[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [xml-dev] use SAX to get "tree" under XML node
> The response from Ron Bourrett was on the mark except
> that I'd have said to always use an XMLWriter of some kind.
> There are too many corner cases you can trip over when you
> try to turn "raw data" into XML text.
I think it depends on XML subset one is using.
When using "Monastic XML" , the only thing
needed is encoding of few symbols into
corresponding &entities; and that should be all (?)
Generating a full-blown XML with
several namespaces becomes a tricky puzzle,
very much different from trivial
_normattr and _normvalue.
> (And I'd not describe turning a partial event stream into text
> to be getting a "tree" ... in that case I'd be more likely to build
> a data structure, maybe even populate a DOM fragment, with
> those events.)
This is very interesting point and I know that
I could not understand your comment, if I read it just
a week ago.
Here is why.
I've been generating XML files for a very long time
with those print " " statements. But now, when
I have Chunks implemented, the way I generate
XML is :
c = Chunk()
c.cput("/doc/title/#value' , "List of books" )
c.cput("/doc/#array/book/#value", books_list )
c.emit() # this has _normattr and _normciontent implemented ;-)
That will dump out
<title>List of books</title>
<book> book 1</book>
<book> book 2</book>
I guess one can do similial things with DOM,
but because XPath is not bi-directional, I'm
not sure that DOM-based code would be as
terse as an (actual) code written above.
My point was to agree that 'populate-then-emit'
strategy is definately viable and may result in
a terse and robust code, comparing it to
'write it out with "print" '