Re: it works after killing esd - doh!

[David O'Toole]

> > too screwed up... cause i think the call to open /dev/dsp in init_sound
> > never returns so the core stops initializing
> > i must definitely learn how to use a debugger :)

A great frontend to GDB that I have been using is GNU DDD. It's available
at http://www.gnu.org/software/ddd I think.

> i feel like a total newbie...
> after using lsof to find out that esd is hogging the oss driver
> and killing esd - octal works for me! cool!!!

Hmm! That's weird. If XMMS is playing thru ESD on my system, and I try to
start Octal, it just gives an error message that it can't open the device,
and exits. I wonder why the call isn't returning on your system? Maybe we
should compare ESD version numbers.

Also, thank you for pointing out the problem with the "make doc" target in
the tarball. I will fix it.

This is kind of ironic, since your problem turned out to be not a deadlock
bug but ESD :-). But anyway, when you first posted it I thought it was a
deadlock bug, and this afternoon I rewrote the communication system
(between the GTK+ GUI and the DSP core). It is much, much better now than
before, and the new approach will be more scalable. So, even though it
turned out not to be a problem, thanks for inspiring me :-). The codebase
actually shrunk by about 100 lines but still does more than last time,
which is always a nice experience.

> dto: how many threads octal uses now? what does each one do?

There are two at the moment --- one for the GUI and another that handles
audio processing and other tasks.


-- 
@@@ david o'toole
@@@ address@hidden
@@@ www.gnu.org/software/octal