Version: kde-3.0.5a for RH8.0 useage (using KDE KDE 3.0.5) Installed from: Unlisted Binary Package OS: Linux This kde install is the special rpm set for RH8.0, for version 3.0.5a. Kwin segfaults with SIGINT_11 or some such, repeatable at will by winning a game of kmahjongg or sishen sho, or closing the last of a multiple windowed gedit session. If it leaves me with an active shell, I can type "kwin &" and restore operation, else its ctrl-alt-bksp and restart X. Which if I have other things going is a bit of a pita sinec all my work is lost at that point. This has been a problem here since 3.0.2 and RH7.3 but is getting worse. The system is stable otherwise and has achieved 2 month uptimes many times. The video card is an nvidia gforce2 mx200, 32 meg video ram.
What happens if you switch your window decoration from BlueCurve to something else? Also, it'd probably be useful if you could provide a backtrace. Hint: you can restart kwin by switching to a plain terminal with Ctrl-Alt-F1, logging in, and running "DISPLAY=:0 kwin &" at the shell prompt.
Subject: Re: constant segfault when closeing apps or winning games On Sunday 12 January 2003 19:43, Maksim Orlovich wrote: >------- You are receiving this mail because: ------- >You reported the bug, or are watching the reporter. > >http://bugs.kde.org/show_bug.cgi?id=52936 > > > > >------- Additional Comments From mo002j@mail.rochester.edu > 2003-01-13 01:43 ------- What happens if you switch your window > decoration from BlueCurve to something else? This crash well predates the bluecurve 'phenomenon'. Which is why I removed the rh8.0 kde install and have since installed 2 updates from the special for rh8.0 drawer at kde.org, now at 3.0.5a. It was set to bluecurve as I hadn't changed that default. The decorations change as they should. I tried them all and left it at KDE-II as some of the others tend to cover up or render the titlebar stuff at very low contrast. >Also, it'd probably > be useful if you could provide a backtrace. I've attached a couple of them. > Hint: you can restart > kwin by switching to a plain terminal with Ctrl-Alt-F1, Thats where I do the startx from in the first place, so it would have to be F2 or greater. > logging > in, and running "DISPLAY=:0 kwin &" at the shell prompt. Nice to know, as that should rescue most of the crashes. I've been trying to file this bug for a long time, 4 or 5 months but mozilla seems to be badly mis-configured here, it kept asking my filesystem for the cgi stuff that only exists on your site. I even downloaded and installed (well, tried to, there is no install instructs in the full package) the whole bugzilla. Then I found that konqeror has come a long ways since the last time it crashed on me, and that worked. Fairly fast, I even got java to work with it last night. But I like the lizards presentation of other things, so if I could get some hints on fixing that cgi problem, I'd appreciate it. It treats them as file instead of http. Now, to see if I can figure out why kmails filters sorted this message to the JunqueMail drawer where I just now found it. I expect a look at all headers will show that reason. Created an attachment (id=736) KWin-crash-backtrace Created an attachment (id=737) kwin-backtrace-2
Well, the two problems could be unrelated. The backtraces you attached sure looks like it's a BlueCurve bug... It'd be very nice to know whether the problem goes away now that you use BII. And if not, please attach the backtraces for the "new" crashes as well. Can't say muich about Mozilla, sorry, don't use it.
Subject: Re: constant segfault when closeing apps or winning games On Monday 13 January 2003 10:59, Maksim Orlovich wrote: >------- You are receiving this mail because: ------- >You reported the bug, or are watching the reporter. > >http://bugs.kde.org/show_bug.cgi?id=52936 > > > > >------- Additional Comments From mo002j@mail.rochester.edu > 2003-01-13 16:59 ------- Well, the two problems could be > unrelated. The backtraces you attached sure looks like it's a > BlueCurve bug... It'd be very nice to know whether the problem > goes away now that you use BII. And if not, please attach the > backtraces for the "new" crashes as well. I will. Not BII, but "KDE2 default" is in effect now. I'll go play a game of shisen sho and see if it crashes when done, that, and kmahjongg are 100% crash generators, but the regular mahjongg doesn't. And I lost a hours work with a gedit quit one of 2 windows operation last night. >Can't say muich about Mozilla, sorry, don't use it. Whats your fav browser?
Subject: Re: constant segfault when closeing apps or winning games On Monday 13 January 2003 10:59, Maksim Orlovich wrote: >------- You are receiving this mail because: ------- >You reported the bug, or are watching the reporter. > >http://bugs.kde.org/show_bug.cgi?id=52936 > > > > >------- Additional Comments From mo002j@mail.rochester.edu > 2003-01-13 16:59 ------- Well, the two problems could be > unrelated. The backtraces you attached sure looks like it's a > BlueCurve bug... It'd be very nice to know whether the problem > goes away now that you use BII. And if not, please attach the > backtraces for the "new" crashes as well. > >Can't say muich about Mozilla, sorry, don't use it. It appears to be the 'bluecurve' bug ok, I've not been able to crash is since switching to 'KDE2 default'. But lemme play for a day and make sure before we cancel it.
I think we can close this bug, I haven't had a crash since switching away from the "bluecurve" window decorations. The question then is: Do we have to go thru all this again for the bluecurve bug, or is that one already cleared and just hasn't made a distribution medium yet?
I'll close this then. The BlueCurve bug doesn't belong here since it's in software written by and distributed by RedHat and not KDE. I've notified one of the RH packagers on IRC, but it might be a good idea to add a note about this in their bugzilla, since you seem to have a good way of reproducing it.
Subject: Re: constant segfault when closeing apps or winning games On Tuesday 14 January 2003 23:01, Maksim Orlovich wrote: >------- You are receiving this mail because: ------- >You reported the bug, or are watching the reporter. > >http://bugs.kde.org/show_bug.cgi?id=52936 >mo002j@mail.rochester.edu changed: > > What |Removed |Added >------------------------------------------------------------------ >---------- Status|UNCONFIRMED |RESOLVED > Resolution| |INVALID > > > >------- Additional Comments From mo002j@mail.rochester.edu > 2003-01-15 05:01 ------- I'll close this then. The BlueCurve bug > doesn't belong here since it's in software written by and > distributed by RedHat and not KDE. I've notified one of the RH > packagers on IRC, but it might be a good idea to add a note about > this in their bugzilla, since you seem to have a good way of > reproducing it. I tried that tonight after rxing this, but their bugzilla is just enough different that I was stuck in a loop of open new bug, get search screen, then list of stuff to file bug against, nothing for kwin or bluecurve there, click on new, or 'report new bug' repeat till disgusted. So I sent bugzilla@redhat.com a quasi-poison pen email, detailing the exact conditions to demo the crash, and the fix, which is apparently get bluecurve 'out of the loop'. I was surprised that konqeror couldn't deal with their bugzilla forms any better than that, and already knew that mozilla was a waste of time because it translates the http: cgi scripts to file: and of course they don't exist on my machine. Thats probably something in the mozilla mimetypes config, but I have NDI (no damned idea) how to go about fixing that. Not even a SWAG. Which is of course a Scientific Wild Assed Guess. I appreciate your looking into this so promptly, its been a pleasure to communicate with someone who actually knows what they're doing for a change. Doubly appreciated since I assume from your name that english possibly is not your first language. I don't speak any other languages, and at my age (68) I won't try, so when someone goes to the effort to learn my language so well, I have to tip my hat to those who do.
bug is now in Red Hat Bugzilla http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=81940
Subject: Re: constant segfault when closeing apps or winning games On Wednesday 15 January 2003 10:57, Harald Hoyer wrote: >------- You are receiving this mail because: ------- >You reported the bug, or are watching the reporter. > >http://bugs.kde.org/show_bug.cgi?id=52936 > > > > >------- Additional Comments From harald@redhat.com 2003-01-15 > 16:57 ------- bug is now in Red Hat Bugzilla >http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=81940 Thanks Harald. But the bigger question to me is why couldn't I file that myself? Unless I could assign it to a package name, it wouldn't let me past the package listing point, but kept looping me back around to that. This was using konqueror since the mime types or something in mozilla want to translate all those http: cgi script stuffs into file: cgi stuffs, and of course they don't exist on my machine. Its almost as if its made difficult just so we can't file bug reports, and I object to that, it should be easy, but its impossible. Things like this do nothing to help your mother to be as comfortable with linux as she is uncomfortable with winderz and its constant emailed viri and crashing.