WMSCapabilities format may fail to correctly read BoundingBox of layers

[erilem]

The WMS specs say:

A BoundingBox inherited from the parent Layer for a particular SRS is replaced by any declaration for the same SRS in the child Layer.

The WMSCapabilities format doesn't comply with that.

Patch to come.

[erilem]

patch-3449-A0.diff fixes it, and adds new tests for both v1_1_1 and v1_3_0. Please review.

[fredj]

while you're tuning the for loop, can you please change:

var key, j;
for (j=0; j<attributes.length; j++) {

to:

var key, j, len;
for (j = 0, len = attributes.length; j < len; j++) {

to match our wiki:"CodingStandards"

[erilem]

(In [12209]) WMSCapabilities format may fail to correctly read BoundingBox of layers, r=fredj (closes #3449)

[pgiraud]

The patch breaks the tests in IE7 and IE8.

Something really weird, we (stupidly) tried to move the "BoundingBox" block above the "AuthorityURL"'s one and the tests pass.

[bartvde]

The issue is that the XML fragments in the tests are not valid according to the DTD, AuthorityURL comes after BoundingBox according to the DTD:

<!-- Nested list of zero or more map Layers offered by this server. -->
<!ELEMENT Layer ( Name?, Title, Abstract?, KeywordList?, SRS*,
                  LatLonBoundingBox?, BoundingBox*, Dimension*, Extent*,
                  Attribution?, AuthorityURL*, Identifier*, MetadataURL*, DataURL*,
                  FeatureListURL*, Style*, ScaleHint?, Layer* ) >

[bartvde]

(In [12343]) fix up tests in IE because of invalid XML fragments in WMS Capabilities 1.1.1 testcase (closes #3449)

[bartvde]

Chris, I'll let you decide if you want to pullup r12343 in the 2.11 branch before the release.

[erilem]

Excellent Bart! Thank you.

It's a test-only thing so we don't need another RC. And I think we should pull this in the 2.11 branch.

[crschmidt]

(In [12353]) Merge a fix for tests (Closes #3449)