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 663122 - Doesn't quit on exit with GLib's git master
Doesn't quit on exit with GLib's git master
Status: RESOLVED FIXED
Product: evolution
Classification: Applications
Component: general
3.4.x (obsolete)
Other Linux
: Normal normal
: ---
Assigned To: Evolution Shell Maintainers Team
Evolution QA team
Depends on:
Blocks:
 
 
Reported: 2011-10-31 20:17 UTC by David Ronis
Modified: 2013-09-13 01:08 UTC
See Also:
GNOME target: ---
GNOME version: ---

Attachments
proposed evo patch (365 bytes, patch)
2011-12-02 10:07 UTC, Milan Crha 		committed Details | Review

Description David Ronis 2011-10-31 20:17:09 UTC 
I just started evolution (today's git/master).   Shortly after the UI mapped, I exited it by pressing
ctrl-c.   The UI disappears, but evo still keeps running in the background (for at least 5 min, in the latest test).

I attached gdb and got the following backtrace:

(gdb) thread apply all bt full

+ Trace 228956







Comment 1


David Ronis





          2011-11-29 14:50:19 UTC
        



After upgrading to the git master,  running evolution from a console, and trying to close it by hitting "x" on the window border, evo hung.  Here's another backtrace:

(gdb) thread apply all bt full





+
Trace
    229172
















Comment 2


Milan Crha





          2011-12-01 10:07:03 UTC
        



Those backtraces are pretty similar. Is it possible this is about used version of glib or gtk3? The threads are mostly in idle:
Thread 1 in gtk_main() called from evolution's main.c
Thread 2 in gdbus_shared_thread_func() from GDBus
Thread 3 in dconf_context_thread() from dconfcontext
Thread 4 in glib_worker_main() from gmain.c

The last one is new to me.

Could you try whether gtk_main_quit() is called, please? I woud run evolution under gdb in offline, let it settle (no actions shown in the status bad) and then quit evolution as you usually do? I would do it with a command like this:
   $ gdb evolution --ex "b gtk_main_quit" --ex "r --offline"

Mine is called when evolution destroys its main window:
(gdb) bt




+
Trace
    229189






    

  • 
#0
gtk_main_quit

    
                at gtkmain.c
                  line
                  1406

    

    • 

  • 
#1
gtk_application_window_removed

    
                at gtkapplication.c
                  line
                  196

    

    • 

  • 
#2
g_closure_invoke

    
                at gclosure.c
                  line
                  774

    

    • 

  • 
#3
signal_emit_unlocked_R

    
                at gsignal.c
                  line
                  3202

    

    • 

  • 
#4
g_signal_emit_valist

    
                at gsignal.c
                  line
                  3003

    

    • 

  • 
#5
g_signal_emit

    
                at gsignal.c
                  line
                  3060

    

    • 

  • 
#6
gtk_window_release_application

    
                at gtkwindow.c
                  line
                  2637

    

    • 

  • 
#7
gtk_window_destroy

    
                at gtkwindow.c
                  line
                  4528

    

    • 

  • 
#8
g_closure_invoke

    
                at gclosure.c
                  line
                  774

    

    • 

  • 
#9
signal_emit_unlocked_R

    
                at gsignal.c
                  line
                  3388

    

    • 

  • 
#10
g_signal_emit_valist

    
                at gsignal.c
                  line
                  3003

    

    • 

  • 
#11
g_signal_emit

    
                at gsignal.c
                  line
                  3060

    

    • 

  • 
#12
gtk_widget_dispose

    
                at gtkwidget.c
                  line
                  10666

    

    • 

  • 
#13
shell_window_dispose

    
                at e-shell-window.c
                  line
                  329

    

    • 

  • 
#14
g_object_run_dispose

    
                at gobject.c
                  line
                  945

    

    • 

  • 
#15
g_list_foreach

    
                at glist.c
                  line
                  938

    

    • 

  • 
#16
shell_ready_for_quit

    
                at e-shell.c
                  line
                  413

    

    • 

  • 
#17
mail_backend_store_operation_done_cb

    
                at e-mail-backend.c
                  line
                  135

    

    • 

  • 
#18
g_simple_async_result_complete

    
                at gsimpleasyncresult.c
                  line
                  749

    

    • 

  • 
#19
complete_in_idle_cb_for_thread

    
                at gsimpleasyncresult.c
                  line
                  817

    

    • 

  • 
#20
g_main_dispatch

    
                at gmain.c
                  line
                  2425

    

    • 

  • 
#21
g_main_context_dispatch

    
                at gmain.c
                  line
                  2995

    

    • 

  • 
#22
g_main_context_iterate

    
                at gmain.c
                  line
                  3073

    

    • 

  • 
#23
g_main_loop_run

    
                at gmain.c
                  line
                  3281

    

    • 

  • 
#24
gtk_main

    
                at gtkmain.c
                  line
                  1362

    

    • 

  • 
#25
main

    
                at main.c
                  line
                  675

    

    • 













Comment 3


David Ronis





          2011-12-01 17:58:22 UTC
        



Hi Milan,

I tried your suggestion.   Interestingly, the break in gtk_main_quit never gets triggered.  Hitting ^C gets me back to the gdb prompt and I got a backtrace.  Here's a log:

gdb evolution --ex "b gtk_main_quit" -ex "r --offline"

Starting program: /opt/garnome-svn-3.2.0/bin/evolution --offline
[Thread debugging using libthread_db enabled]
[New Thread 0xb5f97b70 (LWP 7773)]
[New Thread 0xb576fb70 (LWP 7774)]
[New Thread 0xb4dffb70 (LWP 7775)]
[New Thread 0xa511cb70 (LWP 7781)]
[Thread 0xa511cb70 (LWP 7781) exited]
[New Thread 0xa511cb70 (LWP 7782)]
[New Thread 0xa44d0b70 (LWP 7790)]
[New Thread 0xa3cd0b70 (LWP 7791)]
[Thread 0xa3cd0b70 (LWP 7791) exited]
[Thread 0xa511cb70 (LWP 7782) exited]
[New Thread 0xa511cb70 (LWP 7792)]
[New Thread 0xa3cd0b70 (LWP 7793)]
[New Thread 0xa2effb70 (LWP 7794)]
[Thread 0xa44d0b70 (LWP 7790) exited]
[Thread 0xa2effb70 (LWP 7794) exited]
[New Thread 0xa2effb70 (LWP 7801)]
[New Thread 0xa44d0b70 (LWP 7802)]
[New Thread 0xa1fffb70 (LWP 7803)]
[Thread 0xa44d0b70 (LWP 7802) exited]
[Thread 0xa511cb70 (LWP 7792) exited]
[Thread 0xa1fffb70 (LWP 7803) exited]
[Thread 0xa2effb70 (LWP 7801) exited]
[Thread 0xa3cd0b70 (LWP 7793) exited]
^C
Program received signal SIGINT, Interrupt.
0xb620977c in poll () from /lib/libc.so.6

(gdb) thread apply all bt full





+
Trace
    229195
















Comment 4


Milan Crha





          2011-12-02 09:58:53 UTC
        



Thanks for the update. After some investigation I realized that this is broken after (incomplete) changes in bug #658805, which touched both GLib and gtk.






Comment 5


Milan Crha





          2011-12-02 10:07:15 UTC
        



Created attachment 202580 [details] [review]
proposed evo patch

for evolution;

Could you try with this workaround, please? It may work, but one never knows, these days.






Comment 6


David Ronis





          2011-12-02 17:49:16 UTC
        



The patch works here too.  Thanks Milan.






Comment 7


Milan Crha





          2011-12-02 18:01:06 UTC
        



Thanks for testing.

Created commit 217c83b in evo master (3.3.3+)