[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-bugs-dist
Subject:    Bug#33240: Kmail crashes randomly, normally when idle
From:       Michael =?iso-8859-1?q?H=E4ckel?= <Michael () Haeckel ! Net>
Date:       2001-10-05 8:43:36
[Download RAW message or body]

On Thursday 04 October 2001 00:13, Nick Lawes wrote:
>
> So, is the 2.2 supplied qt-2.3.1 okay, or would you recommend
> rebuilding that as well? Presumably it, and the kde libs should be
> okay, and hopefully contain debug symbols.

Well, the crash seems to appear eventually in kdelibs so recompiling that also 
with "./configure --enable-debug" can't hurt, but I can't see that exactely 
from the backtrace.

> I could then run it from gdb I guess to get a better chance to poke
> around, if that would be useful, or will the stack trace be enough?

Running from gdb doesn't give a better backtrace, it just gives a backtrace at 
all in some cases, where the crash manager doesn't work.

> > Does it also crash, if you disable interval mailchecking and check
> > your mails only manually?
>
> FYI I tend to leave kmail running 24/7, if this makes a difference. I
> do the same at work, and do see similar problems there. I'll rebuild
> kmail from source at work too. Crashes vary between, say, a couple in
> one day (rare), to one every few days.

At least a few other people already reported crashes, when the connection to 
the server fails with interval mail checking. But I can't reproduce that and 
up to now that was always with object prelinking what the SuSE RPMs don't 
seem to use.
Maybe you could try, if interval mail checking crashes, if you intentionally 
enter some wrong information.

> Funnily enough it just "locked up" on me without a stack trace, while I
> was typing in this message, I'd received another mail message, replied
> to it, and hit send from the composer. All windows stopped updating. I
> left it for 10 minutes, then killed it.

This was most likely do to a known clipboard related bug in Qt. "killall -6 
kmail" gives a crash manager in such a case.

Regards,
Michael Häckel

[prev in list] [next in list] [prev in thread] [next in thread]