-03-14 Last modified: 2012-09-24

View all other issues in [ios.base].

View all issues with New status.

Discussion:

The static function ios_base::xalloc() could be called from multiple threads and is not covered by 17.6.4.10 [res.on.objects] and 17.6.5.9 [res.on.data.races]. Adding a thread-safety requirement should not impose a significant burden on implementations, as the function can be easily implemented with hopefully lock-free atomics.

Proposed resolution:

This wording is relative to N3376.

  1. In 27.5.3.5 [ios.base.storage] add a new paragraph after paragraph 1:

    static int xalloc();
    

    -1- Returns: index ++.

    -?- Remarks: Concurrent access to this function by multiple threads shall not result in a data race (1.10 [intro.multithread]).


2144. Missing noexcept specification in type_index

Section: 20.13 [type.index] Status: New Submitter: Daniel Krügler Opened: 2012-03-18 Last modified: 2012-09-24

View all other issues in [type.index].

View all issues with New status.

Discussion:

The class type type_index is a thin wrapper of type_info to adapt it as a valid associative container element. Similar to type_info, all member functions have an effective noexcept(true) specification, with the exception of hash_code() and name(). The actual effects of these functions is a direct call to type_info's hash_code() and name function, but according to 18.7 [support.rtti] these are both noexcept functions, so there is no reason for not declaring them as noexcept, too. In fact, one of the suggested changes of the original proposing paper N2530 specifically was to ensure that type_info would get a hash_code() function that guarantees not to throw exceptions (during that time the hash requirements did not allow to exit with an exception). From this we can conclude that type_index::hash_code() was intended to be nothrow.

It seems both consistent and technically simply to require these functions to be noexcept.

Proposed resolution:

This wording is relative to N3376.

  1. Modify the class type_index synopsis, 20.13.2 [type.index.overview] as indicated:

    namespace std {
      class type_index {
      public:
        type_index(const type_info& rhs) noexcept;
        bool operator==(const type_index& rhs) const noexcept;
        bool operator!=(const type_index& rhs) const noexcept;
        bool operator< (const type_index& rhs) const noexcept;
        bool operator<= (const type_index& rhs) const noexcept;
        bool operator> (const type_index& rhs) const noexcept;
        bool operator>= (const type_index& rhs) const noexcept;
        size_t hash_code() const noexcept;
        const char* name() const noexcept;
      private:
        const type_info* target; // exposition only
        // Note that the use of a pointer here, rather than a reference,
        // means that the default copy/move constructor and assignment
        // operators will be provided and work as expected.
      };
    }
    
  1. Modify the prototype definitions in 20.13.3 [type.index.members] as indicated:

    size_t hash_code() const noexcept;
    

    -8- Returns: target->hash_code()

    const char* name() const noexcept;
    

    -9- Returns: target->name()


2145. error_category default constructor

Section: 19.5.1 [syserr.errcat] Status: New Submitter: Howard Hinnant Opened: 2012-03-21 Last modified: 2012-09-24

View all other issues in [syserr.errcat].

View all issues with New status.

Discussion:

Should error_category have a default constructor?

If you look at the synopsis in 19.5.1.1 [syserr.errcat.overview], it appears the answer is no. There is no default constructor declared and there is another constructor declared (which should inhibit a default constructor).

However in paragraph 1 of the same section, descriptive text says:

Classes may be derived from error_category to support categories of errors in addition to those defined in this International Standard.

How shall classes derived from error_category construct their base?

Jonathan Wakely: In N2066 error_category was default-constructible. That is still the case in N2241, because no other constructor is declared. Then later N2422 (issue 6) declares the copy constructor as deleted, but doesn't add a default constructor, causing it to be no longer default-constructible. That looks like an oversight to me, and I think there should be a public default constructor.

Daniel: A default-constructor indeed should be provided to allow user-derived classes as described by the standard. I suggest this one to be both noexcept and constexpr. The latter allows user-derived non-abstract classes to take advantage of the special constant initialization rule of 3.6.2 [basic.start.init] p2 b2 for objects with static (or thread) storage duration in namespace scope. Note that a constexpr constructor is feasible here, even though there exists a non-trivial destructor and even though error_category is not a literal type (see std::mutex for a similar design choice).

In addition to that the proposed resolution fixes another minor glitch: According to 17.5.2.2 [functions.within.classes] virtual destructors require a semantics description.

Alberto Ganesh Barbati: I would suggest to remove =default from the constructor instead. Please consider that defaulting a constructor or destructor may actually define them as deleted under certain conditions (see 12.1 [class.ctor]/5 and 12.4 [class.dtor]/5). Removing =default is easier than providing wording to ensures that such conditions do not occur.

Proposed resolution:

This wording is relative to N3376.

  1. Modify the class error_category synopsis, 19.5.1.1 [syserr.errcat.overview] as indicated: [Drafting note: According to the general noexcept library guidelines destructors should not have any explicit exception specification. This destructor was overlooked during the paper analysis — end note]

    namespace std {
      class error_category {
      public:
        constexpr error_category() noexcept;
        virtual ~error_category() noexcept;
        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;
      };
    }
    
  2. Before 19.5.1.2 [syserr.errcat.virtuals] p1 insert a new prototype description as indicated:

    virtual ~error_category();
    

    -?- Effects: Destroys an object of class error_category.

  3. Before 19.5.1.3 [syserr.errcat.nonvirtuals] p1 insert a new prototype description as indicated:

    constexpr error_category() noexcept;
    

    -?- Effects: Constructs an object of class error_category.


2146. Are reference types Copy/Move-Constructible/Assignable or Destructible?

Section: 17.6.3.1 [utility.arg.requirements] Status: New Submitter: Nikolay Ivchenkov Opened: 2012-03-23 Last modified: 2012-09-24

View other active issues in [utility.arg.requirements].

View all other issues in [utility.arg.requirements].

View all issues with New status.

Discussion:

According to 17.6.3.1 [utility.arg.requirements] p1

The template definitions in the C++ standard library refer to various named requirements whose details are set out in tables 17–24. In these tables, T is an object or reference type to be supplied by a C++ program instantiating a template; a, b, and c are values of type (possibly const) T; s and t are modifiable lvalues of type T; u denotes an identifier; rv is an rvalue of type T; and v is an lvalue of type (possibly const) T or an rvalue of type const T.

Is it really intended that T may be a reference type? If so, what should a, b, c, s, t, u, rv, and v mean? For example, are "int &" and "int &&" MoveConstructible?

As far as I understand, we can explicitly specify template arguments for std::swap and std::for_each. Can we use reference types there?

  1. #include <iostream>
    #include <utility>
    
    int main()
    {
      int x = 1;
      int y = 2;
      std::swap<int &&>(x, y); // undefined?
      std::cout << x << " " << y << std::endl;
    }
    
  2. #include <algorithm>
    #include <iostream>
    #include <iterator>
    #include <utility>
    
    struct F
    {
      void operator()(int n)
      {
        std::cout << n << std::endl;
        ++count;
      }
      int count;
    } f;
    
    int main()
    {
      int arr[] = { 1, 2, 3 };
      auto&& result = std::for_each<int *, F &&>( // undefined?
        std::begin(arr),
        std::end(arr),
        std::move(f));
      std::cout << "count: " << result.count << std::endl;
    }
    

Are these forms of usage well-defined?

Let's also consider the following constructor of std::thread:

template <class F, class ...Args>
explicit thread(F&& f, Args&&... args);

Requires: F and each Ti in Args shall satisfy the MoveConstructible requirements.

When the first argument of this constructor is an lvalue (e.g. a name of a global function), template argument for F is deduced to be lvalue reference type. What should "MoveConstructible" mean with regard to an lvalue reference type? Maybe the wording should say that std::decay<F>::type and each std::decay<Ti>::type (where Ti is an arbitrary item in Args) shall satisfy the MoveConstructible requirements?

Proposed resolution:


2147. Unclear hint type in Allocator's allocate function

Section: 17.6.3.5 [allocator.requirements] Status: New Submitter: Daniel Krügler Opened: 2012-03-05 Last modified: 2012-09-24

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with New status.

Discussion:

According to Table 28 — "Allocator requirements", the expression

a.allocate(n, u)

expects as second argument a value u that is described in Table 27 as:

a value of type YY::const_pointer obtained by calling YY::allocate, or else nullptr.

This description leaves it open, whether or whether not a value of type YY::const_void_pointer is valid or not. The corresponding wording in C++03 is nearly the same, but in C++03 there did not exist the concept of a general void_pointer for allocators. There is some evidence for support of void pointers because the general allocator_traits template declares

static pointer allocate(Alloc& a, size_type n, const_void_pointer hint);

and the corresponding function for std::allocator<T> is declared as:

pointer allocate(size_type, allocator<void>::const_pointer hint = 0);

As an additional minor wording glitch (especially when comparing with the NullablePointer requirements imposed on const_pointer and const_void_pointer), the wording seems to exclude lvalues of type std::nullptr_t, which looks like an unwanted artifact to me.

Proposed resolution:

This wording is relative to N3376.

  1. Change Table 27 — "Descriptive variable definitions" in 17.6.3.5 [allocator.requirements]:

    Table 27 — Descriptive variable definitions
    Variable Definition
    u a value of type YY::const_pointer obtained by calling YY::allocate, or else nullptrXX::const_void_pointer obtained by conversion from a result value of YY::allocate, or else a value of type (possibly const) std::nullptr_t.

2148. Hashing enums should be supported directly by std::hash

Section: 20.8.12 [unord.hash] Status: New Submitter: Ville Voutilainen Opened: 2012-04-10 Last modified: 2012-09-24

View other active issues in [unord.hash].

View all other issues in [unord.hash].

View all issues with New status.

Discussion:

The paper proposes various hashing improvements. What it doesn't mention is hashing of enums; enums are integral types, and users expect them to have built-in hashing support, rather than having to convert enums to ints for uses with unordered containers and other uses of hashes. Daniel Krügler explains in c++std-lib-32412 that this is not achievable with a SFINAEd hash specialization because it would require a partial specialization with a type parameter and a non-type parameter with a default argument, which is currently not allowed, and hence the fixes in N3333 should be adopted instead.

Proposed resolution:


2149. Concerns about 20.8/5

Section: 20.8 [function.objects] Status: New Submitter: Scott Meyers Opened: 2012-02-15 Last modified: 2012-09-24

View other active issues in [function.objects].

View all other issues in [function.objects].

View all issues with New status.

Discussion:

20.8 [function.objects] p5 says:

To enable adaptors and other components to manipulate function objects that take one or two arguments it is required that the function objects correspondingly provide typedefs argument_type and result_type for function objects that take one argument and first_argument_type, second_argument_type, and result_type for function objects that take two arguments.

I have two concerns about this paragraph. First, the wording appears to prescribe a requirement for all function objects in valid C++ programs, but it seems unlikely that that is the intent. As such, the scope of the requirement is unclear. For example, there is no mention of these typedefs in the specification for closures (5.1.2), and Daniel Krügler has explained in the thread at http://tinyurl.com/856plkn that conforming implementations can detect the difference between closures with and without these typedefs. (Neither gcc 4.6 nor VC10 appear to define typedefs such as result_type for closure types. I have not tested other compilers.)

Second, the requirement appears to be unimplementable in some cases, notably for function objects returned from std::bind, as Howard Hinnant explains in the thread at http://tinyurl.com/6q5bos4.

From what I can tell, the standard already defines which adaptability typedefs must be provided by various kinds of function objects in the specifications for those objects. Examples include the function objects specified in 20.8.3 [refwrap]-20.8.8 [negators]. I therefore suggest that 20.8 [function.objects]/5 simply be removed from the standard. I don't think it adds anything except opportunities for confusion.

Proposed resolution:

This wording is relative to N3376.

Remove 20.8 [function.objects] p5:

To enable adaptors and other components to manipulate function objects that take one or two arguments it is required that the function objects correspondingly provide typedefs argument_type and result_type for function objects that take one argument and first_argument_type, second_argument_type, and result_type for function objects that take two arguments.


2150. Unclear specification of find_end

Section: 25.2.6 [alg.find.end] Status: New Submitter: Andrew Koenig Opened: 2012-03-28 Last modified: 2012-09-24

View all issues with New status.

Discussion:

25.2.6 [alg.find.end] describes the behavior of find_end as returning:

The last iterator i in the range [first1,last1 - (last2 - first2)) such that for any nonnegative integer n < (last2 - first2), the following corresponding conditions hold: *(i + n) == *(first2 + n), pred(*(i + n), *(first2 + n)) != false.

Does "for any" here mean "for every" or "there exists a"? I think it means the former, but it could be interpreted either way.

Daniel: The same problem exists for the following specifications from Clause 25 [algorithms]:

  1. 25.2.13 [alg.search] p2 and p6
  2. 25.3.10 [alg.reverse] p4
  3. 25.3.13 [alg.partitions] p5 and p9
  4. 25.4 [alg.sorting] p5
  5. 25.4.2 [alg.nth.element] p1
  6. 25.4.3.1 [lower.bound] p2
  7. 25.4.3.2 [upper.bound] p2
  8. 25.4.7 [alg.min.max] p21 and p23

Proposed resolution:

This wording is relative to N3376.

  1. Change 25.2.6 [alg.find.end] p2 as indicated:

    template<class ForwardIterator1, class ForwardIterator2>
    ForwardIterator1 
    find_end(ForwardIterator1 first1, ForwardIterator1 last1,
             ForwardIterator2 first2, ForwardIterator2 last2);
    template<class ForwardIterator1, class ForwardIterator2,
             class BinaryPredicate>
    ForwardIterator1
    find_end(ForwardIterator1 first1, ForwardIterator1 last1,
             ForwardIterator2 first2, ForwardIterator2 last2,
             BinaryPredicate pred);
    

    […]

    -2- Returns: The last iterator i in the range [first1,last1 - (last2 - first2)) such that for anyevery nonnegative integer n < (last2 - first2), the following corresponding conditions hold: *(i + n) == *(first2 + n), pred(*(i + n), *(first2 + n)) != false. Returns last1 if [first2,last2) is empty or if no such iterator is found.

  2. Change 25.2.13 [alg.search] p2 and p6 as indicated:

    template<class ForwardIterator1, class ForwardIterator2>
    ForwardIterator1
    search(ForwardIterator1 first1, ForwardIterator1 last1,
           ForwardIterator2 first2, ForwardIterator2 last2);
    template<class ForwardIterator1, class ForwardIterator2,
             class BinaryPredicate>
    ForwardIterator1
    search(ForwardIterator1 first1, ForwardIterator1 last1,
           ForwardIterator2 first2, ForwardIterator2 last2,
           BinaryPredicate pred);
    

    […]

    -2- Returns: The first iterator i in the range [first1,last1 - (last2-first2)) such that for anyevery nonnegative integer n less than last2 - first2 the following corresponding conditions hold: *(i + n) == *(first2 + n), pred(*(i + n), *(first2 + n)) != false. Returns first1 if [first2,last2) is empty, otherwise returns last1 if no such iterator is found.

    […]

    template<class ForwardIterator, class Size, class T>
    ForwardIterator
    search_n(ForwardIterator first, ForwardIterator last, Size count,
             const T& value);
    template<class ForwardIterator, class Size, class T,
             class BinaryPredicate>
    ForwardIterator
    search_n(ForwardIterator first, ForwardIterator last, Size count,
             const T& value, BinaryPredicate pred);
    

    […]

    -6- Returns: The first iterator i in the range [first,last-count) such that for anyevery non-negative integer n less than count the following corresponding conditions hold: *(i + n) == value, pred(*(i + n),value) != false. Returns last if no such iterator is found.

  3. Change 25.3.10 [alg.reverse] p4 as indicated:

    template<class BidirectionalIterator, class OutputIterator>
    OutputIterator
    reverse_copy(BidirectionalIterator first,
                 BidirectionalIterator last, OutputIterator result);
    

    […]

    -4- Effects: Copies the range [first,last) to the range [result,result+(last-first)) such that for anyevery non-negative integer i < (last - first) the following assignment takes place: *(result + (last - first) - i) = *(first + i).

  4. Change 25.3.13 [alg.partitions] p5 and p9 as indicated:

    template<class ForwardIterator, class Predicate>
    ForwardIterator
    partition(ForwardIterator first,
              ForwardIterator last, Predicate pred);
    

    […]

    -5- Returns: An iterator i such that for anyevery iterator j in the range [first,i) pred(*j) != false, and for anyevery iterator k in the range [i,last), pred(*k) == false.

    […]

    template<class BidirectionalIterator, class Predicate>
    BidirectionalIterator
    stable_partition(BidirectionalIterator first,
                     BidirectionalIterator last, Predicate pred);
    

    […]

    -9- Returns: An iterator i such that for anyevery iterator j in the range [first,i), pred(*j) != false, and for anyevery iterator k in the range [i,last), pred(*k) == false. The relative order of the elements in both groups is preserved.

  5. Change 25.4 [alg.sorting] p5 as indicated:

    -5- A sequence is sorted with respect to a comparator comp if for anyevery iterator i pointing to the sequence and anyevery non-negative integer n such that i + n is a valid iterator pointing to an element of the sequence, comp(*(i + n), *i) == false.

  6. Change 25.4.2 [alg.nth.element] p1 as indicated:

    template<class RandomAccessIterator>
    void nth_element(RandomAccessIterator first, RandomAccessIterator nth,
                     RandomAccessIterator last);
    template<class RandomAccessIterator, class Compare>
    void nth_element(RandomAccessIterator first, RandomAccessIterator nth,
                     RandomAccessIterator last, Compare comp);
    

    -1- After nth_element the element in the position pointed to by nth is the element that would be in that position if the whole range were sorted. Also for anyevery iterator i in the range [first,nth) and anyevery iterator j in the range [nth,last) it holds that: !(*i > *j) or comp(*j, *i) == false.

  7. Change 25.4.3.1 [lower.bound] p2 as indicated:

    template<lass ForwardIterator, class T>
    ForwardIterator
    lower_bound(ForwardIterator first, ForwardIterator last,
                const T& value);
    template<class ForwardIterator, class T, class Compare>
    ForwardIterator
    lower_bound(ForwardIterator first, ForwardIterator last,
                const T& value, Compare comp);
    

    […]

    -2- Returns: The furthermost iterator i in the range [first,last] such that for anyevery iterator j in the range [first,i) the following corresponding conditions hold: *j < value or comp(*j, value) != false.

  8. Change 25.4.3.2 [upper.bound] p2 as indicated:

    template<lass ForwardIterator, class T>
    ForwardIterator
    upper_bound(ForwardIterator first, ForwardIterator last,
                const T& value);
    template<class ForwardIterator, class T, class Compare>
    ForwardIterator
    upper_bound(ForwardIterator first, ForwardIterator last,
                const T& value, Compare comp);
    

    […]

    -2- Returns: The furthermost iterator i in the range [first,last] such that for anyevery iterator j in the range [first,i) the following corresponding conditions hold: !(value < *j) or comp(value, *j) == false.

  9. Change 25.4.7 [alg.min.max] p21 and p23 as indicated:

    template<class ForwardIterator>
    ForwardIterator min_element(ForwardIterator first, ForwardIterator last);
    template<class ForwardIterator, class Compare>
    ForwardIterator min_element(ForwardIterator first, ForwardIterator last,
                                Compare comp);
    

    -21- Returns: The first iterator i in the range [first,last) such that for anyevery iterator j in the range [first,last) the following corresponding conditions hold: !(*j < *i) or comp(*j, *i) == false. Returns last if first == last.

    […]

    template<class ForwardIterator>
    ForwardIterator max_element(ForwardIterator first, ForwardIterator last);
    template<class ForwardIterator, class Compare>
    ForwardIterator max_element(ForwardIterator first, ForwardIterator last,
                                Compare comp);
    

    -23- Returns: The first iterator i in the range [first,last) such that for anyevery iterator j in the range [first,last) the following corresponding conditions hold: !(*i < *j) or comp(*i, *j) == false. Returns last if first == last.


2151. basic_string<>::swap semantics ignore allocators

Section: 21.4.1 [string.require] Status: New Submitter: Robert Shearer Opened: 2012-04-13 Last modified: 2012-09-24

View other active issues in [string.require].

View all other issues in [string.require].

View all issues with New status.

Discussion:

In C++11, basic_string is not described as a "container", and is not governed by the allocator-aware container semantics described in sub-clause 23.2 [container.requirements]; as a result, and requirements or contracts for the basic_string interface must be documented in Clause 21 [strings].

Sub-clause 21.4.6.8 [string::swap] defines the swap member function with no requirements, and with guarantees to execute in constant time without throwing. Fulfilling such a contract is not reasonable in the presence of unequal non-propagating allocators.

In contrast, 23.2.1 [container.requirements.general] p7 declares the behavior of member swap for containers with unequal non-propagating allocators to be undefined.

Resolution proposal:

Additional language from Clause 23 [containers] should probably be copied to Clause 21 [strings]. I will refrain from an exactly recommendation, however, as I am raising further issues related to the language in Clause 23 [containers].

Proposed resolution:


2152. Instances of standard container types are not swappable

Section: 17.6.3.2 [swappable.requirements], 23.2.1 [container.requirements.general] Status: New Submitter: Robert Shearer Opened: 2012-04-13 Last modified: 2012-09-24

View other active issues in [swappable.requirements].

View all other issues in [swappable.requirements].

View all issues with New status.

Discussion:

Sub-clause 17.6.3.2 [swappable.requirements] defines two notions of swappability: a binary version defining when two objects are swappable with one another, and a unary notion defining whether an object is swappable (without qualification), with the latter definition requiring that the object satisfy the former with respect to all values of the same type.

Let T be a container type based on a non-propagating allocator whose instances do not necessarily compare equal. Then sub-clause 23.2.1 [container.requirements.general] p7 implies that no object t of type T is swappable (by the unary definition).

Throughout the standard it is the unary definition of "swappable" that is listed as a requirement (with the exceptions of 20.2.2 [utility.swap] p4, 20.3.2 [pairs.pair] p31, 20.4.2.3 [tuple.swap] p2, 25.3.3 [alg.swap] p2, and 25.3.3 [alg.swap] p6, which use the binary definition). This renders many of the mutating sequence algorithms of sub-clause 25.3 [alg.modifying.operations], for example, inapplicable to sequences of standard container types, even where every element of the sequence is swappable with every other.

Note that this concern extends beyond standard containers to all future allocator-based types.

Resolution proposal:

I see two distinct straightforward solutions:

  1. Modify the requirements of algorithms from sub-clause 25.3 [alg.modifying.operations], and all other places that reference the unary "swappable" definition, to instead use the binary "swappable with" definition (over a domain appropriate to the context). The unary definition of "swappable" could then be removed from the standard.
  2. Modify sub-clause 23.2.1 [container.requirements.general] such that objects of standard container types are "swappable" by the unary definition.

I favor the latter solution, for reasons detailed in the following issue.

Proposed resolution:


2153. Narrowing of the non-member swap contract

Section: 20.2.2 [utility.swap], 17.6.3.2 [swappable.requirements], 23.2.1 [container.requirements.general] Status: New Submitter: Robert Shearer Opened: 2012-04-13 Last modified: 2012-09-24

View all issues with New status.

Discussion:

Sub-clause 20.2.2 [utility.swap] defines a non-member 'swap' function with defined behavior for all MoveConstructible and MoveAssignable types. It does not guarantee constant-time complexity or noexcept in general, however this definition does render all objects of MoveConstructible and MoveAssignable type swappable (by the unary definition of sub-clause 17.6.3.2 [swappable.requirements]) in the absence of specializations or overloads.

The overload of the non-member swap function defined in Table 96, however, defines semantics incompatible with the generic non-member swap function, since it is defined to call a member swap function whose semantics are undefined for some values of MoveConstructible and MoveAssignable types.

The obvious (perhaps naive) interpretation of sub-clause 17.6.3.2 [swappable.requirements] is as a guide to the "right" semantics to provide for a non-member swap function (called in the context defined by 17.6.3.2 [swappable.requirements] p3) in order to provide interoperable user-defined types for generic programming. The standard container types don't follow these guidelines.

More generally, the design in the standard represents a classic example of "contract narrowing". It is entirely reasonable for the contract of a particular swap overload to provide more guarantees, such as constant-time execution and noexcept, than are provided by the swap that is provided for any MoveConstructible and MoveAssignable types, but it is not reasonable for such an overload to fail to live up to the guarantees it provides for general types when it is applied to more specific types. Such an overload or specialization in generic programming is akin to an override of an inherited virtual function in OO programming: violating a superclass contract in a subclass may be legal from the point of view of the language, but it is poor design and can easily lead to errors. While we cannot prevent user code from providing overloads that violate the more general swap contract, we can avoid doing so within the library itself.

My proposed resolution is to draw a sharp distinction between member swap functions, which provide optimal performance but idiosyncratic contracts, and non-member swap functions, which should always fulfill at least the contract of 20.2.2 [utility.swap] and thus render objects swappable. The member swap for containers with non-propagating allocators, for example, would offer constant-time guarantees and noexcept but would only offer defined behavior for values with allocators that compare equal; non-member swap would test allocator equality and then dispatch to either member swap or std::swap depending on the result, providing defined behavior for all values (and rendering the type "swappable"), but offering neither the constant-time nor the noexcept guarantees.

Proposed resolution:


2154. What exactly does compile-time complexity imply?

Section: 26.5.1.3 [rand.req.urng] Status: New Submitter: John Salmon Opened: 2012-04-26 Last modified: 2012-09-24

View all issues with New status.

Discussion:

The expressions G::min() and G::max() in Table 116 in 26.5.1.3 [rand.req.urng] are specified as having "compile-time" complexity.

It is not clear what, exactly, this requirement implies. If a URNG has a method:

static int min();

then is the method required to have a constexpr qualifier? I believe the standard would benefit from clarification of this point.

Proposed resolution:


2155. Macro __bool_true_false_are_defined should be removed

Section: 18.10 [support.runtime] Status: New Submitter: Thomas Plum Opened: 2012-04-30 Last modified: 2012-09-24

View other active issues in [support.runtime].

View all other issues in [support.runtime].

View all issues with New status.

Discussion:

Since C99, the C standard describes a macro named __bool_true_false_are_defined.

In the process of harmonizing C++11 with C99, this name became part of the C++ standard.

I propose that all mention of this name should be removed from the C and C++ standards.

Here's the problem: The name was originally proposed as a transition tool, so that the headers for a project could contain lines like the following.

#if !defined(__bool_true_false_are_defined)
#define bool int /* or whatever */
#define true 1
#define false 0
#endif

Then when the project was compiled by a "new" compiler that implemented bool as defined by the evolving C++98 or C99 standards, those lines would be skipped; but when compiled by an "old" compiler that didn't yet provide bool, true, and false, then the #define's would provide a simulation that worked for most purposes.

It turns out that there is an unfortunate ambiguity in the name. One interpretation is as shown above, but a different reading says "bool, true, and false are #define'd", i.e. that the meaning of the macro is to assert that these names are macros (not built-in) ... which is true in C, but not in C++.

In C++11, the name appears in parentheses followed by a stray period, so some editorial change is needed in any event:

18.10 [support.runtime] para 1:

Headers <csetjmp> (nonlocal jumps), <csignal> (signal handling), <cstdalign> (alignment), <cstdarg> (variable arguments), <cstdbool> (__bool_true_false_are_defined). <cstdlib> (runtime environment getenv(), system()), and <ctime> (system clock clock(), time()) provide further compatibility with C code.

However, para 2 says

"The contents of these headers are the same as the Standard C library headers <setjmp.h>, <signal.h>, <stdalign.h>, <stdarg.h>, <stdbool.h>, <stdlib.h>, and <time.h>, respectively, with the following changes:",

and para 8 says

"The header <cstdbool> and the header <stdbool.h> shall not define macros named bool, true, or false."

Thus para 8 doesn't exempt the C++ implementation from the arguably clear requirement of the C standard, to provide a macro named __bool_true_false_are_defined defined to be 1.

Real implementations of the C++ library differ, so the user cannot count upon any consistency; furthermore, the usefulness of the transition tool has faded long ago.

That's why my suggestion is that both C and C++ standards should eliminate any mention of __bool_true_false_are_defined. In that case, the name belongs to implementers to provide, or not, as they choose.

Proposed resolution:


2156. Unordered containers' reserve(n) reserves for n-1 elements

Section: 23.2.5 [unord.req] Status: New Submitter: Daniel James Opened: 2012-05-07 Last modified: 2012-09-24

View other active issues in [unord.req].

View all other issues in [unord.req].

View all issues with New status.

Discussion:

I think that unordered containers' reserve doesn't quite do what it should. I'd expect after calling x.reserve(n) to be able to insert n elements without invalidating iterators. But as the standard is written (I'm looking at n3376), I think the guarantee only holds for n-1 elements.

For a container with max_load_factor of 1, reserve(n) is equivalent to rehash(ceil(n/1)), ie. rehash(n). rehash(n) requires that the bucket count is >= n, so it can be n (Table 103). The rule is that insert shall not affect the validity of iterators if (N + n) < z * B (23.2.5 [unord.req] p15). But for this case the two sides of the equation are equal, so insert can affect the validity of iterators.

Proposed resolution:


2157. How does std::array<T,0> initialization work when T is not default-constructible?

Section: 23.3.2.8 [array.zero] Status: New Submitter: Daryle Walker Opened: 2012-05-08 Last modified: 2012-09-24

View all other issues in [array.zero].

View all issues with New status.

Discussion:

Objects of std::array<T,N> are supposed to be initialized with aggregate initialization (when not the destination of a copy or move). This clearly works when N is positive. What happens when N is zero? To continue using an (inner) set of braces for initialization, a std::array<T,0> implementation must have an array member of at least one element, and let default initialization take care of those secret elements. This cannot work when T has a set of constructors and the default constructor is deleted from that set. Solution: Add a new paragraph in 23.3.2.8 [array.zero]:

The unspecified internal structure of array for this case shall allow initializations like:

array<T, 0> a = { };

and said initializations must be valid even when T is not default-constructible.

Proposed resolution:

This wording is relative to N3376.

Add the following new paragraph between the current 23.3.2.8 [array.zero] p1 and p2:

-1- array shall provide support for the special case N == 0.

-?- The unspecified internal structure of array for this case shall allow initializations like:

array<T, 0> a = { };

and said initializations must be valid even when T is not default-constructible.

-2- In the case that N == 0, begin() == end() == unique value. The return value of data() is unspecified.

-3- The effect of calling front() or back() for a zero-sized array is undefined.

-4- Member function swap() shall have a noexcept-specification which is equivalent to noexcept(true).


2158. Conditional copy/move in std::vector

Section: 23.3.6.3 [vector.capacity] Status: New Submitter: Nikolay Ivchenkov Opened: 2012-05-08 Last modified: 2012-09-24

View other active issues in [vector.capacity].

View all other issues in [vector.capacity].

View all issues with New status.

Discussion:

There are various operations on std::vector that can cause elements of the vector to be moved from one location to another. A move operation can use either rvalue or const lvalue as argument; the choice depends on the value of !is_nothrow_move_constructible<T>::value && is_copy_constructible<T>::value, where T is the element type. Thus, some operations on std::vector (e.g. 'resize' with single parameter, 'reserve', 'emplace_back') should have conditional requirements. For example, let's consider the requirement for 'reserve' in N3376 – 23.3.6.3 [vector.capacity]/2:

Requires: T shall be MoveInsertable into *this.

This requirement is not sufficient if an implementation is free to select copy constructor when !is_nothrow_move_constructible<T>::value && is_copy_constructible<T>::value evaluates to true. Unfortunately, is_copy_constructible cannot reliably determine whether T is really copy-constructible. A class may contain public non-deleted copy constructor whose definition does not exist or cannot be instantiated successfully (e.g., std::vector<std::unique_ptr<int>> has copy constructor, but this type is not copy-constructible). Thus, the actual requirements should be:

Maybe it would be useful to introduce a new name for such conditional requirement (in addition to "CopyInsertable" and "MoveInsertable").

Proposed resolution:


2159. atomic_flag initialization

Section: 29.7 [atomics.flag] Status: New Submitter: Alberto Ganesh Barbati Opened: 2012-05-24 Last modified: 2012-09-24

View other active issues in [atomics.flag].

View all other issues in [atomics.flag].

View all issues with New status.

Discussion:

29.7 [atomics.flag]/4 describes the ATOMIC_FLAG_INIT, but it's not quite clear about a couple of points:

  1. it's said that ATOMIC_FLAG_INIT "can be used to initialize an object of type atomic_flag" and the following example:

    std::atomic_flag guard = ATOMIC_FLAG_INIT;
    

    is presented. It's not clear whether the macro can also be used in the other initialization contexts:

    std::atomic_flag guard ATOMIC_FLAG_INIT; 
    std::atomic_flag guard {ATOMIC_FLAG_INIT};
    
    struct A { std::atomic_flag flag; A(); };
    A::A() : flag (ATOMIC_FLAG_INIT); 
    A::A() : flag {ATOMIC_FLAG_INIT};
    

    Please also note that examples are non-normative, according to the ISO directives, meaning that the wording presents no normative way to use the macro.

  2. it's said that "It is unspecified whether an uninitialized atomic_flag object has an initial state of set or clear.". I believe the use of "uninitialized" is inappropriate. First of all, if an object is uninitialized it is obvious that we cannot assert anything about its state. Secondly, it doesn't address the following cases:

    std::atomic_flag a; // object is "initialized" by trivial default constructor
    std::atomic_flag a {}; // object is value-initialized
    static std::atomic_flag a; // object is zero-initialized
    

    strictly speaking a trivial constructor "initializes" the object, although it doesn't actually initialize the sub-objects.

  3. it's said that "For a static-duration object, that initialization shall be static.". Considering the following example:

    struct A
    {
      A(); // user-provided, not constexpr
    
      std::atomic_flag flag = ATOMIC_FLAG_INIT;
      // possibly other non-static data members
    };
    
    static A a;
    

    The object a.flag (as a sub-object of the object a) has static-duration, yet the initialization has to be dynamic because A::A is not constexpr.

Proposed resolution:

This wording is relative to N3376.

Change 29.7 [atomics.flag]/4 as follows:

The macro 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. The macro can be used in the form:

atomic_flag guard = ATOMIC_FLAG_INIT;

It's unspecified whether the macro can be used in other initialization contexts. For a complete static-duration object, that initialization shall be static. It is unspecified whether an uninitialized Unless initialized with ATOMIC_FLAG_INIT, it is unspecified whether an atomic_flag object has an initial state of set or clear. [ Example:

atomic_flag guard = ATOMIC_FLAG_INIT;

end example ]


2160. Unintended destruction ordering-specification of resize

Section: 23.3.6.3 [vector.capacity] Status: New Submitter: Daniel Krügler Opened: 2012-06-07 Last modified: 2012-09-24

View other active issues in [vector.capacity].

View all other issues in [vector.capacity].

View all issues with New status.

Discussion:

As part of resolving LWG issue 2033 a wording change was done for resize() to respect the problem mentioned in the question:

Does a call to 'void resize(size_type sz)' of std::vector require the element type to be MoveAssignable because the call erase(begin() + sz, end()) mentioned in the Effects paragraph would require the element type to be MoveAssignable?

The wording change was to replace in 23.3.3.3 [deque.capacity] and 23.3.6.3 [vector.capacity]:

-1- Effects: If sz <= size(), equivalent to erase(begin() + sz, end()); […]

by:

-1- Effects: If sz <= size(), equivalent to calling pop_back() size() - sz times. […]

The overlooked side-effect of this wording change is that this implies a destruction order of the removed elements to be in reverse order of construction, but the previous version did not impose any specific destruction order due to the way how the semantics of erase is specified in Table 100.

Given the program:

#include <vector>
#include <iostream>

struct Probe {
  int value;
  Probe() : value(0) {}
  Probe(int value) : value(value) {}
  ~Probe() { std::cout << "~Probe() of " << value << std::endl; }
};

int main() {
  std::vector<Probe> v;
  v.push_back(Probe(1));
  v.push_back(Probe(2));
  v.push_back(Probe(3));
  std::cout << "---" << std::endl;
  v.resize(0);
}

the last three lines of the output for every compiler I tested was:

~Probe() of 1
~Probe() of 2
~Probe() of 3

but a conforming implementation would now need to change this order to

~Probe() of 3
~Probe() of 2
~Probe() of 1

This possible stringent interpretation makes sense, because one can argue that sequence containers (or at least std::vector) should have the same required destruction order of it's elements, as elements of a C array or controlled by memory deallocated with an array delete have. I also learned that libc++ does indeed implement std::vector::resize in a way that the second output form is observed.

While I agree that required reverse-destruction would better mimic the natural behaviour of std::vector this was not required in C++03 and this request may be too strong. My current suggestion would be to restore the effects of the previous wording in regard to the destruction order, because otherwise several currently existing implementations would be broken just because of this additional requirement.

Proposed resolution:


2161. const equivalence of std::map

Section: 23.4 [associative], 23.5 [unord] Status: New Submitter: Bjarne Stroustrup Opened: 2012-06-18 Last modified: 2012-09-24

View other active issues in [associative].

View all other issues in [associative].

View all issues with New 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>;

Proposed resolution:


2162. allocator_traits::max_size missing noexcept

Section: 17.6.3.5 [allocator.requirements], 20.6.8.2 [allocator.traits.members], 20.6.8 [allocator.traits] Status: New Submitter: Bo Persson Opened: 2012-07-03 Last modified: 2012-09-24

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with New status.

Discussion:

N3376 describes in 20.6.8.2 [allocator.traits.members]/7

static size_type max_size(Alloc& a);

Returns: a.max_size() if that expression is well-formed; otherwise, numeric_limits<size_type>::max().

The max_size function is supposed to call one of two functions that are both noexcept. To make this intermediate function useful for containers, it should preserve the noexcept attribute.

Proposed changes:

In 20.6.8 [allocator.traits] and 20.6.8.2 [allocator.traits.members]/7, change the function signature to

static size_type max_size(Alloc& a) noexcept;

[2012-08-05 Daniel comments]

On the first sight this does not seem like a defect of the specification, because the Allocator requirements in 17.6.3.5 [allocator.requirements] (Table 28) do not impose a no-throw requirement onto max_size(); the table just describes the fall-back implementation for max_size() if a given allocator does not provide such a function.

std::allocator as a special model of this concept and is allowed to increase the exception-guarantees for max_size(), but this does not imply a corresponding rules for other allocators.

Furthermore, max_size() of Containers is not specified in terms of Allocator::max_size(), so again this is not a real contradiction.

Nonetheless I think that the following stronger decision should be considered:

  1. Require that for all Allocators (as specified in 17.6.3.5 [allocator.requirements]) max_size() never throws an exception. This would it make much more useful to call this function in situations where no exception should leave the context.

  2. Require that for all Allocators (as specified in 17.6.3.5 [allocator.requirements]) max_size() can be called on const allocator object. Together with the previous item this would allow an implementation of a container's max_size() function to delegate to the allocator's max_size() function.

In regard to the second statement it should be mentioned that there are two current specification deviations from that in the draft:

  1. The synopsis of 20.6.8 [allocator.traits] uses a const allocator argument as part of the signature of the max_size function.

  2. Both the synopsis of 20.12.1 [allocator.adaptor.syn] and the member specification in 20.12.4 [allocator.adaptor.members] p8 declare scoped_allocator_adaptor::max_size as const member function, but this function delegates to

    allocator_traits<OuterAlloc>::max_size(outer_allocator())
    

    where outer_allocator() resolves to the member function overload returning a const outer_allocator_type&.

The question arises whether these current defects actually point to a defect in the Allocator requirements and should be fixed there.

Proposed resolution:


2163. nth_element requires inconsistent post-conditions

Section: 25.4.2 [alg.nth.element] Status: New Submitter: Peter Sommerlad Opened: 2012-07-06 Last modified: 2012-09-24

View all issues with New status.

Discussion:

The specification of nth_element refers to operator< whereas all sorting without a compare function is based on operator<. While it is true that for all regular types both operators should be defined accordingly, all other sorting algorithms only rely on existence of operator<. So I guess the paragraph p1

After nth_element the element in the position pointed to by nth is the element that would be in that position if the whole range were sorted. Also for any iterator i in the range [first,nth) and any iterator j in the range [nth,last) it holds that: !(*i > *j) or comp(*j, *i) == false.

should read

After nth_element the element in the position pointed to by nth is the element that would be in that position if the whole range were sorted. Also for any iterator i in the range [first,nth) and any iterator j in the range [nth,last) it holds that: !(*j < *i) or comp(*j, *i) == false.

Note only "!(*i > *j)" was changed to "!(*j < *i)" and it would be more symmetric with comp(*j, *i) as well.

In theory this might be a semantic change to the spec, but I believe the mistake is unintended.

Proposed resolution:

This wording is relative to N3376.

  1. Change 25.4.2 [alg.nth.element] p1 as indicated:

    template<class RandomAccessIterator>
    void nth_element(RandomAccessIterator first, RandomAccessIterator nth,
                     RandomAccessIterator last);
    template<class RandomAccessIterator, class Compare>
    void nth_element(RandomAccessIterator first, RandomAccessIterator nth,
                     RandomAccessIterator last, Compare comp);
    

    -1- After nth_element the element in the position pointed to by nth is the element that would be in that position if the whole range were sorted. Also for any iterator i in the range [first,nth) and any iterator j in the range [nth,last) it holds that: !(*i > *j)!(*j < *i) or comp(*j, *i) == false.


2164. What are the semantics of vector.emplace(vector.begin(), vector.back())?

Section: 23.3.6.5 [vector.modifiers], 23.2 [container.requirements] Status: New Submitter: Howard Hinnant Opened: 2012-07-07 Last modified: 2012-09-24

View all other issues in [vector.modifiers].

View all issues with New status.

Discussion:

Nikolay Ivchenkov recently brought the following example on the std-discussion newsgroup, asking whether the following program well-defined:

#include <iostream>
#include <vector>

int main()
{
  std::vector<int> v;
  v.reserve(4);
  v = { 1, 2, 3 };
  v.emplace(v.begin(), v.back());
  for (int x : v)
    std::cout << x << std::endl;
}

Nikolay Ivchenkov:

I think that an implementation of vector's 'emplace' should initialize an intermediate object with v.back() before any shifts take place, then perform all necessary shifts and finally replace the value pointed to by v.begin() with the value of the intermediate object. So, I would expect the following output:

3
1
2
3

GNU C++ 4.7.1 and GNU C++ 4.8.0 produce other results:

2
1
2
3

Howard Hinnant:

I believe Nikolay is correct that vector should initialize an intermediate object with v.back() before any shifts take place. As Nikolay pointed out in another email, this appears to be the only way to satisfy the strong exception guarantee when an exception is not thrown by T's copy constructor, move constructor, copy assignment operator, or move assignment operator as specified by 23.3.6.5 [vector.modifiers]/p1. I.e. if the emplace construction throws, the vector must remain unaltered.

That leads to an implementation that tolerates objects bound to the function parameter pack of the emplace member function may be elements or sub-objects of elements of the container.

My position is that the standard is correct as written, but needs a clarification in this area. Self-referencing emplace should be legal and give the result Nikolay expects. The proposed resolution of LWG 760 is not correct.

Proposed resolution:


2165. std::atomic<X> requires X to be nothrow default constructible

Section: 29.5 [atomics.types.generic], 29.6 [atomics.types.operations] Status: New Submitter: Jonathan Wakely Opened: 2012-07-19 Last modified: 2012-09-24

View all other issues in [atomics.types.generic].

View all issues with New status.

Discussion:

As raised in c++std-lib-32781, this fails to compile even though the default constructor is not used:

#include <atomic>

struct X {
  X() noexcept(false) {}
  X(int) { }
};

std::atomic<X> x(3);

This is because atomic<T>'s default constructor is declared to be non-throwing and is explicitly-defaulted on its first declaration:

atomic() noexcept = default;

This is ill-formed if the implicitly-declared default constructor would not be non-throwing.

Possible solutions:

  1. Add nothrow default constructible to requirements for template argument of the generic atomic<T>
  2. Remove atomic<T>::atomic() from the overload set if T is not nothrow default constructible.
  3. Remove noexcept from atomic<T>::atomic(), allowing it to be deduced (but the default constructor is intended to be always noexcept)
  4. Do not default atomic<T>::atomic() on its first declaration (but makes the default constructor user-provided and so prevents atomic<T> being trivial)
  5. A core change to allow the mismatched exception specification if the default constructor isn't used (see c++std-core-21990)

Proposed resolution:


2166. Heap property underspecified?

Section: 25.4.6 [alg.heap.operations] Status: New Submitter: Peter Sommerlad Opened: 2012-07-09 Last modified: 2012-09-24

View all other issues in [alg.heap.operations].

View all issues with New status.

Discussion:

Another similar issue to the operator< vs greater in nth_element but not as direct occurs in 25.4.6 [alg.heap.operations]:

-1- A heap is a particular organization of elements in a range between two random access iterators [a,b). Its two key properties are:

  1. There is no element greater than *a in the range and
  2. *a may be removed by pop_heap(), or a new element added by push_heap(), in O(log(N)) time.

As noted by Richard Smith, it seems that the first bullet should read:

*a is not less than any element in the range

Even better the heap condition could be stated here directly, instead of leaving it unspecified, i.e.,

Each element at (a+2*i+1) and (a+2*i+2) is less than the element at (a+i), if those elements exist, for i>=0.

But may be that was may be intentional to allow other heap organizations?

See also follow-up discussion of c++std-lib-32780.

Proposed resolution:


2167. Copy assignment requirements of Containers

Section: 23.2.1 [container.requirements.general] Status: New Submitter: Dean Michael Berris Opened: 2012-07-13 Last modified: 2012-09-24

View other active issues in [container.requirements.general].

View all other issues in [container.requirements.general].

View all issues with New 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.

Proposed resolution:

This wording is relative to N3376.

  1. Change Table 96 — "Container requirements" in 23.2.1 [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

2168. Inconsistent specification of uniform_real_distribution constructor

Section: 26.5.8.2.2 [rand.dist.uni.real] Status: New Submitter: Marshall Clow Opened: 2012-07-14 Last modified: 2012-09-24

View all issues with New status.

Discussion:

uniform_real says in 26.5.8.2.2 [rand.dist.uni.real] p1:

A uniform_real_distribution random number distribution produces random numbers x, a ≤ x < b,

but also that (26.5.8.2.2 [rand.dist.uni.real] p2):

explicit uniform_real_distribution(RealType a = 0.0, RealType b = 1.0);

-2- Requires: a ≤ b and b - a ≤ numeric_limits<RealType>::max().

If you construct a uniform_real_distribution<RealType>(a, b) where there are no representable numbers between 'a' and 'b' (using RealType's representation) then you cannot satisfy 26.5.8.2.2 [rand.dist.uni.real].

An obvious example is when a == b.

Proposed resolution:


2169. Missing reset() requirements in unique_ptr specialization

Section: 20.7.1.3.3 [unique.ptr.runtime.modifiers] Status: New Submitter: Geoffrey Romer Opened: 2012-07-16 Last modified: 2012-09-24

View all other issues in [unique.ptr.runtime.modifiers].

View all issues with New status.

Discussion:

In 20.7.1.3.3 [unique.ptr.runtime.modifiers]/p1-2 of N3376, the description of reset() in the array specialization of unique_ptr partially duplicates the description of the base template method (as specified in 20.7.1.2.5 [unique.ptr.single.modifiers]/p3-5), but lacks some significant requirements. Specifically, the text introduced in LWG 998, and item 13 of LWG 762, is present only in the base template, not the specialization.

This gives the appearance that these requirements specifically do not apply to the specialization, which I don't believe is correct or intended: the issue of reset() operation order addressed by LWG 998 applies just as much to the derived template as to the base template, and the derived template has just as much need to rely on get_deleter()(get()) being well-defined, well-formed, and not throwing exceptions (arguably some of those properties follow from the fact that T is required to be a complete type, but not all).

Assuming the derived template's reset() semantics are intended to be identical to the base template's, there is no need to explicitly specify the semantics of reset(pointer p) at all (since 20.7.1.3 [unique.ptr.runtime]/3 specifies "Descriptions are provided below only for member functions that have behavior different from the primary template."), and reset(nullptr_t p) can be specified by reference to the 'pointer' overload. This is more concise, and eliminates any ambiguity about intentional vs. accidental discrepancies.

Proposed resolution:

This wording is relative to N3376.

Change 20.7.1.3.3 [unique.ptr.runtime.modifiers] as indicated:

void reset(pointer p = pointer()) noexcept;
void reset(nullptr_t p) noexcept;

-1- Effects: If get() == nullptr there are no effects. Otherwise get_deleter()(get()) Equivalent to reset(pointer()).

-2- Postcondition: get() == p.


2170. Aggregates cannot be DefaultConstructible

Section: 17.6.3.1 [utility.arg.requirements] Status: New Submitter: Daniel Krügler Opened: 2012-07-19 Last modified: 2012-09-24

View other active issues in [utility.arg.requirements].

View all other issues in [utility.arg.requirements].

View all issues with New status.

Discussion:

The lack of the definition of the DefaultConstructible requirements in C++03 was fixed by LWG 724 at a time where the core rules of list-initialization were slightly different than today, at that time value-initialization (shortly) was the primary rule for class types, i.e. just before applying CWG 1301, CWG 1324, and CWG 1368.

The order in 8.5.4 [dcl.init.list] p3 was changed to respect aggregate initialization, but that had the side-effect that formally aggregate types cannot satisfy the DefaultConstructible requirements anymore, because we require that

T u{};

value-initializes the object u.

Of-course exclusion of aggregates was not intended, therefore I suggest to extend the requirements in Table 19 (17.6.3.1 [utility.arg.requirements]) for empty aggregate-initialization as well.

Proposed resolution:

This wording is relative to N3376.

Change Table 19 in 17.6.3.1 [utility.arg.requirements] as indicated:

Table 19 — DefaultConstructible requirements [defaultconstructible]
Expression Post-condition
T t; object t is default-initialized
T u{}; object u is value-initialized or aggregate-initialized
T()
T{}
a temporary object of type T is value-initialized or aggregate-initialized

2171. "swappable" undefined for swapping lvalue and rvalue

Section: 17.6.3.2 [swappable.requirements] Status: New Submitter: Robert Shearer Opened: 2012-07-24 Last modified: 2012-09-24

View other active issues in [swappable.requirements].

View all other issues in [swappable.requirements].

View all issues with New status.

Discussion:

Paragraph 17.6.3.2 [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".

Proposed resolution:

This wording is relative to N3376.

Change 17.6.3.2 [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.


2172. Does atomic_compare_exchange_* accept v == nullptr arguments?

Section: 20.7.2.5 [util.smartptr.shared.atomic] Status: New Submitter: Howard Hinnant Opened: 2012-07-28 Last modified: 2012-09-24

View all other issues in [util.smartptr.shared.atomic].

View all issues with New status.

Discussion:

Looking at 20.7.2.5 [util.smartptr.shared.atomic]/p31

template<class T>
  bool
  atomic_compare_exchange_strong_explicit(shared_ptr<T>* p,
                                          shared_ptr<T>* v,
                                          shared_ptr<T> w,
                                          memory_order success, memory_order failure);

-31- Requires: p shall not be null.

What about v? Can it be null? And if so, what happens?

This looks closely related to C++11 issue LWG 1030, where we gave every signature in this section a:

Requires: p shall not be null.

It looks like a simple oversight to me that we did not add for the atomic_compare_exchange_*:

Requires: p shall not be null and v shall not be null.

Proposed resolution:

This wording is relative to N3376.

Change 20.7.2.5 [util.smartptr.shared.atomic] as indicated:

template<class T>
  bool atomic_compare_exchange_weak(
    shared_ptr<T>* p, shared_ptr<T>* v, shared_ptr<T> w);

-27- Requires: p shall not be null and v shall not be null.

[…]

template<class T>
  bool atomic_compare_exchange_weak_explicit(
    shared_ptr<T>* p, shared_ptr<T>* v, shared_ptr<T> w,
    memory_order success, memory_order failure);
template<class T>
  bool atomic_compare_exchange_strong_explicit(
    shared_ptr<T>* p, shared_ptr<T>* v, shared_ptr<T> w,
    memory_order success, memory_order failure);

-31- Requires: p shall not be null and v shall not be null.

[…]


2173. The meaning of operator + in the description of the algorithms

Section: 25 [algorithms] Status: New Submitter: Nikolay Ivchenkov Opened: 2012-08-01 Last modified: 2012-09-24

View all other issues in [algorithms].

View all issues with New status.

Discussion:

According to 25.1 [algorithms.general]/12,

In the description of the algorithms operators + 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;

There are several places where such operator + is applied to an output iterator — for example, see the description of std::copy:

template<class InputIterator, class OutputIterator>
OutputIterator copy(InputIterator first, InputIterator last,
                    OutputIterator result);

-1- Effects: Copies elements in the range [first,last) into the range [result,result + (last - first)) starting from first and proceeding to last. For each non-negative integer n < (last - first), performs *(result + n) = *(first + n).

std::advance is not supposed to be applicable to output iterators, so we need a different method of description.

See also message c++std-lib-32908.

Proposed resolution:


2174. wstring_convert::converted() should be noexcept

Section: 22.3.3.2.2 [conversions.string] Status: New Submitter: Jonathan Wakely Opened: 2012-08-02 Last modified: 2012-09-24

View other active issues in [conversions.string].

View all other issues in [conversions.string].

View all issues with New status.

Discussion:

There is no reason wstring_convert::converted() shouldn't be noexcept.

It might be possible for wstring_convert::state() and wbuffer_convert::state() to be noexcept too, depending on the requirements on mbstate_t.

Proposed resolution:

This wording is relative to N3376.

  1. Edit in the class template wstring_convert synopsis 22.3.3.2.2 [conversions.string] p2:

    size_t converted() const noexcept;
    
  2. Edit the signature before 22.3.3.2.2 [conversions.string] p6:

    size_t converted() const noexcept;
    

2175. wstring_convert and wbuffer_convert validity

Section: 22.3.3.2.2 [conversions.string], 22.3.3.2.3 [conversions.buffer] Status: New Submitter: Jonathan Wakely Opened: 2012-08-02 Last modified: 2012-09-24

View other active issues in [conversions.string].

View all other issues in [conversions.string].

View all issues with New status.

Discussion:

See discussion following c++std-lib-32710.

It's not specified what happens if wstring_convert and wbuffer_convert objects are constructed with null Codecvt pointers.

Should the constructors have preconditions that the pointers are not null? If not, are conversions expected to fail, or is it undefined to attempt conversions if the pointers are null?

There are no observer functions to check whether objects were constructed with valid Codecvt pointers. If the types are made movable such observers would be necessary even if the constructors require non-null pointers (see also LWG 2176).

Proposed resolution:

This wording is relative to N3376.

  1. Insert a new paragraph before 22.3.3.2.2 [conversions.string] paragraph 16:

    wstring_convert(Codecvt *pcvt = new Codecvt);
    wstring_convert(Codecvt *pcvt, state_type state);
    wstring_convert(const byte_string& byte_err,
      const wide_string& wide_err = wide_string());
    

    -?- Requires: For the first and second constructors pcvt != nullptr.

    -16- Effects: The first constructor shall store pcvt in cvtptr and default values in cvtstate, byte_err_string, and wide_err_string. The second constructor shall store pcvt in cvtptr, state in cvtstate, and default values in byte_err_string and wide_err_string; moreover the stored state shall be retained between calls to from_bytes and to_bytes. The third constructor shall store new Codecvt in cvtptr, state_type() in cvtstate, byte_err in byte_err_string, and wide_err in wide_err_string.

  2. Insert a new paragraph before 22.3.3.2.3 [conversions.buffer] paragraph 10:

    wbuffer_convert(std::streambuf *bytebuf = 0,
      Codecvt *pcvt = new Codecvt, state_type state = state_type());
    

    -?- Requires: pcvt != nullptr.

    -10- Effects: The constructor constructs a stream buffer object, initializes bufptr to bytebuf, initializes cvtptr to pcvt, and initializes cvtstate to state.


2176. Special members for wstring_convert and wbuffer_convert

Section: 22.3.3.2.2 [conversions.string], 22.3.3.2.3 [conversions.buffer] Status: New Submitter: Jonathan Wakely Opened: 2012-08-02 Last modified: 2012-09-24

View other active issues in [conversions.string].

View all other issues in [conversions.string].

View all issues with New status.

Discussion:

See discussion following c++std-lib-32699.

The constructors for wstring_convert and wbuffer_convert should be explicit, to avoid implicit conversions which take ownership of a Codecvt pointer and delete it unexpectedly.

Secondly, 22.3.3.2.3 [conversions.buffer] p11 describes a destructor which is not declared in the class synopsis in p2.

Finally, and most importantly, the definitions in 22.3.3.2.2 [conversions.string] and 22.3.3.2.3 [conversions.buffer] imply implicitly-defined copy constructors and assignment operators, which would do shallow copies of the owned Codecvt objects and result in undefined behaviour in the destructors.

Codecvt is not required to be CopyConstructible, so deep copies are not possible. The wstring_convert and wstring_buffer types could be made move-only, but the proposed resolution below doesn't do so because of the lack of preconditions regarding null Codecvt pointers and the absence of observer functions that would allow users to check preconditions (see also LWG 2175).

Proposed resolution:

This wording is relative to N3376.

  1. Edit the class template wstring_convert synopsis in 22.3.3.2.2 [conversions.string] p2:

    explicit wstring_convert(Codecvt *pcvt = new Codecvt);
    wstring_convert(Codecvt *pcvt, state_type state);
    explicit wstring_convert(const byte_string& byte_err,
                             const wide_string& wide_err = wide_string());
    ~wstring_convert();
    
    wstring_convert(const wstring_convert&) = delete;
    wstring_convert& operator=(const wstring_convert&) = delete;
    				 
    
  2. Edit the signatures before 22.3.3.2.2 [conversions.string] p16:

    explicit wstring_convert(Codecvt *pcvt = new Codecvt);
    wstring_convert(Codecvt *pcvt, state_type state);
    explicit wstring_convert(const byte_string& byte_err,
        const wide_string& wide_err = wide_string());
    
  3. Edit the class template wbuffer_convert synopsis in 22.3.3.2.3 [conversions.buffer] p2:

    explicit wbuffer_convert(std::streambuf *bytebuf = 0,
                             Codecvt *pcvt = new Codecvt,
                             state_type state = state_type());
    
    ~wbuffer_convert();
    
    wbuffer_convert(const wbuffer_convert&) = delete;
    wbuffer_convert& operator=(const wbuffer_convert&) = delete;
    						 
    
  4. Edit the signature before 22.3.3.2.3 [conversions.buffer] p10:

    explicit wbuffer_convert(std::streambuf *bytebuf = 0,
        Codecvt *pcvt = new Codecvt, state_type state = state_type());
    

2177. Requirements on Copy/MoveInsertable

Section: 23.2.1 [container.requirements.general] Status: New Submitter: Loïc Joly Opened: 2012-08-10 Last modified: 2012-09-24

View other active issues in [container.requirements.general].

View all other issues in [container.requirements.general].

View all issues with New status.

Discussion:

See also discussion following c++std-lib-32883 and c++std-lib-32897.

The requirements on CopyInsertable and MoveInsertable are either incomplete, or complete but hard to figure out.

From e-mail c++std-lib-32897:

Pablo Halpern:

I agree that we need semantic requirements for all of the *Insertable concepts analogous to the requirements we have on similar concepts.

Howard Hinnant:

I've come to believe that the standard is actually correct as written in this area. But it is really hard to read. I would have no objection whatsoever to clarifications to CopyInsertable as you suggest (such as the post-conditions on v). And I do agree with you that the correct approach to the clarifications is to confirm that CopyInsertable implies MoveInsertable.

Proposed resolution:

This wording is relative to N3376.

  1. Edit 23.2.1 [container.requirements.general] p13 as indicated:

    -13- […] Given a container type X having an allocator_type identical to A and a value_type identical to T and given an lvalue m of type A, a pointer p of type T*, an expression v of type (possibly const) T, and an rvalue rv of type T, the following terms are defined. If X is not allocator-aware, the terms below are defined as if A were std::allocator<T> — no allocator object needs to be created and user specializations of std::allocator<T> are not instantiated:

    • T is DefaultInsertable into X means that the following expression is well-formed:

      allocator_traits<A>::construct(m, p);
      
    • An element of X is default-inserted if it is initialized by evaluation of the expression

      allocator_traits<A>::construct(m, p);
      

      where p is the address of the uninitialized storage for the element allocated within X.

    • T is CopyMoveInsertable into X means that the following expression is well-formed:

      allocator_traits<A>::construct(m, p, rv);
      

      and when evaluated the following postconditions hold: The value of *p is equivalent to the value of rv before the evaluation. [Note: rv remains a valid object. Its state is unspecified — end note]

    • T is MoveCopyInsertable into X means that, in addition to satisfying the MoveInsertable requirements, the following expression is well-formed:

      allocator_traits<A>::construct(m, p, rv);
      

      and when evaluated the following postconditions hold: The value of v is unchanged and is equivalent to *p.

    • T is EmplaceConstructible into X from args, for zero or more arguments args, means that the following expression is well-formed:

      allocator_traits<A>::construct(m, p, args);
      
    • T is Erasable from X means that the following expression is well-formed:

      allocator_traits<A>::destroy(m, p);
      

    [Note: A container calls allocator_traits<A>::construct(m, p, args) to construct an element at p using args. The default construct in std::allocator will call ::new((void*)p) T(args), but specialized allocators may choose a different definition. — end note]


2178. Allocator requirement changes not mentioned Annex C

Section: 17.6.3.5 [allocator.requirements], C.2 [diff.library] Status: New Submitter: Nevin Liber Opened: 2012-08-14 Last modified: 2012-09-24

View other active issues in [allocator.requirements].

View all other issues in [allocator.requirements].

View all issues with New status.

Discussion:

Given that a number of things were removed from the allocator requirements (reference, const_reference, address() in 17.6.3.5 [allocator.requirements]), it seems that these incompatible changes should be mentioned in Annex C.2 [diff.library], more specifically in [diff.cpp03].

Proposed resolution:


2179. enable_shared_from_this and construction from raw pointers

Section: 20.7.2.4 [util.smartptr.enab], 20.7.2.2.1 [util.smartptr.shared.const] Status: New Submitter: Daniel Krügler Opened: 2012-08-16 Last modified: 2012-09-24

View all issues with New status.

Discussion:

On reflector message c++std-lib-32927, Matt Austern asked whether the following example should be well-defined or not

struct X : public enable_shared_from_this<X> { };
auto xraw = new X;
shared_ptr<X> xp1(xraw);
shared_ptr<X> xp2(xraw);

pointing out that 20.7.2.2.1 [util.smartptr.shared.const] does not seem to allow it, since xp1 and xp2 aren't allowed to share ownership, because each of them is required to have use_count() == 1. Despite this wording it might be reasonable (and technical possible) to implement that request.

On the other hand, there is the non-normative note in 20.7.2.4 [util.smartptr.enab] p11 (already part of TR1):

The shared_ptr constructors that create unique pointers can detect the presence of an enable_shared_from_this base and assign the newly created shared_ptr to its __weak_this member.

Now according to the specification in 20.7.2.2.1 [util.smartptr.shared.const] p3-7:

template<class Y> explicit shared_ptr(Y* p);

the notion of creating unique pointers can be read to be included by this note, because the post-condition of this constructor is unique() == true. Evidence for this interpretation seems to be weak, though.

Howard Hinnant presented the counter argument, that actually the following is an "anti-idiom" and it seems questionable to teach it to be well-defined in any case:

auto xraw = new X;
shared_ptr<X> xp1(xraw);
shared_ptr<X> xp2(xraw);

He also pointed out that the current post-conditions of the affected shared_ptr constructor would need to be reworded.

It needs to be decided, which direction to follow. If this idiom seems too much broken to be supported, the note could be improved. If it should be supported, the constructors in 20.7.2.2.1 [util.smartptr.shared.const] need a careful analysis to ensure that post-conditions are correct.

Several library implementations currently do not support this example, instead they typically cause a crash. Matt points out that there are currently no explicit requirements imposed on shared_ptr objects to own the same underlying object without sharing the ownership. It might be useful to add such a requirement.

Proposed resolution:


2180. Exceptions from std::seed_seq operations

Section: 26.5.7.1 [rand.util.seedseq] Status: New Submitter: Daniel Krügler Opened: 2012-08-18 Last modified: 2012-09-24

View all other issues in [rand.util.seedseq].

View all issues with New status.

Discussion:

26.5.7.1 [rand.util.seedseq] p1 says upfront:

No function described in this section 26.5.7.1 [rand.util.seedseq] throws an exception.

This constraint seems non-implementable to me when looking especially at the members

template<class T>
seed_seq(initializer_list<T> il);

template<class InputIterator>
seed_seq(InputIterator begin, InputIterator end);

which have the effect of invoking v.push_back() for the exposition-only member of type std::vector (or its equivalent) over all elements of the provided range, so out-of-memory exceptions are always possible and the seed_seq object doesn't seem to be constructible this way.

In addition to the potential lack-of-resources problem, the operations of InputIterator might also throw exceptions.

Aside to that it should me mentioned, that a default constructor of vector<uint_least32_t> in theory can also throw exceptions, even though this seems less of a problem to me in this context, because such an implementation could easily use a different internal container in seed_seq that can hold this no-throw exception guarantee.

Secondly, a slightly different problem category related to exceptions occurs for the member templates

template<class RandomAccessIterator>
void generate(RandomAccessIterator begin, RandomAccessIterator end);

template<class OutputIterator>
void param(OutputIterator dest) const;

where the actual operations performed by the implementation would never need to throw, but since they invoke operations of a user-provided customization point, the overall operation, like for example

copy(v.begin(), v.end(), dest);

could also throw exceptions. In this particular example we can just think of a std::back_insert_iterator applied to a container that needs to allocate its elements used as the type for OutputIterator.

Even though Clause 26 [numerics] has mostly stronger exception constraints than other parts of the library the here discussed are overrestrictive, especially since no operation of std::seed_seq except the template generate is actually needed within the library implementation, as mentioned in the discussion of LWG 2124.

I suggest to remove the general no-exception constraints for operations of std::seed_seq except for member size() and the default constructor and to provide specific wording for generate() and param() to ensure that the algorithm itself is a nothrow operation, which is especially for generate() important, because the templates specified in 26.5.3 [rand.eng] and 26.5.4 [rand.adapt] also depend on this property indirectly, which is further discussed in LWG 2181.

Howard:

I suggest to use a different form for the exception specification, something similar to 20.8.9.1.2 [func.bind.bind] p4:

Throws: Nothing unless an operation on RandomAccessIterator throws an exception.

Daniel:

The currently suggested "what and when" form seems a bit more specific and harmonizes with the form used for function template generate_canonical from 26.5.7.2 [rand.util.canonical].

Proposed resolution:

This wording is relative to N3376.

  1. Edit 26.5.7.1 [rand.util.seedseq] p1 as indicated:

    -1- No function described in this section 26.5.7.1 [rand.util.seedseq] throws an exception.

  2. Edit 26.5.7.1 [rand.util.seedseq] around p2 as indicated:

    seed_seq();
    

    -2- Effects: Constructs a seed_seq object as if by default-constructing its member v.

    -?- Throws: Nothing.

  3. Edit 26.5.7.1 [rand.util.seedseq] around p7 as indicated:

    template<class RandomAccessIterator>
      void generate(RandomAccessIterator begin, RandomAccessIterator end);
    

    -7- Requires: RandomAccessIterator shall meet the requirements of a mutable random access iterator (Table 111) type. Moreover, iterator_traits<class RandomAccessIterator>::value_type shall denote an unsigned integer type capable of accommodating 32-bit quantities.

    -8- Effects: Does nothing if begin == end. Otherwise, with s = v.size() and n = end - begin, fills the supplied range [begin, end) according to the following algorithm […]

    -?- Throws: What and when RandomAccessIterator operations of begin and end throw.

  4. Edit 26.5.7.1 [rand.util.seedseq] around p9 as indicated:

    size_t size() const;
    

    -9- Returns: The number of 32-bit units that would be returned by a call to param().

    -??- Throws: Nothing.

    -10- Complexity: constant time.

  5. Edit 26.5.7.1 [rand.util.seedseq] around p11 as indicated:

    template<class OutputIterator>
      void param(OutputIterator dest) const;
    

    -11- Requires: OutputIterator shall satisfy the requirements of an output iterator (Table 108) type. Moreover, the expression *dest = rt shall be valid for a value rt of type result_type.

    -12- Effects: Copies the sequence of prepared 32-bit units to the given destination, as if by executing the following statement:

    copy(v.begin(), v.end(), dest);
    

    -??- Throws: What and when OutputIterator operations of dest throw.


2181. Exceptions from seed sequence operations

Section: 26.5.1.2 [rand.req.seedseq], 26.5.3 [rand.eng], 26.5.4 [rand.adapt] Status: New Submitter: Daniel Krügler Opened: 2012-08-18 Last modified: 2012-09-24

View other active issues in [rand.req.seedseq].

View all other issues in [rand.req.seedseq].

View all issues with New status.

Discussion:

LWG issue 2180 points out some deficiences in regard to the specification of the library-provided type std::seed_seq regarding exceptions, but there is another specification problem in regard to general types satisfying the seed sequence constraints (named SSeq) as described in 26.5.1.2 [rand.req.seedseq].

26.5.3 [rand.eng] p3 and 26.5.4.1 [rand.adapt.general] p3 say upfront:

Except where specified otherwise, no function described in this section 26.5.3 [rand.eng]/26.5.4 [rand.adapt] throws an exception.

This constraint causes problems, because the described templates in these sub-clauses depend on operations of SSeq::generate() which is a function template, that depends both on operations provided by the implementor of SSeq (e.g. of std::seed_seq), and those of the random access iterator type provided by the caller. With class template linear_congruential_engine we have just one example for a user of SSeq::generate() via:

template<class Sseq> 
linear_congruential_engine<>::linear_congruential_engine(Sseq& q);

template<class Sseq> 
void linear_congruential_engine<>::seed(Sseq& q);

None of these operations has an exclusion rule for exceptions.

As described in 2180 the wording for std::seed_seq should and can be fixed to ensure that operations of seed_seq::generate() won't throw except from operations of the provided iterator range, but there is no corresponding "safety belt" for user-provided SSeq types, since 26.5.1.2 [rand.req.seedseq] does not impose no-throw requirements onto operations of seed sequences.

  1. A quite radical step to fix this problem would be to impose general no-throw requirements on the expression q.generate(rb,re) from Table 115, but this is not as simple as it looks initially, because this function again depends on general types that are mutable random access iterators. Typically, we do not impose no-throw requirements on iterator operations and this would restrict general seed sequences where exceptions are not a problem. Furthermore, we do not impose comparable constraints for other expressions, like that of the expression g() in Table 116 for good reasons, e.g. random_device::operator() explicitly states when it throws exceptions.

  2. A less radical variant of the previous suggestion would be to add a normative requirement on the expression q.generate(rb,re) from Table 115 that says: "Throws nothing if operations of rb and re do not throw exceptions". Nevertheless we typically do not describe conditional Throws elements in proper requirement sets elsewhere (Container requirements excluded, they just describe the containers from Clause 23) and this may exclude resonable implementations of seed sequences that could throw exceptions under rare situations.

  3. The iterator arguments provided to SSeq::generate() for operations in templates of 26.5.3 [rand.eng] and 26.5.4 [rand.adapt] are under control of implementations, so we could impose stricter exceptions requirements on SSeq::generate() for SSeq types that are used to instantiate member templates in 26.5.3 [rand.eng] and 26.5.4 [rand.adapt] solely.

  4. We simply add extra wording to the introductive parts of 26.5.3 [rand.eng] and 26.5.4 [rand.adapt] that specify that operations of the engine (adaptor) templates that depend on a template parameter SSeq throw no exception unless SSeq::generate() throws an exception.

Given these options I would suggest to apply the variant described in the fourth bullet.

The proposed resolution attempts to reduce a lot of the redundancies of requirements in the introductory paragraphs of 26.5.3 [rand.eng] and 26.5.4 [rand.adapt] by introducing a new intermediate sub-clause "Engine and engine adaptor class templates" following sub-clause 26.5.2 [rand.synopsis]. This approach also solves the problem that currently 26.5.3 [rand.eng] also describes requirements that apply for 26.5.4 [rand.adapt] (Constrained templates involving the Sseq parameters).

Proposed resolution:

This wording is relative to N3376.

  1. Add a new sub-clause titled "Engine and engine adaptor class templates" following sub-clause 26.5.2 [rand.synopsis] (but at the same level) and add one further sub-clause "General" as child of the new sub-clause as follows:

    Engine and engine adaptor class templates [rand.engadapt]

    General [rand.engadapt.general]

    -?- Throughout this sub-clause general requirements and conventions are described that apply to every class template specified in sub-clause 26.5.3 [rand.eng] and 26.5.4 [rand.adapt]. Phrases of the form "in those sub-clauses" shall be interpreted as equivalent to "in sub-clauses 26.5.3 [rand.eng] and 26.5.4 [rand.adapt]".

    -?- Except where specified otherwise, the complexity of each function specified in those sub-clauses is constant.

    -?- Except where specified otherwise, no function described in those sub-clauses throws an exception.

    -?- Every function described in those sub-clauses that has a function parameter q of type SSeq& for a template type parameter named SSeq that is different from type std::seed_seq throws what and when the invocation of q.generate throws.

    -?- Descriptions are provided in those sub-clauses only for engine operations that are not described in 26.5.1.4 [rand.req.eng], for adaptor operations that are not described in 26.5.1.5 [rand.req.adapt], or for operations where there is additional semantic information. In particular, declarations for copy constructors, for copy assignment operators, for streaming operators, and for equality and inequality operators are not shown in the synopses.

    -?- Each template specified in those sub-clauses requires one or more relationships, involving the value(s) of its non-type template parameter(s), to hold. A program instantiating any of these templates is ill-formed if any such required relationship fails to hold.

    -?- For every random number engine and for every random number engine adaptor X defined in those sub-clauses:

    • if the constructor

      template <class Sseq> explicit X(Sseq& q);
      

      is called with a type Sseq that does not qualify as a seed sequence, then this constructor shall not participate in overload resolution;

    • if the member function

      template <class Sseq> void seed(Sseq& q);
      

      is called with a type Sseq that does not qualify as a seed sequence, then this function shall not participate in overload resolution;

    The extent to which an implementation determines that a type cannot be a seed sequence is unspecified, except that as a minimum a type shall not qualify as a seed sequence if it is implicitly convertible to X::result_type.

  2. Edit the contents of sub-clause 26.5.3 [rand.eng] as indicated:

    -1- Each type instantiated from a class template specified in this section 26.5.3 [rand.eng] satisfies the requirements of a random number engine (26.5.1.4 [rand.req.eng]) type and the general implementation requirements specified in sub-clause [rand.engadapt.general].

    -2- Except where specified otherwise, the complexity of each function specified in this section 26.5.3 [rand.eng] is constant.

    -3- Except where specified otherwise, no function described in this section 26.5.3 [rand.eng] throws an exception.

    -4- Descriptions are provided in this section 26.5.3 [rand.eng] only for engine operations that are not described in 26.5.1.4 [rand.req.eng] […]

    -5- Each template specified in this section 26.5.3 [rand.eng] requires one or more relationships, involving the value(s) of its non-type template parameter(s), to hold. […]

    -6- For every random number engine and for every random number engine adaptor X defined in this subclause (26.5.3 [rand.eng]) and in sub-clause 26.5.3 [rand.eng]: […]

  3. Edit the contents of sub-clause 26.5.4.1 [rand.adapt.general] as indicated:

    -1- Each type instantiated from a class template specified in this section 26.5.3 [rand.eng]26.5.4 [rand.adapt] satisfies the requirements of a random number engine adaptor (26.5.1.5 [rand.req.adapt]) type and the general implementation requirements specified in sub-clause [rand.engadapt.general].

    -2- Except where specified otherwise, the complexity of each function specified in this section 26.5.4 [rand.adapt] is constant.

    -3- Except where specified otherwise, no function described in this section 26.5.4 [rand.adapt] throws an exception.

    -4- Descriptions are provided in this section 26.5.4 [rand.adapt] only for engine operations that are not described in 26.5.1.5 [rand.req.adapt] […]

    -5- Each template specified in this section 26.5.4 [rand.adapt] requires one or more relationships, involving the value(s) of its non-type template parameter(s), to hold. […]


2182. Container::[const_]reference types are misleadingly specified

Section: 26.5.1.2 [rand.req.seedseq] Status: New Submitter: Daniel Krügler Opened: 2012-08-20 Last modified: 2012-09-24

View other active issues in [rand.req.seedseq].

View all other issues in [rand.req.seedseq].

View all issues with New status.

Discussion:

According to Table 96 (Container requirements) the return type of X::reference and X::const_reference is "lvalue of T" and "const lvalue of T", respectively. This does not make much sense, because an lvalue is an expression category, not a type. It could also refer to an expression that has a type, but this doesn't make sense either in this context, because obviously X::[const_]reference are intended to refer to types.

Given the fact that vector<bool> has no real reference type for X::[const_]reference and this definition presumably is intended to cover such situations as well, one might think that the wording is just a sloppy form of "type that represents a [const] lvalue of T". But this is also problematic, because basically all proxy reference expressions are rvalues.

It is unclear what the intention is. A straightward way of fixing this wording could make X::[const_]reference identical to [const] T&. This holds for all Library containers except for vector<bool>.

Another way of solving this definition problem would be to impose a requirement that holds for both references and reference-like proxies. Both X::reference and X::const_reference would need to be convertible to const T&. Additionally X::reference would need to support for a mutable container an assignment expression of the form declval<X::reference>() = declval<T>() (this presentation intentionally does not require declval<X::reference&>() = declval<T>()).

Further, the Table 96 does not impose any relations between X::reference and X::const_reference. It seems that at least X::reference needs to be convertible to X::const_reference.

A related question is whether X::reference is supposed to be a mutable reference-like type, irrespective of whether the container is an immutable container or not. The way, type match_results defines reference identical to const_reference indicates one specific interpretation. Note that this can be a different decision as that for iterator and const_iterator, e.g. for sets the type X::reference still is a mutable reference, even though iterator is described as constant iterator.

The proposed resolution is incomplete in regard to the last question.

Proposed resolution:

This wording is relative to N3376.

  1. Change Table 96 — "Container requirements" as indicated:

    Table p6 — Container requirements
    Expression Return type Operational
    Semantics
    Assertion/note
    pre-/post-condition
    Complexity
    X::reference lvalue oftype that behaves like a reference to T   convertible to X::const_reference
    and to const T&
    compile time
    X::const_reference const lvalue oftype that behaves like a reference to const T   convertible to const T& compile time

2183. Muddled allocator requirements for match_results constructors

Section: 28.10.1 [re.results.const], 28.10.6 [re.results.all] Status: New Submitter: Pete Becker Opened: 2012-08-29 Last modified: 2012-09-24

View other active issues in [re.results.const].

View all other issues in [re.results.const].

View all issues with New status.

Discussion:

28.10.1 [re.results.const] p1 says:

In all match_results constructors, a copy of the Allocator argument shall be used for any memory allocation performed by the constructor or member functions during the lifetime of the object.

There are three constructors:

match_results(const Allocator& = Allocator());
match_results(const match_results& m);
match_results(match_results&& m) noexcept;

The second and third constructors do no have an Allocator argument, so despite the "all match_results constructors", it is not possible to use "the Allocator argument" for the second and third constructors.

The requirements for those two constructors also does not give any guidance. The second constructor has no language about allocators, and the third states that the stored Allocator value is move constructed from m.get_allocator(), but doesn't require using that allocator to allocate memory.

The same basic problem recurs in 28.10.6 [re.results.all], which gives the required return value for get_allocator():

Returns: A copy of the Allocator that was passed to the object's constructor or, if that allocator has been replaced, a copy of the most recent replacement.

Again, the second and third constructors do not take an Allocator, so there is nothing that meets this requirement when those constructors are used.

Proposed resolution:


2184. Muddled allocator requirements for match_results assignments

Section: 28.10.1 [re.results.const], 28.10.6 [re.results.all] Status: New Submitter: Pete Becker Opened: 2012-08-29 Last modified: 2012-09-24

View other active issues in [re.results.const].

View all other issues in [re.results.const].

View all issues with New status.

Discussion:

The effects of the two assignment operators are specified in Table 141. Table 141 makes no mention of allocators, so, presumably, they don't touch the target object's allocator. That's okay, but it leaves the question: match_results::get_allocator() is supposed to return "A copy of the Allocator that was passed to the object's constructor or, if that allocator has been replaced, a copy of the most recent replacement"; if assignment doesn't replace the allocator, how can the allocator be replaced?

Proposed resolution:


2185. Missing throws clause for future/shared_future::wait_for/wait_until

Section: 30.6.6 [futures.unique_future], 30.6.7 [futures.shared_future] Status: New Submitter: Vicente J. Botet Escriba Opened: 2012-09-20 Last modified: 2012-09-24

View all other issues in [futures.unique_future].

View all issues with New status.

Discussion:

The functions future::wait_for, future::wait_until, shared_future::wait_for, and shared_future::wait_for can throw any timeout-related exceptions. It would be better if the wording could be more explicit. This is in line with the changes proposed in LWG 2093's Throws element of condition_variable::wait with predicate.

Proposed resolution:

This wording is relative to N3376.

  1. Change 30.6.6 [futures.unique_future] as indicated:

    template <class Rep, class Period>
    future_status wait_for(const chrono::duration<Rep, Period>& rel_time) const;
    

    -21- Effects: none if the shared state contains a deferred function (30.6.8 [futures.async]), otherwise blocks until the shared state is ready or until the relative timeout (30.2.4 [thread.req.timing]) specified by rel_time has expired.

    -22- Returns: […]

    -??- Throws: timeout-related exceptions (30.2.4 [thread.req.timing]).

    template <class Clock, class Duration>
    future_status wait_until(const chrono::time_point<Clock, Duration>& abs_time) const;
    

    -23- Effects: none if the shared state contains a deferred function (30.6.8 [futures.async]), otherwise blocks until the shared state is ready or until the absolute timeout (30.2.4 [thread.req.timing]) specified by abs_time has expired.

    -24- Returns: […]

    -??- Throws: timeout-related exceptions (30.2.4 [thread.req.timing]).

  2. Change 30.6.7 [futures.shared_future] as indicated:

    template <class Rep, class Period>
    future_status wait_for(const chrono::duration<Rep, Period>& rel_time) const;
    

    -23- Effects: none if the shared state contains a deferred function (30.6.8 [futures.async]), otherwise blocks until the shared state is ready or until the relative timeout (30.2.4 [thread.req.timing]) specified by rel_time has expired.

    -24- Returns: […]

    -??- Throws: timeout-related exceptions (30.2.4 [thread.req.timing]).

    template <class Clock, class Duration>
    future_status wait_until(const chrono::time_point<Clock, Duration>& abs_time) const;
    

    -25- Effects: none if the shared state contains a deferred function (30.6.8 [futures.async]), otherwise blocks until the shared state is ready or until the absolute timeout (30.2.4 [thread.req.timing]) specified by abs_time has expired.

    -26- Returns: […]

    -??- Throws: timeout-related exceptions (30.2.4 [thread.req.timing]).


2186. Incomplete action on async/launch::deferred

Section: 30.6.8 [futures.async] Status: New Submitter: Vicente J. Botet Escriba Opened: 2012-09-20 Last modified: 2012-09-24

View other active issues in [futures.async].

View all other issues in [futures.async].

View all issues with New status.

Discussion:

The description of the effects of async when the launch policy is launch::deferred doesn't state what is done with the result of the deferred function invocation and the possible exceptions as it is done for the asynchronous function when the policy is launch::async.

Proposed resolution:

This wording is relative to N3376.

  1. Change 30.6.8 [futures.async] p3 bullet 2 as indicated:

    template <class F, class... Args>
    future<typename result_of<typename decay<F>::type(typename decay<Args>::type...)>::type>
    async(F&& f, Args&&... args);
    template <class F, class... Args>
    future<typename result_of<typename decay<F>::type(typename decay<Args>::type...)>::type>
    async(launch policy, F&& f, Args&&... args);
    

    -2- Requires: […]

    -3- Effects:: The first function behaves the same as a call to the second function with a policy argument of launch::async | launch::deferred and the same arguments for F and Args. […] The further behavior of the second function depends on the policy argument as follows (if more than one of these conditions applies, the implementation may choose any of the corresponding policies):

    • if policy & launch::async is non-zero […]

    • if policy & launch::deferred is non-zero — Stores DECAY_COPY(std::forward<F>(f)) and DECAY_COPY(std::forward<Args>(args))... in the shared state. These copies of f and args constitute a deferred function. Invocation of the deferred function evaluates INVOKE(std::move(g), std::move(xyz)) where g is the stored value of DECAY_COPY(std::forward<F>(f)) and xyz is the stored copy of DECAY_COPY(std::forward<Args>(args)).... Any return value is stored as the result in the shared state. Any exception propagated from the execution of the deferred function is stored as the exceptional result in the shared state. The shared state is not made ready until the function has completed. The first call to a non-timed waiting function (30.6.4 [futures.state]) on an asynchronous return object referring to this shared state shall invoke the deferred function in the thread that called the waiting function. Once evaluation of INVOKE(std::move(g), std::move(xyz)) begins, the function is no longer considered deferred. [Note: If this policy is specified together with other policies, such as when using a policy value of launch::async | launch::deferred, implementations should defer invocation or the selection of the policy when no more concurrency can be effectively exploited. — end note]


2187. vector<bool> is missing emplace and emplace_back member functions

Section: 23.3.7 [vector.bool] Status: New Submitter: Nevin Liber Opened: 2012-09-21 Last modified: 2012-09-24

View all other issues in [vector.bool].

View all issues with New status.

Discussion:

It should have them so that it more closely matches the vector<T> interface, as this helps when writing generic code.

Proposed resolution:

This wording is relative to N3376.

  1. Change the class template vector<bool> synopsis, 23.3.7 [vector.bool] p1, as indicated:

    namespace std {
      template <class Allocator> class vector<bool, Allocator> {
      public:
        […]
        // modifiers:
        template <class... Args> void emplace_back(Args&&... args);
        void push_back(const bool& x);
        void pop_back();
        template <class... Args> iterator emplace(const_iterator position, Args&&... args);
        iterator insert(const_iterator position, const bool& x);
        […]
      };
    }
    

2188. Reverse iterator does not fully support targets that overload operator&

Section: 24.5.1.3.5 [reverse.iter.opref] Status: New Submitter: Alisdair Meredith Opened: 2012-09-23 Last modified: 2012-09-24

View all other issues in [reverse.iter.opref].

View all issues with New status.

Discussion:

The specification for reverse_iterator::operator-> returns the address of the object yielded by dereferencing with operatator*, but does not have the usual wording about returning the true address of the object. As reverse_iterator requires the adapted iterator have at least the bidirectional iterator category, we know that the returned reference is a true reference, and not a proxy, hence we can use std::addressof on the reference to get the right answer.

This will most likely show itself as an issue with a list or vector of a type with such an overloaded operator, where algorithms are likely to work with a forward iteration, but not with reverse iteration.

Proposed resolution:

Revise 24.5.1.3.5 [reverse.iter.opref] p1, as indicated:

Returns: The true address of the object returned by &(operator*()).


2189. Throwing swap breaks unordered containers' state

Section: 23.2.5.1 [unord.req.except] Status: New Submitter: Alisdair Meredith Opened: 2012-09-23 Last modified: 2012-09-24

View all issues with New status.

Discussion:

The hash functor and key-comparison functor of unordered containers are allowed to throw on swap.

23.2.5.1 [unord.req.except]p3 "For unordered associative containers, no swap function throws an exception unless that exception is thrown by the swap of the container's Hash or Pred object (if any)."

In such a case we must offer the basic exception safety guarantee, where both objects are left in valid but unspecified states, and no resources are leaked. This yields a corrupt, un-usable container if the first swap succeeds, but the second fails by throwing, as the functors form a matched pair.

So our basic scenario is first, swap the allocators if the allocators propagate on swap, according to allocator_traits. Next we swap the pointers to our internal hash table data structures, so that they match the allocators that allocated them. (Typically, this operation cannot throw). Now our containers are back in a safely destructible state if an exception follows.

Next, let's say we swap the hash functor, and that throws. We have a corrupt data structure, in that the buckets are not correctly indexed by the correct functors, lookups will give unpredicatable results etc. We can safely restore a usable state by forcibly clearing each container - which does not leak resources and leaves us with two (empty but) usable containers.

Now let us assume that the hasher swap succeeds. Next we swap the equality comparator functor, and this too could throw. The important point to bear in mind is that these two functors form an important pairing - two objects that compare equal by the equality functor must also hash to the same value. If we swap one without the other, we most likely leave the container in an unusable state, even if we clear out all elements.

1. A colleague pointed out that the solution for this is to dynamically allocate the two functors, and then we need only swap pointers, which is not a throwing operation. And if we don't want to allocate on default construction (a common QoI request), we might consider moving to a dynamically allocated functors whenever swap is called, or on first insertion. Of course, allocating memory in swap is a whole new can of worms, but this does not really sound like the design we had intended.

2. The simplest option is to say that we do not support hasher or equality functors that throw on ADL swap. Note that the requirement is simply to not throw, rather than to be explicitly marked as noexcept. Throwing functors are allowed, so long as we never use values that would actually manifest a throw when used in an unordered container.

Pablo went on to give me several more options, to be sure we have a full set to consider:

3. Disallow one or the other functor from throwing. In that case, the possibly-throwing functor must be swapped first, then the other functor, the allocator, and the data pointer(s) afterwards (in any order -- there was a TC that allocator assignment and swap may not throw if the corresponding propagation trait is true.). Of course, the question becomes: which functor is allowed to throw and which one is not?

4. Require that any successful functor swap be reliably reversible. This is very inventive. I know of no other place in the standard where such a requirement is stated, though I have occasionally wanted such a guarantee.

5. Allow a failed swap to leave the containers in a state where future insertions may fail for reasons other than is currently allowed. Specifically, if the hash and equality functors are out of sync, all insertions will fail. Presumably some "incompletely swapped" exception would be thrown. This is "slightly" inventive, although people have been discussing "radioactive" states for a while.

Proposed resolution: