RE: xsltproc/LibXSLT - non-compliance?

> > Well, in this case the argument *does* evaluate to a string 
> that is a 
> > QName. It consequently matches the quoted description.
> 
> No, it evaluates to a nodeset with one node whose string 
> value matches QName. That is not the same as evaluating to a string.
> 
> > Is this interpretation of your words correct?
> yes
> 
> > And the definitions of element-available() and function-available() 
> > don't say that much, so you can't rely on automatic 
> conversion to just 
> > happen.
> 
> Yes, although to be honest I would expect them to happen in 
> this case (as apparently would implementors of other xslt 
> systems.)

Section 3.2 of XPath 1.0 (function calling) says that the supplied argument
is converted to the required type, which in this case is string. The second
argument of key() is different because the declared type is "object".

Michael Kay
http://www.saxonica.com/