If you look closer, you'll see that we actually stop reading the document
when we get to the contents of the body by virtue of PartialXMLStreamReader.
The idea is that we can have a two step process for those times we want to
process the whole document.
1. Start creating a document from the XMLStreamReader, but stop at the first
node in <soap:Body>
2. Continue reading in the body contents in SAAJInInterceptor if we need to

This makes it easy to support WS-Security UsernameToken scenarios without
parsing the whole message into a DOM. We are just create a DOM (starting at
the Envelope) which doesn't have any Body unless we need it to.

Does that make sense?

- Dan

On 4/12/07, Guillaume Nodet <[EMAIL PROTECTED]> wrote:

While stepping in the ReadHeadersInterceptor, I've just seen a recent
change
in it:


http://fisheye6.cenqua.com/browse/celtixfire/trunk/rt/bindings/soap/src/main/java/org/apache/cxf/binding/soap/interceptor/ReadHeadersInterceptor.java?r1=509004&r2=519759

The main effect is that the whole message is always transformed to a DOM
document,
instead of just the headers element, which seems quite a bad idea imo.

Any reason for that ?

--
Cheers,
Guillaume Nodet
------------------------
Principal Engineer, IONA
Blog: http://gnodet.blogspot.com/




--
Dan Diephouse
Envoi Solutions
http://envoisolutions.com | http://netzooid.com/blog

Reply via email to