[FFmpeg-cvslog] r17409 - in trunk: configure libavcodec/Makefile

[Diego Biurrun]

On Wed, Feb 18, 2009 at 06:12:47PM -0800, Mike Melanson wrote:
> Diego Biurrun wrote:
> > On Wed, Feb 18, 2009 at 09:44:18AM +0000, Robert Swain wrote:
> >> 2009/2/18 Diego Biurrun <diego at biurrun.de>:
> >>> On Wed, Feb 18, 2009 at 01:00:05AM +0000, Carl Eugen Hoyos wrote:
> >>>> diego <subversion <at> mplayerhq.hu> writes:
> >>>>> Log:
> >>>>> Handle conditional compilation of CPU-specific FFT code in the Makefile.
> >>>>> Also replace one coarse MMX condition with more fine-grained
> >>>>> 3DNow!/3DNowExt!/SSE conditions.
> >>>> This did not just break MPlayer compilation, but, according to FATE, also Mac
> >>>> OSX and gcc 2.95.3.
> >>> Did proper bug reports go out of style so much?
> >>>
> >>> I need details and error messages.  It obviously works on my machines,
> >>> else I would not have committed..
> >> The point is, it's an automated testing system.
> > 
> > Carl is not an automated testing system.
> 
> No, but FATE is.

It would still be nice if the habit of providing useful information along
with bug reports was not lost in favor of "look at FATE".

> > Anyway, the build works fine on my box even with gcc 2.95.4
> > 
> > Looking at FATE I wonder why SSE gets disabled for gcc 2.95.4...
> 
> Try enabling it and you will probably find out. Make sure to include all 
> the options that FATE does.

I configured with

./configure --disable-debug --enable-gpl --enable-postproc --enable-avfilter --enable-avfilter-lavf --enable-shared --enable-swscale --enable-sse --cc=gcc-2.95

and the compilation succeeded.  What problem are you experiencing
exactly?

> BTW, it's only 2.95.3. I didn't know there was a 2.95.4. I don't even 
> want to think about how hard that will be to compile.

I think 2.95.4 is Debian-specific.

Diego