[FFmpeg-devel] [PATCH] avcodec/utils: use a minimum 32pixel width in avcodec_align_dimensions2() for H.264

[Michael Niedermayer]

On Fri, Jul 10, 2015 at 10:51:45AM -0400, Ronald S. Bultje wrote:
> Hi,
> 
> On Thu, Jul 9, 2015 at 8:11 PM, Michael Niedermayer <michaelni at gmx.at>
> wrote:
> 
> > Fixes Assertion failure
> > Found-by: Andreas Cadhalpun <andreas.cadhalpun at googlemail.com>
> > Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
> > ---
> >  libavcodec/utils.c |    4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/libavcodec/utils.c b/libavcodec/utils.c
> > index 96f177d..925b3ed 100644
> > --- a/libavcodec/utils.c
> > +++ b/libavcodec/utils.c
> > @@ -428,10 +428,12 @@ void avcodec_align_dimensions2(AVCodecContext *s,
> > int *width, int *height,
> >
> >      *width  = FFALIGN(*width, w_align);
> >      *height = FFALIGN(*height, h_align);
> > -    if (s->codec_id == AV_CODEC_ID_H264 || s->lowres)
> > +    if (s->codec_id == AV_CODEC_ID_H264 || s->lowres) {
> >          // some of the optimized chroma MC reads one line too much
> >          // which is also done in mpeg decoders with lowres > 0
> >          *height += 2;
> > +        *width = FFMAX(*width, 32);
> > +    }
> 
> 
> Would you mind adding a comment why? This codec-specific hackery tends to
> raise lots of eyebrows a few years after, so knowing what exactly it fixes
> is very useful.
> 
> (E.g. what assertion, caused by what file with what special feature that
> technically cause the assert, etc.)

comment added in a moment

thanks

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

No snowflake in an avalanche ever feels responsible. -- Voltaire
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20150710/355a7dfa/attachment.sig>