- From: Florent Georges <fgeorges@fgeorges.org>
- Date: Thu, 17 Dec 2015 21:17:27 +0100
- To: Abel Braaksma <abel.braaksma@xs4all.nl>
- Cc: Christian GrĂ¼n <christian.gruen@gmail.com>, Public Joint XSLT XQuery XPath <public-xsl-query@w3.org>
I guess you can define => in 2 different ways: either by calling RHS once for each item in LHS, something a bit more "object oriented", syntactic sugar for "$s ! f(.)", or by calling RHS exactly once, syntactic sugar for "f($s)". Because we chose the latter, there is no reason why treating the empty sequence in a special way (that'd be dangerous I think). Regards, -- Florent Georges http://fgeorges.org/ http://h2oconsulting.be/ On 17 December 2015 at 19:14, Abel Braaksma wrote: >> >> I agree with Christian. I think a key difference with path expressions is that >> the RHS operand is not called once per item in the result of evaluating the >> LHS operand (which indeed, is zero times for the empty sequence). >> >> With the arrow operator, the RHS is called exactly once, whatever the result >> of evaluating the LHS. >> > > Yes, I think that makes sense, albeit a trifle unintuitive. So the LHS cannot work as a filter (but we have "/" and "!" for that, which *does* work that way, i.e., if LHS is empty, RHS needs no eval) > >
Received on Thursday, 17 December 2015 20:18:15 UTC