[PATCH] Take care also of the KJob::finished signal in KPart - it also fixes Ark's BUG 187538

[Raphael Kubo da Costa]

2009/4/7 David Faure <faure at kde.org>:
> On Tuesday 07 April 2009, David Faure wrote:
>> I hadn't realized that the default for kill() had changed in kde4.
>
> Ah!
> It was changed, but in fact the other job trackers do call kill(EmitResult):
>
> ./kabstractwidgetjobtracker.cpp:80:    job->kill(KJob::EmitResult); // notify that the job has been killed
> ./kstatusbarjobtracker.cpp:143:    job->kill(KJob::EmitResult); // notify that the job has been killed
>
> which proves that the expected behavior hasn't really changed :)
> The bug is only in kuiserverjobtracker.cpp which calls kill() without an EmitResult
> because of a direct signal/slot connection. I'm very much in favour of rakuco's patch
> (with better handling of the race where the jobview isn't in the map anymore,
> we're talking about it on irc now).

I'm attaching the patch that's going to be committed. Q_ASSERT and
keys() were dropped in favour of if (job) and key(). And sorry for the
thread duplication ;)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: kdeui_fix_kill_emitresult.diff
Type: text/x-patch
Size: 1660 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20090407/86561382/attachment.bin>