I've packaged up what I've done so far and you can find it at
http://www.cabbage.uklinux.net/pgxml.tar.gz
The TODO file included indicates what still remains to be done (a lot!).
In particular, it would be good to implement more of the XPath grammar.
However, once we get into the realm of more complex paths there becomes a
question about optimisation of XPath selection. If the documents are
pre-parsed, then XPath query elements can be rewritten as SQL queries and
you get the optimisation of the planner on your side.
I'd like to stick with the current solution if possible, because I think
it delivers a very simple interface to the user and is (code-wise) also
very straightforward. Maybe less efficient queries are a penalty worth paying?
Any thoughts?
Regards
John