CyberNeko HTML Parser

A charset regression was fixed.

The license was changed to Apache 2.0 and the version number was boosted to reflect the maturity of the project. Project files were reorganized to decouple them from the rest of the CyberNeko Tools for XNI. xercesMinimal.jar and source were updated so that NekoHTML compiles using Xerces-J 2.9.1. The default behavior was changed to not normalize attribute values and a new feature was added to allow users to turn on normalization. The build was modified to target compilation for Java 1.3. Suggested paragraph tag balancing was adjusted and various reported bugs were fixed.

A feature to allow a scanner to fix character
entity references for Microsoft Windows characters
was added. The nekohtmlXni.jar file is no longer
built by default. Tag-balancing was changed to
allow headers inside of links. Handling of the
blockquote tag, a tag-balancing bug for unknown
elements, the mapping of the encoding name in meta
tags, various namespace binding bugs, and a
no-such-method exception when using the
augmentations feature with older versions of
Xerces2 were fixed.

This release added features for stripping CDATA delimiters from script and style tags, made augmentations, bugfixes, and performance enhancements, and fixed some tag balancing issues.

This version implements scanning of XML declaration, fixes a
script tag scanning bug, and adds version class and manifest
entries to query product information.