[MPlayer-dev-eng] [PATCH] configure fixes: x11 header dir always included, better CFLAGS for unknown compilers

Michael Niedermayer michaelni at gmx.at
Wed Dec 6 17:13:37 CET 2006


Hi

On Wed, Dec 06, 2006 at 10:43:58AM +0100, Reimar Doeffinger wrote:
> Hello,
> On Tue, Dec 05, 2006 at 10:12:39PM +0100, Dominik 'Rathann' Mierzejewski wrote:
> > On Tuesday, 05 December 2006 at 18:33, Reimar Döffinger wrote:
> > > please say what you think.
> > > First part sets
> > > CFLAGS="-O2 $_march $_mcpu $_pipe"
> > > for non-intel, non-gnu compilers. None are supported so it might be
> > > pointless, but on the other hand it helps with trying some things out.
> > 
> > Just "-O2" is sufficient and doesn't break things. non-intel non-gnu compilers
> > probably don't support -march/-mcpu/-mtune switches.
> 
> But IIRC configure tests if they are available, so they will not hurt
> either. Correct me if I'm wrong.
> 
> > > Maybe a detection for -ffast-math, -fomit-frame-pointer, and -O4 would
> > > be better, but also more complicated.
> > 
> > Hm. I'm tempted to try out ICC. Is there any non-intel non-gnu compiler
> > that can compile mplayer?
> 
> I do not know any non-intel non-gnu one that would even support the
> variable-length macros MPlayer needs :-( (I assume that Apple's compiler
> counts as gnu compiler).
> At least linux ICC does not support __attribute__((used)) correctly so

hmm, what happens if you drop all __attribute__((used)) ? or replace them
by volatile? if either works these should be replaced by attribute_used or
so in svn IMHO

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

If you really think that XML is the answer, then you definitly missunderstood
the question -- Attila Kinali



More information about the MPlayer-dev-eng mailing list