[Kde-games-devel] KGoldrunner problem in Windows

[Christian Ehrlicher]

Luciano Montanaro schrieb:
> Il Sunday 23 December 2007 16:10:30 Christian Ehrlicher ha scritto:
>> Ian Wadham schrieb:
>>>> Andreas Pakulat schrieb:
>>>>> Oh, last but not least: If you find problems that are specific to
>>>>> win32, please send a mail to the kde-windows mailinglist as AFAIK noone
>>>>> (except myself) of the kdewin-people read this list.
>>> OK, I am switching this discussion to kde-windows.  I am the author of
>>> KGoldrunner and would like to find out more about this problem and help
>>> address it and any other problems that come up, although I do not have
>>> a KDE Windows setup.  BTW KGr for KDE 4.0 is on bugzilla and my email
>>> address is also in the code.
>>>
>>> The problem with KGoldrunner was experienced by someone called Patrick
>>> on the KDE 4 Windows team and was reported only indirectly to
>>> kde-games-devel by Mauricio Piacentini.  Quoting Mauricio: "What happens
>>> is that the game runs, but runners are not visible.  The background
>>> renders."
>>>
>>> On Sat, 22 Dec 2007 09:28 pm, Christian Ehrlicher wrote:
>>>> And regarding your kgoldrunner problem - looks like someone forgot to
>>>> read the QPainter docu. I really wonder why it worked on linux ... :)
>>> So what does this mean, Christian?  And what are you implying? ... :-)
>> Just take a look on my last checkin to fix the problem. And see what
>> QPainter::begin() and :end() is for :)
>>
> 
> Well, it would be more helpful if you actually could explain the problem and 
> how you fixed it, and what are the issues to keep in mind with the Windows 
> port.
> 
> I don't know why this patch fixes the problem for you; however, the docs say 
> it is not necessary to call painter.end() since the painter destructor does 
> that automatically.
> 
But the QPainter is not destroyed at all! Only in the end, but not when
a image is 'ready'.
> So the end() in the background() method should not be needed. Is it really 
> needed to see the results?
> 
see above
> If it is so, this would mean that Qt on windows queues the commands and 
> performs them only on end(). But this is not clearly needed by reading the 
> documentation.
> 
Yes - it only works on linux by accident and it's not guaranteed that it
will work every time.


Christian