[PATCH] D53486: [libcxx] Only define __libcpp_is_floating_point<_Float16> for Clang

[Petr Hosek via Phabricator]

phosek added a comment.

In https://reviews.llvm.org/D53486#1272388, @ldionne wrote:

> Thinking about this more, I'm fine with either this patch or one that does not define `__libcpp_is_floating_point<_Float16>` at all. This patch provides the best level of support that we seem to be able to provide, but is inconsistent across compilers and does not follow the standard perfectly. Not providing the definition at all decreases the level of support but increases consistency and simplicity. I'm fine with either (with a small preference towards consistency and simplicity, as a lazy maintainer :-)


A possible reason to go with the current option is `__libcpp_is_floating_point<_Float16>` that has been in libc++ for a few months and has been shipped with 7.0, so there might already be users depending on this trait. Is that a concern? If yes, then current version is probably the right way to go.


Repository:
  rCXX libc++

https://reviews.llvm.org/D53486