Re: Context item dependant function in a step of apathexpression

[Michael Kay]

> As I've said, this makes perfect sense. But... I couldn't 
> find anything in the specs that could explain such a 
> behaviour. Roughly, I couldn't see anything that precisely 
> says that a path expression processing should/could be 
> stopped when a step returns an empty sequence. I conceed that 
> my interpretation of what we are dealing with might be wrong 
> though, and I'd welcome any clarification.

The XPath spec describes it fairly informally: 

Each operation E1/E2 is evaluated as follows: Expression E1 is evaluated,
and if the result is not a (possibly empty) sequence of nodes, a type error
is raised [err:XPTY0019]. Each node resulting from the evaluation of E1 then
serves in turn to provide an inner focus for an evaluation of E2...

So E2 is evaluated once for each node in the result of E1, which means that
if the result of E1 is empty, then E2 is evaluated zero times. 

The formal semantics has a more rigorous treatment.

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