[PATCH] D126907: Deferred Concept Instantiation Implementation Take 2

[Chuanqi Xu via Phabricator via cfe-commits]

ChuanqiXu added a comment.

In D126907#3645198 <https://reviews.llvm.org/D126907#3645198>, @erichkeane wrote:

> In D126907#3644127 <https://reviews.llvm.org/D126907#3644127>, @ChuanqiXu wrote:
>
>> In D126907#3642530 <https://reviews.llvm.org/D126907#3642530>, @erichkeane wrote:
>>
>>> This version passes check-runtimes, so libc++ is fine, and it passes everything I have available.  @ChuanqiXu : Would you be able to do 1 more run over this to make sure it won't break something?  Thanks in advance either way!
>>
>> I've tested some our internal workloads and all of them looks fine. But it is expected since we don't use concept heavily. I had tried to run libcxx before but it looks like my previous config is wrong...
>
> Ok, thanks!  I DID test the libcxx test suite on the above, I figured it out through check-runtimes.  If you were able to take a quick look at the diff and make sure I'm not doing anything horrible, it would be appreciated.

Yeah, things looks good to me basically.



================
Comment at: clang/lib/Sema/SemaTemplateInstantiate.cpp:963-964
+          Entity(Entity),
+          EvaluatingAConstraint(EvaluatingConstraint ||
+                                !SemaRef.CurContext->isDependentContext()) {}
 
----------------
I would like to see this in call site and a comment to tell us that we don't want to evaluate constraints in dependent context.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D126907/new/

https://reviews.llvm.org/D126907