gtk2-2.16.0-1.fc11.x86_64 gedit-2.26.0-1.fc11.x86_64 ibus-1.1.0.20090311-3.fc11.x86_64 ibus-hangul-1.1.0.20090328-1.fc11.x86_64 1) Run gedit. 2) Activate ibus-hangul. 3) Type "wkk" 4) Use hotkeys to cycle through all engines, then back to Hangul. 5) Hit Backspace a few times. 6) gedit locks up.
Please update ibus to 1.1.0.2009331-1, and try again. Thanks
Still broken.
There seem to be numerous other ways in which ibus-hangul can cause the entire GTK+ app to lockup. It happened to me a few times by only Alt-Shift cycling through my active engines. Bumping up priority because this is pretty bad.
I can't reproduce exactly - for me when I cycle back to Hangul it just stops working - ie I can only continue to input ascii even though ibus says hangul is activated but the config button is gone from the toolbar. (Additionally I can get next-engine bindings with Alt to work though this is under virt if it matters.)
It seems the most common way to cause the gtk+ app to lockup is if you next-engine cycle from hangul. I can however get the gtk+ app to lockup sometimes when attempting hanja character conversion. Like: wk then press F9.
ibus-hangul, no next-engine switching, just hitting the Hanja button often locks up here. #0 0x000000358f2d63b8 in *__GI___poll (fds=0x7fff6b5aa9f0, nfds=1, timeout=<value optimized out>) at ../sysdeps/unix/sysv/linux/poll.c:83 #1 0x0000003595e23d40 in socket_do_iteration (transport=0x23b5d50, flags=6, timeout_milliseconds=-1) at dbus-transport-socket.c:1046 #2 0x0000003595e2213d in _dbus_transport_do_iteration (transport=0x23b5d50, flags=<value optimized out>, timeout_milliseconds=<value optimized out>) at dbus-transport.c:956 #3 0x0000003595e0e7ce in _dbus_connection_do_iteration_unlocked (connection=0x23850f0, flags=6, timeout_milliseconds=-1) at dbus-connection.c:1150 #4 0x0000003595e10900 in _dbus_connection_read_write_dispatch (connection=0x23850f0, timeout_milliseconds=-1, dispatch=<value optimized out>) at dbus-connection.c:3446 #5 0x00007fc65a4c6e3d in ibus_input_context_process_key_event (context=<value optimized out>, keyval=65508, state=0) at ibusinputcontext.c:576 #6 0x00007fc65a6eab4b in ibus_im_context_filter_keypress (context=0x21dcd00, event=0x22c6910) at ibusimcontext.c:361 #7 0x00000034ebb2955d in gtk_invoke_key_snoopers (event=<value optimized out>, grab_widget=<value optimized out>) at gtkmain.c:1908 #8 IA__gtk_main_do_event (event=<value optimized out>, grab_widget=<value optimized out>) at gtkmain.c:1593 #9 0x00000034eb64e7cc in gdk_event_dispatch (source=<value optimized out>, callback=<value optimized out>, user_data=<value optimized out>) at gdkevents-x11.c:2364 #10 0x00007fc662a2706e in g_main_dispatch (context=<value optimized out>) at gmain.c:1814 #11 IA__g_main_context_dispatch (context=<value optimized out>) at gmain.c:2367 #12 0x00007fc662a2a7c8 in g_main_context_iterate (context=0x20827b0, block=<value optimized out>, dispatch=<value optimized out>, self=<value optimized out>) at gmain.c:2448 #13 0x00007fc662a2ac65 in IA__g_main_loop_run (loop=0x237d000) at gmain.c:2656 #14 0x00000034ebb297e7 in IA__gtk_main () at gtkmain.c:1205 #15 0x000000000042830c in main () NOTE: If we cannot fix this before Fedora 11, we need to backout to the pre-Hanja ibus-hangul. It sucks, but the desktop is at least usable.
I think it is a bug in ibus. I fixed in ibus-1_1_0_20090407-1_fc11. Please test it.