On Sat, Dec 04, 2004 at 05:03:54PM -0600, akonstam@xxxxxxxxxxx wrote: > > People often suggest filing a bugzilla report. Well I have a perfect > record not a single bugzilla I have submitted has been resolved. For > example: > https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=134344 > > which is very annoying. Have other people had other experiences? I > assume yes. So what am I been doing wrong? In this case you may be asking (bugging) the wrong chunk of code. Ctl+Alt+Backspace will KILL the X-server not log the user out. This is a subtle but important difference. Logging out involves a process where various things are terminated graceful including updates to the wtmp and utmp files that w looks at. Since the desktop starts a pile of things that are 'disconnected' from controlling terminals the chain of child parent signals can get disconnected and the tidy up on exit stuff does not happen. At one level as I read the report it is "not a bug" (IMO). At another level it would be nice to do as you expect which is I suspect why the bug is still open. Ctrl+Alt+Backspace Immediately kills the server -- no questions asked. This can be disabled with the DontZap xf86config(5x) file option. Killing the server is not equal to Logout. I might suggest that you attempt to identify the specific programs that do not tidy up well (fix wtmp/ utmp files on exit) and what signal they are not catching correctly. i.e. It may not be X that is doing the wrong thing but some other specific application. You can trace a 'correct' Logout triggered exit and contrast with your Zap forced exit. In some cases (like kill -9) the signal cannot be caught and there is little that can be done. Processes like esd and arts might be problematic because they sort of 'float' (no controlling terminal) and do their input/ output with pipes. i.e. there is not 'good' way to know that these chaps need killing. But I do not believe they are wtmp/ utmp file writers. -- T o m M i t c h e l l spam unwanted email. SPAM, good eats, and a trademark of Hormel Foods.
