I've committed some code to CVS at src/examples/api/document-markup.
This code demonstrates one method for subclassing Digester to implement the
parsing of "document markup" style XML.
After the 1.6 release would be a good time to push for a solution "built-in" to
Digester. I'm in favour of this in general, but not in favour of holding the
release up to do it now, given that the example seems to show quite a reasonable
Justin: If this is satisfactory, could you please close the bug?
If you have any concerns, or suggestions for better solutions, then please bring
them up here or on the mailing list ASAP, to get them in before 1.6.