[OpenLayers-Users] strange click events with internet explorer

[Christian Sengstock]

After debugging i found the problem to be in the method:

OpenLayers.Events#getMousePosition

Problem is, that i have my map div element in a "dislpay:none" div at
startup (so the map div offsets are 0). However inside of
#getMousePosition there's just once the check for the div offsets at
first method call. After that the offsets are saved as a member var.

So my problem was solved by checking the map div element offsets each
time i call #getMousePosition .. thus:

getMousePosition: function (evt) {
       // if (!this.element.offsets) {   // COMMENTED
            this.element.offsets = OpenLayers.Util.pagePosition(this.element);
            this.element.offsets[0] += (document.documentElement.scrollLeft
                         || document.body.scrollLeft);
            this.element.offsets[1] += (document.documentElement.scrollTop
                         || document.body.scrollTop);
       // } // COMMENTED
        return new OpenLayers.Pixel(
            (evt.clientX + (document.documentElement.scrollLeft
                         || document.body.scrollLeft)) -
this.element.offsets[0],
            (evt.clientY + (document.documentElement.scrollTop
                         || document.body.scrollTop)) - this.element.offsets[1]
        );
    },



I saw that most of the click handling stuff comes from the prototype
library. Wouldn't it be easier using the prototype methods directly (
and make use of all the benefits of a constantly enhanced library) ?

Chris



2007/5/31, percy <percyd at pdx.edu>:
> This sounds similar to a problem I had last October. I think it was
> fixed in version 2.2, I'll paste the email below...
> Percy
>
> -------- Original Message --------
> Subject: Re: [OpenLayers-Users] "interesting" cross browser problem
> with    mouse click position
> Date: Fri, 27 Oct 2006 11:09:34 -0700
> From: percy <percyd at pdx.edu>
> To: Christopher Schmidt <crschmidt at metacarta.com>
> CC: users at openlayers.org
> References: <20061011160003.A6010EDFAF at rainier.metacarta.com>
> <45401072.3020304 at pdx.edu>      <20061026143712.GB15808 at metacarta.com>
>
> that fixed it, thanks!
>
> any idea when 2.2 will become the "stable" release?
>
> I just got the SVN version out yesterday and was playing with that SWEET
> new MousePosition control. I need to wait to put that on our production
> server until it gets blessed, though.
>
> Thanks again for the groovy interface,
> Percy
>
> Christopher Schmidt wrote:
>  > On Wed, Oct 25, 2006 at 06:33:38PM -0700, percy wrote:
>  >
>  >> Hi All,
>  >>
>  >> We are attempting to use an openlayers frontend to a kamap tiled
> backend.
>  >>
>  >> My test page, http://ngmdb.usgs.gov/ol/examples/ngmdb4.html , works
>  >> great in both "major" browsers. However when we integrate it into the
>  >> main search page (which is not yet on a public site, but here's a
>  >> simpler Kamap version: http://ngmdb.usgs.gov/ngm_compsearch_map.html) we
>  >> get errant behaviour with regard to how the mouseclicks are handled by
>  >> OpenLayers in FireFox!
>  >>
>  >> to see this in action, just edit your test file by sticking in this div
>  >> just after the body tag:
>  >>
>  >> <div style="width:600px;height:800px;background-color:#008800;">
>  >> Push the map down the screen
>  >> </div>
>  >>
>  >> and then load the test file in a browser. firefox will exhibit the same
>  >> breaking behavior as it does on the working page!"
>  >>
>  >
>  > Mistakenly sent my response to Percy, rather than to the list: This was
>  > fixed with r1692, http://trac.openlayers.org/changeset/1692 closing bug
>  > #333, http://trac.openlayers.org/ticket/333 , and the patch available at
>  > the bottom of the page on r1692 should allow you to fix this issue for
>  > older OpenLayers instances. Alternatively, this fix is included in the
>  > 2.2-rc1 currently being tested, and will be included in the final 2.2
>  > release.
>  >
>  > Regards,
>  >
>
> Christopher Schmidt wrote:
> > On Thu, May 31, 2007 at 11:00:08AM +0200, Christian Sengstock wrote:
> >> Is this a known problem and is there maybe a workaround? I think the
> >> problem depends on where the map div is located in the DOM, because it
> >> worked properly when i tested the MouseToolbar with a very basic html
> >> page (just the map div).
> >
> > I'm not aware of this problem, or a workaround. (This just means I
> > haven't run into it, largely because I mostly use simple DOMs :))
> >
> > Anyone else run into this issue?
> >
> > Regards,
>
> --
> David Percy
> Geospatial Data Manager
> Geology Department
> Portland State University
> http://gisgeek.pdx.edu
> 503-725-3373
>