Meta LWG issue: 2022-07 meeting

[frederick-vs-ja]

(Previous meta-issue: #2527)

At the July 2022 virtual plenary meeting, the resolutions of the following LWG issues were merged into the C++ Working Paper.

❔ Not yet analyzed

❌ Not applicable

If an issue requires no action from implementers, we mark it as N/A. Categories:

• Pure wording clarifications with nothing to implement (these can be changes to non-normative text like examples and informative notes, or wording cleanups to normative text that don't impact observable behavior)
  • LWG-3659 Consider ATOMIC_FLAG_INIT undeprecation
    • MSVC STL has never deprecated this macro, and it might be impossible to reflect such deprecation in a library implementation
• Something that increases the restrictions placed on users, but implementers aren't expected to enforce those restrictions.
• Fixes for obviously broken wording, where implementers would have done the right thing anyways
  • LWG-3709 LWG-3703 was underly ambitious
  • LWG-3713 Sorted with respect to comparator (only)

😸 Already implemented

Sometimes we cite LWG issues in product code comments as we're implementing their proposed resolutions. When the resolutions are officially accepted, we should remove the citations (as the default assumption is that we're implementing what the Standard says). If something is especially subtle, we can convert the citation to mention the relevant Standard section.

Sometimes we should add test coverage - e.g. when the Standard begins requiring something that we were already doing, but weren't explicitly testing for.

• Already implemented, comments need to be removed and messages need to cite the Standard
  • LWG-3670 Cpp17InputIterators don't have integer-class difference types
  • LWG-3705 Hashability shouldn't depend on basic_string's allocator
• Implemented without comments
  • LWG-3687expected<cv void, E> move constructor should move
  • LWG-3703 Missing requirements for expected<T, E> requires is_void<T>
  • LWG-3704 LWG-2059 added overloads that might be ill-formed for sets
  • LWG-3708 take_while_view::sentinel's conversion constructor should move

🩹 Patches an unimplemented feature

We should record this LWG issue in the GitHub issue tracking the feature. That way, we'll remember to verify it, but it doesn't represent net new work.

• LWG-3692 zip_view::iterator's operator<=> is overconstrained
• LWG-3702 Should zip_transform_view::iterator remove operator<?

🐞 Not yet implemented

• Unprocessed
• Filed a GitHub issue labeled LWG
  • LWG-3564 transform_view::iterator<true>::value_type and iterator_category should use const F&
  • LWG-3617 function/packaged_task deduction guides and deducing this
    • Note that WG21-P1169 (static operator()) also changes these deduction guides.
  • LWG-3701 Make formatter<remove_cvref_t<const charT[N]>, charT> requirement explicit
• PR out for review
  • LWG-3656 Inconsistent bit operations returning a count (implemented in Implement LWG-3656: Inconsistent bit operations returning a count #2880 )
  • LWG-3671 atomic_fetch_xor missing from stdatomic.h (implemented in Implement LWG-3671: atomic_fetch_xor missing from stdatomic.h #2905 )
  • LWG-3672 common_iterator::operator->() should return by value (implemented in Implement LWG-3672: common_iterator::operator->() should return by value #2907 )
  • LWG-3683 operator== for polymorphic_allocator cannot deduce template argument in common cases (implemented in Implement LWG-3683: operator== for polymorphic_allocator cannot deduce template argument in common cases #2879 )
  • LWG-3707 chunk_view::outer-iterator::value_type::size should return unsigned type (implemented in Implement LWG-3707: chunk_view::outer-iterator::value_type::size should return unsigned type #2883 )
  • LWG-3710 The end of chunk_view for input ranges can be const (implemented in Implement LWG-3710: The end of chunk_view for input ranges can be const #2878 )
  • LWG-3711 Missing preconditions for slide_view constructor (implemented in Implement LWG-3715: view_interface::empty is overconstrained #2946)
  • LWG-3712 chunk_view and slide_view should not be default_initializable (implemented in Implement LWG-3712: chunk_view and slide_view should not be default_initializable #2943 )
  • LWG-3715 view_interface::empty is overconstrained (implemented in Implement LWG-3715: view_interface::empty is overconstrained #2946 )
  • LWG-3719 Directory iterators should be usable with default sentinel (implemented in Implement LWG-3719: Directory iterators should be usable with default sentinel #2902 )
  • LWG-3721 Allow an arg-id with a value of zero for width in std-format-spec (implemented in Implement LWG-3721: Allow an arg-id with a value of zero for width in std-format-spec #2906 )
  • LWG-3724 decay-copy should be constrained (implemented in Implement LWG-3724: decay-copy should be constrained #2903 )

[frederick-vs-ja]

@StephanTLavavej @CaseyCarter feel free to edit this if my analyzation for any LWG issue is incorrect, P/R for any issue hasn't been voted into the working paper, or the implementation status changes.

[CaseyCarter]

Wow, thanks @frederick-vs-ja! This is a ton of work. I'm going to tag this (and any relevant PRs) as "blocked" until we actually get approval for the proposed resolutions in the WG21 virtual plenary on 2022-07-25, but folks should feel free to go ahead and submit/review such PRs. I'm only concerned that we not merge before WG21 approves.

[StephanTLavavej]

Agreed, this is amazing, thank you! 😻 I've applied a small edit, unchecking the "Already implemented, comments need to be removed and messages need to cite the Standard" items (as unchecked boxes mean "action needs to be taken", checked boxes mean "no further action required"), and adding sub-bullets with details about the affected files.

[CaseyCarter]

I believe every issue has now been processed. We have PRs out for the majority, I've recorded the two that patch zip in #2252, filed issues for the three that don't yet have pull requests, and submitted a PR to cleanup comments for the two we implemented speculatively with comments.

Thanks to everyone, and again extra-special thanks to @frederick-vs-ja for filing this issue and a majority of the resolution PRs.