[PATCH] D71049: Remove implicit conversion that promotes half to other larger precision types for fp classification builtins

[Ulrich Weigand via Phabricator via cfe-commits]

uweigand added a comment.

This causes build bot failures on SystemZ due to a failed assertion in ConstantFP::getInfinity:
http://lab.llvm.org:8011/builders/clang-s390x-linux/builds/28723/steps/ninja%20check%201/logs/FAIL%3A%20Clang%3A%3Abuiltins.c


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D71049/new/

https://reviews.llvm.org/D71049