[MPlayer-cvslog] r30971 - trunk/cpudetect.c

[Diego Biurrun]

On Sun, Mar 28, 2010 at 10:58:43AM +0200, Reimar Döffinger wrote:
> On Sun, Mar 28, 2010 at 05:43:44AM +0200, Diego Biurrun wrote:
> > On Sat, Mar 27, 2010 at 08:45:59PM +0100, reimar wrote:
> > > 
> > > Log:
> > > Leave disabling SSE to the check_os_katmai_support always regardless of OS,
> > > it can already handle all cases, will always print a proper warning and
> > > probably fixes SSE being disabled on x86_64 Solaris systems.
> > > 
> > > --- trunk/cpudetect.c	Sat Mar 27 17:01:22 2010	(r30970)
> > > +++ trunk/cpudetect.c	Sat Mar 27 20:45:59 2010	(r30971)
> > > @@ -195,18 +195,10 @@ void GetCpuCaps( CpuCaps *caps)
> > >  
> > >  		/* FIXME: Does SSE2 need more OS support, too? */
> > > -#if defined(__linux__) || defined(__FreeBSD__) || defined(__FreeBSD_kernel__) \
> > > -  || defined(__NetBSD__) || defined(__OpenBSD__) || defined(__DragonFly__) \
> > > -  || defined(__APPLE__) || defined(__CYGWIN__) || defined(__MINGW32__) \
> > > -  || defined(__OS2__)
> > >  		if (caps->hasSSE)
> > >  			check_os_katmai_support();
> > >  		if (!caps->hasSSE)
> > >  			caps->hasSSE2 = 0;
> > > -#else
> > > -		caps->hasSSE=0;
> > > -		caps->hasSSE2 = 0;
> > > -#endif
> > 
> > Nice, this was incredibly ugly...
> 
> More annoyingly it was wrong. I just wonder why it was like that, I suspect
> _something_ might break, I guess we'll find out.

Whatever breaks, configure must test for it, this #ifdef madness must stop.

Diego