Hi, Doug. This is a very clear and useful test case demonstrating the issue. I think you are correct. Your sample file should validate against either schema. I am ccing the c-dev@xerces list because it appears to be a bug in Xerces itself.
We've been taking a look at it in the context of icXML, our accelerated version of Xerces incorporating Parabix (parallel bit stream) technology. Unfortunately, icXML-0.95 also has the bug. In essence, I think the problem is that the default processing of unions in Xerces/icXML is to process them as if they have "collapse" as the value of the whiteSpace facet. But this is not in accord with the XML Schema spec as I read it. We'll put it on the roadmap for a fix in icXML and we may be able to submit a patch for Xerces as well. On Mon, Jan 6, 2014 at 7:02 AM, Glidden, Douglass A < douglass.a.glid...@boeing.com> wrote: > Consider the following XML schema and document: > > <xsd:schema targetNamespace="http://example.com/schema"; > elementFormDefault="qualified" > xmlns:xsd="http://www.w3.org/2001/XMLSchema";> > <xsd:element name="CannedComment"> > <xsd:simpleType> > <xsd:restriction base="xsd:string"> > <xsd:enumeration value="This is a canned comment."/> > <xsd:enumeration value="This is a canned comment. Notice > the double space"/> > </xsd:restriction> > </xsd:simpleType> > </xsd:element> > </xsd:schema> > > <CannedComment xmlns="http://example.com/schema";>This is a canned > comment. Notice the double space</CannedComment> > > When Xerces-C++ is used to validate this document against the schema, it > passes validation, as expected; however, consider the following modified > version of the XML schema: > > <xsd:schema targetNamespace="http://example.com/schema"; > elementFormDefault="qualified" > xmlns:xsd="http://www.w3.org/2001/XMLSchema";> > <xsd:element name="CannedComment"> > <xsd:simpleType> > <xsd:union> > <xsd:simpleType> > <xsd:restriction base="xsd:string"> > <xsd:minLength value="1"/> > <xsd:maxLength value="48"/> > </xsd:restriction> > </xsd:simpleType> > <xsd:simpleType> > <xsd:restriction base="xsd:string"> > <xsd:enumeration value="This is a canned > comment."/> > <xsd:enumeration value="This is a canned comment. > Notice the double space"/> > </xsd:restriction> > </xsd:simpleType> > </xsd:union> > </xsd:simpleType> > </xsd:element> > </xsd:schema> > > As far as I can tell, the document above should still successfully > validate against this new schema (and incidentally, oXygen agrees with me); > however, Xerces-C++ fails the document with the following error message: > > value 'This is a canned comment. Notice the double space' does not match > any member types of the union > > Notice that, in the error message, the double space has been condensed to > a single space. Does this indicate an issue with my schema, an issue in > how I am using Xerces, or something else? > > Thanks for any help you can offer, > > Doug Glidden > Software Engineer > The Boeing Company > douglass.a.glid...@boeing.com<mailto:douglass.a.glid...@boeing.com> > >
