[FFmpeg-devel] [PATCH] Video decoder and demuxer for AMV files

[Vladimir Voroshilov]

Now i undestand.
Thanks you for detailed explamation.
I have another question.


According to this:

> > > use SEPERATE variables do NOT hack s->picture.data and s->linesize and then
> > > revert that again a few lines later
> > >

[...]

..and this:
> > Did i undestand you right (see patch)?
>
> argh, no!
>
> int linesize= s->linesize;
> uint8_t *data[3]= {s->picture.data[0], ...
>
> if(amv){
>     linesize= -linesize;
>     for(i=0; i<3; i++)
>         data[i]= ...
> }
>
> use linesize+data
>
> and the patch which introduces the new variables should be seperate
> from the one adding amv support

attached implementation is also not acceptable (and worse than hacking
mjpeg_decode_scan).
Am i right? Should i continue hacking mjpeg_decode_scan as you
suggested or improve attached patch?

-- 
Regards,
Vladimir Voroshilov     mailto:voroshil at gmail.com
JID: voroshil at gmail.com, voroshil at jabber.ru
ICQ: 95587719
-------------- next part --------------
A non-text attachment was scrubbed...
Name: amv_mjpeg_decoder.diff
Type: text/x-diff
Size: 6795 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20070929/f88a31fa/attachment.diff>