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.
Using the new defaults, we don't need to pass options to the Navigation
control anymore. As a side effect, we get ZoomBox, but need to add css for
it to work properly. For users outside Vienna, an OSM fallback layer was
added which gets activated when the viewport does not intersect Vienna.
When the TileManager adds an image to its cache, it might be a backbuffer
image. Backbuffer images have an id with a '_bb' postfix. Regular images
do not have an id at all. In OpenLayers.Layer.Grid, in
addTileMonitoringHooks, a loadend listener is created for each tile. This
listener checks for tie image id, and will remove it from its parent node.
This will cause images to be removed from the layerDiv if the image comes
from the TileManager's tileCache and was on a backbuffer by the time it was
added to the cache.
Simply removing the image's id before assigning it to a tile resolves the
issue.
When a layer sets a cached image's className to something else than
.olTileImage (e.g. by setting .olTileReplacing), we should not keep the
image in the cache any more, because it may no longer be valid.
We need to handle the backbuffer before we fire the loadend event.
Otherwise listeners that call e.g. mergeNewParams() will cause the
backbuffer removal code to fail, because tile.imgDiv (and hence
this._transitionElement) will be null.
This change introduces a new 'replace' mode for tile transitions: when the
resolution does not change, which happens when mergeNewParams is called,
the tile will be marked with the .olTileReplace class. If this class sets
the tile's imgDiv display to 'none', the backbuffer for the tile will
immediately be removed when the tile is loaded.
This allows users to control whether 3d acceleration should be used or not:
Just like with plain web pages, having a stylesheet that sets a transform
on the map's layerContainerDiv will make OpenLayers use translate3d and
scale3d. When no such transform is set in the stylesheet, style.left and
style.top will be used, except for e.g. pinch zoom, where scaling is
needed.
In current Webkit browsers, having translate3d on svg child elements
causes the positioning from the layer not to be inherited by the vector
layer content.
The loadend event of an image is fired before the image is rendered. For
standard 256x256 tiles, this does not matter. But for singleTile layers on
large screens, rendering time needs to be considered. So we add a delay
that depends on the tile size. TODO: make the denominator configurable.
The navigation control gets better defaults, and the MouseWheel handler
gets a new maxDelta option, which can be used to avoid huge zoom level
jumps on heavy wheel/pad movements.
Having the TileManager remove an image from the DOM, then setting the
cached image, and then having to position it felt a bit awkward. With the
new beforeload event, the setImage method and putting renderTile before
positionTile, providing the cached image feels way more natural.
We now reuse tile images by maintaining a cache of image elements with a
simplified LRU expiry policy (by order, not by timestamp). The tile queue
is bypassed for images that are available in the cache, so they can be
rendered immediately. And the tile queue itself loads more than just one
image at a time now (2 per layer url).
Previously, minFrameRate could not be set as option with the start method.
The tests failed to catch this flaw. Now both the start method and the
tests are fixed.
In addition to relying on removeBackBufferDelay, we can remove the
backbuffer earlier without flicker in an ontransitionend listener on the
last loaded tile.
OpenLayers.INCHES_PER_UNIT.m should equal
OpenLayers.INCHES_PER_UNIT.Meters, just like OpenLayers.INCHES_PER_UNIT.km
should equal OpenLayers.Inches_PER_UNIT.Kilometers. This confusion probably
comes from mixing International inches with US Survey inches when compiling
the unit conversion list (1 meter is 39.37007874 International inches, but
39.37 US Survey inches. It may not be obvious, but 'inches'/'Inch' in
OpenLayers means US Survey inch, and 'IInch' means International inch).
This change also fixes offsets caused by incorrect resolution calculations in
OpenLayers.Format.WMTSCapabilities.
The new minFrameRate option is used to make sure that an animation does not
run longer than the time calculated from that frame rate. Time is made up
by skipping frames, i.e. skipping execution of the eachStep callback.
This avoids issues with Elements renderers, where the nodeFactory method
cannot find nodes while the map viewport is transitioning from the mapDiv
to the googleControl.
Because this method is called by OpenLayers.Function.bindAsEventListener,
which conditionally assigns window.event already, e cannot be null or
undefined here.
ahocevar