[FFmpeg-devel] [PATCH 1/2] avformat/mux: Warn if the muxers bitexact flag is not set but it looks as if the user wants it set

Andreas Cadhalpun andreas.cadhalpun at googlemail.com
Sun Aug 23 23:53:36 CEST 2015


Hi,

On 23.08.2015 13:10, Ronald S. Bultje wrote:
> On Sun, Aug 23, 2015 at 5:52 AM, Andreas Cadhalpun <
> andreas.cadhalpun at googlemail.com> wrote:
> 
>> On 23.08.2015 11:09, Michael Niedermayer wrote:
>>> From: Michael Niedermayer <michael at niedermayer.cc>
>>>
>>> Note: I doubt requiring this in the future is a good idea
>>> See: [FFmpeg-devel] [PATCH 3/4] fate: add -fflags +bitexact to the
>> relevant targets
>>> Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
>>> ---
>>>  libavformat/mux.c |    7 ++++++-
>>>  1 file changed, 6 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/libavformat/mux.c b/libavformat/mux.c
>>> index 81c4676..77a24cc 100644
>>> --- a/libavformat/mux.c
>>> +++ b/libavformat/mux.c
>>> @@ -250,10 +250,15 @@ static int init_muxer(AVFormatContext *s,
>> AVDictionary **options)
>>>          (ret = av_opt_set_dict2(s->priv_data, &tmp,
>> AV_OPT_SEARCH_CHILDREN)) < 0)
>>>          goto fail;
>>>
>>> +    if (s->nb_streams && s->streams[0]->codec->flags &
>> AV_CODEC_FLAG_BITEXACT) {
>>> +        if (!(s->flags & AVFMT_FLAG_BITEXACT))
>>> +            av_log(s, AV_LOG_WARNING, "Muxer bitexact flag is not set,
>> please set AVFormatContext.flags |= AVFMT_FLAG_BITEXACT.\n"
>>> +                                      "This will become mandatory with
>> future API cleanup\n"
>>> +            );
>>
>> Such a warning is probably a good idea.
>> However, either it should be removed together with FF_API_LAVF_BITEXACT
>> (which only makes sense if we keep FF_API_LAVF_BITEXACT until next time)
>> or the warning has to change, when FF_API_LAVF_BITEXACT is disabled.
>> Saying 'future API cleanup', when the cleanup already happened seems wrong.
> 
> 
> See FF_API_OLD_FILTER_OPTS_ERROR, is probably a good template.

Yes, doing it similar to that case would be good.

> I'm fine with this, but what do you think of a utility -bitexact option
> that sets all bitexact flags? Most fate test commands can be halved if we
> use that. But your choice.

I think such a -bitexact option would be great for the ffmpeg program,
because it would make many of the fate command lines a lot more readable.
Now someone just has to implement it. ;)
But nonetheless this warning would make sense for API users.

Best regards,
Andreas


More information about the ffmpeg-devel mailing list