[libc-commits] [libc] [libc][math][c23] Temporarily disable float16 on 32-bit Arm (PR #95027)

[via libc-commits]

overmighty wrote:

@nickdesaulniers We have a few ideas that aren't incompatible:

- Writing our own functions to convert between `_Float16` and higher-precision types with bit manipulation, to not depend on the functions from the compiler runtime.
- Using inline assembly for targets that support instructions for converting between `_Float16` and higher-precision types, to not depend on the functions from the compiler runtime.
- Linking with the latest compiler-rt (I haven't yet confirmed that it fixes the issue). It could be an option for the runtime build.


https://github.com/llvm/llvm-project/pull/95027