[PATCH] D46535: Correct warning on Float->Integer conversions.

Eli Friedman via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Wed Oct 3 12:34:50 PDT 2018


efriedma added a comment.

> Every Integer is representable (lossy of course) as a float as far as I know.

Casting a __uint128_t to float can overflow.  And overflowing a _Float16 is easy.  (Of course, both __uint128_t and _Float16 are rare in normal C/C++ code.)


Repository:
  rC Clang

https://reviews.llvm.org/D46535





More information about the cfe-commits mailing list