problem with events on overview map

oddly, this has to do with having a layer switcher and overview map in the map

OpenLayers.Event.stopObservingElement doesn't do so well if it's passed an element which doens't have an ID -- it will unobserve everything that doens't have an ID. Patch incoming.

whew, overview map narrowly escapes again

I like the idea of stopObservingElement returning false if elementId is null

OpenLayers.Event.observers[null] is not likely to belong to the correct element

This patch just bails out, in the same way as if there were no events to stop observing. I'll return the number of removed events instead.

alternative events patch - less voodoo

instead of using the element's 'id' field to index into the hashtable, assign each element a new, unique '_eventCacheID' property, which will be its key into the events cache hashtable. this way, it matters not if an element is observed has an id or not, nor if the user changes the id later. had to bubble out a new, private _stopObservingElementRow() function, but so be it. I even wrote a big test function with lots of tests to give us a gross sense that this is working. a real, thorough test would obviously be better, but this will do for now. tests pass in FF, IE6

ah barf. screwed up checking the replaceexisting attachment checkbox. I wanted to change one little thing in a comment. so here is the final patch.

okay, I don't understand what the heck is going on with this ticket now. can someone explain to me please: (a) which patches are supposed to be applied in which order, and (b) what does the "row" in _stopObservingElementRow() mean, and (c) does this actually fix the problem?

the note from my orig patch (which im deleting for clarity):

instead of using the element's 'id' field to index into the hashtable, assign each element a new, unique '_eventCacheID' property, which will be its key into the events cache hashtable. this way, it matters not if an element is observed has an id or not, nor if the user changes the id later. had to bubble out a new, private _stopObservingElementRow() function, but so be it. I even wrote a big test function with lots of tests to give us a gross sense that this is working. a real, thorough test would obviously be better, but this will do for now. tests pass in FF, IE6

final patch.... take two. updated to work with the latest version of trunk, plus renamed row() function to _removeElementObservers() on smart request from sde. tests pass FF & ie6

ok, this seems plausible now. nice work. go ahead and commit.

r2993