- From: Youenn Fablet <youenn.fablet@crf.canon.fr>
- Date: Thu, 01 Jun 2006 11:35:45 +0200
- To: Jonathan Marsh <jmarsh@microsoft.com>
- Cc: Arthur Ryman <ryman@ca.ibm.com>, www-ws-desc@w3.org, www-ws-desc-request@w3.org
The sparql example is quite interesting for WSDL as it is one of the few documents to use the HTTP binding. The use of schema constructs is on the other hand widely covered in other test cases. Therefore, even if your implementation fails on the schema parts of the sparql example, I see value to compare the http bits that your implementation extracts with the one Woden or CRF parser extract. That is also why, while I do not want generally to modify any real world wsdl test suite examples (and especially the sparql use of the http binding), changing a little bit a schema import is not a big deal to me (in the context of the wsdl test suite of course). Youenn Jonathan Marsh wrote: > > In practice I don’t see much difference between augmenting a testcase > with a schemaLocation attribute, or augmenting it with a catalog file. > Either way seems to be a bit intrusive. This testcase is riddled with > schemaLocation attributes – just missing the one I was looking for. > Let’s ask the DAWG why it was omitted in this one instance. > > FWIW, as this testcase involves multiple levels of schema import, I’m > unlikely to ever pass this test even after implementing support for > parsing the HTTP binding namespace. >
Received on Thursday, 1 June 2006 09:36:27 UTC