[MPlayer-dev-eng] ppc runtime-cpu-detection fails with gcc 3.3

[Rich Felker]

On Sun, Feb 26, 2006 at 01:40:09AM +0100, Balatoni Denes wrote:
> Hi!
> 
> szombat 25 február 2006 23.48-kor Rich Felker ezeket a bolcs gondolatokat 
> fogalmazta meg:
> > On Sat, Feb 25, 2006 at 10:25:55PM +0100, Luca Barbato wrote:
> > > > Then there's the possibility of ditching the C language extensions and
> > > > just writing the altivec code with asm() like the MMX guys do...
> > >
> > > No, please...
> >
> > I for one consider this a much more graceful approach. ASM is a
> > standard language, albeit cpu-specific. Apple's and GCC's hacks and
> > extensions on top of the C language are proprietary, incompatible, and
> > lock users into GCC in the long term or else force other compilers
> > that might compete to duplicate the same disgusting hacks.
> 
> What are you talking about ?? What standard is there for x86 for example ? Do 

I mean that the language is fixed, not necessarily the syntax for
writing it.

> you mean gas' at&t syntax or the intel syntax? And would using gas' at&t 
> syntax in your opinion lock us into gcc&gas as those are the only widely 
> known compiler using it?

No, because it's trivial to implement. A C compiler with all of gcc's
nonstandard quirks and extensions is extremely nontrivial to
implement.

Also there are at least a couple modern non-gas assemblers with at&t
syntax, and all the historical proprietary unix ones used it.

Rich