- From: Savas Parastatidis <Savas.Parastatidis@newcastle.ac.uk>
- Date: Wed, 9 Jul 2003 10:20:10 +0100
- To: "Jeni Tennison" <jeni@jenitennison.com>
- Cc: <xmlschema-dev@w3.org>
Jeni, I feel that we are very close in me being persuaded. But not there yet (I am sure you didn't want to hear this :-) > > > This is the modified version. I just changed the "schemaLocation" to > > point to a file locally. Can you use the file attached? It is the > > Schema-for-Schemas with a slightly modified xs:element complexType > > (the "type" attribute has been commented out). > [snip] > > Now, if your schema validator says that this is valid, I would argue > > that it has ignored the xs:element element definition found in the > > SfS.xsd but has not ignored the xs:element type declaration. Why? > > The schema validator that I used (Xerces-J) reported that the schema > was valid, as expected. SfS.xsd is used in constructing the infoset > for my test.xsd schema that's used to validate my instance document; > it is not used to validate test.xsd to check that it's a valid schema. > The contents of SfS.xsd have absolutely no bearing on whether the > validator thinks that test.xsd is valid or not. > OK, thanks. I understand this bit now. But let's test my understanding... The XML Schema validator checked test.xsd for XML Schema compliance. It found the attribute type="xs:element" but ignored the contents. Then, it started building the new infoset. In the new infoset, xs:element is valid because there is a type declaration. What would have happened if type="xs:string"? Does the XML Schema validator assume that all XML-Schema datatypes are known in the new infoset? My problem with this has always been the fact that there are some datatypes that are seen as built-in in the XML Schema namespace and then this xs:element appears. Thank you, .savas.
Received on Wednesday, 9 July 2003 05:22:13 UTC