[Ffmpeg-devel] moving non-SIMD parts of libswscale to LGPL

[Michael Niedermayer]

Hi

On Wed, Dec 20, 2006 at 10:47:23AM +0100, Luca Abeni wrote:
> Hi Michael,
> 
> On Tue, 2006-12-19 at 17:59 +0100, Michael Niedermayer wrote:
> [...]
> > > > ok if regression tests pass
> > > Well, we don't have regression tests for libswscale (I mean: when ffmpeg
> > > is configured with --enable-swscaler, regression tests always fail
> > > because the reference values are for the "old" libavcodec's scaling and
> > > conversion code). Or did I misunderstand you?
> > 
> > no i was just stupid, forget it, the codes ok commit it
> Ok, good. I am going to commit it shortly. Just one last question (I
> want to be really sure to do everything properly): non-MMX compilation
> of swscale_template.c with gcc 2.95 is broken. The patch I posted some
> time ago contained something like this to fix it:
> Index: swscale_template.c
> ===================================================================
> --- swscale_template.c  (revision 21694)
> +++ swscale_template.c  (working copy)
> @@ -1754,12 +1754,14 @@
>                 : "%"REG_a
>         );
>  #else
> +       {
>         int i;
>         for(i=0; i<width; i++)
>         {
>                 dstU[i]= src1[4*i + 1];
>                 dstV[i]= src1[4*i + 3];
>         }
> +       }
>  #endif
>  }
> [...]
> I plan to commit this before the rest of the patch, then to fix
> indentation in a separate commit, and then to commit the rest of the
> patch. Is this ok? Or do you prefer a different fix, such as moving "int
> i;" at the beginning of the function (in this case, gcc will produce a
> warning if HAVE_MMX is defined)?

i prefer a fix without extra {}

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

I am the wisest man alive, for I know one thing, and that is that I know
nothing. -- Socrates