jruby, a ruby interpreter written in java, makes it much easier to
interface with java libraries from ruby than the ruby java bridge
(where i couldnt, for the life of me, make java accept a pure ruby
java.io.InputStream implementation). I chose woodstox's implementation
of stax (a pull parser that apparently performs better than even the
xerces stream parser) to play with the idea.

As an aside, doesn't the idea of controlling the flow of the parser
programmatically (by calling 'next' on the reader to get an event)
make so much more sense than having the entire flow of your code
controlled by an xml file?


#!/usr/bin/env jruby
#
#  Created by Travis Tilley on 2006-12-09.
#  Copyright (c) 2006. All rights reserved.

require 'java'

# this class can wrap a generic ruby IO duck and make it quack like an
InputStream
include_class 'org.jruby.util.IOInputStream'
# stax
include_class 'javax.xml.stream.XMLInputFactory'
include_class 'javax.xml.stream.XMLStreamException'
include_class 'javax.xml.stream.XMLStreamReader'

# the java interface for accessing XMLStreamConstants doesnt like me
XMLStreamConstants = {
 :START_ELEMENT            => 1,
 :END_ELEMENT              => 2,
 :PROCESSING_INSTRUCTION   => 3,
 :CHARACTERS               => 4,
 :COMMENT                  => 5,
 :SPACE                    => 6,
 :START_DOCUMENT           => 7,
 :END_DOCUMENT             => 8,
 :ENTITY_REFERENCE         => 9,
 :ATTRIBUTE                => 10,
 :DTD                      => 11,
 :CDATA                    => 12,
 :NAMESPACE                => 13,
 :NOTATION_DECLARATION     => 14,
 :ENTITY_DECLARATION       => 15
}

# create a standard ruby File object
xml = File.new('tst.xml')
# create InputStream interface to the IO object
xmlstream = IOInputStream.new(xml)

# create a stax stream reader and hook it to the InputStream
factory = XMLInputFactory.newInstance
reader = factory.createXMLStreamReader(xmlstream)

# implement a "stream parser" on top of the pull parser (kinda)
event = reader.getEventType
while event
 case event
 when XMLStreamConstants[:START_DOCUMENT]
   puts "Start Document"
 when XMLStreamConstants[:START_ELEMENT]
   puts "Start Element: " + reader.getLocalName
   reader.getAttributeCount.times do |count|
     puts "  Attribute: " + reader.getAttributeLocalName(count)
     puts "  -> Value: " + reader.getAttributeValue(count)
   end
 when XMLStreamConstants[:CHARACTERS]
   puts "Text: " + reader.getText if ! reader.isWhiteSpace
 when XMLStreamConstants[:END_ELEMENT]
   puts "End Element: " + reader.getLocalName
 when XMLStreamConstants[:END_DOCUMENT]
   puts "End Document"
 end

 if reader.hasNext
   event = reader.next
 else
   event = nil
 end
end