Re: [AD] x color conversion again (w/ responsiveness patch)

[Evert Glebbeek <eglebbk@xxxxxxxxxx>]

> If SIGALRM is also asynchronous, then that means we'd definately have 
> problems calling X routines from it. Also, the rest of the document 
> leads me to believe that signals do indeed have the same stack trappings 
> as a hardware interrupt.. ie. no FPU stack. I tried asking Bob in 
> #allegro if MMX would have the same problems, but he's not around. It 
> does make sense since I compiled Allegro with -mmmx which enables MMX 
> usage in the produced asm, which includes the signal handlers.

MMX should be affected in the same way as the FPU, I think. It's been a 
while since I looked at the details of the MMX instruction set, but this 
is what I remember.

Evert