Fix the PIX_FMT_* ---> IMGFMT_* conversion (PIX_FMT_RGB32 is IMGFMT_BGR32)

Originally committed as revision 19875 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

libswscale/swscale.c

@@ -221,7 +221,7 @@ static const int fmt_name[PIX_FMT_NB] = {
     [PIX_FMT_BGR24] = IMGFMT_BGR24,     ///< Packed pixel, 3 bytes per pixel, BGRBGR...
     [PIX_FMT_YUV422P] = IMGFMT_422P,   ///< Planar YUV 4:2:2 (1 Cr & Cb sample per 2x1 Y samples)
     [PIX_FMT_YUV444P] = IMGFMT_444P,   ///< Planar YUV 4:4:4 (1 Cr & Cb sample per 1x1 Y samples)
-    [PIX_FMT_RGBA32] = IMGFMT_RGB32,    ///< Packed pixel, 4 bytes per pixel, BGRABGRA..., stored in cpu endianness
+    [PIX_FMT_RGBA32] = IMGFMT_BGR32,    ///< Packed pixel, 4 bytes per pixel, BGRABGRA..., stored in cpu endianness
     [PIX_FMT_YUV410P] = IMGFMT_YVU9,   ///< Planar YUV 4:1:0 (1 Cr & Cb sample per 4x4 Y samples)
     [PIX_FMT_YUV411P] = IMGFMT_411P,   ///< Planar YUV 4:1:1 (1 Cr & Cb sample per 4x1 Y samples)
     [PIX_FMT_RGB565] = IMGFMT_RGB16,    ///< always stored in cpu endianness