[FFmpeg-devel] [PATCH 1/3] avformat/mxfdec: do not ignore bad size errors

[Marton Balint]

On Thu, 22 Aug 2019, Tomas Härdin wrote:

> ons 2019-08-21 klockan 21:54 +0200 skrev Marton Balint:
>> 
>> On Sun, 18 Aug 2019, Tomas Härdin wrote:
>> 
>> > lör 2019-08-17 klockan 21:41 +0200 skrev Marton Balint:
>> > > The return value was unintentionally lost after
>> > > 00a2652df3bf25a27d174cc67ed508b5317cb115.
>> > > 
>> > > Signed-off-by: Marton Balint <cus at passwd.hu>
>> > > ---
>> > >  libavformat/mxfdec.c | 4 ++--
>> > >  1 file changed, 2 insertions(+), 2 deletions(-)
>> > > 
>> > > diff --git a/libavformat/mxfdec.c b/libavformat/mxfdec.c
>> > > index bb72fb9841..397f820b3f 100644
>> > > --- a/libavformat/mxfdec.c
>> > > +++ b/libavformat/mxfdec.c
>> > > @@ -3508,8 +3508,8 @@ static int mxf_read_packet(AVFormatContext *s, AVPacket *pkt)
>> > >                  } else {
>> > >                      if ((size = next_ofs - pos) <= 0) {
>> > >                          av_log(s, AV_LOG_ERROR, "bad size: %"PRId64"\n", size);
>> > > -                        ret = AVERROR_INVALIDDATA;
>> > > -                        goto skip;
>> > > +                        mxf->current_klv_data = (KLVPacket){{0}};
>> > > +                        return AVERROR_INVALIDDATA;
>> > 
>> > Should maybe ask for a sample. Else looks OK
>> 
>> I don't think this can happen with any of the valid files, so the text 
>> saying that ffmpeg is missing a feature probably won't be true. It is a 
>> lot more likely that the user has a broken file if this error is shown.
>> 
>> So I'd rather keep it as is.
>
> Right, the skip had me confused. This is fine of course

Thanks, applied.

Regards,
Marton