[Evolution-hackers] Re: bonobo_main thread ...

[Chris Toshok <toshok ximian com>]

The extra mainloop thread makes programs that don't use the async api
possible, when they don't have another thread running a mainloop.

Like:

int
main(...)
{
....

e_book_get_default_addressbook (&book, NULL);

e_book_get_contact (book, contact_id, &contact, NULL);

printf ("contact name == %s\n", e_contact_get_const (contact,
E_CONTACT_FULL_NAME));

}

We still need a mainloop running somewhere, and (trusting glib's thread
safety when it comes to multiple mainloops running on the same context),
it's safe if you don't care on which thread your GSource callbacks are
executed (which ebook doesn't).

It doesn't seem to adversely affect evolution, fwiw.

Chris

On Wed, 2004-02-04 at 05:40, Michael Meeks wrote:
> Hi guys,
> 
> 	Just debugging a deadlock of some sort (perhaps the ORB not reporting a
> dead connection exception somehow [possibly] when e-d-s crashes). Anyhow
> - I see (again) in my trace this:
> 
> Thread 3 (Thread 16386 (LWP 29821)):
> #0  0x413e6b94 in __pthread_sigsuspend () from /lib/i686/libpthread.so.0
> #1  0x413e69d8 in __pthread_wait_for_restart_signal () from /lib/i686/
> libpthread.so.0
> #2  0x413e2e90 in pthread_cond_wait GLIBC_2 0 () from /lib/i686/
> libpthread.so.0
> #3  0x415d528f in g_main_context_wait () from /usr/lib/libglib-2.0.so.0
> #4  0x415d6ce7 in g_main_loop_run () from /usr/lib/libglib-2.0.so.0
> #5  0x4077c8b7 in bonobo_main () from /opt/gnome/lib/libbonobo-2.so.0
> #6  0x41a8a026 in startup_mainloop (arg=0x0) at e-book.c:2346
> #7  0x413e3f60 in pthread_start_thread () from /lib/i686/libpthread.so.0
> #8  0x41710327 in clone () from /lib/i686/libc.so.6
> 
> 	And I wonder what crack is this ? why do we have a thread sitting there
> running 'bonobo_main' ? - of course, hopefully we'll enter the glib
> mainloop in the main process (blocking that thread from getting access
> to the context on the condition) before anything too bad happens in that
> thread - but; what is the thinking here ? and/or why was that added ? it
> looks hyper-odd to me.
> 
> 	Thanks,
> 
> 		Michael.