Re: GLib-GObject-CRITICAL - not sure how to report it



Christian Kirbach wrote:
On Wed, 11 Jan 2006 09:11:14 +0100, Rod Butcher <rbutcher hyenainternet com> wrote:

GLib-GObject-CRITICAL **: g_value_get_pointer: assertion `G_VALUE_HOLDS_POINTER (value)' failed
aborting...

This occurs in evolution, but is Glib the problem here ?

I've had a brief discussion with Olav yesterday about this :)
It seems like glib is never the culprit, but the applicatio/library calling
it with wrong values.
In order to see who causes this you need to debug the application and examine the stack trace (http://live.gnome.org/GettingTraces). Is is not that difficult
if you know what you have to look for.
Example of a crash in glib:


========================
Backtrace was generated from '/opt/gnome2/bin/gthumb'

[..rubbish...]
0xb726e231 in __waitpid_nocancel () from /lib/tls/libpthread.so.0
#0  0xb726e231 in __waitpid_nocancel () from /lib/tls/libpthread.so.0
#1 0xb7e9ed2d in libgnomeui_segv_handle (signum=-512) at gnome-ui-init.c:749
#2  <signal handler called>
#3  0xb71996d0 in strcmp () from /lib/tls/libc.so.6
#4  0xb729889e in path_list_find_path (list=0x876a9e8,
    path=0x85b7d70 "/home/nazgul/Desktop/100_pana/p1000424.jpg")
    at file-utils.c:1019
[rubbish]
========================

This is the history of a crash *backwards*. Look for "<signal handler called>". It crashes in frame #3 in function strcmp () of libc and then a signal handler is
called that deals with the crashed application. the culprit is the function
path_list_find_path() that is part of "file-utils.c:1019" . If you scan your cvs root for that file you will see that it resides in gthumb/libgthumb - so file the
report against gthumb.

Am I actually doing any good by reporting crashes with cvs versions or just wasting developers' time ?

No, you are doing appreciated work and much needed testing :)
If you can, please check for duplicates. That can save some spam for
developers.
Thanks for the explanation. I'll carry on what I'm doing. Does it matter when I get a lot of "variable .. not available" as in :-
#0  IA__g_logv (log_domain=Variable "log_domain" is not available.
) at gmessages.c:503
as far as I know everything is compiled with option g.
thanks
Rod


--
-----------------------------------------------------
<corporate disclaimer drivel goes here>



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