webkit scriptengine & multiple plasmoids

[Petri Damstén]

On Tuesday 22 December 2009 23:40:19 Aaron J. Seigo wrote:
> On December 20, 2009, Petri Damstén wrote:
> > On Friday 18 December 2009 00:17:19 Marco Martin wrote:
> > > On Thursday 17 December 2009, Petri Damstén wrote:
> > > > Hi,
> > > >
> > > > I'm trying to fix this bug: http://www.kde-
> > > > look.org/content/show.php/Scripted+Image?content=91749 (Error having
> > > > 2 of them). Seems that if there is two (or more) same webkit
> > > > plasmoids on desktop only one of them loads properly on plasma start.
> > > > When same plasmoid B load is started before A load is finished
> > > > javascript functions don't get called. This happens only when
> > > > plasmoids are same e.g. two scripted-image plasmoids. I made attached
> > > > patch to load plasmoids with the same plugin name in queue and it
> > > > seems to solve this but I'm not sure if this right way to do it?
> > > >
> > > > Petri
> > >
> > > seems an hacky workaround, but i can'r really think about a possible
> > > real cause of this...
> > > could be a bug in webkit itself, maybe has some internal sigleton class
> > >  that gets in an inconsistent state when two pages load the same url in
> > > the same moment? seems terribly unlikely but...
> >
> > I can reproduce this bug with a small Qt only app
> > (http://aryhma.oy.cx/damu/webkit-test.tar.gz). It seems that if the same
> > javascript file is included in multiple html files that are loaded at the
> >  same time then loadFinished signal is emitted too early. Should this
> > have a workaround in webkit script engine until this is fixed in
> > Qt/webkit?
> 
> if it has been reported against Qt/Webkit, then yes, we can have a
> (documented) work around in the code that we can remove when Qt/Webkit is
> fixed.

Noticed that if javascript is embedded in html it works, so this could have a 
work around also on those rare webkit applets that are run in multiple 
instances. 

Qt bug tracker seems to recommend webkit bug tracker for webkit bugs but this 
bug is related to qt/webkit signals (of course it still might be a webkit bug) 
so I'm not sure what would be the right place for the report.

Petri
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
Url : http://mail.kde.org/pipermail/plasma-devel/attachments/20091227/6f4c6e6c/attachment.sig