Allow pan-dragging while outside map until mouseup

I am wondering why this is put off for a couple of releases.

Disclaimer: I have zero knowledge of the code base or other priorities.

I offer this up from a user-perspective: using openlayers.org, dragging out of the map was infuritating to use. It was snappy, and slick, and then I'd drag out and get frustrated.

Yeah, we should do this.

Testing trac list.

Testing trac list again.

patches to Handler.Drag welcome. until then, this doesn't belong in a milestone.

actually, would be willing to do this myself in 2.5.

SDE: If you get this done in the next couple weeks, feel free to move back to 2.5, but it's not done now and we're looking at RC1 in 2.5 eweks.

not currently being worked on, afaik.

Getting it work needs a new set of events that are bound to the document instead of to the map. So I introduced map.eventsDocument for events bound to window.document. Functions that should be bound to these events get are suffixed with "Document" (see patch).

Seems to work, but failes some tests

vmx: Thanks for your efforts on this during the sprint. Sorry I wasn't able to provide more help. Please disregard my suggestion to manage this entirely of the level of the control. Of course you will need to make the drag handler aware of the document level event. I assume that the reason for the failing tests is that you replaced the mouseup and mousemove methods of the drag handler with mouseupDocument and mousemoveDocument, instead of adding new methods for these.

I was not able to look into the code in more detail, but if you submit the new patch that you worked on during the sprint I'll try to find some time to do so.

Patch for keeping on dragging even outside of the map

ahocevar: I've finally understood what you mean. I've attached a new working patch (including an example) that passes all tests.

@vmx: Thanks for your new patch. Based on your ideas, I have created a new one which is less invasive and has no performance impact (except for an inexpensive boolean check in the mousemove callback). If this new patch works for you and you like it, it would be cool if you could create unit tests for it.

Also, for some reasons I do not understand, my patch does not work in IE. vmx, do you have an idea? It seems that evt.element never equals window when in the mousemove method of Handler.Drag.

less invasive, no performance impact

@vmx: now also works in IE. Looks like in IE, an Events instance cannot be registered with the window, only with the document.

@ahocevar: wow, great work. Wouldn't be the test included in keepdragging.patch be enough? Could I test anything more?

Replying to vmx:

@ahocevar: wow, great work. Wouldn't be the test included in keepdragging.patch be enough? Could I test anything more?

I think that test would be enough. Thanks!

Test for documentDrag feature

combination of the above two: patch with tests

Tests pass in FF3, IE8 and Chrome. Please review.

Hi Andreas, this is looking good. You can commit if you add a short documentation line to the destroyDocumentEvents and adjustXY methods. Thanks.

(In [9791]) Added documentDrag option to the DragPan control and Drag handler. When set to true, this allow pan-dragging while outside the map. Thanks vmx for the initial patches. r=vmx, bartvde (closes #39)

Reopening this in hopes of one more feature.

This is a very nice feature. To make it a bit more accessible, how about a documentDrag property on the navigation control?

Extending this a bit more, documentDrag (or something similarly named) could be a map property as well. I like the idea of bringing this a bit closer to the user in hopes of getting it used.

Good call. Please commit.

I'd like to change the test to reflect the change (it's a one-line). Either someone does it before committing, or I'll attach a patch the next day (or so).

documentDrag property on the navigation control

(In [9805]) Exposing the documentDrag property on the Navigation control. r=ahocevar (closes #39)

Replying to vmx:

I'd like to change the test to reflect the change (it's a one-line). Either someone does it before committing, or I'll attach a patch the next day (or so).

I think more tests here would be great. In r9805, I added tests relevant to the lib changes in that same revision. If you've got more tests related to r9791 in mind, I think they'd go in the DragPan or Drag handler tests (doesn't look to me like the functionality added is specifically tested).

Let me know if you were thinking of something else.

Replying to tschaub:

I think more tests here would be great. In r9805, I added tests relevant to the lib changes in that same revision. If you've got more tests related to r9791 in mind, I think they'd go in the DragPan or Drag handler tests (doesn't look to me like the functionality added is specifically tested).

I would have updated for the API change only (as you did, thanks). The current test isn't great, but I haven't had a better idea, rather than testing if dragging still works when the documentDrag option is enabled.