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

[Raphael Kubo da Costa]

2009/4/6 Kevin Ottens <ervin at kde.org>:
> And again, looking at the patch it seems that the problem is simply resetting
> the m_job pointer to 0 when the job deletes itself. There's a proper tool for
> that and it is called QPointer. So why not simply using that?
I'm still not fully convinced. Using a QPointer does prevent
ReadOnlyPart from crashing indeed, but _k_slotJobFinished (which not
only sets m_job to 0 but also emits some signals and checks for
errors) still isn't being called when the job is cancelled. Is it
really the desired behaviour?