[FFmpeg-devel] [PATCH 1/2] avcodec/vda_h264: use multichar literal portably

[Hendrik Leppkes]

On Sat, Sep 19, 2015 at 12:35 AM, Michael Niedermayer <michaelni at gmx.at> wrote:
> On Fri, Sep 18, 2015 at 06:16:18PM -0400, Ganesh Ajjanagadde wrote:
>> Multichar literals are implementation defined, and thus trigger -Wmultichar:
>> http://fate.ffmpeg.org/log.cgi?time=20150918202532&log=compile&slot=x86_64-darwin-gcc-5.
>> http://www.zipcon.net/~swhite/docs/computers/languages/c_multi-char_const.html
>> gives a good summary of how to deal with them; in particular this patch
>> results in behavior identical to that generated by GCC (which I assume is correct this case).
>>
>> Signed-off-by: Ganesh Ajjanagadde <gajjanagadde at gmail.com>
>> ---
>>  libavcodec/vda_h264.c | 4 +++-
>>  1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/libavcodec/vda_h264.c b/libavcodec/vda_h264.c
>> index 8c526c0..3279afa 100644
>> --- a/libavcodec/vda_h264.c
>> +++ b/libavcodec/vda_h264.c
>> @@ -339,7 +339,9 @@ int ff_vda_default_init(AVCodecContext *avctx)
>>      CFMutableDictionaryRef buffer_attributes;
>>      CFMutableDictionaryRef io_surface_properties;
>>      CFNumberRef cv_pix_fmt;
>> -    int32_t fmt = 'avc1', pix_fmt = vda_ctx->cv_pix_fmt_type;
>> +#define LE_CHR(a,b,c,d) ( ((a)<<24) | ((b)<<16) | ((c)<<8) | (d) )
>> +    int32_t fmt = LE_CHR( 'a', 'v', 'c', '1');
>
> please use AV_RB32("avc1")
> its simpler and used elsewhere
>

IMHO even better is MKBETAG( 'a', 'v', 'c', '1')