FrameSvg corruption [a bit urgent]

[Marco Martin]

On Friday 16 January 2009, Marco Martin wrote:
> On Friday 16 January 2009, Marco Martin wrote:
> > On Friday 16 January 2009, Ivan Čukić wrote:
> > > > nope
> > > > (yeah, deleted and restarted all had to be :/)
> > >
> > > Bummer... I have absolutely no more ideas what is the problem here...
> > >
> > > Is it possible that the cache has problems when two different processes
> > > access the same cache or something like that? I'm asking because I've
> > > just cleared it, and started only Plasma with one Lancelot Part on it -
> > > thus making it the same process - it was drawn correctly, and then I
> > > started the menu... and it remained screwed.
> >
> > yes, killing the lancelot process and reverting that change makes the
> > problem go away
>
> seems that printing size() as debug in updateSizes()
> both before and after resize() seems that sometimes resize() resizes to
> QSize(-1,-1)
oh, if a renderer is still not created resize() actually does resize to 
QSize(), so do we need to always have one i this case?
i wonder how the performance impact would be, i.e. how many resize() there are 
around that could actually not need to know the size...


-------------- next part --------------
A non-text attachment was scrubbed...
Name: resizeWorking.diff
Type: text/x-patch
Size: 498 bytes
Desc: not available
Url : http://mail.kde.org/pipermail/plasma-devel/attachments/20090116/d7504a78/attachment.diff