After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 149326 - Wnck-CRITICAL when loading applet
Wnck-CRITICAL when loading applet
Status: RESOLVED FIXED
Product: libwnck
Classification: Core
Component: general
2.7.x
Other Linux
: Normal major
: ---
Assigned To: libwnck maintainers
libwnck maintainers
Depends on:
Blocks:
 
 
Reported: 2004-08-04 20:14 UTC by Kjartan Maraas
Modified: 2005-12-13 20:05 UTC
See Also:
GNOME target: ---
GNOME version: 2.7/2.8


Attachments
Fix new warnings (1.12 KB, patch)
2005-01-12 06:35 UTC, Elijah Newren
needs-work Details | Review
Better patch to fix the warnings (885 bytes, patch)
2005-07-20 21:12 UTC, Rodney Dawes
none Details | Review
Proposed patch (2.12 KB, patch)
2005-11-05 20:35 UTC, Vincent Untz
none Details | Review
Other solution (1.83 KB, patch)
2005-11-05 20:38 UTC, Vincent Untz
committed Details | Review

Description Kjartan Maraas 2004-08-04 20:14:24 UTC
Here's the backtrace from --g-fatal-warnings:

GConf-CRITICAL **: file gconf-client.c: line 547 (gconf_client_add_dir):
assertion `gconf_valid_key (dirname, NULL)' failed
aborting...

Program received signal SIGABRT, Aborted.

Thread NaN (LWP 22990)

  • #0 _dl_sysinfo_int80
    from /lib/ld-linux.so.2
  • #1 raise
    from /lib/tls/libc.so.6
  • #2 abort
    from /lib/tls/libc.so.6
  • #3 g_logv
  • #4 g_log
  • #5 gconf_client_add_dir
    at gconf-client.c line 575
  • #6 _gnomeui_gconf_lazy_init
    at gnome-gconf-ui.c line 300
  • #7 _gnome_gconf_get_bool
    at gnome-gconf-ui.c line 318
  • #8 libgnomeui_post_args_parse
    at gnome-ui-init.c line 521
  • #9 gnome_program_postinit
    at gnome-program.c line 1633
  • #10 gnome_program_init_common
    at gnome-program.c line 1865
  • #11 gnome_program_init
    at gnome-program.c line 1683
  • #12 main
    at wncklet.c line 77

Comment 1 Elijah Newren 2004-11-08 21:37:29 UTC
Unique stack trace.
Comment 2 Kjartan Maraas 2005-01-04 00:59:14 UTC
Not seeing those GConf-CRITICAL's any longer, but I do see this when adding the
window list:

(wnck-applet:28264): Wnck-CRITICAL **: wnck_screen_get_windows: assertion
`WNCK_IS_SCREEN (screen)' failed

(wnck-applet:28264): Wnck-CRITICAL **: wnck_screen_get_active_window: assertion
`WNCK_IS_SCREEN (screen)' failed
Comment 3 Elijah Newren 2005-01-12 06:35:39 UTC
Created attachment 35881 [details] [review]
Fix new warnings

I ran across these while trying to debug bug 163343, and these warnings are
caused by some of the related code.  Anyway, here's a patch that fixes the
warnings.  The tasklist was being updated, which depended on knowing the
screen, but the screen wasn't setup until realization time.  So I deferred this
part of the code until the realization happened.
Comment 4 Havoc Pennington 2005-01-15 03:22:35 UTC
Comment on attachment 35881 [details] [review]
Fix new warnings

Anything in realize() should be undone in unrealize() (and anything in init
should be undone in destroy/finalize, so hopefully the code to copy to
unrealize is currently in finalize...)
Comment 5 Luis Villa 2005-01-25 13:47:34 UTC
If this is just warnings now, marking down.
Comment 6 Rodney Dawes 2005-07-20 21:12:36 UTC
Created attachment 49488 [details] [review]
Better patch to fix the warnings

This patch only moves the code that is calling the problem, and mvoes it to the
"new" method, rather than the realize method. I'm not marking the other patch
obsolete though, in case Havoc or someone deems the realize/unrealize method a
better way of doing it.
Comment 7 Rodney Dawes 2005-08-02 20:33:45 UTC
Havoc? Can this be committed?
Comment 8 Havoc Pennington 2005-08-02 21:38:26 UTC
new() has a couple problems I think:
 1. it should only be a convenience method, real stuff should be in init()
    possibly with construct_only properties
 2. you don't have an X screen at new() time, only when realized

I haven't really looked at the specifics here though so I don't know if these
are important points.
Comment 9 Vincent Untz 2005-11-05 20:35:49 UTC
Created attachment 54364 [details] [review]
Proposed patch

Here's my try :-)

We just return if tasklist->window is NULL :-)
Comment 10 Vincent Untz 2005-11-05 20:38:36 UTC
Created attachment 54365 [details] [review]
Other solution

And here's the realize/unrealize way.
Comment 11 Federico Mena Quintero 2005-12-09 03:21:54 UTC
Please commit the patch from comment #10.
Comment 12 Vincent Untz 2005-12-13 20:05:20 UTC
Committed