Include float.h to provide FLT_MAX define. Should fix compilation on windows.

Originally committed as revision 23580 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavcodec/ra144enc.c

@@ -26,6 +26,7 @@
  */
 
 #include <values.h>
+#include <float.h>
 
 #include "avcodec.h"
 #include "put_bits.h"