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 667359 - random crash on config change ...
random crash on config change ...
Status: RESOLVED FIXED
Product: evolution-groupwise
Classification: Deprecated
Component: Calendar
3.2.x
Other Linux
: Normal blocker
: ---
Assigned To: Evolution Groupwise development team
Evolution QA team
Depends on:
Blocks:
 
 
Reported: 2012-01-05 14:07 UTC by Michael Meeks
Modified: 2012-06-27 11:59 UTC
See Also:
GNOME target: ---
GNOME version: ---

Attachments
Patch for evolution-groupwise (586 bytes, patch)
2012-01-17 09:38 UTC, Vibha Yadav 		needs-work Details | Review
Patch for evolution-groupwise master (1.57 KB, patch)
2012-01-17 09:43 UTC, Vibha Yadav 		committed Details | Review
Patch for evolution-groupwise 3-2 branch (4.91 KB, patch)
2012-02-23 11:36 UTC, Vibha Yadav 		needs-work Details | Review
Patch for evolution-groupwise (5.45 KB, patch)
2012-03-02 10:35 UTC, Vibha Yadav 		none Details | Review
Patch for evolution-groupwise (5.60 KB, patch)
2012-03-06 13:51 UTC, Vibha Yadav 		committed Details | Review

Description Michael Meeks 2012-01-05 14:07:38 UTC 
No idea what I did; I didn't knowingly change any config key: and bang.

Program received signal SIGSEGV, Segmentation fault.

+ Trace 229393

Thread 3 (Thread 0xb5cdfb70 (LWP 10571))

  • #0 read
    at ../sysdeps/unix/syscall-template.S line 82
  • #1 read
    at /usr/include/bits/unistd.h line 45
  • #2 unix_signal_helper_thread
    at gmain.c line 4551
  • #3 g_thread_create_proxy
    at gthread.c line 1962
  • #4 start_thread
    at pthread_create.c line 301
  • #5 clone
    at ../sysdeps/unix/sysv/linux/i386/clone.S line 133 






Comment 1


Michael Meeks





          2012-01-05 14:17:13 UTC
        



A trace with (perhaps) more helpful debuginfo for all threads: though my bugzilla helpfully hides the other threads from you (urgh):






+
Trace
    229394






Thread 3
          (Thread 0xb5cdfb70 (LWP 10571))


    

  • 
#0
read

    
                at ../sysdeps/unix/syscall-template.S
                  line
                  82

    

    • 

  • 
#1
read

    
                at /usr/include/bits/unistd.h
                  line
                  45

    

    • 

  • 
#2
unix_signal_helper_thread

    
                at gmain.c
                  line
                  4551

    

    • 

  • 
#3
g_thread_create_proxy

    
                at gthread.c
                  line
                  1962

    

    • 

  • 
#4
start_thread

    
                at pthread_create.c
                  line
                  301

    

    • 

  • 
#5
clone

    
                at ../sysdeps/unix/sysv/linux/i386/clone.S
                  line
                  133

    

    • 













Comment 2


Matthew Barnes





          2012-01-05 14:33:37 UTC
        



evolution-groupwise making GConf calls in Thread > 1.

GConfClient is known to not be thread-safe.






Comment 3


Michael Meeks





          2012-01-05 16:51:42 UTC
        



And now another similar bug; I switched week in week view:

Program received signal SIGSEGV, Segmentation fault.




+
Trace
    229395






Thread 2980051824
          (LWP 14984)


    

  • 
#0
link_before

    
                at dbus-list.c
                  line
                  122

    

    • 

  • 
#1
_dbus_list_prepend

    
                at dbus-list.c
                  line
                  290

    

    • 

  • 
#2
_dbus_list_append

    
                at dbus-list.c
                  line
                  262

    

    • 

  • 
#3
_dbus_validate_signature_with_reason

    
                at dbus-marshal-validate.c
                  line
                  174

    

    • 

  • 
#4
validate_body_helper

    • 

  • 
#5
validate_body_helper

    • 

  • 
#6
validate_body_helper

    
                at dbus-marshal-validate.c
                  line
                  644

    

    • 

  • 
#7
validate_body_helper

    
                at dbus-marshal-validate.c
                  line
                  494

    

    • 

  • 
#8
_dbus_validate_body_with_reason

    
                at dbus-marshal-validate.c
                  line
                  731

    

    • 

  • 
#9
_dbus_header_load

    
                at dbus-marshal-header.c
                  line
                  1007

    

    • 

  • 
#10
load_message

    
                at dbus-message.c
                  line
                  3996

    

    • 

  • 
#11
_dbus_message_loader_queue_messages

    
                at dbus-message.c
                  line
                  4197

    

    • 

  • 
#12
_dbus_transport_get_dispatch_status

    
                at dbus-transport.c
                  line
                  1103

    

    • 

  • 
#13
_dbus_transport_queue_messages

    
                at dbus-transport.c
                  line
                  1130

    

    • 

  • 
#14
do_reading

    
                at dbus-transport-socket.c
                  line
                  851

    

    • 

  • 
#15
do_reading

    
                at dbus-transport-socket.c
                  line
                  706

    

    • 

  • 
#16
socket_do_iteration

    
                at dbus-transport-socket.c
                  line
                  1162

    

    • 

  • 
#17
_dbus_transport_do_iteration

    
                at dbus-transport.c
                  line
                  978

    

    • 

  • 
#18
_dbus_connection_do_iteration_unlocked

    
                at dbus-connection.c
                  line
                  1232

    

    • 

  • 
#19
_dbus_connection_block_pending_call

    
                at dbus-connection.c
                  line
                  2405

    

    • 

  • 
#20
dbus_pending_call_block

    
                at dbus-pending-call.c
                  line
                  707

    

    • 

  • 
#21
dbus_connection_send_with_reply_and_block

    
                at dbus-connection.c
                  line
                  3514

    

    • 

  • 
#22
gconf_engine_get_fuller

    
                at gconf-dbus.c
                  line
                  1201

    

    • 

  • 
#23
gconf_engine_get_entry

    
                at gconf-dbus.c
                  line
                  1284

    

    • 

  • 
#24
get

    
                at gconf-client.c
                  line
                  1493

    

    • 

  • 
#25
gconf_client_get_full

    
                at gconf-client.c
                  line
                  1543

    

    • 

  • 
#26
gconf_client_get_bool

    
                at gconf-client.c
                  line
                  1778

    

    • 

  • 
#27
set_default_alarms

    
                at e-cal-backend-groupwise-utils.c
                  line
                  986

    

    • 

  • 
#28
e_gw_item_to_cal_component

    
                at e-cal-backend-groupwise-utils.c
                  line
                  1424

    

    • 

  • 
#29
get_deltas

    
                at e-cal-backend-groupwise.c
                  line
                  711

    

    • 

  • 
#30
delta_thread

    
                at e-cal-backend-groupwise.c
                  line
                  789

    

    • 

  • 
#31
g_thread_create_proxy

    
                at gthread.c
                  line
                  1962

    

    • 

  • 
#32
start_thread

    
                at pthread_create.c
                  line
                  301

    

    • 

  • 
#33
clone

    
                at ../sysdeps/unix/sysv/linux/i386/clone.S
                  line
                  133

    

    • 








all other threads quiescent - perhaps I'll run that beast under valgrind ...






Comment 4


Michael Meeks





          2012-01-05 17:16:24 UTC
        



confirming it then - we shouldn't use gconf from other threads; hopefully in the shiny new dconf world, that is no longer a problem. It's possible that my gconf is also too old:

gconf2-3.2.0-2.2.2.i586






Comment 5


Vibha Yadav





          2012-01-17 09:38:31 UTC
        



Created attachment 205428 [details] [review]
Patch for evolution-groupwise

Setting up default alarm in main thread






Comment 6


Vibha Yadav





          2012-01-17 09:43:39 UTC
        



Created attachment 205429 [details] [review]
Patch for evolution-groupwise master

Patch using GSetting as gconf is deprecated now.

Not calling set_default_alarms in main thread as dconf is thread safe.






Comment 7


Chenthill P





          2012-02-17 05:44:50 UTC
        



Review of attachment 205429 [details] [review]:

please commit the patch to master and also to gnome-3-2 branch if applicable.






Comment 8


Chenthill P





          2012-02-17 05:46:16 UTC
        



Review of attachment 205428 [details] [review]:

You would need to get the values from gconf once while opening the calendar and later use that for setting default alarms to appointments or meetings. Down side is that any changes to gconf will only be reflected after a restart of eds which I think is livable, considering we have fixed it in upstream.






Comment 9


Vibha Yadav





          2012-02-23 11:36:54 UTC
        



Created attachment 208246 [details] [review]
Patch for evolution-groupwise 3-2 branch

Reworked on review comments






Comment 10


Vibha Yadav





          2012-02-23 11:39:01 UTC
        



Comment on attachment 205429 [details] [review]
Patch for evolution-groupwise master

created commit 444ca986 on master branch.






Comment 11


Frederic Crozat





          2012-02-27 15:08:12 UTC
        



any review for 3.2 branch patch, by any chance ?






Comment 12


Michael Meeks





          2012-02-27 15:46:01 UTC
        



yep - having a working groupwise backend would be really pleasant ;-)






Comment 13


Chenthill P





          2012-02-28 12:37:31 UTC
        



Review of attachment 208246 [details] [review]:

Please make the changes..

::: src/calendar/e-cal-backend-groupwise.c
@@ +1417,3 @@
+	priv->default_reminder = gconf_client_get_bool (gconf, CALENDAR_CONFIG_DEFAULT_REMINDER, NULL);
+	priv->interval = gconf_client_get_int (gconf, CALENDAR_CONFIG_DEFAULT_REMINDER_INTERVAL, NULL);
+	priv->units = gconf_client_get_string (gconf, CALENDAR_CONFIG_DEFAULT_REMINDER_UNITS, NULL);

This is still getting called in a thread as groupwise_open runs in a thread. Please use the main thread to get these values once. Using e_flag to syncronize the threads would be simple.






Comment 14


Vibha Yadav





          2012-03-02 10:35:48 UTC
        



Created attachment 208844 [details] [review]
Patch for evolution-groupwise

Reworked on above comments.
Added the code for getting gconf entries in a function that gets called in main thread.
Using EFlag to synchronize the threads.






Comment 15


Vibha Yadav





          2012-03-06 13:51:24 UTC
        



Created attachment 209085 [details] [review]
Patch for evolution-groupwise

Updated Patch.






Comment 16


Vibha Yadav





          2012-03-06 14:02:07 UTC
        



Comment on attachment 209085 [details] [review]
Patch for evolution-groupwise

created commit 75b99a1 to gnome-3-2 branch