Hey, I've been seeing odd corner cases around CDATA sections too...
James?
>- getting the parser to understand high-order ascii characters
What do you mean by "high-order ascii". If you declare the
encoding as ISO-8859 (or whatever the magic string is) then
I think it works OK, if that's what you mean. If you're using
ISO-Latin or some MS codepage and you're *not* providing such
a declaration, expat is going to think you mean UTF-8 and will
properly boot on any characters > 127.
-Tim