between 10,000 and 40,000. (It's hard to tell, since it crashes the browser
when you try.) The resulting behavior is that the browser crashes. This
wraps these values into Pixel regions, and doesn't draw features which fall
outside of them. (This sucks, but the result is the browser not crashing.)
This closes#669, and should be improved as described in #670 for 2.5.
If for some reason you know that the browser you're on can support larger
maxPixel values, you can adjust the maxPixel private variable on the
renderer. I believe that Opera and WebKit nightlies will work fine with
a larger maxPixel value. Changing this is encouraged *only* if you
are developer working on the code, because it can totally crash browsers
in a production environment.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@3071 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
tschaub. This includes tests (hooray) and fixes issues with zooming a wfs
layer while you have a feature selected.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@3067 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
CSS-based styling of the cursor on the map, rather than forcing 'default' after every
action. Tested in Safari, Opera, Firefox, IE7. Patch from Andreas Hocevar, #635. Thanks!
git-svn-id: http://svn.openlayers.org/trunk/openlayers@3042 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
minor modification: because features by default do nothing at all, change
the default cursor property (keep it at null) so that users aren't confused
when clicking has no affect. Thanks for the patch.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@3040 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
var from OpenLayers.Tile's constructor." This has no affects other than to
let people know now that this behavior will change in future versions of
OpenLayers. #647.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@3034 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
subcontrols. This will allow users to modify properties of the div -- for
example, adding a title element to the div. Access this property via
panel_div on the added control after addControls is called on the panel.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2973 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
always having a calculateOffset (which overwrites offset if it exists),
always have an offset. If people want calculateOffset, they can add it.
Reported by David Bitner.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2962 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
also reported on the users mailing list. The problem appears to be that Firefox
has poor support for circles of very small radius -- below about .0002. Since
units were in geographic units, this just didn't work so well. So:
* Change coordinate space to be pixel based.
* Make all x/y operations divided by resolution
* add getComponentsString, getShortString helpers
* Redraw nodes totally on every 'reprojectNode' call
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2960 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
* Add a 'style' property to the layer.
* When a feature is added to a layer, either uses the style on the
layer, or uses the default feature style.
* Add test to Layer.Vector for this funcionality
* Adds a preFeatureInsert hook to change a feature *before*
drawing it.
* Change openmnnd demo to include the use of preFeatureInsert
and style attribute on layer.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2949 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
the innerHTML still prevents the memory leak, but unregistering in this way
doesn't do the right thing. remove it.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2926 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
memory associated with those images. This completes #581, and reduces the
memory used by applications where adding and removing of many markers
is used.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2925 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
but it's not working for me at the moment, so I'm going to trust that the
rest of our testing process will find this at the moment, since the panzoom
is obvious enough that people will see it if it breaks :)
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2922 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
* code was only ever removing the first event listener from each element
on unload cache (for loop instead of while)
* code was frequently not removing observers because their associated
elements have already been removed
These two are fixed by this patch. Combined with the previous commits to
setting events, this should fix the memory leaks demonstrated by the 'drip'
tool for IE.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2913 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
events. This removes a number of minor memory leaks with creating/destroying
popups, layers, features, and markers.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2894 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
Woodbridge, tests written by Paul Spencer and some reworking done. Thanks
to both Paul and Steve for this code. (Closes #526.)
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2891 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
mousewheel functionality we had in earlier releases in a similar way, rewritten
for handlers. (The benefit of handlers in this case is that handlers are
write-once, use many, so the next mousewheel thing we get will *also* be able
to use this code.)
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2877 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
tests against the current output -- you may not break backwards compaatibility,
but that doesn't mean your functionality works! Fix broken mousetoolbar test.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2783 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
should be used in all library code. This patch replaces all uses of "$()" in
existing code. The primary reason for this is to avoid conflicts with libraries
(like Prototype) which might redefine this function differently.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2541 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
because the new map size was not set until *after* the onMapResize event.
(oops.) This way, untiled layers redraw with the correct size.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2532 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
was approved before 2.3 released, but never applied. Adds a removeMarker
function for the boxes layer, subclassing the marker default.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2302 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
in #506: tests are included to test that the links are properly created
for Atom 1.0 content (with a data sample provided in the patch).
Thanks, Sean!
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2256 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
this change does not affect the input or output of the function it is changing,
no test is neccesary. Patch from tschaub in #481.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2240 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
other than WMS. It should do this for any Grid layer. Ticket #496 includes
a patch for this: it adds a mergeNewParams call at the Layer.Grid level,
(subclassing Layer.HTTPRequest) which reinits the tile grid. This commit
also includes tests for this functionality. Since testing directly
on the Grid layer doesn't work (it's meant to be subclassed), we need to
instead pick a second layer to test -- in this case, ka-Map. Thanks to
Bill Woodall for the patch.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2237 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
is important is that the WMS spec specifies that you should seperate layer
names with a ",", and we were encoding that, so compliant WMSes would attempt
to find a single layer with ","s in the name, instead of grabbing multiple
layers. The new way to specify multiple layers is to set:
layers: ['global_modis','landsat']
or the like.
Includes tests, signoff from SDE.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2232 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
in Tile/Image.js. This is, for the moment, a manual test, though once we solve
the problem it will hopefully be possible to turn this into an automated
test.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2206 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
IE6, FF, Moz, Opera, IE7, Safari.
This may not work in all situations. It's included in util for utility, but
it is possible it may fail in some cases in some environments. If you find
this to be the case, please open a bug and we will work to resolve it.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2162 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
use of relative URLs, urls with port 80 in them, etc. patch originally from
tschaub, revised by euzuro, reviewed by me.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2093 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
support for buffer:0 on grids, and includes changes to Remove extra
rows and columns from the grid in _initTiles when reusing an existing
grid, which is a fix to #357 and #436.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2091 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
mergeNewParams". Since the Untiled layer won't redraw if the map is not
outside the current bounds, we force it with a 'zoomChanged' flag, which
indicates that the layer should redraw regardless.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2074 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
tile' problems where right clicking returns the right tile not have problems
anymore. Note that this is seperate from TileCache related issues where
OpenLayers doesn't request tiles on a grid.
Note that I have no reason why this should work. In fact, I'd say it
*shouldn't* work, but other people are seeing this problem, so I'm glad
to have the code in trunk, since it doesn't change teh default behavior.
Thanks to Tim for the patch.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2048 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
cache from shooting themselves in the foot: This doesn't seem to happen often,
but it has happened in internal code. I'm going to add more debugging info to
#440 as we get it, but want to get this in to prevent people from shooting
themselves in the foot in IE accidentally.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@2024 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
crschmidt