[FFmpeg-devel] [FFmpeg-devel-irc] IRC log for 2010-09-30

[Diego Biurrun]

On Sun, Oct 03, 2010 at 05:14:59PM +0200, Michael Niedermayer wrote:
> On Sun, Oct 03, 2010 at 02:37:55PM +0200, Diego Biurrun wrote:
> > On Sat, Oct 02, 2010 at 06:27:41PM +0200, Michael Niedermayer wrote:
> > > On Sat, Oct 02, 2010 at 05:48:48PM +0200, Diego Biurrun wrote:
> > > > On Fri, Oct 01, 2010 at 11:17:01PM +0200, Michael Niedermayer wrote:
> > > > > On Fri, Oct 01, 2010 at 04:25:32PM -0400, compn wrote:
> > > > > > 
> > > > > > if this broke compilation somewhere it'd be nice for someone to paste
> > > > > > it so that we can end this dumb argument quickly.
> > > > > > 
> > > > > > only thing i can find is that it broke linking in mplayer
> > > > > > [23:55:29] <Dark_Shikari> 09:55 < ubitux> revision 25278 has introduced
> > > > > > some linkage error: ffmpeg/libavutil/libavutil.a(mathematics.o): In
> > > > > > function `av_rescale_rnd' << undefined reference to `assert'
> > > > > 
> > > > > This seems a bug in the MPlayer build system
> > > > > it builds libavutil from within the directory and uses -I.
> > > > > and without -I- that way the local directory will be used to search for
> > > > > #include <...> headers
> > > > > this seems quite incorrect behavior to me
> > > > > 
> > > > > diego?
> > > > 
> > > > Is this still a problem?  What is '-I-'?
> > > 
> > > "problem" in terms of currently breaking maybe not but it can break again
> > 
> > It seems this was fixed by the header rename.
> 
> no that was a workaround not a fix, the mplayer -I. is wrong if iam not
> mistaken it shouldnt be there.
> This can easily break again ...

See the rest of my patch to use FFmpeg as a single svn:externals property
and employ FFmpeg's own build system instead of duplicating it forever.
It will address this issue once fully applied.

Diego