[PATCH] D133425: Silence -Wctad-maybe-unsupported stemming from system headers

David Blaikie via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Thu Sep 8 14:13:29 PDT 2022


dblaikie accepted this revision.
dblaikie added a comment.

In D133425#3777615 <https://reviews.llvm.org/D133425#3777615>, @aaron.ballman wrote:

> In D133425#3777598 <https://reviews.llvm.org/D133425#3777598>, @aaron.ballman wrote:
>
>> In D133425#3775353 <https://reviews.llvm.org/D133425#3775353>, @ldionne wrote:
>>
>>> Re-applying `fcd549a7d8284a8e7c763fee3da2206acd8cdc4f` would not require any Clang changes.
>>
>> I think that would be a good solution if it's workable, at least for the issue we're seeing internally.
>
> It is worth noting that this won't resolve the issue when using libstdc++ and Clang's behavior with this diagnostic differs from GCC's: https://godbolt.org/z/YMznh4Weq

Yeah, seems worth making this work when using libstdc++ too.

So previously/currently-without-this-patch the diagnostic was suppressed if the use of ctad was in a system header (suppression based on the generic/builtin diagnostic suppression infrastructure) & now it'll suppress if that happens, or if the template is defined in a system header.

Seems fair to me. (but welcome not to take my approval as authoritative if you're looking for more/other feedback)


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D133425/new/

https://reviews.llvm.org/D133425



More information about the cfe-commits mailing list