[most of this is already in the ticket - I recapture here] Hi Mats,
On Sun, 08. Jun 2014 at 09:07:33 +0200, Mats Elfström wrote: > Now tell me, is it a deliberate design choice not to make Qgis server OGC > compliant, and if so, why? QGIS server wasn't intentionally made incompliant. It was just extended. I don't know if any clients actually have problems with the service it offers, but those extensions actually make the validation of the GetCapabilities response fail. But the validation is a expensive process anyway and therefore most clients skip to validate and just use the response as is. That should work fine and so the invalidity of the response is probably a non-issue for most applications - and therefore went unnoticed (by wms clients and us) for quite a while. But extending the GetCapabilities response it also covered by the spec, so that itself isn't a problem either. It's just that the response needs to include another pointer to a document that describes the extensions and make validating the response document possible again. Only that part was missing in QGIS. Currently QGIS server offers GetLegendGraphic, GetStyles and GetPrint as extended requests. I'm not an XML expert (the GetCapabilities response is in XML), but including such an reference document seems to always introduce a new namespace and therefore extended requests always need to have a (namespace) prefix. So eg. GetPrint should be qgis:GetPrint instead. That's currently also not the case. So to fix this, we need to add prefixes, but that in turn might break clients that currently rely on the prefixless version. I don't know if there are any - but it makes me hesitant to change it. Anyway, the spec apparently allows to have the OGC schema at a different spot and doesn't clearly state that the copy needs to be identical - just probibits that "the normative content of the schema is changed". The current solution (which admittly has a smell) is that there is a document on qgis.org, that the qgis server response references instead of the OGC original and that document includes the original schema (by reference) and adds the extended requests the usual way, but in the process avoiding the prefix. If that complies with the quote above, we're fine now, although we might be bending the spec a little - but we're not breaking any existing clients and have validity now. We can come up with something better once we more about affected clients. I suppose the actual players in the qgis server field will join next week. Jürgen -- Jürgen E. Fischer norBIT GmbH Tel. +49-4931-918175-31 Dipl.-Inf. (FH) Rheinstraße 13 Fax. +49-4931-918175-50 Software Engineer D-26506 Norden http://www.norbit.de QGIS PSC member (RM) Germany IRC: jef on FreeNode -- norBIT Gesellschaft fuer Unternehmensberatung und Informationssysteme mbH Rheinstrasse 13, 26506 Norden GF: Jelto Buurman, HR: Amtsgericht Emden, HRB 5502 _______________________________________________ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer