[FFmpeg-devel] [PATCH 6/8] avcodec/decode: Do not overwrite AVFrame.pkt_pos if its already set

[James Almer]

On 8/14/2019 2:39 PM, Michael Niedermayer wrote:
> On Mon, Aug 12, 2019 at 09:03:22PM -0300, James Almer wrote:
>> On 8/12/2019 5:31 PM, Michael Niedermayer wrote:
>>> On Mon, Aug 12, 2019 at 09:21:59PM +0200, Vittorio Giovara wrote:
>>>> On Mon, Aug 12, 2019 at 9:19 PM Michael Niedermayer <michael at niedermayer.cc>
>>>> wrote:
>>>>
>>>>> Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
>>>>> ---
>>>>>  libavcodec/decode.c | 2 +-
>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/libavcodec/decode.c b/libavcodec/decode.c
>>>>> index 6c31166ec2..09a509d659 100644
>>>>> --- a/libavcodec/decode.c
>>>>> +++ b/libavcodec/decode.c
>>>>> @@ -435,7 +435,7 @@ static int decode_simple_internal(AVCodecContext
>>>>> *avctx, AVFrame *frame)
>>>>>          if (!(avctx->codec->caps_internal & FF_CODEC_CAP_SETS_PKT_DTS))
>>>>>              frame->pkt_dts = pkt->dts;
>>>>>          if (avctx->codec->type == AVMEDIA_TYPE_VIDEO) {
>>>>> -            if(!avctx->has_b_frames)
>>>>> +            if(!avctx->has_b_frames && frame->pkt_pos < 0)
>>>>>                  frame->pkt_pos = pkt->pos;
>>>>>              //FIXME these should be under if(!avctx->has_b_frames)
>>>>>              /* get_buffer is supposed to set frame parameters */
>>>>> --
>>>>>
>>>>
>>>> sure but why?
>>>
>>> the next 2 patches need this
>>> [FFmpeg-devel] [PATCH 7/8] avcodec/qtrle: return last frame even if unchanged
>>> [FFmpeg-devel] [PATCH 8/8] avcodec/nuv: Avoid duplicating frames
>>>
>>> without it the common code overwrites the pkt_pos set by the decoder
>>
>> What about rawdec.c? It's also setting frame->pkt_pos.
> 
> That looks redundant (will send patch after testing)
> 
> 
>>
>> And cant' this be handled with a new internal codec cap like the pkt_dts
>> one used above?
> 
> yes, if thats preferred ?

I think it would be better, but not sure if others agree.

A well defined way for decoders/encoders to inform the generic code what
to expect is imo nicer than it assuming that a given value in a frame
means that it was not touched, even if said value is the default.
Perhaps it could be simplified into a single codec cap to sell all pkt_*
fields and other props (like size and duration), rather than one cap per
field. And since it's all internal, it can be removed as easily and
quickly as it was added.

> 
> Thanks
> 
> [...]
> 
> 
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> 
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>