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 361145 - Evolution hangs when formatting message
Evolution hangs when formatting message
Status: RESOLVED FIXED
Product: evolution
Classification: Applications
Component: Mailer
3.4.x (obsolete)
Other All
: High major
: ---
Assigned To: evolution-mail-maintainers
Evolution QA team
: 256479 342910 352984 363174 372456 502665 516314 568743 578577 593692 610637 (view as bug list)
Depends on: 578577
Blocks:
 
 
Reported: 2006-10-10 13:48 UTC by Havard Bjastad
Modified: 2013-08-06 02:12 UTC
See Also:
GNOME target: ---
GNOME version: 2.25/2.26


Attachments
Stack traces for all evolution processes at time of hang (59.84 KB, text/plain)
2007-10-10 21:24 UTC, Wayne A Feick
Details
Stack traces for evolution process during another hang (6.17 KB, text/plain)
2007-10-10 23:19 UTC, Wayne A Feick
Details
The 3 debug traces of evolution by thread (2.90 KB, application/x-compressed-tar)
2009-03-30 17:40 UTC, Dragan Tomas
Details
Backtraces from 2.26.0-0ubuntu2 jammed while formatting (10.62 KB, text/plain)
2009-04-09 07:18 UTC, Michael Gratton
Details
back trace from eds, 2.26.0-0ubuntu2 (near svn head) (5.93 KB, text/plain)
2009-04-09 11:10 UTC, Michael Gratton
Details
back trace from evo, 2.26.0-0ubuntu2 (near svn head) (10.63 KB, text/plain)
2009-04-09 11:12 UTC, Michael Gratton
Details

Description Havard Bjastad 2006-10-10 13:48:45 UTC
Please describe the problem:
While browsing through e-mails, evolution appears to get 'stuck' on a
particular message, refusing to update the preview pane when I select
another message. I can still double-click on a message to open it
explicitly in a new window, and I can switch freely to other components.
On coming back to the mail-reader, however, the preview pane remains
stuck. 

Steps to reproduce:
I haven't been able to determine a way of consistently reproducing, but it happens just about every hour...

Actual results:
Evolution hangs when formatting message

Expected results:
Preview view does not freeze.

Does this happen every time?
Just about every hour.

Other information:
Seems to only happen with the Exchange account, not with any of the others.

Maybe related to http://bugzilla.gnome.org/show_bug.cgi?id=259496 and http://bugzilla.gnome.org/show_bug.cgi?id=256479 ?
Comment 1 Havard Bjastad 2006-10-11 13:13:23 UTC
This is actually Evolution 2.6.1 (Ubuntu), but 2.6.0 is the closest in the version list.
Comment 2 Havard Bjastad 2006-10-24 10:05:58 UTC
I thought this only happened to the Exchange account, but it now happened to the local Inbox also :-(
However, I never saw the behaviour until I added the Exchange account, so I'm pretty sure it has to do with the Exchange connector...
Comment 3 Havard Bjastad 2006-11-21 11:06:04 UTC
After upgrading to Ubuntu 6.10 (Evolution 2.8.1), the problem still exists.
Comment 4 Havard Bjastad 2007-06-27 08:20:20 UTC
After upgrading to Ubuntu 7.04 (Evolution 2.10.1), the problem still exists. Is this bugzilla in use, or am I reporting to a repository that nobody reads?
Comment 5 Havard Bjastad 2007-08-10 12:37:42 UTC
The one year anniversary of this bug is approaching, without anyone touching it. I've now migrated 2 of my 3 mailboxes to Thunderbird, and on the one year mark I will migrate the last one...
Comment 6 Wayne A Feick 2007-10-10 20:39:31 UTC
I see this frequently as well on Fedora 7. I don't have debug versions installed, but below is what I get from gdb.

(gdb) info threads
  6 Thread 1115699536 (LWP 11719)  0x000000381d0c82e6 in poll ()
   from /lib64/libc.so.6
  5 Thread 1157925200 (LWP 11724)  0x000000381d0c82e6 in poll ()
   from /lib64/libc.so.6
  4 Thread 1199884624 (LWP 11756)  0x000000381d0c82e6 in poll ()
   from /lib64/libc.so.6
  3 Thread 1094719824 (LWP 12563)  0x000000381e00a486 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib64/libpthread.so.0
  2 Thread 1126189392 (LWP 12583)  0x000000381e00c808 in __lll_mutex_lock_wait
    () from /lib64/libpthread.so.0
  1 Thread 46912496410768 (LWP 11696)  0x000000381d0c82e6 in poll ()
   from /lib64/libc.so.6
(gdb) thread 2
[Switching to thread 2 (Thread 1126189392 (LWP 12583))]#0  0x000000381e00c808 in __lll_mutex_lock_wait () from /lib64/libpthread.so.0
(gdb) where
  • #0 __lll_mutex_lock_wait
    from /lib64/libpthread.so.0
  • #1 _L_mutex_lock_103
    from /lib64/libpthread.so.0
  • #2 pthread_mutex_lock
    from /lib64/libpthread.so.0
  • #3 em_utils_in_addressbook
    from /usr/lib64/evolution/2.10/components/libevolution-mail.so
  • #4 ??
    from /usr/lib64/evolution/2.10/components/libevolution-mail.so
  • #5 ??
    from /usr/lib64/evolution/2.10/components/libevolution-mail.so
  • #6 ??
    from /usr/lib64/evolution/2.10/components/libevolution-mail.so
  • #7 ??
    from /lib64/libglib-2.0.so.0
  • #8 ??
    from /lib64/libglib-2.0.so.0
  • #9 start_thread
    from /lib64/libpthread.so.0
  • #10 clone
    from /lib64/libc.so.6
  • #0 pthread_cond_wait
    from /lib64/libpthread.so.0
  • #1 ??
    from /usr/lib64/libebook-1.2.so.9
  • #2 e_book_get_contacts
    from /usr/lib64/libebook-1.2.so.9
  • #3 em_utils_in_addressbook
    from /usr/lib64/evolution/2.10/components/libevolution-mail.so
  • #4 ??
    from /usr/lib64/evolution/2.10/components/libevolution-mail.so
  • #5 ??
    from /usr/lib64/evolution/2.10/components/libevolution-mail.so
  • #6 ??
    from /usr/lib64/evolution/2.10/components/libevolution-mail.so
  • #7 ??
    from /lib64/libglib-2.0.so.0
  • #8 ??
    from /lib64/libglib-2.0.so.0
  • #9 start_thread
    from /lib64/libpthread.so.0
  • #10 clone
    from /lib64/libc.so.6

Comment 7 Wayne A Feick 2007-10-10 21:24:55 UTC
Created attachment 97033 [details]
Stack traces for all evolution processes at time of hang
Comment 8 Wayne A Feick 2007-10-10 21:27:20 UTC
Evolution has been hanging a lot today, so I installed the debug rpms from Fedora 7 updates and attached stack traces from all of the evolution processes.


Also, doing some searching, bug 471788 looks to be a recently reported duplicate of this.
Comment 9 Wayne A Feick 2007-10-10 23:19:46 UTC
Created attachment 97037 [details]
Stack traces for evolution process during another hang
Comment 10 Wayne A Feick 2007-10-10 23:25:46 UTC
I've added a second attachment, showing stack traces for just the evolution process that occurred later this afternoon (pared down to remove all the info about loading various shared objects). 

One thing to note is that the hang doesn't affect other message viewer windows, so a work around is to open a new window and close the old one.

I'm running two email accounts, one is secure IMAP and the other is using the exchange connector. The hangs seem to only be occurring today when viewing my IMAP inbox.

I'm happy to help debug this further if need be, but I don't know enough (well, anything) about the internals of evolution to be tempted to push ahead on my own.
Comment 11 Havard Bjastad 2007-10-11 07:24:07 UTC
Hi Wayne, seems like we're the only 2 persons interested in this bug - at least the Evo developers won't touch it with a ten foot pole...
One thing I've noticed over the year that's gone since first registering the bug, is that it only seems to happen for HTML emails.
Comment 12 Hok Hie Tjioe 2007-10-11 10:38:50 UTC
(In reply to comment #11)
> Hi Wayne, seems like we're the only 2 persons interested in this bug - at least
> the Evo developers won't touch it with a ten foot pole...
> One thing I've noticed over the year that's gone since first registering the
> bug, is that it only seems to happen for HTML emails.
> 

Hi Wayne and Harvard,

You are not the only two persons. I have also the same problem with the Evolution. I have opened a new bug ID in the first stage (number: 471788), but Wayne has noticed me that this is a dup of this bug id..

Harvard, like you mentioned before I have indeed problem with Exchange server. And it is only apply to HTML emails. This is the only account that I use today. I haven't try to add other account though..

Hopefully that the developer can help us further in this case???

.h2t.
Comment 13 mark.greenbank 2007-10-22 13:50:44 UTC
Hi, several suggestions on the net yield no help. 

One has to set EVOLUTION_NO_BAR=1, the other was to disable loading of images. Neither one has solved the problem for me.
Comment 14 Wayne A Feick 2007-10-22 15:15:15 UTC
I cleared the "Global Address List / Active Directory" setting and this got a lot better. Have you tried that?
Comment 15 mark.greenbank 2007-10-22 15:41:03 UTC
I'm not exactly sure what you mean by Global Address List / Active Directory but I'm not using LDAP or exchange or anything -- all of my accounts are IMAP accounts.
Comment 16 Wayne A Feick 2007-10-22 16:25:48 UTC
Okay, nevermind then I guess. I have one IMAP account and one Exchange account.
Comment 17 Johan Bergström 2007-12-12 13:46:30 UTC
(In reply to comment #14)
> I cleared the "Global Address List / Active Directory" setting and this got a
> lot better. Have you tried that?
> 

Trying that now, looks like it's working alot better.
Comment 18 Michael Meeks 2008-01-07 13:06:31 UTC
I guess, ultimately that the image loading needs to be far more asynchronous - and performing the (potentially nastily slow) addressbook lookup on the sender needs to be pushed out until we actually need to see the image.
Comment 19 Russell Harrison 2008-02-19 19:52:28 UTC
I'm also seeing this problem with evolution 2.12.3

Other bugs which may be duplicates of this one are:
http://bugzilla.gnome.org/show_bug.cgi?id=361145
http://bugzilla.gnome.org/show_bug.cgi?id=342910
http://bugzilla.gnome.org/show_bug.cgi?id=502665

Seems to be a very common problem.

$ rpm -qa | grep evolution
evolution-exchange-2.12.3-1.fc8
evolution-exchange-debuginfo-2.12.3-1.fc8
evolution-data-server-1.12.3-2.fc8
evolution-remove-duplicates-0.0.3-1.fc8
evolution-sharp-debuginfo-0.14.0.1-1.fc8
libopensync-plugin-evolution2-0.22-2.fc8
evolution-remove-duplicates-debuginfo-0.0.3-1.fc8
evolution-2.12.3-1.fc8
evolution-webcal-2.12.0-1.fc8
evolution-help-2.12.3-1.fc8
evolution-data-server-debuginfo-1.12.3-2.fc8
syncevolution-0.6-2.fc8
evolution-sharp-0.14.0.1-1.fc8
libopensync-plugin-evolution2-debuginfo-0.22-2.fc8
evolution-python-0.0.4-2.fc8
evolution-python-debuginfo-0.0.4-2.fc8
evolution-webcal-debuginfo-2.12.0-1.fc8
evolution-debuginfo-2.12.3-1.fc8
beagle-evolution-0.2.18-5.fc8
Comment 20 Matthew Barnes 2008-03-16 21:56:14 UTC
*** Bug 342910 has been marked as a duplicate of this bug. ***
Comment 21 Matthew Barnes 2008-03-16 21:56:47 UTC
*** Bug 502665 has been marked as a duplicate of this bug. ***
Comment 22 Ricardo Fernández Pascual 2008-06-11 17:24:31 UTC
I still see this bug with Evolution 2.22.2 in Fedora 9. I happens very often, always with HTML mail, but it is mostly unpredictable. I don't use the exchange component, so I doubt it has anything to do with this.
Comment 23 Akhil Laddha 2008-07-11 11:38:29 UTC
see bug 524338
Comment 24 Brian J. Murrell 2008-07-11 16:58:19 UTC
(In reply to comment #23)
> see bug 524338

I don't think this is related.
Comment 25 Steven Willis 2008-09-11 17:04:03 UTC
I just wanted to verify that this is still happening in 2.22.3.1 in Ubuntu Hardy (8.04), as I reported here:

https://bugs.launchpad.net/evolution/+bug/175233

This is the relevant portion of the backtrace from my post:

Program received signal SIGINT, Interrupt.
[Switching to Thread 0xb6529740 (LWP 9916)]
0xb7f9e410 in __kernel_vsyscall ()
(gdb) thread apply all bt

Thread 1273 (Thread 0xaedfdb90 (LWP 15161))

  • #0 __kernel_vsyscall
  • #1 pthread_cond_wait
    from /lib/tls/i686/cmov/libpthread.so.0
  • #2 do_get_contacts
    at e-book.c line 2140
  • #3 em_utils_contact_photo
    at em-utils.c line 2101
  • #4 efh_format_message
    at em-format-html.c line 1952
  • #5 efh_format_exec
    at em-format-html.c line 1261
  • #6 mail_msg_proxy
    at mail-mt.c line 523
  • #7 g_thread_pool_thread_proxy
    at /build/buildd/glib2.0-2.16.4/glib/gthreadpool.c line 265
  • #8 g_thread_create_proxy
    at /build/buildd/glib2.0-2.16.4/glib/gthread.c line 635
  • #9 start_thread
    from /lib/tls/i686/cmov/libpthread.so.0
  • #10 clone
    from /lib/tls/i686/cmov/libc.so.6

Comment 26 Bharath Acharya 2008-09-22 06:46:52 UTC
Looks like http://svn.gnome.org/viewvc/evolution?view=revision&revision=33613 is the culprit. Could you also get the traces of evolution-data-server process for this hang? I think that would spill a lot more erroneous data.
Comment 27 Milan Crha 2008-09-22 10:22:27 UTC
There has been added an option in Edit->Preferences->Mail Preference, tab "Headers", to indicate whether only local address books are supposed to be scanned for a contact.

To Havard: I see your pain, but there are similar bugs as yours which could have some info from developers. The problem is that there are only a few people triaging all evolution bug reports and trying to "pair" them.

I also think I did some patch in this area in other bug, but I cannot find it at the moment, I'm sorry.
Comment 28 hh 2008-11-14 09:13:57 UTC
the problem inevitybly appears when changing IP adresses during evolution open. whenever i start vpnc, i get "formatting message ..." even if i wait fo hours. evolution can't be closed regularly but has to be killed. it also appears after weaking my pc from suspend.  
Comment 29 Akhil Laddha 2008-11-17 09:42:59 UTC
changing product version as per comment#25
Comment 30 Jim 2008-11-18 19:15:36 UTC
I also am being severely hampered in using Evolution by this bug.  It seems to almost be reproducible in my case.  I am using Fedora 9 64 bit, connecting to a remote imap server.  If I am doing anything when it is checking for new messages, I seem to have a better chance of it hanging.  It will happen even without doing anything though.

What it feels like, and this is just a hypothesis, is that if the remote server doesn't respond fast enough to requests, it hits a timeout and doesn't know what to do.  I say this, because it seems more prevalent at times of high mail server load.  It also seems to happen more frequently when the default mail message is large.  That is, say you have deleted a bunch of messages, the active message after the operations is loaded by the client.  If this message is large, hangs seem more prone to happen

Additionally, my mailbox that is managed by evolution is 543M.  I would say the mail left on the imap server is probably 50-100M

If I can be of further assistance, I will do what I can.
Comment 31 Milan Crha 2008-11-19 09:34:59 UTC
Hi Jim, thanks for the information. Can you install debug info packages for evolution, evolution-data-server, gtkhtml and when it freezes, attach to the running evolution with gdb and paste here result of 'thread apply all bt', please? It'll show where exactly it froze. Thanks.
Comment 32 Jim 2008-11-19 15:53:29 UTC

Here is what I am doing:


rpm -i 

evolution-data-server-debuginfo-2.22.3-2.fc9.x86_64.rpm
evolution-debuginfo-2.22.3.1-1.fc9.x86_64.rpm 
gtkhtml3-debuginfo-3.18.3-1.fc9.x86_64.rpm

and will run the command when it hangs again.

Thank you.
Comment 33 Jim 2008-11-19 19:46:01 UTC
(gdb) thread apply all bt

Thread 5 (Thread 0x42848950 (LWP 27945))

  • #0 __lll_lock_wait
    from /lib64/libpthread.so.0
  • #1 _L_lock_100
    from /lib64/libpthread.so.0
  • #2 pthread_mutex_lock
    from /lib64/libpthread.so.0
  • #3 g_static_rec_mutex_lock
    from /lib64/libglib-2.0.so.0
  • #4 imap_refresh_info
    at camel-imap-folder.c line 583
  • #5 camel_folder_refresh_info
    at camel-folder.c line 302
  • #6 mail_msg_proxy
    at mail-mt.c line 523
  • #7 ??
    from /lib64/libglib-2.0.so.0
  • #8 ??
    from /lib64/libglib-2.0.so.0
  • #9 start_thread
    from /lib64/libpthread.so.0
  • #10 clone
    from /lib64/libc.so.6

Thread 4 (Thread 0x43249950 (LWP 27946))

  • #0 __lll_lock_wait
    from /lib64/libpthread.so.0
  • #1 _L_lock_100
    from /lib64/libpthread.so.0
  • #2 pthread_mutex_lock
    from /lib64/libpthread.so.0
  • #3 g_static_rec_mutex_lock
    from /lib64/libglib-2.0.so.0
  • #4 camel_folder_refresh_info
    at camel-folder.c line 301
  • #5 do_copy
    at camel-imap-folder.c line 1744
  • #6 imap_transfer_online
    at camel-imap-folder.c line 1844
  • #7 do_move
    at camel-filter-driver.c line 541
  • #8 e_sexp_term_eval
    at e-sexp.c line 710
  • #9 term_eval_begin
    at e-sexp.c line 654
  • #10 e_sexp_term_eval
    at e-sexp.c line 700
  • #11 e_sexp_eval
    at e-sexp.c line 1306
  • #12 camel_filter_driver_filter_message
    at camel-filter-driver.c line 1499
  • #13 filter_filter
    at camel-folder.c line 1717
  • #14 session_thread_proxy
    at camel-session.c line 587
  • #15 ??
    from /lib64/libglib-2.0.so.0
  • #16 ??
    from /lib64/libglib-2.0.so.0
  • #17 start_thread
    from /lib64/libpthread.so.0
  • #18 clone
    from /lib64/libc.so.6

Thread 3 (Thread 0x43c4a950 (LWP 27971))

  • #0 __lll_lock_wait
    from /lib64/libpthread.so.0
  • #1 _L_lock_100
    from /lib64/libpthread.so.0
  • #2 pthread_mutex_lock
    from /lib64/libpthread.so.0
  • #3 g_static_rec_mutex_lock
    from /lib64/libglib-2.0.so.0
  • #4 camel_imap_folder_fetch_data
    at camel-imap-folder.c line 3074
  • #5 get_message_simple
    at camel-imap-folder.c line 2287
  • #6 imap_get_message
    at camel-imap-folder.c line 2427
  • #7 camel_folder_get_message
    at camel-folder.c line 1076
  • #8 get_message_exec
    at mail-ops.c line 1720
  • #9 mail_msg_proxy
    at mail-mt.c line 523
  • #10 ??
    from /lib64/libglib-2.0.so.0
  • #11 ??
    from /lib64/libglib-2.0.so.0
  • #12 start_thread
    from /lib64/libpthread.so.0
  • #13 clone
    from /lib64/libc.so.6

Thread 2 (Thread 0x4464b950 (LWP 28197))

  • #0 __lll_lock_wait
    from /lib64/libpthread.so.0
  • #1 _L_lock_100
    from /lib64/libpthread.so.0
  • #2 pthread_mutex_lock
    from /lib64/libpthread.so.0
  • #3 g_static_rec_mutex_lock
    from /lib64/libglib-2.0.so.0
  • #4 camel_folder_refresh_info
    at camel-folder.c line 301
  • #5 refresh_folders_exec
    at mail-send-recv.c line 825
  • #6 mail_msg_proxy
    at mail-mt.c line 523
  • #7 ??
    from /lib64/libglib-2.0.so.0
  • #8 ??
    from /lib64/libglib-2.0.so.0
  • #9 start_thread
    from /lib64/libpthread.so.0
  • #10 clone
    from /lib64/libc.so.6

Comment 34 Milan Crha 2008-11-20 10:10:53 UTC
(In reply to comment #33)
>...

Thanks for the update. This particular freeze has been fixed recently in bug #558737 and will be included in the 2.24.2.
Comment 35 Jim 2008-11-20 13:47:11 UTC
good news, thank you.  I guess I will wait for that to hit the testing repos
and update then.  After that, if I have more hangs, I'll come back.
Comment 36 Jim 2008-11-20 13:56:07 UTC
(In reply to comment #35)
> good news, thank you.  I guess I will wait for that to hit the testing repos
> and update then.  After that, if I have more hangs, I'll come back.
> 

One other question, from the gdb output, what in particular tells the story of the problem I am having, so that if it hangs again, but it feels different I can run gdb again and look myself to see if it is the same bug or not?

Long question, I realize.  I just don't want to waste anyone's time here for a bug that is fixed, but don't want to miss something that may help this issue.
Comment 37 Milan Crha 2008-11-20 14:21:00 UTC
Jim, no worries, it's sometimes better to pair problems together than overlook one.

This one is about locking, and is because the Thread 4 tries to copy your message in filtering from one folder to another and requests lock on the folder, which holds other thread, probably Thread 2, which is waiting for release of the other lock held in Thread 4. Basically, if you see same functions like in Thread 4, even after the 2.24.2, then something goes wrong and the fix isn't correct.
Comment 38 zioalex 2008-12-15 15:01:59 UTC
Some Problem on FC9
Comment 39 mrintegrity 2009-02-10 13:28:16 UTC
I am experiencing this bug in evolution 2.24.3 fedora 10
Comment 40 Dragan Tomas 2009-02-10 20:53:55 UTC
I am also experiencing this identical bug in the latest build of Evolution 2.24.3 for Ubuntu Intrepid (8.10). It is very strange to me that the original bug report about this very same issue I'm having now was reported to you guys back in 2006. It is also very strange to me that I did not experience this bug with the Ubuntu Hardy (8.04). What in the world is going on with this bug? It comes and goes as its pleased? That sucks!

Each time it happens I have to kill evolution then the evolution-alarm-notify, which pulls down with it the evolution-data-server-2.24 process. This happens at least 3 times a day!

Please help.
Comment 41 Akhil Laddha 2009-02-12 03:30:37 UTC
I have experienced in 2.25.90 though not in exchange back end. I guess it is common for all the back ends.

gdb traces of evolution 


Comment 42 Milan Crha 2009-02-12 11:24:31 UTC
*** Bug 256479 has been marked as a duplicate of this bug. ***
Comment 43 Milan Crha 2009-02-12 11:29:22 UTC
(In reply to comment #41)
> I have experienced in 2.25.90 though not in exchange back end. I guess it is
> common for all the back ends.

Right. Jim's comment #33 confused me, but his is the other issue. I guess, this one will be fixed together with bug #563212, where are patches already.
Comment 44 Milan Crha 2009-02-12 11:33:27 UTC
*** Bug 363174 has been marked as a duplicate of this bug. ***
Comment 45 Dragan Tomas 2009-02-25 17:56:41 UTC
I don't know if it's the Exchange 2003 OWA back end or what, but that's all I'm interested in using the Evolution with. I'm the only user of a Linux machine in the entire company that is primarily setup with M$ servers. This same bug happens in the latest Jaunty with Evolution 2.25.91, although not as frequently as on my primary machine that runs on Intrepid. That may be a poor criteria of a measure since it is not my primary system, hence I'm not using the Evolution in it as much. The OWA is running with a self-issued SSL certificate. This bug now is occurring more and more frequently, up to 10 times a day. I tried changing the fonts from Verdana to Arial, the TTF fonts that is, thinking that maybe my Verdana font is messing it up during video rendering operation of the fonts, but that didn't help. Maybe the entire font render engine is messed up? All my incoming and outgoing messages are HTML formated so I can't tell you at the moment if a plain text message would make any difference.
Comment 46 Milan Crha 2009-02-25 18:20:43 UTC
Tomas, is your stacktrace similar to the one from comment #41? One possible workaround is to disable image loading on HTML Mails, and on Header tab the photo lookup, in Edit->Preferences->Mail Preferences.
Comment 47 Dragan Tomas 2009-02-26 16:30:29 UTC
I can try that workaround, but I never made any gdb stacktrace of Evolution does far so I can't help you there until I do so. What's the best way? Start Evolution from the terminal window and then copy and paste what shows up in it or is there some other trick to get that? I can Google it up, I guess, I'm good at that. BTW, I don't know if turning pictures off will help because most of my emails don't have any pictures in them, yet the bug still happens.

Oh... and one other less important thing, my name is Dragan :-)
Comment 48 Dragan Tomas 2009-03-27 13:19:05 UTC
Milan,

Has any progress been made? I can still see this bug even in the latest Jaunty. Are there any recommended steps to discover what the program is really trying to do at the time of failure, like a debug console or some kind of a tracing tool? This bug is one of the major killers of this otherwise very well thought out application. I am at the point where I have to kill Evolution processes at least 10 times a day and it's only getting worse. Sometime I have to kill it because of a failure to refresh a folder, it simply stops talking to Exchange server. The version that's in Jaunty now is maybe just slightly better in that respect, but still suffers from this bug.

Dragan
Comment 49 Havard Bjastad 2009-03-27 13:41:28 UTC
Dragan, as the original reporter, I feel your pain. I waited almost 2 years for a solution, but eventually moved to Thunderbird. Thunderbird is rock solid for its mail capabilities, but unfortunately the calendar is not up to speed...

I guess one has to prioritize - a stable application with fewer features, or more features with constant instability :(
Comment 50 Milan Crha 2009-03-27 15:12:35 UTC
Hi Dragan, not in this bug, but in some other (I do not have the number handy at the moment). The (unspecified) patch is able to cancel rendering of a message when you select a different one, but as a side effect breaks connection with evolution-exchange, thus is basically unusable at the moment.

If your issue is about access to your address books, then you can try two things:
a) Edit->Preferences->Mail preferences->HTML Messages, do not use "load images if sender in my address book".
b) tab "Headers", uncheck "Show the photograph" or check "only in a local folder"

if your issue is different (the above didn't help), please install debug info packages for gtkhtml, evolution-data-server, evolution, evolution-exchange (if you use), evolution-mapi (if you use) and then, when evolution freezes, get the process ID of it (the PID below) and run this command:
   $ gdb --batch --ex "t a a bt" -pid=PID
and it depends what it'll show what to do next.
Comment 51 Dragan Tomas 2009-03-30 17:40:50 UTC
Created attachment 131717 [details]
The 3 debug traces of evolution by thread

This is a .zip file.
Comment 52 Dragan Tomas 2009-03-30 17:43:05 UTC
Milan,

Thank you for that piece of advise. I don't think it's related to display of images because the bug happens regardless of the e-mail content.

I followed your instructions and you will find the 3 resulting debug files in the attachment. Hope I did it right and that it will be helpful. I ran it 3 times, once for each process: evolution, evolution-alarm-notify and evolution-data-server-2.24.

Dragan
Comment 53 Dragan Tomas 2009-03-30 17:58:18 UTC
Milan,

Just to clarify one thing. My HTML message mail preferences indicate that I am choosing to show animations in messages and to always load images from the Internet. Under "Headers" tab the option to show the photograph of a sender in the message preview is checked. I unchecked it just now to see what happens. I will let you know if things get any better with that option being unchecked.

Dragan
Comment 54 Milan Crha 2009-03-30 18:07:01 UTC
OK, thanks, those looks good. eds and alarm-notify seems to not doing much. The culprit seems to be evolution:

Thread 2 (Thread 0xae3feb90 (LWP 23111))

  • #2 e_flag_wait
    at e-flag.c line 120
  • #3 do_get_contacts
    at e-book.c line 2093
  • #4 e_book_get_contacts
    at e-book.c line 2131
  • #5 em_utils_contact_photo
    at em-utils.c line 2266

Interestingly enough, eds should show some activity on an address book side, like getting contacts, but it isn't. So here should help b) from comment #50.
Please catch me on #evolution IRC channel at irc.gimp.org, for more debugging, when you'll have some time. :) It seems some address book crashed eds, but evolution didn't realize. (It starts new eds automatically, thus hard to notice.)
Comment 55 Dragan Tomas 2009-03-30 18:36:09 UTC
Milan,

That's true, I've noticed that Evolution at times is not able to display contacts when I'm writing a message. In that case I simply exit Evolution, open it again and then it works. I also often see errors in the status bar at the bottom of Evolution that it is unable to fetch new mail or synchronize the folder content with the server. I apply the same close/open method for that too and it recovers.

I will take you up on the offer to chat via IRC, but it's been a decade or more since the last time I've tried using any IRC so I'm pretty rusty. That was Window$ though, I decided not to go there because of hackers and lack of any privacy behind "Uncle Bill'$ Window" :-)

Dragan
Comment 56 Michael Meeks 2009-03-31 09:00:42 UTC
So - I have a duplicate of the formatting hang here; however - it is possible that this is a symptom of the failure to create a new thread in camel - due to address space starvation; and the fact that there is not good error handling for this (extremely unusual case).

I strongly suspect this is just a dup of the existing thread leakage problem, as known & fixed already - though the symptoms appear very different.

For me - when this hang happens, I do:

$ pmap `pidof /usr/bin/evolution`

and the Total is showing north of 3Gb - I just assume it is the thread issue ;-)
Comment 57 Dragan Tomas 2009-04-08 13:49:45 UTC
Michael,

My pmap for evolution PID shows only 166316K around the time it's started and it never increases by much. I do get a bunch of camel assertion errors among other errors at the start of evolution (in the terminal console). I don't know why I'm getting such frequent hangs, but it may be GAL related judging by some messages I get from the evolution-data-server right before the hang occurs. It's hard to pinpoint it. Milan Crha was very kind to offer some help with troubleshooting it using gdb, but the results we got are inconclusive. What complicates things is the fact that apparently Evolution suffers from more than one issue that is plaguing the code so the trace has the mixed number of things. I'm not a programmer, but I'm still here to help this bug squashed.

I came up with this script after getting tired killing the processes manually. It first shuts down and kills everything evolution related and then strts it again. I called it appropriately killstartevo.sh :-)

#! /bin/sh
evolution --force-shutdown
killall -e bonobo-activation-server
evolution

I'm wondering if the programmers working on this project are now all more focused on the new MAPI support and are not willing to spend time on fixing serious bugs in the current 'active sync' look-alike. I'm also disappointed that the Evolution's HTML renderer is doing terrible job of accurately displaying HTML messages, many features are missing from it. For example, when you get a confirmation e-mail from Dell about a certain purchase it looks a whole lot better when viewed in Firefox with all it's gradient colors of table headings and frames, not to mention the fonts. It is not any better in Jaunty, even though it uses the newer version of Evolution. The same problems I'm having with formating hangs are present in it too.
Comment 58 Milan Crha 2009-04-08 15:00:24 UTC
(In reply to comment #57)
> ...snip...
> that the Evolution's HTML renderer is doing terrible job of accurately
> displaying HTML messages, many features are missing from it.
> ...snip...

bug #540362, the first step is rendering, then composing.
Comment 59 Dragan Tomas 2009-04-08 15:13:48 UTC
Yes, the composing, that's what I meant. I stay corrected :-)

Thank you Milan!
Comment 60 Michael Gratton 2009-04-09 07:18:43 UTC
Created attachment 132390 [details]
Backtraces from 2.26.0-0ubuntu2 jammed while formatting

Backtraces from 2.26.0-0ubuntu2, which was actually built from a recent svn head.

I'm getting this a lot with 2.26.0, even after it has only been running for a short time.

It seems to happen a lot when switching between inboxes from different IMAP accounts in the folder list.
Comment 61 Milan Crha 2009-04-09 09:59:20 UTC
(In reply to comment #60)

It's trying to open your Personal address book, to check whether you've there a photo for the sender (see comment #50, point b) ):

  • #6 GNOME_Evolution_Addressbook_BookFactory_getBook
    at Evolution-DataServer-Addressbook-stubs.c line 330
  • #7 e_book_new
    at e-book.c line 3361
  • #8 em_utils_contact_photo
    at em-utils.c line 2239
  • #9 efh_format_message
    at em-format-html.c line 2008

How many contacts do you have in your personal address book? Though I guess evolution-data-server has hard time to open if because of something unknown at the moment.

Could you try to get backtrace of evolution-data-server process in such freeze, please? Also, might get some information, close evolution, do evolution --force-shutdown, then run evolution-data-server on one console (/usr/libexec/evolution-data-server-2.24 (can be 2.26, can be in lib instead of libexec), then run evolution and when it freezes, then look into evolution-data-server console, what it prints there. Might worth to open new bug for this, as it seems like your own issue with a local address book, whereas this bug might address some design issue.
Comment 62 Michael Gratton 2009-04-09 11:10:14 UTC
Created attachment 132401 [details]
back trace from eds, 2.26.0-0ubuntu2 (near svn head)
Comment 63 Michael Gratton 2009-04-09 11:12:00 UTC
Created attachment 132402 [details]
back trace from evo, 2.26.0-0ubuntu2 (near svn head)

This is the evolution bt that accompanies the last attachment.
Comment 64 Michael Gratton 2009-04-09 11:14:19 UTC
stdout from eds was just this over and over (final lines below):

> => 0x9ee0cc0
> impl_GNOME_Evolution_Addressbook_Book_open (0x9ee0cc0)
> impl_GNOME_Evolution_Addressbook_Book_getContactList
> impl_GNOME_Evolution_Addressbook_BookFactory_getBook
>  + file:///home/mjg/.evolution/addressbook/local/system
>  => 0x9ee0db0
> impl_GN

I don't have too many contacts in my personal address book, just over 200.
Comment 65 Milan Crha 2009-04-09 11:30:04 UTC
OK, I do not understand why it does so, but this is a deadlock in eds.
From your traces, thread 2 holds the lock which thread 1 is waiting on, but thread 2 wants to notify on thread 1, which is "busy".

Really, let's move to some new bug, your issue is unrelated to this bug, it's just way how to trigger this. I guess some similar thing might happen when you try to add new contact to this address book from email (by right clicking over the email address in the preview pane of a mail). Maybe not the first try, but might be on the second or so.

Please just give here a new bug reference and I'll move there too.
Comment 66 Akhil Laddha 2009-05-28 08:30:20 UTC
*** Bug 372456 has been marked as a duplicate of this bug. ***
Comment 67 Stanislav Brabec 2009-05-28 13:58:59 UTC
Hanging in the addressbook may be a the same as bug 523135, and also http://developer.pidgin.im/ticket/3820.
Comment 68 Enver ALTIN 2009-06-22 13:48:49 UTC
I have this problem too, I'm on 2.27.3 (Ubuntu Karmic).

I have added my Google/Gmail contacts to Evolution, could that be the problem?
Comment 69 Milan Crha 2009-06-22 15:40:12 UTC
(In reply to comment #68)
> I have added my Google/Gmail contacts to Evolution, could that be the problem?

I do not think so, check comment #50
Comment 70 Pieter Ennes 2009-08-11 12:03:45 UTC
It's indeed a general failure of evolution-data-server which keeps Evolution waiting I guess. If i just kill evolution-data-server and leave evolution running, it's fine again.

The fix should therefore be twofold maybe:

- change evolution so that it can detect the bad behaviour of EDS and restart/kill it (since the time scale of this bug shows that EDS will probably continue to behave like this for long).

- fix whatever is hanging/deadlocking EDS (which is quite a lot of things seen from my situation). Should i run with debug packages? I'm seeing this a lot on jaunty and karmic (2.27.5) now.
Comment 71 Milan Crha 2009-08-31 17:47:20 UTC
*** Bug 593692 has been marked as a duplicate of this bug. ***
Comment 72 Oded Arbel 2009-12-07 00:19:22 UTC
BTW - the launchpad bug report describes a related crash: If the formatting hang occurs, and you open a new window as described in comment #10, then if you close the original window and then try to cancel the stuck "formatting message" job then Evolution immediately crashes. 

A core dump is attached to the launchpad bug report.
Comment 73 Milan Crha 2010-03-26 15:14:36 UTC
*** Bug 352984 has been marked as a duplicate of this bug. ***
Comment 74 Milan Crha 2010-03-26 15:15:22 UTC
*** Bug 568743 has been marked as a duplicate of this bug. ***
Comment 75 Brian J. Murrell 2010-04-14 20:24:25 UTC
Can this bug get some love?  It's now 3.5 years old (the actual problem is way older than that though) and 71 comments long.  I didn't even bother counting the duplicates.

There was a commitment made on the evo list a while ago that there was a periodic (weekly) meeting where bugs were triaged and prioritized based on various criteria, including how prevalent a bug is.  Surely this bug has to qualify.

The issue seems quite reproducible.  I just increased my frequency of it greatly by adding another (remote) LDAP address book and re-enabling address completion from all of my address books.

I'd even be happy to use the fact that I can reproduce quite easily as a test bed for debugging or test patches.
Comment 76 Akhil Laddha 2010-06-10 05:51:20 UTC
*** Bug 610637 has been marked as a duplicate of this bug. ***
Comment 77 Penelope Fudd 2010-07-18 19:32:44 UTC
I've been getting this problem ever since I wrote a short script that emails me a log file once a week.  The text file is 34000 lines long (1.4mbytes), and is sent with this command:  "sendmail user@host.com < /tmp/logfile".  The preview pane shows "formatting message..." for a few minutes, during which time I can select other messages, but I can't view them.
Comment 78 Penelope Fudd 2010-07-18 19:37:00 UTC
This is Evolution 2.28.3, on Fedora 12, but I'm sure it would have happened on earlier versions.
Comment 79 claudia 2010-08-11 10:49:09 UTC
This is a reason to stop using evolution.
Comment 80 Gergely POLONKAI 2010-10-04 08:19:05 UTC
This bug still appears in 2.28.3 under Ubuntu 10.04.
Comment 81 Jean-François Fortin Tam 2011-02-11 18:22:11 UTC
This may be anecdotal, but using Evolution 2.30.3, I started getting such hangs when trying to open local (POP3) mails (and I did wait for everything to be downloaded, unlike bug #640777).

What seems to have worked for me is to disable HTML mails entirely. At the bottom of the Preferences' "HTML Messages" tab, I set the "HTML Mode" setting to "Show plain text if present".

I guess more testing/confirmation would be required to figure out if this is indeed a proper workaround (and not placebo effect).
Comment 82 Jean-François Fortin Tam 2011-02-17 13:49:08 UTC
Replying to my own comment: nevermind about the html/plain text thing... what seems to do the trick for messages that "can't" be opened even in plaintext mode is to activate the menu option "View > Show all mail headers" before opening the message. 
How strange.
Comment 83 André Klapper 2012-04-04 11:14:52 UTC
Does anybody still face this problem in 3.4 or 3.2?
Comment 84 Milan Crha 2012-04-04 13:25:38 UTC
More interesting would be 3.5.1, with webkit :)
Comment 85 Michel Dänzer 2012-04-04 13:33:57 UTC
(In reply to comment #83)
> Does anybody still face this problem in 3.4 or 3.2?

No, not seeing this anymore with 3.2 (and probably some versions before).
Comment 86 André Klapper 2012-04-04 13:35:45 UTC
(In reply to comment #84)
> More interesting would be 3.5.1, with webkit :)

Without warning that 3.5.1 is UNSTABLE that's not a good comment.
Comment 87 Wayne A Feick 2012-04-04 17:30:01 UTC
Honestly, it's been about 4 1/2 years since I added my comment to this bug (it was originally reported over a year before that). I had high hopes for Evolution to be a usable Exchange client, but I gave up on it years ago because bugs like this weren't getting fixed. 

I've gone back to using Thunderbird, and I run Outlook on Windows in a VMWare cage for calendaring.
Comment 88 Milan Crha 2012-05-09 09:25:10 UTC
I just got confirmation of reproducibility from 3.4.1 downstream:
https://bugzilla.redhat.com/show_bug.cgi?id=818864

And it's just about the reason of the cloned bug #578577 where I'm not aware of any other reason causing this stuck state, thus I'm marking the newer bug as a duplicate of this.
Comment 89 Milan Crha 2012-05-09 09:26:17 UTC
*** Bug 578577 has been marked as a duplicate of this bug. ***
Comment 90 Milan Crha 2012-05-18 14:54:56 UTC
*** Bug 516314 has been marked as a duplicate of this bug. ***
Comment 91 Dan Vrátil 2012-06-14 17:55:54 UTC
The new formatter merged for Evolution 3.5.3 is calling em_utils_in_addressbook() asynchronously, thus it does not block GUI anymore.

Closing as fixed.