Bug 578954 – Crash in mapi_object_table_init at libmapi/mapi_object.c line 311

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 578954 - Crash in mapi_object_table_init at libmapi/mapi_object.c line 311


Summary:	Crash in mapi_object_table_init at libmapi/mapi_object.c line 311


Status:	RESOLVED DUPLICATE of bug 627999

Product:	evolution-mapi
Classification:	Applications
Component:	miscellaneous
Version:	0.30.x
Hardware:	Other All

Importance:	High critical
Target Milestone:	---
Assigned To:	evolution-mapi-maint
QA Contact:	evolution-mapi-maint

URL:
Whiteboard:	evolution[mapi]

Duplicates:	529392 608050 (view as bug list)
Depends on:
Blocks:

Reported:	2009-04-14 17:51 UTC by idht4n
Modified:	2010-10-13 07:53 UTC

See Also:
GNOME target:	---
GNOME version:	2.29/2.30

Attachments
oc test change (4.65 KB, text/plain) 2010-06-08 09:06 UTC, Milan Crha	Details

Description idht4n 2009-04-14 17:51:41 UTC

What were you doing when the application crashed?
I had just started evolution when it crashed


Distribution: Fedora release 10.92 (Rawhide)
Gnome Release: 2.26.0 2009-03-16 (Red Hat, Inc)
BugBuddy Version: 2.26.0

System: Linux 2.6.29.1-52.fc11.i586 #1 SMP Mon Apr 6 03:25:10 EDT 2009 i686
X Vendor: The X.Org Foundation
X Vendor Release: 10600000
Selinux: Permissive
Accessibility: Disabled
GTK+ Theme: Nodoka
Icon Theme: Fedora
GTK+ Modules: canberra-gtk-module, pk-gtk-module, gnomebreakpad

Memory status: size: 188821504 vsize: 188821504 resident: 30924800 share: 22278144 rss: 30924800 rss_rlim: 18446744073709551615
CPU usage: start_time: 1239731184 rtime: 149 utime: 124 stime: 25 cutime:0 cstime: 0 timeout: 0 it_real_value: 0 frequency: 100

Backtrace was generated from '/usr/bin/evolution'

[Thread debugging using libthread_db enabled]
[New Thread 0xb0ae7b70 (LWP 2936)]
[New Thread 0xb15eeb70 (LWP 2935)]
[New Thread 0xb2b38b70 (LWP 2896)]
[New Thread 0xb46fcb70 (LWP 2894)]
[New Thread 0xb64ffb70 (LWP 2892)]
[New Thread 0xb50fdb70 (LWP 2890)]
[New Thread 0xb5afeb70 (LWP 2889)]
[New Thread 0xb70ffb70 (LWP 2887)]
[New Thread 0xb7c17b70 (LWP 2886)]
0x00fac422 in __kernel_vsyscall ()

+ Trace 214458

Thread 6 (Thread 0xb64ffb70 (LWP 2892))

#0 __kernel_vsyscall
#1 __lll_lock_wait
at ../nptl/sysdeps/unix/sysv/linux/i386/i486/lowlevellock.S line 142
#2 _L_lock_752
from /lib/libpthread.so.0
#3 __pthread_mutex_lock
at pthread_mutex_lock.c line 61
#4 segv_redirect
at main.c line 435
#5 <signal handler called>
#6 mapi_object_table_init
at libmapi/mapi_object.c line 311
#7 GetHierarchyTable
at libmapi/IMAPIContainer.c line 270
#8 get_child_folders
at exchange-mapi-connection.c line 2455
#9 exchange_mapi_get_folders_list
at exchange-mapi-connection.c line 2723
#10 mapi_folders_sync
at camel-mapi-store.c line 1115
#11 mapi_get_folder_info
at camel-mapi-store.c line 1219
#12 camel_store_get_folder_info
at camel-store.c line 894
#13 get_folderinfo_exec
at mail-ops.c line 1069
#14 mail_msg_proxy
at mail-mt.c line 520
#15 g_thread_pool_thread_proxy
at gthreadpool.c line 265
#16 g_thread_create_proxy
at gthread.c line 635
#17 start_thread
at pthread_create.c line 297
#18 clone
at ../sysdeps/unix/sysv/linux/i386/clone.S line 130

Comment 1 Akhil Laddha 2009-04-15 04:39:01 UTC

are you using mapi provider ?

Comment 2 Milan Crha 2009-04-16 15:22:04 UTC

(In reply to comment #1)
> are you using mapi provider ? 

Obviously :) See Thread 6

Comment 3 Paul Smith 2009-04-22 04:17:26 UTC

I'm seeing something similar as well; see: http://mail.gnome.org/archives/evolution-list/2009-April/msg00113.html

Comment 4 Akhil Laddha 2009-09-02 09:38:09 UTC

*** Bug 529392 has been marked as a duplicate of this bug. ***

Comment 5 Akhil Laddha 2010-01-28 03:47:36 UTC

*** Bug 608050 has been marked as a duplicate of this bug. ***

Comment 6 Akhil Laddha 2010-01-28 03:49:04 UTC

bug 608050 was filed while using libmapi 0.9

Comment 7 dwcatcvg 2010-01-28 16:35:37 UTC

FYI, I built evolution-mapi 0.28.2 locally and ran with the library I compiled -- using evolution * and openchange libmapi.so version 0.9 (you can see in my original bug, 608050, that I first encountered the SEGV using libmapi.so version 0.8.2, but also encountered it with v0.9).

The locally built libexchangemapi from 0.28.2 does NOT core.  If it helps anyone, I've only tried this on i386 Ubuntu 9.10.

Sadly, evolution-mapi doesn't work well with my Exchange 2007 server.  I can see folders, and I can get some folder details (like unread message counts, and folder size), but I cannot see what's in any of the folders.  Evolution never finished fetching summary information or refreshing folders.  I'm also unable to send messages (sits at 100% complete, but never finishes).

Again, openchangeclient works great with my server, so I'm hopeful that evolution-mapi gets there soon.  Obviously the source of the SEGV in the Ubuntu package is still unknown.

Comment 8 Akhil Laddha 2010-04-13 06:02:15 UTC

evolution-mapi 0.31.1

Started evolution after doing force shutdown and it crashed.

Program received signal SIGABRT, Aborted.
[Switching to Thread 0xb27c0b70 (LWP 23297)]
0xffffe430 in __kernel_vsyscall ()
(gdb) t a a bt

+ Trace 221350

Thread 15 (Thread 0xa89cfb70 (LWP 23305))

#0 clone
from /lib/libc.so.6
#1 ??
#2 ??

Thread 1 (Thread 0xb62f1860 (LWP 23289))

#0 IA__g_value_init
at gvalue.c line 168
#1 gtk_list_store_set_valist_internal
at gtkliststore.c line 800
#2 IA__gtk_list_store_set_valist
at gtkliststore.c line 899
#3 IA__gtk_list_store_set
at gtkliststore.c line 939
#4 shortcuts_insert_file
at gtkfilechooserdefault.c line 1664
#5 shortcuts_append_home
at gtkfilechooserdefault.c line 1820
#6 shortcuts_model_create
at gtkfilechooserdefault.c line 2282
#7 gtk_file_chooser_default_constructor
at gtkfilechooserdefault.c line 5201
#8 IA__g_object_newv
at gobject.c line 1261
#9 IA__g_object_new
at gobject.c line 1092
#10 _gtk_file_chooser_default_new
at gtkfilechooserdefault.c line 10095
#11 gtk_file_chooser_widget_constructor
at gtkfilechooserwidget.c line 103
#12 IA__g_object_newv
at gobject.c line 1261
#13 IA__g_object_new
at gobject.c line 1092
#14 gtk_file_chooser_dialog_constructor
at gtkfilechooserdialog.c line 278
#15 IA__g_object_newv
at gobject.c line 1261
#16 IA__g_object_new_valist
at gobject.c line 1377
#17 IA__g_object_new
at gobject.c line 1095
#18 gtk_file_chooser_dialog_new_valist
at gtkfilechooserdialog.c line 419
#19 IA__gtk_file_chooser_dialog_new
at gtkfilechooserdialog.c line 463
#20 gtk_file_chooser_button_constructor
at gtkfilechooserbutton.c line 648
#21 IA__g_object_newv
at gobject.c line 1261
#22 _gtk_builder_construct
at gtkbuilder.c line 438
#23 builder_construct
at gtkbuilderparser.c line 197
#24 end_element
at gtkbuilderparser.c line 981
#25 IA__g_markup_parse_context_parse
at gmarkup.c line 1498
#26 _gtk_builder_parser_parse_buffer
at gtkbuilderparser.c line 1162
#27 IA__gtk_builder_add_from_file
at gtkbuilder.c line 674
#28 e_load_ui_builder_definition
at e-util.c line 333
#29 em_mailer_prefs_construct
at em-mailer-prefs.c line 853
#30 em_mailer_prefs_new
at em-mailer-prefs.c line 1233
#31 mail_shell_backend_init_preferences
at e-mail-shell-backend.c line 199
#32 g_idle_dispatch
at gmain.c line 4065
#33 g_main_dispatch
at gmain.c line 1960
#34 IA__g_main_context_dispatch
at gmain.c line 2513
#35 g_main_context_iterate
at gmain.c line 2591
#36 IA__g_main_loop_run
at gmain.c line 2799
#37 IA__gtk_main
at gtkmain.c line 1219
#38 main
at main.c line 550
#0 __kernel_vsyscall
#1 raise
from /lib/libc.so.6
#2 abort
from /lib/libc.so.6
#3 talloc_abort
at ../lib/talloc/talloc.c line 199
#4 talloc_abort_double_free
at ../lib/talloc/talloc.c line 218
#5 talloc_chunk_from_ptr
at ../lib/talloc/talloc.c line 239
#6 _talloc_free_internal
at ../lib/talloc/talloc.c line 570
#7 _talloc_free_internal
at ../lib/talloc/talloc.c line 631
#8 _talloc_free_internal
at ../lib/talloc/talloc.c line 631
#9 _talloc_free
at ../lib/talloc/talloc.c line 1133
#10 QueryRows
at libmapi/IMAPITable.c line 313
#11 get_child_folders
at exchange-mapi-connection.c line 2780
#12 get_child_folders
at exchange-mapi-connection.c line 2816
#13 get_child_folders
at exchange-mapi-connection.c line 2816
#14 exchange_mapi_connection_get_folders_list
at exchange-mapi-connection.c line 3035
#15 mapi_folders_sync
at camel-mapi-store.c line 1393
#16 mapi_get_folder_info
at camel-mapi-store.c line 1577
#17 camel_store_get_folder_info
at camel-store.c line 895
#18 get_folderinfo_exec
at mail-ops.c line 1090
#19 mail_msg_proxy
at mail-mt.c line 468
#20 g_thread_pool_thread_proxy
at gthreadpool.c line 315
#21 g_thread_create_proxy
at gthread.c line 1893
#22 start_thread
at pthread_create.c line 297
#23 ??
at pthread_create.c line 216

Comment 9 Milan Crha 2010-06-08 09:06:50 UTC

Created attachment 163029 [details]
oc test change

I tried to reproduce this with actual evolution-mapi git master (0.31.3) and openchange svn trunk (at revision 1889), but no luck. I also tried to modify one openchange example (see the attachment) which opens folder and gets folder hierarchy in a loop, but no luck, it doesn't want to crash. I also tried with valgrind and it doesn't report any issue on the example program to me, like is mentioned in bug #608050.

Can anyone able to reproduce this give a try to openchange svn trunk and evolution-mapi git master, please? (If they will release openchange 0.10 early enough we will depend on it with evolution-mapi in Gnome 3.0.)

Comment 10 idht4n 2010-06-08 15:34:09 UTC

I filed that bug 14 months ago, so I don't really remember the details.  Evolution crashes on me all the time, and I don't have time to look through the backtrace and figure out which crashes go with which bug reports.  I use abrt in Fedora now to report some of the crashes I see.  Two that I've seen recently are here:

https://bugzilla.redhat.com/show_bug.cgi?id=591937
https://bugzilla.redhat.com/show_bug.cgi?id=580511

I don't think either of these are the same as that bug I saw 14 months ago, but I'm not sure.  I think this bug (578954) was mapi specific, right?  I stopped using mapi a year ago because it seemed to crash a lot and was extremely slow.  I tried it again in fedora 13 and it seemed better, but I'm not using it regularly because I can't seem to configure the "Sent Messages Folder" (edit->preferences->edit->defaults->Sent Messages Folder has "Sent" as the selection and it is greyed out so I can't change it).  If you let me know of a way to change that, I'll switch from exchange connector to mapi and file any crashes I see with abrt.  Otherwise, I just can't use mapi enough to confidently say if it is crashing on me.

Comment 11 Milan Crha 2010-06-09 08:52:51 UTC

evolution-mapi is using your mapi server's Sent Items folder automatically, that's why the Sent folder is grayed out in preferences.

Comment 12 idht4n 2010-06-14 19:13:50 UTC

(In reply to comment #11)
> evolution-mapi is using your mapi server's Sent Items folder automatically,
> that's why the Sent folder is grayed out in preferences.

OK... I don't really like that "feature" because I like to configure where the sent messages, go, but I'll deal with it (the thing that is grayed out should really say Sent Items and not just Sent because I have a Sent folder too).

In any case, I've been using mapi for about 3 days now and the only crashes I've noticed have been after suspends.  I have been unsuccessful in reporting one of those crashes with abrt so far.

I've also seen cases where evolution seems to have crashed and then automatically restarted itself...  the reason I say this is that I'm occasionally prompted to recover emails I was composing even though I didn't see evolution crash or exit evolution.

Comment 13 Tobias Mueller 2010-07-31 11:41:57 UTC

Reopening as I don't see any open question.

Comment 14 Milan Crha 2010-10-13 07:53:46 UTC

Revisiting this bug I believe this got addressed within bug #627999, where was added global locking for libmapi calls, thus they cannot overlap like in this case. Due to this change there I'm marking this as a duplicate of that bug. Please reopen if you'll see this with 0.32.0 or later evolution-mapi. Thanks in advance.

*** This bug has been marked as a duplicate of bug 627999 ***