[libc++] Undeprecate ATOMIC_FLAG_INIT (LWG3659)

[tambre]

Summary

According to @aaron.ballman this was marked Tentatively Ready as of 2022-07-07.
D129362 implemented the C counterpart.

[philnik]

Please wait until the LWG issue has been accepted.

libcxx/docs/ReleaseNotes.rst

50

We currently don't release-note LWG issues. I don't know if it would make sense to do so. We only started release-noting papers in this release cycle.

libcxx/docs/Status/Cxx2bIssues.csv

159

This wasn't accepted in February 2022. You also forgot to update the number.

[tambre]

Address review comments

[aaron.ballman]

In D129380#3639202, @philnik wrote:

Please wait until the LWG issue has been accepted.

To be clear on the status: LWG has accepted this as tentatively ready, but there is a short window for people to object to that status which would pull the issue back into needing committee discussion. Given the subject matter, it'd be quite surprising if anyone did so, and so this is expected to go into C++23 (and should be treated as DR against C++20).

I'm not certain of libc++'s schedule, but I'm hoping to land the C changes in Clang shortly so that they make it into Clang 15 in an effort to limit user confusion (Clang 14 shipped with the diagnostic enabled).

[tambre]

libcxx/docs/ReleaseNotes.rst

50

I'd say such fixes seem noteworthy enough. This is definitely a smaller one, but I'm keeping it for now unless someone else weighs in.

libcxx/docs/Status/Cxx2bIssues.csv

159

Done, thanks.

[Mordante]

In D129380#3639224, @aaron.ballman wrote:

In D129380#3639202, @philnik wrote:

Please wait until the LWG issue has been accepted.

To be clear on the status: LWG has accepted this as tentatively ready, but there is a short window for people to object to that status which would pull the issue back into needing committee discussion. Given the subject matter, it'd be quite surprising if anyone did so, and so this is expected to go into C++23 (and should be treated as DR against C++20).

I'm not certain of libc++'s schedule, but I'm hoping to land the C changes in Clang shortly so that they make it into Clang 15 in an effort to limit user confusion (Clang 14 shipped with the diagnostic enabled).

Thanks for the additional information. The libc++ policy is not to implement LWG-issue before they are voted in at a plenary, but we make exceptions when needed.
When this gets accepted in the next plenary we definitely should ship this in LLVM 15, however that doesn't mean we need land it before the branching happens, there still time afterwards see
https://discourse.llvm.org/t/llvm-15-0-0-release-schedule/63495/7 and
https://discourse.llvm.org/t/llvm-15-0-0-release-schedule/63495/9

Based on @aaron.ballman's comment I'm in favour to land this after the review comments have been addressed. In the unexpected case it doesn't get voted is we can revert the change.
@ldionne how do you feel about landing this before it's voted in at the next plenary?

libcxx/docs/ReleaseNotes.rst

50

I've no strong opinion whether or not to add it. I don't think we need to list all LWG issues, since some of the are not really "user visible", but this one is.

libcxx/docs/Status/Cxx2bIssues.csv

159

The issues on this page are group by the plenary where they were accepted. Please move this to the bottom of the proper place for the July 2022 plenary.

159

I would like to see an update the note for P0883 with information regarding this LWG issue.

[tambre]

Address review comments

[tambre]

libcxx/docs/Status/Cxx2bIssues.csv

159

Done. Added the note to the P0883's line on the C++20 status page. Hopefully that's what you intended.

[Mordante]

LGTM modulo one nit. I leave the final approval for @ldionne to land this before the issue is voted in.

libcxx/docs/Status/Cxx20.rst

50 ↗

(On Diff #443499)

libcxx/docs/Status/Cxx2bIssues.csv

159

Yes that's what I intended.

[tambre]

Address nit

[ldionne]

I am fine with landing this even though the LWG issue hasn't been voted in officially. We normally don't do that too often, however I think this is a good case to make an exception since this is pretty user-visible.

libcxx/docs/ReleaseNotes.rst

 - P1165R1 (Make stateful allocator propagation more consistent for ``operator+(basic_string)``)
 - P0674R1 (Support arrays in ``make_shared`` and ``allocate_shared``)
 - P0980R1 (Making ``std::string`` constexpr)
 - P2216R3 (std::format improvements)
 - P0174R2 (Deprecating Vestigial Library Parts in C++17)
 - N4190 (Removing auto_ptr, random_shuffle(), And Old <functional> Stuff)
 - P0154R1 (Hardware inference size)
 - P0618R0 (Deprecating <codecvt>)
+- LWG3659 (Consider ``ATOMIC_FLAG_INIT`` undeprecation)
 
 - Marked the following papers as "Complete" (note that some of those might have
   been implemented in a previous release but not marked as such):
 
     - P1207R4 (Movability of Single-pass Iterators);
     - P1474R1 (Helpful pointers for ``ContiguousIterator``);
     - P1522R1 (Iterator Difference Type and Integer Overflow);
     - P1523R1 (Views and Size Types);

libcxx/docs/Status/Cxx2bIssues.csv

 "`3621 <https://wg21.link/LWG3621>`__","Remove feature-test macro ``__cpp_lib_monadic_optional`` ","February 2022","|Complete|","15.0"
 "`3632 <https://wg21.link/LWG3632>`__","``unique_ptr`` ""Mandates: This constructor is not selected by class template argument deduction""","February 2022","|Nothing to do|",""
 "`3643 <https://wg21.link/LWG3643>`__","Missing ``constexpr`` in ``std::counted_iterator`` ","February 2022","",""
 "`3648 <https://wg21.link/LWG3648>`__","``format`` should not print ``bool`` with ``'c'`` ","February 2022","|Complete|","15.0","|format|"
 "`3649 <https://wg21.link/LWG3649>`__","[fund.ts.v2] Reinstate and bump ``__cpp_lib_experimental_memory_resource`` feature test macro","February 2022","",""
 "`3650 <https://wg21.link/LWG3650>`__","Are ``std::basic_string`` 's ``iterator`` and ``const_iterator`` constexpr iterators?","February 2022","|Nothing to do|",""
 "`3654 <https://wg21.link/LWG3654>`__","``basic_format_context::arg(size_t)`` should be ``noexcept`` ","February 2022","|Complete|","15.0","|format|"
 "`3657 <https://wg21.link/LWG3657>`__","``std::hash<std::filesystem::path>`` is not enabled","February 2022","",""
+"`3657 <https://wg21.link/LWG3659>`__","Consider ``ATOMIC_FLAG_INIT`` undeprecation","February 2022","|Complete|","15.0"
 "`3660 <https://wg21.link/LWG3660>`__","``iterator_traits<common_iterator>::pointer`` should conform to §[iterator.traits]","February 2022","|Complete|","14.0"
 "`3661 <https://wg21.link/LWG3661>`__","``constinit atomic<shared_ptr<T>> a(nullptr);`` should work","February 2022","",""
 "","","","",""
 "`3645 <https://wg21.link/LWG3645>`__","``resize_and_overwrite`` is overspecified to call its callback with lvalues","Not voted in","|Complete|","14.0",""
 "`3656 <https://wg21.link/LWG3656>`__","Inconsistent bit operations returning a count","Not voted in","|Complete|","15.0",""
 "","","","",""

libcxx/include/atomic

 typedef atomic<__libcpp_unsigned_lock_free> atomic_unsigned_lock_free;
 #endif
 
 #define ATOMIC_FLAG_INIT {false}
 #define ATOMIC_VAR_INIT(__v) {__v}
 
 #if _LIBCPP_STD_VER > 17 && !defined(_LIBCPP_DISABLE_DEPRECATION_WARNINGS)
 # if defined(_LIBCPP_CLANG_VER) && _LIBCPP_CLANG_VER >= 1400
-#  pragma clang deprecated(ATOMIC_FLAG_INIT)
 #  pragma clang deprecated(ATOMIC_VAR_INIT)
 # endif
 #endif // _LIBCPP_STD_VER > 17 && !defined(_LIBCPP_DISABLE_DEPRECATION_WARNINGS)
 
 _LIBCPP_END_NAMESPACE_STD
 
 #endif // _LIBCPP_ATOMIC

libcxx/test/std/atomics/atomics.flag/init.pass.cpp

 //===----------------------------------------------------------------------===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 
 // XFAIL: c++03
-// ADDITIONAL_COMPILE_FLAGS: -D_LIBCPP_DISABLE_DEPRECATION_WARNINGS
 
 // <atomic>
 
 // struct atomic_flag
 
 // atomic_flag() = ATOMIC_FLAG_INIT;
 
 #include <atomic>