[Kde-games-devel] KDE/kdebase/workspace/ksysguard/gui/SensorDisplayLib

[Burkhard Lehner]

2007/12/11, Ian Wadham <ianw2 at optusnet.com.au>:
>
> On Tue, 11 Dec 2007 01:45 am, Aaron J. Seigo wrote:
> > hm.. what's unfortunate is that Qt::WA_PendingResizeEvent is set to
> false
> > *after* the resize event is sent (line ~5368 in qwidget.cpp) .. imho
> that's
> > incorrect, and makes it pretty much useless for testing whether or not
> to
> > respond to the event. =/
> >
> > ditto fro WA_PendingMoveEvent ...
> >
> Are you sure, Aaron?  It seems to me that it is the state of the *widget*
> that is being set and later tested, not the state of the resize event's
> object.  The event has been queued, in the situation you refer to, and
> should not happen until after the qwidget.cpp code gives up control.
>
> BTW, I am deeply concerned about this issue, which could be a
> showstopper for several KDE 4 games, including mine, but I'm on
> the run at the moment and cannot elaborate.  Please have a look at
> my kde-devel post of Thursday 6 Dec, "SVG Rendering and Main
> Window Startup".  You can get 1, 2 or 3 resize events, it seems,
> depending on config settings at startup time.


Just my 2 Cents:
I think this shows that doing a recalculation on a resize event is the wrong
way.
Maybe doing the recalculation on the first repaint event after a resize
event would be a better idea. One has to make sure of course, that between
the redundant resize events no repaint events are triggered.

Burkhard
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.kde.org/pipermail/kde-games-devel/attachments/20071211/10fd7f7b/attachment.html