It's a real pain that doesn't have a common solution. I
tend to
(a) avoid using entities. Because I only ever use XML via
XSLT, processing-instructions are much more manageable.
(b) if I do use entities, don't rely on them remaining
intact - i.e. there should be no difference in information content between an
entity and its expansion (so you can always re-entitize mechanistically if you
need to).
(c) preprocess, as suggested, to replace the ampersands
by something else: for example <?ent mdash?>.
Michael Kay
http://www.saxonica.com/
If I have data with character
entities such as § or — in the XML, what is the best way
to keep those intact, as is, after a parse. Are there any parsers that
have an option not to resolve entities? What is the best way you've
found to deal with this issue? Do you escape the ampersands (e.g.
&sect;) in a pre-process? Do you address it in an entity
handler to re-write the original entity text? This seems like a real
pain that must have a common solution.
Thanks!
>> Randy
******************************************************************************
Do you get frequent requests for copies of certain sections of your Code?
We can reproduce chapters of your Code in handy pamphlet format - no minimum
quantity required! Order yours today.
