when allOverlays is true a hidden layer can cause a WFS request

[bartvde]

If a vector layer (with a WFS protocol) is the lowest layer in the layer stack, and allOverlays is set to true, a moveTo will be sent to the vector layer even if this layer is not visible. Which will cause a new WFS request to go out unnecessary.

The solution I could think of is when allOverlays is true only call baseLayer.moveTo if it is in range.

this.baseLayer.moveTo(bounds, zoomChanged, dragging);

What do others think?

[bartvde]

Two questions related to my solution:

-normally the baseLayer would always be visible, except for the case where allOverlays is true. So I don't think this extra test for visibility would harm the other cases, so I did not test specifically when allOverlays is true

-is there a need to also check the inRange stuff here as well? I guess it would also make sense to do that check but my patch does not incorporate it as yet.

I will need to make a new testcase for this as well.

[bartvde]

I've got some issues with running svn diff on a Windows system, so I'll post my testcase inline (Map.html):

        map.zoomToMaxExtent();
        t.eq(moveCount, 1, "map.moveTo moves the base layer only once");
        // new testcase
        moveCount = 0;
        a.setVisibility(false);
        map.zoomToMaxExtent();
        t.eq(moveCount, 0, "map.moveTo moves the base layer only if it is visible");
        a.setVisibility(true);
        // end of new testcase
        t.eq(map.getCenter().toString(), "lon=0,lat=0", "a map with all overlays can have a center");

[crschmidt]

I'm against doing another RC for this issue (though obviously I can be outvoted).

[elemoine]

Bart, with your patch, a "move" or "zoomChanged" event is triggered even if the base layer isn't visible. So a BBOX strategy will still get an event and fetch features.

[bartvde]

Hi Eric, that's not what I was seeing in my test app, since I had a vector layer with WFS protocol and BBOX strategy, and the WFS request did not go off anymore.

[crschmidt]

I'd like to do another RC as soon as possible. Can we come to some resolution on this -- either the patch works or doesn't work? Elemoine, did you test this, or just read the code?

[elemoine]

test-fails-with-ticket2072.patch demonstrates the issue I mentioned above.

ticket2072.2.2.patch is a new patch that looks more valid to me.

[elemoine]

Replying to bartvde:

Hi Eric, that's not what I was seeing in my test app, since I had a vector layer with WFS protocol and BBOX strategy, and the WFS request did not go off anymore.

I don't understand why (but have no time to test that right now)

[bartvde]

Eric is right, my patch does not work whereas his patch works fine. I'll attach an HTML example to reproduce. Just change the visiblity of the WFS layer to false using the LayerSwitcher and then zoom to max extent.

[bartvde]

example HTML to reproduce

[tschaub]

Latest patch looks good to me. Confirmed that the example works and tests pass. Thanks for fixing this up.

[elemoine]

(In [9370]) when allOverlays is true a hidden layer can cause a WFS request, testsed by bartvde, r=tschaub (closes #2072)