[Ffmpeg-cvslog] CVS: ffmpeg/tests libav.regression.ref, 1.91, 1.92 regression.sh, 1.93, 1.94

[Diego Biurrun]

On Fri, Mar 17, 2006 at 05:38:50PM +0100, Michael Niedermayer wrote:
> 
> On Fri, Mar 17, 2006 at 02:30:09PM +0100, Diego Biurrun wrote:
> > On Fri, Mar 17, 2006 at 02:35:13AM +0100, Michael Niedermayer wrote:
> > > 
> > > i think i said the nut regression test must not be removed/fail
> > > it will only make it harder to adapt nut later, as it might decay
> > > similar to ffserver ...
> > 
> > I understood that you did not want the regression tests to fail due to
> > disabling NUT.  They don't fail right now.  I removed the above lines
> > from libav.regression.ref because it is not possible to comment something
> > out in this file.  I thought that if somebody reenables NUT the numbers
> > will have to be updated anyway, so I might as well just remove them.
> > 
> > I can add a comment to the effect of "Don't forget to reenable
> > regression test when you reenable NUT." to libavformat/allformats.h,
> > this should take care of the danger of regression tests being forgotten.
> 
> do what you like in nut.c but dont modify libav.regression.ref, regression.sh
>  = please reverse this change = dont disable NUT = dont disable the NUT
> regression test
> 
> you can, if you like to add something like
> 
> if(filesize of file we have != filesize of the file used in the regression test){
>     av_log(..., "NUT is has not been updated to the latest spec and due to
>                  that has been disabled, thank you for using ffmpeg!\n");
>     return -1;
> }

Reenabled for now, I will look into this later.

Diego