On Thu, 16 Jul 2009, m.wolkst...@gmx.de wrote:
>>
> ohhhhhhhh :))) my patch do nothing!
Not true: We're talking about it and thinking critically. That's worth
something. :-)
> i can code everything against buffersice change crash e.g.
> INFOLOG("HeHeHoho"); instead my patch also fix the buffersize segfault.
> why? its allredy fixed into the zombie branch. crazy
Not exactly fixed. I set up a test today and whenever I play around with
the "per-track outputs." I'm getting crashes when using tracking outputs.
If I change the buffer size (while tracking outs are enabled) I think I
can force a crash. However, with tracking outputs OFF, I couldn't cause
it to crash.
I think maybe the problem is that we're caching the pointer to the output
buffers for each of the tracking outputs. I fixed this for the main audio
outs in jack_zombies, but I overlooked the per-track outputs:
class Sampler {
float* __track_out_L[ MAX_INSTRUMENTS ];
float* __track_out_R[ MAX_INSTRUMENTS ];
}
We need to get these fresh from the audio driver on every process() cycle.
Michael, do you think that you could eliminate these class members? If
you look at audioEngine_process_clearAudioBuffers() you'll see how I did
it in hydrogen.cpp.[1]
Peace,
Gabriel
[1] Feel free to leave out the assert(buf) stuff. I did that so
that it would catch the error if I made a dumb mistake.
------------------------------------------------------------------------------
Enter the BlackBerry Developer Challenge
This is your chance to win up to $100,000 in prizes! For a limited time,
vendors submitting new applications to BlackBerry App World(TM) will have
the opportunity to enter the BlackBerry Developer Challenge. See full prize
details at: http://p.sf.net/sfu/Challenge
_______________________________________________
Hydrogen-devel mailing list
Hydrogen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/hydrogen-devel
