[RTcmix-discuss] OSX HW testing
baker at charlieb.com
Sat Jan 29 12:48:37 EST 2005
guys, on my OS X box, I get an error
DEBUG: audio device: peak check: 256 report clip: 0
Reading symbols for shared libraries . done
Program received signal SIGSYS, Bad system call.
0x90012588 in clock_sleep_trap ()
in several of the imbed instruments...poking around under gdb, I see
the following things:
This is always with sleep() calls performed repeatedly in a while(1)
The actual value handed sleep doesn't seem to matter: they all are
but in the imbedded app, the repeated calls to sleep always return a
bad syscall SIG
after a few calls.
I took wanderfreq, which was not working at all for me,
(it was throwing the SIGSYS almost immediately)
and all I did was comment out the sleep(1)
from the while(1) loop that keeps the app running
while(1) ; // sleep(1);
The result was: wanderfreq ran fine.
Is this a darwin error?
Don't I recall something about repeated calls to sleep() inside a mutex?
(or was that under linux/alsa? sheesh...I'm useless, I know...)
Well, I hope there is a simple solution...
Doug, John, Brad, et al.!!
Thanks for all the excellent work! Other than the imbedded instrs, the
worked like a charm! the audio code even avoids the nastiness with the
cheap USB converter
I have where signing in to core audio will always reset the left
channel to 8-bit, sigh (!)
( an earlier version of the RTcmix core audio code did that: Abelton
Live still does! )
One day I'll find a way to afford my own 828MkII!
(hard to do for an average hacker in the labor-unfriendly, cheap
salary, Bush-lovin, intellectual hating ' south! HA!)
More information about the RTcmix-discuss