Using the same technique as Leaflet, this relays MSPointer events to touch
events. Using the simulator, I can verify that even pinch-zoom works with
this change, but the visual experience is a bit jumpy. Would be great if
people could test this on physical touch devices and see how it looks/feels
there.
d2a32d5 accidently changed the number of assertions in the
test_selectFeature function, so the change in behavior that no
beforefeaturemodified event is called any more in non-standalone mode got
unnoticed: Before unhacking ModifyFeature, the SelectFeature control
called the beforeSelectFeature method directly. Now, without a built-in
SelectFeature control, we need to call it from the selectFeature method in
both standalone and non-standalone mode.
Instead of adding and removing the Google container, we just toggle
style.display. This also removes the unused rendered and displayed
properties of the cache.
In OpenLayers 2.12, the backbuffer for each layer was placed right below
the layer itself. With animated zooming, that was changed so backbuffers
are all placed below the base layer. As it turned out, this is only good
for certain kinds of layers (e.g. street labels), so the default now is
to place the backbuffer right below its layer again. But the
transitionEffect option can now have a new value, 'map-resize', which will
cause the backbuffer to be placed below the base layer.
Since we don't rely on the DOM order for layers, we also cannot rely on it
for backbuffers. This change simply adds a sensible z-index to the
backbuffer container.
When an application zooms the map programmatically right after
initialization, there will be an empty backbuffer. This should be avoided,
because it adds an unnecessary DOM element.
The tile.onLoadEnd handler registers a transitionend listener on the image
of the last loaded tile. But if loading was aborted, there will be no
image. The new logic registers the transitionend event on the last child
of the layer div, which is not necessarily the tile we're handling loadend
for. When the backbuffer is empty (i.e. no tile was loaded at the time the
backbuffer was created), it will be removed immediately.
The ol3 version of the library is more advanced and complete. It can also
parse into flat array. It can handle unsigned integers, signed integers
and floats.
With two nested controls (DragFeature, SelectFeature), which - among
others - activated two OpenLayers.Handler.Feature instances, the
ModifyFeature control was fragile and hard to debug.
The issue that @iwillig, @bartvde and myself tried to track down was
that the two Feature handlers interfered with each other, making it hard
to select features on mobile devices, and causing points to jump during
dragging because of not updated last pixel positions.
With this refactoring, there are no more nested controls. All that's left
is a Drag handler. All tests pass. I had to remove one test that checked
for dragging of an unselected point while another was selected, because
now (and partially even before this change, thanks to the ugly drag
handler hack that is now removed) dragging a point will also select it,
saving the user an extra click.
Quick explanation:
Let targetCenterPx be described by PX and PY. Let oldRes and newRes be R0 and R1 respectively. Let centerPx (center after zoom) be described by CX and CY. And assume there is some anchored pixel point out there that represents the same map location before and after zoom. Let this be the origin OX and OY.
We want to recenter the map on the provided box. This means the map distance between the origin and box center at R0 is the same as the map distance between the origin and the map center at R1.
That is,
R0 * (OX - PX) = R1 * (OX - CX), and
R1 * (OY - PY) = R1 * (OY - CY)
Or, solving for OX and OY:
OX = (R0 * PX - R1 * CX) / (R0 - R1), and
OY = (R0 * PY - R1 * CY) / (R0 - R1)
