[MPlayer-dev-eng] [REPOST] [PATCH] Mac OS X fixes + Darwin AltiVec detection

Dan Christiansen danchr at daimi.au.dk
Mon Apr 7 18:39:50 CEST 2003


On Wednesday, April 2, 2003, at 03:32 PM, Romain Dolbeau wrote:

> Dan Christiansen wrote:
>> This seems to have slipped through. I believe it fixes the issues 
>> Alex Beregszaszi mentioned. I wouldn't recommend applying the cpu 
>> detection patch until someone verified whether it's any good or not.
>
> Why detect the _presence_ of altivec at configure time ?
> The current way (if it can be compiled, compile it,
> and detect at runtime) seems better. One binary will
> work for everyone.

Indeed, if it is the current behaviour, it would be better. I couldn't 
find anything which would enable AltiVec on Darwin, however. The only 
thing I could find was based on cpuinfo, which in turn is based on 
/proc/cpuinfo - and it doesn't exist on Darwin. AFAICT, only an 
explicit --enable-altivec would set $_altivec to true in configure.

I'll try to find out if enabling AltiVec on my G3 will work as it 
should, with runtime detection and all, and report back.

- Dan



More information about the MPlayer-dev-eng mailing list