full axis pragmas?

[David Carlisle]

> The problem is of course that because the feature is optional, users who
> want to write portable code are obliged to use workarounds which are likely
> to be inefficient on *all* implementations.

any takers for

(#exquery:following * #) {
(let $here := . return
   root()/descendant::*[. >> $here] except descendant::*)
}


(#exquery:following-sibling foo #) {
(let $here := . return
   ../foo[. >> $here])
}

etc.

or even

(#exquery:following-sibling foo #) {
error("ask your xquery implementor to do the decent thing")
}

seriously http://monet.nag.co.uk/xq2xml/index.html#s6 rewrites queries
to avoid optional axes, would it be worth wrapping the expanded
expressions as extension expression fallbacks so they could be
implemented as axes again on systems that want to play alomg by
supporting some as-yet-to-be-specified common xquery extension
namespace?



David


________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________