] Hi all,
] 
] I have a document that looks something like this:
] 
] <?xm-insertion_mark_start author="Nadia Swaby"
] time="20051202T111856-0500"?>
] <design.consideration></design.consideration>
] 
] I need to do a test to find out if the sibling immediately preceding the
] design.consideration element is an xm-insertion_mark_start PI.
] 
] Basically, I need something like
] preceding-sibling::*[position() = 1 and
] processing-instruction('xm-insertion_mark_start')]
] but that statement doesn't return anything.
] 
] Is there any way of doing this?
] 

Yes you just need to be more careful in your natural language
description of the problem, then the xpath follows naturally.
You show three nodes, a PI, a text node with value a newline character,
and an element node with name design.consideration.

preceding-sibling::* selects preceding _elements_ so if teh current node
is the design.consideration element, it won;t select anything (so the
predicate has no effect. If there was an earlier element selected by the
* , the predicate would select the element if it was the first element
(counting backwards) and if it had a _child_ processing instruction.

So I think what you actually want to ask is if the first node (in
reverse order) that is not white space text is a PI owith the name
xm-insertion_.

so
preceding-sibling::node[not(self::text()[normalize-space()=''])]
                       [1]
                       [self::processing-instruction('xm-insertion_mark_start')]

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
________________________________________________________________________