https://bugs.kde.org/show_bug.cgi?id=368287
David Hallas <da...@davidhallas.dk> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |da...@davidhallas.dk
--- Comment #28 from David Hallas <da...@davidhallas.dk> ---
I have been looking through all these backtraces and they all point to the
KJob::finishJob function and usually the crash is triggered from a call to
isAutoDelete. Prior to the call to isAutoDelete two signals are emitted:
finished and (optionally) result. So I was thinking is it possible that the
receiver of one of these signals would delete the KJob instance? If so, that
would explain the crashes. And to fix it we could simply move the isAutoDelete
check to before emitting the signals?
I can't reproduce the crash myself, but I can cook up a patch that moves this
check if anyone is interested in testing?
--
You are receiving this mail because:
You are watching all bug changes.
