[Ffmpeg-devel] swscale and regression tests

[Michael Niedermayer]

Hi

On Wed, Aug 16, 2006 at 09:36:16AM +0200, Luca Abeni wrote:
> Hi Michael,
> 
> On Mon, 2006-08-14 at 14:02 +0200, Michael Niedermayer wrote:
> [...]
> > > So, how to fix "make test" with swscale? I can think about 2 possible
> > > solutions, but there are probably more:
> > > 1) leave the libavocodec scaling and format conversion routines in
> > > place, and always use them in the tests (even if swscale is enabled)
> > > 2) use different *.ref files whe swscale is enabled
> > 
> > could we make old and sws scaling runtime switchable?
> Yes, of course :) I just need to sit down and think a little bit about
> the best way to do this.
> 
> Currently, when swscale is enabled the ffmpeg rescaling and format
> conversion routines are not compiled. I did it this way because I had
> the impression that there was the will to have swscale as a replacement
> to ffmpeg's routines (if I remember well, this whole story begun because
> someone complained about ffmpeg's routines).
> But maybe I misunderstood; I'll find a way to make the two
> implementation coexist.

wait, thinking about it again, theres really no sense in compiling with
both enabled, what rather should be done is
* always use swscale but disable the optimized scalers if CONFIG_GPL is not
  set
* add a BITEXACT flag to disable optimized code which isnt binary identical
  to the c counterparts in swscale
* make all formats which work currently work in swscale
* update the regression test checksums to the swscale case

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

In the past you could go to a library and read, borrow or copy any book
Today you'd get arrested for mere telling someone where the library is