XOM API for Processing XML with Java

[Elliotte Harold <elharo AT metalab.unc.edu>]

luca wrote:

> I think you
> just have to implement a way to tell XOM to ignore DTDs no
> matter what. This is defying the purpose of XOM being
> the tool for 80% of everyone's XML tasks

That's not going to happen, at least not in 1.0, probably not ever. 
Requiring the external DTD subset is a deliberate design decision that 
IMNSHO falls squarely in the 80%. There are reasons you might want to 
ignore this, but those use cases are less common, and I think on the 20% 
side of the cut. If you really need skipped entities (and it doesn't 
sound like you do) then XOM is not for you, but if all you want is to 
not load external resources, then you can do this by using a SAX 
EntityResolver.

For the time being, you should make a local copy of your DTD, and use a 
SAX EntityResolver to pick out that copy rather than the remote version 
at wapforum.org. You could implement this on top of catalogs using the 
Apache catalog resolver, or you could roll your own. It's not hard. Set 
up an XMLReader with this resolver, and then use that XMLReader to 
construct XOM's Builder. If you use the Apache catalog resolver, the 
code looks like this:

XMLReader reader = new org.apache.xerces.parsers.SAXParser();
reader = new org.apache.xml.resolver.tools.ResolvingXMLFilter(reader);
Builder parser = new Builder(reader);
Document doc = parser.build("http://www.example.org";);

You'd also need to set up the catalog file in the appropriate place.

Post-1.0, I would like to provide built in catalog support so you could 
replace a reference to an external DTD with a reference to a local copy 
(among other cool hacks). However, you'll always have to read something.

--
Elliotte Rusty Harold
elharo AT metalab.unc.edu