[MPlayer-dev-eng] Patch for libmpdemux/demux_lavf.c, (revision 25137), ByteIOContext pointer issue

Ivan Kalvachev ikalvachev at gmail.com
Fri Nov 23 12:04:53 CET 2007


2007/11/23, Michael Niedermayer <michaelni at gmx.at>:
> On Thu, Nov 22, 2007 at 04:08:45PM +0200, Ivan Kalvachev wrote:
> > 2007/11/22, Chris Welton <electrostatic_1 at yahoo.com>:
> > > This fixes the ByteIOContext pointer issue in
> > > libmpdemux/demux_lavf.c so flash files will once again
> > > play.
> > >
> > > There is also a rather spurious cast I ended up with
> > > while debugging that could probably be cleaned up
> > > better...  (I had no idea about the pointer issue when
> > > I started so I just went down the list with the
> > > warnings in demux_lavf.c until there were no more
> > > warnigns and everything worked)
> >
> >
> > I usually stay away from (de)muxers, but yesterday I had to quick fix
> > build error caused by the ByteIOContext change.
> > However my fix wasn't complete, so now I attach patch for muxer_lavf.c
> > similar to the above patch. I hope I haven't missed something else.
>
> ive not checked it too careful but it looks better with than without the
> patch, so apply!

Applied
both mine and Chris patch,

I just removed the warning fix about the const char pointer in the
Chris patch, as it was wrong and not related to the API change fix.
I'll commit the right fix in a little bit. It should looks like this:
@@ -248,4 +248,4 @@
-        char **p = preferred_list;
+        const char **p = preferred_list;

There are other warning, that should be easy to fix. I probably will
send separate patches for them.

BTW, in muxer_lavf,c the structure muxer_priv_t contains ByteIOContext
*pb, int audio/video_streams, int64_t last_pos, but none of them is
used. Are these obsolete left overs or they are planned for future
extensions?



More information about the MPlayer-dev-eng mailing list