fp.c: Suppress float-conversion warnings on FreeBSD.

We used to do this on Windows only, but it appears to affect
multiple platforms when building with certain versions of GCC, and a
common pattern for defining the floating-point classifier functions.

Fixes part of 31687. I'm calling this a bugfux on 31687, when we
started suppressing these warnings on Windows.

changes/ticket31687_1

@@ -0,0 +1,4 @@
+  o Minor bugfixes (compilation):
+    - Suppress spurious float-conversion warnings from GCC when calling
+      floating-point classifier functions on FreeBSD.  Fixes part of bug
+      31687; bugfix on 0.3.1.5-alpha.

src/lib/math/fp.c

@@ -62,12 +62,16 @@ clamp_double_to_int64(double number)
 {
   int exponent;
 
-#if defined(MINGW_ANY) && GCC_VERSION >= 409
+#if (defined(MINGW_ANY)||defined(__FreeBSD__)) && GCC_VERSION >= 409
 /*
   Mingw's math.h uses gcc's __builtin_choose_expr() facility to declare
   isnan, isfinite, and signbit.  But as implemented in at least some
   versions of gcc, __builtin_choose_expr() can generate type warnings
   even from branches that are not taken.  So, suppress those warnings.
+
+  FreeBSD's math.h uses an __fp_type_select() macro, which dispatches
+  based on sizeof -- again, this can generate type warnings from
+  branches that are not taken.
 */
 #define PROBLEMATIC_FLOAT_CONVERSION_WARNING
 DISABLE_GCC_WARNING(float-conversion)