[MPlayer-dev-eng] [PATCH] makes --enable-*-faad behave correctly

Aurelien Jacobs aurel at gnuage.org
Wed Feb 2 16:40:47 CET 2005


On Wed, 02 Feb 2005 08:43:50 +0100
Nico Sabbi <nsabbi at tiscali.it> wrote:

> Aurelien Jacobs wrote:
> 
> >Hi,
> >
> >Currently when using --enable-*-faad, configure still try to compile
> >and execute some code to verify if the right version of faad is
> >available.
> >This patch makes enable really mean enable without any test, to be
> >consistent (and also to ease cross-compilation).
> >
> >  
> >
> 
> I'm against it: faad is very prone to break. If the external library
> doesn't compile (now or in the future) mplayer won't compile either
> because of the removal of this check.

What ? I don't understand what you mean exactly.

If external library doesn't compile, it's not related to MPlayer...
And mplayer won't detect any external faad so won't use it.

Or you mean if external faad API changes and that ad_faad don't
compile anymore ?
My patch modify the test code used when _faad_external=auto, so that
it try to compile faacDecInit() (which argument number changed during
the API change) so that it will fail on older versions of faad2 and
thus disable faad on those broken versions.
If a new API change is introduced, we will anyway have to modify the
configure to take care of the new version.

So the removal of the version check only disable support of very old
faad2 versions but when _faad_external=auto we still verify the the
version found follow the new API.

So what's exactly the problem with that ?

Aurel




More information about the MPlayer-dev-eng mailing list