From: Jordi - I. - M. <jmi...@or...> - 2001-02-22 14:00:12
|
I agree. Pascal wrote: > the addChild() is the first point that creates an actual HTML element/object > (it adds contents to the html page) > > this would suggest that normal javascript objects are freed (also sounds > logicall, seeing the documents found by Mr.Design) using some form of > garbage collection. > > There might be a problem with javascript objects having a reference to real > html objects (layers) and that reference is still valid (so the GC won't > destroy the objects).. > > Pascal Bestebroer (pb...@oi...) > Software ontwikkelaar > Oberon Informatiesystemen b.v. > http://www.oibv.com > > > -----Oorspronkelijk bericht----- > > Van: dyn...@li... > > [mailto:dyn...@li...]Namens Lasse Lindgård > > Verzonden: donderdag 22 februari 2001 13:47 > > Aan: dyn...@li... > > Onderwerp: [Dynapi-Dev] Hunting the memoryhole with a shutgun > > > > > > After several attempts I still can't produce a simple example > > where IE5.5 > > leaks without having DynAPI involved. > > The create example leaks 200kb. > > If you remove the DynAPI.document.addChild(myLayer) the leak stops. > > I wonder what that could be ? > > /Lasse > > > > > > _______________________________________________ > > Dynapi-Dev mailing list > > Dyn...@li... > > http://lists.sourceforge.net/lists/listinfo/dynapi-dev > > > > _______________________________________________ > Dynapi-Dev mailing list > Dyn...@li... > http://lists.sourceforge.net/lists/listinfo/dynapi-dev |