another infoset vs. tree-model question

[Richard Lanyon <rgl@decisionsoft.com>]

As you've probably guessed, I'm trying to define a tree-like XML
document model using the definitions given by the infoset. Obviously
XPath does something similar, but one question I've just realised it
doesn't answer (at least, not in Appendix B) is:

do Unexpanded Entity Reference Information Items break up text nodes?

In other words, is:

foo&entityref;bar

one text node or two, if the entityref isn't expanded? The quote "A
text node comes from a sequence of one or more consecutive character
information items" suggests that any information item which interrupts
the "consecutive" sequence will effectively delimit text nodes.

-- 
Richard Lanyon, Software Engineer     DecisionSoft Ltd.
Telephone: +44-1865-203192            http://www.decisionsoft.com
       "The medium is the message" -- Marshall McLuhan