[MPlayer-dev-eng] [PATCH] negating unsigned in asm code

Michael Niedermayer michaelni at gmx.at
Thu Jun 30 18:48:16 CEST 2005


Hi

On Thursday 30 June 2005 18:37, Rich Felker wrote:
> On Thu, Jun 30, 2005 at 11:31:09AM +0200, Michael Niedermayer wrote:
> > Hi
> >
> > On Thursday 30 June 2005 11:14, Guillaume POIRIER wrote:
> > > Hej,
> > >
> > > On 6/30/05, Reimar Döffinger <Reimar.Doeffinger at stud.uni-karlsruhe.de>
> >
> > wrote:
> > > > Hi,
> > > >
> > > > On Thu, Jun 30, 2005 at 10:56:04AM +0200, Diego Biurrun wrote:
> > > > > Why not go with the signed solution?  Seems cleaner to me...
> > > >
> > > > Yes, but unfortunately these things are all over the place, would be
> > > > a gigantic patch. I guess we might do it together with the int->long
> > > > change *g*
> > > > Any volunteers??
> > >
> > > Rich volunteered for the big s/int/long/ patch ! :-)
> >
> > +#define int long :)
>
> And this from someone who refused to allow #define inline ... ;)

did i refuse it? i dont remember, but anyway #define inline ... is 
problematic, IIRC gcc has a inline limit and always_inline is also counted, 
so if you change one function which gcc didnt inline from inline to 
always_inline then gcc will stop inlining something else, this also creates 
the illusion that code becomes slower if always_inline is added 

[...]
-- 
Michael




More information about the MPlayer-dev-eng mailing list