Change packet_size from signed to unsigned, since it never has a negative

value. Currently, this is only used in qcp.c and mpegenc.c, and they are
practically unaffected by this change.

See "[PATCH] make packet_size in AVFormatContext unsigned" thread on ML.

Originally committed as revision 19250 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavformat/avformat.h

@@ -23,7 +23,7 @@
 
 #define LIBAVFORMAT_VERSION_MAJOR 52
 #define LIBAVFORMAT_VERSION_MINOR 34
-#define LIBAVFORMAT_VERSION_MICRO  0
+#define LIBAVFORMAT_VERSION_MICRO  1
 
 #define LIBAVFORMAT_VERSION_INT AV_VERSION_INT(LIBAVFORMAT_VERSION_MAJOR, \
                                                LIBAVFORMAT_VERSION_MINOR, \
@@ -553,7 +553,7 @@ typedef struct AVFormatContext {
     int index_built;
 
     int mux_rate;
-    int packet_size;
+    unsigned int packet_size;
     int preload;
     int max_delay;