[OpenLayers-Users] Subtle issue with Safari/Opera

Christopher Schmidt crschmidt at metacarta.com
Tue May 20 08:10:29 EDT 2008


On Tue, May 20, 2008 at 11:06:50AM +0200, Diego Guidi wrote:
> I'm working in a issue that compares only in safari and opera
> browsers, and I'm searching for any kind of suggestion...
> I have a monitor component that listens for a loadstart/loadend event
> and shows/hide an animated gif... nothing complicated: in my code
> (that I've modified form a source code published in this mailing list)
> I have a monitor functions like this.
> monitor: function(layer)
> {
> layer.events.register('loadstart', this, this.start);
> layer.events.register('loadcancel', this, this.end);
> layer.events.register('loadend', this, this.end);
> }
> and I use this component to monitor a WMS layer with a singleTile
> option activated.
> 
> All works well in IE and Firefox2, but in Safari 3.1 and Opera 9.2 the
> loadend event is thrown immediately, and not when the load  is
> effectively terminated! I'm pretty sure about this because I've
> verified (via console.log messages) that loadend is generated
> immediately after loadstart.


> This is maybe strange, but the really strange issue is that the same
> monitor component (exactly the same lines of code) works well (in all
> the browsers!) in another website build from my colleague: in this
> website the component is used to monitor the same WMS layer, uses
> google data as background like the mine, but something is different
> (as example, we used different custom layer switchers).
> The main point is that my website uses OL 2.6 (actually, I'm using
> http://openlayers.org/api/OpenLayers.js to grab OL from internet),
> when my colleague uses OL 2.5 from a local repository.
> So maybe something has changed in event model from 2.5 to 2.6?

Perhaps, but I'm not aware of anything changing there that would affect
this intentionally, so presumably it was an unintentional side effect of
the work done on tile transitions. I did a lot of work on maintaining
backwards copatibility, but it's certainly possible that I screwed it
up. 

Regards,
-- 
Christopher Schmidt
MetaCarta



More information about the Users mailing list