Branden Robinson wrote:
On Tue, Aug 31, 2004 at 10:37:48AM +0100, David Greaves wrote:No comment? I have tried to do some debugging and anlysis - am I on the right track? Thanks :)Usually, the Debian Bug Tracking System is used to file bugs. :) If you could provide us with a small program that exercises this bug, it would be helpful.
Hi Thanks for getting back to me. I did consider filing a bug but several things prevented me:* I didn't compile X - I just used the debug versions of the library and dl'ed the src to step through * The smallest test case I currently have is part of Qt - I later tried writing a small test that passed the same params to the same call and it worked fine (grr)
So I thought I'd try and discuss it rather than call it a bug in X when it could be in any of X, Qt or the application (a plugin to designer - the Qt IDE) :) (oh, and it's possible that I have some 'duff' fonts since I've installed a variety of weird 3rd party ones for use in the Gimp etc)
Having said this I did decide it was likely to be a bug in X given it segfaulted when the args passed were clean - it should never do that. I figured something internal to X had been corrupted (possibly by my fonts) BTW, I'm not alone in seeing this behaviour - google found me an obscure report which I think is the exact same thing that dated back to late 2003.
So - next step? File a bug or discuss it further? (The point of discussing it first is to help me close in on the problem - eg you may have a debug routine to dump X's fontcache or something - I wouldn't have a clue where to start with that kind of thing)
DavidPS FWIW I now work around the problem by never using italic fonts in the syntax highlighting in the editor - but I'm still up for putting in the time to help fix the bug.