Re: [Evolution-hackers] e-d-s refusing to exit

[Not Zed <notzed ximian com>]

>   
> Thread 1 (Thread 16384 (LWP 14533)):
> #0  0x4055f5d4 in __pthread_sigsuspend () from /lib/libpthread.so.0
> #1  0x4055f398 in __pthread_wait_for_restart_signal () from /lib/libpthread.so.0#2  0x00000020 in ?? ()
> #3  0xbffff3b4 in ?? ()
> #4  0x404d26bc in ?? () from /usr/lib/libORBit-2.so.0
> #0  0x4070461a in poll () from /lib/libc.so.6
> (gdb)
>
> I see that threads 1 and 3 involve ORBit...
>
> This is where I admit complete ignorance of debugging ORBit/bonobo.

The main problem here is the trace isn't showing a lot of the symbols.  You probably need a debugging build of orbit/bonobo to try and debug anything like this (i.e. build from source), otherwise you're not going to get enough info.

But from what i know about it, it launches multiple g_mainloops in different threads and uses a library 'linc2' to manage most of that and communicate between them.  Depending on which POA is in use, threads can be launched for each method call or object instance too, all managed through the same stuff.