evolution-data-server crashed on startup

Crashes for me too, repeatedly. This is 1.8.0 on Edgy.

Memory status: size: 1699840 vsize: 0 resident: 1699840 share: 0 rss: 487424 rss_rlim: 0
CPU usage: start_time: 1157538647 rtime: 0 utime: 0 stime: 0 cutime:0 cstime: 0 timeout: 0 it_real_value: 0 frequency: 0

Backtrace was generated from '/usr/libexec/evolution-data-server-1.8'

(no debugging symbols found)
Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1".
0xffffe410 in __kernel_vsyscall ()
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7e8a71e in wait4 () from /lib/tls/i686/cmov/libc.so.6
#2 0xb7e8a6f7 in __wait3 () from /lib/tls/i686/cmov/libc.so.6
#3 0x0804ed62 in ?? ()
#4 0xbfe43790 in ?? ()
#5 0x00000000 in ?? ()

How consistently does this happen?

Could you install evolution-data-server-dbg, log out and back in and then follow the steps at https://wiki.ubuntu.com/Backtrace to get a debugging backtrace ?

It did also happen at each login for me.
I get ride of it by removing all references to external calendars, I think I had something weird with my config. I tried once to add a google-calendar source but entered it in the bad section and then I was unable to delete it.
After some gconf cleanup the old not-really-existing-calendar is gone and the crash with it.

Sorry, I do not have any gdb backtrace, I just remember that this backtrace was harder to get than others as evolution ran another instance and did not use the one I attached. The crash at login time is probably triggered by the clock-applet.
When running evo itself, the mail part was working fine and everything went wrong when switching to he calendar part.

For me it happens:

1) Every time I start evolution
2) Every time I click on the clock applet to bring up the calendar, plus
3) Every time I cancel BugBuddy when the clock applet caused the crash (thus, again, and again, and again...)

I *do* have the -dbg package installed. The above trace is all I get. Running /usr/lib/evolution/evolution-data-server-1.8 from the command line gives me the attached report.

BTW, why does it say "Backtrace was generated from '/usr/libexec/evolution-data-server-1.8'"

???

That file does not even exist on my system...

I can confirm what aurelien said: removing two web-based calendars using gconf-editor solves the issue for me. No more crashes.

[but what is that XML blob doing in gconf???]

Could anybody get a backtrace with gdb?
- evolution --force-shutdown
- gdb /usr/lib/evolution/evolution-data-server-1.8
(gdb) run
- get the crash
(gdb) thread apply all bt

Ok, I did:

1. Added secure, authenticated webdav calendar to the evo calendar.
2. evolution --force-shutdown
3. gdb /usr/lib/evolution/evolution-data-server-1.8
4. run
5. Start evo
6. Go online
7. Switch to calendar view -> crash
8:

(gdb) thread apply all bt

Thread 10 (Thread -1267741792 (LWP 12358)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb74e07c3 in *__GI___poll () from /lib/tls/i686/cmov/libc.so.6
#2 0xb75c8883 in g_main_context_iterate (context=0x8095db8, block=1, dispatch=1, self=0x8095e38) at gmain.c:2979
#3 0xb75c8bf9 in IA__g_main_loop_run (loop=0x80c2040) at gmain.c:2881
#4 0xb7ebde62 in startup_mainloop (arg=0x0) at e-book.c:3748
#5 0xb75e341f in g_thread_create_proxy (data=0x8095e38) at gthread.c:553
#6 0xb740f504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#7 0xb74ea4de in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 2 (Thread -1224016992 (LWP 12309)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb74e07c3 in *__GI___poll () from /lib/tls/i686/cmov/libc.so.6
#2 0xb75c8883 in g_main_context_iterate (context=0x8067308, block=1, dispatch=1, self=0x8064b18) at gmain.c:2979
#3 0xb75c8bf9 in IA__g_main_loop_run (loop=0x8067280) at gmain.c:2881
#4 0xb76cf7e0 in link_set_io_thread () from /usr/lib/libORBit-2.so.0
#5 0xb75e341f in g_thread_create_proxy (data=0x8064b18) at gthread.c:553
#6 0xb740f504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#7 0xb74ea4de in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 1 (Thread -1221294416 (LWP 12304)):
#0 0xb7489688 in strcmp () from /lib/tls/i686/cmov/libc.so.6
#1 0xb75e20c4 in IA__g_str_equal (v1=0x0, v2=0xb70bd1d0) at gstring.c:77
#2 0xb70bb525 in begin_retrieval_cb (cbhttp=0x807bb18) at e-cal-backend-http.c:449
#3 0xb75c3b11 in g_idle_dispatch (source=0x8073f28, callback=0xb70bd1d0 <_fini+492>, user_data=0x807bb18) at gmain.c:3926
#4 0xb75c5872 in IA__g_main_context_dispatch (context=0x805ec10) at gmain.c:2045
#5 0xb75c884f in g_main_context_iterate (context=0x805ec10, block=1, dispatch=1, self=0x804ef30) at gmain.c:2677
#6 0xb75c8bf9 in IA__g_main_loop_run (loop=0x8074db0) at gmain.c:2881
#7 0xb7a42a23 in bonobo_main () from /usr/lib/libbonobo-2.so.0
#8 0x0804b9d4 in main (argc=134687768, argv=0xbfabf344) at server.c:393
#9 0xb74328cc in __libc_start_main () from /lib/tls/i686/cmov/libc.so.6
#10 0x0804a441 in _start ()
(gdb)

What debug package am I missing?

$ dpkg -l *-dbg | grep "^ii"
ii evolution-data-server-dbg 1.8.0-0ubuntu1 evolution database backend server with debugging symbols
ii evolution-dbg 2.8.0-0ubuntu1 The groupware suite - with debugging symbols
ii libc6-dbg 2.4-1ubuntu10 GNU C Library: Libraries with debugging symbols
ii libglib2.0-0-dbg 2.12.3-1ubuntu1 The GLib libraries and debugging symbols
ii libgnomevfs2-0-dbg 2.16.0-0ubuntu1 GNOME virtual file-system (debugging libraries)
ii libgtk2.0-0-dbg 2.10.3-0ubuntu1 The GTK+ libraries and debugging symbols

Ah, this patch should fix it... this comes from the fix in this bug:

http://bugzilla.gnome.org/show_bug.cgi?id=267402

Will comment there.

I meant: the fix of the bug in http://bugzilla.gnome.org/show_bug.cgi?id=267402 introduced an error causing this crash.

Fix commited upstream...

One more apport for good measure...

and another one...crashes on desktop startup, this time after clicking on the "report crash log" or something with a similar name icon in the deskbar

Yay. Removing web calendars works.

Any chance we can get the aforementioned patch applied to the Ubuntu package directly while we wait for upstream to make a new release?

Or anyone care to attach a patch here that does so we that want web calendars can rebuild with it while we wait?

fixed to edgy