[libc-commits] [libc] [llvm] [libc][math] Refactor tanhf16 implementation to header-only in src/__support/math folder. (PR #178645)

[Muhammad Bassiouni via libc-commits]

================
@@ -118,6 +118,7 @@ TEST(LlvmLibcSharedMathTest, AllFloat) {
   EXPECT_FP_EQ(0.0f, LIBC_NAMESPACE::shared::sinf(0.0f));
   EXPECT_FP_EQ(0.0f, LIBC_NAMESPACE::shared::sqrtf(0.0f));
   EXPECT_FP_EQ(0.0f, LIBC_NAMESPACE::shared::tanf(0.0f));
+  EXPECT_FP_EQ(0.0f, LIBC_NAMESPACE::shared::tanhf16(0.0f16));
----------------
bassiounix wrote:

this should not be here, we have a dedicated section for `float16`
also both the param type and return type are `float16` not `float`

https://github.com/llvm/llvm-project/pull/178645