staikos at kde.org
Wed Nov 2 00:04:21 CET 2005
On Tuesday 01 November 2005 17:53, Ted Kisner wrote:
> On Tuesday 01 November 2005 14:42, George Staikos wrote:
> > I'm most interested to see when this update is being called, because it
> > should never happen. Then we can determine the best solution.
> the backtrace is attached- the crash is actually happening in the
> KstImage::paint function, not update. Sorry for my confusion...
That's even more odd. Paint should not happen inside load() either. It's
coming from an X event it looks like (though I'm not positive), so I guess we
need to guard paint events while loading. Paints should not cause draws.
Hm. More likely, to provide a smooth user experience, we should just have
the paint() function bail on data related portions if the data inputs are
missing. I think that's what we do with 2dplot+vcurve. What do you think?
KDE Developer http://www.kde.org/
Staikos Computing Services Inc. http://www.staikos.net/
More information about the Kst