Ubuntu bugzilla about this: http://bugzilla.ubuntu.com/show_bug.cgi?id=16426

"Hello,

When using a multihead setup in X (NOT xinerama), the wnck-applet keeps
crashing, also 
when trying to switch workspaces on the second monitor a get a schema error that
the number of workspaces (in my case 4) is not a correctly formatted schema value.

Using xinerama multihead support does work."

Debug backtrace from the Ubuntu bug:

Backtrace was generated from '/usr/libexec/wnck-applet'

Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1".
(no debugging symbols found)
(no debugging symbols found)
[Thread debugging using libthread_db enabled]
[New Thread -1223587296 (LWP 13222)]
0xffffe410 in ?? ()

+ Trace 63228

This is also true for GNOME 2.10.2 and Gentoo.

P.S.: The same can be said for the Window List applet.
P.P.S: Another way to almost deterministically crash the Workspace Switcher
applet is to attempt to bring up its properties.

Moving to libwnck.

So the crash seems to show that it's on the line that reads
  g_return_val_if_fail (WNCK_IS_WINDOW (window), None);
unless the line numbers have changed.  How could that line cause a crash, though?

And, uh, if the gconf value is set to something not in the schema, what's the
output of
  gconftool-2 --get /apps/panel/applets/applet_11/prefs/display_all_workspaces
and
  gconftool-2 --get \
    /schemas/apps/workspace_switcher_applet/prefs/display_all_workspaces

Running in multihead again, with new hardware, stuff still crashes.

To answer your questions:
applet_7 is now my switcher:

gconftool-2 --get /apps/panel/applets/applet_7/prefs/display_all_workspaces
Failed to get value for
`/apps/panel/applets/applet_7/prefs/display_all_workspaces': Failed: Failed:
Schema `/schemas/apps/workspace_switcher_applet/prefs/display_all_workspaces'
specified for `/apps/panel/applets/applet_7/prefs/display_all_workspaces' stores
a non-schema value

%  gconftool-2 --get \
>     /schemas/apps/workspace_switcher_applet/prefs/display_all_workspaces
true

Hope this helps

Uh, why would the schema be 'true'?  That totally doesn't make sense.  It ought
to be the description of what the preference means.  E.g., on my machine:

1002 newren@amr:~$ gconftool-2 --get \
    /schemas/apps/workspace_switcher_applet/prefs/display_all_workspaces
Type: bool
List Type: *invalid*
Car Type: *invalid*
Cdr Type: *invalid*
Default Value: true
Owner: workspace-switcher-applet
Short Desc: Display all workspaces
Long Desc:
            If true, the workspace switcher will show all workspaces. Otherwise
            it will only show the current workspace.


But anyway, I'm guessing that's unrelated to the crash.  The stack trace doesn't
make sense to me, unless it's some kind of memory corruption.  Could someone
with a multi-head setup run under valgrind?  Unfortunately, I only have single
head setups...

With my new hardware I cannot reproduce the crash - so it was (I think) faulty
memory in my old workstation.

However libwnck still isn't usefull in a multihead setup. The workspace switching
only works on the main monitor. On the other monitor I only see only workspace
instead of the 4 I've defined.

*** Bug 319114 has been marked as a duplicate of this bug. ***

*** Bug 318914 has been marked as a duplicate of this bug. ***

reopening due to the dups

I may have found a way to reproduce this

1. Set up your computer in dual head mode for two monitors
2. Put a panel on each monitor, have the "Workspace switcher" and "Window List"
applets on each of those panels
3. Open an application like firefox/gedit
4. Now close that application
5. Now open that same application on the 2nd monitor
6. Crash!
7. Do the same in reverse
8. Open firefox/gedit on second monitor
9. Close the application
10. Open the application on the first monitor
11. Crash!

*** Bug 327063 has been marked as a duplicate of this bug. ***

*** Bug 324165 has been marked as a duplicate of this bug. ***

*** Bug 328506 has been marked as a duplicate of this bug. ***

*** Bug 328057 has been marked as a duplicate of this bug. ***

Hi, I can confirm identical behaviour to bryan ie. open an app on screen 0, close app, open the same app on screen 1 - crash!! see bug 328057 :/

*** Bug 335610 has been marked as a duplicate of this bug. ***

*** Bug 343525 has been marked as a duplicate of this bug. ***

*** Bug 344711 has been marked as a duplicate of this bug. ***

*** Bug 350035 has been marked as a duplicate of this bug. ***

Ubuntu bug with a similar backtrace (Ubuntu edgy): https://launchpad.net/products/libwnck/+bug/72632

Another way:
1: Set up dual head with separate x screens.
2: Set up a workspace switcher on desktop 1 .
3: Select "Show Workspace Names in switcher".
4: Set up a workspace switcher on desktop 2 .
5: Select "Show Workspace Names in switcher".
6: Crash.

new ubuntu bug with the same backtrace and GNOME 2.17: https://launchpad.net/bugs/89597

"Binary package hint: gnome-panel

Have been having a lotttttttt of trouble when trying to right click on the window 'frames' to send to a different desktop and clicking on 'view', in the 'file Browser' to make changes... When right clicking on the window 'frame' it's like nautilus cycles through all the windows on each desktop, until it stops on the one that I started with, trying to move it to a different desktop.... It's been doing this off and on for a month or so now.... I pretty much do the 'updates' for fiesty as they come out... I can't seem to nail this down as to when exactly it began though....
...
libwnck18 2.17.92-0ubuntu1
..."

This crash is still happening, but we need a recent stack trace, or a valgrind log as Elijah already pointed out.

Marking as NEEDINFO until we get some new data.

I've closed the distribution bug for now, I'll reopen if new datas are sent about the crash

Can any of the people who were affected by this bug and who are cc'ed here still reproduce this?

I am currently using Ubuntu Feisty Fawn, my locale is now New Zealand, I am using Evolution 2.10.1

Yes, the bug is still present for me, I tried updating my grandmothers D.O.B. (1914) and it still won't accept it, "Invalid Date Value" is the warning displayed.

(In reply to comment #29)
> I am currently using Ubuntu Feisty Fawn, my locale is now New Zealand, I am
> using Evolution 2.10.1
> 
> Yes, the bug is still present for me, I tried updating my grandmothers D.O.B.
> (1914) and it still won't accept it, "Invalid Date Value" is the warning
> displayed.
> 

sorry for the noise, wrong bug....

*** Bug 454719 has been marked as a duplicate of this bug. ***

I cannot reproduce this crash with GNOME 2.22. However, the following bug remains: Upon login, the applet instance running on the secondary screen does not display the status of the desktops associated with the secondary screen. Instead, it displays the status of the desktops associated with the primary screen. There's a bug about this here, but I can't find it. *scratch head*

(In reply to comment #32)
> I cannot reproduce this crash with GNOME 2.22. However, the following bug
> remains: Upon login, the applet instance running on the secondary screen does
> not display the status of the desktops associated with the secondary screen.
> Instead, it displays the status of the desktops associated with the primary
> screen. There's a bug about this here, but I can't find it. *scratch head*

When you say "upon login", do you mean that after a while, it works correctly?

Here's the comment I made in bug #480710:

=========================
As of GNOME 2.22, the workspace switcher bug is still present:

1. Configure X to use 2 screens separately (:0.0 and :0.1)
2. Create a horizontal bottom panel on each screen.
3. Add a workspace switcher to each panel.
4. In each workspace-switcher, set the number of desktops to a total of 8, in 2
rows of 4.
5. Log out.
6. Log in.

At this point, both workspace switchers display the status of desktops on :0.0,
and neither of them displays the status of :0.1. If you remove the switcher on
:0.1 from its panel, and re-add it, it will work again, showing 8 desktops in a
row, instead of your pre-login choice of 2 rows of 4.

If the panel on :0.1 is a vertical panel, removing and re-adding the switcher
doesn't help. Even after re-adding the switcher, it still displays the status
of :0.0.

Here's hoping GNOME 2.24 will work (again).
=========================

This bug is actually #479959. The comment above is from a similar bug. I guess I put it in the wrong place.

I checked that the version of libwnck I have on my desktop contains a fix to bug #479959, however, on Gentoo at least the fix didn't work for me. I checked that the libwnck tarball Gentoo uses for building contains this change:

http://svn.gnome.org/viewvc/libwnck/trunk/libwnck/pager.c?view=log#rev1556

However, the above outlined steps still lead me to the misbehaviour I described in the above comment,

P.S.: I haven't checked if, after waiting a while, it works correctly. After all, that would mean that the switcher on :0.1 all of a sudden starts to display 1 row of 8 desktops instead of 2 rows of 4. And just to make the link work, it's bug #479959.

Hm. We still have no valgrind log Vincent has asked for :-/ Are we going to close this bug then in a few weeks?
Maybe this issue is fixed anyway?

Well, I haven't observed a crash recently. However, the switcher on :0.1 still doesn't work correctly. Whenever I log in, I have to remove the switcher from the :0.1 panel, then add a new one, then set the number of rows to 2 in its preferences.

May I suggest that we close this bug then and that you open a new one for your issue?

Closing this bug report as no further information has been provided. Please feel free to reopen this bug if you can provide the information asked for.
Thanks!