Looks like an issue in the libgnomekbd code.

Maciej, could you please get libgnomekbd sources, build and run test/gkbd-status-test. Would it crash the same way?

Actually, since that code already changed in git, could you please the version from git instead of 2.30 tarball?

gkbd-status-test seems to do nothing. It does not consume CPU etc. It does not finish. According to strace:
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=9, events=POLLIN|POLLPRI}, {fd=11, events=POLLIN|POLLPRI}, {fd=12, events=POLLIN|POLLPRI}, {fd=13, events=POLLIN|POLLPRI}], 6, -1) = 1 ([{fd=3, revents=POLLIN}])
read(3, "b\2R\1U\222\215\v\3\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0 \1\5\0\0", 4096) = 32
read(3, 0x222b7b4, 4096)                = -1 EAGAIN (Resource temporarily unavailable)
read(3, 0x222b7b4, 4096)                = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=9, events=POLLIN|POLLPRI}, {fd=11, events=POLLIN|POLLPRI}, {fd=12, events=POLLIN|POLLPRI}, {fd=13, events=POLLIN|POLLPRI}], 6, -1) = 1 ([{fd=3, revents=POLLIN}])
read(3, "b\2R\1\357\224\215\v\3\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\1\0 \1\4\0\0", 4096) = 32
read(3, 0x222b7b4, 4096)                = -1 EAGAIN (Resource temporarily unavailable)
read(3, 0x222b7b4, 4096)                = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=9, events=POLLIN|POLLPRI}, {fd=11, events=POLLIN|POLLPRI}, {fd=12, events=POLLIN|POLLPRI}, {fd=13, events=POLLIN|POLLPRI}], 6, -1) = 1 ([{fd=3, revents=POLLIN}])
read(3, "b\2R\1\211\226\215\v\3\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0 \1\5\0\0", 4096) = 32
read(3, 0x222b7b4, 4096)                = -1 EAGAIN (Resource temporarily unavailable)
read(3, 0x222b7b4, 4096)                = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=9, events=POLLIN|POLLPRI}, {fd=11, events=POLLIN|POLLPRI}, {fd=12, events=POLLIN|POLLPRI}, {fd=13, events=POLLIN|POLLPRI}], 6, -1

Stacktrace with libgnomekbd from git:
System: Linux 2.6.33-gentoo-ck1-tuxonice #6 SMP PREEMPT Fri Mar 19 02:50:32 GMT 2010 x86_64
X Vendor: The X.Org Foundation
X Vendor Release: 10799902
Selinux: No
Accessibility: Disabled
GTK+ Theme: Clearlooks
Icon Theme: Tango
GTK+ Modules: canberra-gtk-module, gnomebreakpad

Memory status: size: 354066432 vsize: 354066432 resident: 15962112 share: 11403264 rss: 15962112 rss_rlim: 18446744073709551615
CPU usage: start_time: 1270181834 rtime: 6 utime: 6 stime: 0 cutime:1 cstime: 0 timeout: 0 it_real_value: 0 frequency: 100

Backtrace was generated from '/usr/libexec/gnome-settings-daemon'

Traceback (most recent call last):
  File "/usr/share/gdb/auto-load/usr/lib64/libgobject-2.0.so.0.2400.0-gdb.py", line 9, in <module>
    from gobject import register
  File "/usr/share/glib-2.0/gdb/gobject.py", line 3, in <module>
    import gdb.backtrace
ImportError: No module named backtrace
[Thread debugging using libthread_db enabled]
[New Thread 0x7f234710e710 (LWP 26640)]
0x00007f2359054bcd in __libc_waitpid (pid=26648, 
    stat_loc=<value optimized out>, options=0)
    at ../sysdeps/unix/sysv/linux/waitpid.c:41
	in ../sysdeps/unix/sysv/linux/waitpid.c

+ Trace 221208

	Inferior 1 [process 26611] will be detached.

Quit anyway? (y or n) [answered Y; input not from terminal]


----------- .xsession-errors ---------------------
Evolution.ContactItemSource "Evolution Contacts" encountered an error in UpdateItems: System.TypeInitializationException: An exception was thrown by the type initializer for Evolution.SourceList ---> 
  at (wrapper managed-to-native) Evolution.CalBackendCache:e_cal_backend_cache_get_type ()
  at Evolution.CalBackendCache.get_GType () [0x00000] in <filename unknown>:0 
  at GtkSharp.EvolutionSharp.ObjectManager.Initialize () [0x00000] in <filename unknown>:0 
  at Evolution.SourceList..cctor () [0x00000] in <filename unknown>:0 
  --- End of inner exception stack trace ---
  at Evolution.ContactItemSource.UpdateItems () [0x00000] in <filename unknown>:0 
  at Do.Universe.Safe.SafeItemSource.UpdateItems () [0x00000] in <filename unknown>:0 .
[Debug 06:17:55.179] Acquiring org.freedesktop.DBus session instance
[Debug 06:17:55.180] org.bansheeproject.CollectionIndexer is already started
[Debug 06:17:55.181] Resolving org.bansheeproject.CollectionIndexer.Service (attempt 1)
warning: no loadable sections found in added symbol-file /usr/lib64/debug/usr/lib64/pango/1.6.0/modules/pango-basic-fc.so.debug
41	../sysdeps/unix/sysv/linux/waitpid.c: No such file or directory.
Cannot access memory at address 0xec8148388d49ffff
--------------------------------------------------

It seems to be some race condition. If I log into failsafe terminal and then I execute gnome-session all is ok.

I bisect through all revision after (and including) adding parameter for xklavier 5. All causes problems. It means that:
 - Problem is in gnome-settings-daemon
 - Problem is in xklavier
 - Problem was introduced before this commit
 - It is not sufficient to see in lsof -p `pgrep gnome-settings` that the git copy of library was loaded

> Problem was introduced before this commit
Sorry, what commit do you mean? Do mean some exact point in CVS history?

gnome-status-test is expected to put an indicator into your notification area on the panel. Do you see it?

(In reply to comment #7)
> > Problem was introduced before this commit
> Sorry, what commit do you mean? Do mean some exact point in CVS history?
> 

Git history. I mean your commit d7ff7775e4b1fcfd5d86a88abf49467b97b5ed4b ( http://git.gnome.org/browse/libgnomekbd/commit/?id=d7ff7775e4b1fcfd5d86a88abf49467b97b5ed4b ).

I run git bisect and problem occured in each version tested after this commit. I haven't tested the previous history.

Problen disappeared after downgrading gnome-settings-deamon and libgnomekbd to 2.28.x and libxlavier to 4.x.

> gnome-status-test is expected to put an indicator into your notification area
> on the panel. Do you see it?

No.

Well, the fact that the test does not work is suspicious... Did you build the latest libgnomekbd from git? Can you run the test with XKL_DEBUG=160 and attach the output?

Could you please try the latest version from git? There is chance I fixed it (if it is same as #617635)

Sorry, I meant bug #617643

Working. Yupee.

*** This bug has been marked as a duplicate of bug 617643 ***