as long -- presumably, the tests are less 'interactive', so the longer
wait doesn't hurt anyone, and this should help prevent failures due to
timeouts.l
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5486 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
permission to beat me senseless for not being able to test it there first.
Note that I'm pretty sure this still doens't work with FF3.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5479 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
it or not, and #1218 will cause the behavior to change across the board,
most likely, so we'll just be slightly more accepting, since the XML prolog
is almost definitely unlikely to not matter.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5478 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
the KML comes out namespaceless. Wrap the test up so that it doesn't totally
fail, and Paul is adding this to KnownIssues.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5475 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
FF2 won't work unless the map is loaded 'onload' -- but in Opera, adding
something to onload causes opera to believe that the page never finishes
loading within the test framework. FF3 works fine with this test if we
don't run it in a new window, but FF2 does not. For the time being, make
it work in our most well-supported platform -- FF2 -- and add a
FIXME to make it work in Opera.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5474 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
we still want this to b reported as a success, but log them to debug so that
people are aware the test is being skipped.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5468 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
If we do not do this, then every time we instantiate a new map using the same
div, the clientHeight as reported by the browser grows, and eventually we're
doing tests with a totally different sized window than we intended to.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5457 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
1. Meta-refresh, set to 20 minutes.
2. 20 second page download timeout, instead of the standard 6.
3. After all tests have finished, posts results back to
http://openlayers.org/test/results.cgi , which shows the results. Failures kick
off an email (currently only to me).
To set up auto-testing for your platform, set up a machine with a browser running full
time, and open the URL:
http://openlayers.org/dev/tests/auto-tests.html?run=all
tests will be run automatically on load, and every time the page refreshes, they will
run again, recording their results each time.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5444 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
and the Y transforms just make working with the SVG directly -- for example,
with people modifying it to support text -- more difficult than it should be.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5430 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
that Andreas has been working so hard on. I think this is the single most
awesome commit I've ever had the pleasure of committing. The results of this
commit are described on http://trac.openlayers.org/wiki/Styles: essentially,
this makes it possible to style features in all kinds of fun ways based on
rules, and will also form the underlying basis for #533. Things this patch
adds:
* OL.Rule classes. These classes allow you to do tests against the propertie
of a feature, and set a style based on these properties -- so you can
compare the 'speedlimit' property of a line, and test if it is > 60, and if
it is greater than 60, render it in a different color. You can also test
combinations of rules using the OL.Rule.Logical class, and test featureids
with the FeatureID class.
* OL.Style class: The OL.Style class lets you wrap up Rules into styles that
can be used with drawFeature to draw the feature in the selected style.
* OL.Layer.Vector.drawFeature will check if the given style is an OL.Style
object, and if so, it will draw the feature accordingly.
examples/georss-flickr.html shows usage of these classes.
Many, many thanks go to Andreas for all his hard work on this: this code really
is very pretty, and includes unit tests for all the classes (and we know that I
am a big fan of tests.)
Three cheers for Andreas: Hip hip, hooray! hip hip, hooray! hip hip, hooray!
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5429 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
the SphericalMercator mixin register two transformations to/from EPSG:900913,
EPSG:4326. Thanks to Tim for the feedback and review. (Closes#1210)
This allows us to transform points to/from SphericalMercator
without proj4js support -- and if other projects need similar functionality,
they can write their own custom transformation functions rather than
modifying proj4js to support some custom projection.
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5410 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
go ahead and commit this (relatively lighttweight) patch to the code so that
the projection library base API is there, even though for the most part, it's
not usable yet. This changes map.projection from being a string to being a
class, with a projCode on it. (Closes#1035)
git-svn-id: http://svn.openlayers.org/trunk/openlayers@5401 dc9f47b5-9b13-0410-9fdd-eb0c1a62fdaf
