Hi,

> I'm using egcs-2.91.66 (aka egcs-1.1.2-30 on RH 6.2) and glibc 2.1.3-22.
> 
> It very may well just be your version of gcc..

Very likely. Both systems where this occured were using
gcc version 2.95.3 20010125 (prerelease)gcc version 2.95.3 20010125
(prerelease)
from Debian/unstable (sid). I just wanted to make sure that this problem
isn't more wide-spread...

> Also, I've been digging around the sq3 sound eventgetting lost thing..
> my gut feeling is that the doall(sounds) is clearing the signal before
> the doall(scripts) is running.  I wonder when real SCI calls the sound
> server (or is it the other way around; the sound driver is realtime, and
> it tells SCI to go?)

AFAIK, the sound driver and interpreter run concurrently, using "shared
memory". (The term is not quite appropriate, since the memory is not
exactly explicitly shared, but, rather implicitly not unshared ;-)

IIRC, the sound server runs on Int 8 (or something called from Int 8) on
DOS boxes, and explicitly writes to memory each time something changes.
Since it cannot be interrupted, this action can be considered atomic.


llap,
 Christoph

Reply via email to