[Ffmpeg-cvslog] CVS: ffmpeg/libavcodec/i386 snowdsp_mmx.c, 1.4, 1.5

[Guillaume POIRIER]

Hi,

On 4/18/06, Rich Felker <dalias at aerifal.cx> wrote:
> On Tue, Apr 18, 2006 at 05:52:01AM +0200, Guillaume POIRIER wrote:
> > Hi,
> >
> > On 4/18/06, Guillaume Poirier CVS <gpoirier at mplayerhq.hu> wrote:
> > > Update of /cvsroot/ffmpeg/ffmpeg/libavcodec/i386
> > > In directory mail:/var2/tmp/cvs-serv1059/i386
> > >
> > > Modified Files:
> > >         snowdsp_mmx.c
> > > Log Message:
> > > gcc-2.95 compile fix, patch by Wolfram Gloger <wmglo A dent PIS med PIS uni-muenchen PIS de>
> >
> > To be exact, gcc-2.95 doesn't compile it right, and the resulting
> > binary segfaults.
> > So it's more like a compiler bug work-around, but since we don't do
> > compiler workarounds, let's just call it a compile fix, shall we? ;-)
>
> I don't recall "rm" even being valid in gcc 2.95...

If it's not valid, GCC should yell louder IMHO.
If you run grep -r "\"rm\"" . on your source tree, you'll see that the
following files use "rm" in inline asm:

MPlayer:
./libmpcodecs/vf_spp.c
./libmpcodecs/vf_filmdint.c

FFmpeg:
./libavcodec/i386/snow_mmx_sse2.h
./libavcodec/i386/dsputil_h264_template_mmx.c
./libavcodec/i386/snowdsp_mmx.c

There's far less than I suspected, but nonetheless, they do exist, and
if they don't work, it means that they are really few gcc-2.95 users
these days, otherwise we should have heard about this problem sooner.

Rich, since you use gcc-2.95, could you check if these files are truly
miscompiled?

Guillaume
--
I am disillusioned enough to know that no man's opinion on any subject
is worth a damn unless backed up with enough genuine information to
make him really know what he's talking about.

-- H. P. Lovecraft (about the flamewars on FFmpeg and MPlayer-dev mailing lists)
http://www.brainyquote.com/quotes/quotes/h/hplovecr278144.html