On Tuesday 12 October 2010 5:49:31 pm Benson Margulies wrote: > One of our test schemas says: > > > <element name="inHeader"> > <complexType> > <sequence> > <element name="" > type="xsd:string"/> > </sequence> > </complexType> > </element> > > > Is a null name really a reasonable thing in a sequence? As I read the > spec, name="" is never reasonable. The spec demands an NCName for > name, and either name or ref must be present.
Yea. I don't think that would be valid. Now, the question is, is the test testing what happens with an invalid schema? I suppose that's too much to ask for. :-) -- Daniel Kulp dk...@apache.org http://dankulp.com/blog