Use endianness definitions provided by GCC or Clang if available.

Instead of using all sorts of operating system specific constructs, we
can just query the compiler which byte order is being used. This has the
advantage that the code builds on new platforms without any tweaks.

include/fpmath.h

@@ -38,28 +38,37 @@
 #endif
 #endif
 
-#ifdef __linux
+#if defined(__BYTE_ORDER__) && defined(__ORDER_LITTLE_ENDIAN__)
+
+/* Definitions provided directly by GCC and Clang. */
+#define _LITTLE_ENDIAN    __ORDER_LITTLE_ENDIAN__
+#define _BIG_ENDIAN       __ORDER_BIG_ENDIAN__
+#define _PDP_ENDIAN       __ORDER_PDP_ENDIAN__
+#define _BYTE_ORDER       __BYTE_ORDER__
+
+#elif defined(__linux)
+
 #include <features.h>
 #include <endian.h>
 #define _LITTLE_ENDIAN  __LITTLE_ENDIAN
 #define _BIG_ENDIAN     __BIG_ENDIAN
 #define _PDP_ENDIAN     __PDP_ENDIAN
 #define _BYTE_ORDER     __BYTE_ORDER
-#endif
 
-#ifdef __APPLE__
+#elif defined(__APPLE__)
+
 #include <machine/endian.h>
 #define _LITTLE_ENDIAN  LITTLE_ENDIAN
 #define _BIG_ENDIAN     BIG_ENDIAN
 #define _PDP_ENDIAN     PDP_ENDIAN
 #define _BYTE_ORDER     BYTE_ORDER
-#endif
 
-#ifdef __FreeBSD__
+#elif defined(__FreeBSD__)
+
 #include <machine/endian.h>
-#endif
 
-#ifdef _WIN32
+#elif defined(_WIN32)
+
 #define _LITTLE_ENDIAN 1234
 #define _BIG_ENDIAN    4321
 #define _PDP_ENDIAN    3412
@@ -69,6 +78,7 @@
 #define BIG_ENDIAN     _BIG_ENDIAN
 #define PDP_ENDIAN     _PDP_ENDIAN
 #define BYTE_ORDER     _BYTE_ORDER
+
 #endif
 
 #ifndef _IEEE_WORD_ORDER