[libcxx-commits] [PATCH] D112462: [libc++][NFC] Mark LWG2731 as complete

Tue Oct 26 06:10:49 PDT 2021

Quuxplusone added inline comments.

================
Comment at: libcxx/docs/Status/Cxx2bIssues.csv:14
 "`3464 <https://wg21.link/LWG3464>`__","``istream::gcount()`` can overflow","November 2020","",""
-"`2731 <https://wg21.link/LWG2731>`__","Existence of ``lock_guard<MutexTypes...>::mutex_type`` typedef unclear","November 2020","",""
+"`2731 <https://wg21.link/LWG2731>`__","Existence of ``lock_guard<MutexTypes...>::mutex_type`` typedef unclear","November 2020","|Complete|","5.0"
 "`2743 <https://wg21.link/LWG2743>`__","P0083R3 ``node_handle`` private members missing ""exposition only"" comment","November 2020","",""
----------------
jloser wrote:
> Quuxplusone wrote:
> > Despite the issue title, the issue is actually about `scoped_lock`, not `lock_guard`, and so I don't see how "5.0" could possibly be the right fix-version.
> > Other than that, LGTM, ship it.
> Well, `scoped_lock` was introduced in https://github.com/llvm/llvm-project/commit/6015dd11c875b9467c93221f19e5d59b0522757a (March 23rd, 2017) and then `llvm5` got released on September 7th, 2017. Given that the `mutex_type` type alias was never an issue from its origin, that's why I dated this resolution back to `5.0`. 
> 
> Would you prefer I just mark it as "Nothing to Do" and omit the version value? I don't feel strongly. I'd also be OK if you just want to label this as `14.0`, but that doesn't feel quite right since there was no code change required in this patch.
OK; I didn't guess either that `scoped_lock` was that old or (more to the point) that LLVM 5 was that recent. We've been through 10 major version numbers in 4 years, eh. 😛 Then I'm OK with `5.0` here.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D112462/new/

https://reviews.llvm.org/D112462