Re: [AD] get_audio_stream_buffer

[Serge Semashko <serge@xxxxxxxxxx>]

Elias Pschernig wrote:

> > Can anybody confirm the bug or I am the only one who gets it?
> >
> > My current configuration is:
> > computer 1: AMD Duron 950, Gentoo Linux, kernel 2.6.7, glibc with NPTL, 
> > ALSA with OSS emulation (Allegro is configured to use OSS only), Allegro 
> > 4.1.14, configure option '--disable-asm' used
> > computer 2: AMD Athlon 850, Gentoo Linux, kernel 2.6.7, glibc without 
> > NPTL, ALSA with OSS emulation (Allegro is configured to use OSS only), 
> > Allegro 4.0.3, configure option '--disable-asm' used
> >
> > I will try more different configurations, but it takes much time to test
>
>
> Yes, the time to test is a big problem. And the problem could be with
> DUMB. Using OSS instead of Alsa, and --disable-asm - I'd have thought
> both to be more stable than using Alsa and the asm version. So probably
> it has nothing to do with them. But don't know.

I removed --disable-asm option just for test, it did not help.

> Can you describe the symptoms in some more detail? E.g., what if you
> switch off and on music? Does the game get responsive again? And do also
> other sounds played with play_sample have noise after the problem
> occurs?

Well, I'm using that small test program now:
http://ufo2000.lxnt.info/files/other/music-bug.zip

But the symptoms are the same as in the game. It plays music for some 
time, but then something happens and it only produces hissing noise that 
has nothing common with the music played before.

The memory usage does not grow, so if there are some leaks, the program 
is probably leaking some handles.

As nobody else can reproduce the problem, that could probably be 
compiler optimization bugs. I'm using gcc 3.3.3 (from Gentoo Linux), my 
CFLAGS are '-Os -mcpu=athlon-tbird -march=athlon-tbird'. Can anybody 
else who is using Gentoo Linux try to run this test program (for at 
least some hours)?

PS. I'm short on free time, so I have currently just disabled music 
support in Linux build. But as time allows I will try to investigate 
this problem further.