Using ID constraints in XML Schema

Question

I've been testing out a schema containing a number of key/keyref constraints for validation with the DOM parser and I've been having problems.
Specifically, when I've used predicates as in:
blocklist/block[1],
blocklist/block[position()=1] or
blocklist/block[position()=last()]
the parser hasn't found any nodes.
Does anybody know anything about this - what type of XPath expressions are allowed in selector and field elements and exactly what predicate syntax is recognised (if any!)?

Answer 1

My mistake - it was a namespace problem!!
I'd remembered to use namespace prefixes for the xpath names in the keyrefs but not in the keys!
Because I was getting different messages from the parser when I was defining the keyrefs with and without the predicates, I got the impression that it was working ok without predicates and that it was a predicate problem.
Actually, I don't think that the error messages returned from the parser for broken key constraints are very infomative.
I have three keys and keyrefs defined. If there is a constraint error, the parser doesn't tell you which keyref the error is with, only that there is an error with one of them. Additionally, because the the key and keyref tables are calculated after the rest of the document has been schema verified, the line given as the error line is the current line - ie the last line of document!