[PATCH] D111453: Delay trailing constraing evaluation so we short-circuit

[Erich Keane via Phabricator via cfe-commits]

erichkeane added a comment.

The implementation seems to have tried checking this constraint 2x, and the first time required fully instantiating the constraint.  The only condition in the check is PartialOrdering here: https://github.com/llvm/llvm-project/blob/main/clang/lib/Sema/SemaTemplateDeduction.cpp#L3654

I wasn't able to get that trigger in any case where constraints mattered, and we don't fail any tests here, so I'm at a loss as to in which way this would not work.  I'm hoping to get feedback on the direction if it is wrong.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D111453/new/

https://reviews.llvm.org/D111453