Control should set this.div to null in destroy

[bartvde]

currently Control does not set this.div to null in the destroy function. It should.

[bartvde]

The closes hook seems to be down, so r10789

[ahocevar]

r10789 makes the PanZoom tests fail - that control removes buttons from the div after the div has been destroyed by super.destroy.

[ahocevar]

In fact, r10789 makes the tests for many controls fail, including OpenLayers.Control.

[bartvde]

Oops, let me undo this change until I get time to investigate. Sorry for the noise.

[bartvde]

Undo done in r10790 need to investigate why this is causing problems.

[bartvde]

Okay, it seems the PanZoom destructor is the only one causing the problems, I'll upload a new patch for review :-)

Thanks Andreas for catching this!

[bartvde]

All control tests pass again in FF 3.6 and IE6, please review.

This should teach me to think an innocent one-line patch exists ... :-)

[jorix]

Replying to bartvde:

All control tests pass again in FF 3.6 and IE6, please review. This should teach me to think an innocent one-line patch exists ... :-)

Is best left three test in panel.html. Only need to change line 246 in panel.html removing "panel." leaving:

t.ok (div.innerHTML == ""

and test works.
I neglected to remove the "panel." in that line in my ticket #2835, sorry.

[bartvde]

thanks jorix, good suggestion, I've updated the patch.

[jorix]

Please see:  http://trac.osgeo.org/openlayers/ticket/2835#comment:6 and 7

There has been a clash of tickets!

[jorix]

@bartvde: In attachment:ol-2850-without-collision.patch I removed the effects of the  collision with Panel, and I made a change for compatibility in tests/PanZoomBar.html (due to r11228), no change in the effects of the test before the patch. Is it right for you?

@reviewer: The test of all controls work properly in FF and Safari.

[bartvde]

(In [11631]) Control should set this.div to null in destroy, thanks jorix for the updated patch, p=jorix,me r=me,fredj (closes #2850)