[libc-commits] [libc] [libc][math][c23] Add {ldexp, scalbn, scalbln}f16 C23 math functions (PR #94797)

[via libc-commits]

================
@@ -183,6 +184,10 @@ template <size_t Bits> struct DyadicFloat {
     return r;
   }
 
+  LIBC_INLINE explicit constexpr operator float16() const {
+    return static_cast<float16>(static_cast<float>(*this));
+  }
----------------
overmighty wrote:

The smoke test for `ldexpf16` showed that `operator T()` defined above did not always give correct results for `float16`. At least this part did not: https://github.com/llvm/llvm-project/blob/main/libc/src/__support/FPUtil/dyadic_float.h#L160-L161.

Maybe we could have an exhaustive test for `ldexpf16`.

(Comment moved from #94758.)

https://github.com/llvm/llvm-project/pull/94797