GNOME Bugzilla – Bug 563212
Crash in e-book.c:do_cancel: 2527
Last modified: 2009-04-27 17:48:15 UTC
What were you doing when the application crashed? Reading mails, selecting next mail, it crashing while displaying "formating message..." Distribution: Gentoo Base System release 2.0.0 Gnome Release: 2.24.2 2008-12-01 (Gentoo) BugBuddy Version: 2.24.2 System: Linux 2.6.26-gentoo-r1 #2 SMP PREEMPT Mon Oct 20 17:47:13 CEST 2008 x86_64 X Vendor: The X.Org Foundation X Vendor Release: 10502000 Selinux: No Accessibility: Disabled GTK+ Theme: Glossy Icon Theme: gnome Memory status: size: 732532736 vsize: 732532736 resident: 104222720 share: 20873216 rss: 104222720 rss_rlim: 18446744073709551615 CPU usage: start_time: 1228383516 rtime: 7710 utime: 7036 stime: 674 cutime:0 cstime: 5 timeout: 0 it_real_value: 0 frequency: 100 Backtrace was generated from '/usr/bin/evolution' Really redefine built-in command "frame"? (y or n) [answered Y; input not from terminal] [Thread debugging using libthread_db enabled] [New Thread 0x7ff1ecf52750 (LWP 4827)] [New Thread 0x42358950 (LWP 9098)] [New Thread 0x4439d950 (LWP 5030)] [New Thread 0x42b59950 (LWP 5029)] [New Thread 0x4335a950 (LWP 4843)] 0x00007ff1e994d53f in __libc_waitpid (pid=0x238d, stat_loc=0x7ffff4fa1b60, options=0x0) at ../sysdeps/unix/sysv/linux/waitpid.c:41 in ../sysdeps/unix/sysv/linux/waitpid.c
+ Trace 210469
Thread 2 (Thread 0x42358950 (LWP 9098))
----------- .xsession-errors --------------------- stop id=1 "Comwax"-"Travail Temps Complet" 886:24:52 00-41-41 hours=886 min=53184 secs=3191092 Avertissement du gestionnaire de fenêtres : Buggy client sent a _NET_ACTIVE_WINDOW message with a timestamp of 0 for 0x26002b1 (Restart Id) Avertissement du gestionnaire de fenêtres : meta_window_activate called by a pager with a 0 timestamp; the pager needs to be fixed. Avertissement du gestionnaire de fenêtres : Buggy client sent a _NET_ACTIVE_WINDOW message with a timestamp of 0 for 0x26002b1 (Restart Id) Avertissement du gestionnaire de fenêtres : meta_window_activate called by a pager with a 0 timestamp; the pager needs to be fixed. Avertissement du gestionnaire de fenêtres : Buggy client sent a _NET_ACTIVE_WINDOW message with a timestamp of 0 for 0x26002b1 (Restart Id) Avertissement du gestionnaire de fenêtres : meta_window_activate called by a pager with a 0 timestamp; the pager needs to be fixed. start id=1 "Comwax"-"Travail Temps Complet" 886:28:29 00-45-18 hours=886 min=53188 secs=3191309 (evolution:4827): evolution-mail-WARNING **: Can't get contacts: EBookStatus returned 20 (evolution:4827): evolution-mail-WARNING **: Can't get contacts: EBookStatus returned 20 camel-Message: -- camel-Message: -- 41 ../sysdeps/unix/sysv/linux/waitpid.c: No such file or directory. --------------------------------------------------
*** Bug 563683 has been marked as a duplicate of this bug. ***
*** Bug 564245 has been marked as a duplicate of this bug. ***
*** Bug 568418 has been marked as a duplicate of this bug. ***
(In reply to comment #3) > *** Bug 568418 has been marked as a duplicate of this bug. *** This difference in my above bug which is supposedly a duplicate of this one is that I don't get crashes. I simply get evolution not displaying messages in the preview window any more but rather it displays "Formatting Message..." constantly. Changing to different messages doesn't alter this. It simply adds yet another "Formatting message (...)" to the status bar at the bottom. Here is a stack trace from my most recent example. Note that this is not a crash. This is a stack trace from a running (but useless) evolution, stuck "Formatting Message...": 0xb800b430 in __kernel_vsyscall ()
+ Trace 211886
This happens many times a day. It would be nice to get at least some initial triage going on with it.
Further to comment #4, trying to quit Evolution when it's in this state never works. The UI goes all "grey" as if it's trying to quit but never does. The stack in this state is: 0xb800b430 in __kernel_vsyscall ()
+ Trace 211887
At this point I have to use evolution --force-shutdown.
*** Bug 569103 has been marked as a duplicate of this bug. ***
*** Bug 569621 has been marked as a duplicate of this bug. ***
For what it's worth, with a couple of patches from other bugs applied, this is my main cause of segfaults in evolution now.
As a workaround, you can disable Edit->Preferences->Mail Preferences->tab Headers, check Show the photograph of sender...; you can also try to change your preferences of Loading images section in HTML Messages tab to something else than Load images in messages from contacts. All these are looking to address books which are marked for an autocompletion, if I recall correctly.
So, is this bug completely understood then? Is the solution known? What is interesting is that when I hit this situation/bug, even an evolution --force-shutdown will not complete and I am left with a dead evolution and this stack trace:
+ Trace 212181
Surely that has got to help diagnose this problem if it's not yet completely understood, yes?
I'm trying to find some solution for this. I'll attach some test patch for you here, in case you are able to apply and compile it, it'll be great if you'll be able to test it. (Actually, I touched evolution and evolution-data-server, thus two patches will come here.) Nonetheless, it happened to me that I saw it crashing even with patches applied, but I cannot reproduce it at the moment. :(
(In reply to comment #11) > but I cannot reproduce it at the moment. :( False. It just happened again (I should change between messages up and down like mad to trigger that.). Let's see tomorrow, with some better fix, hopefully.
Created attachment 127923 [details] [review] proposed eds patch for evolution-data-server; Track state of the operation, three states WAITING, PROCESSING, CANCELLING and cancel or process only those WAITING. Free the operation after cancel and trigger always the operation flag here.
Created attachment 127924 [details] [review] proposed evo patch for evolution; Some evo bits necessary to change to let this work better.
Milan, the patches seem fine to me. But the volume of changes, locks etc scare me for trunk. If you can convince Matt or Ross, I would say ok :-)
eds part committed to trunk. Committed revision 10095. evo part committed to trunk. Committed revision 37329. I'm fine with trunk only, this is not so common as other related.