Hi all,
When generating metadata for a dataset that I published in GeoServer I found out that GeoServer uses different naming conventions for layer/featuretype/coverage identifier for the three different services (WMS/WFS/WCS) for the virtual endpoints.
To illustrate the behaviour I published a layer “raster_layer” to GeoServer in the workspace “namespace_test”.
The name element of the layer “raster_layer” in the capabilities document of the WMS service of workspace “namespace_test” has the value “raster_layer”.
(http://localhost:8080/geoserver/namespace_test/wms?request=getcapabilities&service=WMS&version=1.3.0)
The Identifier element of the coverage “raster_layer” in the capabilities document of the WCS service of workspace “namespace_test” has the value “namespace_test:raster_layer”.
(http://localhost:8080/geoserver/namespace_test/wcs?request=getcapabilities&service=WCS&version=1.1.1)
So GeoServer prefixes the name of the workspace to the identifier element of the coverage. The same difference can be seen between WMS and WFS when publishing vector layers (WFS capabilities document also prefixes the workspace name to the featuretype name elements).
When accessing the capabilities document of the global endpoint for the different services GeoServer will always prefix the workspace name for the identifier element (also for WMS).
(http://localhost:8080/geoserver/wms?request=getcapabilities&service=WMS&version=1.3.0)
I would expect GeoServer to omit the workspace name in the identifier when accessing the capabilities document of the virtual endpoint. Does anyone know whether this is expected behaviour or a bug in GeoServer?
Cheers,
Anton Bakker