A somewhat DOM-like library that implements lazy generation of the
node tree, i.e. it's generated from the data upon lookup. There's
also a little bit of XPath evaluation to do queries on the node
tree.
Implementation note: This is generally more pragmatic than
Parser.XML.DOM , meaning it's not so pretty and compliant, but
more efficient.
Implementation status: There's only enough implemented to parse a
node tree from source and access it, i.e. modification functions
aren't implemented. Data hiding stuff like NodeList and
NamedNodeMap is not implemented, partly since it's cumbersome to
meet the "live" requirement. Also, Parser.HTML is used in XML
mode to parse the input. Thus it's too error tolerant to be XML
compliant, and it currently doesn't handle DTD elements, like
"<!DOCTYPE", or the XML declaration (i.e. "<?xml
version='1.0'?>".