SAXConverter no longer needlessly duplicates namespace declarations,
and should be faster. The fat Text class should be working again. A few
small issues have been fixed in Jaxen.


Changes: The latest Unicode normalization tables are used. A lookup method was added in XPathContext, which finds the namespace URI for a prefix. xml:id attributes are no longer checked for NCNames. Xerces was upgraded to 2.8.0, DTD-only version. DOMConverter can accept a NodeFactory to be used in creating the XOM document. Various bugs were fixed, especially in canonicalization and XPath, and one that could cause toXML to generate malformed output when a namespace URI contained an ampersand. The version number is correctly reported by java -jar.


Changes: New features implemented include node factories for DOMConverter and a lookup method in XPathContext. In addition, some obscure bugs have been fixed.


Changes: This release adds XPath, exclusive XML Canonicalization, Document Subset Canonicalization, a setInternalDTDSubset method in DocType, xml:id support, and externally supplied XSL parameters. It ranges from two to four times faster than version 1.0 for many common operations.


Changes: SAXConverter no longer needlessly duplicates namespace declarations, and should be faster. The fat Text class should be working again. A few small issues have been fixed in Jaxen.


Changes: The API is now considered to be stable, and probably won't change before 1.1 final. Beta 3 focuses on performance. This release is measurably faster than beta 2 for many common operations, and probably at least twice as fast as XOM 1.0.