[FFmpeg-devel] [PATCH] Interplay Video decoder 16-bit support

[Kostya]

On Thu, Dec 24, 2009 at 05:15:08PM +0100, Michael Niedermayer wrote:
> On Wed, Dec 23, 2009 at 09:56:29PM +0200, Kostya wrote:
> > Now this is done as extending existing decoder without introducing
> > record-breaking long codec ID and new decoder.
> > 
> > I've tried to follow current formatting conventions, but there's still
> > some old formatting left to fix in that file.
 
[...] 
> > +#define CHECK_MV_PTR(n) \
> > +    if (s->mv_end - s->mv_ptr < n) { \
> > +        av_log(s->avctx, AV_LOG_ERROR, "Interplay video warning: mv_ptr out of bounds (%p >= %p)\n", \
> > +               s->mv_ptr + n, s->mv_end); \
> > +        return -1; \
> > +    }
> 
> duplicate of CHECK_STREAM_PTR
> and there might be more stuff that an be merged
 
Merged.
 
[...]
> > @@ -95,7 +107,8 @@
> >              motion_offset, s->upper_motion_limit_offset);
> >          return -1;
> >      }
> > -    s->dsp.put_pixels_tab[1][0](s->pixel_ptr, src->data[0] + motion_offset, s->stride, 8);
> > +    s->dsp.put_pixels_tab[!s->is_16bpp][0](s->pixel_ptr, src->data[0] + motion_offset,
> > +                                           s->current_frame.linesize[0], 8);
> 
> the switch from stride ->linesize should possibly be seperate

It was.

> [...]
> 
> above is just quick review, ill try to do a more complete one when i have
> more time

Well, it turns out there is a maintainer for that file who is different
Michael than you and I don't think he'll mind having this stuff
committed. In any case I've tried to break it into clean independent
parts committed separately, so it's easier to review for you if you like.

> -- 
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB