[libc-commits] [libc] [libc] Add MPFR testing infra for float128. (PR #119499)

[Nick Desaulniers via libc-commits]

================
@@ -62,16 +61,15 @@ template <> struct make_real<cfloat16> {
   using type = float16;
 };
 #endif
-#if defined(LIBC_TYPES_HAS_CFLOAT128) &&                                       \
-    !defined(LIBC_TYPES_CFLOAT128_IS_COMPLEX_LONG_DOUBLE)
+#ifdef LIBC_TYPES_CFLOAT128_IS_NOT_COMPLEX_LONG_DOUBLE
 template <> struct make_real<cfloat128> {
   using type = float128;
 };
 #endif
 
 template <typename T> using make_real_t = typename make_real<T>::type;
 
-template <typename T> LIBC_INLINE constexpr T conjugate(T c) {
+template <typename T> LIBC_INLINE static constexpr T conjugate(T c) {
----------------
nickdesaulniers wrote:

I think more generally, we can probably remove this pattern throughout the codebase; I've been noticing that we're not super consistent.

https://github.com/llvm/llvm-project/pull/119499