D-Bus crasher from Python (on Ubuntu Edgy, not on Dapper???)

[Nigel Tao]

Sorry to spam the list with my homework, but we just got a bug report
in the GNOME Deskbar Applet that we'd like to fix before the imminent
stable release.

Some highlights from http://bugzilla.gnome.org/show_bug.cgi?id=353548
------------------------------------------
Distribution: Ubuntu 6.10 (edgy)
Gnome Release: 2.15.92 2006-08-22 (Ubuntu)
BugBuddy Version: 2.15.92

Memory status: size: 60628992 vsize: 0 resident: 60628992 share: 0 rss:
29646848 rss_rlim: 0
CPU usage: start_time: 1156947181 rtime: 0 utime: 127 stime: 0 cutime:109
cstime: 0 timeout: 18 it_real_value: 0 frequency: 0

Backtrace was generated from '/usr/libexec/deskbar-applet'

(no debugging symbols found)
Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1".
(no debugging symbols found)
[Thread debugging using libthread_db enabled]
[New Thread -1209747792 (LWP 5139)]
[New Thread -1254401120 (LWP 5162)]
[New Thread -1246008416 (LWP 5157)]
0xffffe410 in __kernel_vsyscall ()
#0  0xffffe410 in __kernel_vsyscall ()
#1  0xb7fbb30b in __waitpid_nocancel ()
   from /lib/tls/i686/cmov/libpthread.so.0
#2  0xb7135ed5 in gnome_gtk_module_info_get () from /usr/lib/libgnomeui-2.so.0
#3  <signal handler called>
#4  0xb6c3cc05 in dbus_connection_has_messages_to_send ()
   from /usr/lib/libdbus-1.so.3
#5  0xb6c4ef81 in dbus_pending_call_unref () from /usr/lib/libdbus-1.so.3
#6  0xb6c3e025 in dbus_connection_get_unix_fd () from /usr/lib/libdbus-1.so.3
#7  0xb6c3e07b in dbus_connection_get_unix_fd () from /usr/lib/libdbus-1.so.3
#8  0xb6c406d8 in dbus_connection_flush () from /usr/lib/libdbus-1.so.3
#9  0xb6c4084b in dbus_connection_flush () from /usr/lib/libdbus-1.so.3
#10 0xb6c4eddd in dbus_pending_call_block () from /usr/lib/libdbus-1.so.3
#11 0xb52bd1e6 in ?? ()
   from /usr/lib/python2.4/site-packages/dbus/dbus_bindings.so
#12 0x0895d5b0 in ?? ()
------------------------------------------
and
------------------------------------------
Initializing Web History
Initializing Files, Folders and Places
Initializing Computer Actions
Attempt to register the same DBusConnection with the message bus, but it is
already registered
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1210230592 (LWP 6431)]
0xb6bc6c05 in dbus_connection_has_messages_to_send ()
   from /usr/lib/libdbus-1.so.3
(gdb) bt full
#0  0xb6bc6c05 in dbus_connection_has_messages_to_send ()
   from /usr/lib/libdbus-1.so.3
No symbol table info available.
#1  0xb6bd8f81 in dbus_pending_call_unref () from /usr/lib/libdbus-1.so.3
No symbol table info available.
#2  0xb6bc8025 in dbus_connection_get_unix_fd () from /usr/lib/libdbus-1.so.3
No symbol table info available.
#3  0xb6bc807b in dbus_connection_get_unix_fd () from /usr/lib/libdbus-1.so.3
No symbol table info available.
#4  0xb6bca6d8 in dbus_connection_flush () from /usr/lib/libdbus-1.so.3
No symbol table info available.
#5  0xb6bca84b in dbus_connection_flush () from /usr/lib/libdbus-1.so.3
No symbol table info available.
#6  0xb6bd8ddd in dbus_pending_call_block () from /usr/lib/libdbus-1.so.3
No symbol table info available.
#7  0xb51631e6 in ?? ()
   from /usr/lib/python2.4/site-packages/dbus/dbus_bindings.so
No symbol table info available.
#8  0x08a51748 in ?? ()
No symbol table info available.
#9  0x00000000 in ?? ()
No symbol table info available.
(gdb)
------------------------------------------

The "Initializing FooBar" messages in the second snippet come from
Deskbar, and it means that we have got up to running this Python file
http://cvs.gnome.org/viewcvs/deskbar-applet/deskbar/handlers/gdmactions.py?rev=1.2&view=markup

Which does indeed start with
------------------------------------------
import dbus, dbus.glib
------------------------------------------
and continues with a bunch of dbus calls.

I am suprised that we can cause a crasher solely through the Python
D-Bus bindings.  Any suggestions?  From the bugzilla report, Raphael
Slinckx tried to reproduce, but failed, possibly because he was
running Deskbar HEAD on Ubuntu Dapper (and whatever version of D-Bus
that that entails) and the reporter was running Edgy, which may be a
clue.