Re: [GnomeMeeting-list] Problem with gconf!



    Starting program: /home/pci2103/gmeeting_0.85.1/src/./gnomemeeting
    [New Thread 1024 (LWP 22176)]
    
    GnomeUI-WARNING **: While connecting to session manager:
    Could not open network socket.
    2002-03-28T11:36:46-0500  2002-03-28T11:36:46-0500
    [New Thread 2049 (LWP 23119)]
    [New Thread 1026 (LWP 23120)]
    Delayed SIGSTOP caught for LWP 23120.
    [New Thread 2051 (LWP 23127)]
    Delayed SIGSTOP caught for LWP 23127.
    [New Thread 3076 (LWP 23128)]
    Delayed SIGSTOP caught for LWP 23128.
    
    Program received signal SIGSEGV, Segmentation fault.
    [Switching to Thread 1024 (LWP 22176)]
    0x409f3c10 in PAbstractList::GetObjectsIndex () from
    /usr/local/lib/libpt_linux_x86_r.so.1

Here it is: I had a mismatch of ptlib libraries, because I wrongly assumed that
my ./configure option "--with-ptlib-libs=$PWLIBDIR/lib" also meant the
executable was using $PWLIBDIR/lib - stupid me. As a side note, I was pretty
scared of gdb, but should try to overcome this from now on.

> I suspect it is still a gconf problem.

So it wasn't for the "crash at startup" part.

> I don't remember that. I don't think there are races with GM, but I
> think that your Gnome installation and your setup are completely broken,
> or at least very weird.

Well, all my compilation problems were related to gconf. But the last few days
concerning the "crash at startup" have been mostly about my lack of insight.
Admittedly, running gconf over SUSE is kinda weird because it doesn't work out
of the box (gconf's fault, in my opinion).

> I do not see, and do not understand why adding a sleep (2) in
> gnomemeeting.cpp could change in main_window.cpp, that makes no sense to
> me, and believe me, I'm adding code everywhere since a long time without
> the need to put things like a sleep (2) here and there. Sorry, there are
> no "bad fixes" like that in GnomeMeeting.

This was a sleep(1) just after gnomemeeting_init(gw, pw, ...) in
GnomeMeeting::Main ()
Agreed, it doesn't make sense at all. And as of GM 0.12.2 it was still saying
right under my kludge:

  /* Quick hack to make the GUI refresh even on high load from the other
     threads */
  gtk_idle_add ((GtkFunction) gnome_idle_timer, gw);

  gtk_timeout_add (1000, (GtkFunction) AppbarUpdate, 
		   gw->statusbar);

> Moreover, why didn't you contribute your code?

I didn't contribute it because this was mostly a hack used for me to make some
audio measurements remotely on a H.323 compliant host (and thus my total
satisfaction with a kludge). In essence, read audio from a file, send it to the
remote host, and record what's coming back onto another file. (Please note that
the corresponding executable wasn't distributed beyond my own HD, thus
respecting the terms of the GPL).

Don't take me wrong, I am 100% for open souce software, and started working on
code regarding two aspects of the OpenH323/PwLib stack that will be contributed
back to Equivalence, as it should in the open source spirit (and despite the
less stringent terms of their license).

As of today, I have no plans regarding GM except for testing the OpenH323/PwLib
stack.

Please let me thank you here for your assistance in getting GM working. This
was much appreciated.

Guilhem.


__________________________________________________
Do You Yahoo!?
Yahoo! Greetings - send holiday greetings for Easter, Passover
http://greetings.yahoo.com/



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]