Doc. no. | D???? |
Date: | 2022-11-24 |
Project: | Programming Language C++ |
Reply to: | Jonathan Wakely <lwgchair@gmail.com> |
Revised 2022-11-24 at 09:59:48 UTC
Reference ISO/IEC IS 14882:2020(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 14882:2020(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 and Accepted Issues 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 14882:2020(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.
Issues reported to the LWG transition through a variety of statuses, indicating their progress towards a resolution. Typically, most issues will flow through the following stages.
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.
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).
Typically, an issue must have a proposed resolution in the currently published issues list, whose wording does not change during LWG review, to move to the Ready status.
Voting - This status should not be seen in a published issues list, but is a marker for use during meetings to indicate an issues was Ready in the pre-meeting mailing, the Proposed Resolution is correct, and the issue will be offered to the working group at the end of the current meeting to apply to the current working paper (WP) or to close in some other appropriate manner. This easily distinguishes such issues from those moving to Ready status during the meeting itself, that should not be forwarded until the next meeting. If the issue does not move forward, it should fall back to one of the other open states before the next list is published.
Immediate - This status should not be seen in a published issues list, but is a marker for use during meetings to indicate an issues was not Ready in the pre-meeting mailing, but the Proposed Resolution is correct, and the issue will be offered to the working group at the end of the current meeting to apply to the current working paper (WP) or to close in some other appropriate manner. This status is used only rarely, typically for fixes that are both small and obvious, and usually within a meeting of the expected publication of a revised standard. If the issue does not move forward, it should fall back to one of the other open states before the next list is published.
In addition, there are a few ways to categorise and issue that remains open to a resolution within the library, but is not actively being worked on.
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.
Core - The LWG has discussed the issue, and feels that some key part of resolving the issue is better handled by a cleanup of the language in the Core part of the standard. The issue is passed to the Core Working Group, which should ideally open a corresponding issue that can be linked from the library issue. Such issues will be revisitted after Core have made (or declined to make) any changes.
EWG - The LWG has discussed the issue, and wonder that some key part of resolving the issue is better handled by some (hopefully small) extension to the language. The issue is passed to the Evolution Working Group, which should ideally open a corresponding issue that can be linked from the library issue. Such issues will be revisitted after Evoltion have made (or declined to make) any recommendations. Positive recommendations from EWG will often mean the issue transition to Core status while we wait for some proposed new feature to land in the working paper.
LEWG - The LWG has discussed the issue, and deemd the issue is either an extension, however small, or changes the library design in some fundamental way, and so has delegated the initial work to the Library Evolution Working Group.
Ultimately, all issues should reach closure with one of the following statuses.
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.
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.
C++20 - (C++ Standard, as revised for 2020) - The full WG21/PL22.16 committee has voted to accept the Defect Report's Proposed Resolution into the published 2020 revision to the C++ standard, ISO/IEC IS 14882:2020(E).
C++17 - (C++ Standard, as revised for 2017) - The full WG21/PL22.16 committee has voted to accept the Defect Report's Proposed Resolution into the published 2017 revision to the C++ standard, ISO/IEC IS 14882:2017(E).
C++14 - (C++ Standard, as revised for 2014) - The full WG21/PL22.16 committee has voted to accept the Defect Report's Proposed Resolution into the published 2014 revision to the C++ standard, ISO/IEC IS 14882:2014(E).
C++11 - (C++ Standard, as revised for 2011) - The full WG21/PL22.16 committee has voted to accept the Defect Report's Proposed Resolution into the published 2011 revision to the C++ standard, ISO/IEC IS 14882:2011(E).
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.
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.
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.
TS - (TS - various) - The full WG21/PL22.16 committee has voted to accept the Defect Report's Proposed Resolution into a published Technical Specification.
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.
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).
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.
The following statuses have been retired, but may show up on older issues lists.
NAD Future - In addition to the regular status, the LWG believes that this issue should be revisited at the next revision of the standard. That is now an ongoing task managed by the Library Evolution Working Group, and most issues in this status were reopended with the status LEWG.
NAD Concepts - This status reflects an evolution of the language during the development of C++11, where a new feature entered the language, called concepts, that fundamentally changed the way templates would be specified and written. While this language feature was removed towards the end of the C++11 project, there is a clear intent to revisit this part of the language design. During that development, a number of issues were opened against the updated library related to use of that feature, or requesting fixes that would require explicit use of the concepts feature. All such issues have been closed with this status, and may be revisitted should this or a similar language feature return for a future standard.
NAD Arrays - This status reflects an evolution of the language during the development of C++14/17, where work on a Technical Specification, called the Arrays TS was begun. In early 2016, this work was abandoned, and the work item was officially withdrawn. During development of the TS, a number of issues were opened the features in the TS. All such issues have been closed with this status, and may be revisitted should this or a similar language feature return for a future standard.
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 PL22.16 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), an updated standard (C++11, C++14), or are closed without action other than a Record of Response (Resolved) where the desired effect has already been achieved by some other process. The intent of this LWG process is that only issues which are truly defects in the Standard move to the formal ISO DR status.
Section: 31 [input.output] Status: Open Submitter: Martin Sebor Opened: 2003-09-18 Last modified: 2018-12-09
Priority: 3
View all other issues in [input.output].
View all issues with Open status.
Discussion:
A third party test suite tries to exercise istream::ignore(N) with a negative value of N and expects that the implementation will treat N as if it were 0. Our implementation asserts that (N >= 0) holds and aborts the test.
I can't find anything in section 27 that prohibits such values but I don't see what the effects of such calls should be, either (this applies to a number of unformatted input functions as well as some member functions of the basic_streambuf template).
[ 2009-07 Frankfurt ]
This is related to LWG 255.
Move to NAD Future.
[LEWG Kona 2017]
Recommend Open: We agree that we should require N >= 0 for the selected functions
[2018-12-04 Reflector prioritization]
Set Priority to 3
Proposed resolution:
I propose that we add to each function in clause 27 that takes an argument, say N, of type streamsize a Requires clause saying that "N >= 0." The intent is to allow negative streamsize values in calls to precision() and width() but disallow it in calls to streambuf::sgetn(), istream::ignore(), or ostream::write().
[Kona: The LWG agreed that this is probably what we want. However, we need a review to find all places where functions in clause 27 take arguments of type streamsize that shouldn't be allowed to go negative. Martin will do that review.]
Section: 25.3.5.3 [input.iterators] Status: Open Submitter: Chris Jefferson Opened: 2004-09-16 Last modified: 2018-01-27
Priority: 2
View other active issues in [input.iterators].
View all other issues in [input.iterators].
View all issues with Open status.
Discussion:
From comp.std.c++:
I note that given an input iterator a for type T, then *a only has to be "convertable to T", not actually of type T.
Firstly, I can't seem to find an exact definition of "convertable to T". While I assume it is the obvious definition (an implicit conversion), I can't find an exact definition. Is there one?
Slightly more worryingly, there doesn't seem to be any restriction on the this type, other than it is "convertable to T". Consider two input iterators a and b. I would personally assume that most people would expect *a==*b would perform T(*a)==T(*b), however it doesn't seem that the standard requires that, and that whatever type *a is (call it U) could have == defined on it with totally different symantics and still be a valid inputer iterator.
Is this a correct reading? When using input iterators should I write T(*a) all over the place to be sure that the object I'm using is the class I expect?
This is especially a nuisance for operations that are defined to be "convertible to bool". (This is probably allowed so that implementations could return say an int and avoid an unnessary conversion. However all implementations I have seen simply return a bool anyway. Typical implemtations of STL algorithms just write things like while(a!=b && *a!=0). But strictly speaking, there are lots of types that are convertible to T but that also overload the appropriate operators so this doesn't behave as expected.
If we want to make code like this legal (which most people seem to expect), then we'll need to tighten up what we mean by "convertible to T".
[Lillehammer: The first part is NAD, since "convertible" is well-defined in core. The second part is basically about pathological overloads. It's a minor problem but a real one. So leave open for now, hope we solve it as part of iterator redesign.]
[ 2009-07-28 Reopened by Alisdair. No longer solved by concepts. ]
[ 2009-10 Santa Cruz: ]
Mark as NAD Future. We agree there's an issue, but there is no proposed solution at this time and this will be solved by concepts in the future.
[2017-02 in Kona, LEWG recommends NAD]
Has been clarified by 14. By design. Ranges might make it go away. Current wording for input iterators is more constrained.
[2017-06-02 Issues Telecon]
Move to Open. This is very similar to 2962, possibly a duplicate.
Marshall to research
[2017-07 Toronto Thurs Issue Prioritization]
Priority 2; same as 2962.
Proposed resolution:
Rationale:
[ San Francisco: ]
Solved by N2758.
Section: 32 [re] Status: Open Submitter: Eric Niebler Opened: 2005-07-01 Last modified: 2020-07-17
Priority: 4
View other active issues in [re].
View all other issues in [re].
View all issues with Open status.
Discussion:
A problem with TR1 regex is currently being discussed on the Boost developers list. It involves the handling of case-insensitive matching of character ranges such as [Z-a]. The proper behavior (according to the ECMAScript standard) is unimplementable given the current specification of the TR1 regex_traits<> class template. John Maddock, the author of the TR1 regex proposal, agrees there is a problem. The full discussion can be found at http://lists.boost.org/boost/2005/06/28850.php (first message copied below). We don't have any recommendations as yet.
-- Begin original message --
The situation of interest is described in the ECMAScript specification (ECMA-262), section 15.10.2.15:
"Even if the pattern ignores case, the case of the two ends of a range is significant in determining which characters belong to the range. Thus, for example, the pattern /[E-F]/i matches only the letters E, F, e, and f, while the pattern /[E-f]/i matches all upper and lower-case ASCII letters as well as the symbols [, \, ], ^, _, and `."
A more interesting case is what should happen when doing a case-insensitive match on a range such as [Z-a]. It should match z, Z, a, A and the symbols [, \, ], ^, _, and `. This is not what happens with Boost.Regex (it throws an exception from the regex constructor).
The tough pill to swallow is that, given the specification in TR1, I don't think there is any effective way to handle this situation. According to the spec, case-insensitivity is handled with regex_traits<>::translate_nocase(CharT) — two characters are equivalent if they compare equal after both are sent through the translate_nocase function. But I don't see any way of using this translation function to make character ranges case-insensitive. Consider the difficulty of detecting whether "z" is in the range [Z-a]. Applying the transformation to "z" has no effect (it is essentially std::tolower). And we're not allowed to apply the transformation to the ends of the range, because as ECMA-262 says, "the case of the two ends of a range is significant."
So AFAICT, TR1 regex is just broken, as is Boost.Regex. One possible fix is to redefine translate_nocase to return a string_type containing all the characters that should compare equal to the specified character. But this function is hard to implement for Unicode, and it doesn't play nice with the existing ctype facet. What a mess!
-- End original message --
[ John Maddock adds: ]
One small correction, I have since found that ICU's regex package does implement this correctly, using a similar mechanism to the current TR1.Regex.
Given an expression [c1-c2] that is compiled as case insensitive it:
Enumerates every character in the range c1 to c2 and converts it to it's case folded equivalent. That case folded character is then used a key to a table of equivalence classes, and each member of the class is added to the list of possible matches supported by the character-class. This second step isn't possible with our current traits class design, but isn't necessary if the input text is also converted to a case-folded equivalent on the fly.
ICU applies similar brute force mechanisms to character classes such as [[:lower:]] and [[:word:]], however these are at least cached, so the impact is less noticeable in this case.
Quick and dirty performance comparisons show that expressions such as "[X-\\x{fff0}]+" are indeed very slow to compile with ICU (about 200 times slower than a "normal" expression). For an application that uses a lot of regexes this could have a noticeable performance impact. ICU also has an advantage in that it knows the range of valid characters codes: code points outside that range are assumed not to require enumeration, as they can not be part of any equivalence class. I presume that if we want the TR1.Regex to work with arbitrarily large character sets enumeration really does become impractical.
Finally note that Unicode has:
Three cases (upper, lower and title). One to many, and many to one case transformations. Character that have context sensitive case translations - for example an uppercase sigma has two different lowercase forms - the form chosen depends on context(is it end of a word or not), a caseless match for an upper case sigma should match either of the lower case forms, which is why case folding is often approximated by tolower(toupper(c)).
Probably we need some way to enumerate character equivalence classes, including digraphs (either as a result or an input), and some way to tell whether the next character pair is a valid digraph in the current locale.
Hoping this doesn't make this even more complex that it was already,
[ Portland: Alisdair: Detect as invalid, throw an exception. Pete: Possible general problem with case insensitive ranges. ]
[ 2009-07 Frankfurt ]
We agree that this is a problem, but we do not know the answer.
We are going to declare this NAD until existing practice leads us in some direction.
No objection to NAD Future.
Move to NAD Future.
[LEWG Kona 2017]
Recommend Open: Tim Shen proposes: forbid use of case-insensitive ranges with regex traits other than std::regex_traits<{char, wchar_t, char16_t, char32_t}> when regex_constants::collate is specified.
[2020-07-17; Priority set to 4 in telecon]
Proposed resolution:
Section: 22.4.9 [tuple.rel], 99 [tr.tuple.rel] Status: LEWG Submitter: David Abrahams Opened: 2005-11-29 Last modified: 2016-01-28
Priority: Not Prioritized
View other active issues in [tuple.rel].
View all other issues in [tuple.rel].
View all issues with LEWG status.
Duplicate of: 348
Discussion:
Where possible, tuple comparison operators <,<=,=>, and > ought to be defined in terms of std::less rather than operator<, in order to support comparison of tuples of pointers.
[ 2009-07-28 Reopened by Alisdair. No longer solved by concepts. ]
[ 2009-10 Santa Cruz: ]
If we solve this for tuple we would have to solve it for pair algorithms, etc. It is too late to do that at this time. Move to NAD Future.
Proposed resolution:
change 6.1.3.5/5 from:
Returns: The result of a lexicographical comparison between t and u. The result is defined as: (bool)(get<0>(t) < get<0>(u)) || (!(bool)(get<0>(u) < get<0>(t)) && ttail < utail), where rtail for some tuple r is a tuple containing all but the first element of r. For any two zero-length tuples e and f, e < f returns false.
to:
Returns: The result of a lexicographical comparison between t and u. For any two zero-length tuples e and f, e < f returns false. Otherwise, the result is defined as: cmp( get<0>(t), get<0>(u)) || (!cmp(get<0>(u), get<0>(t)) && ttail < utail), where rtail for some tuple r is a tuple containing all but the first element of r, and cmp(x,y) is an unspecified function template defined as follows.
Where T is the type of x and U is the type of y:
if T and U are pointer types and T is convertible to U, returns less<U>()(x,y)
otherwise, if T and U are pointer types, returns less<T>()(x,y)
otherwise, returns (bool)(x < y)
[ Berlin: This issue is much bigger than just tuple (pair, containers, algorithms). Dietmar will survey and work up proposed wording. ]
Rationale:
Recommend NAD. This will be fixed with the next revision of concepts.
[ San Francisco: ]
Solved by N2770.
Section: 24.3.7 [array] Status: Open Submitter: Bo Persson Opened: 2006-12-30 Last modified: 2022-11-12
Priority: 3
View all other issues in [array].
View all issues with Open status.
Discussion:
The <array> header is given under 24.3 [sequences]. 24.3.7 [array]/paragraph 3 says:
"Unless otherwise specified, all array operations are as described in 24.2 [container.requirements]".
However, array isn't mentioned at all in section 24.2 [container.requirements]. In particular, Table 82 "Sequence requirements" lists several operations (insert, erase, clear) that std::array does not have in 24.3.7 [array].
Also, Table 83 "Optional sequence operations" lists several operations that std::array does have, but array isn't mentioned.
[ 2009-07 Frankfurt ]
The real issue seems to be different than what is described here. Non-normative text says that std::array is a sequence container, but there is disagreement about what that really means. There are two possible interpretations:
- a sequence container is one that satisfies all sequence container requirements
- a sequence container is one that satisfies some of the sequence container requirements. Any operation that the container supports is specified by one or more sequence container requirements, unless that operation is specifically singled out and defined alongside the description of the container itself.
Move to Tentatively NAD.
[ 2009-07-15 Loïc Joly adds: ]
The section 24.2.4 [sequence.reqmts]/1 states that array is a sequence. 24.2.4 [sequence.reqmts]/3 introduces table 83, named Sequence container requirements. This seems to me to be defining the requirements for all sequences. However, array does not follow all of this requirements (this can be read in the array specific section, for the standard is currently inconsistent).
Proposed resolution 1 (minimal change):
Say that array is a container, that in addition follows only some of the sequence requirements, as described in the array section:
The library provides
fivethree basic kinds of sequence containers:array, vector,forward_list, list, and deque. In addition, array and forward_list follows some of the requirements of sequences, as described in their respective sections.Proposed resolution 2 (most descriptive description, no full wording provided):
Introduce the notion of a Fixed Size Sequence, with it requirement table that would be a subset of the current Sequence container. array would be the only Fixed Size Sequence (but dynarray is in the queue for TR2). Sequence requirements would now be requirements in addition to Fixed Size Sequence requirements (it is currently in addition to container).
[ 2009-07 Frankfurt: ]
Move to NAD Editorial
[ 2009 Santa Cruz: ]
This will require a lot of reorganization. Editor doesn't think this is really an issue, since the description of array can be considered as overriding what's specified about sequences. Move to NAD.
[2022-10-27; Hubert Tong comments and requests to reopen]
This issue appears to be unresolved (should not be NAD).
As noted in 24.3.7.1 [array.overview] paragraph 3, array does not meet 24.2.2.2 [container.reqmts] paragraph 10. This means that array does not meet the container requirements, never mind the requirements for sequence containers or contiguous containers. However, there is wording that claims the opposite. 24.2.4 [sequence.reqmts] paragraph 1:In addition, array is provided as a sequence container which provides limited sequence operations because it has a fixed number of elements.
(Perhaps the above should be worded with "except".)
24.3.1 [sequences.general] paragraph 1:The headers <array> […] define class templates that meet the requirements for sequence containers.
24.3.7.1 [array.overview] paragraph 1:
[…] An array is a contiguous container (24.2.2.1 [container.requirements.general]).
In this comment, Casey suggests that the requirements be changed so that array does meet the requirements.
[Kona 2022-11-12; Set Priority to 3]
Proposed resolution:
Section: 33.6.4 [thread.mutex.requirements] Status: LEWG Submitter: Pete Becker Opened: 2008-12-05 Last modified: 2017-03-01
Priority: Not Prioritized
View other active issues in [thread.mutex.requirements].
View all other issues in [thread.mutex.requirements].
View all issues with LEWG status.
Duplicate of: 961
Discussion:
33.6.4 [thread.mutex.requirements] describes the requirements for a type to be a "Mutex type". A Mutex type can be used as the template argument for the Lock type that's passed to condition_variable_any::wait (although Lock seems like the wrong name here, since Lock is given a different formal meaning in 33.6.5 [thread.lock]) and, although the WD doesn't quite say so, as the template argument for lock_guard and unique_lock.
The requirements for a Mutex type include:
Also, a Mutex type "shall not be copyable nor movable".
The latter requirement seems completely irrelevant, and the three requirements on return types are tighter than they need to be. For example, there's no reason that lock_guard can't be instantiated with a type that's copyable. The rule is, in fact, that lock_guard, etc. won't try to copy objects of that type. That's a constraint on locks, not on mutexes. Similarly, the requirements for void return types are unnecessary; the rule is, in fact, that lock_guard, etc. won't use any returned value. And with the return type of bool, the requirement should be that the return type is convertible to bool.
[ Summit: ]
Move to open. Related to conceptualization and should probably be tackled as part of that.
- The intention is not only to place a constraint on what types such as lock_guard may do with mutex types, but on what any code, including user code, may do with mutex types. Thus the constraints as they are apply to the mutex types themselves, not the current users of mutex types in the standard.
- This is a low priority issue; the wording as it is may be overly restrictive but this may not be a real issue.
[ Post Summit Anthony adds: ]
Section 33.6.4 [thread.mutex.requirements] conflates the requirements on a generic Mutex type (including user-supplied mutexes) with the requirements placed on the standard-supplied mutex types in an attempt to group everything together and save space.
When applying concepts to chapter 30, I suggest that the concepts Lockable and TimedLockable embody the requirements for *use* of a mutex type as required by unique_lock/lock_guard/condition_variable_any. These should be relaxed as Pete describes in the issue. The existing words in 33.6.4 [thread.mutex.requirements] are requirements on all of std::mutex, std::timed_mutex, std::recursive_mutex and std::recursive_timed_mutex, and should be rephrased as such.
[2017-03-01, Kona]
SG1: Agreement that we need a paper.
Proposed resolution:
Section: 33.6.4 [thread.mutex.requirements] Status: LEWG Submitter: Pete Becker Opened: 2009-01-07 Last modified: 2017-03-01
Priority: Not Prioritized
View other active issues in [thread.mutex.requirements].
View all other issues in [thread.mutex.requirements].
View all issues with LEWG status.
Duplicate of: 936
Discussion:
33.6.4 [thread.mutex.requirements] describes required member functions of mutex types, and requires that they throw exceptions under certain circumstances. This is overspecified. User-defined types can abort on such errors without affecting the operation of templates supplied by standard-library.
[ Summit: ]
Move to open. Related to conceptualization and should probably be tackled as part of that.
[ 2009-10 Santa Cruz: ]
Would be OK to leave it as is for time constraints, could loosen later.
Mark as NAD Future.
[2017-03-01, Kona]
SG1: Agreement that we need a paper.
Proposed resolution:
Section: 24.3.11.3 [vector.capacity] Status: Open Submitter: Daniel Krügler Opened: 2009-04-20 Last modified: 2020-07-17
Priority: 3
View other active issues in [vector.capacity].
View all other issues in [vector.capacity].
View all issues with Open status.
Discussion:
I have the impression that even the wording of current draft N2857 does insufficiently express the intent of vector's reallocation strategy. This has produced not too old library implementations which release memory in the clear() function and even modern articles about C++ programming cultivate the belief that clear is allowed to do exactly this. A typical example is something like this:
const int buf_size = ...; std::vector<T> buf(buf_size); for (int i = 0; i < some_condition; ++i) { buf.resize(buf_size); write_or_read_data(buf.data()); buf.clear(); // Ensure that the next round get's 'zeroed' elements }
where still the myth is ubiquitous that buf might be allowed to reallocate it's memory inside the for loop.
IMO the problem is due to the fact, that
the effects clause of std::vector's erase overloads in 24.3.11.5 [vector.modifiers]/4 is silent about capacity changes. This easily causes a misunderstanding, because the counter parting insert functions described in 24.3.11.5 [vector.modifiers]/2 explicitly say, that
Causes reallocation if the new size is greater than the old capacity. If no reallocation happens, all the iterators and references before the insertion point remain valid.
It requires a complex argumentation chain about four different places in the standard to provide the — possibly weak — proof that calling clear() also does never change the capacity of the std::vector container. Since std::vector is the de-facto replacement of C99's dynamic arrays this type is near to a built-in type and it's specification should be clear enough that usual programmers can trust their own reading.
[ Batavia (2009-05): ]
Bill believes paragraph 1 of the proposed resolution is unnecessary because it is already implied (even if tortuously) by the current wording.
Move to Review.
[ 2009-10 Santa Cruz: ]
Mark as NAD. Rationale: there is no consensus to clarify the standard, general consensus that the standard is correct as written.
[2020-05-08; Reopen after reflector discussions]
"correct as written" has been disputed.
[2020-07-17; Priority set to 3 in telecon]
Proposed resolution:
[ This is a minimum version. I also suggest that the wording explaining the allocation strategy of std::vector in 24.3.11.3 [vector.capacity]/3 and /6 is moved into a separate sub paragraph of 24.3.11.3 [vector.capacity] before any of the prototype's are discussed, but I cannot provide reasonable wording changes now. ]
Change 24.3.11.3 [vector.capacity]/6 as follows:
It is guaranteed that no reallocation takes place during insertions or erasures that happen after a call to reserve() until the time when an insertion would make the size of the vector greater than the value of capacity().
Change 24.3.11.5 [vector.modifiers]/4 as follows:
Effects: The capacity shall remain unchanged and no reallocation shall happen. Invalidates iterators and references at or after the point of the erase.
Section: 24.2.8 [unord.req] Status: Open Submitter: Pablo Halpern Opened: 2009-07-17 Last modified: 2020-09-06
Priority: 3
View other active issues in [unord.req].
View all other issues in [unord.req].
View all issues with Open 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
[2012, Portland: Move to Open]
The wording still does not call out Pablo's original concern, that the element constructor is called no more than N times, and that the N squared term applies to moves during rehash.
Inconsistent use of O(n)+O(N) vs. O(n+N), with a preference for the former.
AJM to update wording with a reference to "no more than N element constructor calls".
Matt concerned that calling out the O(n) requirements is noise, and dangerous noise in suggesting a precision we do not mean. The cost of constructing a bucket is very different to constructing an element of user-supplied type.
AJM notes that if there are multiple rehashes, the 'n' complexity is probably not linear.
Matt suggests back to Open, Pablo suggests potentially NAD if we keep revisitting without achieving a resolution.
Matt suggests complexity we are concerned with is the number of operations, such as constructing elements, moving nodes, and comparing/hashing keys. We are less concerned with constructing buckets, which are generally noise in this bigger picture.
[2015-01-29 Telecon]
AM: essentially correct, but do we want to complicate the spec?
HH: Pablo has given us permission to NAD it JM: when I look at the first change in the P/R I find it mildly disturbing that the existing wording says you have a constant time constructor with a single element even if your n is 10^6, so I think adding this change makes people aware there might be a large cost in initializing the hash table, even though it doesn't show up in user-visible constructions. HH: one way to avoid that problem is make the default ctor noexcept. Then the container isn't allowed to create an arbitrarily large hash table AM: but this is the constructor where the user provides n MC: happy with the changes, except I agree with the editorial recommendation to keep the two 𝒪s separate. JW: yes, the constant 'k' is different in 𝒪(n) and 𝒪(N) GR: do we want to talk about buckets at all JM: yes, good to highlight that bucket construction might be a significant cost HH: suggest we take the suggestion to split 𝒪(n+N) to 𝒪(n)+𝒪(N) and move to Tentatively Ready GR: 23.2.1p2 says all complexity requirements are stated solely in terms of the number of operations on the contained object, so we shouldn't be stating complexity in terms of the hash table initialization HH: channeling Pete, there's an implicit "unless otherwise specified" everywhere. VV: seem to be requesting modifications that render this not Tentatively Ready GR: I think it can't be T/R AM: make the editorial recommendation, consider fixing 23.2.1/3 to give us permission to state complexity in terms of bucket initialization HH: only set it to Review after we get new wording to review[2015-02 Cologne]
Update wording, revisit later.
Previous resolution [SUPERSEDED]:
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)… Modify 24.5.4.2 [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 unspecified
impldefdefault 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 unspecified
impldefdefault 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).Modify 24.5.5.2 [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 unspecified
impldefdefault 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 unspecified
impldefdefault 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).Modify 24.5.6.2 [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 unspecified
impldefdefault 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 unspecified
impldefdefault 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).Modify 24.5.7.2 [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 unspecified
impldefdefault 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 unspecified
impldefdefault 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).
[2019-03-17; Daniel comments and provides revised wording]
The updated wording ensures that we can now specify complexity requirements for containers even when they are not expressed in terms of the number on the contained objects by an exception of the rule. This allows us to say that 𝒪(n) describes the complexity in terms of bucket initialization instead.
Proposed resolution:
This wording is relative to N4810.
Modify 24.2.2.1 [container.requirements.general] as indicated:
-2- Unless otherwise specified,
All of thecomplexity requirements in this Clause are stated solely in terms of the number of operations on the contained objects. [Example: The copy constructor of type vector<vector<int>> has linear complexity, even though the complexity of copying each contained vector<int> is itself linear. — end example]
Modify 24.2.8 [unord.req] as indicated:
-11- In Table 70:
(11.1) — […]
[…]
(11.23) — […]
(11.?) — Notwithstanding the complexity requirements restrictions of 24.2.2.1 [container.requirements.general], the complexity form 𝒪(n) describes the number of operations on buckets.
Modify the following rows in Table 70 — "Unordered associative container requirements" to add the explicit bucket allocation overhead of some constructions.
[Drafting note: It is kindly suggested to the Project Editor not to shorten the sum 𝒪(n) + 𝒪(N) to 𝒪(n + N), because two different work units are involved. — end drafting note]
Table 70 — Unordered associative container requirements (in addition to container) Expression Return type Assertion/note pre-/post-condition Complexity … X()
X a;X Expects: […]
Effects: Constructs an empty container with an unspecified number n of
buckets, using hasher() as the hash function and key_equal() as the key
equality predicate.constant𝒪(n)X(i, j, n, hf, eq)
X a(i, j, n, hf, eq)X Expects: […]
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) + 𝒪(N) (N
is distance(i, j)), worst case
𝒪(n) + 𝒪(N2)X(i, j, n, hf)
X a(i, j, n, hf)X Expects: […]
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) + 𝒪(N) (N
is distance(i, j)), worst case
𝒪(n) + 𝒪(N2)X(i, j, n)
X a(i, j, n)X Expects: […]
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) + 𝒪(N) (N
is distance(i, j)), worst case
𝒪(n) + 𝒪(N2)X(i, j)
X a(i, j)X Expects: […]
Effects: Constructs an empty container with an unspecified number n of
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) + 𝒪(N) (N
is distance(i, j)), worst case
𝒪(n) + 𝒪(N2)…
Modify 24.5.4.1 [unord.map.overview], class template unordered_map, as indicated:
// 24.5.4.2 [unord.map.cnstr], construct/copy/destroy […] template <class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); […] unordered_map(initializer_list<value_type> il, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); […]
Modify 24.5.4.2 [unord.map.cnstr] as indicated:
unordered_map() : unordered_map(size_type(see belowunspecified)) { } explicit unordered_map(size_type n, 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 predicate, and allocator, and using at least n buckets.
-?- Ensures: max_load_factor() == 1.0 -2- Complexity:For the default constructor, the number of buckets is implementation-defined.max_load_factor() returns 1.0.ConstantLinear in the number of buckets.template <class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); unordered_map(initializer_list<value_type> il, size_type n =see belowunspecified, 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 predicate, and allocator, and using at least n buckets.
-?- Ensures: max_load_factor() == 1.0 -4- Complexity:If n is not provided, the number of buckets is implementation-defined.Then inserts elements from the range [f, l) for the first form, or from the range [il.begin(), il.end()) for the second form.max_load_factor() returns 1.0.Average case linear, worst case quadraticLinear in the number of buckets, plus 𝒪(N) (average case) or 𝒪(N2) (worst case) where N is the number of insertions.
Modify 24.5.5.1 [unord.multimap.overview], class template unordered_multimap, as indicated:
// 24.5.5.2 [unord.multimap.cnstr], construct/copy/destroy […] template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); […] unordered_multimap(initializer_list<value_type> il, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); […]
Modify 24.5.5.2 [unord.multimap.cnstr] as indicated:
unordered_multimap() : unordered_multimap(size_type(see belowunspecified)) { } explicit unordered_multimap(size_type n, 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 predicate, and allocator, and using at least n buckets.
-?- Ensures: max_load_factor() == 1.0 -2- Complexity:For the default constructor, the number of buckets is implementation-defined. max_load_factor() returns 1.0.ConstantLinear in the number of buckets.template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); unordered_multimap(initializer_list<value_type> il, size_type n =see belowunspecified, 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 predicate, and allocator, and using at least n buckets.
-?- Ensures: max_load_factor() == 1.0 -4- Complexity:If n is not provided, the number of buckets is implementation-defined.Then inserts elements from the range [f, l) for the first form, or from the range [il.begin(), il.end()) for the second form.max_load_factor() returns 1.0.Average case linear, worst case quadraticLinear in the number of buckets, plus 𝒪(N) (average case) or 𝒪(N2) (worst case) where N is the number of insertions.
Modify 24.5.6.1 [unord.set.overview], class template unordered_set, as indicated:
// 24.5.6.2 [unord.set.cnstr], construct/copy/destroy […] template <class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); […] unordered_set(initializer_list<value_type> il, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); […]
Modify 24.5.6.2 [unord.set.cnstr] as indicated:
unordered_set() : unordered_set(size_type(see belowunspecified)) { } explicit unordered_set(size_type n, 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 predicate, and allocator, and using at least n buckets.
-?- Ensures: max_load_factor() == 1.0 -2- Complexity:For the default constructor, the number of buckets is implementation-defined.max_load_factor() returns 1.0.ConstantLinear in the number of buckets.template <class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); unordered_set(initializer_list<value_type> il, size_type n =see belowunspecified, 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 predicate, and allocator, and using at least n buckets.
-?- Ensures: max_load_factor() == 1.0 -4- Complexity:If n is not provided, the number of buckets is implementation-defined.Then inserts elements from the range [f, l) for the first form, or from the range [il.begin(), il.end()) for the second form.max_load_factor() returns 1.0.Average case linear, worst case quadraticLinear in the number of buckets, plus 𝒪(N) (average case) or 𝒪(N2) (worst case) where N is the number of insertions.
Modify 24.5.6.1 [unord.set.overview], class template unordered_multiset, as indicated:
// 24.5.7.2 [unord.multiset.cnstr], construct/copy/destroy […] template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); […] unordered_multiset(initializer_list<value_type> il, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); […]
Modify 24.5.7.2 [unord.multiset.cnstr] as indicated:
unordered_multiset() : unordered_multiset(size_type(see belowunspecified)) { } explicit unordered_multiset(size_type n, 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 predicate, and allocator, and using at least n buckets.
-?- Ensures: max_load_factor() == 1.0 -2- Complexity:For the default constructor, the number of buckets is implementation-defined.max_load_factor() returns 1.0.ConstantLinear in the number of buckets.template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n =see belowunspecified, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); unordered_multiset(initializer_list<value_type> il, size_type n =see belowunspecified, 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 predicate, and allocator, and using at least n buckets.
-?- Ensures: max_load_factor() == 1.0 -4- Complexity:If n is not provided, the number of buckets is implementation-defined.Then inserts elements from the range [f, l) for the first form, or from the range [il.begin(), il.end()) for the second form.max_load_factor() returns 1.0.Average case linear, worst case quadraticLinear in the number of buckets, plus 𝒪(N) (average case) or 𝒪(N2) (worst case) where N is the number of insertions.
Section: 25.3 [iterator.requirements] Status: Open Submitter: Daniel Krügler Opened: 2009-09-19 Last modified: 2016-01-28
Priority: 4
View all other issues in [iterator.requirements].
View all issues with Open status.
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 25.3 [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. 27.11 [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 99 [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
[2014-02-20 Re-open Deferred issues as Priority 4]
Consider to await the paper.
Proposed resolution:
Section: 27 [algorithms] Status: Open Submitter: Alisdair Meredith Opened: 2009-10-15 Last modified: 2020-09-06
Priority: 3
View other active issues in [algorithms].
View all other issues in [algorithms].
View all issues with Open status.
Discussion:
The library has many algorithms that take a source range represented by a pair of iterators, and the start of some second sequence given by a single iterator. Internally, these algorithms will produce undefined behaviour if the second 'range' is not as large as the input range, but none of the algorithms spell this out in Requires clauses, and there is no catch-all wording to cover this in clause 17 or the front matter of 25.
There was an attempt to provide such wording in paper n2944 but this seems incidental to the focus of the paper, and getting the wording of this issue right seems substantially more difficult than the simple approach taken in that paper. Such wording will be removed from an updated paper, and hopefully tracked via the LWG issues list instead.
It seems there are several classes of problems here and finding wording to solve all in one paragraph could be too much. I suspect we need several overlapping requirements that should cover the desired range of behaviours.
Motivating examples:
A good initial example is the swap_ranges algorithm. Here there is a clear requirement that first2 refers to the start of a valid range at least as long as the range [first1, last1). n2944 tries to solve this by positing a hypothetical last2 iterator that is implied by the signature, and requires distance(first2,last2) < distance(first1,last1). This mostly works, although I am uncomfortable assuming that last2 is clearly defined and well known without any description of how to obtain it (and I have no idea how to write that).
A second motivating example might be the copy algorithm. Specifically, let us image a call like:
copy(istream_iterator<int>(is),istream_iterator(),ostream_iterator<int>(os));
In this case, our input iterators are literally simple InputIterators, and the destination is a simple OutputIterator. In neither case am I happy referring to std::distance, in fact it is not possible for the ostream_iterator at all as it does not meet the requirements. However, any wording we provide must cover both cases. Perhaps we might deduce last2 == ostream_iterator<int>{}, but that might not always be valid for user-defined iterator types. I can well imagine an 'infinite range' that writes to /dev/null and has no meaningful last2.
The motivating example in n2944 is std::equal, and that seems to fall somewhere between the two.
Outlying examples might be partition_copy that takes two output iterators, and the _n algorithms where a range is specified by a specific number of iterations, rather than traditional iterator pair. We should also not accidentally apply inappropriate constraints to std::rotate which takes a third iterator that is not intended to be a separate range at all.
I suspect we want some wording similar to:
For algorithms that operate on ranges where the end iterator of the second range is not specified, the second range shall contain at least as many elements as the first.
I don't think this quite captures the intent yet though. I am not sure if 'range' is the right term here rather than sequence. More awkwardly, I am not convinced we can describe an Output sequence such as produce by an ostream_iterator as "containing elements", at least not as a precondition to the call before they have been written.
Another idea was to describe require that the trailing iterator support at least distance(input range) applications of operator++ and may be written through the same number of times if a mutable/output iterator.
We might also consider handling the case of an output range vs. an input range in separate paragraphs, if that simplifies how we describe some of these constraints.
[ 2009-11-03 Howard adds: ]
Moved to Tentatively NAD Future after 5 positive votes on c++std-lib.
[LEWG Kona 2017]
Recommend Open: The design is clear here; we just need wording
[2019-01-20 Reflector prioritization]
Set Priority to 3
Rationale:
Does not have sufficient support at this time. May wish to reconsider for a future standard.
Proposed resolution:
Section: 24.3.12 [vector.bool] Status: Open Submitter: BSI Opened: 2010-08-25 Last modified: 2020-09-06
Priority: 3
View other active issues in [vector.bool].
View all other issues in [vector.bool].
View all issues with Open status.
Discussion:
Addresses GB-118
vector<bool> iterators are not random access iterators because their reference type is a special class, and not bool &. All standard libary operations taking iterators should treat this iterator as if it was a random access iterator, rather than a simple input iterator.
[ Resolution proposed in ballot comment ]
Either revise the iterator requirements to support proxy iterators (restoring functionality that was lost when the Concept facility was removed) or add an extra paragraph to the vector<bool> specification requiring the library to treat vector<bool> iterators as-if they were random access iterators, despite having the wrong reference type.
[ Rapperswil Review ]
The consensus at Rapperswil is that it is too late for full support for proxy iterators, but requiring the library to respect vector<bool> iterators as-if they were random access would be preferable to flagging this container as deliberately incompatible with standard library algorithms.
Alisdair to write the note, which may become normative Remark depending on the preferences of the project editor.
[ Post-Rapperswil Alisdair provides wording ]
Initial wording is supplied, deliberately using Note in preference to Remark although the author notes his preference for Remark. The issue of whether iterator_traits<vector<bool>>::iterator_category is permitted to report random_access_iterator_tag or must report input_iterator_tag is not addressed.
[ Old Proposed Resolution: ]
Insert a new paragraph into 24.3.12 [vector.bool] between p4 and p5:
[Note All functions in the library that take a pair of iterators to denote a range shall treat vector<bool> iterators as-if they were random access iterators, even though the reference type is not a true reference.-- end note]
[ 2010-11 Batavia: ]
Closed as NAD Future, because the current iterator categories cannot correctly describe vector<bool>::iterator. But saying that they are Random Access Iterators is also incorrect, because it is not too hard to create a corresponding test that fails. We should deal with the more general proxy iterator problem in the future, and see no benefit to take a partial workaround specific to vector<bool> now.
[2017-02 in Kona, LEWG recommends NAD]
D0022 Proxy Iterators for the Ranges Extensions - as much a fix as we’re going to get for vector<bool>.
[2017-06-02 Issues Telecon]
P0022 is exploring a resolution. We consider this to be fairly important issue
Move to Open, set priority to 3
Proposed resolution:
Rationale:
No consensus to make this change at this time.
Section: 33.5.4 [atomics.order] Status: LEWG Submitter: Canada Opened: 2010-08-25 Last modified: 2016-01-28
Priority: Not Prioritized
View other active issues in [atomics.order].
View all other issues in [atomics.order].
View all issues with LEWG status.
Duplicate of: 1458
Discussion:
Addresses CA-21, GB-131
33.5.5 [atomics.lockfree] p.8 states:
An atomic store shall only store a value that has been computed from constants and program input values by a finite sequence of program evaluations, such that each evaluation observes the values of variables as computed by the last prior assignment in the sequence.
... but 6.9.1 [intro.execution] p.13 states:
If A is not sequenced before B and B is not sequenced before A, then A and B are unsequenced. [ Note: The execution of unsequenced evaluations can overlap. — end note ]
Overlapping executions can make it impossible to construct the sequence described in 33.5.5 [atomics.lockfree] p.8. We are not sure of the intention here and do not offer a suggestion for change, but note that 33.5.5 [atomics.lockfree] p.8 is the condition that prevents out-of-thin-air reads.
For an example, suppose we have a function invocation f(e1,e2). The evaluations of e1 and e2 can overlap. Suppose that the evaluation of e1 writes y and reads x whereas the evaluation of e2 reads y and writes x, with reads-from edges as below (all this is within a single thread).
e1 e2 Wrlx y-- --Wrlx x rf\ /rf X / \ Rrlx x<- ->Rrlx y
This seems like it should be allowed, but there seems to be no way to produce a sequence of evaluations with the property above.
In more detail, here the two evaluations, e1 and e2, are being executed as the arguments of a function and are consequently not sequenced-before each other. In practice we'd expect that they could overlap (as allowed by 6.9.1 [intro.execution] p.13), with the two writes taking effect before the two reads. However, if we have to construct a linear order of evaluations, as in 33.5.5 [atomics.lockfree] p.8, then the execution above is not permited. Is that really intended?
[ Resolution proposed by ballot comment ]
Please clarify.
[2011-03-09 Hans comments:]
I'm not proud of 33.5.4 [atomics.order] p9 (formerly p8), and I agree with the comments that this isn't entirely satisfactory. 33.5.4 [atomics.order] p9 was designed to preclude out-of-thin-air results for races among memory_order_relaxed atomics, in spite of the fact that Java experience has shown we don't really know how to do that adequately. In the long run, we probably want to revisit this.
However, in the short term, I'm still inclined to declare this NAD, for two separate reasons:6.9.1 [intro.execution] p15 states: "If a side effect on a scalar object is unsequenced relative to either another side effect on the same scalar object or a value computation using the value of the same scalar object, the behavior is undefined." I think the examples presented here have undefined behavior as a result. It's not completely clear to me whether examples can be constructed that exhibit this problem, and don't have undefined behavior.
This comment seems to be using a different meaning of "evaluation" from what is used elsewhere in the standard. The sequence of evaluations here doesn't have to consist of full expression evaluations. They can be evaluations of operations like lvalue to rvalue conversion, or individual assignments. In particular, the reads and writes executed by e1 and e2 in the example could be treated as separate evaluations for purposes of producing the sequence. The definition of "sequenced before" in 6.9.1 [intro.execution] makes little sense if the term "evaluation" is restricted to any notion of complete expression. Perhaps we should add yet another note to clarify this? 33.5.4 [atomics.order] p10 probably leads to the wrong impression here.
An alternative resolution would be to simply delete our flakey attempt at preventing out-of-thin-air reads, by removing 33.5.4 [atomics.order] p9-11, possibly adding a note that explains that we technically allow, but strongly discourage them. If we were starting this from scratch now, that would probably be my preference. But it seems like too drastic a resolution at this stage.[2011-03-24 Madrid]
Moved to NAD Future
Proposed resolution:
Section: 33.4.3 [thread.thread.class] Status: LEWG Submitter: INCITS Opened: 2010-08-25 Last modified: 2017-03-01
Priority: Not Prioritized
View all issues with LEWG status.
Discussion:
Addresses US-183
There is no way to join a thread with a timeout.
[ Resolution proposed by ballot comment: ]
Add join_for and join_until. Or decide one should never join a thread with a timeout since pthread_join doesn't have a timeout version.
[ 2010 Batavia ]
The concurrency working group deemed this an extension beyond the scope of C++0x.
Rationale:
The LWG does not wish to make a change at this time.
[2017-03-01, Kona]
SG1 recommends: Close as NAD
There has not been much demand for it, and it would usually be difficult to deal with thread_local destructor races. It can be approximated with a condition variable wait followed by an unconditional join. Adding it would create implementation issues on Posix. As always, this may be revisited if we have a paper exploring the issues in detail.Proposed resolution:
Section: 33.6 [thread.mutex] Status: LEWG Submitter: INCITS Opened: 2010-08-25 Last modified: 2017-03-01
Priority: Not Prioritized
View all other issues in [thread.mutex].
View all issues with LEWG status.
Discussion:
Addresses US-185
Cooperate with WG14 to improve interoperability between the C++0x and C1x threads APIs. In particular, C1x mutexes should be conveniently usable with a C++0x lock_guard. Performance overheads for this combination should be considered.
[ Resolution proposed by ballot comment: ]
Remove C++0x timed_mutex and timed_recursive_mutex if that facilitates development of more compatible APIs.
[ 2010 Batavia ]
The concurrency sub-group reviewed the options, and decided that closer harmony should wait until both standards are published.
Rationale:
The LWG does not wish to make any change at this time.
[2017-03-01, Kona]
SG1 recommends: Close as NAD
Papers about C compatibility are welcome, but there may be more pressing issues. C threads are not consistently available at this point, so there seems to be little demand to fix this particular problem.Proposed resolution:
Section: 33.6.4 [thread.mutex.requirements] Status: LEWG Submitter: INCITS Opened: 2010-08-25 Last modified: 2017-03-01
Priority: Not Prioritized
View other active issues in [thread.mutex.requirements].
View all other issues in [thread.mutex.requirements].
View all issues with LEWG status.
Discussion:
Addresses US-189
mutex and recursive_mutex should have an is_locked() member function. is_locked allows a user to test a lock without acquiring it and can be used to implement a lightweight try_try_lock.
[ Resolution proposed by ballot comment: ]
Add a member function:
bool is_locked() const;to std::mutex and std::recursive_mutex. These functions return true if the current thread would not be able to obtain a mutex. These functions do not synchronize with anything (and, thus, can avoid a memory fence).
[ 2010 Batavia ]
The Concurrency subgroup reviewed this issue and deemed it to be an extension to be handled after publishing C++0x.
Rationale:
The LWG does not wish to make a change at this time.
[2017-03-01, Kona]
SG1 recommends: Close as NAD
Several participants voiced strong objections, based on either memory model issues or lock elision. No support. It is already possible to write a wrapper that explicitly tracks ownership for testing in the owning thread, which may have been part of the intent here.Proposed resolution:
Section: 24.2.2.1 [container.requirements.general] Status: Open Submitter: Mike Spertus Opened: 2010-10-16 Last modified: 2019-01-20
Priority: 3
View other active issues in [container.requirements.general].
View all other issues in [container.requirements.general].
View all issues with Open status.
Discussion:
Addresses US-104, US-141
The standard doesn't say that containers should use abstract pointer types internally. Both Howard and Pablo agree that this is the intent. Further, it is necessary for containers to be stored, for example, in shared memory with an interprocess allocator (the type of scenario that allocators are intended to support).
In spite of the (possible) agreement on intent, it is necessary to make this explicit:
An implementations may like to store the result of dereferencing the pointer (which is a raw reference) as an optimization, but that prevents the data structure from being put in shared memory, etc. In fact, a container could store raw references to the allocator, which would be a little weird but conforming as long as it has one by-value copy. Furthermore, pointers to locales, ctypes, etc. may be there, which also prevents the data structure from being put in shared memory, so we should make explicit that a container does not store raw pointers or references at all.
[ Pre-batavia ]
This issue is being opened as part of the response to NB comments US-104/141. See paper N3171 in the pre-Batavia mailing.
[2011-03-23 Madrid meeting]
Deferred
[ 2011 Batavia ]
This may be an issue, but it is not clear. We want to gain a few years experience with the C++11 allocator model to see if this is already implied by the existing specification.
[LEWG Kona 2017]
Status to Open: Acknowledged, need wording: (N4618 numbering) 23.2.1 container.requirements.general p8 first sentence. Replace non-normative note with requirement.
See discussion on LEWG Wiki
[2019-01-20 Reflector prioritization]
Set Priority to 3
Proposed resolution:
Add to the end of 24.2.2.1 [container.requirements.general] p. 8:
[..] 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. The container may not store internal objects whose types are of the form T * or T & except insofar as they are part of the item type or members.
Section: 25.3.5.4 [output.iterators] Status: Open Submitter: Daniel Krügler Opened: 2011-02-27 Last modified: 2016-01-28
Priority: 3
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:
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 25.3.1 [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.
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.[2015-02 Cologne]
The matter is complicated, Daniel volunteers to write a paper.
Proposed resolution:
Add a reference to 25.3.1 [iterator.requirements.general] to the following parts of the library preceding Clause 24 Iterators library: (I stopped from 24.2.8 [unord.req] on, because the remaining references are the concrete containers)
16.4.4.3 [swappable.requirements] p5:
-5- A type X satisfying any of the iterator requirements (24.2) is ValueSwappable if, for any dereferenceable (25.3.1 [iterator.requirements.general]) object x of type X, *x is swappable.
16.4.4.6 [allocator.requirements], Table 27 — "Descriptive variable definitions", row with the expression c:
a dereferenceable (25.3.1 [iterator.requirements.general]) pointer of type C*
20.2.3.3 [pointer.traits.functions]:
Returns: The first template function returns a dereferenceable (25.3.1 [iterator.requirements.general]) pointer to r obtained by calling Ptr::pointer_to(r); […]
23.4.3.4 [string.iterators] p. 2:
Returns: An iterator which is the past-the-end value (25.3.1 [iterator.requirements.general]).
30.4.6.2.3 [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 (25.3.1 [iterator.requirements.general]).
24.2.2.1 [container.requirements.general] p. 6:
[…] end() returns an iterator which is the past-the-end (25.3.1 [iterator.requirements.general]) value for the container. […]
24.2.4 [sequence.reqmts] p. 3:
[…] q denotes a valid dereferenceable (25.3.1 [iterator.requirements.general]) const iterator to a, […]
24.2.7 [associative.reqmts] p. 8 (I omit intentionally one further reference in the same sub-clause):
[…] q denotes a valid dereferenceable (25.3.1 [iterator.requirements.general]) const iterator to a, […]
24.2.8 [unord.req] p. 10 (I omit intentionally one further reference in the same sub-clause):
[…] q and q1 are valid dereferenceable (25.3.1 [iterator.requirements.general]) const iterators to a, […]
Edit 25.3.1 [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. […]
Modify the column contents of Table 106 — "Iterator requirements", 25.3.5.2 [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.
Modify the column contents of Table 107 — "Input iterator requirements", 25.3.5.3 [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 specification — end rationale]:
Table 107 — Input iterator requirements (in addition to Iterator) Expression Return type Operational semantics Assertion/note
pre-/post-conditiona != 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; }
Modify the column contents of Table 108 — "Output iterator requirements", 25.3.5.4 [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 specification — end 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 operationRemark: After this operation
r is not required to be
dereferenceable.
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-endincrementable.
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.
Modify the column contents of Table 109 — "Forward iterator requirements", 25.3.5.5 [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-conditionr++ convertible to const X& { X tmp = r;
++r;
return tmp; }*r++ reference { reference tmp = *r;
++r;
return tmp; }
Modify the column contents of Table 110 — "Bidirectional iterator requirements", 25.3.5.6 [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 isdereferenceableincrementable.
--(++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; }
Section: 25.3.5.4 [output.iterators] Status: Open Submitter: Pete Becker Opened: 2011-02-27 Last modified: 2016-01-28
Priority: 3
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 16.4.4.3 [swappable.requirements] there is a mention of "dereferenceable object"; in 16.4.4.6 [allocator.requirements] the table of Descriptive variable definitions refers to a "dereferenceable pointer"; 20.2.3.3 [pointer.traits.functions] refers to a "dereferenceable pointer"; in 30.4.6.2.3 [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:
Section: 21.3.5.4 [meta.unary.prop] Status: Open Submitter: Daniel Krügler Opened: 2011-08-20 Last modified: 2016-01-28
Priority: 3
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:
Section: 17.10.5 [exception.terminate] Status: Open Submitter: Daniel Krügler Opened: 2011-09-25 Last modified: 2016-01-28
Priority: 3
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 17.10.5.4 [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:
- In case std::thread is assigned to or destroyed while being joinable there is no throw-expression involved.
- In case std::unexpected is called by the program, std::terminate is triggered by the implementation - no throw-expression involved.
- In case a destructor throws during stack unwinding we have two throw-expressions involved.
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, 17.10.5.1 [terminate.handler], paragraph 2 saysRequired 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 17.10.5.4 [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 17.10.5.1 [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:
Section: 33.10.6 [futures.promise], 33.10.10 [futures.task] Status: LEWG Submitter: Jonathan Wakely Opened: 2011-11-01 Last modified: 2019-06-03
Priority: 4
View other active issues in [futures.promise].
View all other issues in [futures.promise].
View all issues with LEWG 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{} };
[2012, Portland]
This is an allocator issue, and should be dealt with directly by LWG.
[2013-03-06]
Jonathan suggests to make the new constructors non-explicit and makes some representational improvements.
[2013-09 Chicago]
Move to deferred.
This issue has much in common with similar problems with std::function that are being addressed by the polymorphic allocators proposal currently under evaluation in LEWG. Defer further discussion on this topic until the final outcome of that paper and its proposed resolution is known.
[2014-02-20 Re-open Deferred issues as Priority 4]
[2016-08 Chicago]
Fri PM: Send to LEWG - and this also applies to function in LFTS.
[2019-06-03 Jonathan Wakely provides updated wording]
Jonathan updates wording post-2976 and observes that this resolution conflicts with 3003.
Previous resolution [SUPERSEDED]:
[This wording is relative to the FDIS.]
Add to 33.10.6 [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(); […] }; […] }Change 33.10.6 [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]Add to 33.10.10 [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 Allocator> packaged_task(allocator_arg_t, const Allocator& a) noexcept; template <class F> explicit packaged_task(F&& f); template <class F, class Allocator> explicit packaged_task(allocator_arg_t, const Allocator& a, F&& f); ~packaged_task(); // no copy packaged_task(const packaged_task&) = delete; template<class Allocator> packaged_task(allocator_arg_t, const Allocator& a, const packaged_task&) = delete; packaged_task& operator=(const packaged_task&) = delete; // move support packaged_task(packaged_task&& rhs) noexcept; template <class Allocator> packaged_task(allocator_arg_t, const Allocator& a, packaged_task&& rhs) noexcept; packaged_task& operator=(packaged_task&& rhs) noexcept; void swap(packaged_task& other) noexcept; […] }; […] }Change 33.10.10.2 [futures.task.members] as indicated:
packaged_task() noexcept; template <class Allocator> 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> 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]
Proposed resolution:
[This wording is relative to N4810.]
Add to 33.10.6 [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(); […] }; […] }
Change 33.10.6 [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]
Section: 28.6.8 [template.mask.array] Status: Open Submitter: Thomas Plum Opened: 2011-12-10 Last modified: 2016-01-28
Priority: 4
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 28.6.8 [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, 28.6.8.1 [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 HERErefer 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
[....]
This template is a helper template used by the mask subscript operator: mask_array<T> valarray<T>::operator[](const valarray<bool>&).
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:
Section: 21.3.5.4 [meta.unary.prop] Status: Tentatively NAD Submitter: Dave Abrahams Opened: 2011-12-09 Last modified: 2022-10-10
Priority: 3
View other active issues in [meta.unary.prop].
View all other issues in [meta.unary.prop].
View all issues with Tentatively NAD 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:A defaulted copy/move constructor for a class X is defined as deleted (9.5.3 [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 destructor 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.
[2017-01-27 Telecon]
Gave the issue a better title
This issue interacts with 2827
Ville would like "an evolution group" to take a look at this issue.
[2020-08; LWG reflector]
A poll was taken to close the issue as NAD, but only gained three votes in favour (and one vote against, which was subsequently withdrawn).
[2022-03; LWG reflector]
A poll was taken to close the issue as NAD, with six votes in favour. (and one vote against, subsequently withdrawn).
"Write a paper if you want something else. These traits have well established meaning now." "Minimizing requirements is not as important a concern for standard library concepts as as minimizing the number of concepts. Requirements like 'I need to construct but not destroy an object' are niche enough that we don't need to support them."
Proposed resolution:
Section: 30.4.3.3.3 [facet.num.put.virtuals], 99 [ios::fmtflags], 31.5.5.1 [fmtflags.manip] Status: Open Submitter: Benjamin Kosnik Opened: 2011-12-15 Last modified: 2016-04-16
Priority: 3
View other active issues in [facet.num.put.virtuals].
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)
[Issaquah 2014-02-10-12: Move to LEWG]
Since this issue was filed, we have grown a new working group that is better placed to handle feature requests.
We will track such issues with an LEWG status until we get feedback from the Library Evolution Working Group.
[Issaquah 2014-02-12: LEWG discussion]
SF | F | N | A | SA |
2 | 4 | 1 | 0 | 0 |
Think about the ABI break for adding a flag. But this could be mitigated by putting the data into an iword instead of a flag.
This needs to change Stage 2 in [facet.num.put.virtuals].
Previous resolution, which needs the above corrections:
This wording is relative to the FDIS.
Insert in 30.4.3.3.3 [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));Change header <ios> synopsis, [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); […] }Change class ios_base synopsis, 31.5.2 [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 ; […] }; }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 […] After 99 [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.
Proposed resolution:
Section: 16.3.2 [structure] Status: Open Submitter: Jens Maurer Opened: 2012-03-08 Last modified: 2021-06-25
Priority: 3
View all issues with Open 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.
[ 2012-10 Portland: Move to Open ]
Consensus that we do not clearly say this, and that we probably should. A likely location to describe the guarantees of postconditions could well be a new sub-clause following 99 [res.on.required] which serves the same purpose for requires clauses. However, we need such wording before we can make progress.
Also, see 2137 for a suggestion that we want to see a paper resolving both issues together.
[2015-05-06 Lenexa: EirkWF to write paper addressing 2136 and 2137]
MC: Idea is to replace all such "If no exception" postconditions with "Exception safety" sections.
[2021-06-20; Daniel comments]
An informal editorial change suggestion has recently been made whose editorial implementation would promote the idea that the default assumption is that Postconditions: are only met if the function doesn't exit with an exception.
After analyzing all current existing Postconditions: elements the following seems to hold: Affected by this issue are only non-noexcept functions and mostly non-constructor functions (unless the Postconditions: element says something about the value of its arguments). Most existing Postconditions seem to be intended to apply only in non-exceptional cases. I found some where this is presumably not intended, namely those of the expressions os << x and is >> v in Tables [tab:rand.req.eng] and [tab:rand.req.dist], maybe also 29.11.2.4 [time.zone.db.remote] p4. Nonetheless, the editorial change seems to be applicable even without having this issue resolved, because it doesn't actually change the normative state by itself.Proposed resolution:
Section: 32.7.3 [re.regex.assign] Status: Open Submitter: Jonathan Wakely Opened: 2012-03-08 Last modified: 2021-06-20
Priority: 3
View all other issues in [re.regex.assign].
View all issues with Open status.
Discussion:
The post-conditions of basic_regex<>::assign 32.7.3 [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.
[ 2012-10 Portland: Move to Open ]
A favorable resolution clearly depends on a favorable resolution to 2136. There is also a concern that this is just one example of where we would want to apply such a wording clean-up, and which is really needed to resolve both this issue and 2136 is a paper providing the clause 17 wording that gives the guarantee for postcondition paragraphs, and then reviews clauses 18-30 to apply that guarantee consistently. We do not want to pick up these issues piecemeal, as we risk opening many issues in an ongoing process.
[2015-05-06 Lenexa: EricWF to write paper addressing 2136 and 2137]
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.
Section: 16.4.4.2 [utility.arg.requirements] Status: Open Submitter: Nikolay Ivchenkov Opened: 2012-03-23 Last modified: 2020-10-02
Priority: 3
View all other issues in [utility.arg.requirements].
View all issues with Open status.
Discussion:
According to 16.4.4.2 [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?
#include <iostream>
#include <utility>
int main()
{
int x = 1;
int y = 2;
std::swap<int &&>(x, y); // undefined?
std::cout << x << " " << y << std::endl;
}
#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?
[2013-03-15 Issues Teleconference]
Moved to Open.
The questions raised by the issue are real, and should have a clear answer.
[2015-10, Kona Saturday afternoon]
STL: std::thread needs to be fixed, and anything behaving like it needs to be fixed, rather than reference types. std::bind gets this right. We need to survey this. GR: That doesn't sound small to me. STL: Seach for CopyConstructible etc. It may be a long change, but not a hard one.
MC: It seems that we don't have a PR. Does anyone have one? Is anyone interested in doing a survey?
[2016-03, Jacksonville]
Casey volunteers to make a survey
[2016-06, Oulu]
During an independent survey performed by Daniel as part of the analysis of LWG 2716, some overlap was found between these two issues. Daniel suggested to take responsibility for surveying LWG 2146 and opined that the P/R of LWG 2716 should restrict to forwarding references, where the deduction to lvalue references can happen without providing an explicit template argument just by providing an lvalue function argument.
[2018-06, Rapperwsil]
Jonathan says that this will be covered by his Omnibus requirements paper.
[2019 Cologne Wednesday night]
Daniel will start working on this again; Marshall to provide rationale why some of the examples are not well-formed.
[2020-10-02; Issue processing telecon: change from P2 to P3]
For the examples given in the original report, the for_each case is now banned, because 27.2 [algorithms.requirements] p15 forbids explicit template argument lists. std::thread's constructor has also been fixed to describe the requirements on decay_t<T> instead of T.
We believe we're more careful these days about using remove_cvref or decay as needed, but there are still places where we incorrectly state requirements in terms of types that might be references. The swap case still needs solving. Still need a survey.
Proposed resolution:
Section: 16.4.4.3 [swappable.requirements], 24.2.2.1 [container.requirements.general] Status: LEWG Submitter: Robert Shearer Opened: 2012-04-13 Last modified: 2020-09-06
Priority: 3
View all other issues in [swappable.requirements].
View all issues with LEWG status.
Discussion:
Sub-clause 16.4.4.3 [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 24.2.2.1 [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 22.2.2 [utility.swap] p4, 22.3.2 [pairs.pair] p31, 22.4.4.3 [tuple.swap] p2, 27.7.3 [alg.swap] p2, and 27.7.3 [alg.swap] p6, which use the binary definition). This renders many of the mutating sequence algorithms of sub-clause 27.7 [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:I favor the latter solution, for reasons detailed in the following issue.
[ 2012-10 Portland: Move to Open ]
The issue is broader than containers with stateful allocotors, although they are the most obvious example contained within the standard itself. The basic problem is that once you have a stateful allocator, that does not propagate_on_swap, then whether two objects of this type can be swapped with well defined behavior is a run-time property (the allocators compare equal) rather than a simple compile-time property that can be deduced from the type. Strictly speaking, any type where the nature of swap is a runtime property does not meet the swappable requirements of C++11, although typical sequences of such types are going to have elements that are all swappable with any other element in the sequence (using our other term of art for specifying requirements) as the common case is a container of elements who all share the same allocator.
The heart of the problem is that the swappable requirments demand that any two objects of the same type be swappable with each other, so if any two such objects would not be swappable with each other, then the whole type is never swappable. Many algorithms in clause 25 are specified in terms of swappable which is essentially an overspecification as all they actually need is that any element in the sequence is swappable with any other element in the sequence.
At this point Howard joins the discussion and points out that the intent of introducing the two swap-related terms was to support vector<bool>::reference types, and we are reading something into the wording that was never intended. Consuses is that regardless of the intent, that is what the words today say.
There is some support to see a paper reviewing the whole of clause 25 for this issue, and other select clauses as may be necessary.
There was some consideration to introducing a note into the front of clause 25 to indicate swappable requirements in the clause should be interpreted to allow such awkward types, but ultimately no real enthusiasm for introducing a swappable for clause 25 requirement term, especially if it confusingly had the same name as a term used with a subtly different meaning through the rest of the standard.
There was no enthusiasm for the alternate resolution of requiring containers with unequal allocators that do not propagate provide a well-defined swap behavior, as it is not believed to be possible without giving swap linear complexity for such values, and even then would require adding the constraint that the container element types are CopyConstructible.
Final conclusion: move to open pending a paper from a party with a strong interest in stateful allocators.
[2016-03 Jacksonville]
Alisdair says that his paper P0178 addresses this.
[2016-06 Oulu]
P0178 reviewed, and sent back to LEWG for confirmation.
Thursday Morning: A joint LWG/LEWG meeting declined to adopt P0178.
[2017-02 in Kona, LEWG responds]
Note in the issue that this is tracked here
[2017-06-02 Issues Telecon]
Leave as LEWG; priority 3
Proposed resolution:
Apply P0178.
Section: 22.2.2 [utility.swap], 16.4.4.3 [swappable.requirements], 24.2.2.1 [container.requirements.general] Status: LEWG Submitter: Robert Shearer Opened: 2012-04-13 Last modified: 2020-10-02
Priority: 2
View all other issues in [utility.swap].
View all issues with LEWG status.
Discussion:
Sub-clause 22.2.2 [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 16.4.4.3 [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 16.4.4.3 [swappable.requirements] is as a guide to the "right" semantics to provide for a non-member swap function (called in the context defined by 16.4.4.3 [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 22.2.2 [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.[2013-03-15 Issues Teleconference]
Moved to Open.
This topic deserves more attention than can be given in the telecon, and there is no proposed resolution.
[2016-03 Jacksonville]
Alisdair says that his paper P0178 addresses this.
[2016-08 Chicago]
Send to LEWG
[2016-06 Oulu]
P0178 reviewed, and sent back to LEWG for confirmation.
Thursday Morning: A joint LWG/LEWG meeting declined to adopt P0178.
[2020-10-02; remove P0178 as Proposed Resolution]
Proposed resolution:
Section: 24.3.7.5 [array.zero] Status: Open Submitter: Daryle Walker Opened: 2012-05-08 Last modified: 2021-03-14
Priority: 3
View all other issues in [array.zero].
View all issues with Open 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 24.3.7.5 [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.
[2012, Portland: Move to Open]
Some discussion to understand the issue, which is that implementations currently have freedom to implement an empty array by holding a dummy element, and so might not support value initialization, which is surprising when trying to construct an empty container. However, this is not mandated, it is an unspecified implementation detail.
Jeffrey points out that the implication of 24.3.7.1 [array.overview] is that this initialization syntax must be supported by empty array objects already. This is a surprising inference that was not obvious to the room, but consensus is that the reading is accurate, so the proposed resolution is not necessary, although the increased clarity may be useful.
Further observation is that the same clause effectively implies that T must always be DefaultConstructible, regardless of N for the same reasons - as an initializer-list may not supply enough values, and the remaining elements must all be value initialized.
Concern that we are dancing angels on the head of pin, and that relying on such subtle implications in wording is not helpful. We need a clarification of the text in this area, and await wording.
[2015-02 Cologne]
DK: What was the outcome of Portland? AM: Initially we thought we already had the intended behaviour. We concluded that T must always be DefaultConstructible, but I'm not sure why. GR: It's p2 in std::array, "up to N". AM: That wording already implies that "{}" has to work when N is zero. But the wording of p2 needs to be fixed to make clear that it does not imply that T must be DefaultConstructible.
Conclusion: Update wording, revisit later.[2015-10, Kona Saturday afternoon]
MC: How important is this? Can you not just use default construction for empty arrays?
TK: It needs to degenerate properly from a pack. STL agrees.
JW: Yes, this is important, and we have to make it work.
MC: I hate the words "initialization like".
JW: I'll reword this.
WEB: Can I ask that once JW has reworded this we move it to Review rather than Open?
MC: We'll try to review it in a telecon and hopefully get it to tentatively ready.
STL: Double braces must also work: array<T, 0> a = {{}};.
Jonathan to reword.
[2018-03-14 Wednesday evening issues processing]
Jens suggested that we remove the requirement that begin() == end() == unique-value, specifically the unique value part.
Previous resolution [SUPERSEDED]:
This wording is relative to N3376.
Add the following new paragraph between the current 24.3.7.5 [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).
[2018-06-14, Jonathan Wakely comments and provides revised wording]
The new wording does not address the 2018-03-14 suggestion from Jens to remove the unique value. It wasn't clear to me that there was consensus to make that change, and it would be a change in behaviour not just a clarification of the existing wording.
Previous resolution [SUPERSEDED]:
This wording is relative to N4750.
Modify 24.3.7.5 [array.zero] as indicated:
-1- array
-?- A zero-sized array type is an aggregate that meets the DefaultConstructible (Table 22) and CopyConstructible (Table 24) requirements. There is a single element of the aggregate, of an unspecified DefaultConstructible type. [Note: This allows initialization of the form array<T, 0> a = {{}};. There is no requirement for T to be DefaultConstructible. — end note] -2-shallprovides support for the special case of a zero-sized array that is always empty, i.e. N == 0, with the properties described in this subclause.In the case that N == 0, begin() == end() == unique valuebegin() and end() return non-dereferenceable iterators such that begin() == end() and a.begin() != b.begin() where a and b are distinct objects of the same zero-sized array type. 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 havehas constant complexity and a non-throwing exception specification.
[2018-08-30, Jonathan revises wording following feedback from Daniel Kruegler and Tim Song.]
Daniel noted that it's undefined to compare iterators from different containers,
so a.begin() != b.begin() can't be used. That means whether the iterators
from different containers are unique is unobservable anyway.
We can say they don't share the same underlying sequence, which tells users they can't compare them
and tells implementors they can't return value-initialized iterators.
Tim noted that it's not sufficient to say the unspecified type in a zero-sized array is DefaultConstructible,
it also needs to be constructible from = {}. Also, a zero-sized array should be CopyAssignable.
Previous resolution [SUPERSEDED]:
This wording is relative to N4762.
Modify 24.3.7.5 [array.zero] as indicated:
-1- array
-?- A zero-sized array type is an aggregate that meets the Cpp17DefaultConstructible (Table 24) and Cpp17CopyConstructible (Table 26) and Cpp17CopyAssignable (Table 28) requirements. There is a single element of the aggregate, of an unspecified Cpp17DefaultConstructible type that is copy-list-initializable from an empty list. [Note: This allows initialization of the form array<T, 0> a = {{}};. There is no requirement for T to be Cpp17DefaultConstructible. — end note] -2-shallprovides support for the special case of a zero-sized array that is always empty, i.e. N == 0, with the properties described in this subclause.In the case that N == 0, begin() == end() == unique valuebegin() and end() return non-dereferenceable iterators such that begin() == end(). When a and b are distinct objects of the same zero-sized array type, a.begin() and b.begin() are not iterators over the same underlying sequence. [Note: Therefore begin() does not return a value-initialized iterator — end note]. 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 havehas constant complexity and a non-throwing exception specification.
[2021-03-14; Johel Ernesto Guerrero Peña comments and provides improved wording]
The currently proposed wording specifies:
There is a single element of the aggregate, of an unspecified Cpp17DefaultConstructible type that is copy-list-initializable from an empty list.
This doesn't specify which expressions involving zero-sized array specializations are constant expressions. 24.3.7.1 [array.overview] p4 specifies array<T, 0> to be a structural type when T is a structural type. This requires that its single element, let's call it single-element, be a structural type. But that says nothing about which of the special member functions of single-element are constant expressions. By being a structural type, single-element is permitted to be implemented as a literal class type. To meet this requirement, single-element can be implemented to have one constexpr constructor that is not a copy or move constructor (6.8.1 [basic.types.general] p10), so its default constructor needn't be constexpr. This is unlike non-zero-sized array specializations, which inherit these properties from T. Furthermore, this permits an implementation of single-element whose default constructor stores the result of std::source_location::current() in a data member (as exemplified in the specification for current). Cpp17DefaultConstructible doesn't require the default constructor to produce equal values. The simplest way to solve these issues and any other that might arise from future changes and oversights would be to specify single-element as an empty aggregate type. Then the wording from 24.3.7.2 [array.cons] p1 makes it clear that all the special member functions are constant expressions. It would also mean that the default constructor produces template-argument-equivalent values.
Proposed resolution:
This wording is relative to N4878.
Modify 24.3.7.5 [array.zero] as indicated:
-1- array
-?- A zero-sized array type is an aggregate that meets the Cpp17DefaultConstructible (Table 29 [tab:cpp17.defaultconstructible]) and Cpp17CopyConstructible (Table 31 [tab:cpp17.copyconstructible]) and Cpp17CopyAssignable (Table 33 [tab:cpp17.copyassignable]) requirements. There is a single element of the aggregate, of an unspecified empty aggregate type. [Note: This allows initialization of the form array<T, 0> a = {{}};. There is no requirement for T to be Cpp17DefaultConstructible. — end note] -2-shallprovides support for the special case of a zero-sized array that is always empty, i.e. N == 0, with the properties described in this subclause.In the case that N == 0, begin() == end() == unique valuebegin() and end() return non-dereferenceable iterators such that begin() == end(). When a and b are distinct objects of the same zero-sized array type, a.begin() and b.begin() are not iterators over the same underlying sequence. [Note: Therefore begin() does not return a value-initialized iterator — end note].. 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 havehas constant complexity and a non-throwing exception specification.
Section: 24.3.11.3 [vector.capacity] Status: Open Submitter: Nikolay Ivchenkov Opened: 2012-05-08 Last modified: 2022-11-06
Priority: 3
View other active issues in [vector.capacity].
View all other issues in [vector.capacity].
View all issues with Open 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 – 24.3.11.3 [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:
if !is_nothrow_move_constructible<T>::value && is_copy_constructible<T>::value then T shall be CopyInsertable into *this;
otherwise T shall be MoveInsertable into *this.
Maybe it would be useful to introduce a new name for such conditional requirement (in addition to "CopyInsertable" and "MoveInsertable").
[2016-08 Chicago]
The problem does not appear to be as severe as described. The MoveInsertable requirements are consistently correct, but an issue may arise on the exception-safety guarantees when we check for is_copy_constructible_v<T>. The problem, as described, is typically for templates that appear to have a copy constructor, but one that fails to compile once instantiated, and so gives a misleading result for the trait.
In general, users should not provide such types, and the standard would not serve users well by trying to address support for such types. However, the standard should not be providing such types either, such as vector<unique_ptr<T>>. A possible resolution would be to tighten the constraints in Table 80 — Container Requirements, so that if the Requirements for the copy constructor/assingment operator of a container are not satisfied, that operation shall be deleted.
A futher problem highlighted by this approach is that there are no constraints on the copy-assignment operator, so that vector<unique_ptr<T>> should be CopyAssignable! However, we can lift the equivalent constraints from the Allocator-aware container requirements.
[08-2016, Chicago]
Fri PM: Move to Open
[2017-11 Albuquerque Saturday issues processing]
There's a bunch of uses of "shall" here that are incorrect. Also, CopyInsertable contains some semantic requirements, which can't be checked at compile time, so 'ill-formed' is not possible for detecting that.
[2018-06 Rapperswil Wednesday issues processing]
Daniel to provide updated wording.
[2018-06-12, Daniel provides revised wording]
Previous resolution [SUPERSEDED]:
This wording is relative to N4606.
24.2.2.1 [container.requirements.general] Table 80 — Container requirements Expression Return type Operational semantics Assertion/note/pre-/post-condition Complexity X(a) Requires: T is CopyInsertable into X (see below) ., otherwise this expression shall be ill-formed.
post: a == X(a).linear X u(a)
X u = a;Requires: T is CopyInsertable into X (see below) ., otherwise this expression shall be ill-formed.
post: u == a.linear ... ... ... ... ... r = a X& Requires: T is CopyInsertable into X and CopyAssignable, otherwise this expression shall be ill-formed.
post: r == a.linear
24.2.2.1 [container.requirements.general] Table 83 — Allocator-aware container requirements Expression Return type Operational semantics Assertion/note/pre-/post-condition Complexity a = t X& Requires: T is CopyInsertable into X and CopyAssignable ., otherwise this expression shall be ill-formed
post: r == a.linear
[2018-08-23 Batavia Issues processing. Priority to 3]
Changed CopyInsertable -> Cpp17CopyInsertable in the resolution.
Tim says that the wording is not quite right — it imposes additional requirements.
[2022-11-06; Daniel comments]
This issue has considerable overlap with LWG 3758.
Proposed resolution:
This wording is relative to N4750.
The revised wording below uses the new Mandates: element introduced by adopting P0788R3 at the Rapperswil meeting 2018 and which will become a new term of art with Jonathan's omnibus paper throughout the Standard Library.
Expression | Return type | Operational semantics | Assertion/note/pre-/post-condition | Complexity |
X(a) |
Mandates: Syntactic requirements of T is Cpp17CopyInsertable into X (see below). Requires: T is Cpp17CopyInsertable into X post: a == X(a). |
linear | ||
X u(a) X u = a; |
Mandates: Syntactic requirements of T is Cpp17CopyInsertable into X (see below). Requires: T is Cpp17CopyInsertable into X post: u == a. |
linear | ||
... | ... | ... | ... | ... |
r = a | X& |
Mandates: Syntactic requirements of T is Cpp17CopyInsertable into X (see below) and CopyAssignable. Requires: T is Cpp17CopyInsertable into X and CopyAssignable. post: r == a. |
linear |
Expression | Return type | Operational semantics | Assertion/note/pre-/post-condition | Complexity |
a = t | X& |
Mandates: Syntactic requirements of T is Cpp17CopyInsertable into X and CopyAssignable. Requires: T is Cpp17CopyInsertable into X and CopyAssignable. post: r == a. |
linear |
operator +
in the description of the algorithmsSection: 27 [algorithms] Status: Open Submitter: Nikolay Ivchenkov Opened: 2012-08-01 Last modified: 2018-06-12
Priority: 4
View other active issues in [algorithms].
View all other issues in [algorithms].
View all issues with Open status.
Discussion:
According to 27.1 [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.[2014-06-07 Daniel comments and provides wording]
The specification for output iterators is somewhat tricky, because here a sequence of increments is required to be combined with intervening assignments to the dereferenced iterator. I tried to respect this fact by using a conceptual assignment operation as part of the specification.
Another problem in the provided as-if-code is the question which requirements are imposed on n. Unfortunately, the corresponding function advance is completely underspecified in this regard, so I couldn't borrow wording from it. We cannot even assume here that n is the difference type of the iterator, because for output iterators there is no requirements for this associated type to be defined. The presented wording attempts to minimize assumptions, but still can be considered as controversial.[2018-06 Rapperswil Wednesday issues processing]
Status to Open
Proposed resolution:
This wording is relative to N4606.
Change 27.1 [algorithms.general] around p12 as indicated:
-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;when X meets the input iterator requirements (25.3.5.3 [input.iterators]), otherwise it is the same as that of
X tmp = a; for (auto i = n; i; ++tmp, (void) --i) *tmp = Expr(i); return tmp;where Expr(i) denotes the (n-i)th expression that is assigned to for the corresponding algorithm; and that of b-a is the same as of
return distance(a, b);
Section: 24.2.8.2 [unord.req.except] Status: Open Submitter: Alisdair Meredith Opened: 2012-09-23 Last modified: 2019-07-22
Priority: 3
View all issues with Open status.
Discussion:
The hash functor and key-comparison functor of unordered containers are allowed to throw on swap.
24.2.8.2 [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.
[2013-03-15 Issues Teleconference]
Moved to Open.
[2019 Cologne Wednesday night]
Billy to write resolution for option #2. This may require a paper.
Proposed resolution:
Section: 32.9 [re.results] Status: Ready Submitter: Daniel Krügler Opened: 2012-10-06 Last modified: 2022-11-10
Priority: 3
View all other issues in [re.results].
View all issues with Ready status.
Discussion:
The requirement expressed in 32.9 [re.results] p2
The class template match_results shall satisfy the requirements of an allocator-aware container and of a sequence container, as specified in 24.2.4 [sequence.reqmts], except that only operations defined for const-qualified sequence containers are supported.
can be read to require the existence of the described constructors from as well, but they do not exist in the synopsis.
The missing sequence constructors are:match_results(initializer_list<value_type>); match_results(size_type, const value_type&); template<class InputIterator> match_results(InputIterator, InputIterator);
The missing allocator-aware container constructors are:
match_results(const match_results&, const Allocator&); match_results(match_results&&, const Allocator&);
It should be clarified, whether (a) constructors are an exception of above mentioned operations or (b) whether at least some of them (like those accepting a match_results value and an allocator) should be added.
As visible in several places of the standard (including the core language), constructors seem usually to be considered as "operations" and they certainly can be invoked for const-qualified objects. The below given proposed resolution applies only the minimum necessary fix, i.e. it excludes constructors from above requirement.[2013-04-20, Bristol]
Check current implementations to see what they do and, possibly, write a paper.
[2013-09 Chicago]
Ask Daniel to update the proposed wording to include the allocator copy and move constructors.
[2014-01-18 Daniel changes proposed resolution]
Previous resolution from Daniel [SUPERSEDED]:
Change 32.9 [re.results] p2 as indicated:
The class template match_results shall satisfy the requirements of an allocator-aware container and of a sequence container, as specified in 24.2.4 [sequence.reqmts], except that only operations defined for const-qualified sequence containers that are not constructors are supported.
[2015-05-06 Lenexa]
MC passes important knowledge to EF.
VV, RP: Looks good.
TK: Second form should be conditionally noexcept
JY: Sequence constructors are not here, but mentioned in the issue writeup. Why?
TK: That would have been fixed by the superseded wording.
JW: How does this interact with Mike Spertus' allocator-aware regexes? [...] Perhaps it doesn't.
JW: Can't create match_results, want both old and new resolution.
JY: It's problematic that users can't create these, but not this issue.
VV: Why conditional noexcept?
MC: Allocator move might throw.
JW: Update superseded wording to "only non-constructor operations that are"?
MC: Only keep superseded, but append "and the means of constructing match_results are limited to [...]"?
JY: Bullet 4 paragraph 2 needs to address the allocator constructor.
Assigned to JW for drafting.
[2015-10, Kona Saturday afternoon]
STL: I want Mike Spertus to be aware of this issue.
Previous resolution from Daniel [SUPERSEDED]:This wording is relative to N3936.
Change 32.9 [re.results] p4, class template match_results synopsis, as indicated:
[…] // 28.10.1, construct/copy/destroy: explicit match_results(const Allocator& a = Allocator()); match_results(const match_results& m); match_results(const match_results& m, const Allocator& a); match_results(match_results&& m) noexcept; match_results(match_results&& m, const Allocator& a) noexcept; […]Change 32.9.2 [re.results.const] as indicated: [Drafting note: Paragraph 6 as currently written, makes not much sense, because the noexcept does not allow any exception to propagate. Further-on, the allocator requirements do not allow for throwing move constructors. Deleting it seems to be near to editorial — end drafting note]
match_results(const match_results& m); match_results(const match_results& m, const Allocator& a);-4- Effects: Constructs an object of class match_results, as a copy of m.
match_results(match_results&& m) noexcept; match_results(match_results&& m, const Allocator& a) noexcept;-5- Effects: Move-constructs an object of class match_results from m satisfying the same postconditions as Table 142.
AdditionallyFor the first form, the stored Allocator value is move constructed from m.get_allocator().-6- Throws: Nothing if the allocator's move constructor throws nothing.
[2019-03-27 Jonathan updates proposed resolution]
Previous resolution [SUPERSEDED]:
This wording is relative to N4810.
These edits overlap with the proposed resolution of 2191 but it should be obvious how to resolve the conflicts. Both resolutions remove the word "Additionally" from p4. Issue 2191 removes the entire Throws: element in p5 but this issue replaces it with different text that applies to the new constructor only.
Change 32.9 [re.results] p4, class template match_results synopsis, as indicated:
[…] // 30.10.1, construct/copy/destroy: explicit match_results(const Allocator& a = Allocator()); match_results(const match_results& m); match_results(const match_results& m, const Allocator& a); match_results(match_results&& m) noexcept; match_results(match_results&& m, const Allocator& a); […]Change 32.9.2 [re.results.const] as indicated:
match_results(const match_results& m); match_results(const match_results& m, const Allocator& a);-3- Effects: Constructs an object of class match_results, as a copy of m. For the second form, the stored Allocator value is constructed from a.
match_results(match_results&& m) noexcept; match_results(match_results&& m, const Allocator& a);-4- Effects: Move-constructs an object of class match_results from m satisfying the same postconditions as Table 128.
AdditionallyFor the first form, the stored Allocator value is move constructed from m.get_allocator(). For the second form, the stored Allocator value is constructed from a.-6- Throws:
Nothing.The second form throws nothing if a == m.get_allocator().
[2022-11-06; Daniel syncs wording with recent working draft]
To ensure that all constructors are consistent in regard to the information about how the stored allocator is constructed, more wording is added. This harmonizes with the way how we specify the individual container constructors (Such as vector) even though 24.2.2.5 [container.alloc.reqmts] already provides some guarantees. For the copy-constructor we intentionally refer to 24.2.2.2 [container.reqmts] so that we don't need to repeat what is said there.
[Kona 2022-11-08; Move to Ready]
Proposed resolution:
This wording is relative to N4917.
Change 32.9 [re.results], class template match_results synopsis, as indicated:
[…] // 32.9.2 [re.results.const], construct/copy/destroy: match_results() : match_results(Allocator()) {} explicit match_results(const Allocator& a); match_results(const match_results& m); match_results(const match_results& m, const Allocator& a); match_results(match_results&& m) noexcept; match_results(match_results&& m, const Allocator& a); […]
Change 32.9.2 [re.results.const] as indicated:
explicit match_results(const Allocator& a);-?- Effects: The stored Allocator value is constructed from a.
-2- Postconditions: ready() returns false. size() returns 0.match_results(const match_results& m); match_results(const match_results& m, const Allocator& a);-?- Effects: For the first form, the stored Allocator value is obtained as specified in 24.2.2.2 [container.reqmts]. For the second form, the stored Allocator value is constructed from a.
-3- Postconditions: As specified in Table 142 [tab:re.results.const].match_results(match_results&& m) noexcept; match_results(match_results&& m, const Allocator& a);-4- Effects: For the first form, t
-5- Postconditions: As specified in Table 142 [tab:re.results.const]. -?- Throws: The second form throws nothing if a == m.get_allocator() is true.The stored Allocator value is move constructed from m.get_allocator(). For the second form, the stored Allocator value is constructed from a.
Section: 24.2.8 [unord.req] Status: Open Submitter: Alisdair Meredith Opened: 2012-10-09 Last modified: 2016-12-10
Priority: 3
View other active issues in [unord.req].
View all other issues in [unord.req].
View all issues with Open status.
Discussion:
The user cannot specify a max_load_factor for their unordered container at construction, it must be supplied after the event, when the container is potentially not empty. The contract for this method is deliberately vague, not guaranteeing to use the value supplied by the user, and any value actually used will be used as a ceiling that the container will attempt to respect.
The only guarantee we have is that, if user requests a max_load_factor that is less than the current load_factor, then the operation will take constant time, thus outlawing an implementation that chooses to rehash and so preserve as a class invariant that load_factor < max_load_factor.
Reasonable options conforming to the standard include ignoring the user's request if the requested value is too low, or deferring the rehash to the next insert operation and allowing the container to have a strange state (wrt max_load_factor) until then - and there is still the question of rehashing if the next insert is for a duplicate key in a unique container.
Given the deliberate vagueness of the current wording, to support a range of reasonable (but not perfect) behaviors, it is not clear why the equally reasonable rehash to restore the constraint should be outlawed. It is not thought that this is a performance critical operation, where users will be repeatedly setting low load factors on populated containers, in a tight or (less unlikely) an instant response scenario.
[2013-03-15 Issues Teleconference]
Moved to Open.
Alisdair to provide wording.
[2016-11-12, Issaquah]
Sat PM: Howard to provide wording
[2016-11-17 Howard provided wording.]
The provided wording is consistent with LWG discussion in Issaquah. An implementation of the proposed wording would be setting max_load_factor() to max(z, load_factor()). This preserves the container invariant:
load_factor() <= max_load_factor()And it preserves the existing behavior that no rehash is done by this operation.
If it is desired to change the max_load_factor() to something smaller than the current load_factor() that can be done by first reducing the current load_factor() by either increasing bucket_count() (via rehash or reserve), or decreasing size() (e.g. erase), and then changing max_load_factor().
This resolution reaffirms that load_factor() <= max_load_factor() is a container invariant which can never be violated.
[2016-11-27, Nico comments]
Current implementations behave differently.
In regard to the sentence"The only guarantee we have is that, if user requests a max_load_factor that is less than the current load_factor, then the operation will take constant time, thus outlawing an implementation that chooses to rehash and so preserve as a class invariant that load_factor < max_load_factor."Note that the current spec says that there is constant complexity without any precondition. So, rehashing to keep the invariant would violate the spec (which is probably not be the intention). This issue is related to LWG 2199.
Proposed resolution:
Modify Table 87 as follows:
Expression | Return type | Assertion/note pre-/post-condition | Complexity |
---|---|---|---|
a.max_load_factor(z) | void |
Pre: z shall be positive. May change the container's maximum load factor, uing z as a hint. Post: a.load_factor() <= a.max_load_factor() Note: a.load_factor() is not modified by this operation. |
Constant |
Section: 33.10.9 [futures.async] Status: Deferred Submitter: Detlef Vollmann Opened: 2012-10-19 Last modified: 2016-01-28
Priority: 4
View other active issues in [futures.async].
View all other issues in [futures.async].
Discussion:
promise, packaged_task, and async are the only places where a shared state is actually supposed to be allocated. Accordingly, promise and packaged_task are "allocator-aware". But function template async provides no way to provide an allocator.
[2013-09 Chicago]
Matt: deprecate async
Nico: read my paper Alisdair: defer issues to wait for polymorphic allocators Alisdair: defer, active topic of research Deferred[2014-02-20 Re-open Deferred issues as Priority 4]
[2015-05 Lenexa, SG1 response]
We want whatever status approximates: "will not fix; we're working on a replacement facility and don't want to add features to a broken one"
Proposed resolution:
Section: 24.2.4 [sequence.reqmts], 24.2.7 [associative.reqmts], 24.2.8 [unord.req], 28.5.3.2 [rand.req.seedseq] Status: Open Submitter: Jeffrey Yasskin Opened: 2012-10-21 Last modified: 2020-09-06
Priority: 3
View other active issues in [sequence.reqmts].
View all other issues in [sequence.reqmts].
View all issues with Open status.
Discussion:
In 24.2.4 [sequence.reqmts] p3, we have "il designates an object of type initializer_list<value_type>", and then several functions that take 'il' as an argument. However, an expression like {1, 2, 'a'} is not an object of type initializer_list<int> unless it's used to initialize an explicitly-typed variable of that type. I believe we want:
std::vector<int> v; v = {1, 2, 'a'};
to compile portably, so we should say something different when defining 'il'. The same phrasing happens in 24.2.7 [associative.reqmts], 24.2.8 [unord.req], and 28.5.3.2 [rand.req.seedseq].
This may just be an editorial issue because the actual class synopses declare the functions to take initializer_list<exact_type>.[2013-03-15 Issues Teleconference]
Moved to Open.
This is definitely not NAD
Should copy the suggested wording as the proposed resolution.
[2019-03-26; Daniel comments and provides wording]
The 2013-03-15 comment is confusing, since it recommends to "copy the suggested wording as the proposed resolution". I couldn't find such wording in the issue nor in the associated wiki, so I provided that wording out of myself. The tricky part is to define which kind of braced-init-list we want to allow. As Tim Song pointed out, we still need the existing support for std::initializer_list<value_type> as well, because otherwise existing semantics based on expressions such as li.begin() won't work anymore. The below suggested wording restricts supported braced-init-lists to every initializer list that can be used to copy-list-initialize an object of type std::initializer_list<value_type> by saying:
"bil designates any braced-init-list suitable to copy-list-initialize an object of type initializer_list<value_type> (9.4.5 [dcl.init.list])"
As a drive-by fix, the provided wording adds another initialization "expression" that makes the construction of the form
std::vector<int> v = {1, 2, 'a'};
valid (We just miss a copy-initialization case).
Proposed resolution:
This wording is relative to N4810.
[Drafting note: We need to special-case the "expression" X u = bil; below, because for empty braced-init-list the effects are those of calling the default constructor. — end drafting note]
Modify 24.2.4 [sequence.reqmts] as indicated:
-3- In Tables 66 and 67, […] il designates a
n objectvalue of type initializer_list<value_type>, bil designates any braced-init-list suitable to copy-list-initialize an object of type initializer_list<value_type> (9.4.5 [dcl.init.list]), […]
Modify Table 66 — "Sequence container requirements (in addition to container)" as indicated:
Table 66 — Sequence container requirements (in addition to container) Expression Return type Assertion/note
pre-/post-condition[…] X(il)
X u = il;Equivalent to X(il.begin(), il.end())
or X u(il.begin(), il.end());, respectivelyX(bil) Equivalent to X(initializer_list<value_type>(bil)) X u = bil; If bil is empty, equivalent to X u;, otherwise
equivalent to X u = initializer_list<value_type>(bil);a = il X& […] a = bil X& Equivalent to a = initializer_list<value_type>(bil) […] a.insert(p, il) iterator […] a.insert(p, bil) iterator Equivalent to a.insert(p, initializer_list<value_type>(bil)) […] a.assign(il) void […] a.assign(bil) void Equivalent to a.assign(initializer_list<value_type>(bil)) […]
Modify 24.2.7 [associative.reqmts] as indicated:
-8- In Table 69, […] il designates a
n objectvalue of type initializer_list<value_type>, bil designates any braced-init-list suitable to copy-list-initialize an object of type initializer_list<value_type> (9.4.5 [dcl.init.list]), […]
Modify Table 69 — "Associative container requirements (in addition to container)" as indicated:
Table 69 — Associative container requirements (in addition to container) Expression Return type Assertion/note
pre-/post-conditionComplexity […] X(il)
X u = il;same as X(il.begin(), il.end())
or X u(il.begin(), il.end());, respectivelysame as X(il.begin(), il.end())
or X u(il.begin(), il.end());, respectivelyX(bil) Equivalent to X(initializer_list<value_type>(bil)) X u = bil; If bil is empty, equivalent to X u;, otherwise
equivalent to X u = initializer_list<value_type>(bil);X(il,c) same as X(il.begin(), il.end(), c) same as X(il.begin(), il.end(), c) X(bil, c) Equivalent to X(initializer_list<value_type>(bil), c) a = il X& […] […] a = bil X& Equivalent to a = initializer_list<value_type>(bil) […] a.insert(il) void equivalent to a.insert(il.begin(), il.end()) a.insert(bil) void Equivalent to a.insert(initializer_list<value_type>(bil)) […] a.assign(il) void […] a.assign(bil) void Equivalent to a.assign(initializer_list<value_type>(bil)) […]
Modify 24.2.8 [unord.req] p11's bullet list as indicated:
-11- In Table 70:
(11.1) — […]
[…]
(11.14) — il denotes a value of type initializer_list<value_type>,
(11.?) — bil denotes any braced-init-list suitable to copy-list-initialize an object of type initializer_list<value_type> (9.4.5 [dcl.init.list]),
[…]
Modify Table 70 — "Unordered associative container requirements (in addition to container)" as indicated:
[Drafting note: There is a preexisting issue with Table 70, that there is no symbol u specified ("u denotes the name of a variable being declared"), so existing initialization forms with a named variable are currently always written as "X a[…]" where a is defined as "a denotes a value of type X", the wording below follows this existing practice but the author of this wording would like to kindly ask the Project Editor to introduce said symbol u and apply it to all existing and new such named initialization forms instead. — end drafting note]
Table 70 — Unordered associative container requirements (in addition to container) Expression Return type Assertion/note
pre-/post-conditionComplexity […] X(il)
X a = il;X Same as X(il.begin(), il.end())
or X a(il.begin(), il.end());, respectivelySame as X(il.begin(), il.end())
or X a(il.begin(), il.end());, respectivelyX(bil) X Equivalent to X(initializer_list<value_type>(bil)) X a = bil; X If bil is empty, equivalent to X a;, otherwise
equivalent to X a = initializer_list<value_type>(bil);X(il, n) X Same as X(il.begin(), il.end(), n) Same as X(il.begin(), il.end(), n) X(bil, n) X Equivalent to X(initializer_list<value_type>(bil), n) X(il, n, hf) X Same as X(il.begin(), il.end(), n, hf) Same as X(il.begin(), il.end(), n, hf) X(bil, n, hf) X Equivalent to X(initializer_list<value_type>(bil), n, hf) X(il, n, hf, eq) X Same as X(il.begin(), il.end(), n, hf, eq) Same as X(il.begin(), il.end(), n, hf, eq) X(bil, n, hf, eq) X Equivalent to X(initializer_list<value_type>(bil), n, hf, eq) […] a = il X& […] […] a = bil X& Equivalent to a = initializer_list<value_type>(bil) […] a.insert(il) void Same as a.insert(il.begin(), il.end()). Same as a.insert(il.begin(), il.end()). a.insert(bil) void Equivalent to a.insert(initializer_list<value_type>(bil)) […]
Modify 28.5.3.2 [rand.req.seedseq] p2's bullet list as indicated:
-2- A class S satisfies the requirements of a seed sequence if the expressions shown in Table 82 are valid and have the indicated semantics, and […] In that Table and throughout this subclause:
(2.1) — […]
(2.?) — u denotes the name of a variable being declared,
[…]
(2.6) — il is a value of initializer_list<T>
.;(2.?) — bil denotes any braced-init-list suitable to copy-list-initialize an object of type initializer_list<T> (9.4.5 [dcl.init.list]).
Modify Table 82 — "Seed sequence requirements" as indicated:
Table 82 — Seed sequence requirements Expression Return type Pre/post-condition Complexity […] S(il)
S u = il;Same as S(il.begin(), il.end())
or S u(il.begin(), il.end());, respectivelysame as S(il.begin(), il.end())
or S u(il.begin(), il.end());, respectivelyS(bil) Equivalent to S(initializer_list<T>(bil)) S u = bil; If bil is empty, equivalent to S u;, otherwise
equivalent to S u = initializer_list<T>(bil);[…]
Section: 31.5.4.2 [basic.ios.cons] Status: Open Submitter: Andrey Semashev Opened: 2012-11-09 Last modified: 2021-07-31
Priority: 4
View all other issues in [basic.ios.cons].
View all issues with Open status.
Discussion:
There is an ambiguity in how std::basic_ios::init method (31.5.4.2 [basic.ios.cons]) can be used in the derived class. The Standard only specify the state of the basic_ios object after the call completes. However, in basic_ios default constructor description (31.5.4.2 [basic.ios.cons]) there is this sentence:
Effects: Constructs an object of class basic_ios (31.5.2.8 [ios.base.cons]) leaving its member objects uninitialized. The object shall be initialized by calling basic_ios::init before its first use or before it is destroyed, whichever comes first; otherwise the behavior is undefined.
This restriction hints that basic_ios::init should be called exactly once before the object can be used or destroyed, because basic_ios::init may not know whether it was called before or not (i.e. whether its members are actually uninitialized or are initialized by the previous call to basic_ios::init). There is no such restriction in the basic_ios::init preconditions so it is not clear whether it is allowed to call basic_ios::init multiple times or not.
This problem has already affected publicly available implementations. For example, Microsoft Visual C++ STL introduces a memory leak if basic_ios::init is called multiple times, while GCC 4.7 and STLPort reinitialize the basic_ios object correctly without memory leak or any other undesired effects. There was a discussion of this issue on Boost developers mailing list, and there is a test case that reproduces the problem. The test case is actually a bug report for my Boost.Log library, which attempts to cache basic_ostream-derived objects internally to avoid expensive construction and destruction. My stream objects allowed resetting the stream buffer pointers the stream is attached to, without requiring to destroy and construct the stream. My personal view of the problem and proposed resolution follows. While apparently the intent of basic_ios::init is to provide a way to initialize basic_ios after default construction, I see no reason to forbid it from being called multiple times to reinitialize the stream. Furthermore, it is possible to implement a conforming basic_ios that does not have this restriction. The quoted above section of the Standard that describes the effects of the default constructor is misleading. The Standard does not mandate any data members of basic_ios or ios_base (31.5.2 [ios.base]), which it derives from. This means that the implementation is allowed to use non-POD data members with default constructors that initialize the members with particular default values. For example, in the case of Microsoft Visual C++ STL the leaked memory is an std::locale instance that is dynamically allocated during basic_ios::init, a raw pointer to which is stored within ios_base. It is possible to store e.g. an unique_ptr instead of a raw pointer as a member of ios_base, the smart pointer will default initialize the underlying raw pointer on default construction and automatically destroy the allocated object upon being reset or destroyed, which would eliminate the leak and allow basic_ios::init to be called multiple times. This leads to conclusion that the default constructor of basic_ios cannot leave "its member objects uninitialized" but instead performs default initialization of the member objects, which would mean the same thing in case of POD types. However, I feel that restricting ios_base and basic_ios members to non-POD types is not acceptable. Since multiple calls to basic_ios::init are not forbidden by the Standard, I propose to correct the basic_ios default constructor description so that it is allowed to destroy basic_ios object without calling basic_ios::init. This would imply that any raw members of basic_ios and ios_base should be initialized to values suitable for destruction (essentially, this means only initializing raw pointers to NULL). The new wording could look like this:Effects: Constructs an object of class basic_ios (31.5.2.8 [ios.base.cons]) initializing its member objects to unspecified state, only suitable for basic_ios destruction. The object shall be initialized by calling basic_ios::init before its first use; otherwise the behavior is undefined.
This would remove the hint that basic_ios::init must be called exactly once. Also, this would remove the requirement for basic_ios::init to be called at all before the destruction. This is also an important issue because the derived stream constructor may throw an exception before it manages to call basic_ios::init (for example, if the streambuf constructor throws), and in this case the basic_ios destructor has undefined behavior.
To my mind, the described modification is sufficient to resolve the issue. But to emphasize the possibility to call basic_ios::init multiple times, a remark or a footnote for basic_ios::init postconditions could be added to explicitly state the semantics of calling it multiple times. The note could read as follows:The function can be called multiple times during the object lifetime. Each subsequent call reinitializes the object to the described in postconditions initial state.
[2013-04-20, Bristol]
Alisdair: The current wording is unclear but the proposed resolution is wrong
Solution: Clarify that init must be called once and only once. Move then to review.[2021-07-29 Tim comments]
The requirement that "init must be called once and only once" conflicts with the disposition of LWG 135.
Proposed resolution:
This wording is relative to N3485.
Edit 31.5.4.2 [basic.ios.cons] as indicated:
basic_ios();-2- Effects: Constructs an object of class basic_ios (31.5.2.8 [ios.base.cons])
leaving its member objects uninitializedinitializing its member objects to unspecified state, only suitable for basic_ios destruction. The object shall be initialized by calling basic_ios::init before its first useor before it is destroyed, whichever comes first; otherwise the behavior is undefined.void init(basic_streambuf<charT,traits>* sb);Postconditions: The postconditions of this function are indicated in Table 128.
-?- Remarks: The function can be called multiple times during the object lifetime. Each subsequent call reinitializes the object to the described in postconditions initial state.
Section: 24.2.7 [associative.reqmts], 24.2.8 [unord.req] Status: Open Submitter: Alisdair Meredith Opened: 2012-11-14 Last modified: 2015-10-22
Priority: 3
View other active issues in [associative.reqmts].
View all other issues in [associative.reqmts].
View all issues with Open status.
Discussion:
The requirements on the functors used to arrange elements in the various associative and unordered containers are given by a set of expressions in tables 102 — Associative container requirements, and 103 — Unordered associative container requirements. In keeping with Library convention these expressions make the minimal requirements necessary on their types. For example, we have the following 3 row extracts for the unordered containers:
Expression | Assertion/note pre-/post-condition |
X(n, hf, eq) X a(n, hf, eq) |
Requires: hasher and key_equal are CopyConstructible. |
X(n, hf) X a(n, hf) |
Requires: hasher is CopyConstructible and key_equal is DefaultConstructible. |
X(n) X a(n) |
Requires: hasher and key_equal are DefaultConstructible. |
However, the signature for each class template requires that the functors must effectively be CopyConstructible for each of these expressions:
template <class Key, class T, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<std::pair<const Key, T> > > class unordered_map { ... // construct/destroy/copy explicit unordered_map(size_type n = see below, const hasher& hf = hasher(), const key_equal& eql = key_equal(), const allocator_type& a = allocator_type()); ... }
The letter of the standard can be honored as long as implementors recognize their freedom to split this one signature into multiple overloads, so that the documented default arguments (requiring a CopyConstructible functor) are not actually passed as default arguments.
As we look into the requirements for the copy constructor and copy-assignment operator, the requirements are even more vague, as the explicit requirements on the functors are not called out, other than saying that the functors are copied.
Must the functors be CopyAssignable? Or is CopyConstructible sufficient in this case? Do we require that the functors be Swappable so that the copy-swap idiom can be deployed here? Note that a type that is both CopyConstructible and CopyAssignable is still not guaranteed to be Swappable as the user may delete the swap function for their type in their own namespace, which would be found via ADL.
Some clean-up of the requirements table looks necessary, to at least document the assignment behavior. In addition, we should have clear guidance on whether these functors should always be CopyConstructible, as suggested by the class template definitions, or if the requirement tables are correct and we should explicitly split up the constructors in the (unordered) associative containers to no longer use default (function) arguments to obtain their defaulted functors.
I recommend the simplest solution would be to always require that the functors for (unordered) associative containers be CopyConstructible, above the requirements tables themselves, so that the issue need not be addressed within the tables. I suggest that the assignment operators for these containers add the requirement that the functors be Swappable, rather than forwarding the corresponding Assignable requirement.
[2013-03-15 Issues Teleconference]
Moved to Open.
Alisdair to propose wording.
[2014-06-08, Daniel comments]
The area of this issue partially overlaps what LWG 2227 addresses.
[2015-10-20, Daniel comments]
The revised resolution of LWG 2227 should resolve this issue as well. It follows the recommendations of the submitter to require CopyConstructible requirements for the function objects owned by containers, but it does not impose any further fundamental requirements.
Proposed resolution:
See the resolution of LWG 2227.
Section: 32.10.4 [re.alg.replace] Status: New Submitter: Jeffrey Yasskin Opened: 2012-11-26 Last modified: 2016-01-28
Priority: 3
View all other issues in [re.alg.replace].
View all issues with New status.
Discussion:
template <class traits, class charT, class ST, class SA> basic_string<charT, ST, SA> regex_replace(const basic_string<charT, ST, SA>& s, const basic_regex<charT, traits>& e, const charT* fmt, regex_constants::match_flag_type flags = regex_constants::match_default);
and friends are documented as
Constructs an empty string result of type basic_string<charT, ST, SA> and calls regex_replace(back_inserter(result), s.begin(), s.end(), e, fmt, flags).
This appears to require the result to have a default-constructed allocator, which isn't even possible for all allocator types. I suspect the allocator should be copied from 's' instead. Possibly there should be an additional defaulted argument to override the allocator of the result.
Proposed resolution:
Section: 32.11.2.3 [re.tokiter.comp] Status: Open Submitter: Pete Becker Opened: 2012-11-21 Last modified: 2018-08-23
Priority: 3
View all issues with Open status.
Discussion:
Consider the following example:
std::string str0("x"); std::regex rg0("a"); std::regex_token_iterator it0(str0.begin(), str0.end(), rg0, -1); // points at "x" in str0 std::string str1("x"); std::regex rg1("b"); std::regex_token_iterator it1(str1.begin(), str1.end(), rg1, -1); // points at "x" in str1
32.11.2.3 [re.tokiter.comp] p1 says that it0.operator==(it1) returns true "if *this and right are both suffix iterators and suffix == right.suffix"; both conditions are satisfied in this example. It does not say that they must both be iterators into the same sequence, nor does it say (as general iterator requirements do) that they must both be in the domain of == in order for the comparison to be meaningful. It's a simple statement: they're equal if the strings they point at compare equal. Given this being a valid comparison, the obtained result of "true" looks odd.
The problem is that for iterator values prior to the suffix iterator, equality means the same regular expression and the same matched sequence (both uses of "same" refer to identity, not equality); for the suffix iterator, equality means that the matched sequences compare equal.[2014-02-10]
Priority set to 2
[2018-08-20 Casey adds a proposed resolution]
Priority changed to 3.
Marshall notes that iterator comparisons typically require the iterators to denote elements of the same sequence.Previous resolution [SUPERSEDED]:
This wording is relative to N4762.
Modify 32.11.2.3 [re.tokiter.comp] as follows:
bool operator==(const regex_token_iterator& right) const;-?- Expects: *this and right are both end-of-sequence iterators or both have the same underlying sequence.
-1- Returns: true if *this and right are both end-of-sequence iterators, or if […]
bool operator!=(const regex_token_iterator& right) const;-?- Expects: *this and right are both end-of-sequence iterators or both have the same underlying sequence.
-2- Returns: !(*this == right).
[2018-08-23 Casey revises the P/R in response to LWG feedback]
Proposed resolution:
This wording is relative to N4762.
Modify 32.11.2.3 [re.tokiter.comp] as follows:
bool operator==(const regex_token_iterator& right) const;-?- Expects: At least one of *this and right is an end-of-sequence iterator, or both *this and right have the same underlying sequence.
-1- Returns: true if *this and right are both end-of-sequence iterators, or if […]
bool operator!=(const regex_token_iterator& right) const;-?- Expects: At least one of *this and right is an end-of-sequence iterator, or both *this and right have the same underlying sequence.
-2- Returns: !(*this == right).
Section: 24.2.7 [associative.reqmts] Status: Open Submitter: Juan Soulie Opened: 2012-12-19 Last modified: 2019-04-23
Priority: 3
View other active issues in [associative.reqmts].
View all other issues in [associative.reqmts].
View all issues with Open status.
Discussion:
Table 102 in 24.2.7 [associative.reqmts]/8 states on expression a.key_comp() that it "returns the comparison object out of which a was constructed". At the same time, 24.2.2.1 [container.requirements.general]/8 states (starting in the third line) that "...Any Compare, Pred, or Hash objects belonging to a and b shall be swappable and shall be exchanged by unqualified calls to non-member swap...". This is problematic for any compliant implementation, since once swapped the container cannot return the comparison object out of which it was constructed unless incurring in storing an otherwise needless object.
The simple solution is to correct that statement in Table 102, but I believe this is part of a larger problem of underspecified behavior: The new standard has made an effort in regards to allocators and now fully specifies what happens to stateful allocator objects. It has even specified what happens to stateful hasher and key_equal members of unordered containers (they propagate), but it says nothing about stateful comparison objects of (ordered) associative containers, except for the statement in 24.2.2.1 [container.requirements.general]/8 referred above and only related to swap. For example, it is unclear to me what is specified to happen on an assignment: should the comparison object be copied/moved along with the elements, or should the left-hand side object keep its own? Maybe this has been intentionally left unspecified with the purpose of compatibility with C++98, which I understand it specified that comparison objects were kept for the entire life of the container (like allocators) — an unfortunate choice. But anyway, the segment of 24.2.2.1 [container.requirements.general] quoted above seems to break any possible backwards compatibility with C++98 in this regard. Therefore, taking into consideration consistency with how this is dealed with for unordered associative containers, I propose that Table 102 is modified as follows:The row for expression a.key_comp() is changed so that its "assertion/note pre-/post-condition" reads "Returns a's comparison object."
A new row is added at the appropriate location (which I believe would be after "X(il)" row), with:
Table 102 — Associative container requirements (in addition to container) Expression Return type Assertion/note pre-/post-condition Complexity X(b)
X a(b)X Copy constructor. In addition to
the requirements of Table 96, copies
the comparison object.Linear in b.size() a = b X& Copy assignment operator. In addition to
the requirements of Table 96, copies the
comparison object.Linear in a.size() and b.size()
[2013-03-15 Issues Teleconference]
Moved to Review.
[2013-04-18, Bristol]
STL: can't believe we don't specify this already. this is totally necessary
Alisdair: how does it do this? copy construction? assignment? Also need it for move. STL: we already specify this for constructing from a comparator, not during copy construction though. Jonathan: don't like wording, should say "key_compare is CopyConstructible. Uses b.key_comp() as a comparison object." STL: we get it right for unordered! Jonathan: can't wordsmith this now, but I think implementations do the right thing. Alisdair: not sure what right thing is for moves. Also we say nothing about propagating allocators to functors.Moved to Open.
[2015-02 Cologne]
TK: There's no need for fine-grained propagate/not-propagate control. If you don't want to propagate the predicate, you can simply construct or insert from an iterator range.
VV: libstdc++ already implements the resolution of this issue. GR: There are a couple of other problems. We don't specify move constructor and move assignment for maps. Those are just general. TK: General container requirements already describe the semantics for {copy,move}-{construction,assignment}, so it doesn't seem that there's room for choice in std::map assignments. unordered_map is different, though. [Note: Check what general container requirements say about container equality.] DK will draft wording. The decision is to unambiguously make all {copy,move}-{construction,assignment} operations endow the LHS with the exact state of the RHS, including all predicates and hash function states. Conclusion: Update wording, revisit later.[2015-05-06 Lenexa: Waiting for updated wording]
Previous resolution [SUPERSEDED]:
This wording is relative to N3485.
Change Table 102 as indicated:
Table 102 — Associative container requirements (in addition to container) Expression Return type Assertion/note pre-/post-condition Complexity … X(il) Same as X(il.begin(), il.end()). same as X(il.begin(), il.end()). X(b)
X a(b)Copy constructor. In addition to
the requirements of Table 96, copies
the comparison object.Linear in b.size() a = b X& Copy assignment operator. In addition to
the requirements of Table 96, copies the
comparison object.Linear in a.size() and b.size() … a.key_comp() X::key_compare rReturnsthea's comparison object
out of which a was constructed.constant
[2015-10-19 Daniel comments and provides alternative wording]
The current standard is especially unclear in regard to what effects move operations of unordered/associative containers should have. We have one example that is standardized exactly in this way by looking at 24.6.7.3 [priqueue.cons.alloc] p7:
template <class Alloc> priority_queue(priority_queue&& q, const Alloc& a);-7- Effects: Initializes c with std::move(q.c) as the first argument and a as the second argument, and initializes comp with std::move(q.comp)
A similarly comparable example are the move-operations of std::unique_ptr in regard to the deleter (when this is no a reference), which also respect move-capabilities of that function object.
We have wording from C++98 for associative containers (but not for unordered containers!) that was never adjusted to C++11 move-semantics in 24.2.7 [associative.reqmts] p12:When an associative container is constructed by passing a comparison object the container shall not store a pointer or reference to the passed object, even if that object is passed by reference. When an associative container is copied, either through a copy constructor or an assignment operator, the target container shall then use the comparison object from the container being copied, as if that comparison object had been passed to the target container in its constructor.
The second sentence of this wording is problematic for several reasons:
It only talks about copy operations, not about move operations, except that the term "assignment" without leading "copy" is a bit ambigious (albeit it seems clear in the complete context).
It is not really clear how to interpret "as if that comparison object had been passed to the target container in its constructor" for an assignment operation. A possible but not conclusive interpretation could be that this is wording supporting a "copy-via-swap" idiom.
There does not exist similar wording for unordered containers, except that Table 102 provides entries for copy construction and copy assignment of the containers whose wording just talks of "copies" in either case.
Existing implementations differ already:
Visual Studio 2015 uses copy construction and copy assignment for the two copy operations but uses swap operations for the move operations.
GCC's libstdc++ performs copy construction and copy assignment for the two copy operations and for the two move operations, respectively
clang++'s libc++ performs copy/move construction and copy/move assignment for the corresponding four copy/move operations
The alternative wording provided below attempts to clarify that container copy/move operations perform the corresponding copy/move operations on the owned function objects.
In addition the wording also resolves LWG 2215: I believe that the current wording should require that container function objects should meet the CopyConstructible requirements. Adding this general requirement also fixes the underspecified requirements of the accessor functions key_comp() and value_comp(). I don't think that a general requirement for Swappable is needed, only the member swap function currently requires this. Nonetheless the wording below does support stateful functors that are also moveable or move-assignable, therefore the specified semantics in terms of move operations. I should add the following warning, though: If this proposed wording would be accepted, there is a little chance of code breakage, because the current wording can be read that in general there is no requirement that the container functors are CopyConstructible. The following code example is accepted by gcc + libstd++:#include <map> #include <utility> #include <iostream> struct Cmp { Cmp() = default; Cmp(const Cmp&) = delete; Cmp(Cmp&&) = delete; Cmp& operator=(const Cmp&) = delete; Cmp& operator=(Cmp&&) = delete; template<class T> bool operator()(const T& x, const T& y) const { return x < y; } }; typedef std::map<int, int, Cmp> MyMap; int main() { MyMap m; std::cout << (m.find(12) == m.end()) << std::endl; }
Previous resolution [SUPERSEDED]:
This wording is relative to N4527.
Change 24.2.7 [associative.reqmts] p8 as indicated:
-8- In Table 101, X denotes an associative container class, a denotes a value of type X, b denotes a possibly const value of type X, rv denotes a non-const rvalue of type X, u denotes the name of a variable being declared, […]
Change Table 101 as indicated:
Table 101 — Associative container requirements (in addition to container) Expression Return type Assertion/note pre-/post-condition Complexity … X::key_compare Compare Requires: Compare is CopyConstructible.
defaults to less<key_type>compile time X(c)
X u(c);Requires: key_compare is CopyConstructible.Effects: Constructs an empty container.
Uses a copy of c as a comparison object.[…] … X(i,j,c)
X u(i,j,c);Requires: key_compare is CopyConstructible.value_type is EmplaceConstructible into X from *i.
Effects: Constructs an empty container and inserts elements
from the range [i, j) into it; uses c as a comparison object.[…] … X(il) Same as X(il.begin(), il.end()). same as X(il.begin(), il.end()). X(b)
X a(b)(In addition to the requirements of Table 95)
Effects: Copy constructs the comparison object of a from
the comparison object of b.Linear in b.size() X(rv)
X a(rv)(In addition to the requirements of Table 95 and Table 98)
Effects: Move constructs the comparison object of a from
the comparison object of rv.constant a = b X& (In addition to the requirements of Table 95 and Table 98)
Requires: key_compare is CopyAssignable.
Effects: Copy assigns the comparison object of b
to the comparison object of a.Linear in a.size() and b.size() a = rv X& (In addition to the requirements of Table 95 and Table 98)
Requires: key_compare is MoveAssignable.
Effects: Move assigns from the comparison object of rv
to the comparison object of a.Linear … a.key_comp() X::key_compare rReturnsthea's comparison object
out of which a was constructed.constant Change 24.2.7 [associative.reqmts] p12 as indicated:
-12- When an associative container is constructed by passing a comparison object the container shall not store a pointer or reference to the passed object, even if that object is passed by reference.
When an associative container is copied, either through a copy constructor or an assignment operator, the target container shall then use the comparison object from the container being copied, as if that comparison object had been passed to the target container in its constructor.Change 24.2.8 [unord.req] p11 as indicated:
-11- In Table 102: X denotes an unordered associative container class, a denotes a value of type X, b denotes a possibly const value of type X, rv denotes a non-const rvalue of type X, […]
Change Table 102 as indicated:
Table 102 — Unordered associative container requirements (in addition to container) Expression Return type Assertion/note pre-/post-condition Complexity … X::hasher Hash Requires: Hash is CopyConstructible.
Hash shall be a unary function object type
such that the expression hf(k) has type std::size_t.compile time X::key_equal Pred Requires: Pred is CopyConstructible.
Pred shall be a binary predicate that takes
two arguments of type Key.
Pred is an equivalence relation.compile time … X(n, hf, eq)
X a(n, hf, eq)X Requires: hasher and key_equal are CopyConstructible.Effects: […][…] X(n, hf)
X a(n, hf)X Requires: hasher is CopyConstructible andkey_equal is DefaultConstructible.
Effects: […][…] … X(i, j, n, hf, eq)
X a(i, j, n, hf, eq)X Requires: hasher and key_equal are CopyConstructible.value_type is EmplaceConstructible into X from *i.
Effects: […][…] X(i, j, n, hf)
X a(i, j, n, hf)X Requires: hasher is CopyConstructible andkey_equal is DefaultConstructible.
value_type is EmplaceConstructible into X from *i.
Effects: […][…] … X(b)
X a(b)X Copy constructor. In addition(In addition to the requirements of Table 95)
to the requirements of Table 95,
copies the hash function,
predicate, and maximum load
factor.
Effects: Copy constructs the hash function, predicate, and maximum load factor
of a from the corresponding objects of b.Average case linear in
b.size(),
worst case quadratic.X(rv)
X a(rv)X (In addition to the requirements of Table 95 and Table 98)
Effects: Move constructs the hash function, predicate, and maximum load factor
of a from the corresponding objects of rv.constant a = b X& Copy assignment operator. In(In addition to the requirements of Table 95 and Table 98)
addition to the requirements of
Table 95, copies the hash
function, predicate, and
maximum load factor.
Requires: hasher and key_equal are CopyAssignable.
Effects: Copy assigns the hash function, predicate, and maximum load factor
of b to the corresponding objects of a.Average case linear in
b.size(),
worst case quadratic.a = rv X& (In addition to the requirements of Table 95 and Table 98)
Requires: hasher and key_equal are MoveAssignable.
Effects: Move assigns the hash function, predicate, and maximum load factor
from rv to the corresponding objects of a.Linear …
[2016-08-07]
Daniel removes the previously proposed wording to work on revised wording.
[2019-04-22, Billy comments]
In addition to the Cpp17CopyConstructible discussion going on there, I think we need to require that calling the comparison function when Compare itself is const needs to produce the same answer as if Compare is non-const.
Proposed resolution:
Section: 33.5.2 [atomics.syn], 33.5.4 [atomics.order] Status: SG1 Submitter: Daniel Krügler Opened: 2013-01-19 Last modified: 2016-01-28
Priority: Not Prioritized
View other active issues in [atomics.syn].
View all other issues in [atomics.syn].
View all issues with SG1 status.
Discussion:
The "magic" kill_dependency function is a function without any constraints on the template parameter T and is specified as
template <class T> T kill_dependency(T y) noexcept;-14- Effects: The argument does not carry a dependency to the return value (1.10).
-15- Returns: y.
I wonder whether the unconditional noexcept is really intended here: Assume we have some type U that has a potentially throwing move constructor (or it has a potentially throwing copy constructor and no move constructor), for any "normal" function template with the same signature and the same effects (modulo the dependency magic) this would mean that it cannot safely be declared noexcept because of the return statement being part of the complete function call affected by noexcept (The by-value function argument is irrelevant in this context). In other words it seems that a function call such as
struct S { ... S(const S& r) { if(some condition) throw Something(); } ... }; int main() { S s1 = ...; S s2 = std::kill_dependency(s1); }
would be required to call std::terminate if the copy constructor of S throws during the return of std::kill_dependency.
To require copy elision for this already magic function would look like a low-hanging fruit to solve this problem, but this case is not covered by current copy elision rules see 12.8 p31 b1: "— in a return statement in a function with a class return type, when the expression is the name of a non-volatile automatic object (other than a function or catch-clause parameter) with the same cv-unqualified type as the function return type, the copy/move operation can be omitted by constructing the automatic object directly into the function's return value". Some options come into my mind:Make the exception-specification a constrained one in regard via std::is_nothrow_move_constructible:
template <class T> T kill_dependency(T y) noexcept(see below);
This is similar to the approach taken for function templates such as std::swap.
Use perfect forwarding (This needs further wording to correct the effects):
template <class T> T&& kill_dependency(T&& y) noexcept;
Impose constraints on the template arguments in regard to throwing exceptions while copying/moving.
Keep the state as it is but possibly add a note about a call of std::terminate in above scenario.
A second problem is that the current wording is not clear whether it is well-defined to call the function with types that are reference types, such as in the following example:
#include <atomic> int main() { int a = 12; int& b = std::kill_dependency<int&>(a); }
It is unclear what kind of dependency is killed here. This is presumably a core language problem, but could affect the possible resolutions of the problem.
[2014-11 Urbana]
Recommend using a revised example:
int lookup(class D* p) { class E* q = p->a.load(memory_order_consume); int y = std::kill_dependency(q->y); }
[2015-02 Cologne]
Handed over to SG1.
Proposed resolution:
Section: 23.5 [c.strings] Status: New Submitter: Jason Merrill Opened: 2013-01-29 Last modified: 2016-01-28
Priority: 4
View other active issues in [c.strings].
View all other issues in [c.strings].
View all issues with New status.
Discussion:
Apparently C1X changes __STDC_UTF_16__ and __STDC_UTF_32__ from macros defined in uchar.h (and reflected in C++ by Table 79) to be predefined by the compiler. Do we want to do the same?
Proposed resolution:
Section: 23.5 [c.strings] Status: Open Submitter: Johannes Schaub Opened: 2013-02-02 Last modified: 2016-08-09
Priority: 3
View other active issues in [c.strings].
View all other issues in [c.strings].
View all issues with Open status.
Discussion:
The non-explicit nature of the iterator-pair constructor of containers, such a
template <class InputIterator> vector(InputIterator first, InputIterator last, const Allocator& = Allocator());
can be selected in unexpected situations, leading to a hard runtime error, as demonstrated by the following example:
#include <vector> void f(std::vector<char> v){ /* ... */} int main() { f({"A", "B"}); }
The actually intended initializer-list constructor isn't feasible here, so the best match is the constructor template
template <class InputIterator> vector(InputIterator first, InputIterator last, const Allocator& = Allocator());
This compiles, but will result in code running amok. The potential trap (that cannot be easily detected by the library implementation) could be reduced by making this constructor explicit. It would still have the effect to be selected here, but the code would be ill-formed, so the programmer gets a clear message here.
[2014-06 Rapperswil]
JW: can't fix this, don't want to touch this, Do The Right Thing clause has been a source of tricky issues. only really happens with string literals, that's the only way to create an array that isn't obviously an array
GR: want to see paper AM: is it only string literals, or also UDLs? STL: maybe, but we don't need to deal with that. This is only a problem in a very specific case Leave as Open.Proposed resolution:
Section: 17.3.5 [numeric.limits] Status: New Submitter: Pete Becker Opened: 2013-03-08 Last modified: 2018-11-08
Priority: 4
View other active issues in [numeric.limits].
View all other issues in [numeric.limits].
View all issues with New status.
Discussion:
This member should probably be named "is_ieee754". Or at least the standard should explain that IEC-559 no longer exists, and that it's been superseded by IEEE-754.
[2016-06, Oulu]
The ISO version of the standard is ISO/IEC/IEEE 60559:2011,
which C11 Annex F refers to as IEC 60559
(although C still refers to it as IEC 559 in the __STDC_IEC_559__
macro).
Proposed resolution:
Section: 20.3.1.3 [unique.ptr.single] Status: Open Submitter: Rob Desbois Opened: 2013-05-15 Last modified: 2017-03-21
Priority: 3
View other active issues in [unique.ptr.single].
View all other issues in [unique.ptr.single].
View all issues with Open status.
Discussion:
N3337 20.3.1.3.6 [unique.ptr.single.modifiers] contains 2 non-normative notes stating:
[para 4]: "The order of these operations is significant because the call to get_deleter() may destroy *this."
[para 5]: "The postcondition does not hold if the call to get_deleter() destroys *this since this->get() is no longer a valid expression."
It seems this wording was created to resolve 998 due to the possibility that a unique_ptr may be destroyed through deletion of its stored pointer where that directly or indirectly refers to the same unique_ptr. If unique_ptr is required to support circular references then it seems this must be normative text: an implementation is currently allowed to operate on *this after the assignment and deletion specified in para 4, since this is only 'disallowed' by the non-normative note.
I propose the following draft rewording:
[para 4]: Effects: assigns p to the stored pointer, and then if the old value of the stored pointer, old_p, was not
equal to nullptr, calls get_deleter()(old_p). No operation shall be performed after the call to
get_deleter()(old_p) that requires *this to be valid, because the deletion may destroy *this if it is
referred to directly or indirectly by the stored pointer. [Note: The order of these operations is significant
because the call to get_deleter() may destroy *this. — end note]
I expect it will also be necessary to amend the requirements for a deleter, so in addition:
20.3.1.3 [unique.ptr.single] [para 1]: The default type for the template parameter D is default_delete. A client-supplied template argument D shall be a function object type (20.10), lvalue-reference to function, or lvalue-reference to function object type for which, given a value d of type D and a value ptr of type unique_ptr<T, D>::pointer, the expression d(ptr) is valid and has the effect of disposing of the pointer as appropriate for that deleter. Where D is not an lvalue reference type, d(ptr) shall be valid if ptr refers directly or indirectly to the invoking unique_ptr object.
[2013-10-05, Stephan T. Lavavej comments and provides alternative wording]
In Chicago, we determined that the original proposed change to 20.3.1.3 [unique.ptr.single]/1 was insufficient, because d might be a reference to a deleter functor that's destroyed during self-destruction.
We believed that 20.3.1.3.6 [unique.ptr.single.modifiers]/4 was already sufficiently clear. The Standard occasionally prevents implementations of X from doing various things, through the principle of "nothing allows X to fail in that situation". For example, v.push_back(v[0]) is required to work for non-empty vectors because nothing allows that to fail. In this case, the intent to allow self-destruction is already clear. Additionally, we did not believe that 20.3.1.3.6 [unique.ptr.single.modifiers]/5 had to be changed. The current note is slightly squirrely but it does not lead to confusion for implementers or users.Previous resolution from Rob Desbois:
Edit 20.3.1.3 [unique.ptr.single] p1 as indicated:
The default type for the template parameter D is default_delete. A client-supplied template argument D shall be a function object type (20.10), lvalue-reference to function, or lvalue-reference to function object type for which, given a value d of type D and a value ptr of type unique_ptr<T, D>::pointer, the expression d(ptr) is valid and has the effect of disposing of the pointer as appropriate for that deleter. Where D is not an lvalue reference type, d(ptr) shall be valid if ptr refers directly or indirectly to the invoking unique_ptr object.
Edit 20.3.1.3.6 [unique.ptr.single.modifiers] p4+5 as indicated:
void reset(pointer p = pointer()) noexcept;-3- Requires: The expression get_deleter()(get()) shall be well formed, shall have well-defined behavior, and shall not throw exceptions.
-4- Effects: assigns p to the stored pointer, and then if the old value of the stored pointer, old_p, was not equal to nullptr, calls get_deleter()(old_p). No operation shall be performed after the call to get_deleter()(old_p) that requires *this to be valid, because the deletion may destroy *this if it is referred to directly or indirectly by the stored pointer.[Note: The order of these operations is significant because the call to get_deleter() may destroy *this. — end note]-5- Postconditions: If the call get_deleter()(old_p) destroyed *this, none. Otherwise, get() == p.[Note: The postcondition does not hold if the call to get_deleter() destroys *this since this->get() is no longer a valid expression. — end note]
Previous resolution [SUPERSEDED]:
This wording is relative to N3691.
Edit 20.3.1.3 [unique.ptr.single] p1 as indicated:
The default type for the template parameter D is default_delete. A client-supplied template argument D shall be a function object type (20.10), lvalue-reference to function, or lvalue-reference to function object type for which, given a value d of type D and a value ptr of type unique_ptr<T, D>::pointer, the expression d(ptr) is valid and has the effect of disposing of the pointer as appropriate for that deleter. d(ptr) shall be valid even if it triggers the destruction of d or (if D is an lvalue reference to function object type) the function object that d refers to.
[2015-05, Lenexa]
After some discussion in Lenexa there was some wavering on if the added sentence is necessary. Here is example code that demonstrates why the extra sentence is necessary. In this example the call to d(ptr) is valid, however the deleter references *this after destructing its element:
#include <cassert> #include <memory> #include <iostream> class Deleter { int state_ = 0; enum { destructed = -4, self_move_assigned = -3, move_assigned_from = -2, move_constructed_from = -1 }; public: ~Deleter() {state_ = destructed;} Deleter() = default; Deleter(Deleter const&) = default; Deleter& operator=(Deleter const&) = default; Deleter(Deleter&& a) noexcept : state_(a.state_) {a.state_ = move_constructed_from;} Deleter& operator=(Deleter&& a) noexcept { if (this == &a) state_ = self_move_assigned; else { state_ = a.state_; a.state_ = move_assigned_from; } return *this; } Deleter(int state) : state_(state) { assert(state >= 0); } template <class T> void operator()(T* t) const { std::cout << "Deleter beginning operator()(T*)\n"; std::cout << "The deleter = " << *this << '\n'; std::cout << "Deleter about to destruct the X.\n"; delete t; std::cout << "Deleter has destructed the X.\n"; std::cout << "The deleter = " << *this << '\n'; std::cout << "Deleter ending operator()(T*)\n"; } friend std::ostream& operator<<(std::ostream& os, const Deleter& a) { switch (a.state_) { case destructed: os << "**destructed**"; break; case self_move_assigned: os << "self_move_assigned"; break; case move_assigned_from: os << "move_assigned_from"; break; case move_constructed_from: os << "move_constructed_from"; break; default: os << a.state_; break; } return os; } }; struct X { Deleter deleter_{1}; }; int main() { auto xp = new X; { std::unique_ptr<X, Deleter&> p(xp, xp->deleter_); std::cout << "unique_ptr is constructed.\n"; std::cout << "The deleter = " << p.get_deleter() << '\n'; std::cout << "Destructing unique_ptr...\n"; } std::cout << "unique_ptr is destructed.\n"; }
Which outputs:
unique_ptr is constructed. The deleter = 1 Destructing unique_ptr... Deleter beginning operator()(T*) The deleter = 1 Deleter about to destruct the X. Deleter has destructed the X. The deleter = **destructed** Deleter ending operator()(T*) unique_ptr is destructed.
The line "The deleter = **destructed**" represents the deleter referencing itself after it has been destructed by the d(ptr) expression, but prior to that call returning.
Suggested alternative to the current proposed wording:The expression d(ptr) shall not refer to the object d after it executes ptr->~T().
[2015-07, Telecon]
Geoffrey: Deleter may or may not execute ~T().
Alisdair: After the destructor after the element has run. Say it in words instead of code.
Howard will provide updated wording. Perhaps need both normative and non-normative wording.
[2015-08-03, Howard updates P/R per telecon discussion.]
[2017-03-04, Kona]
This is related to 2751, which has been suggested NAD.
STL wants "Effects equivalent to" here - say it in code. Marshall to research.
Proposed resolution:
This wording is relative to N4431.
Edit 20.3.1.3 [unique.ptr.single] p1 as indicated:
The default type for the template parameter D is default_delete. A client-supplied template argument D shall be a function object type (20.9), lvalue-reference to function, or lvalue-reference to function object type for which, given a value d of type D and a value ptr of type unique_ptr<T, D>::pointer, the expression d(ptr) is valid and has the effect of disposing of the pointer as appropriate for that deleter. The expression d(ptr), if it destructs the object referred to by ptr, shall not refer to the object d after it destructs *ptr. [Note: The object being destructed may control the lifetime of d. — end note]
Section: 33.5.4 [atomics.order] Status: Open Submitter: Brian Demsky Opened: 2013-06-17 Last modified: 2016-01-28
Priority: 4
View other active issues in [atomics.order].
View all other issues in [atomics.order].
View all issues with Open status.
Discussion:
I believe that the following variation on IRIW should admit executions in which c1 = d1 = 5 and c2 = d2 = 0. If this is allowed, then what is sequence of program evaluations for 33.5.4 [atomics.order] p9 that justifies the store to z? It seems that 33.5.4 [atomics.order] p9 should not allow this execution because one of the stores to x or y has to appear earlier in the sequence, each of the fetch_adds reads the previous load in the thread (and thus must appear later in the sequence), and 33.5.4 [atomics.order] p9 states that each load must read from the last prior assignment in the sequence.
atomic_int x; atomic_int y; atomic_int z; int c1, c2, d1, d2; static void a(void* obj) { atomic_store_explicit(&x, 5, memory_order_relaxed); } static void b(void* obj) { atomic_store_explicit(&y, 5, memory_order_relaxed); } static void c(void* obj) { c1 = atomic_load_explicit(&x, memory_order_relaxed); // this could also be an atomic load if the address depends on c1: c2 = atomic_fetch_add_explicit(&y, c1, memory_order_relaxed); } static void d(void* obj) { d1 = atomic_load_explicit(&y, memory_order_relaxed); d2 = atomic_fetch_add_explicit(&x, d1, memory_order_relaxed); } int user_main(int argc, char** argv) { thrd_t t1, t2, t3, t4; atomic_init(&x, 0); atomic_init(&y, 0); printf("Main thread: creating 4 threads\n"); thrd_create(&t1, (thrd_start_t)&a, NULL); thrd_create(&t2, (thrd_start_t)&b, NULL); thrd_create(&t3, (thrd_start_t)&c, NULL); thrd_create(&t4, (thrd_start_t)&d, NULL); thrd_join(t1); thrd_join(t2); thrd_join(t3); thrd_join(t4); printf("c1=%d c2=%d\n",c1,c2); printf("d1=%d d2=%d\n",d1,d2); // Can this store write 1000 (i.e., c1=d1=5, c2=d2=0)? atomic_store(&z, (c1+d1)*100+c2+d2); printf("Main thread is finished\n"); return 0; }
It seems that the easiest fix is to allow a load in 33.5.4 [atomics.order] p9 to read from any prior store in the evaluation order.
That said, I would personally advocate the following: It seems to me that C/C++ atomics are in a bit of different situation than Java because:People are expected to use relaxed C++ atomics in potentially racy situations, so it isn't clear that semantics as complicated as the JMM's causality would be sane.
People who use C/C++ atomics are likely to be experts and use them in a very controlled fashion. I would be really surprised if compilers would find any real wins by optimizing the use of atomics.
Why not do something like:
There is satisfaction DAG of all program evaluations. Each evaluation observes the values of variables as computed by some prior assignment in the DAG. There is an edge x->y between two evaluations x and y if:the evaluation y observes a value computed by the evaluation x or
the evaluation y is an atomic store, the evaluation x is an atomic load, and there is a condition branch c that may depend (intrathread dependence) on x and x-sb->c and c-sb->y.
This seems to allow reordering of relaxed atomics that processors do without extra fence instructions, allows most reorderings by the compiler, and gets rid of satisfaction cycles.
[2015-02 Cologne]
Handed over to SG1.
[2015-05 Lenexa, SG1 response]
This was partially addressed (weasel-worded) in C++14 (See N3786). The remainder is an open research problem. N3710 outlines a "solution" that doesn't have a consensus behind it because it costs performance. We have no better solution at the moment.
Proposed resolution:
Section: 27.8.2.4 [partial.sort.copy] Status: New Submitter: Matt Austern Opened: 2013-06-26 Last modified: 2016-01-28
Priority: 3
View all issues with New status.
Discussion:
The signature of this function is:
template<class InputIterator, class RandomAccessIterator> RandomAccessIterator partial_sort_copy(InputIterator first, InputIterator last, RandomAccessIterator result_first, RandomAccessIterator result_last);
(and the usual overload for an explicitly provided comparison function). The standard says nothing about requirements in the case where the input type (iterator_traits<InputIterator>::value_type) and the output type (iterator_traits<RandomAccessIterator>::value_type) are different.
Presumably the input type must be convertible to the output type. What's less clear is what the requirements are on the comparison operator. Does the algorithm only perform comparisons on two values of the output type, or does it also perform comparisons on values of the input type, or might it even perform heterogeneous comparisons?Proposed resolution:
Section: 24.2.2.1 [container.requirements.general] Status: New Submitter: Matt Austern Opened: 2013-06-26 Last modified: 2016-01-28
Priority: 4
View other active issues in [container.requirements.general].
View all other issues in [container.requirements.general].
View all issues with New status.
Discussion:
Consider the following code snippet:
#include <vector> #include <algorithm> int main() { std::vector<int> v1(100, 3); std::vector<int> v2(100); copy(v1.begin(), v1.end(), v2.begin()); }
It compiles without error on my desktop. Is it required to? I can't find evidence from the standard that it is. In my test std::copy was found by argument-dependent lookup because the implementation I used made std::vector<int>::iterator a user-defined type defined in namespace std. But the standard only requires std::vector<int>::iterator to be an implementation specified random access iterator type. I can't find anything requiring it to be a user-defined type at all (and in fact there are reasonable implementation where it isn't), let alone a user defined type defined in a specific namespace.
Since the defining namespace of container iterators is visible to users, should the standard say anything about what that namespace is?
Proposed resolution:
Section: 31.8.2.5 [stringbuf.virtuals] Status: Open Submitter: Sergey Zubkov Opened: 2013-08-29 Last modified: 2018-06-12
Priority: 4
View other active issues in [stringbuf.virtuals].
View all other issues in [stringbuf.virtuals].
View all issues with Open status.
Discussion:
In 31.8.2.5 [stringbuf.virtuals]/1, basic_stringbuf::underflow() is specified to unconditionally return traits::eof() when a read position is not available.
The semantics of basic_stringbuf require, and existing libraries implement it so that this function makes a read position available if possible to do so, e.g. if some characters were inserted into the stream since the last call to overflow(), resulting in pptr() > egptr(). Compare to the conceptually similar D.12.2.4 [depr.strstreambuf.virtuals]/15.[2018-06-06, Billy argues for NAD]
The existing "Any character in the underlying buffer which has been initialized is considered to be part of the input sequence." sentence already describes what the stringbuf is supposed to do to the get area. The specific mechanism that the stringbuf uses to alter the get area is unspecified because the mechanism by which the stringbuf remembers the "high water mark" is unspecified.
Consider the following:stringstream s; s << "Hello"; s.seekp(0); string x; s >> x;
Before this P/R, this will store Hello in x, because the characters Hello are initialized. After this P/R, the "written put area" is empty, so it will store the empty string in x.
Saying that the initialized part of the string is used already describes what needs to happen here.[2018-06 Rapperswil Wednesday issues processing]
Billy to provide rationale for closing as NAD.
Proposed resolution:
This wording is relative to N3691.
Change 31.8.2.5 [stringbuf.virtuals] as indicated:
int_type underflow();-1- Returns: If the input sequence has a read position available or the function makes a read position available (as described below), returns traits::to_int_type(*gptr()). Otherwise, returns traits::eof(). Any character in the underlying buffer which has been initialized is considered to be part of the input sequence.
-?- The function can make a read position available only if (mode & ios_base::in) != 0 and if the write next pointer pptr() is not null and is greater than the current read end pointer egptr(). To make a read position available, the function alters the read end pointer egptr() to equal pptr().
Section: 22.3.2 [pairs.pair], 22.4.4.1 [tuple.cnstr], 29.5 [time.duration] Status: Open Submitter: Daniel Krügler Opened: 2013-09-09 Last modified: 2020-06-13
Priority: 3
View other active issues in [pairs.pair].
View all other issues in [pairs.pair].
View all issues with Open status.
Discussion:
During the acceptance of N3471 and some similar constexpr papers, specific wording was added to pair, tuple, and other templates that were intended to impose implementation constraints that ensure that the observable constexpr "character" of a defaulted function template is solely determined by the required expressions of the user-provided types when instantiated, for example:
The defaulted move and copy constructor, respectively, of pair shall be a constexpr function if and only if all required element-wise initializations for copy and move, respectively, would satisfy the requirements for a constexpr function.
This wording doesn't require enough, especially since the core language via CWG 1358 does now support constexpr function template instantiations, even if such function cannot appear in a constant expression (as specified in 7.7 [expr.const]) or as a constant initializer of that object (as specified in [basic.start.init]). The wording should be improved and should require valid uses in constant expressions and as constant initializers instead.
[Lenexa 2015-05-05]
STL : notice order of move/copy and copy/move with "respectively".
General word-smithing; ask for updated wording
Are we happy with this with changes we are suggesting?
unanimous
[2016-12-14, Daniel comments]
LWG 2833 overlaps considerably and both should be resolved together.
Previous resolution from Daniel [SUPERSEDED]:This wording is relative to N3691.
Change 22.3.2 [pairs.pair] p2 as indicated:
-2-
The defaulted move and copy constructor, respectively, of pair shall be a constexpr function if and only if all required element-wise initializations for copy and move, respectively, would satisfy the requirements for a constexpr functionAn invocation of the move or copy constructor of pair shall be a constant expression (7.7 [expr.const]) if all required element-wise initializations would be constant expressions. An invocation of the move or copy constructor of pair shall be a constant initializer for that pair object ( [basic.start.init]) if all required element-wise initializations would be constant initializers for the respective subobjects.Change 22.4.4.1 [tuple.cnstr] p2 as indicated:
-2-
The defaulted move and copy constructor, respectively, of tuple shall be a constexpr function if and only if all required element-wise initializations for copy and move, respectively, would satisfy the requirements for a constexpr function. The defaulted move and copy constructor of tuple<> shall be constexpr functionsAn invocation of the move or copy constructor of tuple shall be a constant expression (7.7 [expr.const]) if all required element-wise initializations would be constant expressions. An invocation of the move or copy constructor of tuple shall be a constant initializer for that tuple object ( [basic.start.init]) if all required element-wise initializations would be constant initializers for the respective subobjects. An invocation of the move or copy constructor of tuple<> shall be a constant expression, or a constant initializer for that tuple<> object, respectively, if the function argument would be constant expression.Change 29.5 [time.duration] p7 as indicated:
-7- Remarks:
The defaulted copy constructor of duration shall be a constexpr function if and only if the required initialization of the member rep_ for copy and move, respectively, would satisfy the requirements for a constexpr function.An invocation of the copy constructor of duration shall be a constant expression (7.7 [expr.const]) if the required initialization of the member rep_ would be a constant expression. An invocation of the copy constructor of duration shall be a constant initializer for that duration object ( [basic.start.init]) if the required initialization of the member rep_ would be constant initializers for this subobject.
[2020-06-08 Nina Dinka Ranns comments]
The revised wording provided by LWG 2833 should resolve this issue as well.
Proposed resolution:
Section: 21 [meta] Status: Open Submitter: Daniel Krügler Opened: 2013-09-02 Last modified: 2016-01-28
Priority: 3
View other active issues in [meta].
View all other issues in [meta].
View all issues with Open status.
Discussion:
The current library specification uses at several places wording that is intended to refer to core language template deduction failure at the top-level of expressions (aka "SFINAE"), for example:
The expression declval<T>() = declval<U>() is well-formed when treated as an unevaluated operand (Clause 5). Access checking is performed as if in a context unrelated to T and U. Only the validity of the immediate context of the assignment expression is considered. [Note: The compilation of the expression can result in side effects such as the instantiation of class template specializations and function template specializations, the generation of implicitly-defined functions, and so on. Such side effects are not in the "immediate context" and can result in the program being ill-formed. — end note]
Similar wording can be found in the specification of result_of, is_constructible, and is_convertible, being added to resolve an NB comment by LWG 1390 and 1391 through N3142.
This wording is necessary to limit speculative compilations needed to implement these traits, but it is also lengthy and repetitive.[2014-05-19, Daniel suggests a descriptive term]
constrictedly well-formed expression:
An expression e depending on a set of types A1, ..., An which is well-formed when treated as an unevaluated operand (Clause 5). Access checking is performed as if in a context unrelated to A1, ..., An. Only the validity of the immediate context of e is considered. [Note: The compilation of the expression can result in side effects such as the instantiation of class template specializations and function template specializations, the generation of implicitly-defined functions, and so on. Such side effects are not in the "immediate context" and can result in the program being ill-formed. — end note][2014-05-20, Richard and Jonathan suggest better terms]
Richard suggested "locally well-formed"
Jonathan suggested "contextually well-formed" and then "The expression ... is valid in a contrived argument deduction context"[2014-06-07, Daniel comments and improves wording]
The 2014-05-19 suggestion did only apply to expressions, but there are two important examples that are not expressions, but instead are involving an object definition (std::is_constructible) and a function definition (std::is_convertible), respectively, instead. Therefore I suggest to rephrase the usage of "expression" into "program construct" in the definition of Jonathan's suggestion of "valid in a contrived argument deduction context".
I would like to point out that given the new definition of "valid in a contrived argument deduction context", there are several other places of the Library specification that could take advantage of this wording to improve the existing specification, such as 22.10.17.3 [func.wrap.func] p2, most functions in 20.2.9.3 [allocator.traits.members], and the **Insertable, EmplaceConstructible, and Erasable definitions in 24.2.2.1 [container.requirements.general], but given that these are not fully described in terms of the aforementioned wording yet, I would recommend to fix them by a separate issue once the committee has agreed on following the suggestion presented by this issue.[2015-05-05 Lenexa: Move to Open]
...
MC: I think we like the direction but it isn't quite right: it needs some work
JW: I'm prepared to volunteer to move that further, hopefully with the help of Daniel
Roger Orr: should this be Core wording because it doesn't really have anything to do with libraries - the term could then just be used here
AM: Core has nothing to deal with that, though
HT: it seems there is nothing to imply that allows dropping out with an error - maybe that's a separate issue
MC: I'm not getting what you are getting at: could you write an issue? - any objection to move to Open?
...
Proposed resolution:
This wording is relative to N3936.
Add the following new definition to [definitions] as indicated:
valid in a contrived argument deduction context [defns.valid.contr.context]
A program construct c depending on a set of types A1, ..., An, and treated as an unevaluated operand (Clause 5) when c is an expression, which is well-formed. Access checking is performed as if in a context unrelated to A1, ..., An. Only the validity of the immediate context (13.10.3 [temp.deduct]) of c is considered. [Note: The compilation of c can result in side effects such as the instantiation of class template specializations and function template specializations, the generation of implicitly-defined functions, and so on. Such side effects are not in the "immediate context" and can result in the program being ill-formed. — end note].Change Table 49 ("Type property predicates") as indicated:
Table 49 — Type property predicates Template Condition Preconditions … template <class T, class U>
struct is_assignable;The expression declval<T>() =
declval<U>() is valid in a
contrived argument deduction context
([defns.valid.contr.context]) for types
T and U.well-formed when treated
as an unevaluated operand
(Clause 5). Access
checking is performed as if
in a context unrelated to T
and U. Only the validity of
the immediate context of
the assignment expression
is considered. [Note: The
compilation of the
expression can result in
side effects such as the
instantiation of class
template specializations
and function template
specializations, the
generation of
implicitly-defined
functions, and so on. Such
side effects are not in the
"immediate context" and
can result in the program
being ill-formed. — end
note][…] …
Change 21.3.5.4 [meta.unary.prop] p7 as indicated:
-7- Given the following function prototype:
template <class T> add_rvalue_reference_t<T> create() noexcept;the predicate condition for a template specialization is_constructible<T, Args...> shall be satisfied if and only if the following variable definition
would be well-formedfor some invented variable t would be valid in a contrived argument deduction context ([defns.valid.contr.context]) for types T and Args...:T t(create<Args>()...);[Note: These tokens are never interpreted as a function declaration. — end note]
Access checking is performed as if in a context unrelated to T and any of the Args. Only the validity of the immediate context of the variable initialization is considered. [Note: The evaluation of the initialization can result in side effects such as the instantiation of class template specializations and function template specializations, the generation of implicitly-defined functions, and so on. Such side effects are not in the "immediate context" and can result in the program being ill-formed. — end note]
Change Table 57 ("Other transformations") as indicated:
Table 57 — Other transformations Template Condition Comments … template <class Fn, class... ArgTypes>
struct result_of<Fn(ArgTypes...)>;[…] If the expression
INVOKE(declval<Fn>(),
declval<ArgTypes>()...) is
valid in a contrived argument deduction
context ([defns.valid.contr.context]) for types
Fn and ArgTypes...well, the
formed when treated as an
unevaluated operand (Clause 5)
member typedef type shall name the
type
decltype(INVOKE(declval<Fn>(),
declval<ArgTypes>()...));
otherwise, there shall be no member
type.Access checking is performed as
if in a context unrelated to Fn and
ArgTypes. Only the validity of the
immediate context of the expression is
considered. [Note: The compilation of
the expression can result in side
effects such as the instantiation of
class template specializations and
function template specializations, the
generation of implicitly-defined
functions, and so on. Such side effects
are not in the "immediate context"
and can result in the program being
ill-formed. — end note]…
Change 21.3.7 [meta.rel] p4 as indicated:
-4- Given the following function prototype:
template <class T> add_rvalue_reference_t<T> create() noexcept;the predicate condition for a template specialization is_convertible<From, To> shall be satisfied if and only if the return expression in the following code would be
well-formedvalid in a contrived argument deduction context ([defns.valid.contr.context]) for types To and From, including any implicit conversions to the return type of the function:To test() { return create<From>(); }[Note: This requirement gives well defined results for reference types, void types, array types, and function types. — end note]
Access checking is performed as if in a context unrelated to To and From. Only the validity of the immediate context of the expression of the return-statement (including conversions to the return type) is considered. [Note: The evaluation of the conversion can result in side effects such as the instantiation of class template specializations and function template specializations, the generation of implicitly-defined functions, and so on. Such side effects are not in the "immediate context" and can result in the program being ill-formed. — end note]
Section: 30.3.1.3 [locale.cons] Status: Ready Submitter: Juan Soulie Opened: 2013-09-04 Last modified: 2022-11-09
Priority: 3
View other active issues in [locale.cons].
View all other issues in [locale.cons].
View all issues with Ready status.
Discussion:
30.3.1.3 [locale.cons] p14 ends with:
"[…] If f is null, the resulting object is a copy of other."
but the next line p15 says:
"Remarks: The resulting locale has no name."
But both can't be true when other has a name and f is null.
I've tried it on two implementations (MSVC,GCC) and they are inconsistent with each other on this.Daniel Krügler:
As currently written, the Remarks element applies unconditionally for all cases and thus should "win". The question arises whether the introduction of this element by LWG 424 had actually intended to change the previous Note to a Remarks element. In either case the wording should be improved to clarify this special case.[2022-02-14; Daniel comments]
This issue seems to have some overlap with LWG 3676 so both should presumably be resolved in a harmonized way.
[2022-11-01; Jonathan provides wording]
This also resolves 3673 and 3676.
[2022-11-04; Jonathan revises wording after feedback]
Revert an incorrect edit to p8, which was incorrectly changed to:
"If cats is equal to locale::none, the resulting locale has the same name as locale(std_name). Otherwise, the locale has a name if and only if other has a name."
[Kona 2022-11-08; Move to Ready status]
Proposed resolution:
This wording is relative to N4917.
Modify 30.3.1.3 [locale.cons] as indicated:
explicit locale(const char* std_name);-2- Effects: Constructs a locale using standard C locale names, e.g., "POSIX". The resulting locale implements semantics defined to be associated with that name.
-3- Throws: runtime_error if the argument is not valid, or is null.
-4- Remarks: The set of valid string argument values is "C", "", and any implementation-defined values.
explicit locale(const string& std_name);-5- Effects:
The same asEquivalent to locale(std_name.c_str()).locale(const locale& other, const char* std_name, category cats);-?- Preconditions: cats is a valid category value (30.3.1.2.1 [locale.category]).
-6- Effects: Constructs a locale as a copy of other except for the facets identified by the category argument, which instead implement the same semantics as locale(std_name).
-7- Throws: runtime_error if the second argument is not valid, or is null.
-8- Remarks: The locale has a name if and only if other has a name.
locale(const locale& other, const string& std_name, category cats);-9- Effects:
The same asEquivalent to locale(other, std_name.c_str(), cats).template<class Facet> locale(const locale& other, Facet* f);-10- Effects: Constructs a locale incorporating all facets from the first argument except that of type Facet, and installs the second argument as the remaining facet. If f is null, the resulting object is a copy of other.
-11- Remarks: If f is null, the resulting locale has the same name as other. Otherwise, the
Theresulting locale has no name.locale(const locale& other, const locale& one, category cats);-?- Preconditions: cats is a valid category value.
-12- Effects: Constructs a locale incorporating all facets from the first argument except for those that implement cats, which are instead incorporated from the second argument.
-13- Remarks: If cats is equal to locale::none, the resulting locale has a name if and only if the first argument has a name. Otherwise, the
Thelocale has a name if and only if the first two arguments both have names.
Section: 17.7.3.4 [new.delete.placement] Status: New Submitter: Daniel Krügler Opened: 2013-09-18 Last modified: 2016-01-28
Priority: 3
View all other issues in [new.delete.placement].
View all issues with New status.
Discussion:
The library gives explicit permission in 16.4.5.2.1 [namespace.std] p2 that user code may explicitly instantiate a library template provided that the instantiations depend on at least one user-defined type:
A program may explicitly instantiate a template defined in the standard library only if the declaration depends on the name of a user-defined type and the instantiation meets the standard library requirements for the original template.
But it seems that the C++11 library is not specified in a way that guarantees such an instantiation to be well-formed if the minimum requirements of the library is not satisfied.
For example, in general, the first template parameter of std::vector is not required to be DefaultConstructible in general, but due to the split of the single C++03 member function with default argumentvoid resize(size_type sz, T c = T());
into
void resize(size_type sz); void resize(size_type sz, const T& c);
the effect is now that for a type ND that is not DefaultConstructible, such as
struct NP { NP(int); };
the explicit instantiation of std::vector<ND> is no longer well-formed, because the attempt to instantiate the single-argument overload of resize cannot not succeed, because this function imposes the DefaultInsertable requirements and given the default allocator this effectively requires DefaultConstructible.
But DefaultConstructible is not the only point, what about CopyConstructible versus MoveConstructible alone? It turns out that currently the second resize overload would fail during an explicit instantiation for a type like
struct MO { MO() = default; MO(MO&&) = default; };
because it imposes CopyInsertable requirements that end up being equivalent to the CopyConstructible requirements for the default allocator.
Technically a library can solve these issues: For special member functions by defining them in some base class, for others by transforming them effectively into a function template due to the great feature of default template arguments for function templates (At the very moment the validity of the latter approach depends on a resolution of core language issue CWG 1635, though). E.g. the here mentioned resize functions of std::vector could be prevented from instantiation by defining them like this with an implementation:template<class = void> void resize(size_type sz) { […] } template<class = void> void resize(size_type sz, const T& c) { […] }
In this case, these functions could also be defined in a base class, but the latter approach won't work in all cases.
Basically such an implementation is required to constrain all member functions that are not covered by the general requirements imposed on the actual library template parameters. I tested three different C++11 library implementations and but none could instantiate for example std::list, std::vector, or std::deque with value types that are not DefaultConstructible or only MoveConstructible. This issue is raised to clarify the current situation in regard to the actual requirements imposed on user-provided types that are used to explicitly instantiate Library-provided templates. For example, the current Container requirements impose very little requirements on the actual value type and it is unclear to which extend library implementations have to respect that. The minimum solution of this issue should be to at least realize that there is no fundamental requirement on DefaultConstructible for value types of library containers, because we have since C++03 the general statement of 16.4.4.2 [utility.arg.requirements] ("In general, a default constructor is not required."). It is unclear whether CopyConstructible should be required for an explicit instantiation request, but given the careful introduction of move operations in the library it would seem astonishing that a MoveConstructible type wouldn't suffice for value types of the container types. In any case I can envision at least two approaches to solve this issue:As indicated in LWG 2292, those function could get an explicit "Template Constraints:" element, albeit this promises more than needed to solve this issue.
The library could introduce a completely new element form, such as "Instantiation Constraints:" that would handle this situation for explicit instantiation situations. This would allow for simpler techniques to solve the issue when explicit instantiation is required compared to the first bullet, because it would not (necessarily) guarantee SFINAE-friendly expression-wellformedness, such as inspecting the expression std::declval<std::vector<ND>&>.resize(0) in an unevaluated context.
It should be noted that the 2013-08-27 comment to LWG 2193 could be resolved by a similar solution as indicated in this issue here.
Proposed resolution:
Section: 24 [containers] Status: LEWG Submitter: Zhihao Yuan Opened: 2013-09-26 Last modified: 2018-11-12
Priority: 2
View other active issues in [containers].
View all other issues in [containers].
View all issues with LEWG status.
Discussion:
LWG 2193 yields explicit for default ctors to allow {}, but not for all cases of uniform initialization. For example:
explicit vector(size_type count, const Allocator& alloc = Allocator());
This prevents {n, alloc()}. Although this use is relatively rare, but the behavior is inconsistent with that of
vector(size_type count, const T& value, const Allocator& alloc = Allocator());
[Urbana 2014-11-07: Move to Open]
[2018-08 Batavia Monday issue discussion]
This really needs a paper; splitting a lot of constructors. Nevin to write paper.
[2018-11 San Diego Thursday night issue processing]
LEWG has rejected Nevin's paper, so they need to formulate a policy.
Proposed resolution:
Section: 24.2.2.1 [container.requirements.general] Status: Open Submitter: Stephan T. Lavavej Opened: 2013-09-21 Last modified: 2018-08-24
Priority: 3
View other active issues in [container.requirements.general].
View all other issues in [container.requirements.general].
View all issues with Open status.
Discussion:
24.2.2.1 [container.requirements.general]/10 says that unless otherwise specified, "no swap() function invalidates any references, pointers, or iterators referring to the elements of the containers being swapped. [Note: The end() iterator does not refer to any element, so it may be invalidated. — end note]". However, move constructors and move assignment operators aren't given similar invalidation guarantees. The guarantees need several exceptions, so I do not believe that blanket language like /11 "Unless otherwise specified (either explicitly or by defining a function in terms of other functions), invoking a container member function or passing a container as an argument to a library function shall not invalidate iterators to, or change the values of, objects within that container." is applicable.
[2014-02-13 Issaquah]
General agreeement on intent, several wording nits and additional paragraphs to hit.
STL to provide updated wording. Move to Open.
[2015-02 Cologne]
AM: in the proposed wording, I'd like to mention that the iterators now refer to elements of a different container. I think we're saying something like this somewhere. JY: There's some wording like that for swap I think. TK: It's also in list::splice(). DK to JY: 23.2.1p9.
VV: The issue says that STL was going to propose new wording. Has he done that? AM: I believe we're looking at that. GR: The request touches on multiple paragraphs, and this PR has only one new paragraph, so this looks like it's not up-to-date. MC: This was last updated a year ago in Issaquah. Conclusion: Skip, not up to date.[2015-06, Telecon]
Still waiting for updated wording
[2015-08 Chicago]
Still waiting for updated wording
[2018-08-23 Batavia Issues processing]
Priority to 3
Proposed resolution:
This wording is relative to N3691.
In 24.2.2.1 [container.requirements.general]/10 change as indicated:
-10- Unless otherwise specified (see 23.2.4.1, 23.2.5.1, 23.3.3.4, and 23.3.7.5) all container types defined in this Clause meet the following additional requirements:
[…]
no copy constructor or assignment operator of a returned iterator throws an exception.
no move constructor (or move assignment operator when allocator_traits<allocator_type>::propagate_on_container_move_assignment::value is true) of a container (except for array) invalidates any references, pointers, or iterators referring to the elements of the source container. [Note: The end() iterator does not refer to any element, so it may be invalidated. — end note]
no swap() function throws an exception.
no swap() function invalidates any references, pointers, or iterators referring to the elements of the containers being swapped. [Note: The end() iterator does not refer to any element, so it may be invalidated. — end note]
Section: 32.4.2 [re.synopt] Status: Open Submitter: Stephan T. Lavavej Opened: 2013-09-21 Last modified: 2016-01-28
Priority: 3
View all other issues in [re.synopt].
View all issues with Open status.
Discussion:
The table in 32.4.2 [re.synopt]/1 says that regex_constants::collate "Specifies that character ranges of the form "[a-b]" shall be locale sensitive.", but 32.12 [re.grammar]/14 says that it affects individual character comparisons too.
[2012-02-12 Issaquah : recategorize as P3]
Marshall Clow: 28.13/14 only applies to ECMAScript
All: we're unsure
Jonathan Wakely: we should ask John Maddock
Move to P3
[2014-5-14, John Maddock response]
The original intention was the original wording: namely that collate only made character ranges locale sensitive. To be frank it's a feature that's probably hardly ever used (though I have no real hard data on that), and is a leftover from early POSIX standards which required locale sensitive collation for character ranges, and then later changed to implementation defined if I remember correctly (basically nobody implemented locale-dependent collation).
So I guess the question is do we gain anything by requiring all character-comparisons to go through the locale when this bit is set? Certainly it adds a great deal to the implementation effort (it's not what Boost.Regex has ever done). I guess the question is are differing code-points that collate identically an important use case? I guess there might be a few Unicode code points that do that, but I don't know how to go about verifying that. STL: If this was unintentional, then 32.4.2 [re.synopt]/1's table should be left alone, while 32.12 [re.grammar]/14 should be changed instead. Jeffrey Yasskin: This page mentions that [V] in Swedish should match "W" in a perfect world. However, the most recent version of TR18 retracts both language-specific loose matches and language-specific ranges because "for most full-featured regular expression engines, it is quite difficult to match under code point equivalences that are not 1:1" and "tailored ranges can be quite difficult to implement properly, and can have very unexpected results in practice. For example, languages may also vary whether they consider lowercase below uppercase or the reverse. This can have some surprising results: [a-Z] may not match anything if Z < a in that locale." ECMAScript doesn't include collation at all. IMO, +1 to changing 28.13 instead of 28.5.1. It seems like we'd be on fairly solid ground if we wanted to remove regex_constants::collate entirely, in favor of named character classes, but of course that's not for this issue.Proposed resolution:
This wording is relative to N3691.
In 32.4.2 [re.synopt]/1, Table 138 — "syntax_option_type effects", change as indicated:
Table 138 — syntax_option_type effects Element Effect(s) if set … collate Specifies that character ranges of the form "[a-b]"comparisons and character range comparisons shall be locale sensitive.…
Section: 32.6 [re.traits], 30.3.1.2.2 [locale.facet] Status: Open Submitter: Sergey Zubkov Opened: 2013-10-15 Last modified: 2016-02-01
Priority: 3
View all other issues in [re.traits].
View all issues with Open status.
Discussion:
32.6 [re.traits]/7, begins with "if typeid(use_facet<collate<charT> >) == typeid(collate_byname<charT>)", which appears to be pseudocode with the intention to convey that the collate facet has not been replaced by the user. Cf. the wording in N1429 "there is no portable way to implement transform_primary in terms of std::locale, since even if the sort key format returned by std::collate_byname<>::transform is known and can be converted into a primary sort key, the user can still install their own custom std::collate implementation into the locale object used, and that can use any sort key format they see fit.".
Taken literally, 32.6 [re.traits]/7 appears to imply that named locales are required to hold their collate facets with dynamic type std::collate_byname<charT>, which is in fact true in some implementations (e.g libc++), but not others (e.g. libstdc++). This does not follow from the description of _byname in 30.3.1.2.2 [locale.facet]/4, which is only required to provide equivalent semantics, to the named locale's facet, not to actually be one.[2015-05-06 Lenexa: Move to Open]
MC, RP: Consequence of failing to follow the rule is UB.
MC: Tightening of requirements.
RP: It should be this way, we just didn't impose it before.
MC: Second change is a bug fix, original code didn't work.
TK: Doesn't seem to make things worse.
Bring up in larger group tomorrow.
JW arrives.
JW: libstdc++ violates this due to two std::string ABIs.
JW: This prevents installing a type derived from Facet_byname, constrains the implementor from using a smarter derived class version.
JW: Can't look at facet id to detect replacement, because replacements have the same id.
RP: Can you give it multiple ids through multiple inheritance?
JW: No, the facet mechanism wouldn't like that.
JW: We should also ask Martin Sebor, he's implemented this stuff recently.
MC: Sounds like this resolution doesn't work, need a better solution.
JW: Write in words "if the facet has not been replaced by the user", the implementation knows how to detect that, but not like this.
RP: User RE traits need to detect this too.
JW: =(
Move to Open, JW will invite Martin Sebor to join LWG for discussion.
Later ...
JW: This is not needed for user specializations after all.
MC: Agree, [re.traits]/7 only applies to the stdlib traits.
NM: Effects: doesn't make sense.
JW, NM, Martin Sebor to come up with new wording.
Proposed resolution:
This wording is relative to N3691.
Modify 30.3.1.2.2 [locale.facet]/4 as indicated:
For some standard facets a standard "..._byname" class, derived from it, implements the virtual function semantics
equivalent toprovided by that facet of the locale constructed by locale(const char*) with the same name. Each such facet provides a constructor that takes a const char* argument, which names the locale, and a refs argument, which is passed to the base class constructor. Each such facet also provides a constructor that takes a string argument str and a refs argument, which has the same effect as calling the first constructor with the two arguments str.c_str() and refs. If there is no "..._byname" version of a facet, the base class implements named locale semantics itself by reference to other facets. For any locale loc constructed by locale(const char*) and facet Facet that has a corresponding standard Facet_byname class, typeid(use_facet<Facet>(loc)) == typeid(Facet_byname).
Modify 32.6 [re.traits]/7 as indicated:
template <class ForwardIterator> string_type transform_primary(ForwardIterator first, ForwardIterator last) const;-7- Effects: if typeid(use_facet<collate<charT> >(getloc())) == typeid(collate_byname<charT>) and the form of the sort key returned by collate_byname<charT>::transform(first, last) is known and can be converted into a primary sort key then returns that key, otherwise returns an empty string.
Section: 31.7.6.2 [ostream] Status: New Submitter: Alf P. Steinbach Opened: 2013-10-29 Last modified: 2016-01-28
Priority: 4
View all other issues in [ostream].
View all issues with New status.
Discussion:
For wide streams argument types wchar_t const* and wchar_t are supported only as template parameters. User defined conversions are not considered for template parameter matching. Hence inappropriate overloads of operator<< are selected when an implicit conversion is required for the argument, which is inconsistent with the behavior for char const* and char, is unexpected, and is a useless result.
Demonstration:#include <iostream> struct Byte_string { operator char const*() const { return "Hurray, it works!"; } }; struct Wide_string { operator wchar_t const*() const { return L"Hurray, it works!"; } }; struct Byte_ch { operator char() const { return 'X'; } }; struct Wide_ch { operator wchar_t() const { return L'X'; } }; auto main() -> int { using namespace std; wcout << "'X' as char value : " << Byte_ch() << endl; wcout << "'X' as wchar_t value: " << Wide_ch() << endl; wcout << "Byte string pointer : " << Byte_string() << endl; wcout << "Wide string pointer : " << Wide_string() << endl; }
Example output:
'X' as char value : X 'X' as wchar_t value: 88 Byte string pointer : Hurray, it works! Wide string pointer : 000803C8
Proposed resolution:
This wording is relative to N3797.
Modify 31.7.6.2 [ostream], class template basic_ostream synopsis, as indicated:
namespace std { […] // 27.7.3.6.4 character inserters template<class charT, class traits> basic_ostream<charT,traits>& operator<<(basic_ostream<charT,traits>&, charT); template<class charT, class traits> basic_ostream<charT,traits>& operator<<(basic_ostream<charT,traits>&, char); template<class traits> basic_ostream<char,traits>& operator<<(basic_ostream<char,traits>&, char); template<class traits> basic_ostream<wchar_t,traits>& operator<<(basic_ostream<wchar_t,traits>&, wchar_t); […] template<class charT, class traits> basic_ostream<charT,traits>& operator<<(basic_ostream<charT,traits>&, const charT*); template<class charT, class traits> basic_ostream<charT,traits>& operator<<(basic_ostream<charT,traits>&, const char*); template<class traits> basic_ostream<char,traits>& operator<<(basic_ostream<char,traits>&, const char*); template<class traits> basic_ostream<wchar_t,traits>& operator<<(basic_ostream<wchar_t,traits>&, const wchar_t*); […] }
Modify 31.7.6.3.4 [ostream.inserters.character] as indicated: [Drafting note: The replacement of os by out in p1 and the insertion of "out." in p4 just fix two obvious typos — end drafting note]
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); template<class traits> basic_ostream<wchar_t,traits>& operator<<(basic_ostream<wchar_t,traits>& out, wchar_t 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 as a formatted output function (31.7.6.3.1 [ostream.formatted.reqmts]) of out. Constructs a character sequence seq. If c has type char and the character type of the stream is not char, then seq consists of out.widen(c); otherwise seq consists of c. Determines padding for seq as described in 31.7.6.3.1 [ostream.formatted.reqmts]. Inserts seq into out. Calls
-2- Returns: out.osout.width(0).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<wchar_t,traits>& operator<<(basic_ostream<wchar_t,traits>& out, const wchar_t* 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);-3- Requires: s shall not be a null pointer.
-4- Effects: Behaves like a formatted inserter (as described in 31.7.6.3.1 [ostream.formatted.reqmts]) of out. Creates a character sequence seq of n characters starting at s, each widened using out.widen() (27.5.5.3), where n is the number that would be computed as if by:
traits::length(s) for the following overloads:
where the first argument is of type basic_ostream<charT, traits>& and the second is of type const charT*,
and also for the overloadwhere the first argument is of type basic_ostream<char, traits>& and the second is of type const char*,where the first argument is of type basic_ostream<wchar_t, traits>& and the second is of type const wchar_t*,
std::char_traits<char>::length(s) for the overload where the first argument is of type basic_ostream<charT, traits>& and the second is of type const char*,
traits::length(reinterpret_cast<const char*>(s)) for the other two overloads.
Determines padding for seq as described in 31.7.6.3.1 [ostream.formatted.reqmts]. Inserts seq into out. Calls out.width(0).
-5- Returns: out.
Section: 22.9.2 [template.bitset], 31.7.9 [quoted.manip] Status: Open Submitter: Zhihao Yuan Opened: 2013-12-02 Last modified: 2016-01-28
Priority: 3
View all other issues in [template.bitset].
View all issues with Open status.
Discussion:
Example: char16_t('1') != u'1' is possible.
The numeric value of char16_t is defined to be Unicode code point, which is same to the ASCII value and UTF-8 for 7-bit chars. However, char is not guaranteed to have an encoding which is compatible with ASCII. For example, '1' in EBCDIC is 241. I found three places in the standard casting narrow char literals: bitset::bitset, bitset::to_string and quoted. PJ confirmed this issue and says he has a solution used in their <filesystem> implementation, and he may want to propose it to the standard. The solution in my mind, for now, is to make those default arguments magical, where the "magic" can be implemented with a C11 _Generic selection (works in clang):#define _G(T, literal) _Generic(T{}, \ char: literal, \ wchar_t: L ## literal, \ char16_t: u ## literal, \ char32_t: U ## literal) _G(char16_t, '1') == u'1'
[Lenexa 2015-05-05: Move to Open]
Ask for complete PR (need quoted, to string, et al.)
Will then take it up again
Expectation is that this is correct way to fix this
Proposed resolution:
This wording is relative to N3797.
[Drafting note: This is a sample wording fixing only one case; I'm just too lazy to copy-paste it before we discussed whether the solution is worth and sufficient (for example, should the other `charT`s like `unsigned char` just don't compile without supplying those arguments? I hope so). — end drafting note]Modify 22.9.2 [template.bitset] p1, class template bitset synopsis, as indicated:
namespace std { template <size_t N> class bitset { public: […] template<class charT, class traits, class Allocator> explicit bitset( const basic_string<charT,traits,Allocator>& str, typename basic_string<charT,traits,Allocator>::size_type pos = 0, typename basic_string<charT,traits,Allocator>::size_type n = basic_string<charT,traits,Allocator>::npos, charT zero =charT('0')see below, charT one =charT('1')see below); […] }; […] }
Modify 22.9.2.2 [bitset.cons] as indicated:
template<class charT, class traits, class Allocator> explicit bitset(const basic_string<charT, traits, Allocator>& str, typename basic_string<charT, traits, Allocator>::size_type pos = 0, typename basic_string<charT, traits, Allocator>::size_type n = basic_string<charT, traits, Allocator>::npos, charT zero =charT('0')see below, charT one =charT('1')see below);-?- The default values of zero and one compare equal to the character literals 0 and 1 of type charT, respectively.
-3- Requires:: pos <= str.size(). […]
Section: 21.3.5.4 [meta.unary.prop] Status: Open Submitter: Richard Smith Opened: 2014-02-01 Last modified: 2017-02-02
Priority: 3
View other active issues in [meta.unary.prop].
View all other issues in [meta.unary.prop].
View all issues with Open status.
Discussion:
The 'Condition' for std::is_empty is listed as:
"T is a class type, but not a union type, with no non-static data members other than bit-fields of length 0, no virtual member functions, no virtual base classes, and no base class B for which is_empty<B>::value is false."
This is incorrect: there is no such thing as a non-static data member that is a bit-field of length 0, since bit-fields of length 0 must be unnamed, and unnamed bit-fields are not members (see 11.4.10 [class.bit] p2).
It also means that classes such as:struct S { int : 3; };
are empty (because they have no non-static data members). There's implementation divergence on the value of is_empty<S>::value.
I'm not sure what the purpose of is_empty is (or how it could be useful), but if it's desirable for the above type to not be treated as empty, something like this could work:"T is a class type, but not a union type, with no non-static data members
other than, no unnamed bit-fields of non-zero length0, no virtual member functions, no virtual base classes, and no base class B for which is_empty<B>::value is false."
and if the above type should be treated as empty, then this might be appropriate:
"T is a class type, but not a union type, with no (named) non-static data members
other than bit-fields of length 0, no virtual member functions, no virtual base classes, and no base class B for which is_empty<B>::value is false."
[2016-08 Chicago]
Walter says: We want is_empty_v<S> to produce false as a result. Therefore, we recommend adoption of the first of the issue's suggestions.
Tuesday AM: Moved to Tentatively Ready
Previous resolution [SUPERSEDED]:
[2016-10 by Marshall - this PR incorrectly highlighted changed portions]
Modify Table 38 — Type property predicates for is_empty as follows:
T is a non-union class type with no non-static data members
other than, no unnamed bit-fields of non-zero length0, no virtual member functions, no virtual base classes, and no base class B for which is_empty_v<B> is false.
[2016-10 Telecon]
Should probably point at section 1.8 for some of this. Status back to 'Open'
Proposed resolution:
Modify Table 38 — Type property predicates for is_empty as follows:
T
is a class type, but not a union type,is a non-union class type with no non-static data membersother than, no unnamed bit-fields of non-zero length0, no virtual member functions, no virtual base classes, and no base class B for which is_empty_v<B> is false.
Section: 24.2.7 [associative.reqmts], 24.2.8 [unord.req] Status: New Submitter: Jeffrey Yasskin Opened: 2014-02-15 Last modified: 2015-09-23
Priority: 3
View other active issues in [associative.reqmts].
View all other issues in [associative.reqmts].
View all issues with New status.
Discussion:
a_uniq.emplace(args) is specified as:
Effects: Inserts a value_type 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.
However, we occasionally find code of the form:
std::unique_ptr<Foo> p(new Foo); auto res = m.emplace("foo", std::move(p));
where we'd like to avoid destroying the Foo if the insertion doesn't take place (if the container already had an element with the specified key).
N3873 includes a partial solution to this in the form of a new emplace_stable member function, but LEWG's discussion strongly agreed that we'd rather have emplace() Just Work: Should map::emplace() be guaranteed not to move/copy its arguments if the insertion doesn't happen? SF: 8 F: 3 N: 0 A: 0 SA: 0 This poll was marred by the fact that we didn't notice or call out that emplace() must construct the key before doing the lookup, and it must not then move the key after it determines whether an insert is going to happen, and the mapped_type instance must live next to the key. The very similar issue 2006 was previously marked NAD, with N3178 as discussion. However, given LEWG's interest in the alternate behavior, we should reopen the question in this issue. We will need a paper that describes how to implement this before we can make more progress.Proposed resolution:
Section: 25.6.4 [istreambuf.iterator] Status: New Submitter: Hyman Rosen Opened: 2014-02-19 Last modified: 2014-06-17
Priority: 3
View other active issues in [istreambuf.iterator].
View all other issues in [istreambuf.iterator].
View all issues with New status.
Discussion:
Given the following code,
#include <sstream> std::stringbuf buf; std::istreambuf_iterator<char> begin(&buf); std::istreambuf_iterator<char> end;
it is not clear from the wording of the Standard whether begin.equal(end) must be true. In at least one implementation it is not (CC: Sun C++ 5.10 SunOS_sparc Patch 128228-25 2013/02/20) and in at least one implementation it is (gcc version 4.3.2 x86_64-unknown-linux-gnu).
25.6.4 [istreambuf.iterator] says that end is an end-of-stream iterator since it was default constructed. It also says that an iterator becomes equal to an end-of-stream iterator when end of stream is reached by sgetc() having returned eof(). 99 [istreambuf.iterator::equal] says that equal() returns true iff both iterators are end of stream or not end of stream. But there seems to be no requirement that equal check for end-of-stream by calling sgetc(). Jiahan Zi at BloombergLP discovered this issue through his code failing to work correctly. Dietmar Kühl has opined in a private communication that the iterators should compare equal.Proposed resolution:
Section: 29.5.9 [time.duration.literals] Status: Open Submitter: Jonathan Wakely Opened: 2014-05-16 Last modified: 2014-11-08
Priority: 3
View all issues with Open status.
Discussion:
29.5.9 [time.duration.literals] p3 says:
If any of these suffixes are applied to an integer literal and the resulting chrono::duration value cannot be represented in the result type because of overflow, the program is ill-formed.
Ill-formed requires a diagnostic at compile-time, but there is no way to detect the overflow from unsigned long long to the signed duration<>::rep type.
Overflow could be detected if the duration integer literals were literal operator templates, otherwise overflow can either be undefined or a run-time error, not ill-formed.[Urbana 2014-11-07: Move to Open]
Proposed resolution:
Section: 3.37 [defns.ntcts], 30.3.1.2.1 [locale.category], 31.2.3 [iostreams.limits.pos], 31.7.6.3.1 [ostream.formatted.reqmts], 31.7.6.3.4 [ostream.inserters.character] Status: New Submitter: Jeffrey Yasskin Opened: 2014-06-01 Last modified: 2017-04-22
Priority: 3
View all issues with New status.
Discussion:
The term "character type" is used in 3.37 [defns.ntcts], 30.3.1.2.1 [locale.category], 31.2.3 [iostreams.limits.pos], 31.7.6.3.1 [ostream.formatted.reqmts], and 31.7.6.3.4 [ostream.inserters.character], but the core language only defines "narrow character types" (6.8.2 [basic.fundamental]).
"wide-character type" is used in D.22 [depr.locale.stdcvt], but the core language only defines a "wide-character set" and "wide-character literal".Proposed resolution:
Section: 17.8.3 [type.info] Status: Open Submitter: Stephan T. Lavavej Opened: 2014-06-14 Last modified: 2016-08-06
Priority: 3
View all other issues in [type.info].
View all issues with Open status.
Discussion:
type_info's destructor is depicted as being virtual, which is nearly unobservable to users (since they can't construct or copy this class, they can't usefully derive from it). However, it's technically observable (via is_polymorphic and has_virtual_destructor). It also imposes real costs on implementations, requiring them to store one vptr per type_info object, when RTTI space consumption is a significant concern.
Making this implementation-defined wouldn't affect users (who can observe this only if they're specifically looking for it) and wouldn't affect implementations who need virtual here, but it would allow other implementations to drop virtual and improve their RTTI space consumption. Richard Smith: It's observable in a few other ways.std::map<void*, something> m; m[dynamic_cast<void*>(&typeid(blah))] = stuff;
... is broken by this change, because you can't dynamic_cast a non-polymorphic class type to void*.
type_info& f(); typeid(f());
... evaluates f() at runtime without this change, and might not do so with this change.
These are probably rare things, but I can imagine at least some forms of the latter being used in SFINAE tricks.[Lenexa 2015-05-05: Move to Open]
Marshall to poll LEWG for their opinion
[2016-06]
On the reflector, STL wrote:
We'll prototype this change and report back with data in the future.
[2016-08 Chicago]
No update from STL. Set priority to P3
Proposed resolution:
This wording is relative to N3936.
Change 17.8.3 [type.info] as indicated:
namespace std { class type_info { public:virtualsee below ~type_info(); […] }; }-1- The class type_info describes type information generated by the implementation. Objects of this class effectively store a pointer to a name for the type, and an encoded value suitable for comparing two types for equality or collating order. The names, encoding rule, and collating sequence for types are all unspecified and may differ between programs. Whether ~type_info() is virtual is implementation-defined.
Section: 19.3 [assertions] Status: New Submitter: David Krauss Opened: 2014-06-25 Last modified: 2014-11-03
Priority: 4
View other active issues in [assertions].
View all other issues in [assertions].
View all issues with New status.
Discussion:
When NDEBUG is defined, assert must expand exactly to the token sequence ((void)0), with no whitespace (C99 §7.2/1 and also C11 §7.2/1). This is a lost opportunity to pass the condition along to the optimizer.
The user may observe the token sequence using the stringize operator or discriminate it by making a matching #define directive. There is little chance of practical code doing such things. It's reasonable to allow any expansion that is a void expression with no side effects or semantic requirements, for example, an extension keyword or an attribute-specifier finagled into the context. Conforming optimizations would still be limited to treating the condition as hint, not a requirement. Nonconformance on this point is quite reasonable though, given user preferences. Anyway, it shouldn't depend on preprocessor quirks. As for current practice, Darwin OS <assert.h> provides a GCC-style compiler hint __builtin_expect but only in debug mode. Shouldn't release mode preserve hints? Daniel: The corresponding resolution should take care not to conflict with the intention behind LWG 2234.Proposed resolution:
Section: 16.4.6.9 [reentrancy] Status: Open Submitter: Stephan T. Lavavej Opened: 2014-07-01 Last modified: 2021-07-31
Priority: 3
View all other issues in [reentrancy].
View all issues with Open status.
Discussion:
N3936 16.4.6.9 [reentrancy]/1 talks about "functions", but that doesn't address the scenario of calling different member functions of a single object. Member functions often have to violate and then re-establish invariants. For example, vectors often have "holes" during insertion, and element constructors/destructors/etc. shouldn't be allowed to observe the vector while it's in this invariant-violating state. The [reentrancy] Standardese should be extended to cover member functions, so that implementers can either say that member function reentrancy is universally prohibited, or selectively allowed for very specific scenarios.
(For clarity, this issue has been split off from LWG 2382.)[2014-11-03 Urbana]
AJM confirmed with SG1 that they had no special concerns with this issue, and LWG should retain ownership.
AM: this is too overly broad as it also covers calling the exact same member function on a different objectMove to Open
[2015-07 Telecon Urbana]
Marshall to ping STL for updated wording.
[2016-05 email from STL]
I don't have any better suggestions than my original PR at the moment.
Previous resolution [SUPERSEDED]:
This wording is relative to N3936.
Change 16.4.6.9 [reentrancy] p1 as indicated:
-1- Except where explicitly specified in this standard, it is implementation-defined which functions (including different member functions called on a single object) in the Standard C++ library may be recursively reentered.
[2021-07-29 Tim suggests new wording]
The "this pointer" restriction is modeled on 11.9.5 [class.cdtor] p2. It allows us to continue to specify a member function f as calling some other member function g, since any such call would use something obtained from the first member function's this pointer.
In all other cases, this wording disallows such "recursion on object" unless both member functions are const (or are treated as such for the purposes of data race avoidance). Using "access" means that we also cover direct access to the object representation, such as the following pathological example from Arthur O'Dwyer, which is now undefined:std::string s = "hello world"; char *first = (char*)&s; char *last = (char*)(&s + 1); s.append(first, last);
Proposed resolution:
This wording is relative to N4892.
Add the following paragraph to 16.4.6.9 [reentrancy]:
-?- During the execution of a standard library non-static member function F on an object, if that object is accessed through a glvalue that is not obtained, directly or indirectly, from the this pointer of F, in a manner that can conflict (6.9.2.2 [intro.races]) with any access that F is permitted to perform (16.4.6.10 [res.on.data.races]), the behavior is undefined unless otherwise specified.
Section: 20.2.5 [ptr.align] Status: New Submitter: Melissa Mears Opened: 2014-08-06 Last modified: 2014-11-03
Priority: 3
View all other issues in [ptr.align].
View all issues with New status.
Discussion:
The specification of std::align does not appear to specify what happens when the value of the size parameter is 0. (The question of what happens when alignment is 0 is mentioned in another Defect Report, 2377; it would change the behavior to be undefined rather than potentially implementation-defined.)
The case of size being 0 is interesting because the result is ambiguous. Consider the following code's output:#include <cstdio> #include <memory> int main() { alignas(8) char buffer[8]; void *ptr = &buffer[1]; std::size_t space = sizeof(buffer) - sizeof(char[1]); void *result = std::align(8, 0, ptr, space); std::printf("%d %td\n", !!result, result ? (static_cast<char*>(result) - buffer) : std::ptrdiff_t(-1)); }
There are four straightforward answers as to what the behavior of std::align with size 0 should be:
The behavior is undefined because the size is invalid.
The behavior is implementation-defined. This seems to be the status quo, with current implementations using #3.
Act the same as size == 1, except that if size == 1 would fail but would be defined and succeed if space were exactly 1 larger, the result is a pointer to the byte past the end of the ptr buffer. That is, the "aligned" version of a 0-byte object can be one past the end of an allocation. Such pointers are, of course, valid when not dereferenced (and a "0-byte object" shouldn't be), but whether that is desired is not specified in the Standard's definition of std::align, it appears. The output of the code sample is "1 8" in this case.
Act the same as size == 1; this means that returning "one past the end" is not a possible result. In this case, the code sample's output is "0 -1".
The two compilers I could get working with std::align, Visual Studio 2013 and Clang 3.4, implement #3. (Change %td to %Id on Visual Studio 2013 and earlier. 2014 and later will have %td.)
Proposed resolution:
Section: 28.6.5 [template.slice.array], 28.6.7 [template.gslice.array], 28.6.8 [template.mask.array], 28.6.9 [template.indirect.array] Status: New Submitter: Akira Takahashi Opened: 2014-08-12 Last modified: 2014-11-03
Priority: 4
View all other issues in [template.slice.array].
View all issues with New status.
Discussion:
I found a missing specification of the copy constructor of the following class templates:
slice_array (28.6.5 [template.slice.array])
gslice_array (28.6.7 [template.gslice.array])
mask_array (28.6.8 [template.mask.array])
indirect_array (28.6.9 [template.indirect.array])
Proposed resolution:
Before 28.6.5.2 [slice.arr.assign] insert a new sub-clause as indicated:
-?- slice_array constructors [slice.arr.cons]
slice_array(const slice_array&);-?- Effects: The constructed slice refers to the same valarray<T> object to which the argument slice refers.
Before 28.6.7.2 [gslice.array.assign] insert a new sub-clause as indicated:
-?- gslice_array constructors [gslice.array.cons]
gslice_array(const gslice_array&);-?- Effects: The constructed slice refers to the same valarray<T> object to which the argument slice refers.
Before 28.6.8.2 [mask.array.assign] insert a new sub-clause as indicated:
-?- mask_array constructors [mask.array.cons]
mask_array(const mask_array&);-?- Effects: The constructed slice refers to the same valarray<T> object to which the argument slice refers.
Before 28.6.9.2 [indirect.array.assign] insert a new sub-clause as indicated:
-?- indirect_array constructors [indirect.array.cons]
indirect_array(const indirect_array&);-?- Effects: The constructed slice refers to the same valarray<T> object to which the argument slice refers.
Section: 32.2 [re.req] Status: New Submitter: Jonathan Wakely Opened: 2014-09-30 Last modified: 2020-04-16
Priority: 3
View other active issues in [re.req].
View all other issues in [re.req].
View all issues with New status.
Discussion:
The requirements on the traits class in 32.2 [re.req] do not say whether a regular expression traits class is required to be DefaultConstructible, CopyConstructible, CopyAssignable etc.
The std::regex_traits class appears to be all of the above, but can basic_regex assume that for user-defined traits classes? Should the following statements all leave u in equivalent states?X u{v}; X u; u = v; X u; u.imbue(v.getloc();
Whether they are equivalent has implications for basic_regex copy construction and assignment.
[2020-04-16, Jonathan adds that 32.7.5 [re.regex.locale] requires the traits type to be default-initialized, despite no guarantee that the traits type is default constructible. ]
Proposed resolution:
Section: 17.11 [support.initlist] Status: Tentatively NAD Submitter: David Krauss Opened: 2014-09-30 Last modified: 2015-05-22
Priority: 2
View other active issues in [support.initlist].
View all other issues in [support.initlist].
View all issues with Tentatively NAD status.
Discussion:
std::initializer_list::operator= 17.11 [support.initlist] is horribly broken and it needs deprecation:
std::initializer_list<foo> a = {{1}, {2}, {3}}; a = {{4}, {5}, {6}}; // New sequence is already destroyed.
Assignability of initializer_list isn't explicitly specified, but most implementations supply a default assignment operator. I'm not sure what 16.3 [description] says, but it probably doesn't matter.
[Lenexa 2015-05-05: Send to EWG as discussed in Telecon]
[2022-08-24; Reflector poll]
Set status to Tentatively NAD after reflector poll in October 2021.
"If somebody wants to revisit it, they'll need to write a paper to demonstrate what the change would break, whether that would be a problem in practice, and convince the evolution groups to make a change. But it's not an LWG issue."
Proposed resolution:
Edit 17.11 [support.initlist] p1, class template initializer_list synopsis, as indicated:
namespace std { template<class E> class initializer_list { public: […] constexpr initializer_list() noexcept; initializer_list(const initializer_list&) = default; initializer_list(initializer_list&&) = default; initializer_list& operator=(const initializer_list&) = delete; initializer_list& operator=(initializer_list&&) = delete; constexpr size_t size() const noexcept; […] }; […] }
Section: 21 [meta] Status: Core Submitter: Hubert Tong Opened: 2014-11-04 Last modified: 2015-10-21
Priority: 3
View other active issues in [meta].
View all other issues in [meta].
Discussion:
The BaseCharacteristic for is_constructible is defined in terms of the well-formedness of a declaration for an invented variable. The well-formedness of the described declaration itself may change for the same set of arguments because of the introduction of default arguments.
In the following program, there appears to be conflicting definitions of a specialization of std::is_constructible; however, it seems that this situation is caused without a user violation of the library requirements or the ODR. There is a similar issue with is_convertible, result_of and others. a.cc:#include <type_traits> struct A { A(int, int); }; const std::false_type& x1 = std::is_constructible<A, int>(); int main() { }
b.cc:
#include <type_traits> struct A { A(int, int); }; inline A::A(int, int = 0) { } const std::true_type& x2 = std::is_constructible<A, int>();
Presumably this program should invoke undefined behaviour, but the Library specification doesn't say that.
[2015-02 Cologne]
Core wording should say "this kind of thing is ill-formed, no diagnostic required"
Proposed resolution:
Section: 17.11 [support.initlist], 25.7 [iterator.range] Status: New Submitter: Richard Smith Opened: 2014-11-11 Last modified: 2021-06-06
Priority: 3
View other active issues in [support.initlist].
View all other issues in [support.initlist].
View all issues with New status.
Discussion:
These sections define helper functions, some of which apply to initializer_list<T>. And they're available if you include one of a long list of header files, many of which include <initializer_list>. But they are not available if you include <initializer_list>. This seems very odd.
#include <initializer_list> auto x = {1, 2, 3}; const int *p = data(x); // error, undeclared #include <vector> const int *q = data(x); // ok
Proposed resolution:
Section: 25.7 [iterator.range] Status: New Submitter: Janez Žemva Opened: 2014-11-16 Last modified: 2015-02-23
Priority: 3
View other active issues in [iterator.range].
View all other issues in [iterator.range].
View all issues with New status.
Discussion:
The following code:
#include <algorithm> #include <iterator> #include <iostream> #include <cassert> int main() { int a[2][3][4] = { { { 1, 2, 3, 4}, { 5, 6, 7, 8}, { 9, 10, 11, 12} }, { {13, 14, 15, 16}, {17, 18, 19, 20}, {21, 22, 23, 24} } }; int b[2][3][4]; assert(std::distance(std::begin(a), std::end(a)) == 2 * 3 * 4); std::copy(std::begin(a), std::end(a), std::begin(b)); std::copy(std::begin(b), std::end(b), std::ostream_iterator<int>(std::cout, ",")); }
does not compile.
A possible way to remedy this would be to add the following overloads of begin, end, rbegin, and rend to 25.7 [iterator.range], relying on recursive evaluation:namespace std { template <typename T, size_t M, size_t N> constexpr remove_all_extents_t<T>* begin(T (&array)[M][N]) { return begin(*array); } template <typename T, size_t M, size_t N> constexpr remove_all_extents_t<T>* end(T (&array)[M][N]) { return end(array[M - 1]); } template <typename T, size_t M, size_t N> reverse_iterator<remove_all_extents_t<T>*> rbegin(T (&array)[M][N]) { return decltype(rbegin(array))(end(array[M - 1])); } template <typename T, size_t M, size_t N> reverse_iterator<remove_all_extents_t<T>*> rend(T (&array)[M][N]) { return decltype(rend(array))(begin(*array)); } }
Proposed resolution:
Section: 16.4.4.6 [allocator.requirements], 24.3.11.3 [vector.capacity], 24.3.11.5 [vector.modifiers] Status: New Submitter: dyp Opened: 2014-12-06 Last modified: 2015-06-10
Priority: 3
View other active issues in [allocator.requirements].
View all other issues in [allocator.requirements].
View all issues with New status.
Discussion:
When resizing a vector, the accessibility and exception specification of the value type's constructors determines whether the elements are copied or moved to the new buffer. However, the copy/move is performed via the allocator's construct member function, which is assumed, but not required, to call the copy/move constructor and propagate only exceptions from the value type's copy/move constructor. The issue might also affect other classes.
The current wording in N4296 relevant here is from Table 28 — "Allocator requirements" in 16.4.4.6 [allocator.requirements]:
Table 28 — Allocator requirements Expression Return type Assertion/note
pre-/post-conditionDefault … a.construct(c, args) (not used) Effect: Constructs an object of type C at c ::new ((void*)c) C(forward<Args>(args)...) …
and from 16.4.4.6 [allocator.requirements] p9:
An allocator may constrain the types on which it can be instantiated and the arguments for which its construct member may be called. If a type cannot be used with a particular allocator, the allocator class or the call to construct may fail to instantiate.
I conclude the following from the wording:
The allocator is not required to call the copy constructor if the arguments (args) is a single (potentially const) lvalue of the value type. Similarly for a non-const rvalue + move constructor. See also 24.2.2.1 [container.requirements.general] p15 which seems to try to require this, but is not sufficient: That paragraph specifies the semantics of the allocator's operations, but not which constructors of the value type are used, if any.
The allocator may throw exceptions in addition to the exceptions propagated by the constructors of the value type; it can also propagate exceptions from constructors other than a copy/move constructor.
This leads to an issue with the wording of the exception safety guarantees for vector modifiers in 24.3.11.5 [vector.modifiers] p1:
[…]
void push_back(const T& x); void push_back(T&& x);Remarks: Causes reallocation if the new size is greater than the old capacity. If no reallocation happens, all the iterators and references before the insertion point remain valid. If an exception is thrown other than by the copy constructor, move constructor, assignment operator, or move assignment operator of T or by any InputIterator operation there are no effects. If an exception is thrown while inserting a single element at the end and T is CopyInsertable or is_nothrow_move_constructible<T>::value is true, there are no effects. Otherwise, if an exception is thrown by the move constructor of a non-CopyInsertable T, the effects are unspecified.
The wording leads to the following problem: Copy and move assignment are invoked directly from vector. For intermediary objects (see 2164), vector also directly invokes the copy and move constructor of the value type. However, construction of the actual element within the buffer is invoked via the allocator abstraction. As discussed above, the allocator currently is not required to call a copy/move constructor. If is_nothrow_move_constructible<T>::value is true for some value type T, but the allocator uses modifying operations for MoveInsertion that do throw, the implementation is required to ensure that "there are no effects", even if the source buffer has been modified.
Similarly, the vector capacity functions specify exception safety guarantees referring to the move constructor of the value type. For example, vector::resize in 24.3.11.3 [vector.capacity] p14:Remarks: If an exception is thrown other than by the move constructor of a non-CopyInsertable T there are no effects.
The wording leads to the same issue as described above.
Code example:template<class T> class allocator; class pot_reg_type // a type which creates // potentially registered instances { private: friend class allocator<pot_reg_type>; struct register_t {}; static std::set<pot_reg_type*>& get_registry() { static std::set<pot_reg_type*> registry; return registry; } void enregister() noexcept(false) { get_registry().insert(this); } void deregister() { get_registry().erase(this); } public: pot_reg_type(void ) noexcept(true) {} pot_reg_type(pot_reg_type const&) noexcept(true) {} pot_reg_type(pot_reg_type&& ) noexcept(true) {} private: pot_reg_type(register_t ) noexcept(false) { enregister(); } pot_reg_type(register_t, pot_reg_type const&) noexcept(false) { enregister(); } pot_reg_type(register_t, pot_reg_type&& ) noexcept(false) { enregister(); } }; template<class T> class allocator { public: using value_type = T; value_type* allocate(std::size_t p) { return (value_type*) ::operator new(p); } void deallocate(value_type* p, std::size_t) { ::operator delete(p); } void construct(pot_reg_type* pos) { new((void*)pos) pot_reg_type((pot_reg_type::register_t())); } void construct(pot_reg_type* pos, pot_reg_type const& source) { new((void*)pos) pot_reg_type(pot_reg_type::register_t(), source); } template<class... Args> void construct(T* p, Args&&... args) { new((void*)p) T(std::forward<Args>(args)...); } };
The construct member function template is only required for rebinding, which can be required e.g. to store additional debug information in the allocated memory (e.g. VS2013).
Even though the value type has an accessible and noexcept(true) move constructor, this allocator won't call that constructor for rvalue arguments. In any case, it does not call a constructor for which vector has formulated its requirements. An exception thrown by a constructor called by this allocator is not covered by the specification in 24.3.11.5 [vector.modifiers] and therefore is guaranteed not to have any effect on the vector object when resizing. For an example how this might invalidate the exception safety guarantees, see this post on the std-discussion mailing list. Another problem arises for value types whose constructors are private, but may be called by the allocator e.g. via friendship. Those value types are not MoveConstructible (is_move_constructible is false), yet they can be MoveInsertable. It is not possible for vector to create intermediary objects (see 2164) of such a type by directly using the move constructor. Current implementations of the single-element forms of vector::insert and vector::emplace do create intermediary objects by directly calling one of the value type's constructors, probably to allow inserting objects from references that alias other elements of the container. As far as I can see, Table 100 — "Sequence container requirements" in 24.2.4 [sequence.reqmts] does not require that the creation of such intermediare objects can be performed by containers using the value type's constructor directly. It is unclear to me if the allocator's construct function could be used to create those intermediary objects, given that they have not been allocated by the allocator. Two possible solutions:Add the following requirement to the allocator_traits::construct function: If the parameter pack args consists of a single parameter of the type value_type&&, the function may only propagate exceptions if is_nothrow_move_constructible<value_type>::value is false.
Requiring alloctor_traits::construct to call a true copy/move constructor of the value type breaks std::scoped_allocator_adapter, as pointed out by Casey Carter in a post on the std-discussion mailing list.Change vector's criterion whether to move or copy when resizing:
Instead of testing the value type's constructors via is_move_constructible, check the value of noexcept( allocator_traits<Allocator>::construct(alloc, ptr, rval) ) where alloc is an lvalue of type Allocator, ptr is an expression of type allocator_traits<Allocator>::pointer and rval is a non-const rvalue of type value_type.A short discussion of the two solutions:
Solution 1 allows keeping is_nothrow_move_constructible<value_type> as the criterion for vector to decide between copying and moving when resizing. It restricts what can be done inside the construct member function of allocators, and requires implementers of allocators to pay attention to the value types used. One could conceive allocators checking the following with a static_assert: If the value type is_nothrow_move_constructible, then the constructor actually called for MoveInsertion within the construct member function is also declared as noexcept. Solution 2 requires changing both the implementation of the default allocator (add a conditional noexcept) and vector (replace is_move_constructible with an allocator-targeted check). It does not impose additional restrictions on the allocator (other than 24.2.2.1 [container.requirements.general] p15), and works nicely even if the move constructor of a MoveInsertable type is private or deleted (the allocator might be a friend of the value type). In both cases, an addition might be required to provide the basic exception safety guarantee. A short discussion on this topic can be found in the std-discussion mailing list. Essentially, if allocator_traits<Allocator>::construct throws an exception, the object may or may not have been constructed. Two solutions are mentioned in that discussion:allocator_traits<Allocator>::construct needs to tell its caller whether or not the construction was successful, in case of an exception.
If allocator_traits<Allocator>::construct propagates an exception, it shall either not have constructed an object at the specified location, or that object shall have been destroyed (or it shall ensure otherwise that no resources are leaked).
[2015-05-23, Tomasz Kamiński comments]
Solution 1 discussed in this issue also breaks support for the polymorphic_allocator proposed in the part of the Library Fundamentals TS v1, in addition to already mentioned std::scoped_allocator_adapter. Furthermore there is unknown impact on the other user-defined state-full allocators code written in the C++11.
In addition the library resolution proposed in the LWG issues 2089 and N4462, will break the relation between the std::allocator_trait::construct method and copy/move constructor even for the standard std::allocator. As example please consider following class:struct NonCopyable { NonCopyable() = default; NonCopyable(NonCopyable const&) = delete; NonCopyable(NonCopyable&&) = delete; }; struct InitListConstructor : NonCopyable { InitListConstructor() = default; InitListConstructor(std::initializer_list<int>); operator int() const; };
For the above declarations following expression are ill-formed:
InitListConstructor copy(std::declval<InitListConstructor const&>()); InitListConstructor move(std::declval<InitListConstructor&&>());
So the class is not CopyConstructible nor MoveConstructible. However the following are well formed:
InitListConstructor copy{std::declval<InitListConstructor const&>()}; InitListConstructor move{std::declval<InitListConstructor&&>()};
And will be used by std::allocator<InitListConstructor>::construct in case of move-insertion and copy-insertion, after appliance of the resolution proposed in mentioned papers:
The gist of the proposed library fix is simple:
if is_constructible_v<TargetType, Args...>, use direct-nonlist-initialization
otherwise, use brace-initialization.
As consequence the requirement proposed in the Solution 1:
If the parameter pack args consists of a single parameter of the type value_type&&, the function may only propagate exceptions if is_nothrow_move_constructible<value_type>::value is false.
Will no longer hold for the std::allocator.
Proposed resolution:
Section: 27.7.1 [alg.copy] Status: Open Submitter: Jonathan Wakely Opened: 2015-01-28 Last modified: 2018-06-22
Priority: 3
View other active issues in [alg.copy].
View all other issues in [alg.copy].
View all issues with Open status.
Discussion:
It's unspecified how many times copy_n increments the InputIterator. uninitialized_copy_n is specified to increment it exactly n times, which means if an istream_iterator is used then the next character after those copied is read from the stream and then discarded, losing data.
I believe all three of Dinkumware, libc++ and libstdc++ implement copy_n with n - 1 increments of the InputIterator, which avoids reading and discarding a character when used with istream_iterator, but is inconsistent with uninitialized_copy_n and causes surprising behaviour with istreambuf_iterator instead, because copy_n(in, 2, copy_n(in, 2, out)) is not equivalent to copy_n(in, 4, out)[2016-08 Chicago]
Tues PM: refer to LEWG
[LEWG Kona 2017]
This is a mess. Append to Effects: If the InputIterator is not a forward iterator, increments n-1 times. Otherwise the number of increments is not more than n. (ncm) The preceding proposition is unsatisfactory, because it is wrong for istreambuf_iterator, which is much more useful than istream_iterator. Proposing instead: Append to Effects: If InputIterator is istream_iterator for some T, increments n-1 times. Otherwise, increments n times. Want a paper exploring what the implementations actually do, and what non-uniformity is "right".
Status to Open
Proposed resolution:
Section: 22.4.9 [tuple.rel], 20.2.10.3 [allocator.globals], 20.3.1.6 [unique.ptr.special], 20.3.2.2.8 [util.smartptr.shared.cmp], 29.5.7 [time.duration.comparisons], 29.6.7 [time.point.comparisons], 20.5.5 [scoped.adaptor.operators], 25.5.1.8 [reverse.iter.cmp], 25.5.4.8 [move.iter.op.comp] Status: New Submitter: Richard Smith Opened: 2015-02-07 Last modified: 2021-06-06
Priority: 3
View other active issues in [tuple.rel].
View all other issues in [tuple.rel].
View all issues with New status.
Discussion:
The standard library specifies a lot of heterogeneous comparison operators. For instance:
template<class... TTypes, class... UTypes> constexpr bool operator!=(const tuple<TTypes...>&, const tuple<UTypes...>&);
This has an unfortunate consequence:
#include <tuple> #include <utility> using namespace std::rel_ops; std::tuple<int> a(0); bool b = a != a;
The last line here is ill-formed due to ambiguity: it might be rel_ops::operator!=, and it might be the heterogeneous tuple operator!=. These are not partially ordered, because they have different constraints: rel_ops requires the types to match, whereas the tuple comparison requires both types to be tuples (but not to match). The same thing happens for user code that defines its own unconstrained 'template<typename T> operator!=(const T&, const T&)' rather than using rel_ops.
One straightforward fix would be to add a homogeneous overload for each heterogeneous comparison:template<class... TTypes> constexpr bool operator!=(const tuple<TTypes...>&, const tuple<TTypes...>&);This is then unambiguously chosen over the other options in the preceding case. FWIW, libstdc++ already does this in some cases.
Proposed resolution:
Section: D.23.2 [depr.conversions.string] Status: New Submitter: Jonathan Wakely Opened: 2015-03-04 Last modified: 2017-04-22
Priority: 4
View other active issues in [depr.conversions.string].
View all other issues in [depr.conversions.string].
View all issues with New status.
Discussion:
How do wstring_convert::from_bytes and wstring_convert::to_bytes use the cvtstate member?
Is it passed to the codecvt member functions? Is a copy of it passed to the member functions? "Otherwise it shall be left unchanged" implies a copy is used, but if that's really what's intended there are simpler ways to say so.Proposed resolution:
Section: D.23.3 [depr.conversions.buffer] Status: New Submitter: Jonathan Wakely Opened: 2015-03-04 Last modified: 2017-04-22
Priority: 4
View other active issues in [depr.conversions.buffer].
View all other issues in [depr.conversions.buffer].
View all issues with New status.
Discussion:
How does wbuffer_convert use the cvtstate member?
Is the same conversion state object used for converting both the get and put areas? That means a read which runs out of bytes halfway through a multibyte character will leave some shift state in cvtstate, which would then be used by a following write, even though the shift state of the get area is unrelated to the put area.Proposed resolution:
Section: D.23.3 [depr.conversions.buffer] Status: New Submitter: Jonathan Wakely Opened: 2015-03-04 Last modified: 2017-04-22
Priority: 4
View other active issues in [depr.conversions.buffer].
View all other issues in [depr.conversions.buffer].
View all issues with New status.
Discussion:
If a codecvt conversion returns codecvt_base::error should that be treated as EOF? An exception? Should all the successfully converted characters before a conversion error be available to the users of the wbuffer_convert and/or the internal streambuf, or does a conversion error lose information?
Proposed resolution:
Section: D.23.2 [depr.conversions.string] Status: New Submitter: Jonathan Wakely Opened: 2015-03-04 Last modified: 2017-04-22
Priority: 4
View other active issues in [depr.conversions.string].
View all other issues in [depr.conversions.string].
View all issues with New status.
Discussion:
Paragraph 4 of D.23.2 [depr.conversions.string] introduces byte_err_string as "a byte string to display on errors". What does display mean? The string is returned on error, it's not displayed anywhere.
Paragraph 14 says "Otherwise, if the object was constructed with a byte-error string, the member function shall return the byte-error string." The term byte-error string is not used anywhere else. Paragraph 17 talks about storing "default values in byte_err_string". What default value? Is "Hello, world!" allowed? If it means default-construction it should say so. If paragraph 14 says it won't be used what does it matter how it's initialized? The end of the paragraph refers to storing "byte_err in byte_err_string". This should be more clearly related to the wording in paragraph 14. It might help if the constructor (and destructor) was specified before the other member functions, so it can more formally define the difference between being "constructed with a byte-error string" and not. All the same issues apply to the wide_err_string member.Proposed resolution:
Section: 32 [re] Status: New Submitter: Stephan T. Lavavej Opened: 2015-03-27 Last modified: 2015-05-05
Priority: 3
View other active issues in [re].
View all other issues in [re].
View all issues with New status.
Discussion:
Only 4 functions are marked noexcept in all of Clause 28. Many more need to be marked — for example, regex_error::code(), basic_regex::swap(), and sub_match::length().
Proposed resolution:
Section: 22.10.8 [comparisons] Status: New Submitter: Agustín K-ballo Bergé Opened: 2015-04-01 Last modified: 2021-04-10
Priority: 3
View other active issues in [comparisons].
View all other issues in [comparisons].
View all issues with New status.
Discussion:
It is not entirely clear if and when the specializations of std::less (and friends) for pointer types can be used in a constant expression. Consider the following code:
#include <functional> struct foo {}; foo x, y; constexpr bool b = std::less<foo*>{}(&x, &y); // [1] foo z[] = {{}, {}}; constexpr bool ba = std::less<foo*>{}(&z[0], &z[1]); // [2]
Comparing the address of unrelated objects is not a constant expression since the result is unspecified, so it could be expected for [1] to fail and [2] to succeed. However, std::less specialization for pointer types is well-defined and yields a total order, so it could just as well be expected for [1] to succeed. Finally, since the implementation of such specializations is not mandated, [2] could fail as well (This could happen, if an implementation would provide such a specialization and if that would use built-in functions that would not be allowed in constant expressions, for example). In any case, the standard should be clear so as to avoid implementation-defined constexpr-ness.
[2017-01-22, Jens provides rationale and proposed wording]
std::less<T*> is required to deliver a total order on pointers. However, the layout of global objects is typically determined by the linker, not the compiler, so requiring std::less<T*> to provide an ordering at compile-time that is consistent with run-time would need results from linking to feed back to the compiler, something that C++ has traditionally not required.
Previous resolution [SUPERSEDED]:This wording is relative to N4618.
Add at the end of 22.10.8 [comparisons]:
-2- For templates less, greater, less_equal, and greater_equal, […], if the call operator calls a built-in operator comparing pointers, the call operator yields a strict total order that is consistent among those specializations and is also consistent with the partial order imposed by those built-in operators. Relational comparisons of pointer values are not required to be usable as constant expressions.
[2021-04-05; Jiang An comments and provides alternative wording]
The libc++ and MSVC STL implementations only support flat address spaces, and always use comparison operators. The libstdc++ implementation casts pointer values to uintptr_t if the direct comparison result is unusable in constant evaluation.
So, I think that we can specify that the implementation-defined strict total order (3.27 [defns.order.ptr]) generates a core constant expression if and only if the corresponding underlying comparison expression comparing pointer values is a core constant expression. No any other case should be a core constant expression, otherwise we should also make the underlying comparison expression a core constant expression. IMO the proposed resolution is already implemented in libc++, libstdc++, and MSVC STL, and implementable on compilers that either support flat address spaces only or have implemented intrinsics needed for transparent comparison operators and std::is_constant_evaluated.Proposed resolution:
This wording is relative to N4885.
Add at the end of 22.10.8 [comparisons] p2:
-2- For templates less, greater, less_equal, and greater_equal, the specializations for any pointer type yield a result consistent with the implementation-defined strict total order over pointers (3.27 [defns.order.ptr]). [Note 1: If a < b is well-defined for pointers a and b of type P, then (a < b) == less<P>()(a, b), (a > b) == greater<P>()(a, b), and so forth. — end note] For template specializations less<void>, greater<void>, less_equal<void>, and greater_equal<void>, if the call operator calls a built-in operator comparing pointers, the call operator yields a result consistent with the implementation-defined strict total order over pointers. A comparison result of pointer values is a core constant expression if and only if the corresponding built-in comparison expression is a core constant expression.
Add at the end of 22.10.9 [range.cmp] (3.1):
-3- Effects:
(3.1) — If the expression std::forward<T>(t) == std::forward<U>(u) results in a call to a built-in operator == comparing pointers: returns false if either (the converted value of) t precedes u or u precedes t in the implementation-defined strict total order over pointers (3.27 [defns.order.ptr]) and otherwise true. The result is a core constant expression if and only if std::forward<T>(t) == std::forward<U>(u) is a core constant expression.
(3.2) — Otherwise, equivalent to: return std::forward<T>(t) == std::forward<U>(u);
Add at the end of 22.10.9 [range.cmp] (7.1):
-7- Effects:
(7.1) — If the expression std::forward<T>(t) < std::forward<U>(u) results in a call to a built-in operator < comparing pointers: returns true if (the converted value of) t precedes u in the implementation-defined strict total order over pointers (3.27 [defns.order.ptr]) and otherwise false. The result is a core constant expression if and only if std::forward<T>(t) < std::forward<U>(u) is a core constant expression.
(7.2) — Otherwise, equivalent to: return std::forward<T>(t) < std::forward<U>(u);
Add at the end of 22.10.8.8 [comparisons.three.way] (3.1):
-3- Effects:
(3.1) — If the expression std::forward<T>(t) <=> std::forward<U>(u) results in a call to a built-in operator <=> comparing pointers: returns strong_ordering::less if (the converted value of) t precedes u in the implementation-defined strict total order over pointers (3.27 [defns.order.ptr]), strong_ordering::greater if u precedes t, and otherwise strong_ordering::equal. The result is a core constant expression if and only if std::forward<T>(t) <=> std::forward<U>(u) is a core constant expression.
(3.2) — Otherwise, equivalent to: return std::forward<T>(t) <=> std::forward<U>(u);
Section: 17.11 [support.initlist] Status: New Submitter: David Krauss Opened: 2015-04-27 Last modified: 2015-05-05
Priority: 4
View other active issues in [support.initlist].
View all other issues in [support.initlist].
View all issues with New status.
Discussion:
The typical use-case of std::initializer_list<T> is for a pass-by-value parameter of T's constructor. However, this contravenes 16.4.5.8 [res.on.functions]/2.5 because initializer_list doesn't specifically allow incomplete types (as do for example std::unique_ptr (20.3.1 [unique.ptr]/5) and std::enable_shared_from_this (20.3.2.5 [util.smartptr.enab]/2)).
A resolution would be to copy-paste the relevant text from such a paragraph.Proposed resolution:
Section: 21.3.5.4 [meta.unary.prop] Status: New Submitter: Hubert Tong Opened: 2015-05-07 Last modified: 2015-08-03
Priority: 3
View other active issues in [meta.unary.prop].
View all other issues in [meta.unary.prop].
View all issues with New status.
Discussion:
I do not believe that the wording in 21.3.5.4 [meta.unary.prop] paragraph 3 allows for the following program to be ill-formed:
#include <type_traits> template <typename T> struct B : T { }; template <typename T> struct A { A& operator=(const B<T>&); }; std::is_assignable<A<int>, int> q;
In particular, I do not see where the wording allows for the "compilation of the expression" declval<T>() = declval<U>() to occur as a consequence of instantiating std::is_assignable<T, U> (where T and U are, respectively, A<int> and int in the example code).
Instantiating A<int> as a result of requiring it to be a complete type does not trigger the instantiation of B<int>; however, the "compilation of the expression" in question does.Proposed resolution:
Section: 31.7.6.2.4 [ostream.sentry] Status: New Submitter: Roger Orr Opened: 2015-05-08 Last modified: 2020-09-06
Priority: 3
View all other issues in [ostream.sentry].
View all issues with New status.
Discussion:
In the current 31.7.6.2.4 [ostream.sentry], p4 refers to the now deprecated std::uncaught_exception(): D.9 [depr.uncaught].
If ((os.flags() & ios_base::unitbuf) && !uncaught_exception() && os.good()) is true, calls os.rdbuf()->pubsync().
This needs to be changed, for example to use std::uncaught_exceptions() and to capture the value on entry and compare with the saved value on exit.
[2015-06, Telecon]
JW: I already added an 's' here to make it use the new function, but that doesn't resolve Roger's suggestion to capture the value on entry and check it.
[2019-03-21; Daniel comments and provides wording]
The wording below implements Roger's suggestion.
Proposed resolution:
This wording is relative to N4810.
Modify 31.7.6.2.4 [ostream.sentry], class basic_ostream::sentry synopsis, as indicated:
namespace std { template<class charT, class traits = char_traits<charT>> class basic_ostream<charT, traits>::sentry { bool ok_; // exposition only int uncaught_ = uncaught_exceptions(); // exposition only public: explicit sentry(basic_ostream<charT, traits>& os); ~sentry(); explicit operator bool() const { return ok_; } sentry(const sentry&) = delete; sentry& operator=(const sentry&) = delete; }; }[…]
~sentry();-4- If (os.flags() & ios_base::unitbuf) &&
!uncaught_exceptions() <= uncaught_ && os.good() is true, calls os.rdbuf()->pubsync(). If that function returns -1, sets badbit in os.rdstate() without propagating an exception.
Section: 31.6.3 [streambuf] Status: New Submitter: Jonathan Wakely Opened: 2015-05-28 Last modified: 2015-08-03
Priority: 3
View all other issues in [streambuf].
View all issues with New status.
Discussion:
31.6.3 [streambuf] p1 says:
The class template basic_streambuf<charT, traits> serves as an abstract base class for deriving various stream buffers whose objects each control two character sequences: […]
The term "abstract base class" is not defined in the standard, but "abstract class" is (11.7.4 [class.abstract]).
According to the synopsis basic_streambuf has no pure virtual functions so is not an abstract class and none of libstdc++, libc++, or dinkumware implement it as an abstract class. I don't believe the wording was ever intended to require it to be an abstract class, but it could be read that way. I suggest the wording be changed to "polymorphic base class" or something else that can't be seen to imply a normative requirement to make it an abstract class.Proposed resolution:
Section: 6.9.2 [intro.multithread], 33.5.8 [atomics.types.generic], 17.14 [support.runtime] Status: SG1 Submitter: Geoffrey Romer Opened: 2015-05-29 Last modified: 2018-03-15
Priority: 3
View all other issues in [intro.multithread].
View all issues with SG1 status.
Discussion:
The concurrency libraries specified in clauses 29 and 30 do not adequately specify how they relate to the concurrency model specified in 6.9.2 [intro.multithread]. In particular:
6.9.2 [intro.multithread] specifies "atomic objects" as having certain properties. I can only assume that instances of the classes defined in Clause 29 are intended to be "atomic objects" in this sense, but I can't find any wording to specify that, and it's genuinely unclear whether Clause 30 objects are atomic objects. In fact, on a literal reading the C++ Standard doesn't appear to provide any portable way to create an atomic object, or even determine whether an object is an atomic object. (It's not clear if the term "atomic object" is actually needed, given that atomic objects can have non-atomic operations, and non-atomic objects can have atomic operations. But even if the term itself goes away, there still needs to be some indication that Clause 29 objects have the properties currently attributed to atomic objects). Similarly, 6.9.2 [intro.multithread] uses "atomic operation" as a term of art, but the standard never unambiguously identifies any operation as an "atomic operation" (although in one case it unambiguously identifies an operation that is not atomic). It does come close in a few cases, but not close enough:6.9.2 [intro.multithread]/p7 could be read to imply that "synchronization operations" in Clauses 29 and 30 are also atomic operations. However, that's vague and indirect, and somewhat belied by 33.6.4.2 [thread.mutex.requirements.mutex]/p5, which specifies that mutex lock and unlock operations "behave as atomic operations", but only "for purposes of determining the existence of a data race". Furthermore, not a single operation in Clause 29 explicitly identifies itself as a "synchronization operation".
33.5.8 [atomics.types.generic]/p4 states in part that "There shall be a specialization atomic<bool> which provides the general atomic operations as specified in 29.6.1", but read in context, "general atomic operations" appears to be a loose synonym for "general operations on atomic types" as defined in [atomics.types.operations.general], rather than a use of "atomic object" as Words of Power. Incidentally, "atomic type" is never satisfactorily defined either (although the <atomic> synopsis comes close).
17.14 [support.runtime]/p10 specifies exactly which operations are "plain lock-free atomic operations", but in a standard where an "integral constant expression" isn't necessarily a "constant expression", I do not feel safe assuming that a "plain lock-free atomic operation" is an "atomic operation".
Hans Boehm tells me the operations with "atomically" in the Effects element are intended to be atomic operations, but since "atomic operation" is a term of art (e.g. in 6.9.2 [intro.multithread]/p27.4), I think this needs to be spelled out rather than assumed. Furthermore, this does not help with 33.5.11 [atomics.fences], or anything in Clause 30.
[2018-03 JAX; Geoffrey comments in behalf of SG1]
SG1 consensus is that operations outside clause 32 are not "atomic operations", and objects of types defined outside clause 32 are not "atomic objects". "Synchronization operations" are operations which act as endpoints of primitive edges of partial orders other than sequenced-before, but it may make more sense to just drop that term and inline the definition, so to speak.
We would welcome a paper to make those definitions more explicit, and revise the wording as needed to be consistent with those definitions.Proposed resolution:
Section: D.22 [depr.locale.stdcvt] Status: New Submitter: Jonathan Wakely Opened: 2015-06-08 Last modified: 2017-04-22
Priority: 3
View all other issues in [depr.locale.stdcvt].
View all issues with New status.
Discussion:
The enumeration type codecvt_mode is effectively a bitmask type (16.3.3.3.4 [bitmask.types]) with three elements, but isn't defined as such.
This harms usability because bitmask types are required to work well with bitwise operators, but codecvt_mode doesn't have overloaded operators, making it very inconvenient to combine values:std::codecvt_utf16<char32_t, 0x10FFFF, static_cast<std::codecvt_mode>(std::little_endian|std::generate_header)> cvt;
The static_cast harms readability and should not be necessary.
I suggest that either codecvt_mode is specified to be a bitmask type, or as a minimal fix we provide an overloaded operator| that returns the right type.Proposed resolution:
Section: 17.7.3.5 [new.delete.dataraces] Status: New Submitter: Hans Boehm Opened: 2015-06-09 Last modified: 2016-02-01
Priority: 3
View all other issues in [new.delete.dataraces].
View all issues with New status.
Discussion:
17.7.3.5 [new.delete.dataraces] uses obsolete wording.
It should introduce a "synchronizes with" relationship. "Happens before" is too weak, since that may not composes with sequenced before. The "shall not introduce a data race" wording is probably not technically correct either. These may race with other (non-allocation/deallocation) concurrent accesses to the object being allocated or deallocated.Proposed resolution:
Section: 30.4.6.2.3 [locale.time.get.virtuals] Status: Open Submitter: Hubert Tong Opened: 2015-06-19 Last modified: 2017-09-07
Priority: 4
View other active issues in [locale.time.get.virtuals].
View all other issues in [locale.time.get.virtuals].
View all issues with Open status.
Discussion:
I recently encountered a failure related to questionable use of do_get_year. The platform where the code happened to work had an implementation which handled certain three-digit "year identifiers" as the number of years since 1900 (this article describes such an implementation).
30.4.6.2.3 [locale.time.get.virtuals] makes it implementation defined whether two-digit years are accepted, etc., but does not say anything specifically about three-digit years. The implementation freedom to not report errors in 30.4.6.2 [locale.time.get] paragraph 1 also seems to be too broad. See also the discussion following c++std-lib-38042.[2016-08 Chicago]
Wed PM: This has been this way since C++98. Don't think it's a P2.
Change to P4, and move to Open.
Proposed resolution:
Section: 23.1 [strings.general] Status: New Submitter: Jonathan Wakely Opened: 2015-06-26 Last modified: 2020-09-06
Priority: 4
View all other issues in [strings.general].
View all issues with New status.
Discussion:
The allocator-aware container requirements in Table 98 impose no MoveAssignable requirements on the value_type when propagate_on_container_move_assignment is true, because typically the container's storage would be moved by just exchanging some pointers.
However for a basic_string using the small string optimization move assignment may need to assign individual characters into the small string buffer, even when the allocator propagates. The only requirement on the char-like objects stored in a basic_string are that they are non-array POD types and Destructible, which means that a POD type with a deleted move assignment operator should be usable in a basic_string, despite it being impossible to move assign:#include <string> struct odd_pod { odd_pod() = default; odd_pod& operator=(odd_pod&&) = delete; }; static_assert(std::is_pod<odd_pod>::value, "POD"); int main() { using S = std::basic_string<odd_pod>; S s; s = S{}; // fails }
Using libstdc++ basic_string<odd_pod> cannot even be default-constructed because the constructor attempts to assign the null terminator to the first element of the small string buffer.
Similar problems exist with POD types with a deleted default constructor. I believe that basic_string should require its value_type to be at least DefaultConstructible and MoveAssignable.[2016-06, Oulu]
This should be resolved by P0178
Note: P0178 was sent back to LEWG in Oulu.
Proposed resolution:
Section: 28.5.9.4.2 [rand.dist.pois.exp] Status: Open Submitter: Michael Prähofer Opened: 2015-08-20 Last modified: 2018-08-22
Priority: 2
View all issues with Open status.
Discussion:
Original title was: exponential_distribution<float> sometimes returns inf.
The random number distribution class template exponential_distribution<float> may return "inf" as can be seen from the following example program:
// compiled with // g++ -std=c++11 Error_exp_distr.cpp #include <iostream> #include <random> #include <bitset> int main(){ unsigned long long h; std::mt19937_64 mt1(1); std::mt19937_64 mt2(1); mt1.discard(517517); mt2.discard(517517); std::exponential_distribution<float> dis(1.0); h = mt2(); std::cout << std::bitset<64>(h) << " " << (float) -log(1 - h/pow(2, 64)) << " " << -log(1 - (float) h/pow(2, 64)) << " " << dis(mt1) << std::endl; h = mt2(); std::cout << std::bitset<64>(h) << " " << (float) -log(1 - h/pow(2, 64)) << " " << -log(1 - (float) h/pow(2, 64)) << " " << dis(mt1) << std::endl; }
output:
0110010110001001010011000111000101001100111110100001110011100001 0.505218 0.505218 0.505218 1111111111111111111111111101010011000110011110011000110101100110 18.4143 inf inf
The reason seems to be that converting a double x in the range [0, 1) to float may result in 1.0f if x is close enough to 1. I see two possibilities to fix that:
use internally double (or long double?) and then convert the result at the very end to float.
take only 24 random bits and convert them to a float x in the range [0, 1) and then return -log(1 - x).
I have not checked if std::exponential_distribution<double> has the same problem: For float on the average 1 out of 224 (~107) draws returns "inf", which is easily confirmed. For double on the average 1 out of 253 (~1016) draws might return "inf", which I have not tested.
Marshall:
I don't think the problem is in std::exponential_distribution; but rather in generate_canonical.
Consider:
which outputs:std::mt19937_64 mt2(1); mt2.discard(517517); std::cout << std::hexfloat << std::generate_canonical<float, std::numeric_limits<float>::digits>(mt2) << std::endl; std::cout << std::hexfloat << std::generate_canonical<float, std::numeric_limits<float>::digits>(mt2) << std::endl; std::cout << std::hexfloat << std::generate_canonical<float, std::numeric_limits<float>::digits>(mt2) << std::endl;
but generate_canonical is defined to return a result in the range [0, 1).0x1.962532p-2 0x1p+0 0x1.20d0cap-3
[2015-10, Kona Saturday afternoon]
Options:
WEB: The one thing we cannot tolerate is any output range other than [0, 1).
WEB: I believe there may be a documented algorithm for the generator, and perhaps it's possible to discover en-route that the algorithm produces the wrong result and fix it.
MC: No. I analyzed this once, and here it is: the algorithm is in [rand.util.canonical], and it's all fine until p5. The expression S/R^k is mathematically less than one, but it may round to one.
GR: Could we change the rounding mode for the computation?
HH: No, because the rounding mode is global, not thread-local.
AM: SG1 wants to get rid of the floating point environment.
STL: The problem is that the standard specifies the implementation, and the implementation doesn't work.
MC: I'm not sure if nudging it down will introduce a subtle bias.
EF: I worry about how the user's choice of floating point environment affects the behaviour.
MS offers to run the topic past colleagues.
MC: Will set the status to open. STL wants to rename the issue. WEB wants to be able to find the issue by its original name still.
Mike Spertus to run the options past his mathematical colleagues, and report back.
[2017-11 Albuquerque Wednesday issue processing]
Choice: Rerun the algorithm if it gets 1.0.
Thomas K to provide wording; Marshall and STL to review.
[2018-08 Batavia Monday issue discussion]
Davis has a paper P0952 which resolves this.
Proposed resolution:
Section: 22.4.4.1 [tuple.cnstr] Status: New Submitter: Brian Rodriguez Opened: 2015-08-25 Last modified: 2017-02-19
Priority: 3
View other active issues in [tuple.cnstr].
View all other issues in [tuple.cnstr].
View all issues with New status.
Discussion:
The std::tuple order of element construction is unspecified. It is either in the same order of the type list or in reverse.
Consider the following program:#include <iostream> #include <tuple> struct X { X(int) { std::cout << "X constructor\n"; } }; struct Y { Y(int) { std::cout << "Y constructor\n"; } }; int main() { std::tuple<X, Y> t(1, 2); }
Here is a link to two sample compilations. The first uses libstdc++ and constructs in reverse order, and the second uses libc++ and constructs in in-order.
A std::tuple mimics both a struct and type-generic container and should thus follow their standards. Construction is fundamentally different from a function call, and it has been historically important for a specific order to be guaranteed; namely: whichever the developer may decide. Mandating construction order will allow developers to reference younger elements later on in the chain as well, much like a struct allows you to do with its members. There are implementation issues as well. Reversed lists will require unnecessary overhead for braced-initializer-list initialization. Since lists are evaluated from left to right, the initializers must be placed onto the stack to respect the construction order. This issue could be significant for large tuples, deeply nested tuples, or tuples with elements that require many constructor arguments. I propose that the std::tuple<A, B, ..., Y, Z>'s constructor implementation be standardized, and made to construct in the same order as its type list e.g. A{}, B{}, ..., Y{}, Z{}.Daniel:
When N3140 became accepted, wording had been added that gives at least an indication of requiring element initialization in the order of the declaration of the template parameters. This argumentation can be based on 22.4.4.1 [tuple.cnstr] p3 (emphasize mine):-3- In the constructor descriptions that follow, let i be in the range [0,sizeof...(Types)) in order, Ti be the ith type in Types, and Ui be the ith type in a template parameter pack named UTypes, where indexing is zero-based.
But the current wording needs to be improved to make that intention clearer and an issue like this one is necessary to be sure that the committee is agreeing (or disagreeing) with that intention, especially because N3140 didn't really point out the relevance of the element construction order in the discussion, and because not all constructors explicitly refer to the ordered sequence of numbers generated by the variable i (The move constructor does it right, but most other don't do that).
[2017-02-12, Alisdair comments]
Note that this issue should not be extended to cover the assignment operators, as implementations may want the freedom to re-order member-wise assignment so that, for example, all potentially-throwing assignments are performed before non-throwing assignments (as indicated by the noexcept operator).
Proposed resolution:
Section: 33.10.5 [futures.state] Status: Open Submitter: Agustín K-ballo Bergé Opened: 2015-09-03 Last modified: 2016-08-06
Priority: 3
View all other issues in [futures.state].
View all issues with Open status.
Discussion:
When a shared-state is released, it may be necessary to execute user defined code for the destructor of a stored value or exception. It is unclear whether the execution of said destructor constitutes an observable side effect.
While discussing N4445 in Lenexa, Nat Goodspeed pointed out that 33.10.5 [futures.state]/5.1 does not explicitly mention the destruction of the result, so implementations should be allowed to release (or reuse) a shared state ahead of time under the "as-if" rule.
The standard should clarify whether the execution of destructors is a visible side effect of releasing a shared state.[2016-08-03 Chicago]
This is related to 2532
Fri AM: Moved to Open
Proposed resolution:
Section: 33.10.6 [futures.promise] Status: Open Submitter: Agustín K-ballo Bergé Opened: 2015-09-03 Last modified: 2016-08-06
Priority: 3
View other active issues in [futures.promise].
View all other issues in [futures.promise].
View all issues with Open status.
Discussion:
promise::set_value_at_thread_exit and promise::set_exception_at_thread_exit operate on a shared state at thread exit, without making the thread participate in the ownership of such shared state.
Consider the following snippet:std::promise<int>{}.set_value_at_thread_exit(42);
Arguably, since the promise abandons its shared state without actually making it ready, a broken_promise error condition should be stored in the shared state. Implementations diverge, they either crash at thread exit by dereferencing an invalid pointer, or keep the shared state around until thread exit.
[2016-08-03 Chicago]
This is related to 2530
[2016-08-03, Billy O'Neal suggests concrete wording]
Fri AM: Moved to Open
Proposed resolution:
This wording is relative to N4606.
Change 33.10.5 [futures.state] p7 as indicated:
-7- When an asynchronous provider is said to abandon its shared state, it means:
(7.1) — first, if that state is not ready or scheduled to be made ready at thread exit, the provider
(7.1.1) — stores an exception object of type future_error with an error condition of broken_promise within its shared state; and then
(7.1.2) — makes its shared state ready;
Change 33.10.5 [futures.state] p10 as indicated:
-10- Some functions (e.g., promise::set_value_at_thread_exit)
delay making the shared state ready untilschedule the shared state to be made ready when the calling thread exits. This associates a reference to the shared state with the calling thread. The destruction of each of that thread's objects with thread storage duration (6.7.5.3 [basic.stc.thread]) is sequenced before making that shared state ready. When the calling thread makes the shared state ready, if the thread holds the last reference to the shared state, the shared state is destroyed. [Note: This means that the shared state may not become ready until after the asynchronous provider has been destroyed. — end note]
Section: 99 [concurr.ts::futures.unique.future] Status: SG1 Submitter: Agustín K-ballo Bergé Opened: 2015-09-03 Last modified: 2021-06-06
Priority: Not Prioritized
View all issues with SG1 status.
Discussion:
Addresses: concurr.ts
In N4538, the continuation given to future::then can be run "on an unspecified thread of execution". This is too broad, as it allows the continuation to be run on the main thread, a UI thread, or any other thread. In comparison, functions given to async run "as if in a new thread of execution", while the Parallelism TS gives less guarantees by running "in either the invoking thread or in a thread implicitly created by the library to support parallel algorithm execution". The threads on which the continuation given to future::then can run should be similarly constrained.
[2017-03-01, Kona, SG1]
Agreement that this is a problem. Suggested addition to the issue is below. We have no immediate delivery vehicle for a fix at the moment, but we would like to make the intended direction clear.
There is SG1 consensus that .then continuations should, by default, and in the absence of executors, be run only in the following ways:If the future is not ready when .then() is called, the .then argument may be run on the execution agent that fulfills the promise.
In all cases, the .then argument may be run on an implementation-provided thread, i.e. a thread that is neither the main thread nor explicitly created by the user.
In the absence of an executor argument (which currently cannot be supplied), running of the .then() continuation will not block the thread calling .then(), even if the future is ready at the time.
Straw polls: SF | F | N | A | SA For the default behaviour: "1. Run on completed task or new execution agent" 0 | 7 | 5 | 1 | 0 "2. Run on completed task or .then caller" 0 | 0 | 5 | 5 | 3 "3. Leave as implementation defined" 1 | 2 | 4 | 3 | 3 "4. Always new execution agent" 2 | 3 | 6 | 2 | 0 The actual conclusion was to allow either (1) or (4) for now, since they are quite close, but present a very different programming mode from (2).Proposed resolution:
Section: 32.12 [re.grammar] Status: New Submitter: Hubert Tong Opened: 2015-10-08 Last modified: 2015-10-21
Priority: 4
View other active issues in [re.grammar].
View all other issues in [re.grammar].
View all issues with New status.
Discussion:
In 32.12 [re.grammar] paragraph 2:
basic_regex member functions shall not call any locale dependent C or C++ API, including the formatted string input functions. Instead they shall call the appropriate traits member function to achieve the required effect.
Yet, the required interface for a regular expression traits class (32.2 [re.req]) does not appear to have any reliable method for determining whether a character as encoded for the locale associated with the traits instance is the same as a character represented by a UnicodeEscapeSequence, e.g., assuming a sane ru_RU.koi8r locale:
#include <stdio.h> #include <stdlib.h> #include <regex> const char data[] = "\xB3"; const char matchCyrillicCaptialLetterYo[] = R"(\u0401)"; int main(void) { try { std::regex myRegex; myRegex.imbue(std::locale("ru_RU.koi8r")); myRegex.assign(matchCyrillicCaptialLetterYo, std::regex_constants::ECMAScript); printf("(%s)\n", std::regex_replace(std::string(data), myRegex, std::string("E")).c_str()); myRegex.assign("[[:alpha:]]", std::regex_constants::ECMAScript); printf("(%s)\n", std::regex_replace(std::string(data), myRegex, std::string("E")).c_str()); } catch (std::regex_error& e) { abort(); } return 0; }
The implementation I tried prints:
(Ё) (E)
Which means that the character class matching worked, but not the matching to the UnicodeEscapeSequence.
Proposed resolution:
Section: 22.10.8 [comparisons], 24.2.2.1 [container.requirements.general], 33.4.3.2 [thread.thread.id] Status: New Submitter: Matt Austern Opened: 2015-10-08 Last modified: 2015-10-21
Priority: 3
View other active issues in [comparisons].
View all other issues in [comparisons].
View all issues with New status.
Discussion:
A number of places in the library, including 22.10.8 [comparisons]/14, the Optional container requirements in 24.2.2.1 [container.requirements.general], and 33.4.3.2 [thread.thread.id]/8, use the phrase "total order". Unfortunately, that phrase is ambiguous. In mathematics, the most common definition is that a relation ≤ is a total order if it's total, transitive, and antisymmetric in the sense that x≤y ∧ y≤x ⇒ x=y. What we really want is a strict total order: a relation < is a strict total order if it's total, transitive, and antisymmetric in the sense that exactly one of x<y, y<x, and x=y holds.
The non-normative note in 27.8 [alg.sorting]/4 correctly uses the phrase "strict total ordering" rather than simply "total ordering".
We could address this issue by replacing "total order" with "strict total order" everywhere it appears, since I think there are no cases where we actually want a non-strict total order, or we could add something in Clause 17 saying that we always mean strict total order whenever we say total order.
Proposed resolution:
Section: 29.2 [time.syn] Status: New Submitter: Andy Giese Opened: 2016-02-05 Last modified: 2016-05-08
Priority: 4
View all other issues in [time.syn].
View all issues with New status.
Discussion:
Currently 29.2 [time.syn] states
// convenience typedefs typedef duration<signed integer type of at least 64 bits, nano> nanoseconds; typedef duration<signed integer type of at least 55 bits, micro> microseconds; typedef duration<signed integer type of at least 45 bits, milli> milliseconds; typedef duration<signed integer type of at least 35 bits > seconds; typedef duration<signed integer type of at least 29 bits, ratio< 60>> minutes; typedef duration<signed integer type of at least 23 bits, ratio<3600>> hours;
However, a duration_cast<minutes>(seconds::max()) would cause overflow if the underlying signed integers only met the minimums specified.
The standard should specify that implementations guarantee that a duration_cast from any smaller duration in these "convenience typedefs" will not overflow any larger duration. That is, hours should be able to hold the maximum of minutes, which should be able to hold the maximum of seconds and so on. More formally, if the ratio typedef A and typedef B is 1:Y where Y > 1 (e.g., 1 : 60 in case of minutes : seconds), then #bitsA-1 must be at least ceil(log2(2#bitsB-1)/Y)). In the case of minutes : seconds, X = 1, Y = 60. Let #bitsseconds = 32. Therefore:2(#bitsseconds - 1) = 231 = 2147483648
ceil(log2(231 / 60) = 26
#bitsminutes - 1 = 26
#bitsminutes = 27
Therefore, a minimum of 27 bits would be needed to store minutes if 32 were used to store seconds.
I propose to change the definitions of the convenience typedefs as follows:// convenience typedefs typedef duration<signed integer type of at least 64 bits, nano> nanoseconds; typedef duration<signed integer type of at least 55 bits, micro> microseconds; typedef duration<signed integer type of at least 46 bits, milli> milliseconds; typedef duration<signed integer type of at least 37 bits > seconds; typedef duration<signed integer type of at least 32 bits, ratio< 60>> minutes; typedef duration<signed integer type of at least 27 bits, ratio<3600>> hours;
These bits were chosen to satisfy the above formula. Note that minimums only increased, so larger ranges could be held. A nice outcome of this choice is that minutes does not go above 32 bits.
[2016-04-23, Tim Song comments]
The P/R of LWG 2592 doesn't fix the issue it wants to solve, because the actual underlying type will likely have more bits than the specified minimum.
Consider seconds, which the P/R requires to have at least 37 bits. On a typical system this implies using a 64-bit integer. To ensure that casting from seconds::max() to minutes doesn't overflow in such a system, it is necessary for the latter to have at least 59 bits (which means, in practice, 64 bits too), not just 32 bits. Thus, just changing the minimum number of bits will not be able to provide the desired guarantee that casting from a smaller unit to a larger one never overflow. If such a guarantee is to be provided, it needs to be spelled out directly. Note that the difference here is 9 bits (for the 1000-fold case) and 5 bits (for the 60-fold case), which is less than the size difference between integer types on common systems, so such a requirement would effectively require those convenience typedefs to use the same underlying integer type.Proposed resolution:
This wording is relative to N4567.
Change 29.2 [time.syn], header <chrono> synopsis, as indicated
[…] // convenience typedefs typedef duration<signed integer type of at least 64 bits, nano> nanoseconds; typedef duration<signed integer type of at least 55 bits, micro> microseconds; typedef duration<signed integer type of at least 4645bits, milli> milliseconds; typedef duration<signed integer type of at least 3735bits > seconds; typedef duration<signed integer type of at least 3229bits, ratio< 60>> minutes; typedef duration<signed integer type of at least 2723bits, ratio<3600>> hours; […]
Section: 20.3.2.2 [util.smartptr.shared] Status: New Submitter: Kazutoshi Satoda Opened: 2016-02-20 Last modified: 2016-06-20
Priority: 3
View all other issues in [util.smartptr.shared].
View all issues with New status.
Discussion:
Latest draft (N4567) 20.3.2.2 [util.smartptr.shared] p1 says:
A shared_ptr object is empty if it does not own a pointer.
Please note that it says "own a pointer". This definition was added as the resolution for LWG defect 813.
20.3.2.2.2 [util.smartptr.shared.const] p8 says about the effect of shared_ptr(nullptr_t p, D d):Effects: Constructs a shared_ptr object that owns the object p and the deleter d.
Please note that it says "owns the object". This was intentionally changed from "the pointer" as a part of resolution for LWG defect 758, to cover nullptr_t case.
Since shared_ptr(nullptr, d) owns an object of type nullptr_t, but does not own a pointer, it is said as "empty" by a strict reading of the above mentioned definition in 20.3.2.2 [util.smartptr.shared] p1. These cause a contradiction: 20.3.2.2.2 [util.smartptr.shared.const] p9 sets a postcondition use_count() == 1 on shared_ptr(nullptr, d). But 20.3.2.2.6 [util.smartptr.shared.obs] p7 says that the return value of use_count() is "0 when *this is empty".Proposed wording changes:
Replace the last 2 words in 20.3.2.2 [util.smartptr.shared] p1 from[…] empty if it does not own a pointer.
to
[…] empty if it does not own an object.
Note that shared_ptr(nullptr_t) is defined to be empty in synopsis in 20.3.2.2 [util.smartptr.shared].
constexpr shared_ptr(nullptr_t) noexcept : shared_ptr() { }
It could be less confusing if shared_ptr(nullptr, d) could be defined to be empty. But it seems too late to change that (which means changing whether the deleter is called or not, see this Stackoverflow article). Then I'm proposing just fix the contradiction.
Proposed resolution:
This wording is relative to N4594.
Change 20.3.2.2 [util.smartptr.shared] p1 as indicated:
-1- The shared_ptr class template stores a pointer, usually obtained via new. shared_ptr implements semantics of shared ownership; the last remaining owner of the pointer is responsible for destroying the object, or otherwise releasing the resources associated with the stored pointer. A shared_ptr object is empty if it does not own an object
a pointer.
Section: 25.5.1.2 [reverse.iterator], 25.5.1.6 [reverse.iter.elem] Status: New Submitter: Robert Haberlach Opened: 2016-02-28 Last modified: 2021-06-06
Priority: 3
View all other issues in [reverse.iterator].
View all issues with New status.
Discussion:
Issue 386 changed the return type of reverse_iterator::operator[] to unspecified. However, as of N3066, the return type of a random access iterator's operator[] shall be convertible to reference; thus the return type of reverse_iterator::operator[] should be reference (and it is in all common implementations).
Suggested resolution: Adjust 25.5.1.2 [reverse.iterator]'s synopsis and 25.5.1.6 [reverse.iter.elem] to use reference instead of unspecified.[2021-06-06 Tim syncs wording to current working draft]
Proposed resolution:
This wording is relative to N4885.
Edit 25.5.1.2 [reverse.iterator], class template synopsis, as indicated:
namespace std { template <class Iterator> class reverse_iterator { public: […] using reference = iter_reference_t<Iterator>; […] constexpr referenceunspecifiedoperator[](difference_type n) const; […] }; }
Change 25.5.1.6 [reverse.iter.elem] before p3 as indicated:
constexpr referenceunspecifiedoperator[](difference_type n) const;
Section: 22.2.6 [declval], 20.3.1 [unique.ptr], 20.3.1.2.1 [unique.ptr.dltr.general], 20.3.2.2 [util.smartptr.shared], 20.3.2.3 [util.smartptr.weak], 20.3.2.5 [util.smartptr.enab] Status: New Submitter: Zhihao Yuan Opened: 2016-03-08 Last modified: 2016-04-16
Priority: 3
View all issues with New status.
Discussion:
Currently the phrase to grant this permission is:
The template parameter T of LibraryTemplate may be an incomplete type.
Two problems:
The timing is unclear. We always allow specializations like LibraryTemplate<Incomp>* p;
To the users of a template, the correct terminology should be "argument" rather than "parameter".
Suggested resolution:
In an instantiation of LibraryTemplate, an incomplete type may be used as the template argument for the template parameter T.
as shown here.
Or, to copy N4510's wording:An incomplete type T may be used when instantiating LibraryTemplate.
Proposed resolution:
Section: 31.5.2.7 [ios.base.callback] Status: New Submitter: David Krauss Opened: 2016-03-14 Last modified: 2019-06-15
Priority: 3
View all issues with New status.
Discussion:
register_callback allocates memory and so it can fail, but the case is unspecified. libc++ sets badbit, which is consistent with iword and pword. libstdc++ throws std::bad_alloc.
[2019-06-13; Billy comments]
Just as an additional data point: MSVC++ agrees with libstdc++ and also throws std::bad_alloc.
Proposed resolution:
Section: 30.4.7.4 [locale.moneypunct] Status: New Submitter: Hubert Tong Opened: 2016-04-12 Last modified: 2016-05-22
Priority: 3
View all other issues in [locale.moneypunct].
View all issues with New status.
Discussion:
The description of money_base::space is that "at least one space is required at that position." (N4582 subclause 22.4.6.3 [locale.moneypunct] paragraph 2)
When formatting for output (30.4.7.3.2 [locale.money.put.virtuals]), it is not clear that"the number of characters generated for the specified format" (excluding fill padding) includes exactly one character for money_base::space (if present), and
all characters corresponding to money_base::space (excluding fill padding) are copies of fill.
In particular, there is implementation divergence over point (b) as to whether U+0020 or fill should be used. Further, should a character other than fill be used, it is unclear when "the fill characters are placed where none or space appears in the formatting pattern", whether the fill characters are placed at the beginning or the end of the "space field".
I believe that a strict interpretation of the current wording supports U+0020; however, fill is more likely to be the pragmatic choice.Proposed resolution:
This wording is relative to N4582.
Change 30.4.7.4 [locale.moneypunct] paragraph 2 as indicated:
-2- Where none or space appears, white space is permitted in the format, except where none appears at the end, in which case no white space is permitted. For input, the value space indicates that at least one space is required at that position. For output, the value space indicates one instance of the fill character (30.4.7.3.2 [locale.money.put.virtuals]).
The value space indicates that at least one space is required at that position. Where symbol appears, the sequence of characters returned by curr_symbol() is permitted, and can be required. Where sign appears, the first (if any) of the sequence of characters returned by positive_sign() or negative_sign() (respectively as the monetary value is non-negative or negative) is required. Any remaining characters of the sign sequence are required after all other format components. Where value appears, the absolute numeric monetary value is required.
Section: 16.4.6.5 [member.functions] Status: New Submitter: Hubert Tong Opened: 2016-04-15 Last modified: 2017-02-02
Priority: 3
View all other issues in [member.functions].
View all issues with New status.
Discussion:
In N4582 subclause 17.6.5.5 [member.functions], the requirement that:
any call to the member function that would select an overload from the set of declarations described in this standard behaves as if that overload were selected
is unclear in the extent of the "as if". For example, in providing:
basic_string(const charT* s);
for a one-argument call to:
basic_string(const charT* s, const Allocator& a = Allocator());
it can be read that an implementation may be required to call the copy constructor for the allocator since the core language rules for copy elision would not allow the "a" argument to be constructed directly into the member used to store the allocator.
Clarification (even if just a note) would be appreciated.[2016-05 Issues Telecon]
This is related to issue 2563.
Proposed resolution:
Section: 30.4.3.3.3 [facet.num.put.virtuals] Status: New Submitter: Hubert Tong Opened: 2016-05-07 Last modified: 2016-05-22
Priority: 3
View other active issues in [facet.num.put.virtuals].
View all other issues in [facet.num.put.virtuals].
View all issues with New status.
Discussion:
The call to do_put(out, str, fill, (int)val) in N4582 subclause 30.4.3.3.3 [facet.num.put.virtuals] paragraph 6 cannot select a best viable function in overload resolution given the overloads listed for do_put in 30.4.3.3 [locale.nm.put].
There is implementation divergence:Some implementations call the long overload (as overriden);
some implementations call the unsigned long overload (as overriden);
some implementations call something else.
It appears that the resolution to DR 359 attempted a fix; however, the relevant portion of the change was not applied to the WP.
Proposed resolution:
Section: 30.4.3.3.3 [facet.num.put.virtuals] Status: New Submitter: Hubert Tong Opened: 2016-05-07 Last modified: 2018-02-11
Priority: 3
View other active issues in [facet.num.put.virtuals].
View all other issues in [facet.num.put.virtuals].
View all issues with New status.
Discussion:
N4582 subclause 30.4.3.3.3 [facet.num.put.virtuals] paragraph 6 makes no provision for fill-padding in its specification of the behaviour when (str.flags() & ios_base::boolalpha) != 0.
[2017-07-06, Marshall comments]
All the other cases from num_put — int, long, etc all are covered in 30.4.3.3.3 [facet.num.put.virtuals] p1 .. p5, which describe how to align and pad the output. (Specifically, stage 3) p6 does not.
With this description:cout << std::setw(15) << false;
outputs:
0 ﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎// Column counter
but
cout << std::setw(15) << boolalpha << false;
outputs:
false
libc++ implements this exactly.
Dinkumware, libstdc++ and MSVC apply padding and alignment.
Proposed resolution:
Section: 24.2.4 [sequence.reqmts] Status: New Submitter: Kazutoshi Satoda Opened: 2016-05-08 Last modified: 2020-09-06
Priority: 3
View other active issues in [sequence.reqmts].
View all other issues in [sequence.reqmts].
View all issues with New status.
Discussion:
Please look through the following modifications on a value v of type vector<T>:
assert(v.size() > 0); v.push_back(v[0]); v.insert(v.begin(), v[0]); v.resize(v.size() * 2, v[0]); v.assign(v.size() * 2, v[0]);
All of these use an element of itself which may be moved or destroyed by the modification.
From what I see so far, the first three are required to work. Please see library issue 526 for validity of them. But only the last one is undefined because it violates a precondition of a sequence container operation. I think this is too subtle. Should it be like that, really? The precondition is in Table 107 "Sequence container requirements" at the next of 24.2.4 [sequence.reqmts] p3.In Tables 107 and 108, X denotes a sequence container class, a denotes a value of X containing elements of type T, […] n denotes a value of X::size_type, […] t denotes an lvalue or a const rvalue of X::value_type, […]
[…]
Table 107 — Sequence container requirements (in addition to container) Expression Return type Assertion/note
pre-/post-condition[…] a.assign(n, t) void Requires: T shall be CopyInsertable into X and CopyAssignable.
pre: t is not a reference into a.
Replaces elements in a with n copies of t.
I looked into the following implementations:
libc++ relies on the precondition.
It deallocates first on n > capacity() case, see here.libstdc++ doesn't rely on the precondition.
It creates temporary vector(n, t) and swap() on n > capacity() case, see here.MSVC relies on the precondition.
It unconditionally does clear() and then insert(begin(), n, t). I looked into my local "%PROGRAMFILES(X86)%/Microsoft Visual Studio 14.0/VC/include/vector".One drawback of libstdc++ implementation, I could find so far, is possibly increased peek memory usage (both old and new buffer exist at the same time). But, because the same can happen on the most other modifications, it seems a reasonable trade-off to remove the precondition to fill the subtle gap. Users who really needs less memory usage can do clear() and insert() by themselves.
I also found that basic_string::assign(n, c) is safe on this point. At 23.4.3.7.3 [string.assign] p17:basic_string& assign(size_type n, charT c);Effects: Equivalent to assign(basic_string(n, c)).
Returns: *this.
This can be seen as another gap.
Looking back on the history, I found that the definition of assign(n, t) was changed at C++14 for library issue 2209. There were more restricting definitions like this:void assign(size_type n, const T& t);Effects:
erase(begin(), end()); insert(begin(), n, t);
I think the precondition was probably set to accept this old definition and is not required inherently. And if the less memory usage was really intended, the standard is now underspecifying about that.
[2016-05 Issues Telecon]
Howard believes this should be NAD, but we tabled the discussion.
Previous resolution [SUPERSEDED]:This wording is relative to N4582.
In 24.2.4 [sequence.reqmts], edit Table 107 (Sequence container requirements) as indicated:
Table 107 — Sequence container requirements (in addition to container) Expression Return type Assertion/note
pre-/post-condition[…] a.assign(n, t) void Requires: T shall be CopyInsertable into X and CopyAssignable.
pre: t is not a reference into a.
Replaces elements in a with n copies of t.
[2020-04-25, Daniel syncs current wording with recent working draft]
Proposed resolution:
This wording is relative to N4861.
In 24.2.4 [sequence.reqmts], edit Table [tab:container.seq.req] (Sequence container requirements) as indicated:
Table 77 — Sequence container requirements (in addition to container) [tab:container.seq.req] Expression Return type Assertion/note
pre-/post-condition[…] a.assign(n, t) void Preconditions: T is Cpp17CopyInsertable into X
and Cpp17CopyAssignable.t is not a reference into a.
Effects: Replaces elements in a with n copies of t.
Invalidates all references, pointers and iterators referring to the elements of a.
For vector and deque, also invalidates the past-the-end iterator.
Section: 31.12.12.2 [fs.rec.dir.itr.members] Status: Open Submitter: Eric Fiselier Opened: 2016-05-09 Last modified: 2018-08-24
Priority: 2
View all other issues in [fs.rec.dir.itr.members].
View all issues with Open status.
Discussion:
The current specification of recursion_pending() says (31.12.12.2 [fs.rec.dir.itr.members]/24):
Returns: true if disable_recursion_pending() has not been called subsequent to the prior construction or increment operation, otherwise false.
This language does not take into account cases where the prior construction was a copy construction from a iterator, it, where it.recursion_pending() == false.
[2016-08 Chicago]
Wed AM: Move to Open
[2018-1-26 issues processing telecon]
Status to 'Tentatively Ready'; Casey will explore whether making recursion_pending an exposition-only member makes this clearer.
Previous resolution from Eric [SUPERSEDED]:This wording is relative to N4582.
Change 31.12.12.2 [fs.rec.dir.itr.members] as indicated:
explicit recursive_directory_iterator(const path& p); recursive_directory_iterator(const path& p, directory_options options); recursive_directory_iterator(const path& p, directory_options options, error_code& ec) noexcept; recursive_directory_iterator(const path& p, error_code& ec) noexcept;[…]
-3- Postcondition:options() == options for the signatures with a directory_options argument, otherwise options() == directory_options::none.
options() == options for the signatures with a directory_options argument, otherwise options() == directory_options::none.
recursion_pending() == true.
[…]
[Drafting note: The following changes the specification of recursion_pending() seemingly recursive. Perhaps it would be easier to specify recursion_pending() in terms of a exposition only member in recursive_directory_iterator.]
bool recursion_pending() const;[…]
-24- Returns:true if disable_recursion_pending() has not been called subsequent to the prior construction or increment operation, otherwise falsefalse if disable_recursion_pending() has been called subsequent to the prior construction or increment operation, otherwise the value of recursion_pending() set by that operation. […]recursive_directory_iterator& operator++(); recursive_directory_iterator& increment(error_code& ec) noexcept;[…]
-27- Effects: As specified by Input iterators (24.2.3), except that: […] -?- Postcondition: recursion_pending() == true.
[2018-01-29: Casey provides a PR with an exposition-only member]
Status to 'Review'.
Previous resolution from Casey [SUPERSEDED]:This wording is relative to N4713.
Change [fs.rec.dir.itr] as indicated:
[…] // other members as required by 25.3.5.3 [input.iterators], input iterators private: bool recurse_; // exposition-only }; }Change 31.12.12.2 [fs.rec.dir.itr.members] as indicated:
explicit recursive_directory_iterator(const path& p); recursive_directory_iterator(const path& p, directory_options options); recursive_directory_iterator(const path& p, directory_options options, error_code& ec) noexcept; recursive_directory_iterator(const path& p, error_code& ec) noexcept;[…]
-3- Postconditions:
options() == options for the signatures with a directory_options argument, otherwise options() == directory_options::none.
options() == options for the signatures with a directory_options argument, otherwise options() == directory_options::none.
recurse_ == true.
[…]
recursive_directory_iterator(const recursive_directory_iterator& rhs);[…]
-8- Postconditions:
[…]
(8.3) —
recursion_pending() == rhs.recursion_pending()recurse_ == rhs.recurse_recursive_directory_iterator(recursive_directory_iterator&& rhs) noexcept;[…]
-10- Postconditions: options(), depth(), and
recursion_pending()recurse_ have the values that rhs.options(), rhs.depth(), andrhs.recursion_pending()rhs.recurse_, respectively, had before the function call.recursive_directory_iterator& operator=(const recursive_directory_iterator& rhs);[…]
-12- Postconditions:
[…]
(12.3) —
recursion_pending() == rhs.recursion_pending()recurse_ == rhs.recurse_[…]
recursive_directory_iterator& operator=(recursive_directory_iterator&& rhs) noexcept;[…]
-15- Postconditions: options(), depth(), and
recursion_pending()recurse_ have the values that rhs.options(), rhs.depth(), andrhs.recursion_pending()rhs.recurse_, respectively, had before the function call.[…]
bool recursion_pending() const;-21- Returns:
true if disable_recursion_pending() has not been called subsequent to the prior construction or increment operation, otherwise falserecurse_.[…]
recursive_directory_iterator& operator++(); recursive_directory_iterator& increment(error_code& ec) noexcept;-23- Effects: As specified for the prefix increment operation of Input iterators ( [iterators.input]), except that:
[…]
-?- Postcondition: recurse_ == true.
void disable_recursion_pending();-28- Postcondition
s:recursion_pending()recurse_ == false.[…]
[2018-05-23: Casey restores the intended design with an expansion of the original PR]
The intended design is that all copies of a single recursive_directory_iterator share a common block of state which includes the values returned by options, depth, and recursion_pending - hence the mandate that those functions not be called on a non-dereferenceable iterator in 31.12.12 [fs.class.rec.dir.itr] para 2. To allow an implementation with such shared state, it's necessary to make changes to the value returned by recursion_pending() visible to all copies of the same dereferenceable iterator.
Also:
pop notionally calls increment repeatedly until the current directory is exhausted, pop should affect the value of recursion_pending similarly to increment.
options is not valid for all constructor signatures described by 31.12.12.2 [fs.rec.dir.itr.members] para 2.
the copies and moves don't specify what they actually do
it's not quite kosher for the copies and moves to have postconditions on the value of expressions that have UB if the iterator copied/moved from is not dereferenceable.
[2018-06, Rapperswil, Wednesday evening]
JW: p21 currently can just say "unspecified"
BO: if we are OK with only remote implementations we can remove the unspecifiedness
BO: the problematic business is the "recursion pending" bit
JW: I want time to work on this
[2018-08-23 Batavia Issues processing]
General agreement that flag should be shared; Casey to reword.
Proposed resolution:
This wording is relative to N4750.
Change 31.12.12.2 [fs.rec.dir.itr.members] as indicated:
explicit recursive_directory_iterator(const path& p); recursive_directory_iterator(const path& p, directory_options options); recursive_directory_iterator(const path& p, directory_options options, error_code& ec) noexcept; recursive_directory_iterator(const path& p, error_code& ec) noexcept;-?- For the signatures with no parameter options, let options be directory_options::none.
-2- Effects: […]
-3- Postconditions:
options() == options for the signatures with a directory_options argument, otherwise options() == directory_options::none.
this->options() == options
recursion_pending() == true
[…]
recursive_directory_iterator(const recursive_directory_iterator& rhs);-7- Effects: Constructs an
object of class recursive_directory_iteratoriterator that denotes the same directory entry as rhs, if any..-8- Postconditions: If rhs is dereferenceable,
[…]
recursive_directory_iterator(recursive_directory_iterator&& rhs) noexcept;-9- Effects: Constructs an
object of class recursive_directory_iteratoriterator that denotes the directory entry denoted by rhs before the function call, if any..-10- Postconditions: If rhs is dereferenceable, […]
recursive_directory_iterator& operator=(const recursive_directory_iterator& rhs);-11- Effects:
If *this and rhs are the same object, the member has no effect.Causes *this to denote the same directory entry denoted by rhs, if any.-12- Postconditions: If rhs is dereferenceable,
[…]
recursive_directory_iterator& operator=(recursive_directory_iterator&& rhs) noexcept;-14- Effects:
If *this and rhs are the same object, the member has no effect.Causes *this to denote the directory entry denoted by rhs before the function call, if any.-15- Postconditions: If rhs was dereferenceable before the function call, […]
-16- Returns: *this.
-x- Remarks: If *this and rhs do not refer to the same object, the resulting state of rhs is unspecified (16.4.6.15 [lib.types.movedfrom]).
directory_options options() const;-17- Returns: The value
of the argument passed to the constructor for the options parameter, if present, otherwise directory_options::noneestablished by the most recently called member that has a postcondition for options().[…]
bool recursion_pending() const;-21- Returns:
true if disable_recursion_pending() has not been called subsequent to the prior construction or increment operation, otherwise false.If disable_recursion_pending() has been called on a copy of *this, an unspecified value. Otherwise, the value established for recursion_pending() by the postcondition of the most recent construction, assignment, increment, or disable_recursion_pending operation.[…]
recursive_directory_iterator& operator++(); recursive_directory_iterator& increment(error_code& ec);-23- Effects: As specified for the prefix increment operation of Input iterators (25.3.5.3 [input.iterators]), except that: […]
-?- Postconditions: If *this is dereferenceable, recursion_pending() == true.
[…]
void pop(); void pop(error_code& ec);-26- Effects: If depth() == 0, set *this to recursive_directory_iterator(). […]
-?- Postconditions: If *this is dereferenceable, recursion_pending() == true.
[…]
Section: 24.5 [unord] Status: New Submitter: Billy Robert O'Neal III Opened: 2016-05-20 Last modified: 2022-07-16
Priority: 3
View all other issues in [unord].
View all issues with New status.
Discussion:
The resolution of LWG 2210 missed constructors accepting a range or initializer list and allocator.
Previous resolution [SUPERSEDED]:
This wording is relative to N4582.
Add to the synopsis in 24.5.4.1 [unord.map.overview] p3:
namespace std { template <class Key, class T, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<std::pair<const Key, T> > > { class unordered_map { public: […] unordered_map(size_type n, const hasher& hf, const allocator_type& a) : unordered_map(n, hf, key_equal(), a) { } template <class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a) : unordered_map(f, l, see below, hasher(), key_equal(), a) { } template <class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_map(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_map(f, l, n, hf, key_equal(), a) { } unordered_map(initializer_list<value_type> il, const allocator_type& a) : unordered_map(il, see below, hasher(), key_equal(), a) { } unordered_map(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_map(il, n, hasher(), key_equal(), a) { } […] }; }Add to the synopsis in 24.5.5.1 [unord.multimap.overview] p3:
namespace std { template <class Key, class T, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<std::pair<const Key, T> > > { class unordered_multimap { public: […] unordered_multimap(size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(n, hf, key_equal(), a) { } template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multimap(f, l, see below, hasher(), key_equal(), a) { } template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multimap(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(f, l, n, hf, key_equal(), a) { } unordered_multimap(initializer_list<value_type> il, const allocator_type& a) : unordered_multimap(il, see below, hasher(), key_equal(), a) { } unordered_multimap(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multimap(il, n, hasher(), key_equal(), a) { } […] }; }Add to the synopsis in 24.5.6.1 [unord.set.overview] p3:
namespace std { template <class Key, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<Key> > { class unordered_set { public: […] unordered_set(size_type n, const hasher& hf, const allocator_type& a) : unordered_set(n, hf, key_equal(), a) { } template <class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a) : unordered_set(f, l, see below, hasher(), key_equal(), a) { } template <class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_set(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_set(f, l, n, hf, key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a) : unordered_set(il, see below, hasher(), key_equal(), a) { } unordered_set(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_set(il, n, hasher(), key_equal(), a) { } […] }; }Add to the synopsis in 24.5.7.1 [unord.multiset.overview] p3:
namespace std { template <class Key, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<Key> > { class unordered_multiset { public: […] unordered_multiset(size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(n, hf, key_equal(), a) { } template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multiset(f, l, see below, hasher(), key_equal(), a) { } template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multiset(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(f, l, n, hf, key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, const allocator_type& a) : unordered_multiset(il, see below, hasher(), key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multiset(il, n, hasher(), key_equal(), a) { } […] }; }
[2016-06, Oulu — Daniel comments and provides new wording]
During the LWG discussion of this issue it has been observed, that the interpretation of the embedded see below is not really clear and that we should split declaration and definition of the new overloads, so that we have a place that allows us to specify what "see below" stands for. In addition, the new wording wraps the "see below" as "size_type(see below)" to clarify the provided expression type, similar as we did for the default constructor of unordered_map.
[Oulu, 2016-06]
Alisdair to review wording.
Previous resolution [SUPERSEDED]:
This wording is relative to N4594.
Add to the synopsis in 24.5.4.1 [unord.map.overview] p3:
namespace std { template <class Key, class T, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<std::pair<const Key, T> > > { class unordered_map { public: […] unordered_map(size_type n, const hasher& hf, const allocator_type& a) : unordered_map(n, hf, key_equal(), a) { } template <class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a); template <class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_map(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_map(f, l, n, hf, key_equal(), a) { } unordered_map(initializer_list<value_type> il, const allocator_type& a); unordered_map(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_map(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.4.2 [unord.map.cnstr] p2
template <class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a) : unordered_map(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_map(initializer_list<value_type> il, const allocator_type& a) : unordered_map(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The number of buckets is implementation-defined.
Add to the synopsis in 24.5.5.1 [unord.multimap.overview] p3:
namespace std { template <class Key, class T, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<std::pair<const Key, T> > > { class unordered_multimap { public: […] unordered_multimap(size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(n, hf, key_equal(), a) { } template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a); template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multimap(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(f, l, n, hf, key_equal(), a) { } unordered_multimap(initializer_list<value_type> il, const allocator_type& a); unordered_multimap(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multimap(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.5.2 [unord.multimap.cnstr] p2
template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multimap(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_multimap(initializer_list<value_type> il, const allocator_type& a) : unordered_multimap(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The number of buckets is implementation-defined.
Add to the synopsis in 24.5.6.1 [unord.set.overview] p3:
namespace std { template <class Key, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<Key> > { class unordered_set { public: […] unordered_set(size_type n, const hasher& hf, const allocator_type& a) : unordered_set(n, hf, key_equal(), a) { } template <class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a); template <class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_set(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_set(f, l, n, hf, key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a); unordered_set(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_set(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.6.2 [unord.set.cnstr] p2
template <class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a) : unordered_set(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a) : unordered_set(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The number of buckets is implementation-defined.
Add to the synopsis in 24.5.7.1 [unord.multiset.overview] p3:
namespace std { template <class Key, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<Key> > { class unordered_multiset { public: […] unordered_multiset(size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(n, hf, key_equal(), a) { } template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a); template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multiset(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(f, l, n, hf, key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, const allocator_type& a); unordered_multiset(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multiset(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.7.2 [unord.multiset.cnstr] p2
template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multiset(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, const allocator_type& a) : unordered_multiset(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The number of buckets is implementation-defined.
[2017-08-04, Daniel and Alisdair finetune wording]
We decided to improve the added Remarks: elements by changing from the previous form:
Remarks: The number of buckets is implementation-defined.
to the more elaborate form:
Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
[2020-11-29; Reflector discussions]
It has been pointed out that this issue is related to LWG 1199, LWG 2210, and LWG 3506.
Previous resolution [SUPERSEDED]:
This resolution is relative to N4687.
Add to the synopsis in 24.5.4.1 [unord.map.overview] p3:
namespace std { template <class Key, class T, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<std::pair<const Key, T> > > { class unordered_map { public: […] unordered_map(size_type n, const hasher& hf, const allocator_type& a) : unordered_map(n, hf, key_equal(), a) { } template <class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a); template <class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_map(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_map(f, l, n, hf, key_equal(), a) { } unordered_map(initializer_list<value_type> il, const allocator_type& a); unordered_map(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_map(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.4.2 [unord.map.cnstr] p2
template <class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a) : unordered_map(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_map(initializer_list<value_type> il, const allocator_type& a) : unordered_map(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.5.1 [unord.multimap.overview] p3:
namespace std { template <class Key, class T, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<std::pair<const Key, T> > > { class unordered_multimap { public: […] unordered_multimap(size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(n, hf, key_equal(), a) { } template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a); template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multimap(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(f, l, n, hf, key_equal(), a) { } unordered_multimap(initializer_list<value_type> il, const allocator_type& a); unordered_multimap(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multimap(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.5.2 [unord.multimap.cnstr] p2
template <class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multimap(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_multimap(initializer_list<value_type> il, const allocator_type& a) : unordered_multimap(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.6.1 [unord.set.overview] p3:
namespace std { template <class Key, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<Key> > { class unordered_set { public: […] unordered_set(size_type n, const hasher& hf, const allocator_type& a) : unordered_set(n, hf, key_equal(), a) { } template <class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a); template <class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_set(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_set(f, l, n, hf, key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a); unordered_set(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_set(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.6.2 [unord.set.cnstr] p2
template <class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a) : unordered_set(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a) : unordered_set(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.7.1 [unord.multiset.overview] p3:
namespace std { template <class Key, class Hash = hash<Key>, class Pred = std::equal_to<Key>, class Allocator = std::allocator<Key> > { class unordered_multiset { public: […] unordered_multiset(size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(n, hf, key_equal(), a) { } template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a); template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multiset(f, l, n, hasher(), key_equal(), a) { } template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(f, l, n, hf, key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, const allocator_type& a); unordered_multiset(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multiset(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.7.2 [unord.multiset.cnstr] p2
template <class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multiset(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, const allocator_type& a) : unordered_multiset(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
[2022-07-10; Daniel comments]
It is has been noticed by Daniel Eiband on [std-discussion] that the following deduction guides for the following constructors of the set types std::unordered_set and std::unordered_multiset are missing:
unordered_set(InputIterator, InputIterator, Allocator); unordered_set(initializer_list<T>, Allocator); unordered_multiset(InputIterator, InputIterator, Allocator); unordered_multiset(initializer_list<T>, Allocator);
Since this issue is adding these missing constructors it should also add the associated deduction guides. The proposed wording has been updated to this effect and also rebased to N4910.
Previous resolution [SUPERSEDED]:
This resolution is relative to N4910.
Add to the synopsis in 24.5.4.1 [unord.map.overview] p3:
namespace std { template<class Key, class T, class Hash = hash<Key>, class Pred = equal_to<Key>, class Allocator = allocator<pair<const Key, T>>> { class unordered_map { public: […] unordered_map(size_type n, const hasher& hf, const allocator_type& a) : unordered_map(n, hf, key_equal(), a) { } template<class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a); template<class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_map(f, l, n, hasher(), key_equal(), a) { } template<class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_map(f, l, n, hf, key_equal(), a) { } […] unordered_map(initializer_list<value_type> il, const allocator_type& a); unordered_map(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_map(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.4.2 [unord.map.cnstr] p2
template<class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a) : unordered_map(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_map(initializer_list<value_type> il, const allocator_type& a) : unordered_map(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.5.1 [unord.multimap.overview] p3:
namespace std { template<class Key, class T, class Hash = hash<Key>, class Pred = equal_to<Key>, class Allocator = allocator<pair<const Key, T>>> { class unordered_multimap { public: […] unordered_multimap(size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(n, hf, key_equal(), a) { } template<class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a); template<class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multimap(f, l, n, hasher(), key_equal(), a) { } template<class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(f, l, n, hf, key_equal(), a) { } […] unordered_multimap(initializer_list<value_type> il, const allocator_type& a); unordered_multimap(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multimap(il, n, hasher(), key_equal(), a) { } […] }; }Insert the following new prototype specification just after 24.5.5.2 [unord.multimap.cnstr] p2
template<class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multimap(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_multimap(initializer_list<value_type> il, const allocator_type& a) : unordered_multimap(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.6.1 [unord.set.overview] p3:
namespace std { template<class Key, class Hash = hash<Key>, class Pred = equal_to<Key>, class Allocator = allocator<Key>> { class unordered_set { public: […] unordered_set(size_type n, const hasher& hf, const allocator_type& a) : unordered_set(n, hf, key_equal(), a) { } template<class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a); template<class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_set(f, l, n, hasher(), key_equal(), a) { } template<class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_set(f, l, n, hf, key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a); unordered_set(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_set(il, n, hasher(), key_equal(), a) { } […] }; […] template<class T, class Allocator> unordered_set(initializer_list<T>, typename see below::size_type, Allocator) -> unordered_set<T, hash<T>, equal_to<T>, Allocator>; template<class T, class Hash, class Allocator> unordered_set(initializer_list<T>, typename see below::size_type, Hash, Allocator) -> unordered_set<T, Hash, equal_to<T>, Allocator>; template<class InputIterator, class Allocator> unordered_set(InputIterator, InputIterator, Allocator) -> unordered_set<iter-value-type<InputIterator>, hash<iter-value-type<InputIterator>>, equal_to<iter-value-type<InputIterator>>, Allocator>; template<class T, class Allocator> unordered_set(initializer_list<T>, Allocator) -> unordered_set<T, hash<T>, equal_to<T>, Allocator>; }Insert the following new prototype specification just after 24.5.6.2 [unord.set.cnstr] p2
template<class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a) : unordered_set(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a) : unordered_set(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.7.1 [unord.multiset.overview] p3:
namespace std { template<class Key, class Hash = hash<Key>, class Pred = equal_to<Key>, class Allocator = allocator<Key>> { class unordered_multiset { public: […] unordered_multiset(size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(n, hf, key_equal(), a) { } template<class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a); template<class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multiset(f, l, n, hasher(), key_equal(), a) { } template<class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(f, l, n, hf, key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, const allocator_type& a); unordered_multiset(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multiset(il, n, hasher(), key_equal(), a) { } […] }; […] template<class T, class Allocator> unordered_multiset(initializer_list<T>, typename see below ::size_type, Allocator) -> unordered_multiset<T, hash<T>, equal_to<T>, Allocator>; template<class T, class Hash, class Allocator> unordered_multiset(initializer_list<T>, typename see below ::size_type, Hash, Allocator) -> unordered_multiset<T, Hash, equal_to<T>, Allocator>; template<class InputIterator, class Allocator> unordered_multiset(InputIterator, InputIterator, Allocator) -> unordered_multiset<iter-value-type<InputIterator>, hash<iter-value-type<InputIterator>>, equal_to<iter-value-type<InputIterator>>, Allocator>; template<class T, class Allocator> unordered_multiset(initializer_list<T>, Allocator) -> unordered_multiset<T, hash<T>, equal_to<T>, Allocator>; }Insert the following new prototype specification just after 24.5.7.2 [unord.multiset.cnstr] p2
template<class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multiset(f, l, size_type(see below), hasher(), key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, const allocator_type& a) : unordered_multiset(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
[2022-07-15; Casey comments]
P1206R7 added from_range_t constructors corresponding to existing iterator pair constructors for the standard containers. For consistency, this issue should add from_range_t constructors corresponding to each new iterator pair constructor.
[2022-07-16; Daniel comments and updates wording]
The new from_range_t constructors have been added for each added new iterator pair constructor. Note that the corresponding deduction guides already exist.
Proposed resolution:
This resolution is relative to N4910.
Add to the synopsis in 24.5.4.1 [unord.map.overview] p3:
namespace std { template<class Key, class T, class Hash = hash<Key>, class Pred = equal_to<Key>, class Allocator = allocator<pair<const Key, T>>> { class unordered_map { public: […] unordered_map(size_type n, const hasher& hf, const allocator_type& a) : unordered_map(n, hf, key_equal(), a) { } template<class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a); template<class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_map(f, l, n, hasher(), key_equal(), a) { } template<class InputIterator> unordered_map(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_map(f, l, n, hf, key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_map(from_range_t, R&& rg, const allocator_type& a); template<container-compatible-range<value_type> R> unordered_map(from_range_t, R&& rg, size_type n, const allocator_type& a) : unordered_map(from_range, std::forward<R>(rg), n, hasher(), key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_map(from_range_t, R&& rg, size_type n, const hasher& hf, const allocator_type& a) : unordered_map(from_range, std::forward<R>(rg), n, hf, key_equal(), a) { } unordered_map(initializer_list<value_type> il, const allocator_type& a); unordered_map(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_map(il, n, hasher(), key_equal(), a) { } […] }; }
Insert the following new prototype specification just after 24.5.4.2 [unord.map.cnstr] p2
template<class InputIterator> unordered_map(InputIterator f, InputIterator l, const allocator_type& a) : unordered_map(f, l, size_type(see below), hasher(), key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_map(from_range_t, R&& rg, const allocator_type& a) : unordered_map(from_range, std::forward<R>(rg), size_type(see below), hasher(), key_equal(), a) { } unordered_map(initializer_list<value_type> il, const allocator_type& a) : unordered_map(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.5.1 [unord.multimap.overview] p3:
namespace std { template<class Key, class T, class Hash = hash<Key>, class Pred = equal_to<Key>, class Allocator = allocator<pair<const Key, T>>> { class unordered_multimap { public: […] unordered_multimap(size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(n, hf, key_equal(), a) { } template<class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a); template<class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multimap(f, l, n, hasher(), key_equal(), a) { } template<class InputIterator> unordered_multimap(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multimap(f, l, n, hf, key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_multimap(from_range_t, R&& rg, const allocator_type& a); template<container-compatible-range<value_type> R> unordered_multimap(from_range_t, R&& rg, size_type n, const allocator_type& a) : unordered_multimap(from_range, std::forward<R>(rg), n, hasher(), key_equal(), a) { } […] unordered_multimap(initializer_list<value_type> il, const allocator_type& a); unordered_multimap(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multimap(il, n, hasher(), key_equal(), a) { } […] }; }
Insert the following new prototype specification just after 24.5.5.2 [unord.multimap.cnstr] p2
template<class InputIterator> unordered_multimap(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multimap(f, l, size_type(see below), hasher(), key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_multimap(from_range_t, R&& rg, const allocator_type& a) : unordered_multimap(from_range, std::forward<R>(rg), size_type(see below), hasher(), key_equal(), a) { } unordered_multimap(initializer_list<value_type> il, const allocator_type& a) : unordered_multimap(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.6.1 [unord.set.overview] p3:
namespace std { template<class Key, class Hash = hash<Key>, class Pred = equal_to<Key>, class Allocator = allocator<Key>> { class unordered_set { public: […] unordered_set(size_type n, const hasher& hf, const allocator_type& a) : unordered_set(n, hf, key_equal(), a) { } template<class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a); template<class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_set(f, l, n, hasher(), key_equal(), a) { } template<class InputIterator> unordered_set(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_set(f, l, n, hf, key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a); unordered_set(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_set(il, n, hasher(), key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_set(from_range_t, R&& rg, const allocator_type& a); template<container-compatible-range<value_type> R> unordered_set(from_range_t, R&& rg, size_type n, const allocator_type& a) : unordered_set(from_range, std::forward<R>(rg), n, hasher(), key_equal(), a) { } […] }; […] template<class T, class Allocator> unordered_set(initializer_list<T>, typename see below::size_type, Allocator) -> unordered_set<T, hash<T>, equal_to<T>, Allocator>; template<class T, class Hash, class Allocator> unordered_set(initializer_list<T>, typename see below::size_type, Hash, Allocator) -> unordered_set<T, Hash, equal_to<T>, Allocator>; template<class InputIterator, class Allocator> unordered_set(InputIterator, InputIterator, Allocator) -> unordered_set<iter-value-type<InputIterator>, hash<iter-value-type<InputIterator>>, equal_to<iter-value-type<InputIterator>>, Allocator>; template<class T, class Allocator> unordered_set(initializer_list<T>, Allocator) -> unordered_set<T, hash<T>, equal_to<T>, Allocator>; }
Insert the following new prototype specification just after 24.5.6.2 [unord.set.cnstr] p2
template<class InputIterator> unordered_set(InputIterator f, InputIterator l, const allocator_type& a) : unordered_set(f, l, size_type(see below), hasher(), key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_set(from_range_t, R&& rg, const allocator_type& a) : unordered_set(from_range, std::forward<R>(rg), size_type(see below), hasher(), key_equal(), a) { } unordered_set(initializer_list<value_type> il, const allocator_type& a) : unordered_set(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Add to the synopsis in 24.5.7.1 [unord.multiset.overview] p3:
namespace std { template<class Key, class Hash = hash<Key>, class Pred = equal_to<Key>, class Allocator = allocator<Key>> { class unordered_multiset { public: […] unordered_multiset(size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(n, hf, key_equal(), a) { } template<class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a); template<class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const allocator_type& a) : unordered_multiset(f, l, n, hasher(), key_equal(), a) { } template<class InputIterator> unordered_multiset(InputIterator f, InputIterator l, size_type n, const hasher& hf, const allocator_type& a) : unordered_multiset(f, l, n, hf, key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_multiset(from_range_t, R&& rg, const allocator_type& a); template<container-compatible-range<value_type> R> unordered_multiset(from_range_t, R&& rg, size_type n, const allocator_type& a) : unordered_multiset(from_range, std::forward<R>(rg), n, hasher(), key_equal(), a) { } […] unordered_multiset(initializer_list<value_type> il, const allocator_type& a); unordered_multiset(initializer_list<value_type> il, size_type n, const allocator_type& a) : unordered_multiset(il, n, hasher(), key_equal(), a) { } […] }; […] template<class T, class Allocator> unordered_multiset(initializer_list<T>, typename see below ::size_type, Allocator) -> unordered_multiset<T, hash<T>, equal_to<T>, Allocator>; template<class T, class Hash, class Allocator> unordered_multiset(initializer_list<T>, typename see below ::size_type, Hash, Allocator) -> unordered_multiset<T, Hash, equal_to<T>, Allocator>; template<class InputIterator, class Allocator> unordered_multiset(InputIterator, InputIterator, Allocator) -> unordered_multiset<iter-value-type<InputIterator>, hash<iter-value-type<InputIterator>>, equal_to<iter-value-type<InputIterator>>, Allocator>; template<class T, class Allocator> unordered_multiset(initializer_list<T>, Allocator) -> unordered_multiset<T, hash<T>, equal_to<T>, Allocator>; }
Insert the following new prototype specification just after 24.5.7.2 [unord.multiset.cnstr] p2
template<class InputIterator> unordered_multiset(InputIterator f, InputIterator l, const allocator_type& a) : unordered_multiset(f, l, size_type(see below), hasher(), key_equal(), a) { } template<container-compatible-range<value_type> R> unordered_multiset(from_range_t, R&& rg, const allocator_type& a)) : unordered_multiset(from_range, std::forward<R>(rg), size_type(see below), hasher(), key_equal(), a) { } unordered_multiset(initializer_list<value_type> il, const allocator_type& a) : unordered_multiset(il, size_type(see below), hasher(), key_equal(), a) { }-?- Remarks: The initial number of buckets supplied by the size_type argument is implementation-defined.
Section: 28.4.6 [complex.ops] Status: New Submitter: Tim Song Opened: 2016-05-23 Last modified: 2018-10-16
Priority: 3
View all other issues in [complex.ops].
View all issues with New status.
Discussion:
The specification of operator>>(istream&, complex<T>&) is extremely short on details. It currently reads, in its entirety (28.4.6 [complex.ops]/12-15):
template<class T, class charT, class traits> basic_istream<charT, traits>& operator>>(basic_istream<charT, traits>& is, complex<T>& x);Effects: Extracts a complex number x of the form: u, (u), or (u,v), where u is the real part and v is the imaginary part (31.7.5.3 [istream.formatted]).
Requires: The input values shall be convertible to T. If bad input is encountered, calls is.setstate(ios_base::failbit) (which may throw ios::failure (31.5.4.4 [iostate.flags])). Returns: is. Remarks: This extraction is performed as a series of simpler extractions. Therefore, the skipping of whitespace is specified to be the same for each of the simpler extractions.
It is completely unclear:
Previous resolution [SUPERSEDED]:
Drafting note: the following wording is based on:
- Characters are extracted using operator>> and compared using traits::eq.
- Mismatched characters are returned to the stream.
This wording is relative to N4582.
Replace 28.4.6 [complex.ops]/12-15 with the following paragraphs:
template<class T, class charT, class traits> basic_istream<charT, traits>& operator>>(basic_istream<charT, traits>& is, complex<T>& x);-?- Effects: First, extracts a character from is.
In the description above, characters are extracted from is as if by operator>> (31.7.5.3.3 [istream.extractors]), and returned to the stream as if by basic_istream::putback (31.7.5.4 [istream.unformatted]). Character equality is determined using traits::eq. An object t of type T is extracted from is as if by is >> t. If any extraction operation fails, no further operation is performed and the whole extraction fails. On failure, calls is.setstate(ios_base::failbit) (which may throw ios::failure (31.5.4.4 [iostate.flags])). -?- Returns: is. -?- [Note: This extraction is performed as a series of simpler extractions. Therefore, the skipping of whitespace is specified to be the same for each of the simpler extractions. — end note]
- If the character extracted is equal to is.widen('('), extracts an object u of type T from is, then extracts a character from is.
- If this character is equal to is.widen(')'), then assigns complex<T>(u) to x.
- Otherwise, if this character is equal to is.widen(','), extracts an object v of type T from is, then extracts a character from is. If this character is equal to is.widen(')'), then assigns complex<T>(u, v) to x; otherwise returns the character to is and the extraction fails.
- Otherwise, returns the character to is and the extraction fails.
- Otherwise, returns the character to is, extracts an object u of type T from is, and assigns complex<T>(u) to x.
[2017-12-13 Tim Song adjusts the P/R to avoid relying on putback.]
Proposed resolution:
Drafting note: the following wording assumes that:
- Characters are extracted using operator>> and compared using traits::eq.
- Mismatched characters are not extracted.
- x is assigned a value-initialized complex on failure for consistency with the arithmetic extractors (compare LWG 696).
This wording is relative to N4778.
Replace 28.4.6 [complex.ops]/12-16 with the following paragraphs:
template<class T, class charT, class traits> basic_istream<charT, traits>& operator>>(basic_istream<charT, traits>& is, complex<T>& x);-?- Effects: Let PEEK(is) be a formatted input function (31.7.5.3.1 [istream.formatted.reqmts]) of is that returns the next character that would be extracted from is by operator>>. [Note: The sentry object is constructed and destroyed, but the returned character is not extracted from the stream. — end note]
In the description above, characters are extracted from is as if by operator>> (31.7.5.3.3 [istream.extractors]), character equality is determined using traits::eq, and an object t of type T is extracted from is as if by is >> t. If any extraction operation fails, no further operation is performed and the whole extraction fails. On failure, assigns complex<T>() to x and calls is.setstate(ios_base::failbit) (which may throw ios::failure (31.5.4.4 [iostate.flags])). -?- Returns: is. -?- [Note: This extraction is performed as a series of simpler extractions. Therefore, the skipping of whitespace is specified to be the same for each of the simpler extractions. — end note]
- If PEEK(is) is not equal to is.widen('('), extracts an object u of type T from is, and assigns complex<T>(u) to x.
- Otherwise, extracts that character from is, then extracts an object u of type T from is, then:
- If PEEK(is) is equal to is.widen(')'), then extracts that character from is and assigns complex<T>(u) to x.
- Otherwise, if it is equal to is.widen(','), then extracts that character from is and then extracts an object v of type T from is, then:
- If PEEK(is) is equal to is.widen(')'), then extracts that character from is and assigns complex<T>(u, v) to x.
- Otherwise, the extraction fails.
- Otherwise, the extraction fails.
Section: 17.3.5 [numeric.limits] Status: Open Submitter: Richard Smith Opened: 2016-06-09 Last modified: 2018-11-12
Priority: 3
View other active issues in [numeric.limits].
View all other issues in [numeric.limits].
View all issues with Open status.
Discussion:
I've received this report at the project editor mail alias, and it seems like it may be worthy of a LWG issue:
I recently had this problem:
- I was storing data in a vector of __uint128_ts
- I used a sorting library which used numeric_limits<T>::max() as a sentinel value
- GCC's libstdc++ provides a numeric_limits specialisation for that type, but
- Clang's libc++ does not.
This broke the sorting for me on different platforms, and it was quite difficult to determine why. If the default numeric_limits didn't default to 0s and false values (18.3.2.4 of N4582), and instead static_asserted, causing my code to not compile, I would have found the solution immediately.
I know that __uint128_t is non-standard, so neither GCC nor Clang is doing the wrong thing nor the right thing here. I could just submit a patch to libc++ providing the specialisations, but it doesn't fix the problem at its core. I am wondering, what is the rationale behind the defaults being 0 and false? It seems like it is inviting a problem for any future numeric types, whether part of a library, compiler extension, and possibly even future updates to C++'s numeric types. I think it would be much better to prevent code that tries to use unspecified numeric_limits from compiling.
An alternative to this suggestion would be to still define the primary template, but not provide any of the members except is_specialized. Either way, this would make numeric_limits members SFINAEable.
Along the same lines, one might wonder why the members that only make sense for floating-point types are required to be defined to nonsense values for integer types.[2016-11-12, Issaquah]
Sat PM: This looks like a good idea. Jonathan and Marshall will do post C++17 implementations and report back.
[2018-11 San Diego Thursday night issue processing]
See Walter's paper P0437 for ideas and/or future directions.
Proposed resolution:
Section: 17.7.3.2 [new.delete.single] Status: New Submitter: Clark Nelson Opened: 2016-06-21 Last modified: 2016-08-01
Priority: 3
View other active issues in [new.delete.single].
View all other issues in [new.delete.single].
View all issues with New status.
Discussion:
It should be considered whether the description of the single-object allocation functions should say "or smaller", like the array allocation functions. For example, according to 17.7.3.2 [new.delete.single] p1 (emphasis mine):
The allocation function (3.7.4.1) called by a new-expression (5.3.4) to allocate size bytes of storage suitably aligned to represent any object of that size.
In contrast to this, 17.7.3.3 [new.delete.array] p1 says (emphasis mine):
The allocation function (3.7.4.1) called by the array form of a new-expression (5.3.4) to allocate size bytes of storage suitably aligned to represent any array object of that size or smaller. (footnote: It is not the direct responsibility of operator new[](std::size_t) or operator delete[](void*) to note the repetition count or element size of the array. Those operations are performed elsewhere in the array new and delete expressions. The array new expression, may, however, increase the size argument to operator new[](std::size_t) to obtain space to store supplemental information.)
Proposed resolution:
Section: 22.5.3.4 [optional.assign], 22.6.3.5 [variant.mod], 22.7.4.4 [any.modifiers] Status: New Submitter: Richard Smith Opened: 2016-07-13 Last modified: 2020-05-10
Priority: 3
View all issues with New status.
Discussion:
Referring to N4604:
In [optional.object.assign]: emplace (normal form) has a Requires that the construction works.
Requires: is_constructible_v<T, Args&&...> is true.
emplace (initializer_list form) has a SFINAE condition:
Remarks: […] unless is_constructible_v<T, initializer_list<U>&, Args&&...> is true.
In 22.7.4.4 [any.modifiers]: emplace (normal form) has a Requires that the construction works:
Requires: is_constructible_v<T, Args...> is true.
emplace (initializer_list form) has a SFINAE condition:
Remarks: […] unless is_constructible_v<T, initializer_list<U>&, Args...> is true.
In 22.6.3.5 [variant.mod]: emplace (T, normal form) has a SFINAE condition:
Remarks: […] unless is_constructible_v<T, Args...> is true, and T occurs exactly once in Types....
emplace (Idx, normal form) has a both a Requires and a SFINAE condition:
Requires: I < sizeof...(Types)
Remarks: […] unless is_constructible_v<T, Args...> is true, and T occurs exactly once in Types....
emplace (T, initializer_list form) has a SFINAE condition:
Remarks: […] unless is_constructible_v<T, initializer_list<U>&, Args...> is true, and T occurs exactly once in Types....
emplace (Idx, initializer_list form) has a both a Requires and a SFINAE condition:
Requires: I < sizeof...(Types)
Remarks: […] unless is_constructible_v<T, Args...> is true, and T occurs exactly once in Types....
Why the inconsistency? Should all the cases have a SFINAE requirement?
I see that variant has an additional requirement (T occurs exactly once in Types...), but that only agues that it must be a SFINAE condition — doesn't say that the other cases (any/variant) should not. map/multimap/unordered_map/unordered_multimap have SFINAE'd versions of emplace that don't take initializer_lists, but they don't have any emplace versions that take ILs. Suggested resolution: Add SFINAE requirements to optional::emplace(Args&&... args) and any::emplace(Args&&... args);[2016-08 Chicago]
During issue prioritization, people suggested that this might apply to any as well.
Ville notes that 2746, 2754 and 2756 all go together.
[2020-05-10; Daniel comments and provides wording]
The inconsistency between the two any::emplace overloads have been removed by resolving LWG 2754 to use Constraints: elements. The last Mandating paper (P1460R1), adopted in Prague, changed the Requires: elements for variant::emplace, "I < sizeof...(Types)" to Mandates:, but that paper was focused on fixing inappropriate preconditions, not searching for consistency here. Given that the in_place_index_t constructors of variant uses SFINAE-conditions for this form of static precondition violation, I recommend that its emplace functions use the same style, which would bring them also in consistency with their corresponding type-based emplace forms that are Mandates:-free but delegate to the index-based forms.
Proposed resolution:
This wording is relative to N4861.
Modify 22.5.3.4 [optional.assign], as indicated:
template<class... Args> T& emplace(Args&&... args);-29-
[…]MandatesConstraints: is_constructible_v<T, Args...> is true.template<class U, class... Args> T& emplace(initializer_list<U> il, Args&&... args);-35- Constraints: is_constructible_v<T, initializer_list<U>&, Args...> is true.
[…]
Modify 22.6.3.5 [variant.mod], as indicated:
template<class T, class... Args> T& emplace(Args&&... args);-1- Constraints: is_constructible_v<T, Args...> is true, and T occurs exactly once in Types.
[…]template<class T, class U, class... Args> T& emplace(initializer_list<U> il, Args&&... args);-3- Constraints: is_constructible_v<T, initializer_list<U>&, Args...> is true, and T occurs exactly once in Types.
[…]template<size_t I, class... Args> variant_alternative_t<I, variant<Types...>>& emplace(Args&&... args);-6- Constraints: is_constructible_v<TI, Args...> is true and I < sizeof...(Types) is true. […]
-5- Mandates: I < sizeof...(Types).template<size_t I, class U, class... Args> variant_alternative_t<I, variant<Types...>>& emplace(initializer_list<U> il, Args&&... args);-13- Constraints: is_constructible_v<TI, initializer_list<U>&, Args...> is true and I < sizeof...(Types) is true. […]
-12- Mandates: I < sizeof...(Types).
Section: 20.3.2.2.3 [util.smartptr.shared.dest] Status: New Submitter: Aaron Jacobs Opened: 2016-07-21 Last modified: 2017-10-08
Priority: 4
View all other issues in [util.smartptr.shared.dest].
View all issues with New status.
Discussion:
The C++14 standard contains no language that guarantees the deleter run by a shared_ptr will see all associated weak_ptr instances as expired. For example, the standard doesn't appear to guarantee that the assertion in the following snippet won't fire:
std::weak_ptr<Foo> weak; std::shared_ptr<Foo> strong{ new Foo, [&weak] (Foo* f) { assert(weak.expired()); delete f; }, }; weak = strong; strong.reset();
It seems clear that the intent is that associated weak_ptrs are expired, because otherwise shared_ptr deleters could resurrect a reference to an object that is being deleted.
Suggested fix: 20.3.2.2.3 [util.smartptr.shared.dest] should specify that the decrease in use_count() caused by the destructor is sequenced before the call to the deleter or the call to delete p.[2016-11-08, Jonathan and STL suggest NAD]
STL and Jonathan feel that the example has unspecified behaviour, and the assertion is allowed to fire, and that's OK (the program's expectation is not reasonable). Otherwise it's necessary to move-construct a copy of the deleter and use that copy to destroy the owned pointer. We do not want to be required to do that.
See also 2262.
[2017-09-20, Jonathan comments]
I'd like to withdraw my NAD suggestion. The value of use_count() is already observable during the destructor via shared_ptr and weak_ptr objects that share ownership, so specifying when it changes ensures correct behaviour.
Proposed resolution:
Section: 22.3.3 [pairs.spec], 22.4.12 [tuple.special], 22.5.9 [optional.specalg], 22.6.10 [variant.specalg], 20.3.1.6 [unique.ptr.special], 24.3.7.4 [array.special], 24.6.6.6 [queue.special], 24.6.7.5 [priqueue.special], 24.6.8.7 [stack.special] Status: New Submitter: Agustín K-ballo Bergé Opened: 2016-08-15 Last modified: 2020-09-06
Priority: 3
View all other issues in [pairs.spec].
View all issues with New status.
Discussion:
Related: 2748 swappable traits for optionals, 2749 swappable traits for variants.
The adoption of P0185R1 "Adding [nothrow-]swappable traits" makes certain non-swappable types indirectly swappable. Consider a type defined as follows:
struct non_swappable { friend void swap(non_swappable&, non_swappable&) = delete; }; non_swappable ns1, ns2; using std::swap; swap(ns1, ns2); // ill-formed static_assert(std::is_swappable_v<non_swappable> == false); // holds
Lvalues of type non_swappable are not swappable, as defined by 16.4.4.3 [swappable.requirements], overload resolution selects the deleted function. Consistently, is_swappable_v<non_swappable> yields false. It should be noted that since non_swappable is move constructible and move assignable, a qualified call to std::swap would be well-formed, even under P0185. Now consider the following snippet:
std::tuple<non_swappable> tns1, tns2; using std::swap; swap(tns1, tns2); // previously ill-formed, now well-formed static_assert(std::is_swappable_v<std::tuple<non_swappable>> == false); // fires
Before P0185, this snippet would violate the implicit requirement of specialized swap for tuples that each tuple element be swappable. After P0185, this specialized swap overload for tuples would be SFINAEd away, resulting in overload resolution selecting the base swap overload, and performing the exchange via move construction and move assignment of tuples.
This issue affects all of pair, tuple, unique_ptr, array, queue, priority_queue, stack, and should eventually also apply to optional and variant. Previous resolution [SUPERSEDED]:This wording is relative to N4606, except when otherwise noted.
Modify 22.3.3 [pairs.spec] as indicated:
template<class T1, class T2> void swap(pair<T1, T2>& x, pair<T1, T2>& y) noexcept(noexcept(x.swap(y)));-7- Effects: As if by x.swap(y).
-8- Remarks: This function shallnot participate in overload resolutionbe defined as deleted unless is_swappable_v<T1> is true and is_swappable_v<T2> is true.Modify 22.4.12 [tuple.special] as indicated:
template <class... Types> void swap(tuple<Types...>& x, tuple<Types...>& y) noexcept(see below);-1- Remarks: This function shall
not participate in overload resolutionbe defined as deleted unless is_swappable_v<Ti> is true for all i, where 0 <= i and i < sizeof...(Types). The expression inside noexcept is equivalent to:noexcept(x.swap(y))-2- Effects: As if by x.swap(y).
Modify 20.3.1.6 [unique.ptr.special] as indicated:
template <class T, class D> void swap(unique_ptr<T, D>& x, unique_ptr<T, D>& y) noexcept;-1- Remarks: This function shall
-2- Effects: Calls x.swap(y).not participate in overload resolutionbe defined as deleted unless is_swappable_v<D> is true.Modify 24.3.7.4 [array.special] as indicated:
template <class T, size_t N> void swap(array<T, N>& x, array<T, N>& y) noexcept(noexcept(x.swap(y)));-1- Remarks: This function shall
-2- Effects: As if by x.swap(y). […]not participate in overload resolutionbe defined as deleted unless N == 0 or is_swappable_v<T> is true.Modify 24.6.6.6 [queue.special] as indicated:
template <class T, class Container> void swap(queue<T, Container>& x, queue<T, Container>& y) noexcept(noexcept(x.swap(y)));-1- Remarks: This function shall
-2- Effects: As if by x.swap(y).not participate in overload resolutionbe defined as deleted unless is_swappable_v<Container> is true.Modify 24.6.7.5 [priqueue.special] as indicated:
template <class T, class Container, class Compare> void swap(priority_queue<T, Container, Compare>& x, priority_queue<T, Container, Compare>& y) noexcept(noexcept(x.swap(y)));-1- Remarks: This function shall
-2- Effects: As if by x.swap(y).not participate in overload resolutionbe defined as deleted unless is_swappable_v<Container> is true and is_swappable_v<Compare> is true.Modify 24.6.8.7 [stack.special] as indicated:
template <class T, class Container> void swap(stack<T, Container>& x, stack<T, Container>& y) noexcept(noexcept(x.swap(y)));-1- Remarks: This function shall
-2- Effects: As if by x.swap(y).not participate in overload resolutionbe defined as deleted unless is_swappable_v<Container> is true.Modify 22.5.9 [optional.specalg] as indicated:
This change should be performed if and only if LWG 2748 is accepted and is against the wording of 2748:
template <class T> void swap(optional<T>& x, optional<T>& y) noexcept(noexcept(x.swap(y)));-1- Effects: Calls x.swap(y).
-2- Remarks: This function shallnot participate in overload resolutionbe defined as deleted unless is_move_constructible_v<T> is true and is_swappable_v<T> is true.Modify 22.6.10 [variant.specalg] as indicated:
This change should be performed if and only if LWG 2749 is accepted and is against the wording of 2749:
template <class... Types> void swap(variant<Types...>& v, variant<Types...>& w) noexcept(see below);-1- Effects: Equivalent to v.swap(w).
-2- Remarks: This function shallnot participate in overload resolutionbe defined as deleted unless is_move_constructible_v<Ti> && is_swappable_v<Ti> is true for all i. The expression inside noexcept is equivalent to noexcept(v.swap(w)).
[2019-04-17 Jonathan updates proposed resolution based on Ville's 2016-11-17 observation that the container adaptors always require swappable sequences anyway. The new proposed resolution is based on the latest WP, "de-shalled", and Remarks elements are repositioned after the Effects.]
Proposed resolution:
This wording is relative to N4810.
Modify 22.3.3 [pairs.spec] as indicated:
template<class T1, class T2> constexpr void swap(pair<T1, T2>& x, pair<T1, T2>& y) noexcept(noexcept(x.swap(y)));-7- Effects: As if by x.swap(y).
-8- Remarks: This functionshall not participate in overload resolutionis defined as deleted unless is_swappable_v<T1> is true and is_swappable_v<T2> is true.
Modify 22.4.12 [tuple.special] as indicated:
template <class... Types> constexpr void swap(tuple<Types...>& x, tuple<Types...>& y) noexcept(see below);-?- Effects: As if by x.swap(y).
-1- Remarks: This function
shall not participate in overload resolutionis defined as deleted unless is_swappable_v<Ti> is true for all i, where 0 <= i and i < sizeof...(Types). The expression inside noexcept is equivalent to:noexcept(x.swap(y))-2-
Effects: As if by x.swap(y).
Modify 22.5.9 [optional.specalg] as indicated:
template <class T> void swap(optional<T>& x, optional<T>& y) noexcept(noexcept(x.swap(y)));-1- Effects: Calls x.swap(y).
-2- Remarks: This functionshall not participate in overload resolutionis defined as deleted unless is_move_constructible_v<T> is true and is_swappable_v<T> is true.
Modify 22.6.10 [variant.specalg] as indicated:
template <class... Types> void swap(variant<Types...>& v, variant<Types...>& w) noexcept(see below);-1- Effects: Equivalent to v.swap(w).
-2- Remarks: This functionshall not participate in overload resolutionis defined as deleted unless is_move_constructible_v<Ti> && is_swappable_v<Ti> is true for all i. The expression inside noexcept is equivalent to noexcept(v.swap(w)).
Modify 20.3.1.6 [unique.ptr.special] as indicated:
template <class T, class D> void swap(unique_ptr<T, D>& x, unique_ptr<T, D>& y) noexcept;-?- Effects: Calls x.swap(y).
-1- Remarks: This function
shall not participate in overload resolutionis defined as deleted unless is_swappable_v<D> is true.-2-
Effects: Calls x.swap(y).
Modify 24.3.7.4 [array.special] as indicated:
template <class T, size_t N> void swap(array<T, N>& x, array<T, N>& y) noexcept(noexcept(x.swap(y)));-1-
-2- Effects: As if by x.swap(y). -3- Complexity: Linear in N. -?- Remarks: This function is defined as deleted unless N == 0 or is_swappable_v<T> is true.Constraints: N == 0 or is_swappable_v<T> is true.
Section: 22.5.3.2 [optional.ctor], 22.5.3.4 [optional.assign], 22.6.3.2 [variant.ctor], 22.7.4.2 [any.cons], 22.7.4.4 [any.modifiers] Status: New Submitter: Tim Song Opened: 2016-10-29 Last modified: 2020-06-13
Priority: 3
View all other issues in [optional.ctor].
View all issues with New status.
Discussion:
Throughout optional/variant/any's specification references are made to "the selected constructor of T". For example, 22.5.3.2 [optional.ctor]/16 says of the constructor from const T&:
-16- Remarks: If T's selected constructor is a constexpr constructor, this constructor shall be a constexpr constructor.
Similarly, the in-place constructor has this wording (22.5.3.2 [optional.ctor]/25-26):
-25- Throws: Any exception thrown by the selected constructor of T.
-26- Remarks: If T's constructor selected for the initialization is a constexpr constructor, this constructor shall be a constexpr constructor.
If T is a scalar type, it has no constructor at all. Moreover, even for class types, the in-place constructor wording ignores any implicit conversion done on the argument before the selected constructor is called, which 1) may not be valid in constant expressions and 2) may throw an exception; such exceptions aren't thrown "by the selected constructor of T" but outside it.
The wording should probably be recast to refer to the entire initialization.[Issues Telecon 16-Dec-2016]
Priority 3; Jonathan to provide wording.
[2020-06-11; Nina Dinka Ranns comments and provides initial wording]
This wording depends on the current resolution for LWG 2833, which covers the constexpr portion of this issue.
Proposed resolution:
This wording is relative to N4861.
Modify 22.5.3.2 [optional.ctor] as indicated:
constexpr optional(const optional& rhs);-3- […]
-4- […] -5- Throws: Any exception thrown by theselected constructor of Tinitialization of the contained value.constexpr optional(optional&& rhs) noexcept(see below);-7- […]
[…] -10- Throws: Any exception thrown by theselected constructor of Tinitialization of the contained value.template<class... Args> constexpr explicit optional(in_place_t, Args&&... args);-12- […]
[…] -15- Throws: Any exception thrown by theselected constructor of Tinitialization of the contained value.template<class U, class... Args> constexpr explicit optional(in_place_t, initializer_list<U> il, Args&&... args);-17- […]
[…] -20- Throws: Any exception thrown by theselected constructor of Tinitialization of the contained value.template<class U = T> constexpr explicit(see below) optional(U&& v);-22- […]
[…] -25- Throws: Any exception thrown by theselected constructor of Tinitialization of the contained value.template<class U> explicit(see below) optional(const optional<U>& rhs);-27- […]
[…] -30- Throws: Any exception thrown by theselected constructor of Tinitialization of the contained value.template<class U> explicit(see below) optional(optional<U>&& rhs);-32- […]
[…] -35- Throws: Any exception thrown by theselected constructor of Tinitialization of the contained value.
Modify 22.6.3.2 [variant.ctor] as indicated:
template<class T, class... Args> constexpr explicit variant(in_place_type_t<T>, Args&&... args);-20- […]
[…] -23- Throws: Any exception thrown bycalling the selected constructor of Tthe initialization of the contained value.template<class T, class U, class... Args> constexpr explicit variant(in_place_type_t<T>, initializer_list<U> il, Args&&... args);-25- […]
[…] -28- Throws: Any exception thrown bycalling the selected constructor of Tthe initialization of the contained value.template<size_t I, class... Args> constexpr explicit variant(in_place_index_t<I>, Args&&... args);-30 […]
[…] -33- Throws: Any exception thrown bycalling the selected constructor of TIthe initialization of the contained value.
Modify 22.7.4.2 [any.cons] as indicated:
any(const any& other);[…]-2- Effects: […]
-3- Throws: Any exceptionsarising from calling the selected constructor forthrown by the initialization of the contained value.template<class T> any(T&& value);-5- […]
[…] -9- Throws: Any exception thrown by theselected constructor of VTinitialization of the contained value.template<class T, class... Args> explicit any(in_place_type_t<T>, Args&&... args);-10- […]
[…] -15- Throws: Any exception thrown by theselected constructor of VTinitialization of the contained value.template<class T, class U, class... Args> explicit any(in_place_type_t<T>, initializer_list<U> il, Args&&... args);-16- […]
[…] -21- Throws: Any exception thrown by theselected constructor of VTinitialization of the contained value.
Modify 22.7.4.4 [any.modifiers] as indicated:
template<class T, class... Args> decay_t<T>& emplace(Args&&... args);-1- […]
[…] -7- Throws: Any exception thrown by theselected constructor of VTinitialization of the contained value. -8- […]template<class T, class U, class... Args> decay_t<T>& emplace(initializer_list<U> il, Args&&... args);-9- […]
[…] -15- Throws: Any exception thrown by theselected constructor of VTinitialization of the contained value.
Section: 17.6 [support.start.term] Status: New Submitter: Jean-François Bastien Opened: 2016-11-07 Last modified: 2020-09-06
Priority: 3
View other active issues in [support.start.term].
View all other issues in [support.start.term].
View all issues with New status.
Discussion:
While SG1 was processing NB comments CA1 and LATE2 regarding P0270R1, we decided to remove the proposed guarantee that quick_exit be made signal safe.
Our reasoning is that functions registered with at_quick_exit aren't forbidden from calling quick_exit, but quick_exit implementations likely acquire some form of a lock before processing all registered functions (because a note forbids the implementation from introducing data races). The following code can therefore deadlock:#include <cstdlib> int main() { std::at_quick_exit([] () { std::quick_exit(0); }); std::quick_exit(1); return 0; }
The same applies if a function registered in at_quick_exit handles a signal, and that signal calls quick_exit. SG1 believes that both issues (same thread deadlock, and signal deadlock) can be resolved in the same manner. Either:
Option 2. seems preferable, and can be implemented along the lines of:
#include <array> #include <atomic> #include <cstddef> namespace { typedef void (*func)(); std::array<func, 32> quick_exit_functions; const auto* quick_exit_functions_ptr = &quick_exit_functions; std::atomic_flag lock = ATOMIC_FLAG_INIT; struct scope { scope() { while (lock.test_and_set(std::memory_order_acquire)) ; } ~scope() { lock.clear(std::memory_order_release); } }; } namespace std { extern "C" void quick_exit(int status) noexcept { decltype(quick_exit_functions_ptr) f; { scope s; f = quick_exit_functions_ptr; quick_exit_functions_ptr = nullptr; } if (f) { size_t pos = f->size(); while (pos > 0) (*f)[--pos](); } _Exit(status); } extern "C++" int at_quick_exit(func f) noexcept { scope s; if (!quick_exit_functions_ptr || quick_exit_functions.size() == quick_exit_functions.max_size()) return -1; quick_exit_functions[quick_exit_functions.size()] = f; return 0; } }
Ideally, the resolution would also add back the wording which SG1 dropped from P0270R1:
Add at new element to the end of 17.6 [support.start.term] p13 (quick_exit()):
Remarks: The function quick_exit() is signal-safe (17.14.4 [csignal.syn]). [Note: It might still be unsafe to call quick_exit() from a handler, because the functions registered with at_quick_exit() might not be signal-safe. — end note]
[Issues Telecon 16-Dec-2016]
Priority 3
Proposed resolution:
This wording is relative to N4606.
Add at new element to the end of 17.6 [support.start.term] p13 (quick_exit()):
[[noreturn]] void quick_exit(int status) noexcept;-13- Effects: Functions registered by calls to at_quick_exit are called in the reverse order of their registration, except that a function shall be called after any previously registered functions that had already been called at the time it was registered. Objects shall not be destroyed as a result of calling quick_exit. If control leaves a registered function called by quick_exit because the function does not provide a handler for a thrown exception, std::terminate() shall be called. [Note: at_quick_exit may call a registered function from a different thread than the one that registered it, so registered functions should not rely on the identity of objects with thread storage duration. — end note] After calling registered functions, quick_exit shall call _Exit(status). [Note: The standard file buffers are not flushed. See: ISO C 7.22.4.5. — end note]
-?- Remarks: The function quick_exit() is signal-safe (17.14.4 [csignal.syn]). [Note: It might still be unsafe to call quick_exit() from a handler, because the functions registered with at_quick_exit() might not be signal-safe. — end note]
Section: 33.2.5 [thread.req.lockable], 33.6.4 [thread.mutex.requirements] Status: New Submitter: Agustín K-ballo Bergé Opened: 2016-11-12 Last modified: 2022-11-06
Priority: 3
View all issues with New status.
Discussion:
The current draft contains 14 occurrences of a Return type: clause. That clause is not covered by 16.3.2.4 [structure.specifications] p3. This was reported as editorial request #266.
[Issues Telecon 16-Dec-2016]
Priority 3; Jonathan to provide wording.
[12-May-2020, Jonathan provides wording to correct the 13 occurrences.]
Previous resolution [SUPERSEDED]:
This wording is relative to N4681.
Modify 33.2.5.3 [thread.req.lockable.req] as indicated:
-1- A type
L
meets the Cpp17Lockable requirements if it meets the Cpp17BasicLockable requirements and the following expressions are well-formed, have typebool
, and have the specified semantics (m
denotes a value of typeL
).
m.try_lock()
-2- Effects: [...]
-3- Return type:bool
.-4- Returns:
true
if the lock was acquired,false
otherwise.Modify 33.2.5.4 [thread.req.lockable.timed] as indicated:
-1- A type
L
meets the Cpp17TimedLockable requirements if it meets the Cpp17BasicLockable requirements and the following expressions are well-formed, have typebool
, and have the specified semantics (m
denotes a value of typeL
,rel_time
denotes a value of an instantiation ofduration
(29.5 [time.duration]), andabs_time
denotes a value of an instantiation oftime_point
(29.6 [time.point])).
m.try_lock_for(rel_time)
-2- Effects: [...]
-3- Return type:bool
.-4- Returns:
true
if the lock was acquired,false
otherwise.
m.try_lock_until(abs_time)
-2- Effects: [...]
-3- Return type:bool
.-4- Returns:
true
if the lock was acquired,false
otherwise.Modify 33.6.4.2 [thread.mutex.requirements.mutex] as indicated:
-6- The expression
m.lock()
is well-formed, has typevoid
, and has the following semantics:-7- Preconditions: [...]
-8- Effects: [...]
-9- Postconditions: [...]
-10- Return type:void
.-11- Synchronization: [...]
-12- Throws: [...]
-13- Error conditions: [...]
-14- The expression
m.try_lock()
is well-formed, has typebool
, and has the following semantics:-15- Preconditions: [...]
-16- Effects: [...]
-17- Return type:bool
.-18- Returns:
true
if ownership of the mutex was obtained for the calling thread, otherwisefalse
.-19- Synchronization: [...]
-20- Throws: Nothing.
-21- The expression
m.unlock()
is well-formed, has typevoid
, and has the following semantics:-22- Preconditions: [...]
-23- Effects: [...]
-24- Return type:void
.-25- Synchronization: [...]
-26- Throws: Nothing.
Modify 33.6.4.3 [thread.timedmutex.requirements] as indicated:
-1- The timed mutex types are the standard library types [...]
-2- The timed mutex types meet the Cpp17TimedLockable requirements (33.2.5.4 [thread.req.lockable.timed]).
-3- The expression
m.try_lock_for(rel_time)
is well-formed, has typebool
, and has the following semantics:-4- Preconditions: [...]
-5- Effects: [...]
-6- Return type:bool
.-7- Returns:
true
if the shared lock was acquired,false
otherwise.-8- Synchronization: [...]
-9- Throws: [...]
-10- The expression
m.try_lock_until(abs_time)
is well-formed, has typebool
, and has the following semantics:-11- Preconditions: [...]
-12- Effects: [...]
-13- Return type:bool
.-14- Returns:
true
if ownership was obtained, otherwisefalse
.-15- Synchronization: [...]
-16- Throws: [...]
Modify 33.6.4.4 [thread.sharedmutex.requirements] as indicated:
-1- The standard library types
shared_mutex
andshared_timed_mutex
are shared mutex types. [...]-2- In addition to the exclusive lock ownership mode [...]
-3- The expression
m.lock_shared()
is well-formed, has typevoid
, and has the following semantics:-4- Preconditions: [...]
-5- Effects: [...]
-6- Postconditions: [...]
-7- Return type:void
.-8- Synchronization: [...]
-9- Throws: [...]
-10- Error conditions: [...]
-11- The expression
m.unlock_shared()
is well-formed, has typevoid
, and has the following semantics:-12- Preconditions: [...]
-13- Effects: [...]
-14- Return type:void
.-15- Synchronization: [...]
-16- Throws: [...]
-17- The expression
m.try_lock_shared()
is well-formed, has typebool
, and has the following semantics:-18- Preconditions: [...]
-19- Effects: [...]
-20- Return type:bool
.-21- Returns:
true
if the shared ownership lock was acquired,false
otherwise.-22- Synchronization: [...]
-23- Throws: [...]
Modify 33.6.4.5 [thread.sharedtimedmutex.requirements] as indicated:
-1- The standard library type
shared_timed_mutex
is a shared timed mutex type. [...]-2- The expression
m.try_lock_shared_for(rel_time)
is well-formed, has typebool
, and has the following semantics:-3- Preconditions: [...]
-4- Effects: [...]
-5- Return type:bool
.-6- Returns:
true
if the shared lock was acquired,false
otherwise.-7- Synchronization: [...]
-8- Throws: [...]
-9- The expression
m.try_lock_shared_until(abs_time)
is well-formed, has typebool
, and has the following semantics:-10- Preconditions: [...]
-11- Effects: [...]
-12- Return type:bool
.-13- Returns:
true
if the shared lock was acquired,false
otherwise.-14- Synchronization: [...]
-15- Throws: [...]
[2022-11-06; Daniel comments and provides alternative wording]
Now that we have the new element Result: specified in 16.3.2.4 [structure.specifications], we can simply replace all occurrences of the Return type: by this element.
Proposed resolution:
This wording is relative to N4917.
Modify 33.2.5.3 [thread.req.lockable.req] as indicated:
-1- A type L meets the Cpp17Lockable requirements if it meets the Cpp17BasicLockable requirements and the following expressions are well-formed and have the specified semantics (m denotes a value of type L).
m.try_lock()
-2- Effects: [...]
-3-
Return typeResult: bool.-4- Returns: true if the lock was acquired, otherwise false.
Modify 33.2.5.4 [thread.req.lockable.timed] as indicated:
-1- A type L meets the Cpp17TimedLockable requirements if it meets the Cpp17Lockable requirements and the following expressions are well-formed and have the specified semantics (m denotes a value of type L, rel_time denotes a value of an instantiation of duration (29.5 [time.duration]), and abs_time denotes a value of an instantiation of time_point (29.6 [time.point])).
m.try_lock_for(rel_time)
-2- Effects: [...]
-3-
Return typeResult: bool.-4- Returns: true if the lock was acquired, otherwise false.
m.try_lock_until(abs_time)
-5- Effects: [...]
-6-
Return typeResult: bool.-7- Returns: true if the lock was acquired, otherwise false.
Modify 33.6.4.2.1 [thread.mutex.requirements.mutex.general] as indicated:
-5- The expression m.lock() is well-formed and has the following semantics:
-6- Preconditions: [...]
-7- Effects: [...]
-8- Synchronization: [...]
-9- Postconditions: [...]
-10-
Return typeResult: void.-11- Throws: [...]
-12- Error conditions: [...]
-13- The expression m.try_lock() is well-formed and has the following semantics:
-14- Preconditions: [...]
-15- Effects: [...]
-16- Synchronization: [...]
-17-
Return typeResult: bool.-18- Returns: true if ownership was obtained, otherwise false.
-19- Throws: Nothing.
-20- The expression m.unlock() is well-formed and has the following semantics:
-21- Preconditions: [...]
-22- Effects: [...]
-23-
Return typeResult: void.-24- Synchronization: [...]
-25- Throws: Nothing.
Modify 33.6.4.3.1 [thread.timedmutex.requirements.general] as indicated:
-1- The timed mutex types are the standard library types […]
-2- The expression m.try_lock_for(rel_time) is well-formed and has the following semantics:
-3- Preconditions: [...]
-4- Effects: [...]
-5- Synchronization: [...]
-6-
Return typeResult: bool.-7- Returns: true if ownership was obtained, otherwise false.
-8- Throws: [...]
-9- The expression m.try_lock_until(abs_time) is well-formed and has the following semantics:
-10- Preconditions: [...]
-11- Effects: [...]
-12- Synchronization: [...]
-13-
Return typeResult: bool.-14- Returns: true if ownership was obtained, otherwise false.
-15- Throws: [...]
Modify 33.6.4.4.1 [thread.sharedmutex.requirements.general] as indicated:
-1- The standard library types shared_mutex and shared_timed_mutex are shared mutex types. [...]
-2- In addition to the exclusive lock ownership mode [...]
-3- The expression m.lock_shared() is well-formed and has the following semantics:
-4- Preconditions: [...]
-5- Effects: [...]
-6- Synchronization: [...]
-7- Postconditions: [...]
-8-
Return typeResult: void.-9- Throws: [...]
-10- Error conditions: [...]
-11- The expression m.unlock_shared() is well-formed and has the following semantics:
-12- Preconditions: [...]
-13- Effects: [...]
-14-
Return typeResult: void.-15- Synchronization: [...]
-16- Throws: [...]
-17- The expression m.try_lock_shared() is well-formed and has the following semantics:
-18- Preconditions: [...]
-19- Effects: [...]
-20- Synchronization: [...]
-21-
Return typeResult: bool.-22- Returns: true if the shared lock was acquired, otherwise false.
-23- Throws: [...]
Modify 33.6.4.5.1 [thread.sharedtimedmutex.requirements.general] as indicated:
-1- The standard library type shared_timed_mutex is a shared timed mutex type. [...]
-2- The expression m.try_lock_shared_for(rel_time) is well-formed and has the following semantics:
-3- Preconditions: [...]
-4- Effects: [...]
-5- Synchronization: [...]
-6-
Return typeResult: bool.-7- Returns: true if the shared lock was acquired, otherwise false.
-8- Throws: [...]
-9- The expression m.try_lock_shared_until(abs_time) is well-formed and has the following semantics:
-10- Preconditions: [...]
-11- Effects: [...]
-12- Synchronization: [...]
-13-
Return typeResult: bool.-14- Returns: true if the shared lock was acquired, otherwise false.
-15- Throws: [...]
Section: 24.3.7.1 [array.overview] Status: Open Submitter: Robert Haberlach Opened: 2016-11-16 Last modified: 2018-03-19
Priority: 3
View other active issues in [array.overview].
View all other issues in [array.overview].
View all issues with Open status.
Discussion:
LWG 2590's resolution is incomplete:
std::array<int, 1> arr{{0}};
should be fine, but isn't guaranteed, since {0} has no type. We should rather go for implicit conversion:
An array is an aggregate (9.4.2 [dcl.init.aggr]) that can be list-initialized with up to N elements
whose types are convertible to Tthat can be implicitly converted to T.
[2016-11-26, Tim Song comments]
This is not possible as written, because due to the brace elision rules for aggregate initialization, std::array<int, 2> arr{{0}, {1}}; will never work: the {0} is taken as initializing the inner array, and the {1} causes an error.
[2017-01-27 Telecon]
Priority 2; consensus is that the P/R is not quite right.
[2018-3-14 Wednesday evening issues processing; priority to 3; move to Open]
Jens: There's nothing you can do about the double braces in std::array. That's a core thing.
STL to write paper to resolve this.
Proposed resolution:
This wording is relative to N4606.
Change 24.3.7.1 [array.overview] p2 as indicated:
-2- An array is an aggregate (9.4.2 [dcl.init.aggr]) that can be list-initialized with up to N elements
whose types are convertiblethat can be implicitly converted to T.
Section: 21.3.5.4 [meta.unary.prop] Status: New Submitter: Richard Smith Opened: 2016-11-17 Last modified: 2020-01-25
Priority: 3
View other active issues in [meta.unary.prop].
View all other issues in [meta.unary.prop].
View all issues with New status.
Discussion:
struct S { ~S(); // non-trivial }; static_assert(std::is_trivially_constructible<S>::value, "");
Should the assert pass? Implementations disagree.
Per 21.3.5.4 [meta.unary.prop]'s Table 38, this trait looks at whether the following variable definition is known to call no operation that is not trivial:S t(create<Args>()...);
... where Args is an empty pack in this case. That variable definition results in a call to the S destructor. Should that call be considered by the trait?
[2017-01-27 Telecon]
Priority 3
This issue interacts with 2116
[2020-01-24; Peter Dimov comments]
std::is_trivially_copy_constructible_v<D>, where D is
struct D { ~D() {} };
reports false. This is because the definition of is_trivially_copy_constructible requires the invented variable definition T t(declval<Args>()...);, which in our case is D t(declval<D>());, to not call any nontrivial operations.
This is interpreted by implementations to include the destructor call, presumably for consistency with is_nothrow_copy_constructible. But that's wrong; the copy constructor is trivial. As a consequence, variant<D> also doesn't have a trivial copy constructor, which causes (completely unnecessary) inefficiencies when said variant is copied.Proposed resolution:
Section: 22.5.3.6 [optional.observe] Status: Open Submitter: Richard Smith Opened: 2016-11-24 Last modified: 2020-06-13
Priority: 2
View other active issues in [optional.observe].
View all other issues in [optional.observe].
View all issues with Open status.
Discussion:
After applying LWG 2740, we have:
constexpr const T* operator->() const; constexpr T* operator->();-1- Requires: *this contains a value.
-2- Returns: val. -3- Throws: Nothing. -4- Remarks: These functions shall be constexpr functions.
Paragraph 4 is completely superfluous. We already said these functions were constexpr in the synopsis. Can it be removed?
[Issues Telecon 16-Dec-2016]
Priority 2
Jonathan notes: Although Richard is correct, I suggest we don't strike the paragraph, so that we remember to fix it as part of 2833, when we know how to say this properly.
[2018-06 Rapperswil Thursday issues processing]
Status to Open; also see 7.7 [expr.const]/6 and 2289.
[2020-06-08 Nina Dinka Ranns comments]
The revised wording provided by LWG 2833 should resolve this issue as well.
Proposed resolution:
Section: 22.6.3.2 [variant.ctor] Status: Open Submitter: Richard Smith Opened: 2016-11-28 Last modified: 2020-12-19
Priority: 2
View other active issues in [variant.ctor].
View all other issues in [variant.ctor].
View all issues with Open status.
Discussion:
The library has lots of functions declared constexpr, but it's not clear what that means. The constexpr keyword implies that there needs to be some invocation of the function, for some set of template arguments and function arguments, that is valid in a constant expression (otherwise the program would be ill-formed, with no diagnostic required), along with a few side conditions. I suspect the library intends to require something a lot stronger than that from implementations (something along the lines of "all calls that could reasonably be constant subexpressions are in fact constant subexpressions, unless otherwise stated").
[variant.ctor]/1 contains this, which should also be fixed:"This function shall be constexpr if and only if the value-initialization of the alternative type T0 would satisfy the requirements for a constexpr function."
This is the wrong constraint: instead of constraining whether the function is constexpr, we should constrain whether a call to it is a constant subexpression.
Daniel:
This is has some considerable overlap with LWG 2289 but is phrased in a more general way.[2016-12-16, Issues Telecon]
Priority 2; this is also the general case of 2829.
[2017-02-20, Alisdair comments and suggests concrete wording]
Below is is draft wording I was working on at Issaquah to try to address both issues.
[2017-11 Albuquerque Wednesday issue processing]
Status to Open; really needs a paper.
STL says "What about plus<T>?" plus<int> needs to be usable in a constexpr context, but plus<string> can't be.
[2017-11 Albuquerque Saturday issues processing]
Geoffrey to write a paper resolving this.
[2018-06 Rapperswil Thursday issues processing]
Geoffrey has been unable to write this paper due to time constraints. He wrote up his progress here. Daniel has offered to help someone to write this paper; he's willing to be a co-author.
[2018-08-23 Batavia Issues processing]
Michael Wong to investigate.
Previous resolution from Daniel [SUPERSEDED]:This wording is relative to N4640.
Modify 16.4.6.7 [constexpr.functions] as indicated:
17.6.5.6 constexpr functions and constructors [constexpr.functions]
-1- This International Standard explicitly requires that certain standard library functions are constexpr (9.2.6 [dcl.constexpr]). If the specification for a templated entity requires that it shall be a constexpr templated entity, then that templated entity shall be usable in a constant expression.. An implementation
shall notmay declareanyadditional standard library function signature as constexprexcept for those where it is explicitly required. Within any header that provides any non-defining declarations of constexpr functions or constructors an implementation shall provide corresponding definitions.
[2020-06-08 Nina Dinka Ranns comments and provides alternative wording]
The revised wording draft also resolves LWG 2289, LWG 2829, and LWG 3215.
Previous resolution [SUPERSEDED]:This wording is relative to N4861.
1. Modify 16.4.6.7 [constexpr.functions] as indicated:
-1- This document explicitly requires that certain standard library functions are constexpr (9.2.6 [dcl.constexpr]). An implementation shall not declare any standard library function signature as constexpr except for those where it is explicitly required. Within any header that provides any non-defining declarations of constexpr functions or constructors an implementation shall provide corresponding definitions.
-?- Let F denote a standard library function template or member function of a class template. If the specification of F declares it to be constexpr, unless otherwise specified, then F can be used in a constant expression if and only if all the expressions that are evaluated as specified in the description of F's semantics can be used in a constant expression.2. - 10. […] // Remainder of Nina's update
[2020-10-02 Jens Maurer improves wording]
Specifically the wording for 16.4.6.7 [constexpr.functions] needs improvement and is updated below.
[2020-10-02 Tim Song comments]
The new wording doesn't cover the following example:
// global scope int x; int y; constexpr int j = (std::swap(x, y), 0); // error
swap is a "standard library function template...declared constexpr"
x and y are (lvalue) constant expressions
std::swap(x, y) is plainly not a constant expression
[2020-10-04 Jens Maurer comments]
Yes, we're still lacking text for that (and maybe Nina's old text helps for that).
[2020-12-14; Jiang An comments]
The item "constexpr functions" is also used in 24.2.2.1 [container.requirements.general]/14 and 25.3.1 [iterator.requirements.general]/16, and such usage should also be modified by this issue here.
Proposed resolution:
This wording is relative to N4861.
Modify 16.4.6.7 [constexpr.functions] as indicated:
-1- This document explicitly requires that certain standard library functions are constexpr (9.2.6 [dcl.constexpr]). An implementation shall not declare any standard library function signature as constexpr except for those where it is explicitly required. Within any header that provides any non-defining declarations of constexpr functions or constructors an implementation shall provide corresponding definitions.
-?- Let F denote a standard library function template or member function of a class template declared constexpr. Unless otherwise specified, a function call expression (7.6.1.3 [expr.call]) whose postfix-expression names F is a constant expression if all of the argument subexpressions are constant expressions.
Modify 22.3.2 [pairs.pair] as indicated:
-2- The defaulted
move and copyconstructors, respectively,of pairis a constexpr functioncan be used in a constant expression if and only if all required element-wise initializationsfor move and copy, respectively, would satisfy the requirements for a constexpr functioncan be used in a constant expression.
Modify 22.4.4.1 [tuple.cnstr] as indicated:
-3- The defaulted
move and copyconstructors, respectively,of tupleis a constexpr functioncan be used in a constant expression if and only if all required element-wise initializationsfor move and copy, respectively, would satisfy the requirements for a constexpr functioncan be used in a constant expression. The defaultedmove and copyconstructors of tuple<>are constexpr functionscan be used in a constant expression.
Modify 22.5.3.2 [optional.ctor] as indicated:
constexpr optional() noexcept; constexpr optional(nullopt_t) noexcept;[…]-1- […]
-2- Remarks: No contained value is initialized.For every object type T these constructors are constexpr constructors (9.2.6 [dcl.constexpr]).template<class... Args> constexpr explicit optional(in_place_t, Args&&... args);-12- […]
-13- […] -14- […] -15- […]-16- Remarks: If T's constructor selected for the initialization is a constexpr constructor, this constructor is a constexpr constructor.template<class U, class... Args> constexpr explicit optional(in_place_t, initializer_list<U> il, Args&&... args);-17- […]
-18- […] -19- […] -20- […]-21- Remarks: If T's constructor selected for the initialization is a constexpr constructor, this constructor is a constexpr constructor.template<class U = T> constexpr explicit(see below) optional(U&& v);-22- […]
-23- […] -24- […] -25- […] -26- Remarks:If T's constructor selected for the initialization is a constexpr constructor, this constructor is a constexpr constructor.The expression inside explicit is equivalent to:!is_convertible_v<U, T>
Modify 22.5.3.6 [optional.observe] as indicated:
constexpr const T* operator->() const; constexpr T* operator->();-1- […]
-2- […] -3- […]-4- Remarks: These functions are constexpr functions.constexpr const T& operator*() const&; constexpr T& operator*() &;[…]-5- […]
-6- […] -7- […]-8- Remarks: These functions are constexpr functions.constexpr explicit operator bool() const noexcept;-11- Returns: true if and only if *this contains a value.
-12- Remarks: This function is a constexpr function.constexpr bool has_value() const noexcept;-13- Returns: true if and only if *this contains a value.
-14- Remarks: This function is a constexpr function.
Modify 22.5.6 [optional.relops] as indicated:
template<class T, class U> constexpr bool operator==(const optional<T>& x, const optional<U>& y);-1- […]
-2- […]-3- Remarks: Specializations of this function template for which *x == *y is a core constant expression are constexpr functions.template<class T, class U> constexpr bool operator!=(const optional<T>& x, const optional<U>& y);-4- […]
-5- […]-6- Remarks: Specializations of this function template for which *x != *y is a core constant expression are constexpr functions.template<class T, class U> constexpr bool operator<(const optional<T>& x, const optional<U>& y);-7- […]
-8- […]-9- Remarks: Specializations of this function template for which *x < *y is a core constant expression are constexpr functions.template<class T, class U> constexpr bool operator>(const optional<T>& x, const optional<U>& y);-10- […]
-11- […]-12- Remarks: Specializations of this function template for which *x > *y is a core constant expression are constexpr functions.template<class T, class U> constexpr bool operator<=(const optional<T>& x, const optional<U>& y);-13- […]
-14- […]-15- Remarks: Specializations of this function template for which *x <= *y is a core constant expression are constexpr functions.template<class T, class U> constexpr bool operator>=(const optional<T>& x, const optional<U>& y);-16- […]
-17- […]-18- Remarks: Specializations of this function template for which *x >= *y is a core constant expression are constexpr functions.template<class T, three_way_comparable_with<T> U> constexpr compare_three_way_result_t<T,U> operator<=>(const optional<T>& x, const optional<U>& y);-19- Returns: If x && y, *x <=> *y; otherwise bool(x) <=> bool(y).
-20- Remarks: Specializations of this function template for which *x <=> *y is a core constant expression are constexpr functions.
Modify 22.6.3.2 [variant.ctor] as indicated:
constexpr variant() noexcept(see below);[…]-1- […]
-2- […] -3- […] -4- […] -5- […] -6- Remarks:This function is constexpr if and only if the value-initialization of the alternative type T0 would satisfy the requirements for a constexpr function.The expression inside noexcept is equivalent to is_nothrow_default_constructible_v<T0>. [Note: See also class monostate. — end note]template<class T> constexpr variant(T&& t) noexcept(see below);-14- […]
[…] -19- Remarks: The expression inside noexcept is equivalent to is_nothrow_constructible_v<Tj, T>.If Tj's selected constructor is a constexpr constructor, this constructor is a constexpr constructor.template<class T, class... Args> constexpr explicit variant(in_place_type_t<T>, Args&&... args);-20- […]
[…]-24- Remarks: If T's selected constructor is a constexpr constructor, this constructor is a constexpr constructor.template<class T, class U, class... Args> constexpr explicit variant(in_place_type_t<T>, initializer_list<U> il, Args&&... args);-25- […]
[…]-29- Remarks: If T's selected constructor is a constexpr constructor, this constructor is a constexpr constructor.template<size_t I, class... Args> constexpr explicit variant(in_place_index_t<I>, Args&&... args);-30- […]
[…]-34- Remarks: If TI's selected constructor is a constexpr constructor, this constructor is a constexpr constructor.template<size_t I, class U, class... Args> constexpr explicit variant(in_place_index_t<I>, initializer_list<U> il, Args&&... args);-35- […]
[…]-38- Remarks: If TI's selected constructor is a constexpr constructor, this constructor is a constexpr constructor.
Modify 25.5.4.11 [move.sent.ops] as indicated:
constexpr move_sentinel();-1- Effects: Value-initializes last.
If is_trivially_default_constructible_v<S> is true, then this constructor is a constexpr constructor.
Modify 22.15.3 [bit.cast] as indicated:
template<class To, class From> constexpr To bit_cast(const From& from) noexcept;-1- […]
-3- Remarks: This functionis constexprcan be used in a constant expression if and only if To, From, and the types of all subobjects of To and From are types T such that:
(3.1) — is_union_v<T> is false;
(3.2) — is_pointer_v<T> is false;
(3.3) — is_member_pointer_v<T> is false;
(3.4) — is_volatile_v<T> is false; and
(3.5) — T has no non-static data members of reference type.
Modify 29.5 [time.duration] as indicated:
-5- The defaulted
copyconstructors of durationshall be a constexpr functioncan be used in a constant expression if and only if the required initialization of the member rep_for copy and move, respectively, would satisfy the requirements for a constexpr functioncan be used in a constant expression.
Section: 24.2.7 [associative.reqmts], 24.2.8 [unord.req] Status: Open Submitter: Matt Austern Opened: 2016-12-14 Last modified: 2020-02-14
Priority: 3
View other active issues in [associative.reqmts].
View all other issues in [associative.reqmts].
View all issues with Open status.
Discussion:
If we write a_uniq.insert(i, j) and [i, j) has multiple elements with keys that compare equivalent, which ones get inserted? Consider, for example, inserting into a map<string, int> with
m.insert({{"red", 5}, {"green", 3}, {"red", 7}, {"blue", 2}, {"pink", 6}});
Which value for "red" will the map have?
On my implementation we got "red" -> 5, and I suspect that's true on most or all implementations, but I don't believe that's guaranteed by anything in the requirements. The wording in Table 90 just says that it "inserts each element from the range [i, j) if and only if there is no element with key equivalent to the key of that element", but that doesn't tell us what happens if [i, j) contains duplicate keys because it doesn't say what order the insertions are performed in. The standard should either guarantee that the first value is the one that gets inserted, or explicitly say that this is unspecified. The same issue applies to the range constructor, and to the unordered associative containers.[2017-01-27 Telecon]
Priority 3; Nico to provide wording.
[2020-02-14, Prague]
LWG discussion. Suggestion to specify that we like the direction of the wording for insert of unordered containers, but would also like to clarify that the loop is meant to be "in order" of the sequence elements.
Daniel K. volunteered to provide such wording.Proposed resolution:
Section: 21.3.2 [meta.rqmts] Status: New Submitter: Tim Song Opened: 2016-12-14 Last modified: 2017-02-02
Priority: 3
View all other issues in [meta.rqmts].
View all issues with New status.
Discussion:
[meta.rqmts]/3 defines TransformationTrait as follows:
A TransformationTrait modifies a property of a type. It shall be a class template that takes one template type argument and, optionally, additional arguments that help define the modification. It shall define a publicly accessible nested type named type, which shall be a synonym for the modified type.
enable_if, result_of and common_type do not necessarily "define a publicly accessible nested type named type". aligned_storage takes no template type argument (it only has two non-type parameters). Yet [meta.trans]/2 says that they are all TransformationTraits.
Incidentally, with the exception of decay, it's not clear that any of the traits in [meta.trans.other] could really be described as "modify[ing] a property of a type".[2017-01-27 Telecon]
Priority 3
Proposed resolution:
Section: 28.4.9 [cmplx.over], 28.7.1 [cmath.syn] Status: New Submitter: Jens Maurer Opened: 2016-12-15 Last modified: 2019-03-18
Priority: 3
View all other issues in [cmplx.over].
View all issues with New status.
Discussion:
In [cmplx.over] and [cmath.syn], when talking about "sufficient additional overloads", we use the phrase "effectively cast", but that is not a defined term.
A hostile interpretation could read "reinterpret_cast" here. Likely we mean "apply floating-point promotions, floating-integral conversions, and floating-point conversions", but that should be spelled out somewhere, e.g. in the library definitions section. (Source: Editorial issue #1248)[2017-01-27 Telecon]
Priority 3
[2019-03-16; Daniel comments and provides wording]
I decided to use the form "implicitly converted" and to refer to 7.3 [conv] for 28.7.1 [cmath.syn] and 28.4.9 [cmplx.over], because those conversions can all be done implicitly. This also holds for the pow specification 28.4.9 [cmplx.over] p3, because the described conversions of complex<T> to complex<U> involve only the need of non-explicit constructors.
Proposed resolution:
This wording is relative to N4810.
Change 28.4.9 [cmplx.over], as indicated:
-2- The additional overloads shall be sufficient to ensure:
-3 Function template pow shall have additional overloads sufficient to ensure, for a call with at least one argument of type complex<T>:
(2.1) — If the argument has type long double, then it is
effectively castimplicitly converted (7.3 [conv]) to complex<long double>.(2.2) — Otherwise, if the argument has type double or an integer type, then it is
effectively castimplicitly converted to complex<double>.(2.3) — Otherwise, if the argument has type float, then it is
effectively castimplicitly converted to complex<float>.
(3.1) — If either argument has type complex<long double> or type long double, then both arguments are
effectively castimplicitly converted (7.3 [conv]) to complex<long double>.(3.2) — Otherwise, if either argument has type complex<double>, double, or an integer type, then both arguments are
effectively castimplicitly converted to complex<double>.(3.3) — Otherwise, if either argument has type complex<float> or float, then both arguments are
effectively castimplicitly converted to complex<float>.
Change 28.7.1 [cmath.syn], as indicated:
-2- For each set of overloaded functions within <cmath>, with the exception of abs, there shall be additional overloads sufficient to ensure:
If any argument of arithmetic type corresponding to a double parameter has type long double, then all arguments of arithmetic type (6.8.2 [basic.fundamental]) corresponding to double parameters are
effectively castimplicitly converted (7.3 [conv]) to long double.Otherwise, if any argument of arithmetic type corresponding to a double parameter has type double or an integer type, then all arguments of arithmetic type corresponding to double parameters are
effectively castimplicitly converted to double.Otherwise, all arguments of arithmetic type corresponding to double parameters have type float.
Section: 28.7.1 [cmath.syn] Status: New Submitter: Jens Maurer Opened: 2016-12-15 Last modified: 2020-09-06
Priority: 3
View other active issues in [cmath.syn].
View all other issues in [cmath.syn].
View all issues with New status.
Discussion:
With P0175R1, we now show in [cmath.syn] three overloads for the sin function: One taking a float, one taking a double, and one taking a long double. However, there is no statement that sin(long double) should actually invoke sinl, presumably delivering extra precision.
An implementation likeinline long double sin(long double x) { return sinf(x); }
seems to satisfy the "effectively cast" requirement, but is certainly unintentional.
The same issue arises for all math functions inherited from C. (Source: Editorial issue #1247)[2017-01-27 Telecon]
Priority 3
Proposed resolution:
Section: 20.4.5.2 [mem.res.pool.options] Status: New Submitter: Jens Maurer Opened: 2016-12-15 Last modified: 2017-02-02
Priority: 3
View all issues with New status.
Discussion:
20.4.5.2 [mem.res.pool.options] p3 talks about a "pass-through-threshold".
First, the phrase is not defined and it seems it could be easily avoided given the context. Second, given the phrasing here, it seems the implementation is essentially allowed to ignore the value largest_required_pool_block as it sees fit. It is unclear whether that is the intention.[2017-01-27 Telecon]
Priority 3; Jonathan will ask Alisdair for wording.
Proposed resolution:
Section: 25.5.1 [reverse.iterators] Status: New Submitter: Hubert Tong Opened: 2017-01-28 Last modified: 2020-09-06
Priority: 4
View all other issues in [reverse.iterators].
View all issues with New status.
Discussion:
Further to LWG 2472, the case of reverse_iterator comparisons is a regression introduced by LWG 280.
Consider the following program:#include <utility> #include <iterator> using namespace std::rel_ops; bool f(std::reverse_iterator<int *> it) { return it != it; }
Under C++03, the operator!= in lib.reverse.iterator is more specialized than the operator!= in std::rel_ops.
Following LWG 280, neither operator!= candidate is more specialized than the other. The program is observed to fail with libc++. Online compiler example, see here. Suggested resolution: Reintroduce the homogeneous comparison operators from C++03 alongside the new ones.[2017-03-04, Kona]
Set priority to 4. STL to write a paper deprecating relops Alisdair to provide an example for Annex C.
Proposed resolution:
This wording is relative to N4618.
Modify 25.5.1 [reverse.iterators], class template reverse_iterator synopsis, as indicated:
template <class Iterator1, class Iterator2> constexpr bool operator==( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator1, class Iterator2> constexpr bool operator<( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator1, class Iterator2> constexpr bool operator!=( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator1, class Iterator2> constexpr bool operator>( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator1, class Iterator2> constexpr bool operator>=( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator1, class Iterator2> constexpr bool operator<=( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator> constexpr bool operator==( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y); template <class Iterator> constexpr bool operator<( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y); template <class Iterator> constexpr bool operator!=( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y); template <class Iterator> constexpr bool operator>( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y); template <class Iterator> constexpr bool operator>=( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y); template <class Iterator> constexpr bool operator<=( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y);
Modify [reverse.iter.op==] as indicated:
template <class Iterator1, class Iterator2> constexpr bool operator==( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator> constexpr bool operator==( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y);-1- Returns: x.current == y.current.
Modify [reverse.iter.op<] as indicated:
template <class Iterator1, class Iterator2> constexpr bool operator<( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator> constexpr bool operator<( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y);-1- Returns: x.current > y.current.
Modify [reverse.iter.op!=] as indicated:
template <class Iterator1, class Iterator2> constexpr bool operator!=( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator> constexpr bool operator!=( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y);-1- Returns: x.current != y.current.
Modify [reverse.iter.op>] as indicated:
template <class Iterator1, class Iterator2> constexpr bool operator>( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator> constexpr bool operator>( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y);-1- Returns: x.current < y.current.
Modify [reverse.iter.op>=] as indicated:
template <class Iterator1, class Iterator2> constexpr bool operator>=( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator> constexpr bool operator>=( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y);-1- Returns: x.current <= y.current.
Modify [reverse.iter.op<=] as indicated:
template <class Iterator1, class Iterator2> constexpr bool operator<=( const reverse_iterator<Iterator1>& x, const reverse_iterator<Iterator2>& y); template <class Iterator> constexpr bool operator<=( const reverse_iterator<Iterator>& x, const reverse_iterator<Iterator>& y);-1- Returns: x.current >= y.current.
Section: 22.6.3 [variant.variant] Status: New Submitter: Switzerland Opened: 2017-02-03 Last modified: 2020-09-06
Priority: 3
View all other issues in [variant.variant].
View all issues with New status.
Discussion:
Addresses CH 7Consider making the variant statically !valueless_by_exception() for cases where is_nothrow_move_constructible_v<T_i> for all alternative types Ti
Proposed change: Adopt section III of P0308R0.
[2017-07 Toronto Thurs Issue Prioritization]
Priority 3. This is similar to 2904, Casey to investigate
Proposed resolution:
Section: 23.3 [string.view] Status: LEWG Submitter: Switzerland Opened: 2017-02-03 Last modified: 2017-07-16
Priority: Not Prioritized
View all other issues in [string.view].
View all issues with LEWG status.
Discussion:
Addresses CH 9The standard library should provide string_view parameters instead or in addition for functions defined with char const * or string const & as parameter types. Most notably in cases where both such overloads exist or where an internal copy is expected anyway.
It might be doubted that the non-null termination of string_view could be an issue with functions that pass the char * down to OS functions, such as fstream_buf::open() etc. and those shouldn't provide it and favour generating a std::string temporary instead in that case. However, std::path demonstrates it is usable to have string_view overloads and there might be many places where it can be handy, or even better.
Proposed change: Provide the overloads for std::regex, the exception classes, std::bitset, std::locale and more.
[Post-Kona 2017]
Most (all?) of these changes were proposed in P0506r1, which was discussed by LEWG in Kona.
[2017-07 Toronto Thurs Issue Prioritization]
Status LEWG - they're already looking at this.
Proposed resolution:
Section: 24 [containers], 22 [utilities] Status: LEWG Submitter: Finland Opened: 2017-02-03 Last modified: 2017-06-27
Priority: Not Prioritized
View other active issues in [containers].
View all other issues in [containers].
View all issues with LEWG status.
Discussion:
Addresses FI 16Relational operators for containers should sfinae; if the underlying type is not comparable, neither should the container be. Same applies to tuple and pair.
Proposed change: Make the relational operators of containers and utility components reflect the validity of the underlying element types.
[ 2017-06-27 Moved to LEWG after 5 positive votes on c++std-lib. ]
Proposed resolution:
Section: 24 [containers], 22 [utilities] Status: LEWG Submitter: Finland Opened: 2017-02-03 Last modified: 2017-07-16
Priority: Not Prioritized
View other active issues in [containers].
View all other issues in [containers].
View all issues with LEWG status.
Discussion:
Addresses FI 17The relational operators of optional and variant completely reflect the semantics of the element types; this is inconsistent with other types in the library, like pair, tuple and containers. If we believe it's important that we don't synthesize relational operators for wrapper types, we should believe it's important for other types as well. Otherwise comparing containers of floating-point types and tuples/pairs etc. of floating point types will give incorrect answers.
Proposed change: Make the relational operators of containers and utility components reflect the semantics of the operators for the underlying element types.
[2017-07 Toronto Thurs Issue Prioritization]
Move to LEWG
Proposed resolution:
Section: 20.3.2.2.2 [util.smartptr.shared.const] Status: New Submitter: United States Opened: 2017-02-03 Last modified: 2017-07-16
Priority: 3
View all other issues in [util.smartptr.shared.const].
View all issues with New status.
Discussion:
Addresses US 130There is no ability to supply an allocator for the control block when constructing a shared_ptr from a unique_ptr. Note that no further shared_ptr constructors need an allocator, as they all have pre-existing control blocks that are shared, or already have the allocator overload.
Proposed change: Add an additional shared_ptr constructor, template <class Y, class D, class A> shared_ptr(unique_ptr<Y, D>&& r, A alloc), with the same semantics as the existing constructor taking a unique_ptr, but using the alloc argument to supply memory as required.
[2017-07 Toronto Thurs Issue Prioritization]
Priority 3; Alisdair to provide wording
Proposed resolution:
Section: 21.3.3 [meta.type.synop] Status: LEWG Submitter: United States Opened: 2017-02-03 Last modified: 2017-07-16
Priority: Not Prioritized
View other active issues in [meta.type.synop].
View all other issues in [meta.type.synop].
View all issues with LEWG status.
Discussion:
Addresses US 171The *_constant<> templates (including the proposed addition, bool_constant<>) do not make use of the new template<auto> feature.
Proposed change: Add a constant<> (subject to bikeshedding) template which uses template<auto>.
Define integral_constant<> as using integral_constant<T, V> = constant<T(V)> or integral_constant<T, V> = constant<V>.
Either remove bool_constant, define it as using bool_constant = constant<bool(B)> or using bool_constant = constant<B>.
[2017-03-03, Kona, LEWG]
Straw polls:
constant | 3 |
numeric_constant | 8 |
static_constant | 1 |
scalar_constant | 7 |
integer_constant (Over LWG's dead body) | 1 |
auto_constant | 4 |
integral_c | 7 |
int_ | 0 |
| |
scalar_constant | 6 |
numeric_constant | 3 |
integral_c | 5 |
Accept P0377 with "scalar_constant" for C++17 to address LWG 2922 and US 171:
SF | F | N | A | SA 0 | 1 | 3 | 7 | 5[2017-07 Toronto Thurs Issue Prioritization]
Status LEWG with P0377
Proposed resolution:
Section: 28.7.1 [cmath.syn] Status: New Submitter: United States Opened: 2017-02-03 Last modified: 2017-07-16
Priority: 4
View other active issues in [cmath.syn].
View all other issues in [cmath.syn].
View all issues with New status.
Discussion:
Addresses US 172noexcept is inconsistently applied across headers which import components of the C standard library into the C++ library; some functions (std::abort(), std::_Exit(), etc) are defined as noexcept in some places, but not in others. Some functions which seem like they should be noexcept (std::abs(), std::div(), etc) are not defined as noexcept.
Proposed change: Make the majority of the C library functions (with exceptions such as std::qsort() and std::bsearch(), which can call user code) noexcept.
[2017-07 Toronto Thurs Issue Prioritization]
Priority 4
Proposed resolution:
Section: 25.4.3 [iterator.operations] Status: Open Submitter: Morwenn Opened: 2017-02-04 Last modified: 2018-12-03
Priority: 3
View other active issues in [iterator.operations].
View all other issues in [iterator.operations].
View all issues with Open status.
Discussion:
It seems that std::next is missing an optimization opportunity when taking a single parameter. The standard mandates that std::next shall call std::advance on the passed iterator and return it. For random-access iterators, it means that operator+= will be called on the iterator. However, if a single-argument overload was added to std::next, it could call ++it directly instead of std::advance(it, 1), which means that operator++ would be called instead of operator+=. This might make a small performance difference for complicated iterators such as std::deque's ones, where operator++ has a simpler logic than operator+=.
An equivalent optimization could be allowed by adding a single-argument overload to std::prev too.[2017-03-04, Kona]
Set priority to 3. Alisdair to provide wording.
[2018-11-30, Jonathan comments, recommending NAD]
Jonathan suggested NAD, because the proposed "just use increment when n==1" optimization can be done in std::next (and/or std::advance, and/or complicated iterators like deque::iterator) without adding an overload. Billy said the overload would avoid metaprogramming costs for dispatching to the right std::advance, and help in non-optimized builds. Zhihao said the overload would make it clear to users that the n==1 case is optimized (Jonathan thinks this is irrelevant as there's no requirement that we tell users what we optimize).
Proposed resolution:
Section: 22.4.5 [tuple.creation] Status: New Submitter: Eric Fiselier Opened: 2017-02-06 Last modified: 2020-09-06
Priority: 3
View all other issues in [tuple.creation].
View all issues with New status.
Discussion:
The current PR for LWG 2773 changes std::ignore to be a constexpr variable. However it says nothing about whether using std::ignore in std::tie is a constant expression. I think the intent was clearly to allow this. Therefore I suggest we update the resolution to explicitly call this out in a note. (I don't think new normative wording is needed).
I propose we update the current PR as follows:Keep the current changes proposed by the PR.
Add a note after [tuple.creation]/p7 (std::tie):
[Note: The constructors and assignment operators provided by ignore shall be constexpr]
Perhaps LWG feels the existing wording is clear enough, but if not I think the above changes sufficiently clarify it.
The ability to constexpr assign to std::ignore can be important: Here is an extremely contrived example:constexpr bool foo() { auto res = std::tie(std::ignore); std::get<0>(res) =42; return true; } static_assert(foo());
[2017-03-04, Kona]
Set priority to 3. P/R is incorrect; it should NOT be a note. Marshall to work with Eric to get better wording. STL says "use an exposition-only class".
Proposed resolution:
This wording is relative to N4640.
Modify 22.4.5 [tuple.creation] as indicated:
template<class... TTypes> constexpr tuple<TTypes&...> tie(TTypes&... t) noexcept;-7- Returns: […]
-?- [Note: The constructors and assignment operators provided by ignore shall be constexpr. — end note] -8- [Example: […] — end example]
Section: 21.3.3 [meta.type.synop] Status: New Submitter: Daniel Krügler Opened: 2017-03-02 Last modified: 2020-02-13
Priority: 2
View other active issues in [meta.type.synop].
View all other issues in [meta.type.synop].
View all issues with New status.
Discussion:
LWG 2797 (RU 2) suggests that certain type-traits should be required to diagnose violations of their pre-conditions. The basic idea is founded and I see no problems for requiring this for the mentioned traits alignment_of or is_base_of, for example. But if we want to require this diagnostics for some other traits, such as is_convertible, is_constructible (and friends), or is_callable (and possibly some others), we really should be sure that our current requirements are OK.
Unfortunately, there exists some cases, where we currently overspecify imposing complete type requirements where they are not actually required. For example, for the following situation the answer of the trait could be given without ever needing the complete type of X:struct X; // Never defined static_assert(std::is_convertible_v<X, const X&>);
Unfortunately we cannot always allow incomplete types, because most type constructions or conversions indeed require a complete type, so generally relaxing the current restrictions is also not an option.
The core language has a solution for this "small" gap of situations, where the response of the compiler might depend on type completeness: Undefined behaviour. So, I believe we need a somewhat more detailled form to express the intend here. Informally, I would suggest that the program should only be ill-formed in the situation described by LWG 2797, if there exists the possibility that the compiler would require complete types for the considered operation. The example shown above, std::is_convertible_v<X, const X&>, would never require the need to complete X, so here no violation should exist. The presented example might seem a tiny one, but the Standard Library type traits are extreme fundamental tools and we should try to not give the impression that an approximate rule of the current type constraints breaks reasonable code. It is correct, that above example has currently undefined behaviour due to the breakage of pre-conditions, therefore this issue suggests to fix the current situation before enforcing a diagnostic for such valid situations.[2017-03-04, Kona]
Set priority to 2. Is related to 2797, but really needs an audit of the type traits.
[2018-08 Batavia Monday issue discussion]
Issues 2797, 2939, 3022, and 3099 are all closely related. Walter to write a paper resolving them.
[2020-02 Prague Thursday issue discussion]
Two of the issues (2797 and 3022) had been resolved by the acceptance of P1285R0.
Proposed resolution:
Section: 31.12.8.1 [fs.enum.path.format], 99 [fs.class.directory_iterator], 31.12.13.4 [fs.op.copy] Status: New Submitter: Thomas Köppe Opened: 2017-03-14 Last modified: 2020-09-06
Priority: 3
View all issues with New status.
Discussion:
During the application of P0430R2, we came across several terms that seem insufficiently clear and lacking proper definitions.
We would like clarification on what those terms mean, and we would welcome wording suggestions, or alternatively a verbose explanation and dispensation to change the presentation editorially. The items in question are:What does it mean to 'always interpret [the character sequence] in the same way'? ([fs.enum.path.format], which is [path.format] in the paper)
What is a "directory-like" file type? ([fs.class.directory_iterator])
What is an "implementation-defined file type"? (e.g. [fs.op.copy])
[2017-07 Toronto Wed Issue Prioritization]
Priority 3
[2018-01-16, Jonathan comments]
In 31.12.8.1 [fs.enum.path.format] "always interpreted in the same way" means irrespective of the path::format value, or the content of the string. Maybe add ", rather than depending on the path::format value or the content of the character sequence".
In 99 [fs.class.directory_iterator] an "implementation-defined directory-like file type" is an implementation-defined file type (see [fs.enum.file_type] and Table 115) that is treated like a directory by the special rules that the OS has for non-regular files (see 31.12.6.2 [fs.path.generic]). In 31.12.13.4 [fs.op.copy], an "implementation-defined file type" is exactly that, see [fs.enum.file_type] and Table 115 again. I don't see what isn't clear about that. Maybe add a cross-reference to [fs.enum.file_type].Proposed resolution:
Section: 16 [library] Status: New Submitter: Jens Maurer Opened: 2017-03-20 Last modified: 2017-07-15
Priority: 4
View other active issues in [library].
View all other issues in [library].
View all issues with New status.
Discussion:
This is from editorial issue #1088:
It is not always made explicit whether the requirement is referring to time or space complexity, or both."Linear time." vs. "Linear."
"Constant time." vs. "Constant."
16.3.2.4 [structure.specifications] says that the Complexity element specifies "the time and/or space complexity of the function", so being explicit about this would be good.
Examples:[sequence.reqmts] a.clear()
[deque.cons]
[vector.cons]
[map.cons]
[multimap.cons]
[set.cons]
[multiset.cons]
[unord.map.cnstr]
[unord.multimap.cnstr]
etc.
[2017-07 Toronto Wed Issue Prioritization]
Priority 4; Robert to look at
Proposed resolution:
Section: 23.2.4.4 [char.traits.specializations.char16.t] Status: New Submitter: Jonathan Wakely Opened: 2017-05-05 Last modified: 2019-04-02
Priority: 3
View all other issues in [char.traits.specializations.char16.t].
View all issues with New status.
Discussion:
The standard requires that char_traits<char16_t>::int_type is uint_least16_t, so when that has the same representation as char16_t there are no bits left to represent the eof value.
23.2.4.4 [char.traits.specializations.char16.t] says:— The member eof() shall return an implementation-defined constant that cannot appear as a valid UTF-16 code unit.
Existing practice is to use the "noncharacter" u'\uffff' for this value, but the Unicode spec is clear that U+FFFF and other noncharacters are valid, and their appearance in a UTF-16 string does not make it ill-formed. See here and here:
The fact that they are called "noncharacters" and are not intended for open interchange does not mean that they are somehow illegal or invalid code points which make strings containing them invalid.
In practice this means there's no way to tell if basic_streambuf<char16_t>::sputc(u'\uffff') succeeded or not. If it can insert the character it returns to_int_type(u'\uffff') and otherwise it returns eof(), which is the same value.
I believe that char_traits<char16_t>::to_int_type(char_type c) can be defined to transform U+FFFF into U+FFFD, so that the invariant eq_int_type(eof(), to_int_type(c)) == false holds for any c (and the return value of sputc will be distinct from eof). I don't think any implementation currently meets that invariant. I think at the very least we need to correct the statement "The member eof() shall return an implementation-defined constant that cannot appear as a valid UTF-16 code unit", because there are no such constants if sizeof(uint_least16_t) == sizeof(char16_t). This issue is closely related to LWG 1200, but there it's a slightly different statement of the problem, and neither the submitter's recommendation nor the proposed resolution solves this issue here. It seems that was closed as NAD before the Unicode corrigendum existed, so at the time our standard just gave "surprising results" but wasn't strictly wrong. Now it makes a normative statement that conflicts with Unicode.[2017-07 Toronto Wed Issue Prioritization]
Priority 3
Proposed resolution:
Section: 25.3.5.3 [input.iterators] Status: Open Submitter: Gašper Ažman Opened: 2017-05-10 Last modified: 2022-04-25
Priority: 2
View other active issues in [input.iterators].
View all other issues in [input.iterators].
View all issues with Open status.
Discussion:
In Table 95 in 25.3.5.3 [input.iterators], it is specified that the expression *a returns reference, which must be convertible to value_type. This is not true for move-only types, which incidentally means that std::vector<std::unique_ptr<int>> does not possess even a lowly InputIterator, which is, of course, absurd.
With the advent of concepts as first-class citizens in the language, getting this right as soon as possible is a priority. This issue seems to be similar to both LWG 448 and LWG 484, but not the same. The proposed resolution stems from two considerations outlined below:Convertibility is too strong for all algorithms
No algorithm in the standard library requires convertibility to value_type. If algorithms require things that smell of that, they specify the assignment or constructibility flavor they need directly. I checked this by going through the specification of each and every one of them in <algorithm> and <numeric>, which highlighted several issues unrelated to this one. These issues are presented in Algorithms with underspecified iterator requirements (LWG 2963). reference needs to be related to value_type Algorithms need this for the following reasons:
lifetime-extension: served as adequately by T const& as by T. Also works for iterators that return by value. T&& also correctly binds to T const&.
passing to predicates: again, served adequately by T const&
writing to *result: not provided by the requirement anyway.
capture-by-copy: currently implicitly guaranteed, but unused in the standard library (always specified separately). A separate specification can always be made for algorithms that need to capture-by-copy.
We must give due consideration to code that so far required its inputs to be CopyConstructible implicitly by requiring convertibility to T. This is done in the issue LWG 2963, which presents the results of a comb-through of <algorithm> and <numeric> to find algorithms that have this requirement, but where it is not specified. While related issues have been identified, no algorithms seems to require more than T const& convertibility without separately requiring convertibility to T.
Since such code is already compiling today, relaxing this requirement does not break code. The only code this could possibly break is if, in a concept checking library, the InputIterator concept requirement on reference being convertible to value_type gets relaxed. Such a library, if it offered overloading based on most-specific modeled concept, could now, once fixed, resolve the call to a different algorithm, which could break user code that uses a hypothetical algorithm with a move-only container and was relying to select some other overload for move-only types based on the implicit CopyConstructible assertion provided by the iterator. In our internal concepts-checking library, we have had this issue "fixed" since the very beginning — move-only types were too important for our internal algorithms library, and also no algorithm in it seems to require something like Iterator::value_type x = *first without also requiring copy-constructibility anyway.
[2017-07 Toronto Monday issue prioritization]
Priority 2; also could affect the ranges TS
Previous resolution [SUPERSEDED]:
This wording is relative to N4659.
Change Table 95 — "Input iterator requirements", 25.3.5.3 [input.iterators] as indicated:
Table 107 — Input iterator requirements (in addition to Iterator) Expression Return type Operational
semanticsAssertion/note pre-/post-condition … *a reference,
convertible to T
that binds to const T&[…] … *r++ convertible to T
that binds to const T&{ Tauto&& tmp = *r;
++r;
return tmp; }
[2018-04-20; Eric Niebler provides improved wording]
The revised wording makes it clear that you can only rely on those operational semantics when the value type is constructible from the reference type and is movable. When those conditions aren't met, we can make no guarantees about the operational semantics of *r++ (which is why *r++ is no longer a required expression of the InputIterator concept in the Ranges TS).
Really, no generic code should be doing *r++ on input iterators. Another option would be to simply deprecate this requirement for input iterators, but that might need a paper. (For forward iterators, *r++ is already required to return reference exactly, and the multi-pass guarantee gives it the proper semantics.) I also now have a question about the proposed return type of *a and *r++, which says they must be something that "binds to const T&". Does this mean that an iterator with a reference type reference-to-[const?]-volatile-T is no longer considered an iterator? I don't think that's what we want to say. Perhaps these should read "binds to const volatile T& instead, except that has the problem for InputIterators that return prvalues that a prvalue is not bindable to a volatile reference.[2018-11 San Diego Thursday night issue processing]
Look at Ranges; EricWF to investigate. Status to Open
Previous resolution [SUPERSEDED]:
This wording is relative to N4741.
Change Table 89 — "Input iterator requirements", 25.3.5.3 [input.iterators] as indicated:
Table 89 — Input iterator requirements (in addition to Iterator) Expression Return type Operational
semanticsAssertion/note pre-/post-condition … *a reference,
convertible to T
that binds to const T&[…] … *r++ convertible to T
that binds to const T&When T tmp = *r is well-formed and
T is MoveConstructible, then
{ T tmp = *r;
++r;
return tmp; }
[2022-04-25; Daniel rebases wording on N4910]
Proposed resolution:
This wording is relative to N4910.
Change 25.3.5.3 [input.iterators], Table 83 — "Cpp17InputIterator requirements (in addition to Cpp17Iterator) [tab:inputiterator]" as indicated:
Table 83 — Cpp17InputIterator requirements (in addition to Cpp17Iterator) [tab:inputiterator] Expression Return type Operational
semanticsAssertion/note pre-/post-condition … *a reference,
convertible to T
that binds to const T&[…] … *r++ convertible to T
that binds to const T&When T tmp = *r is well-formed and
T is Cpp17MoveConstructible, then
{ T tmp = *r;
++r;
return tmp; }
Section: 27 [algorithms], 28 [numerics] Status: New Submitter: Gašper Ažman Opened: 2017-05-10 Last modified: 2017-07-12
Priority: 3
View other active issues in [algorithms].
View all other issues in [algorithms].
View all issues with New status.
Discussion:
While researching whether the proposed resolution of Iterators of Containers of move-only types do not model InputIterator (LWG 2962), I came across several algorithms that underspecify their requirements, mostly with regard to some associated type of the iterator type they operate on. A list can be found below.
The list of algorithms with underspecified requirements from <algorithm> and <numeric> follows. With the advent of concepts, these algorithms will need better specifications if we are ever hoping to be allowed to overload based on them. I want this issue to bring the standard algorithms closer to having their concept requirements directly transcribable to library annotations. Suggested resolution:copy, copy_if, copy_n, copy_backward
Add to description: *result shall be assignable from *first.
move, move_backward
Add to description: *result shall be move-assignable from *first.
transform
Add to description: The result of the expression op(*first) or binary_op(*first1, *first2) shall be writable to result.
rotate_copy
Add to description: *first shall be writable to result.
merge
Add to description: *first1 and *first2 shall be writable to result..
set_union, set_intersection, set_difference, set_symmetric_difference
Add to description: *first1 and *first2 shall be writable to result.
partial_sum
acc is not defined.
Change description: acc, a variable of InputIterator's value type, shall be constructibleadjacent_difference
acc is not defined.
Change description: acc, a variable of InputIterator's value type, shall be MoveAssignable and shall be constructible from the type of *first.iota
iota is mis-specified. Since the expression we need to support is *first = value: *first is required to be of type InputIterator::reference, and value is an lvalue of type T. The current specification allows calling iota with a const output iterator!
[2017-07 Toronto Monday issue prioritization]
Priority 3; Marshall to work with Gaspar to improve wording.
Proposed resolution:
Section: 27.8.6 [alg.merge] Status: LEWG Submitter: Billy Robert O'Neal III Opened: 2017-06-08 Last modified: 2020-09-06
Priority: Not Prioritized
View all other issues in [alg.merge].
View all issues with LEWG status.
Discussion:
At the moment, inplace_merge requires exactly N - 1 comparisons, if enough additional memory is available (and in practice enough additional memory is always available). However, this prohibits implementing the merge operation using forms of binary search, as in Timsort's 'Galloping Mode', a useful optimization for non-uniform input data. It's not really useful to prohibit standard libraries from trying a few extra speculative compares like this, given that users must be prepared for the fallback "not enough memory" 𝒪(N lg N) algorithm.
[2017-07 Toronto Monday issue prioritization]
Status to LEWG
Proposed resolution:
This wording is relative to N4659.
Edit 27.8.6 [alg.merge] as indicated:
template<class BidirectionalIterator> void inplace_merge(BidirectionalIterator first, BidirectionalIterator middle, BidirectionalIterator last); template<class ExecutionPolicy, class BidirectionalIterator> void inplace_merge(ExecutionPolicy&& exec, BidirectionalIterator first, BidirectionalIterator middle, BidirectionalIterator last); template<class BidirectionalIterator, class Compare> void inplace_merge(BidirectionalIterator first, BidirectionalIterator middle, BidirectionalIterator last, Compare comp); template<class ExecutionPolicy, class BidirectionalIterator, class Compare> void inplace_merge(ExecutionPolicy&& exec, BidirectionalIterator first, BidirectionalIterator middle, BidirectionalIterator last, Compare comp);[…]
-8- Complexity: Let N = last - first:
(8.1) — For the overloads with no ExecutionPolicy, if enough additional memory is available,
exactlyN - 1 comparisons on average, 𝒪(N) comparisons in the worst case.(8.2) — For the overloads with no ExecutionPolicy if no additional memory is available, 𝒪(N log N) comparisons.
(8.3) — For the overloads with an ExecutionPolicy, 𝒪(N log N) comparisons.
-9- Remarks: Stable (16.4.6.8 [algorithm.stable]).
Section: 30.4.7.3.2 [locale.money.put.virtuals] Status: New Submitter: Jonathan Wakely Opened: 2017-06-21 Last modified: 2017-06-27
Priority: 3
View all other issues in [locale.money.put.virtuals].
View all issues with New status.
Discussion:
Whether you get ".99" or "0.99" for the following depends on the implementation:
std::cout.imbue(std::locale("en_US")); std::cout << std::put_money(99.L);
I don't see any justification in [locale.money.put.virtuals] for the leading 0, although that seems more useful.
If we want the leading zero, we should say so.[2017-06-27, Jonathan comments and provides wording]
I suggest that we require a leading zero. The wording below is similar to how C specifies the %f format specifier for fprintf.
Proposed resolution:
This wording is relative to N4659.
Edit 30.4.7.3.2 [locale.money.put.virtuals] as indicated:
iter_type do_put(iter_type s, bool intl, ios_base& str, char_type fill, long double units) const; iter_type do_put(iter_type s, bool intl, ios_base& str, char_type fill, const string_type& digits) const;[…]
-2- Remarks: The currency symbol is generated if and only if (str.flags() & str.showbase) is nonzero. If the format specifies a decimal point, at least one digit character appears before it. If the number of characters generated for the specified format is less than the value returned by str.width() on entry to the function, then copies of fill are inserted as necessary to pad to the specified width. For the value af equal to (str.flags() & str.adjustfield), if (af == str.internal) is true, the fill characters are placed where none or space appears in the formatting pattern; otherwise if (af == str.left) is true, they are placed after the other characters; otherwise, they are placed before the other characters.
Section: 31.7.8 [ext.manip] Status: New Submitter: Jonathan Wakely Opened: 2017-06-22 Last modified: 2017-06-26
Priority: 3
View all other issues in [ext.manip].
View all issues with New status.
Discussion:
[ext.manip] p5 says:
Requires: The type moneyT shall be either long double or a specialization of the basic_string template (Clause 24).
This means that put_money(99), put_money(99.), put_money("99"), and put_money(string_view{"99"}) are all undefined, when in practice they will compile fine and do the right thing, converting the argument to long double or std::string as needed.
We could change it to be "otherwise the program is ill-formed", or to remove the function templates from overload resolution when the argument is not long double or a std::basic_string, but that will unnecessarily break code that works fine today. We should accept types convertible to long double or the relevant money_put facet's string_type (which is not known until we attempt to write the unspecified type to an ostream). The requirement is also insufficient, because cout << put_money(wstring(L"99")) won't compile on any implementation, despite the argument type being a specialization of basic_string. This same problem exists for std::get_money.[2017-06-24, Daniel comments and provides wording]
The wording changes below are supposed to support all moneyT types that are convertible to either long double or to money_put/get<Ch, o/istreambuf_iterator<Ch, Tr>>::string_type (but not to both), where Ch and Tr are determined by the concrete instantiated specialization of the exposition-only function template f that is used to specify the semantics of put_money and get_money, respectively. XOR-ing the requirements outlaws types that are convertible to both, which would cause an ambiguity unless we would provide wording that would introduce an ordered application of these convertibility constraints. This is the rationale for the seemingly odd new Remarks formulation. Note also, that the wording provided below intentionally attempts to distinguish between the statically testable conditions based on the is_convertible_v expressions within the Remarks: element and the well-defined runtime behaviour requirement of the actually provided argument of deduced type moneyT within the pre-existing Requires: element. Another point worth pointing out is that the wording attempts to fix an currently existing ambiguity of the meaning of the type moneyT (and to a lesser extend for charT and traits) as either the template parameter of put/get_money or that of the corresponding template argument of the exposition-only f templates. The revised form makes it clearer that it refers to the latter.
It should be emphasized that this extension of the current wording would provide support for put_money(99), put_money(99.), and put_money("99"), but not yet for put_money(string_view{"99"}), because string_view is not convertible to string. To realize support for the latter, this wording approach could be extended by referring to is_constructible instead of is_convertible, though.Proposed resolution:
This wording is relative to N4659.
Edit 31.7.8 [ext.manip] as indicated:
template <class moneyT> unspecified get_money(moneyT& mon, bool intl = false);-?- For an expression in >> get_money(mon, intl) described below, let Mo, Ch, and Tr be the deduced template argument types of the template parameters moneyT, charT, and traits, respectively, of the instantiated specialization of the template f.
-2- Requires:The type moneyT shall be either long double or a specialization of the basic_string template (Clause 23 [strings])Mo shall be either convertible to long double or shall be convertible to money_get<Ch, istreambuf_iterator<Ch, Tr>>::string_type. -?- Remarks: If is_convertible_v<Mo, long double> == is_convertible_v<Mo, money_get<Ch, istreambuf_iterator<Ch, Tr>>::string_type>, the program is ill-formed. -3- Effects: The expression in >> get_money(mon, intl) described below behaves as a formatted input function (31.7.5.3.1 [istream.formatted.reqmts]). -4- Returns: An object of unspecified type such that if in is an object of type basic_istream<charTCh,traitsTr> then the expression in >> get_money(mon, intl) behaves as if it called f(in, mon, intl), where the function f is defined as:template <class charT, class traits, class moneyT> void f(basic_ios<charT, traits>& str, moneyT& mon, bool intl) { using Iter = istreambuf_iterator<charT, traits>; using MoneyGet = money_get<charT, Iter>; ios_base::iostate err = ios_base::goodbit; const MoneyGet& mg = use_facet<MoneyGet>(str.getloc()); mg.get(Iter(str.rdbuf()), Iter(), intl, str, err, mon); if (ios_base::goodbit != err) str.setstate(err); }The expression in >> get_money(mon, intl) shall have type basic_istream<
charTCh,traitsTr>& and value in.template <class moneyT> unspecified put_money(const moneyT& mon, bool intl = false);-?- For an expression out << put_money(mon, intl) described below, let Mo, Ch, and Tr be the deduced template argument types of the template parameters moneyT, charT, and traits, respectively, of the instantiated specialization of the template f.
-5- Requires:The type moneyT shall be either long double or a specialization of the basic_string template (Clause 23 [strings])Mo shall be either convertible to long double or shall be convertible to money_put<Ch, ostreambuf_iterator<Ch, Tr>>::string_type. -?- Remarks: If is_convertible_v<Mo, long double> == is_convertible_v<Mo, money_put<Ch, ostreambuf_iterator<Ch, Tr>>::string_type>, the program is ill-formed. -6- Returns: An object of unspecified type such that if out is an object of type basic_ostream<charTCh,traitsTr> then the expression out << put_money(mon, intl) behaves as a formatted output function (31.7.6.3.1 [ostream.formatted.reqmts]) that calls f(out, mon, intl), where the function f is defined as:template <class charT, class traits, class moneyT> void f(basic_ios<charT, traits>& str, const moneyT& mon, bool intl) { using Iter = ostreambuf_iterator<charT, traits>; using MoneyPut = money_put<charT, Iter>; const MoneyPut& mp = use_facet<MoneyPut>(str.getloc()); const Iter end = mp.put(Iter(str.rdbuf()), intl, str, str.fill(), mon); if (end.failed()) str.setstate(ios::badbit); }The expression out << put_money(mon, intl) shall have type basic_ostream<
charTCh,traitsTr>& and value out.
Section: 27.7.10 [alg.reverse] Status: LEWG Submitter: Billy O'Neal III Opened: 2017-06-24 Last modified: 2018-04-03
Priority: Not Prioritized
View all other issues in [alg.reverse].
View all issues with LEWG status.
Discussion:
The fine folks on our backend team suggested that we special case std::reverse of 1/2/4/8 to take advantage of vector units. Unfortunately, at present std::reverse says it does N/2 iter_swaps, which doesn't permit our vector implementation even if the iterator inputs are pointers to trivially copyable Ts.
The vectorized version for pointers to shorts is ~8x faster on Skylake than the serial version, and about 7x faster for unsigned long longs; and users don't actually care whether or not we call swap here.[2017-07 Toronto Monday issue prioritization]
Status to LEWG; this is similar to 2973
[2018-04-02, Billy comments]
This issue should be resolved by P0551, because it prohibits user specialization of std::swap and std::iter_swap, which means the proposed vectorization optimization for pointers-to-trivially-copyable is now implementable without changes to reverse's specification (We can detect if the user has provided an alternate swap in their own namespace, but not if they explicitly specialized swap or iter_swap).
Proposed resolution:
This wording is relative to N4659.
Edit 27.7.10 [alg.reverse] as indicated:
template<class BidirectionalIterator> void reverse(BidirectionalIterator first, BidirectionalIterator last); template<class ExecutionPolicy, class BidirectionalIterator> void reverse(ExecutionPolicy&& exec, BidirectionalIterator first, BidirectionalIterator last);-1- Requires: *first shall be swappable (16.4.4.3 [swappable.requirements]).
-2- Effects: For each non-negative integer i < (last - first) / 2, applies iter_swap to all pairs of iterators first + i, (last - i) - 1. If is_trivially_copyable_v<typename iterator_traits<BidirectionalIterator>::value_type> is true, an implementation may permute the elements by making temporary copies, rather than by calling iter_swap. [Note: this allows the implementation to be vectorized. — end note]
Section: 32.12 [re.grammar] Status: New Submitter: Hubert Tong Opened: 2017-06-25 Last modified: 2017-07-12
Priority: 4
View other active issues in [re.grammar].
View all other issues in [re.grammar].
View all issues with New status.
Discussion:
In N4660 subclause 31.13 [re.grammar] paragraph 5:
The productions ClassAtomExClass, ClassAtomCollatingElement and ClassAtomEquivalence provide functionality equivalent to that of the same features in regular expressions in POSIX.
The broadness of the above statement makes it sound like it is merely a statement of intent; however, this appears to be a necessary normative statement insofar as identifying the general semantics to be associated with the syntactic forms identified. In any case, if it is meant for ClassAtomCollatingElement to provide functionality equivalent to a collating symbol in a POSIX bracket expression, multi-character collating elements need to be considered.
In [re.grammar] paragraph 14:The behavior of the internal finite state machine representation when used to match a sequence of characters is as described in ECMA-262. The behavior is modified according to any match_flag_type flags specified when using the regular expression object in one of the regular expression algorithms. The behavior is also localized by interaction with the traits class template parameter as follows: [bullets 14.1 to 14.4]
In none of the bullets does the wording handle multi-character collating elements in a clear manner:
14.1 deals in characters.
14.2 deals in characters (traits_inst.translate accepts only a single character).
14.3 might handle a multi-character collating element; however, there is no specification of how such a collating element is to be identified from the sequence of characters. Additionally, the definition of primary equivalence class specifies that it is a set of characters (not of collating elements).
14.4 deals in characters.
The ECMA-262 specification for ClassRanges also deals in characters.
[2017-07 Toronto Monday issue prioritization]
Priority 4
Proposed resolution:
Section: 32.12 [re.grammar] Status: New Submitter: Hubert Tong Opened: 2017-06-25 Last modified: 2017-07-12
Priority: 3
View other active issues in [re.grammar].
View all other issues in [re.grammar].
View all issues with New status.
Discussion:
For a user to implement a regular expression traits class meaningfully, the relationship between the return value of traits_inst.lookup_collatename to the behaviour of the finite state machine corresponding to a regular expression needs to be better specified.
From N4660 subclause 31.13 [re.grammar], traits_inst.lookup_collatename only feeds clearly into two operations:a test if the returned string is empty ([re.grammar]/8), and
a test if the result of traits_inst.transform_primary, with the returned string, is empty ([re.grammar]/10).
Note: It is unclear if bullet 14.3 in [re.grammar]/14 refers to the result of traits_inst.lookup_collatename when it refers to a "collating element"; and if it does, it is unclear what input is to be used.
It is therefore unclear what the effect is if traits_inst.lookup_collatename substitutes another member of the equivalence class as its output. For example, when processing "[[.AA.]]" as a pattern under a locale da_DK.utf8, what is the expected behaviour difference (if any) should traits_inst.lookup_collatename return, for "AA", "\u00C5" (where U+00C5 is A with ring, which sorts the same as "AA")?[2017-07 Toronto Monday issue prioritization]
Priority 3
Proposed resolution:
Section: 22.5.3 [optional.optional] Status: Open Submitter: Casey Carter Opened: 2017-06-27 Last modified: 2018-01-28
Priority: 3
View all other issues in [optional.optional].
View all issues with Open status.
Discussion:
optional<T>::value_type is T, which can be a cv-qualified object type. This is inconsistent with the uses of the name value_type elsewhere in the standard. We should either require optional<T>::value_type to be remove_cv_t<T> — a true value type — or rename the type alias to element_type.
[2017-07 Toronto Tuesday PM issue prioritization]
Priority 3; may also affect array
[2018-1-26 issues processing telecon]
Status to 'Open'
Proposed resolution:
This wording is relative to N4659.
Edit 22.5.3 [optional.optional], class template optional synopsis, as indicated:
template <class T> class optional { public: using value_type = remove_cv_t<T>; […] };
Section: 22.6.3.2 [variant.ctor] Status: LEWG Submitter: Peter Dimov Opened: 2017-06-27 Last modified: 2017-07-12
Priority: Not Prioritized
View other active issues in [variant.ctor].
View all other issues in [variant.ctor].
View all issues with LEWG status.
Discussion:
The copy constructor of std::variant is not conditionally noexcept (I think it was in the original proposal.)
It should be, for two reasons: first, this would be consistent with the other three constructorsconstexpr variant() noexcept(see below); variant(variant&&) noexcept(see below); template <class T> constexpr variant(T&&) noexcept(see below);
and second, variant itself makes use of is_nothrow_copy_constructible, so it's inconsistent for it to take a stance against it.
[2017-07 Toronto Tuesday PM issue prioritization]
Status to LEWG
Proposed resolution:
This wording is relative to N4659.
Edit 22.6.3 [variant.variant], class template variant synopsis, as indicated:
template <class... Types> class variant { public: // 23.7.3.1, constructors constexpr variant() noexcept(see below); variant(const variant&) noexcept(see below); variant(variant&&) noexcept(see below); […] };
Edit 22.6.3.2 [variant.ctor] as indicated:
variant(const variant& w) noexcept(see below);[…]
-8- Remarks: This function shall not participate in overload resolution unless is_copy_constructible_v<Ti> is true for all i. The expression inside noexcept is equivalent to the logical AND of is_nothrow_copy_constructible_v<Ti> for all i.
Section: 23.2 [char.traits], 23.4.3.2 [string.require], 23.3.3 [string.view.template] Status: Open Submitter: Gennaro Prota Opened: 2017-07-03 Last modified: 2018-11-25
Priority: 3
View all other issues in [char.traits].
View all issues with Open status.
Discussion:
basic_string and basic_string_view involve undefined behavior in a few cases where it's simple for the implementation to add a static_assert and make the program ill-formed.
With regards to basic_string, 23.2 [char.traits]/3 states:Traits::char_type shall be the same as CharT.
Here, the implementation can add a static_assert using the is_same type trait. Similar issues exist in 23.4.3.2 [string.require] and, for basic_string_view, in 23.3.3 [string.view.template]/1.
[2017-07 Toronto Tuesday PM issue prioritization]
Priority 3; need to check general container requirements
Partially by the adoption of P1148 in San Diego.
Tim opines: "the remainder deals with allocator value type mismatch, which I think is NAD."
Proposed resolution:
This wording is relative to N4659.
Edit 23.2 [char.traits] as indicated:
-3- To specialize those templates to generate a string or iostream class to handle a particular character container type CharT, that and its related character traits class Traits are passed as a pair of parameters to the string or iostream template as parameters charT and traits. If Traits::char_type
shall be the sameis not the same type as CharT, the program is ill-formed.
Edit 23.4.3.2 [string.require] as indicated:
-3- In every specialization basic_string<charT, traits, Allocator>, if
the typeallocator_traits<Allocator>::value_typeshall name the same typeis not the same type as charT, the program is ill-formed. Every object of type basic_string<charT, traits, Allocator> shall use an object of type Allocator to allocate and free storage for the contained charT objects as needed. The Allocator object used shall be obtained as described in 24.2.2.1 [container.requirements.general]. In every specialization basic_string<charT, traits, Allocator>, the type traits shall satisfy the character traits requirements (23.2 [char.traits]). If, and the typetraits::char_typeshall name the same typeis not the same type as charT, the program is ill-formed.
Edit 23.3.3 [string.view.template] as indicated:
-1- In every specialization basic_string_view<charT, traits>, the type traits shall satisfy the character traits requirements (23.2 [char.traits]). If
, and the typetraits::char_typeshall name the same typeis not the same type as charT, the program is ill-formed.
Section: 33.10.6 [futures.promise] Status: Open Submitter: Billy O'Neal III Opened: 2017-07-16 Last modified: 2020-09-06
Priority: 2
View other active issues in [futures.promise].
View all other issues in [futures.promise].
View all issues with Open status.
Discussion:
In Toronto Saturday afternoon LWG discussed LWG 2976 which finishes the job of removing allocator support from packaged_task. LWG confirmed that, despite the removal of packaged_task allocators "because it looks like std::function" was incorrect, they wanted to keep the allocator removals anyway, in large part due to this resolution being a response to an NB comment.
If we don't want the type erased allocator situation at all, then we should remove them from the remaining place they exist in <future>, namely, in promise. This change also resolves potential implementation divergence on whether allocator::construct is intended to be used on elements constructed in the shared state, and allows the emplace-construction-in-future paper, P0319, to be implemented without potential problems there.[28-Nov-2017 Mailing list discussion - set priority to P2]
Lots of people on the ML feel strongly about this; the suggestion was made that a paper would be welcomed laying out the rationale for removing allocator support here (and in other places).
[2018-1-26 issues processing telecon]
Status to 'Open'; Billy to write a paper.
[2019-06-03]
Jonathan observes that this resolution conflicts with 2095.
Proposed resolution:
This resolution is relative to N4659.
Edit 33.10.6 [futures.promise], class template promise synopsis, as indicated:
template<class R> class promise { public: promise();[…]template <class Allocator> promise(allocator_arg_t, const Allocator& a);[…] }; template <class R> void swap(promise<R>& x, promise<R>& y) noexcept;template <class R, class Alloc> struct uses_allocator<promise<R>, Alloc>;template <class R, class Alloc> struct uses_allocator<promise<R>, Alloc> : true_type { };
-3- Requires: Alloc shall be an Allocator (16.4.4.6 [allocator.requirements]).promise();template <class Allocator> promise(allocator_arg_t, const Allocator& a);-4- Effects: constructs a promise object and a shared state.
The second constructor uses the allocator a to allocate memory for the shared state.
Section: 19.3 [assertions] Status: Open Submitter: Jonathan Wakely Opened: 2017-08-18 Last modified: 2018-08-20
Priority: 2
View other active issues in [assertions].
View all other issues in [assertions].
View all issues with Open status.
Discussion:
The C standard says that the expression in an assert must have a scalar type, and implies (or at least allows) that the condition is tested by comparison to zero. C++ says that the expression is a constant subexpression if it can be contextually converted to bool. Those ways to test the condition are not equivalent.
It's possible to have expressions that meet the C++ requirements for a constant subexpression, but fail to meet the C requirements, and so don't compile.#include <stdlib.h> // A toy implementation of assert: #define assert(E) (void)(((E) != 0) || (abort(), 0)) struct X { constexpr explicit operator bool() const { return true; } }; constexpr bool f(const X& x) { assert(x); return true; }
C++ says that assert(x) is a constant subexpression, but as it doesn't have scalar type it's not even a valid expression.
I think either 19.3.2 [cassert.syn] or 19.3.3 [assertions.assert] should repeat the requirement from C that E has scalar type, either normatively or in a note. We should also consider whether "contextually converted to bool" is the right condition, or if we should use comparison to zero instead.[2017-11 Albuquerque Wednesday night issues processing]
Priority set to 2; status to Open
Jonathan is discussing this with WG14
[2018-08-20, Jonathan comments]
This was reported to WG14 as N2207.
Proposed resolution:
Section: 19.5.3.4 [syserr.errcat.derived] Status: New Submitter: Thomas Köppe Opened: 2017-09-20 Last modified: 2017-11-09
Priority: 3
View all issues with New status.
Discussion:
The presentation of section [syserr.errcat.derived] is currently somewhat problematic:
It is not clear why this section exists and what it is specifying. Presumably, it is the subject of the phrase "in this subclause" of the sibling section [syserr.errcat.overview], but that's confusing (because this would require the interpretation of "this subclause" as the containing superclause). It would be an improvement to say "shall behave as specified in [syserr.errcat.derived]" in the introduction.
The current wording of [syserr.errcat.derived] requires that derived classes keep the name member function pure-virtual, making it impossible to have non-abstract derived classes. This appears to be an editorial error. Surely name should just not be required to be pure-virtual.
There seems to be no requirement concerning the message virtual member function.
We should use override rather than virtual.
I would welcome a short, introductory paragraph in [syserr.errcat.derived] that states (non-redundantly) that the following are requirements on users' derived classes.
Partial wording proposal:
In 19.5.3.1 [syserr.errcat.overview] p1, change:
-1- The class error_category serves as a base class for types used to identify the source and encoding of a particular category of error code. Classes may be derived from error_category to support categories of errors in addition to those defined in this International Standard. Such classes shall behave as specified in
this subclause19.5.3.4 [syserr.errcat.derived]. [Note: error_category objects are passed by reference, and two such objects are equal if they have the same address. This means that applications using custom error_category types should create a single object of each such type. — end note]
In 19.5.3.4 [syserr.errcat.derived], change:
virtualconst char* name() const noexcept override= 0;-1- Returns: A string naming the error category.
virtualerror_condition default_error_condition(int ev) const noexcept override;-2- Returns: An object of type error_condition that corresponds to ev.
virtualbool equivalent(int code, const error_condition& condition) const noexcept override;-3- Returns: true if, for the category of error represented by *this, code is considered equivalent to condition; otherwise, false.
virtualbool equivalent(const error_code& code, int condition) const noexcept override;-4- Returns: true if, for the category of error represented by *this, code is considered equivalent to condition; otherwise, false.
[2017-11 Albuquerque Wednesday night issues processing]
Priority set to 3.
Jonathan to talk to Chris K and Walter about writing a paper describing the use of error_code, error_condition and defining your own.
Proposed resolution:
Section: 16.2.2 [networking.ts::buffer.reqmts.constbuffersequence] Status: New Submitter: Vinnie Falco Opened: 2017-09-20 Last modified: 2020-09-06
Priority: 3
View all issues with New status.
Discussion:
Addresses: networking.ts
The post-condition buffer sequence requirements mandate pointer equivalence. This means that a copies of buffer sequences must point to the same pieces of underlying memory. While this is appropriate for MutableBufferSequence, it is unnecessary for ConstBufferSequence and can actually prevent useful implementation strategies such as the following constant buffer sequence which avoids dynamic allocations:
/// A buffer sequence containing a chunk-encoding header class chunk_size { public: // Storage for the longest hex string we might need class value_type { friend class chunk_size; // First byte holds the length char buf_[1 + 2 * sizeof(std::size_t)]; template<class = void> void prepare(std::size_t n); template<class OutIter> static OutIter to_hex(OutIter last, std::size_t n) { if (n == 0) { *--last = '0'; return last; } while (n) { *--last = "0123456789abcdef"[n & 0xf]; n >>= 4; } return last; } public: operator boost::asio::const_buffer() const { return { buf_ + sizeof(buf_) - buf_[0], static_cast(buf_[0]) }; } }; using const_iterator = value_type const*; chunk_size(chunk_size const& other) = default; /** Construct a chunk header @param n The number of octets in this chunk. */ chunk_size(std::size_t n) { value_.prepare(n); } const_iterator begin() const { return &value_; } const_iterator end() const { return begin() + 1; } private: value_type value_; };
Proposed resolution:
This wording is relative to N4588.
Modify 16.2.2 [networking.ts::buffer.reqmts.constbuffersequence] Table 13 "ConstBufferSequence requirements" as indicated:
Table 13 — ConstBufferSequence requirements expression return type assertion/note
pre/post-condition[…] X u(x); post:
equal( net::buffer_sequence_begin(x), net::buffer_sequence_end(x), net::buffer_sequence_begin(u), net::buffer_sequence_end(u), [](const typename X::value_type& v1, const typename X::value_type& v2) { const_buffer b1(v1); const_buffer b2(v2);return b1.data() == b2.data() && b1.size() == b2.size()return b1.size() == b2.size() && memcmp(b1.data(), b2.data(), b1.size()) == 0; })
Section: 16.2.4 [networking.ts::buffer.reqmts.dynamicbuffer] Status: New Submitter: Vinnie Falco Opened: 2017-10-16 Last modified: 2020-09-06
Priority: 3
View other active issues in [networking.ts::buffer.reqmts.dynamicbuffer].
View all other issues in [networking.ts::buffer.reqmts.dynamicbuffer].
View all issues with New status.
Discussion:
Addresses: networking.ts
The current wording for the DynamicBuffer prepare member function implies that std::length_error is the only allowable thrown exception. This should be changed to reflect that any exception may be thrown, with std::length_error thrown in particular when size() + n exceeds max_size().
[2017-11-08]
Priority set to 3 after five votes on mailing list
Proposed resolution:
This wording is relative to N4588.
Change 16.2.4 [networking.ts::buffer.reqmts.dynamicbuffer], Table 14 "DynamicBuffer requirements", as indicated:
Table 14 — DynamicBuffer requirements expression return type assertion/note pre/post-condition […] x.prepare(n) X::mutable_buffers_type Returns a mutable buffer sequence u
representing the writable bytes, and where
buffer_size(u) == n . The dynamic buffer
reallocates memory as required. All constant or
mutable buffer sequences previously obtained using
data() or prepare() are invalidated.
Throws: length_error if size() + n
exceeds max_size() or any other exception
if the request cannot otherwise be satisfied.
Section: 27.8.8.3 [pop.heap] Status: Open Submitter: Mathias Stearn Opened: 2017-11-04 Last modified: 2020-09-06
Priority: 3
View all issues with Open status.
Discussion:
The spec for <algorithms> pop_heap includes
-1- Requires: The range [first, last) shall be a valid non-empty heap.
This has the unfortunate consequence that to pop a value and push a new value is substantially less efficient than necessary. The popped value must be extracted by pop_heap (using up to 2 log N compares and swaps), and then, in push_heap, the new value must be inserted (for up to N compares and swaps, but more usually something like log N).
Simply relaxing the requirement to-1- Requires: The range [first, last - 1) shall be a valid heap.
enables use of pop_heap in an integrated push-and-pop operation, with less than half the number of expected compare and swap operations. Furthermore, if, as is often the case, the newly pushed value would have ended up at position first, the push/pop operation could complete in time 𝒪(1), instead of (3 log N).
The effect of the proposed relaxation on existing library implementations would be minimal in the extreme, and on existing user code nil. The base algorithm code remains exactly identical. The only changes needed would be to any instrumentation in a debugging version of the library, which would just need to relax its check, and to test suites that should exercise the newly tolerated input. Users today are tempted to get the improved performance by relying on existing implementations' tacit tolerance of input that only satisfies the proposed, relaxed requirements. In fact, the cppreference.com page on pop_heap offers no hint that this usage is not already allowed. This change would bless such reliance as formally permitted. After this change, minor extensions to std::priority_queue would enable it to take advantage of the newly efficient operation, perhaps:void pop_push(const Type&); void pop_push(Type&&); template <class... Args> void pop_emplace(Args&&... args);
These will appear in a formal proposal if the resolution is accepted.
[2017-11 Albuquerque Wednesday night issues processing]
Priority set to 3
[2017-11 Albuquerque Saturday issues processing]
status to Open; Marshall to review
Proposed resolution:
This wording is relative to N4700.
Change 27.8.8.3 [pop.heap] as indicated:
template<class RandomAccessIterator> void pop_heap(RandomAccessIterator first, RandomAccessIterator last); template<class RandomAccessIterator, class Compare> void pop_heap(RandomAccessIterator first, RandomAccessIterator last, Compare comp);-1- Requires: The range [first, last - 1) shall be a valid
non-emptyheap. RandomAccessIterator shall satisfy the requirements of ValueSwappable (16.4.4.3 [swappable.requirements]). The type of *first shall satisfy the requirements of MoveConstructible (Table 23) and of MoveAssignable (Table 25).
Section: 27.8.8 [alg.heap.operations] Status: Ready Submitter: Robert Douglas Opened: 2017-11-08 Last modified: 2022-11-10
Priority: 3
View all other issues in [alg.heap.operations].
View all issues with Ready status.
Discussion:
In discussion of D0202R3 in Albuquerque, it was observed that pop_heap and sort_heap had constexpr removed for their requirement of ValueSwappable. It was then observed that push_heap and make_heap were not similarly marked as having the ValueSwappable requirement. The room believed this was likely a specification error, and asked to open an issue to track it.
[2017-11 Albuquerque Wednesday night issues processing]
Priority set to 3; Marshall to investigate
Previous resolution [SUPERSEDED]:
This wording is relative to N4700.
Change 27.8.8.2 [push.heap] as indicated:
template<class RandomAccessIterator> void push_heap(RandomAccessIterator first, RandomAccessIterator last); template<class RandomAccessIterator, class Compare> void push_heap(RandomAccessIterator first, RandomAccessIterator last, Compare comp);-1- Requires: The range [first, last - 1) shall be a valid heap. RandomAccessIterator shall satisfy the requirements of ValueSwappable (16.4.4.3 [swappable.requirements]). The type of *first shall satisfy the MoveConstructible requirements (Table 23) and the MoveAssignable requirements (Table 25).
Change 27.8.8.4 [make.heap] as indicated:
template<class RandomAccessIterator> void make_heap(RandomAccessIterator first, RandomAccessIterator last); template<class RandomAccessIterator, class Compare> void make_heap(RandomAccessIterator first, RandomAccessIterator last, Compare comp);-1- Requires: RandomAccessIterator shall satisfy the requirements of ValueSwappable (16.4.4.3 [swappable.requirements]). The type of *first shall satisfy the MoveConstructible requirements (Table 23) and the MoveAssignable requirements (Table 25).
[2022-11-06; Daniel comments and syncs wording with recent working draft]
For reference, the finally accepted paper was P0202R3 and the constexpr-ification of swap-related algorithms had been realized later by P0879R0 after resolution of CWG 1581 and more importantly CWG 1330.
[Kona 2022-11-09; Move to Ready]
Proposed resolution:
This wording is relative to N4917.
Change 27.8.8.2 [push.heap] as indicated:
template<class RandomAccessIterator> constexpr void push_heap(RandomAccessIterator first, RandomAccessIterator last); template<class RandomAccessIterator, class Compare> constexpr void push_heap(RandomAccessIterator first, RandomAccessIterator last, Compare comp); template<random_access_iterator I, sentinel_for<I> S, class Comp = ranges::less, class Proj = identity> requires sortable<I, Comp, Proj> constexpr I ranges::push_heap(I first, S last, Comp comp = {}, Proj proj = {}); template<random_access_range R, class Comp = ranges::less, class Proj = identity> requires sortable<iterator_t<R>, Comp, Proj> constexpr borrowed_iterator_t<R> ranges::push_heap(R&& r, Comp comp = {}, Proj proj = {});-1- Let comp be less{} and proj be identity{} for the overloads with no parameters by those names.
-2- Preconditions: The range [first, last - 1) is a valid heap with respect to comp and proj. For the overloads in namespace std, RandomAccessIterator meets the Cpp17ValueSwappable requirements (16.4.4.3 [swappable.requirements]) and the type of *first meets the Cpp17MoveConstructible requirements (Table 32) and the Cpp17MoveAssignable requirements (Table 34). -3- Effects: Places the value in the location last - 1 into the resulting heap [first, last). -4- Returns: last for the overloads in namespace ranges. -5- Complexity: At most log(last - first) comparisons and twice as many projections.
Change 27.8.8.4 [make.heap] as indicated:
template<class RandomAccessIterator> constexpr void make_heap(RandomAccessIterator first, RandomAccessIterator last); template<class RandomAccessIterator, class Compare> constexpr void make_heap(RandomAccessIterator first, RandomAccessIterator last, Compare comp); template<random_access_iterator I, sentinel_for<I> S, class Comp = ranges::less, class Proj = identity> requires sortable<I, Comp, Proj> constexpr I ranges::make_heap(I first, S last, Comp comp = {}, Proj proj = {}); template<random_access_range R, class Comp = ranges::less, class Proj = identity> requires sortable<iterator_t<R>, Comp, Proj> constexpr borrowed_iterator_t<R> ranges::make_heap(R&& r, Comp comp = {}, Proj proj = {});-1- Let comp be less{} and proj be identity{} for the overloads with no parameters by those names.
-2- Preconditions: For the overloads in namespace std, RandomAccessIterator meets the Cpp17ValueSwappable requirements (16.4.4.3 [swappable.requirements]) and the type of *first meets the Cpp17MoveConstructible (Table 32) and Cpp17MoveAssignable (Table 34) requirements. -3- Effects: Constructs a heap with respect to comp and proj out of the range [first, last). -4- Returns: last for the overloads in namespace ranges. -5- Complexity: At most 3(last - first) comparisons and twice as many projections.
Section: 16.4.4.6 [allocator.requirements] Status: New Submitter: Jon Cohen Opened: 2017-12-06 Last modified: 2022-04-25
Priority: 3
View other active issues in [allocator.requirements].
View all other issues in [allocator.requirements].
View all issues with New status.
Discussion:
Table 31 in the C++17 standard specifies X::max_size() (where X is an allocator type) as "The largest value that can meaningfully be passed to X::allocate()". Noticeably missing is the statement "Throws: Nothing".
As an example of why this is an issue, note that vector::max_size() and allocator_traits::max_size() are both marked noexcept. We must then interpret max_size() as being allowed to sometimes call std::terminate, or else {vector, allocator_traits, ...}::max_size() must be allowed to directly calculate numeric_limits<size_type>::max() / sizeof(value_type) instead of querying the allocator, even if Alloc::max_size() exists. This seems like a bug in the wording for the requirements of max_size() in an allocator type. I think an issue should be opened on this subject to add Throws: Nothing or similar to the requirements of max_size() for an allocator. As an example consider writing up a framework to test the exception-safety of types in a given framework, since they were all written in an exception-free environment. One of the types in the framework is an allocator which, in a controlled way, can throw an exception at any point where it is allowed by the standard. It's important that the test framework be as pedantic as possible, so the allocator type throws on max_size(), since it is currently allowed to by the standard. When a reasonable vector implementation (at least those in libstdc++ and msvc) is, for example, asked to construct a vector from an initializer_list, it will call allocator_traits<Alloc>::max_size(), which will terminate the program because the exception thrown in Alloc::max_size() propagated through the noexcept traits function. Although this is conformant behavior, I think it's a bug in the standard that a function as benign as max_size() can terminate the program in this manner, and I think the fix is that a conformant allocator should be required to supply a non-throwing max_size() member function. Daniel: This problem was shortly discussed during review of LWG 2162 (see comment 2012-08-05). At that time the more drastic but also more consistent requirement that an allocator's max_size function shall not throw exceptions has not been added. IMO this position should be reconsidered to follow the spirit of the new issue LWG 3044.[2018-01; Priority set to 3 after mailing list discussion]
[2018-08-21, Jonathan comments and provides wording]
The phrase "the largest value that can meaningfully be passed to X::allocate()" is meaningless. Is it a requirement on the caller, so that larger values must not be passed? Or a hint from the allocator implementor that larger values will produce a bad_alloc exception? Can the return value change dynamically, based on the free memory available to the allocator?! — LWG 197 says it can't change.
As noted in the LWG 2162 comments, we don't currently guarantee it can be called on a const object (so allocator_traits will not use the allocator's max_size() if it's non-const, although that was unclear before DR 2284). In addition to adding "Throws: nothing" we should ensure it's callable on const lvalues, and clarify what "meaningfully" means and who is supposed to care about it. My proposed resolution doesn't achieve all of this, but is a start.Previous resolution [SUPERSEDED]:
This wording is relative to N4762.
Change 16.4.4.6 [allocator.requirements], Table 32 — "Descriptive variable definitions", as indicated:
Table 32 — Descriptive variable definitions Variable Definition T, U, C any cv-unqualified object type (3.9) … a, a1, a2 lvalues of type X a3 an lvalue of type const X … Change 16.4.4.6 [allocator.requirements], Table 33 — "Cpp17Allocator requirements", as indicated:
Table 33 — Cpp17Allocator requirements Expression Return type Assertion/note
pre-/post-conditionDefault … a3.max_size() X::size_type the largest value that can
meaningfully be passed to
X::allocate().
[Note: Larger values might cause
an exce