Description of Problem: With redhat-config-kickstart-2.3.2-10, I can get the application to crash by changing the mouse type and previewing the configuration file. It doesn't happen every time, and I haven't really been able to pin down the steps which are required to reproduce it, but I've forced the crash a number of times. What is left on the screen after the crash is: Gtk-ERROR **: file gtktextbtree.c: line 4509 (gtk_text_line_destroy): assertion failed: (view != NULL) aborting... /usr/sbin/redhat-config-kickstart: line 4: 4346 Aborted /usr/bin/python /usr/share/redhat-config-kickstart/redhat-config-kickstart.py Probably isn't that high of a priority unless someone knows off the top of their head how to fix it. Version-Release number of selected component (if applicable): How Reproducible: Steps to Reproduce: 1. 2. 3. Actual Results: Expected Results: Additional Information:
I spent about 5 minutes beating this one over the head and I couldn't make it crash. Can you be more specific as to the steps required to break it?
Well, just made it happen again with redhat-config-kickstart-2.3.2-12. Basically what I do is change back and forth from Intellimouse (PS/2) to Generic (PS/2) to "Probe for mouse" switching the 'emulate 3 buttons' checkbox from time to time. After every mouse type change, preview the config file. This probably isn't going to happen too often. The only reason that I found it is I was testing out all of the different variations, so I was previewing the file after every change. Most people are going to get it right the first or second time.
I've tested just about every mouse choice, toggling 3 button emulation, and previewing permutation on two different machines and I can't reproduce this problem. My suspicion is that jturner eats too many Fruit Loops for breakfast and it's affecting his mind. I did find a bug with the mouse selection after all, though. If the user selected "No Mouse", but then checked 3 button emulation, the program would write out: "mouse --emulthree none" instead of "mouse none" I'm not sure exactly how anaconda would interpret that, but I'd rather not find out. The right thing to do is probably make the checkbox insensitive if the user selects "No mouse", but given the time I've decided to fix it in an easier way. The program just checks to see if no mouse was selected at Preview time and if it was, it will not write out the --emulthree line. For the next release, we should probably replace all the Combo widgets with GtkMenus like Havoc has suggested for other apps. As far as the original bug, I don't know what else to do except close as WORKSFORME. There seems to be nothing I can do to reproduce it.
/me takes my Fruit Loops and goes home :-( The fix for emulthree and no mouse is confirmed with redhat-config-kickstart-2.3.2-16.