Ticket #3242 (closed bug: wontfix)

Opened 2 years ago

Last modified 2 years ago

wfs-protocol-transactions.html: When draw the insertion point jump and follows the mouse away

Reported by: jorix Owned by:
Priority: critical Milestone: 2.11 Release
Component: general Version: SVN
Keywords: Cc:
State:

Description

Use http://openlayers.org/dev/examples/wfs-protocol-transactions.html

  • Select draw
  • Move the mouse from right to left and see the insertion point
  • ops! point jump left, and follows the mouse away.

Tried with FF, Chrome, Safari

Change History

  Changed 2 years ago by erilem

Yes, critical. I also note that we can't really move the map in this example. At this point I have no clue about this issue.

  Changed 2 years ago by erilem

The problem is related to max extent. Do this in FireBug to see the max extent:

box = new OpenLayers.Feature.Vector(map.getMaxExtent().toGeometry());
map.getLayersByClass('OpenLayers.Layer.Vector')[0].addFeatures(box);

  Changed 2 years ago by erilem

Add a MousePosition control to the Map in this example, and you'll see that the position is wrong when crossing the max extent boundaries. The problem lies in the pixel to lonlat conversion.

  Changed 2 years ago by tschaub

(In [11872]) Things break when map maxExtent does not match the Google maxExtent and a Google layer is used as the base layer (see #3242).

  Changed 2 years ago by tschaub

This broke with r10554 (it was broken in 2.10 as well).

follow-up: ↓ 7   Changed 2 years ago by tschaub

Ok, the relevant part of r10554 was changing wrapDateLine to true. If you add a MousePosition control to the example at r10553, you'll see that the display position is wrong there too.

in reply to: ↑ 6   Changed 2 years ago by tschaub

Replying to tschaub:

Ok, the relevant part of r10554 was changing wrapDateLine to true. If you add a MousePosition control to the example at r10553, you'll see that the display position is wrong there too.

Scratch that. There is more at play.

  Changed 2 years ago by tschaub

  • status changed from new to closed
  • resolution set to wontfix

Ok, wild goose chase.

The issue is that in r10554, the default for Google, Yahoo, and Virtual Earth layers was changed to wrapDateLine: true. Since in OpenLayers, the "date line" is considered the maxExtent of a layer (not the map), if you set the max extent of your base layer to something other than the world, wrapDateLine has no meaning.

We should have put something in the 2.10 release notes about this. We could put something in the 2.11 release notes if that sounds good.

A solution for the wfs-protocol-transactions.html is demonstrated in r11872. A (perhaps simpler) one is to set wrapDateLine: false in the Google layer constructor for that example. Both work.

Please reopen this ticket if you think there is something else we should do about it.

Note: See TracTickets for help on using tickets.