I tried running my Tomcat 3.1 code that uses the xml-tr package with Tomcat 3.2Beta7 with no success. Any of my code that parsed XML documents failed under 3.2. After further investigation, I've determined what the problem is. With the Sun parser/dom implementation that comes with xml-tr, ignoreable whitespace is ignored by default. This means that you will not see Text nodes that are just whitespace that is used for document formating within the DOM model. With the xml package with 3.2, the default is just the opposite; ignorable whitespace is not ignored and shows up as Text nodes. If you've written your DOM processing code not to expect Text nodes that represent ignoreable whitespace, your code will fail miserably. If you get a hold of the XmlDocumentBuilder instance, you can call setIgnoringLexicalInfo(true) to get back the old behavior. Just in case someone else runs into this problem.
