[FFmpeg-devel] [PATCH] Swscale YUV2RGB table generator

[Kostya]

On Sun, Feb 22, 2009 at 11:06:39AM +0000, Robert Swain wrote:
> 2009/2/22 Kostya <kostya.shishkov at gmail.com>:
> > On Sun, Feb 22, 2009 at 11:50:50AM +0100, Diego Biurrun wrote:
> >> On Sun, Feb 22, 2009 at 12:20:16PM +0200, Kostya wrote:
> >> > On Thu, Feb 12, 2009 at 11:40:59PM +0100, Michael Niedermayer wrote:
> >> > > On Thu, Feb 12, 2009 at 10:26:21AM +0200, Kostya wrote:
> >> > [...]
> >> > > > > maybe a few const could be added?
> >> > > > >
> >> > > > > ok except these
> >> > > >
> >> > > > Since it's not wholly identical with the old one, regressions are changed too.
> >> > > > Attached an example of mine obtained at PPC.
> >> > >
> >> > > the regression change looks harmless after a quick look ...
> >> >
> >> > I took the liberty to apply it after all.
> >>
> >> yuv2rgb.c is not compiled but still around.  IMO you should just have
> >> replaced one file with the other and not add a new one with a 2 suffix.
> >
> > I could do that but I'm unsure of consequences to SVN history - this is a new
> > implementation, not some patch to the old file after all.
> 
> Is there anything else that's GPL that must be present for swscale to
> function properly? Can the ASM just be disabled using some ifdef based
> on GPL and the available instruction sets? Indeed the question
> remains, what needs to be done? :)

I think now only MMX and AltiVec code is GPLed (and it's #if'ed out in that case).
Also I have some LGPLed builds with swscale (at work) - seems to work properly.
 
> Nice work Kostya! :D

Now only to deal with the patch that gives Benjamin some pics along with IMC audio
(aka i263 PB-frame support)
 
> Regards,
> Rob