From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5b) Gecko/20030723
Description of problem:
Sometimes, gtk2 applications (it mostly happens with those using antialiased
fonts) will get into a mode where they will accept mouse input but will not
accept keyboard input. It seems to happen sporadically when I'm typing in one
window of the application and a new window appears while I'm typing.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Log in in simplified chinese locale
2. run a program that can create multiple windows (e.g. gaim)
3. type into text fields while new windows are being created.
Actual Results: All windows will get into a state where they do not accept
keyboard input of any kind. Mouse input still works.
Expected Results: should continue to work normally.
Sadly, I don't think we'll be able to do anything unless you can
provide exact details about a set of steps to reproduce the problem,
at the level of "what keys to press", just running logged in in Chinese
for a couple of hours isn't really an option for me...
The key to reproducing the bug seems to be typing at the same time that the same
application maps another window.
I just reproduced it now (Firebird 0.6) by:
and then, before the window actually mapped, spamming the keyboard with
keystrokes in this text field. I'm entering the rest of this comment via cut
and paste, since I can't type in firebird anymore.
I couldn't reproduce it in mozilla or gedit. If it is about gtk+2 it should able to
reproduce on all of the gtk+2 applications?
I've seen the bug so far in gaim, mozilla and gedit, as well as with SWT
programs like eclipse, mainly because those applications are the ones that I
often type in. It may not be a GTK2 bug. Any ideas how to isolate the part of
Can you reproduce on any locales (including en_US.UTF-8) Or is there a specific
locale that you are having problem?
I am using RH9's simplified chinese locale, which is zh_CN.GB18030
Try it in other locales like en_US.UTF-8 and see if you got the same problem
I do not get the same problem in en_US.UTF-8
At first I thought I could reproduce it, but focusing on another window and then
back on the window that had the problem seemed to fix it completely.
Is this the same as bug 106807?
The problem that I've observed applies only to a single process. if, for
example, eclipse gets stuck, killing and restarting eclipse fixes the problem.
If gedit gets stuck, killing and restarting gedit fixes the problem.
If it is the case that bug 106807 is a bug in chinput, that suggests these two
bugs may be different.
Sorry I still cannot reproduce it. Can you test on Fedora Core and see
if that still a problem to you?
I've partially upgraded my desktop to Fedora, but haven't used it
extensively enough to know if the problem is still there.
I cannot reproduce in FC2 test2. Let me know if this problem still
occur for you in FC2.