In our product we use AXIOM for XML processing and pass XMLin non-coalescing mode ("javax.xml.stream.isCoalescing",false). As a result of this when we get a soap message which has a lengthy CDATA block axiom breaks down it into multiple CDATA chunks.
In your email  I noted you have stated that there is a way of configuring Woodstox to report CDATA sections without splitting text nodes into chunks. Will you please be able to point a document or any other reference guid on how to do it.
Your prompt response would be so much appreciated.