[MPlayer-cvslog] CVS: main ChangeLog,1.222,1.223

[Jindrich Makovicka]

Oded Shimon wrote:
> On Mon, Jun 27, 2005 at 05:26:11PM +0200, Guillaume POIRIER wrote:
> 
>>Hej,
>>
>>On 6/27/05, Oded Shimon <ods15 at ods15.dyndns.org> wrote:
>>
>>>On Mon, Jun 27, 2005 at 03:33:42PM +0200, Guillaume Poirier CVS wrote:
>>>
>>>>+    * full gcc 4 support for IA-32 and AMD-64 ports
>>>
>>>.. and broke gcc 3.x in the process.. :)
>>
>>All of them? that's odd, how come does it work on my IA-32 machine with gcc-3.4?
>>
>>How should we fix this? #ifdefs?
>>
>>We should find a solution quickly to fix all the mess I made!
> 
> 
> You REALLY didn't know?? It's been mentioned on the mailing list several 
> times, atleast once by me, with backtrace and cpu info. I was wondering why 
> you still weren't fixing it.. It's an ENTIRE broken and extremely important 
> feature in MPlayer's CVS!!
> 
> http://mplayerhq.hu/pipermail/mplayer-dev-eng/2005-June/035338.html

it looks like gcc-3.3 optimizes out the assembly code at the end of
hcscale_MMX2, despite the volatile keyword. strange...

-- 
Jindrich Makovicka