Review Request: Fix crash in Weather plasma engine on resume from ram

[Aurélien Gâteau]

> On 2010-02-10 00:04:50, Shawn Starr wrote:
> > Please commit, the other ions will also suffer this. Thanks for catching it. Im puzzled why its calling destructor twice though.

I doubt the destructor is called twice. What might happen is reset() being called and then the destructor, or reset() being called twice. The later could happen because of multiple network status updates in less than 6 seconds (I think you should not use QTimer::singleShot() to trigger the reset, but keep a pointer to the timer to avoid multiple "singleshots")


- Aurélien


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.kde.org/r/2868/#review4115
-----------------------------------------------------------


On 2010-02-09 23:45:58, Aurélien Gâteau wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviewboard.kde.org/r/2868/
> -----------------------------------------------------------
> 
> (Updated 2010-02-09 23:45:58)
> 
> 
> Review request for Plasma.
> 
> 
> Summary
> -------
> 
> When resuming from ram, Weather engine crashes because it tries to free the forecast vector more than one time. The change here ensures one can run UKMETIon::reset() multiple times and one can call UKMETIon::reset() immediatly followed by UKMETIon::~UKMETIon().
> 
> 
> This addresses bug 219036.
>     https://bugs.kde.org/show_bug.cgi?id=219036
> 
> 
> Diffs
> -----
> 
>   trunk/KDE/kdebase/workspace/plasma/generic/dataengines/weather/ions/bbcukmet/ion_bbcukmet.h 1088007 
>   trunk/KDE/kdebase/workspace/plasma/generic/dataengines/weather/ions/bbcukmet/ion_bbcukmet.cpp 1088007 
> 
> Diff: http://reviewboard.kde.org/r/2868/diff
> 
> 
> Testing
> -------
> 
> Replaced ion_bbcukmet.so
> Restarted plasma-desktop
> Suspended computer to ram
> Resumed, no crash
> 
> 
> Thanks,
> 
> Aurélien
> 
>