<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE article PUBLIC "-//OASIS//DTD Simplified DocBook XML V1.1b1//EN" "/usr/share/xml/docbook/schema/dtd/4.5/docbookx.dtd">
<!--
Copyright (c) 2001-2014, Jeff Martin, Tim Bacon
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:

    * Redistributions of source code must retain the above copyright
      notice, this list of conditions and the following disclaimer.
    * Redistributions in binary form must reproduce the above
      copyright notice, this list of conditions and the following
      disclaimer in the documentation and/or other materials provided
      with the distribution.
    * Neither the name of the xmlunit.sourceforge.net nor the names
      of its contributors may be used to endorse or promote products
      derived from this software without specific prior written
      permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.
-->

<article>
  <title>XMLUnit Java User's Guide</title>
  <articleinfo>
    <authorgroup>
      <author>
        <firstname>Tim</firstname>
        <surname>Bacon</surname>
      </author>
      <author>
        <firstname>Stefan</firstname>
        <surname>Bodewig</surname>
      </author>
    </authorgroup>
    <revhistory>
      <revision>
        <revnumber>1.0</revnumber>
        <date>January 2003</date>
        <author>
          <firstname>Tim</firstname>
          <surname>Bacon</surname>
        </author>
        <revremark>Documentation for XMLUnit Java 1.0</revremark>
      </revision>
      <revision>
        <revnumber>1.1</revnumber>
        <date>April 2007</date>
        <revremark>Documentation for XMLUnit Java 1.1</revremark>
      </revision>
      <revision>
        <revnumber>1.2</revnumber>
        <date>June 2008</date>
        <revremark>Documentation for XMLUnit Java 1.2</revremark>
      </revision>
      <revision>
        <revnumber>1.3</revnumber>
        <date>September 2009</date>
        <revremark>Documentation for XMLUnit Java 1.3</revremark>
      </revision>
      <revision>
        <revnumber>1.4</revnumber>
        <date>February 2013</date>
        <revremark>Documentation for XMLUnit Java 1.4</revremark>
      </revision>
      <revision>
        <revnumber>1.5</revnumber>
        <date>September 2013</date>
        <revremark>Documentation for XMLUnit Java 1.5</revremark>
      </revision>
      <revision>
        <revnumber>1.6</revnumber>
        <date>December 2014</date>
        <revremark>Documentation for XMLUnit Java 1.6</revremark>
      </revision>
    </revhistory>
  </articleinfo>

  <section id="A Tour of XMLUnit"><title>A Tour of XMLUnit</title>

    <para>This first section contains a tour through XMLUnit's
    features, the next sections will cover them in more detail.</para>

    <para>Note that it has a strong focus on using the
    <literal>XMLTestCase</literal> class which is one option to use
    XMLUnit, but not the only one.  XMLUnit's features can be fully
    used without any dependency on JUnit at all.</para>

    <section id="What is XMLUnit?"><title>What is XMLUnit?</title>

      <para>XMLUnit enables JUnit-style assertions to be made about
      the content and structure of XML<footnote id="more on
      JUnit"><para>For more information on JUnit see <ulink
      url="http://www.junit.org">http://www.junit.org</ulink></para></footnote>. It
      is an open source project hosted at <ulink
      url="http://xmlunit.sourceforge.net/">http://xmlunit.sourceforge.net/</ulink>
      that grew out of a need to test a system that generated and
      received custom XML messages. The problem that we faced was how
      to verify that the system generated the correct message from a
      known set of inputs. Obviously we could use a DTD or a schema to
      validate the message output, but this approach wouldn't allow us
      to distinguish between valid XML with correct content (e.g.
      element <literal><![CDATA[<foo>bar</foo>]]></literal>) and valid
      XML with incorrect content (e.g.  element
      <literal><![CDATA[<foo>baz</foo>]]></literal>). What we really
      wanted was an <literal>assertXMLEqual()</literal> method, so we
      could compare the message that we expected the system to
      generate and the message that the system actually generated. And
      that was the beginning of XMLUnit.</para>
    </section>
    <section id="Quick Tour"><title>Quick tour</title>

      <para>XMLUnit provides a single JUnit extension class,
      <literal>XMLTestCase</literal>, and a set of supporting classes
      that allow assertions to be made about:</para>

      <itemizedlist>
        <listitem>The differences between two pieces of XML (via
        <literal>Diff</literal> and <literal>DetailedDiff</literal>
        classes)</listitem>

        <listitem>The validity of a piece of XML (via
        <literal>Validator</literal> class)</listitem>

        <listitem> The outcome of transforming a piece of XML using
        XSLT (via <literal>Transform</literal> class)</listitem>

        <listitem>The evaluation of an XPath expression on a piece of
        XML (via classes implementing the
        <literal>XpathEngine</literal> interface)</listitem>

        <listitem>Individual nodes in a piece of XML that are exposed
        by DOM Traversal (via <literal>NodeTest</literal>
        class)</listitem>
      </itemizedlist>

      <para>XMLUnit can also treat HTML content, even badly-formed
      HTML, as valid XML to allow these assertions to be made about
      web pages (via the <literal>HTMLDocumentBuilder</literal>
      class).</para>
    </section>

    <section id="Glossary"><title>Glossary</title>

      <para>As with many projects some words in XMLUnit have
      particular meanings so here is a quick overview. A
      <emphasis>piece</emphasis> of XML is a DOM Document, a String
      containing marked-up content, or a Source or Reader that allows
      access to marked-up content within some resource. XMLUnit
      compares the expected <emphasis>control</emphasis> XML to some
      actual <emphasis>test</emphasis> XML. The comparison can reveal
      that two pieces of XML are <emphasis>identical</emphasis>,
      <emphasis>similar</emphasis> or
      <emphasis>different</emphasis>. The unit of measurement used by
      the comparison is a <emphasis>difference</emphasis>, and
      differences can be either <emphasis>recoverable</emphasis> or
      <emphasis>unrecoverable</emphasis>. Two pieces of XML are
      <emphasis>identical</emphasis> if there are <emphasis>no
      differences</emphasis> between them,
      <emphasis>similar</emphasis> if there are <emphasis>only
      recoverable differences</emphasis> between them, and
      <emphasis>different</emphasis> if there are <emphasis>any
      unrecoverable differences</emphasis> between them.</para>
    </section>

    <section id="Configuring XMLUnit"><title>Configuring XMLUnit</title>

      <para>There are many Java XML parsers available, and XMLUnit
      should work with any JAXP compliant parser library, such as
      Xerces-J <footnote id="xerces-link"><para><ulink
      url="http://xerces.apache.org/">http://xerces.apache.org/</ulink></para></footnote>
      from the Apache Software Foundation. To use the XSLT and XPath
      features of XMLUnit a Trax (the XSLT portion of JAXP) compliant
      transformation engine is required, such as Xalan-J<footnote
      id="xalan-link"><para><ulink
      url="http://xalan.apache.org/">http://xalan.apache.org/</ulink></para></footnote>,
      from the Apache Software Foundation.  To configure XMLUnit to
      use a specific parser and transformation engine set three System
      properties before any tests are run, e.g.</para>

      <example><title>Configuring JAXP via System Properties</title>
        <programlisting language="Java"><![CDATA[
System.setProperty("javax.xml.parsers.DocumentBuilderFactory",
    "org.apache.xerces.jaxp.DocumentBuilderFactoryImpl");
System.setProperty("javax.xml.parsers.SAXParserFactory",
    "org.apache.xerces.jaxp.SAXParserFactoryImpl");
System.setProperty("javax.xml.transform.TransformerFactory",
    "org.apache.xalan.processor.TransformerFactoryImpl");
]]></programlisting>
      </example>

      <para>You may want to read <xref linkend="JAXP"/> for more
      details - in particular if you are using Java 1.4 or
      later.</para>

      <para>Alternatively there are static methods on the XMLUnit
      class that can be called directly. The advantage of this
      approach is that you can specify a different parser class for
      control and test XML and change the current parser class at any
      time in your tests, should you need to make assertions about the
      compatibility of different parsers.</para>

      <example><title>Configuring JAXP via XMLUnit class</title>
        <programlisting language="Java"><![CDATA[
XMLUnit.setControlParser("org.apache.xerces.jaxp.DocumentBuilderFactoryImpl");
XMLUnit.setTestParser("org.apache.xerces.jaxp.DocumentBuilderFactoryImpl");
XMLUnit.setSAXParserFactory("org.apache.xerces.jaxp.SAXParserFactoryImpl");
XMLUnit.setTransformerFactory("org.apache.xalan.processor.TransformerFactoryImpl");
]]></programlisting>
      </example>

    </section>

    <section id="Writing XML comparison tests">
      <title>Writing XML comparison tests</title>

      <para>Let's say we have two pieces of XML that we wish to
      compare and assert that they are equal. We could write a simple
      test class like this:</para>

      <example><title>A simple comparison test</title>
        <programlisting language="Java"><![CDATA[
public class MyXMLTestCase extends XMLTestCase {
    public MyXMLTestCase(String name) {
        super(name);
    }

    public void testForEquality() throws Exception {
        String myControlXML = "<msg><uuid>0x00435A8C</uuid></msg>";
        String myTestXML = "<msg><localId>2376</localId></msg>";
        assertXMLEqual("Comparing test xml to control xml",
                       myControlXML, myTestXML);
    }
}]]></programlisting></example>

      <para>The <literal>assertXMLEqual</literal> test will pass if
      the control and test XML are either similar or
      identical. Obviously in this case the pieces of XML are
      different and the test will fail. The failure message indicates
      both what the difference is and the XPath locations of the nodes
      that were being compared:</para>

      <programlisting><![CDATA[
Comparing test xml to control xml
[different] Expected element tag name 'uuid' but was 'localId' - comparing <uuid...> at /msg[1]/uuid[1] to <localId...> at /msg[1]/localId[1]
]]></programlisting>

      <para>When comparing pieces of XML, the
      <literal>XMLTestCase</literal> actually creates an instance of
      the <literal>Diff</literal> class. The <literal>Diff</literal>
      class stores the result of an XML comparison and makes it
      available through the methods <literal>similar()</literal> and
      <literal>identical()</literal>. The
      <literal>assertXMLEqual()</literal> method tests the value of
      <literal>Diff.similar()</literal> and the
      <literal>assertXMLIdentical()</literal> method tests the value
      of <literal>Diff.identical()</literal>.</para>

      <para>It is easy to create a <literal>Diff</literal> instance
      directly without using the <literal>XMLTestCase</literal> class
      as below:</para>

      <example><title>Creating a <literal>Diff</literal>
      instance</title>
        <programlisting language="Java"><![CDATA[
public void testXMLIdentical()throws Exception {
    String myControlXML =
        "<struct><int>3</int><boolean>false</boolean></struct>";
    String myTestXML =
        "<struct><boolean>false</boolean><int>3</int></struct>";
    Diff myDiff = new Diff(myControlXML, myTestXML);
    assertTrue("XML similar " + myDiff.toString(),
               myDiff.similar());
    assertTrue("XML identical " + myDiff.toString(),
               myDiff.identical());
}]]></programlisting></example>

      <para>This test fails as two pieces of XML are similar but not
      identical if their nodes occur in a different sequence. The
      failure message reported by JUnit from the call to
      <literal>myDiff.toString()</literal> looks like this:</para>

      <programlisting><![CDATA[
[not identical] Expected sequence of child nodes '0' but was '1' - comparing <int...> at /struct[1]/int[1] to <int...> at /struct[1]/int[1]
]]></programlisting>

      <para>For efficiency reasons a <literal>Diff</literal> stops the
      comparison process as soon as the first difference is found. To
      get all the differences between two pieces of XML an instance of
      the <literal>DetailedDiff</literal> class, a subclass of
      <literal>Diff</literal>, is required. Note that a
      <literal>DetailedDiff</literal> is constructed using an existing
      <literal>Diff</literal> instance.</para>

      <para>Consider this test that uses a DetailedDiff:</para>

      <example><title>Using <literal>DetailedDiff</literal></title>
        <programlisting language="Java"><![CDATA[
public void testAllDifferences() throws Exception {
    String myControlXML = "<news><item id=\"1\">War</item>"
        + "<item id=\"2\">Plague</item>"
        + "<item id=\"3\">Famine</item></news>";
    String myTestXML = "<news><item id=\"1\">Peace</item>"
        + "<item id=\"2\">Health</item>"
        + "<item id=\"3\">Plenty</item></news>";
    DetailedDiff myDiff = new DetailedDiff(new Diff(myControlXML, myTestXML));
    List allDifferences = myDiff.getAllDifferences();
    assertEquals(myDiff.toString(), 2, allDifferences.size());
}]]></programlisting></example>

      <para>This test fails with the message below as each of the 3
      news items differs between the control and test XML:</para>

      <programlisting><![CDATA[
[different] Expected text value 'War' but was 'Peace' - comparing <item...>War</item> at /news[1]/item[1]/text()[1] to <item...>Peace</item> at /news[1]/item[1]/text()[1]
[different] Expected text value 'Plague' but was 'Health' - comparing <item...>Plague</item> at /news[1]/item[2]/text()[1] to <item...>Health</item> at /news[1]/item[2]/text()[1]
[different] Expected text value 'Famine' but was 'Plenty' - comparing <item...>Famine</item> at /news[1]/item[3]/text()[1] to <item...>Plenty</item> at /news[1]/item[3]/text()[1]
expected <2> but was <3>
]]></programlisting>

      <para>The List returned from the
      <literal>getAllDifferences()</literal> method contains
      <literal>Difference</literal> instances. These instances
      describe both the type<footnote id="DifferenceConstants"><para>A
      full set of prototype <literal>Difference</literal> instances -
      one for each type of difference - is defined using final static
      fields in the <literal>DifferenceConstants</literal>
      class.</para></footnote> of difference found between a control
      node and test node and the <literal>NodeDetail</literal> of
      those nodes (including the XPath location of each
      node). <literal>Difference</literal> instances are passed at
      runtime in notification events to a registered
      <literal>DifferenceListener</literal>, an interface whose
      default implementation is provided by the
      <literal>Diff</literal> class.</para>

      <para>However it is possible to override this default behaviour
      by implementing the interface in your own class.  The
      <literal>IgnoreTextAndAttributeValuesDifferenceListener</literal>
      class is an example of how to implement a custom
      <literal>DifferenceListener</literal>. It allows an XML
      comparison to be made that ignores differences in the values of
      text and attribute nodes, for example when comparing a skeleton
      or outline piece of XML to some generated XML.</para>

      <para>The following test illustrates the use of a custom
      <literal>DifferenceListener</literal>:</para>

      <example><title>Using a custom
          <literal>DifferenceListener</literal></title>
        <programlisting language="Java"><![CDATA[
public void testCompareToSkeletonXML() throws Exception {
    String myControlXML = "<location><street-address>22 any street</street-address><postcode>XY00 99Z</postcode></location>";
    String myTestXML = "<location><street-address>20 east cheap</street-address><postcode>EC3M 1EB</postcode></location>";
    DifferenceListener myDifferenceListener = new IgnoreTextAndAttributeValuesDifferenceListener();
    Diff myDiff = new Diff(myControlXML, myTestXML);
    myDiff.overrideDifferenceListener(myDifferenceListener);
    assertTrue("test XML matches control skeleton XML",
               myDiff.similar());
}]]></programlisting></example>

      <para>The <literal>DifferenceEngine</literal> class generates
      the events that are passed to a
      <literal>DifferenceListener</literal> implementation as two
      pieces of XML are compared. Using recursion it navigates through
      the nodes in the control XML DOM, and determines which node in
      the test XML DOM qualifies for comparison to the current control
      node. The qualifying test node will match the control node's
      node type, as well as the node name and namespace (if defined
      for the control node).</para>

      <para>However when the control node is an
      <literal>Element</literal>, it is less straightforward to
      determine which test <literal>Element</literal> qualifies for
      comparison as the parent node may contain repeated child
      <literal>Element</literal>s with the same name and namespace. So
      for <literal>Element</literal> nodes, an instance of the
      <literal>ElementQualifier</literal> interface is used determine
      whether a given test <literal>Element</literal> node qualifies
      for comparison with a control <literal>Element</literal>
      node. This separates the decision about whether two
      <literal>Elements</literal> should be compared from the decision
      about whether those two <literal>Elements</literal> are
      considered similar. By default an
      <literal>ElementNameQualifier</literal> class is used that
      compares the nth child <literal><![CDATA[<abc>]]></literal> test
      element to the nth child <literal><![CDATA[<abc>]]></literal>
      control element, i.e. the sequence of the child elements in the
      test XML is important. However this default behaviour can be
      overridden using an
      <literal>ElementNameAndTextQualifier</literal> or
      <literal>ElementNameAndAttributesQualifier</literal>.</para>

      <para>The test below demonstrates the use of a custom
      <literal>ElementQualifier</literal>:</para>

      <example><title>Using a custom
          <literal>ElementQualifier</literal></title>
        <programlisting language="Java"><![CDATA[
public void testRepeatedChildElements() throws Exception {
    String myControlXML = "<suite>"
        + "<test status=\"pass\">FirstTestCase</test>"
        + "<test status=\"pass\">SecondTestCase</test></suite>";
    String myTestXML = "<suite>"
        + "<test status=\"pass\">SecondTestCase</test>"
        + "<test status=\"pass\">FirstTestCase</test></suite>";
    assertXMLNotEqual("Repeated child elements in different sequence order are not equal by default",
                      myControlXML, myTestXML);
    Diff myDiff = new Diff(myControlXML, myTestXML);
    myDiff.overrideElementQualifier(new ElementNameAndTextQualifier());
    assertXMLEqual("But they are equal when an ElementQualifier controls which test element is compared with each control element",
                    myDiff, true);
}]]></programlisting></example>

      <para>Note: calling <literal>toString</literal> on an instance
      of <literal>Diff</literal> or <literal>DetailedDiff</literal>
      will perform the comparision and cache its result immediately.
      If you change the <literal>DifferenceListener</literal> or
      <literal>ElementQualifier</literal> after calling
      <literal>toString</literal> it won't have any effect.</para>

    </section>

    <section id="Comparing XML Transformations">
      <title>Comparing XML Transformations</title>

      <para>XMLUnit can test XSLT transformations at a high level using
      the <literal>Transform</literal> class that wraps an
      <literal>javax.xml.transform.Transformer</literal>
      instance. Knowing the input XML, input stylesheet and expected
      output XML we can assert that the output of the transformation
      matches the expected output as follows:</para>

      <example><title>Testing the Result of a Transformation</title>
        <programlisting language="Java"><![CDATA[
public void testXSLTransformation() throws Exception {
    String myInputXML = "...";
    File myStylesheetFile = new File("...");
    Transform myTransform = new Transform(myInputXML, myStylesheetFile);
    String myExpectedOutputXML = "...";
    Diff myDiff = new Diff(myExpectedOutputXML, myTransform);
    assertTrue("XSL transformation worked as expected", myDiff.similar());
}]]></programlisting></example>

      <para>The <literal>getResultString()</literal> and
      <literal>getResultDocument()</literal> methods of the
      <literal>Transform</literal> class can be used to access the
      result of the XSLT transformation programmatically if required,
      for example as below:</para>

      <example><title>Using <literal>Transform</literal>
          programmatically</title>
        <programlisting language="Java"><![CDATA[
public void testAnotherXSLTransformation() throws Exception {
    File myInputXMLFile = new File("...");
    File myStylesheetFile = new File("...");
    Transform myTransform = new Transform(
        new StreamSource(myInputXMLFile),
        new StreamSource(myStylesheetFile));
    Document myExpectedOutputXML =
       XMLUnit.buildDocument(XMLUnit.getControlParser(),
                             new FileReader("..."));
    Diff myDiff = new Diff(myExpectedOutputXML,
    myTransform.getResultDocument());
    assertTrue("XSL transformation worked as expected", myDiff.similar());
}]]></programlisting></example>

    </section>

    <section id="Validation Tests"><title>Validation Tests</title>

      <para>XML parsers that validate a piece of XML against a DTD are
      common, however they rely on a DTD reference being present in
      the XML, and they can only validate against a single DTD. When
      writing a system that exchanges XML messages with third parties
      there are times when you would like to validate the XML against
      a DTD that is not available to the recipient of the message and
      so cannot be referenced in the message itself. XMLUnit provides
      a <literal>Validator</literal> class for this purpose.</para>

      <example><title>Validating Against a DTD</title>
        <programlisting language="Java"><![CDATA[
public void testValidation() throws Exception {
    XMLUnit.getTestDocumentBuilderFactory().setValidating(true);
    // As the document is parsed it is validated against its referenced DTD
    Document myTestDocument = XMLUnit.buildTestDocument("...");
    String mySystemId = "...";
    String myDTDUrl = new File("...").toURL().toExternalForm();
    Validator myValidator = new Validator(myTestDocument, mySystemId,
                                          myDTDUrl);
    assertTrue("test document validates against unreferenced DTD",
               myValidator.isValid());
}]]></programlisting></example>

      <para>Starting with XMLUnit 1.1, the
      <literal>Validator</literal> class can also validate against one
      or more XML Schema definitions.  See <xref
      linkend="XML Schema Validation"/> for details.</para>

      <para>XMLUnit 1.2 introduces a new <literal>Validator</literal>
        class that relies on JAXP
        1.3's <literal>javax.xml.validation</literal> package.  This
        Validator can validate against W3C XML Schema, but may support
        different Schema languages like RELAX NG if your JAXP
        implementation supports it.  See
        <xref linkend="JAXP 1.3 Validation"/> for details.</para>

    </section>

    <section id="Xpath Tests"><title>XPath Tests</title>

      <para>One of the strengths of XML is the ability to
      programmatically extract specific parts of a document using
      XPath expressions. The <literal>XMLTestCase</literal> class
      offers a number of XPath related assertion methods, as
      demonstrated in this test:</para>

      <example><title>Using XPath Tests</title>
        <programlisting language="Java"><![CDATA[
public void testXPaths() throws Exception {
    String mySolarSystemXML = "<solar-system>"
        + "<planet name='Earth' position='3' supportsLife='yes'/>"
        + "<planet name='Venus' position='4'/></solar-system>";
    assertXpathExists("//planet[@name='Earth']", mySolarSystemXML);
    assertXpathNotExists("//star[@name='alpha centauri']",
                         mySolarSystemXML);
    assertXpathsEqual("//planet[@name='Earth']",
                      "//planet[@position='3']", mySolarSystemXML);
    assertXpathsNotEqual("//planet[@name='Venus']",
                         "//planet[@supportsLife='yes']",
                         mySolarSystemXML);
}]]></programlisting></example>

      <para>When an XPath expression is evaluated against a piece of
      XML a <literal>NodeList</literal> is created that contains the
      matching <literal>Node</literal>s. The methods in the previous
      test <literal>assertXpathExists</literal>,
      <literal>assertXpathNotExists</literal>,
      <literal>assertXpathsEqual</literal>, and
      <literal>assertXpathsNotEqual</literal> use these
      <literal>NodeList</literal>s. However, the contents of a
      <literal>NodeList</literal> can be flattened (or
      <literal>String</literal>-ified) to a single value, and XMLUnit
      also allows assertions to be made about this single value, as in
      this test<footnote id="XpathEngine note"><para>Each of the
      <literal>assertXpath...()</literal> methods uses an
      implementation of the <literal>XpathEngine</literal> interface
      to evaluate an XPath expression.</para></footnote>:</para>

      <example><title>Testing XPath Values</title>
        <programlisting language="Java"><![CDATA[
public void testXPathValues() throws Exception {
    String myJavaFlavours = "<java-flavours>"
        + "<jvm current='some platforms'>1.1.x</jvm>"
        + "<jvm current='no'>1.2.x</jvm>"
        + "<jvm current='yes'>1.3.x</jvm>"
        + "<jvm current='yes' latest='yes'>1.4.x</jvm></javaflavours>";
    assertXpathEvaluatesTo("2", "count(//jvm[@current='yes'])",
                           myJavaFlavours);
    assertXpathValuesEqual("//jvm[4]/@latest", "//jvm[4]/@current",
                           myJavaFlavours);
    assertXpathValuesNotEqual("//jvm[2]/@current",
                              "//jvm[3]/@current", myJavaFlavours);
}]]></programlisting></example>

      <para>XPaths are especially useful where a document is made up
      largely of known, unchanging content with only a small amount of
      changing content created by the system. One of the main areas
      where constant "boilerplate" markup is combined with system
      generated markup is of course in web applications. The power of
      XPath expressions can make testing web page output quite
      trivial, and XMLUnit supplies a means of converting even very
      badly formed HTML into XML to aid this approach to
      testing.</para>

      <para>The <literal>HTMLDocumentBuilder</literal> class uses the
      Swing HTML parser to convert marked-up content to Sax
      events. The <literal>TolerantSaxDocumentBuilder</literal> class
      handles the Sax events to build up a DOM document in a tolerant
      fashion i.e. without mandating that opened elements are
      closed. (In a purely XML world this class would have no purpose
      as there are plenty of Sax event handlers that can build DOM
      documents from well formed content). The test below illustrates
      how the use of these classes:</para>

      <example><title>Working with non well-formed HTML</title>
        <programlisting language="Java"><![CDATA[
public void testXpathsInHTML() throws Exception {
    String someBadlyFormedHTML = "<html><title>Ugh</title>"
        + "<body><h1>Heading<ul>"
        + "<li id='1'>Item One<li id='2'>Item Two";
    TolerantSaxDocumentBuilder tolerantSaxDocumentBuilder =
        new TolerantSaxDocumentBuilder(XMLUnit.getTestParser());
    HTMLDocumentBuilder htmlDocumentBuilder =
        new HTMLDocumentBuilder(tolerantSaxDocumentBuilder);
    Document wellFormedDocument =
        htmlDocumentBuilder.parse(someBadlyFormedHTML);
    assertXpathEvaluatesTo("Item One", "/html/body//li[@id='1']",
                           wellFormedDocument);
}]]></programlisting></example>

      <para>One of the key points about using XPaths with HTML content
      is that extracting values in tests requires the values to be
      identifiable. (This is just another way of saying that testing
      HTML is easier when it is written to be testable.) In the
      previous example id attributes were used to identify the list
      item values that needed to be testable, however class attributes
      or span and div tags can also be used to identify specific
      content for testing.</para>

    </section>

    <section id="Testing by Tree Walking">
      <title>Testing by Tree Walking</title>

      <para>The DOM specification allows a <literal>Document</literal>
      to optionally implement the <literal>DocumentTraversal</literal>
      interface. This interface allows an application to iterate over
      the <literal>Node</literal>s contained in a
      <literal>Document</literal>, or to "walk the DOM tree". The
      XMLUnit <literal>NodeTest</literal> class and
      <literal>NodeTester</literal> interface make use of
      <literal>DocumentTraversal</literal> to expose individual
      <literal>Node</literal>s in tests: the former handles the
      mechanics of iteration, and the latter allows custom test
      strategies to be implemented. A sample test strategy is supplied
      by the <literal>CountingNodeTester</literal> class that counts
      the nodes presented to it and compares the actual count to an
      expected count. The test below illustrates its use:</para>

      <example><title>Using <literal>CountingNodeTester</literal></title>
        <programlisting language="Java"><![CDATA[
public void testCountingNodeTester() throws Exception {
    String testXML = "<fibonacci><val>1</val><val>2</val><val>3</val>"
        + "<val>5</val><val>9</val></fibonacci>";
    CountingNodeTester countingNodeTester = new CountingNodeTester(4);
    assertNodeTestPasses(testXML, countingNodeTester, Node.TEXT_NODE);
}]]></programlisting></example>

      <para>This test fails as there are 5 text nodes, and JUnit
      supplies the following message:</para>

      <programlisting>
Expected node test to pass, but it failed! Counted 5 node(s) but
expected 4
      </programlisting>

      <para>Note that if your DOM implementation does not support the
      <literal>DocumentTraversal</literal> interface then XMLUnit will
      throw an <literal>IllegalArgumentException</literal> informing
      you that you cannot use the <literal>NodeTest</literal> or
      <literal>NodeTester</literal> classes. Unfortunately even if
      your DOM implementation does support
      <literal>DocumentTraversal</literal>, attributes are not exposed
      by iteration: however they can be examined from the
      <literal>Element</literal> node that contains them.</para>

      <para>While the previous test could have been easily performed
      using XPath, there are times when <literal>Node</literal>
      iteration is more powerful. In general, this is true when there
      are programmatic relationships between nodes that can be more
      easily tested iteratively. The following test uses a custom
      <literal>NodeTester</literal> class to illustrate the
      potential:</para>

      <example><title>Using a Custom <literal>NodeTester</literal></title>
        <programlisting language="Java"><![CDATA[
public void testCustomNodeTester() throws Exception {
    String testXML = "<fibonacci><val>1</val><val>2</val><val>3</val>"
        + "<val>5</val><val>9</val></fibonacci>";
    NodeTest nodeTest = new NodeTest(testXML);
    assertNodeTestPasses(nodeTest, new FibonacciNodeTester(),
                         new short[] {Node.TEXT_NODE,
                                      Node.ELEMENT_NODE},
                         true);
}

private class FibonacciNodeTester extends AbstractNodeTester {
    private int nextVal = 1, lastVal = 1, priorVal = 0;

    public void testText(Text text) throws NodeTestException {
        int val = Integer.parseInt(text.getData());
        if (nextVal != val) {
            throw new NodeTestException("Incorrect value", text);
        }
        nextVal = val + lastVal;
        priorVal = lastVal;
        lastVal = val;
    }

    public void testElement(Element element) throws NodeTestException {
        String name = element.getLocalName();
        if ("fibonacci".equals(name) || "val".equals(name)) {
            return;
        }
        throw new NodeTestException("Unexpected element", element);
    }

    public void noMoreNodes(NodeTest nodeTest) throws NodeTestException {
    }
}]]></programlisting></example>

      <para>The test fails because the XML contains the wrong value
      for the last number in the sequence:</para>

      <programlisting>
Expected node test to pass, but it failed! Incorrect value [#text: 9]
      </programlisting>

    </section>
  </section>

  <section id="Using XMLUnit">
    <title>Using XMLUnit</title>

    <section id="Requirements">
      <title>Requirements</title>

      <para>XMLUnit requires a JAXP compliant XML parser virtually
      everywhere.  Several features of XMLUnit also require a JAXP
      compliant XSLT transformer.  If it is available, a JAXP
      compliant XPath engine will be used for XPath tests.</para>

      <para>To build XMLUnit at least JAXP 1.2 is required, this is
      the version provided by the Java class library in JDK 1.4.  The
      JAXP 1.3 (i.e. Java5 and above) XPath engine can only be built
      when JAXP 1.3 is available.</para>

      <para>As long as you don't require support for XML Namespaces or
      XML Schema, any JAXP 1.1 compliant implementations should work
      at runtime.  For namespace and schema support you will need a
      parser that complies to JAXP 1.2 and supports the required
      feature.  The XML parser shipping with JDK 1.4 (a version of
      Apache Crimson) for example is compliant to JAXP 1.2 but doesn't
      support Schema validation.</para>

      <para>XMLUnit is supposed to build and run on any Java version
      after 1.3 (at least no new hard JDK 1.4 dependencies have been
      added in XMLUnit 1.1), but it has only been tested on JDK 1.4.2
      and above.</para>

      <para>To build XMLUnit JUnit 3.x (only tested with JUnit 3.8.x)
      is required.  It is not required at runtime unless you intend to
      use the <literal>XMLTestCase</literal> or
      <literal>XMLAssert</literal> classes.</para>
    </section>

    <section id="Basic Usage">
      <title>Basic Usage</title>

      <para>XMLUnit consists of a few classes all living in the
      <literal>org.custommonkey.xmlunit</literal> package.  You can
      use these classes directly from your code, no matter whether you
      are writing a unit test or want to use XMLUnit's features for
      any other purpose.</para>

      <para>This section provides a few hints of where to start if you
      want to use a certain feature of XMLUnit, more details can be
      found in the more specific sections later in this
      document.</para>

      <section id="Basic: Comparing Pieces of XML">
        <title>Comparing Pieces of XML</title>

        <para>Heart and soul of XMLUnit's comparison engine is
        <literal>DifferenceEngine</literal> but most of the time you
        will use it indirectly via the <literal>Diff</literal>
        class.</para>

        <para>You can influence the engine by providing (custom)
        implementations for various interfaces and by setting a couple
        of options on the <literal>XMLUnit</literal> class.</para>

        <para>More information is available in <xref
        linkend="Comparing Pieces of XML"/>.</para>

      </section>

      <section id="Basic: Validating"><title>Validating</title>

        <para>All validation happens in the
        <literal>Validator</literal> class.  The default is to
        validate against a DTD, but XML Schema validation can be
        enabled by an option (see
        <literal>Validator.useXMLSchema</literal>).</para>

        <para>Several options of the <literal>XMLUnit</literal> class
        affect validation.</para>

        <para>More information is available in <xref
        linkend="Validating XML Documents"/>.</para>

      </section>

      <section id="XSLT Transformations"><title>XSLT Transformations</title>

        <para>The <literal>Transform</literal> class provides an easy
        to use layer on top of JAXP's transformations.  An instance of
        this class is initialized with the source document and a
        stylesheet and the result of the transformation can be
        retrieved as a <literal>String</literal> or DOM
        <literal>Document</literal>.</para>

        <para>The output of <literal>Transform</literal> can be used
        as input to comparisons, validations, XPath tests and so on.
        There is no detailed sections on transformations since they
        are really only a different way to create input for the rest
        of XMLUnit's machinery.  Examples can be found in <xref
        linkend="Comparing XML Transformations"/>.</para>

        <para>It is possible to provide a custom
        <literal>javax.xml.transform.URIResolver</literal> via the
        <literal>XMLUnit.setURIResolver</literal> method.</para>

        <para>You can access the underlying XSLT transformer via
        <literal>XMLUnit.getTransformerFactory</literal>.</para>
      </section>

      <section id="XPath Engine">
        <title>XPath Engine</title>

        <para>The central piece of XMLUnit's XPath support is the
        <literal>XpathEngine</literal> interface.  Currently two
        implementations of the interface exist,
        <literal>SimpleXpathEngine</literal> and
        <literal>org.custommonkey.xmlunit.jaxp13.Jaxp13XpathEngine</literal>.</para>

        <para><literal>SimpleXpathEngine</literal> is a very basic
        implementation that uses your XSLT transformer under the
        covers.  This also means it will expose you to the bugs found
        in your transformer like the transformer claiming a stylesheet
        couldn't be compiled for very basic XPath expressions.  This
        has been reported to be the case for JDK 1.5.</para>

        <para><literal>org.custommonkey.xmlunit.jaxp13.Jaxp13XpathEngine</literal>
        uses JAXP 1.3's <literal>javax.xml.xpath</literal> package and
        seems to work more reliable, stable and performant than
        <literal>SimpleXpathEngine</literal>.</para>

        <para>You use the <literal>XMLUnit.newXpathEngine</literal>
        method to obtain an instance of the
        <literal>XpathEngine</literal>.  As of XMLUnit 1.1 this will
        try to use JAXP 1.3 if it is available and fall back to
        <literal>SimpleXpathEngine</literal>.</para>

        <para>Instances of <literal>XpathEngine</literal> can return
        the results of XPath queries either as DOM
        <literal>NodeList</literal> or plain
        <literal>String</literal>s.</para>

        <para>More information is available in <xref
        linkend="XPath Tests"/>.</para>
      </section>

      <section id="Basic: DOM Tree Walking">
        <title>DOM Tree Walking</title>

        <para>To test pieces of XML by traversing the DOM tree you use
        the <literal>NodeTester</literal> class.  Each DOM
        <literal>Node</literal> will be passed to a
        <literal>NodeTester</literal> implementation you provide.  The
        <literal>AbstractNodeTester</literal> class is provided as a
        NullObject Pattern base class for implementations of your
        own.</para>

        <para>More information is available in <xref
        linkend="DOM Tree Walking"/>.</para>
      </section>
    </section>

    <section id="Using XMLUnit With JUnit 3.x">
      <title>Using XMLUnit With JUnit 3.x</title>

      <para>Initially XMLUnit was tightly coupled to JUnit and the
      recommended approach was to write unit tests by inheriting from
      the <literal>XMLTestCase</literal> class.
      <literal>XMLTestCase</literal> provides a pretty long list of
      <literal>assert...</literal> methods that may simplify your
      interaction with XMLUnit's internals in many common
      cases.</para>

      <para>The <literal>XMLAssert</literal> class provides the same
      set of <literal>assert...</literal>s as static methods.  Use
      <literal>XMLAssert</literal> instead of
      <literal>XMLTestCase</literal> for your unit tests if you can't
      or don't want to inherit from
      <literal>XMLTestCase</literal>.</para>

      <para>All power of XMLUnit is available whether you use
      <literal>XMLTestCase</literal> and/or
      <literal>XMLAssert</literal> or the underlying API directly.  If
      you are using JUnit 3.x then using the specific classes may prove
      to be more convenient.</para>

    </section>

    <section id="Common Configuration Options">
      <title>Common Configuration Options</title>

      <section id="JAXP">
        <title>JAXP</title>

        <para>If you are using a JDK 1.4 or later, your Java class
        library already contains the required XML parsers and XSLT
        transformers.  Still you may want to use a different
        parser/transformer than the one of your JDK - in particular
        since the versions shipping with some JDKs are known to
        contain serious bugs.</para>

        <para>As described in <xref linkend="Configuring XMLUnit"/>
        there are two main approaches to choose the XML parser of XSLT
        transformer: System properties and setters in the
        <literal>XMLUnit</literal> class.</para>

        <para>If you use system properties you have the advantage that
        your choice affects the whole JAXP system, whether it is used
        inside of XMLUnit or not.  If you are using JDK 1.4 or later
        you may also want to review the <ulink
        url="http://java.sun.com/j2se/1.4.2/docs/guide/standards/">Endorsed
        Standards Override Mechanism</ulink> to use a different
        parser/transformer than the one shipping with your JDK.</para>

        <para>The second option - using the <literal>XMLUnit</literal>
        class - allows you to use different parsers for control and
        test documents, it even allows you to use different parsers
        for different test cases, if you really want to stretch it
        that far.  It may also work for JDK 1.4 and above, even if you
        don't override the endorsed standards libraries.</para>

        <para>You can access the underlying JAXP parser by
        <literal>XMLUnit.newControlParser</literal>,
        <literal>XMLUnit.newTestParser</literal>,
        <literal>XMLUnit.getControlDocumentBuilderFactory</literal>,
        <literal>XMLUnit.getTestDocumentBuilderFactory</literal> and
        <literal>XMLUnit.getSAXParserFactory</literal> (used by
        <literal>Validator</literal>).  Note that all these methods
        return factories or parsers that are namespace aware.</para>

        <para>The various <literal>build...</literal> methods in
        <literal>XMLUnit</literal> provide convenience layers for
        building DOM <literal>Document</literal>s using the configured
        parsers.</para>

        <para>You can also set the class name for the
        <literal>XPathFactory</literal> to use when using JAXP 1.3 by
        passing the class name to
        <literal>XMLUnit.setXPathFactory</literal>.</para>

      </section>

      <section id="EntityResolver">
        <title><literal>EntityResolver</literal></title>

        <para>You can provide a custom
        <literal>org.xml.sax.EntityResolver</literal> for the control
        and test parsers via
        <literal>XMLUnit.setControlEntityResolver</literal> and
        <literal>XMLUnit.setTestEntityResolver</literal>.
        <literal>Validator</literal> uses the resolver set via
        <literal>setControlEntityResolver</literal> as well.</para>
      </section>

      <section id="Basic: Element Content Whitespace">
        <title>Element Content Whitespace</title>

        <para>Element content whitespace - also known as ignorable
        whitespace - is whitespace contained in elements whose content
        model doesn't allow text content.  I.e. the newline and space
        characters between <literal><![CDATA[<foo>]]></literal> and
        <literal><![CDATA[<bar>]]></literal> in the following example
        could belong into this category.</para>

        <programlisting language="XML"><![CDATA[
<foo>
  <bar/></foo>
]]></programlisting>

        <para>Using <literal>XMLUnit.setIgnoreWhitespace</literal> it
        is possible to make the test and control parser ignore this
        kind of whitespace.</para>

        <para>Note that setting this property to
        <literal>true</literal> usually doesn't have any effect since
        it only works on validating parsers and XMLUnit doesn't enable
        validation by default.  It does have an effect when comparing
        pieces of XML, though, since the same flag is used for a
        different purpose as well in that case.  See <xref
        linkend="Whitespace Handling"/> for more details.</para>

      </section>

      <section id="XSLT Stylesheet Version">
        <title>XSLT Stylesheet Version</title>

        <para>Some features of XMLUnit use XSLT stylesheets under the
        covers, in particular XSLT will be used to strip element
        content whitespace or comments as well as by
        <literal>SimpleXpathEngine</literal>.  These stylesheets only
        require a XSLT transformer that supports XSLT 1.0 and will say
        so in the <literal>stylesheet</literal> element.</para>

        <para>If your XSLT transformer supports XSLT 2.0 or newer it
        may<footnote><para>The W3C recommendation says it
        SHOULD.</para></footnote> issue a warning for these stylesheets
        which can be annoying.  You can use
        <literal>XMLUnit.setXSLTVersion</literal> to make XMLUnit
        change the version attribute to a different value.  Note that
        XMLUnit hasn't been tested with a value other than
        <literal>"1.0"</literal>.</para>
      </section>
    </section>

    <section id="Providing Input to XMLUnit">
      <title>Providing Input to XMLUnit</title>

      <para>Most methods in XMLUnit that expect a piece of XML as
      input provide several overloads that obtain their input from
      different sources.  The most common options are:</para>

      <itemizedlist>

        <listitem>A DOM <literal>Document</literal>.

          <para>Here you have all control over the document's
          creation.  Such a <literal>Document</literal> could as well
          be the result of an XSLT transformation via the
          <literal>Transform</literal> class.</para>
        </listitem>

        <listitem>A SAX <literal>InputSource</literal>.

          <para>This is the most generic way since
          <literal>InputSource</literal> allows you to read from
          arbitrary <literal>InputStream</literal>s or
          <literal>Reader</literal>s.  Use an
          <literal>InputStream</literal> wrapped by an
          <literal>InputSource</literal> if you want the XML parser to
          pick up the proper encoding from the XML declaration.</para>
        </listitem>

        <listitem>A <literal>String</literal>.

          <para>Here a DOM <literal>Document</literal> is built from
          the input <literal>String</literal> using the JAXP parser
          specified for control or test documents - depending on
          whether the input is a control or test piece of XML.</para>

          <para>Note that using a <literal>String</literal> assumes
          that your XML has already been converted from its XML
          encoding to a Java <literal>String</literal> upfront.</para>
        </listitem>

        <listitem>A <literal>Reader</literal>.

          <para>Here a DOM <literal>Document</literal> is built from
          the input <literal>Reader</literal> using the JAXP parser
          specified for control or test documents - depending on
          whether the input is a control or test piece of XML.</para>
          
          <para>Note that using a <literal>Reader</literal> is a
          bad choice if your XML encoding is different from your
          platform's default encoding since Java's IO system won't
          read your XML declaration.  It is a good practice to use one
          of the other overloads rather than the
          <literal>Reader</literal> version to ensure encoding has
          been dealt with properly.</para>
        </listitem>
      </itemizedlist>
    </section>

  </section>

  <section id="Comparing Pieces of XML">
    <title>Comparing Pieces of XML</title>

    <section id="The Difference Engine">
      <title>The Difference Engine</title>

      <para>At the center of XMLUnit's support for comparisons is the
      <literal>DifferenceEngine</literal> class.  In practice you
      rarely deal with it directly but rather use it via instances of
      <literal>Diff</literal> or <literal>DetailedDiff</literal>
      classes (see <xref linkend="Diff"/>).</para>

      <para>The <literal>DifferenceEngine</literal> walks two trees of
      DOM <literal>Node</literal>s, the control and the test tree, and
      compares the nodes.  Whenever it detects a difference, it sends
      a message to a configured <literal>DifferenceListener</literal>
      (see <xref linkend="DifferenceListener"/>) and asks a
      <literal>ComparisonController</literal> (see <xref
      linkend="ComparisonController"/>) whether the current comparison
      should be halted.</para>

      <para>In some cases the order of elements in two pieces of XML
      may not be significant.  If this is true, the
      <literal>DifferenceEngine</literal> needs help to determine
      which <literal>Element</literal>s to compare.  This is the job
      of an <literal>ElementQualifier</literal> (see <xref
      linkend="ElementQualifier"/>).</para>

      <para>The types of differences
      <literal>DifferenceEngine</literal> can detect are enumerated in
      the <literal>DifferenceConstants</literal> interface and
      represented by instances of the <literal>Difference</literal>
      class.</para>

      <para>A <literal>Difference</literal> can be recoverable;
      recoverable <literal>Difference</literal>s make the
      <literal>Diff</literal> class consider two pieces of XML similar
      while non-recoverable <literal>Difference</literal>s render the
      two pieces different.</para>

      <para>The types of <literal>Difference</literal>s that are
      currently detected are listed in <xref linkend="docleveldiff"/>
      to <xref linkend="otherdiff"/> (the first two columns refer to
      the <literal>DifferenceConstants</literal> class).</para>

      <table frame="all" rules="all" pgwide="1" id="docleveldiff">
        <title>Document level <literal>Difference</literal>s detected by
        <literal>DifferenceEngine</literal></title>
        <tgroup cols="4">
          <colspec colname="id" align="center"/>
          <colspec colname="constant" align="center"/>
          <colspec colname="recoverable" align="center"/>
          <colspec colname="description" align="left"/>

          <thead>
            <row>
              <entry><literal>ID</literal></entry>
              <entry><literal>Constant</literal></entry>
              <entry><literal>recoverable</literal></entry>
              <entry align="center">Description</entry>
            </row>
          </thead>

          <tbody>
            <row>
              <entry><literal>HAS_DOCTYPE_DECLARATION_ID</literal></entry>
              <entry><literal>HAS_DOCTYPE_DECLARATION</literal></entry>
              <entry><literal>true</literal></entry>
              <entry>One piece of XML has a DOCTYPE declaration while
                the other one has not.</entry>
            </row>
            <row>
              <entry><literal>DOCTYPE_NAME_ID</literal></entry>
              <entry><literal>DOCTYPE_NAME</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>Both pieces of XML contain a DOCTYPE declaration
                but the declarations specify different names for the
                root element.</entry>
            </row>
            <row>
              <entry><literal>DOCTYPE_PUBLIC_ID_ID</literal></entry>
              <entry><literal>DOCTYPE_PUBLIC_ID</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>Both pieces of XML contain a DOCTYPE declaration
                but the declarations specify different PUBLIC
                identifiers.</entry>
            </row>
            <row>
              <entry><literal>DOCTYPE_SYSTEM_ID_ID</literal></entry>
              <entry><literal>DOCTYPE_SYSTEM_ID</literal></entry>
              <entry><literal>true</literal></entry>
              <entry>Both pieces of XML contain a DOCTYPE declaration
                but the declarations specify different SYSTEM
                identifiers.</entry>
            </row>
            <row>
              <entry><literal>NODE_TYPE_ID</literal></entry>
              <entry><literal>NODE_TYPE</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The test piece of XML contains a different type
              of node than was expected.  This type of difference will
              also occur if either the root control or test
              <literal>Node</literal> is <literal>null</literal> while
              the other is not.</entry>
            </row>
            <row>
              <entry><literal>NAMESPACE_PREFIX_ID</literal></entry>
              <entry><literal>NAMESPACE_PREFIX</literal></entry>
              <entry><literal>true</literal></entry>
              <entry>Two nodes use different prefixes for the same
                XML Namespace URI in the two pieces of XML.</entry>
            </row>
            <row>
              <entry><literal>NAMESPACE_URI_ID</literal></entry>
              <entry><literal>NAMESPACE_URI</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>Two nodes in the two pieces of XML share the same
              local name but use different XML Namespace URIs.</entry>
            </row>
            <row>
              <entry><literal>SCHEMA_LOCATION_ID</literal></entry>
              <entry><literal>SCHEMA_LOCATION</literal></entry>
              <entry><literal>true</literal></entry>
              <entry>Two nodes have different values for the
              <literal>schemaLocation</literal> attribute of the
              XMLSchema-Instance namespace.  The attribute could be
              present on only one of the two nodes.</entry>
            </row>
            <row>
              <entry><literal>NO_NAMESPACE_SCHEMA_LOCATION_ID</literal></entry>
              <entry><literal>NO_NAMESPACE_SCHEMA_LOCATION</literal></entry>
              <entry><literal>true</literal></entry>
              <entry>Two nodes have different values for the
              <literal>noNamespaceSchemaLocation</literal> attribute
              of the XMLSchema-Instance namespace.  The attribute
              could be present on only one of the two nodes.</entry>
            </row>
          </tbody>
        </tgroup>
      </table>

      <table frame="all" rules="all" pgwide="1" id="elementleveldiff">
        <title>Element level <literal>Difference</literal>s detected by
        <literal>DifferenceEngine</literal></title>
        <tgroup cols="4">
          <colspec colname="id" align="center"/>
          <colspec colname="constant" align="center"/>
          <colspec colname="recoverable" align="center"/>
          <colspec colname="description" align="left"/>

          <thead>
            <row>
              <entry><literal>ID</literal></entry>
              <entry><literal>Constant</literal></entry>
              <entry><literal>recoverable</literal></entry>
              <entry align="center">Description</entry>
            </row>
          </thead>

          <tbody>
            <row>
              <entry><literal>ELEMENT_TAG_NAME_ID</literal></entry>
              <entry><literal>ELEMENT_TAG_NAME</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The two pieces of XML contain elements with
                different tag names.</entry>
            </row>
            <row>
              <entry><literal>ELEMENT_NUM_ATTRIBUTES_ID</literal></entry>
              <entry><literal>ELEMENT_NUM_ATTRIBUTES</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The two pieces of XML contain a common element,
                but the number of attributes on the element is
                different.</entry>
            </row>
            <row>
              <entry><literal>HAS_CHILD_NODES_ID</literal></entry>
              <entry><literal>HAS_CHILD_NODES</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>An element in one piece of XML has child nodes
                while the corresponding one in the other has not.</entry>
            </row>
            <row>
              <entry><literal>CHILD_NODELIST_LENGTH_ID</literal></entry>
              <entry><literal>CHILD_NODELIST_LENGTH</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>Two elements in the two pieces of XML differ by
                their number of child nodes.</entry>
            </row>
            <row>
              <entry><literal>CHILD_NODELIST_SEQUENCE_ID</literal></entry>
              <entry><literal>CHILD_NODELIST_SEQUENCE</literal></entry>
              <entry><literal>true</literal></entry>
              <entry>Two elements in the two pieces of XML contain the
                same child nodes but in a different order.</entry>
            </row>
            <row>
              <entry><literal>CHILD_NODE_NOT_FOUND_ID</literal></entry>
              <entry><literal>CHILD_NODE_NOT_FOUND</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>A child node in one piece of XML couldn't be
                matched against any other node of the other piece.</entry>
            </row>
            <row>
              <entry><literal>ATTR_SEQUENCE_ID</literal></entry>
              <entry><literal>ATTR_SEQUENCE</literal></entry>
              <entry><literal>true</literal></entry>
              <entry>The attributes on an element appear in different
              order<footnote><para>Note that the order of attributes
              is not significant in XML, different parsers may return
              attributes in a different order even if parsing the same
              XML document.  There is an option to turn this check off
              - see <xref linkend="Comparing: Configuration"/> - but it is on
              by default for backwards compatibility
              reasons</para></footnote> in the two pieces of
              XML.</entry>
            </row>
          </tbody>
        </tgroup>
      </table>

      <table frame="all" rules="all" pgwide="1" id="attributeleveldiff">
        <title>Attribute level <literal>Difference</literal>s detected by
        <literal>DifferenceEngine</literal></title>

        <tgroup cols="4">
          <colspec colname="id" align="center"/>
          <colspec colname="constant" align="center"/>
          <colspec colname="recoverable" align="center"/>
          <colspec colname="description" align="left"/>

          <thead>
            <row>
              <entry><literal>ID</literal></entry>
              <entry><literal>Constant</literal></entry>
              <entry><literal>recoverable</literal></entry>
              <entry align="center">Description</entry>
            </row>
          </thead>

          <tbody>
            <row>
              <entry><literal>ATTR_VALUE_EXPLICITLY_SPECIFIED_ID</literal></entry>
              <entry><literal>ATTR_VALUE_EXPLICITLY_SPECIFIED</literal></entry>
              <entry><literal>true</literal></entry>
              <entry>An attribute that has a default value according
              to the content model of the element in question has been
              specified explicitly in one piece of XML but not in the
              other.<footnote><para>In order for this difference to be
              detected the parser must have been in validating mode
              when the piece of XML was parsed and the DTD or XML
              Schema must have been available.</para></footnote></entry>

            </row>
            <row>
              <entry><literal>ATTR_NAME_NOT_FOUND_ID</literal></entry>
              <entry><literal>ATTR_NAME_NOT_FOUND</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>One piece of XML contains an attribute on an
                element that is missing in the other.</entry>
            </row>
            <row>
              <entry><literal>ATTR_VALUE_ID</literal></entry>
              <entry><literal>ATTR_VALUE</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The value of an element's attribute is different
                in the two pieces of XML.</entry>
            </row>
          </tbody>
        </tgroup>
      </table>

      <table frame="all" rules="all" pgwide="1" id="otherdiff">
        <title>Other <literal>Difference</literal>s detected by
        <literal>DifferenceEngine</literal></title>

        <tgroup cols="4">
          <colspec colname="id" align="center"/>
          <colspec colname="constant" align="center"/>
          <colspec colname="recoverable" align="center"/>
          <colspec colname="description" align="left"/>

          <thead>
            <row>
              <entry><literal>ID</literal></entry>
              <entry><literal>Constant</literal></entry>
              <entry><literal>recoverable</literal></entry>
              <entry align="center">Description</entry>
            </row>
          </thead>

          <tbody>
            <row>
              <entry><literal>COMMENT_VALUE_ID</literal></entry>
              <entry><literal>COMMENT_VALUE</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The content of two comments is different in the
                two pieces of XML.</entry>
            </row>
            <row>
              <entry><literal>PROCESSING_INSTRUCTION_TARGET_ID</literal></entry>
              <entry><literal>PROCESSING_INSTRUCTION_TARGET</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The target of two processing instructions is
                different in the two pieces of XML.</entry>
            </row>
            <row>
              <entry><literal>PROCESSING_INSTRUCTION_DATA_ID</literal></entry>
              <entry><literal>PROCESSING_INSTRUCTION_DATA</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The data of two processing instructions is
                different in the two pieces of XML.</entry>
            </row>

            <row>
              <entry><literal>CDATA_VALUE_ID</literal></entry>
              <entry><literal>CDATA_VALUE</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The content of two CDATA sections is different in
                the two pieces of XML.</entry>
            </row>
            <row>
              <entry><literal>TEXT_VALUE_ID</literal></entry>
              <entry><literal>TEXT_VALUE</literal></entry>
              <entry><literal>false</literal></entry>
              <entry>The value of two texts is different in the two
                pieces of XML.</entry>
            </row>
          </tbody>
        </tgroup>
      </table>

      <para>Note that some of the differences listed may be ignored by
      the <literal>DifferenceEngine</literal> if certain configuration
      options have been specified.  See <xref
      linkend="Comparing: Configuration"/> for details.</para>

      <para><literal>DifferenceEngine</literal> passes differences
      found around as instances of the <literal>Difference</literal>
      class.  In addition to the type of of difference this class also
      holds information on the nodes that have been found to be
      different.  The nodes are described by
      <literal>NodeDetail</literal> instances that encapsulate the DOM
      <literal>Node</literal> instance as well as the XPath expression
      that locates the <literal>Node</literal> inside the given piece
      of XML.  <literal>NodeDetail</literal> also contains a "value"
      that provides more information on the actual values that have
      been found to be different, the concrete interpretation depends
      on the type of difference as can be seen in <xref
      linkend="diffvalue"/>.</para>

      <table frame="all" rules="all" pgwide="0" id="diffvalue">
        <title>Contents of <literal>NodeDetail.getValue()</literal>
        for <literal>Difference</literal>s</title>

        <tgroup cols="2">
          <colspec colname="id" align="center"/>
          <colspec colname="value" align="left"/>

          <thead>
            <row>
              <entry><literal>Difference.getId()</literal></entry>
              <entry align="center"><literal>NodeDetail.getValue()</literal></entry>
            </row>
          </thead>

          <tbody>
            <row>
              <entry><literal>HAS_DOCTYPE_DECLARATION_ID</literal></entry>
              <entry><literal>"not null"</literal> if the document has
              a DOCTYPE declaration, <literal>"null"</literal>
              otherwise.</entry>
            </row>
            <row>
              <entry><literal>DOCTYPE_NAME_ID</literal></entry>
              <entry>The name of the root element.</entry>
            </row>
            <row>
              <entry><literal>DOCTYPE_PUBLIC_ID</literal></entry>
              <entry>The PUBLIC identifier.</entry>
            </row>
            <row>
              <entry><literal>DOCTYPE_SYSTEM_ID</literal></entry>
              <entry>The SYSTEM identifier.</entry>
            </row>
            <row>
              <entry><literal>NODE_TYPE_ID</literal></entry>
              <entry>If one node was absent: <literal>"not
              null"</literal> if the node exists,
              <literal>"null"</literal> otherwise.  If the node types
              differ the value will be a string-ified version of
              <literal>org.w3c.dom.Node.getNodeType()</literal>.</entry>
            </row>
            <row>
              <entry><literal>NAMESPACE_PREFIX_ID</literal></entry>
              <entry>The Namespace prefix.</entry>
            </row>
            <row>
              <entry><literal>NAMESPACE_URI_ID</literal></entry>
              <entry>The Namespace URI.</entry>
            </row>
            <row>
              <entry><literal>SCHEMA_LOCATION_ID</literal></entry>
              <entry>The attribute's value or "[attribute absent]" if
              it has not been specified.</entry>
            </row>
            <row>
              <entry><literal>NO_NAMESPACE_SCHEMA_LOCATION_ID</literal></entry>
              <entry>The attribute's value or "[attribute absent]" if
              it has not been specified.</entry>
            </row>
            <row>
              <entry><literal>ELEMENT_TAG_NAME_ID</literal></entry>
              <entry>The tag name with any Namespace information
              stripped.</entry>
            </row>
            <row>
              <entry><literal>ELEMENT_NUM_ATTRIBUTES_ID</literal></entry>
              <entry>The number of attributes present turned into a
              <literal>String</literal>.</entry>
            </row>
            <row>
              <entry><literal>HAS_CHILD_NODES_ID</literal></entry>
              <entry><literal>"true"</literal> if the element has
              child nodes, <literal>"false"</literal>
              otherwise.</entry>
            </row>
            <row>
              <entry><literal>CHILD_NODELIST_LENGTH_ID</literal></entry>
              <entry>The number of child nodes present turned into a
              <literal>String</literal>.</entry>
            </row>
            <row>
              <entry><literal>CHILD_NODELIST_SEQUENCE_ID</literal></entry>
              <entry>The sequence number of this child node turned into a
              <literal>String</literal>.</entry>
            </row>
            <row>
              <entry><literal>CHILD_NODE_NOT_FOUND_ID</literal></entry>
              <entry>The name of the unmatched node or
              <literal>"null"</literal>.  If the node is an element
              inside an XML namespace the name will be
              Java5-<literal>QName</literal>-like
              <literal>{NS-URI}LOCAL-NAME</literal> - in all other
              cases it is the node's local name.</entry>
            </row>
            <row>
              <entry><literal>ATTR_SEQUENCE_ID</literal></entry>
              <entry>The attribute's name.</entry>
            </row>
            <row>
              <entry><literal>ATTR_VALUE_EXPLICITLY_SPECIFIED_ID</literal></entry>
              <entry><literal>"true"</literal> if the attribute has
              been specified, <literal>"false"</literal>
              otherwise.</entry>
            </row>
            <row>
              <entry><literal>ATTR_NAME_NOT_FOUND_ID</literal></entry>
              <entry>The attribute's name or
              <literal>"null"</literal>.  If the attribute belongs to
              an XML namespace the name will be
              Java5-<literal>QName</literal>-like
              <literal>{NS-URI}LOCAL-NAME</literal> - in all other
              cases it is the attribute's local name.</entry>
            </row>
            <row>
              <entry><literal>ATTR_VALUE_ID</literal></entry>
              <entry>The attribute's value.</entry>
            </row>
            <row>
              <entry><literal>COMMENT_VALUE_ID</literal></entry>
              <entry>The actual comment.</entry>
            </row>
            <row>
              <entry><literal>PROCESSING_INSTRUCTION_TARGET_ID</literal></entry>
              <entry>The processing instruction's target.</entry>
            </row>
            <row>
              <entry><literal>PROCESSING_INSTRUCTION_DATA_ID</literal></entry>
              <entry>The processing instruction's data.</entry>
            </row>

            <row>
              <entry><literal>CDATA_VALUE_ID</literal></entry>
              <entry>The content of the CDATA section.</entry>
            </row>
            <row>
              <entry><literal>TEXT_VALUE_ID</literal></entry>
              <entry>The actual text.</entry>
            </row>
          </tbody>
        </tgroup>
      </table>

      <para>As said in the first paragraph you won't deal with
      <literal>DifferenceEngine</literal> directly in most cases.  In
      cases where <literal>Diff</literal> or
      <literal>DetailedDiff</literal> don't provide what you need
      you'd create an instance of <literal>DifferenceEngine</literal>
      passing a <literal>ComparisonController</literal> in the
      constructor and invoke <literal>compare</literal> with your DOM
      trees to compare as well as a
      <literal>DifferenceListener</literal> and
      <literal>ElementQualifier</literal>.  The listener will be
      called on any differences while the <literal>control</literal>
      method is executing.</para>

      <example>
        <title>Using <literal>DifferenceEngine</literal>
          Directly</title>
        <programlisting language="Java"><![CDATA[
class MyDifferenceListener implements DifferenceListener {
    private boolean calledFlag = false;
    public boolean called() { return calledFlag; }

    public int differenceFound(Difference difference) {
        calledFlag = true;
        return RETURN_ACCEPT_DIFFERENCE;
    }

    public void skippedComparison(Node control, Node test) {
    }
}

DifferenceEngine engine = new DifferenceEngine(myComparisonController);
MyDifferenceListener listener = new MyDifferenceListener();
engine.compare(controlNode, testNode, listener,
               myElementQualifier);
System.err.println("There have been "
                   + (listener.called() ? "" : "no ")
                   + "differences.");
]]></programlisting></example>
    </section>

    <section id="ComparisonController">
      <title><literal>ComparisonController</literal></title>

      <para>The <literal>ComparisonController</literal>'s job is to
      decide whether a comparison should be halted after a difference
      has been found.  Its interface is:</para>

      <programlisting language="Java"><![CDATA[
    /**
     * Determine whether a Difference that the listener has been notified of
     *  should halt further XML comparison. Default behaviour for a Diff
     *  instance is to halt if the Difference is not recoverable.
     * @see Difference#isRecoverable
     * @param afterDifference the last Difference passed to <code>differenceFound</code>
     * @return true to halt further comparison, false otherwise
     */
    boolean haltComparison(Difference afterDifference);
]]></programlisting>

      <para>Whenever a difference has been detected by the
      <literal>DifferenceEngine</literal> the
      <literal>haltComparison</literal> method will be called
      immediately after the <literal>DifferenceListener</literal> has
      been informed of the difference.  This is true no matter what
      type of <literal>Difference</literal> has been found or which
      value the <literal>DifferenceListener</literal> has
      returned.</para>

      <para>The only implementations of
      <literal>ComparisonController</literal> that ship with XMLUnit
      are <literal>Diff</literal> and <literal>DetailedDiff</literal>,
      see <xref linkend="Diff"/> for details about them.</para>

      <para>A <literal>ComparisonController</literal> that halted the
      comparison on any non-recoverable difference could be
      implemented as:</para>

      <example>
        <title>A Simple
          <literal>ComparisonController</literal></title>
        <programlisting language="Java"><![CDATA[
public class HaltOnNonRecoverable implements ComparisonController {
    public boolean haltComparison(Difference afterDifference) {
        return !afterDifference.isRecoverable();
    }
}
]]></programlisting></example>
    </section>

    <section id="DifferenceListener">
      <title><literal>DifferenceListener</literal></title>

      <para><literal>DifferenceListener</literal> contains two
      callback methods that are invoked by the
      <literal>DifferenceEngine</literal> when differences are
      detected:</para>

      <programlisting language="Java"><![CDATA[
    /**
     * Receive notification that 2 nodes are different.
     * @param difference a Difference instance as defined in {@link
     * DifferenceConstants DifferenceConstants} describing the cause
     * of the difference and containing the detail of the nodes that
     * differ
     * @return int one of the RETURN_... constants describing how this
     * difference was interpreted
     */
    int differenceFound(Difference difference);

    /**
     * Receive notification that a comparison between 2 nodes has been skipped
     *  because the node types are not comparable by the DifferenceEngine
     * @param control the control node being compared
     * @param test the test node being compared
     * @see DifferenceEngine
     */
    void skippedComparison(Node control, Node test);
]]></programlisting>

      <para><literal>differenceFound</literal> is invoked by
      <literal>DifferenceEngine</literal> as soon as a difference has
      been detected.  The return value of that method is completely
      ignored by <literal>DifferenceEngine</literal>, it becomes
      important when used together with <literal>Diff</literal>,
      though (see <xref linkend="Diff"/>).  The return value should be
      one of the four constants defined in the the
      <literal>DifferenceListener</literal> interface:</para>

      <programlisting language="Java"><![CDATA[
    /** 
     * Standard return value for the <code>differenceFound</code> method.
     * Indicates that the <code>Difference</code> is interpreted as defined 
     * in {@link DifferenceConstants DifferenceConstants}.
     */
    int RETURN_ACCEPT_DIFFERENCE;
    /** 
     * Override return value for the <code>differenceFound</code> method.
     * Indicates that the nodes identified as being different should be 
     * interpreted as being identical.
     */
    int RETURN_IGNORE_DIFFERENCE_NODES_IDENTICAL;
    /** 
     * Override return value for the <code>differenceFound</code> method.
     * Indicates that the nodes identified as being different should be 
     * interpreted as being similar.
     */
    int RETURN_IGNORE_DIFFERENCE_NODES_SIMILAR;
    /** 
     * Override return value for the <code>differenceFound</code> method.
     * Indicates that the nodes identified as being similar should be 
     * interpreted as being different.
     */
    int RETURN_UPGRADE_DIFFERENCE_NODES_DIFFERENT = 3;
]]></programlisting>

      <para>The <literal>skippedComparison</literal> method is
      invoked if the <literal>DifferenceEngine</literal> encounters
      two <literal>Node</literal>s it cannot compare.  Before invoking
      <literal>skippedComparison</literal>
      <literal>DifferenceEngine</literal> will have invoked
      <literal>differenceFound</literal> with a
      <literal>Difference</literal> of type
      <literal>NODE_TYPE</literal>.</para>

      <para>A custom <literal>DifferenceListener</literal> that
      ignored any DOCTYPE related differences could be written
      as:</para>

      <example>
        <title>A <literal>DifferenceListener</literal> that Ignores
          DOCTYPE Differences</title>
        <programlisting language="Java"><![CDATA[
public class IgnoreDoctype implements DifferenceListener {
    private static final int[] IGNORE = new int[] {
        DifferenceConstants.HAS_DOCTYPE_DECLARATION_ID,
        DifferenceConstants.DOCTYPE_NAME_ID,
        DifferenceConstants.DOCTYPE_PUBLIC_ID_ID,
        DifferenceConstants.DOCTYPE_SYSTEM_ID_ID
    };

    static {
        Arrays.sort(IGNORE);
    }

    public int differenceFound(Difference difference) {
        return Arrays.binarySearch(IGNORE, difference.getId()) >= 0
            ? RETURN_IGNORE_DIFFERENCE_NODES_IDENTICAL
            : RETURN_ACCEPT_DIFFERENCE;
    }
    
    public void skippedComparison(Node control, Node test) {
    }
}
]]></programlisting></example>

      <para>Apart from <literal>Diff</literal> and
      <literal>DetailedDiff</literal> XMLUnit ships with an additional
      implementation of <literal>DifferenceListener</literal>.</para>

      <section id="IgnoreTextAndAttributeValuesDifferenceListener">
        <title><literal>IgnoreTextAndAttributeValuesDifferenceListener</literal></title>

        <para><literal>IgnoreTextAndAttributeValuesDifferenceListener</literal>
        doesn't do anything in <literal>skippedComparison</literal>.
        It "downgrades" <literal>Difference</literal>s of type
        <literal>ATTR_VALUE</literal>,
        <literal>ATTR_VALUE_EXPLICITLY_SPECIFIED</literal> and
        <literal>TEXT_VALUE</literal> to recoverable
        differences.</para>

        <para>This means if instances of
        <literal>IgnoreTextAndAttributeValuesDifferenceListener</literal>
        are used together with <literal>Diff</literal> then two pieces
        of XML will be considered similar if they have the same basic
        structure.  They are not considered identical, though.</para>

        <para>Note that the list of ignored differences doesn't cover
        all textual differences.  You should configure XMLUnit to
        ignore comments and whitespace and to consider CDATA sections
        and text nodes to be the same (see <xref
        linkend="Comparing: Configuration"/>) in order to cover
        <literal>COMMENT_VALUE</literal> and
        <literal>CDATA_VALUE</literal> as well.</para>

      </section>
    </section>

    <section id="ElementQualifier">
      <title><literal>ElementQualifier</literal></title>

      <para>When <literal>DifferenceEngine</literal> encounters a list
      of DOM <literal>Element</literal>s as children of another
      <literal>Element</literal> it will ask the configured
      <literal>ElementQualifier</literal> which
      <literal>Element</literal> of the control piece of XML should be
      compared to which of the test piece.  Its contract is:</para>

      <programlisting language="Java"><![CDATA[
    /**
     * Determine whether two elements are comparable
     * @param control an Element from the control XML NodeList
     * @param test an Element from the test XML NodeList
     * @return true if the elements are comparable, false otherwise
     */
    boolean qualifyForComparison(Element control, Element test); 
]]></programlisting>

      <para>For any given <literal>Element</literal> in the control
      piece of XML <literal>DifferenceEngine</literal> will cycle
      through the corresponding list of <literal>Element</literal>s in
      the test piece of XML until
      <literal>qualifyForComparison</literal> has returned
      <literal>true</literal> or the test document is
      exhausted.</para>

      <para>When using <literal>DifferenceEngine</literal> or
      <literal>Diff</literal> it is completely legal to set the
      <literal>ElementQualifier</literal> to <literal>null</literal>.
      In this case any kind of <literal>Node</literal> is compared to
      the test <literal>Node</literal> that appears at the same
      position in the sequence.</para>

      <example id="eq-nodelist-example">
        <title>Example Nodes for <literal>ElementQualifier</literal>
          (the comments are not part of the example)</title>
        <programlisting language="XML"><![CDATA[
<!-- control piece of XML -->
<parent>
  <child1/>                        <!-- control node 1 -->
  <child2/>                        <!-- control node 2 -->
  <child2 foo="bar">xyzzy</child2> <!-- control node 3 -->
  <child2 foo="baz"/>              <!-- control node 4 -->
</parent>

<!-- test piece of XML -->
<parent>
  <child2 foo="baz"/>              <!-- test node 1 -->
  <child1/>                        <!-- test node 2 -->
  <child2>xyzzy</child2>           <!-- test node 3 -->
  <child2 foo="bar"/>              <!-- test node 4 -->
</parent>
]]></programlisting></example>

      <para>Taking <xref linkend="eq-nodelist-example"/> without any
      <literal>ElementQualifier</literal>
      <literal>DifferenceEngine</literal> will compare control node
      <literal>n</literal> to test node <literal>n</literal> for
      <literal>n</literal> between 1 and 4.  In many cases this is
      exactly what is desired, but sometimes
      <literal><![CDATA[<a><b/><c/></a>]]></literal> should be similar
      to <literal><![CDATA[<a><c/><b/></a>]]></literal> because the
      order of elements doesn't matter - this is when you'd use a
      different <literal>ElementQualifier</literal>.  XMLUnit ships
      with several implementations.</para>

      <section id="ElementNameQualifier">
        <title><literal>ElementNameQualifier</literal></title>

        <para>Only <literal>Element</literal>s with the same name -
        and Namespace URI if present - qualify.</para>

        <para>In <xref linkend="eq-nodelist-example"/> this means
        control node 1 will be compared to test node 2.  Then control
        node 2 will be compared to test node 3 because
        <literal>DifferenceEngine</literal> will start to search for
        the matching test <literal>Element</literal> at the second
        test node, the same sequence number the control node is at.
        Control node 3 is compared to test node 3 as well and control
        node 4 to test node 4.</para>
      </section>

      <section id="ElementNameAndAttributeQualifier">
        <title><literal>ElementNameAndAttributeQualifier</literal></title>

        <para>Only <literal>Element</literal>s with the same name -
        and Namespace URI if present - as well as the same values for
        all attributes given in
        <literal>ElementNameAndAttributeQualifier</literal>'s
        constructor qualify.</para>

        <para>Let's say <literal>"foo"</literal> has been passed to
        <literal>ElementNameAndAttributeQualifier</literal>'s
        constructor when looking at <xref
        linkend="eq-nodelist-example"/>.  This again means control
        node 1 will be compared to test node 2 since they do have the
        same name and no value at all for attribute
        <literal>"foo"</literal>.  Then control node 2 will be
        compared to test node 3 - again, no value for
        <literal>"foo"</literal>.  Control node 3 is compared to test
        node 4 as they have the same value <literal>"bar"</literal>.
        Finally control node 4 is compared to test node 1; here
        <literal>DifferenceEngine</literal> searches from the
        beginning of the test node list after test node 4 didn't
        match.</para>

        <para>There are three constructors in
        <literal>ElementNameAndAttributeQualifier</literal>.  The
        no-arg constructor creates an instance that compares all
        attributes while the others will compare a single attribute or
        a given subset of all attributes.</para>
      </section>

      <section id="ElementNameAndTextQualifier">
        <title><literal>ElementNameAndTextQualifier</literal></title>

        <para>Only <literal>Element</literal>s with the same name -
        and Namespace URI if present - as well as the same text
        content nested into them qualify.</para>

        <para>In <xref linkend="eq-nodelist-example"/> this means
        control node 1 will be compared to test node 2 since they both
        don't have any nested text at all.  Then control node 2 will
        be compared to test node 4.  Control node 3 is compared to
        test node 3 since they have the same nested text and control
        node 4 to test node 4.</para>
      </section>

      <section id="RecursiveElementNameAndTextQualifier">
        <title><literal>org.custommonkey.xmlunit.examples.RecursiveElementNameAndTextQualifier</literal></title>

        <para>All <literal>ElementQualifier</literal>s seen so far
        only looked at the <literal>Element</literal>s themselves and
        not at the structure nested into them at a deeper level.  A
        frequent user question has been which
        <literal>ElementQualifier</literal> should be used if the
        pieces of XML in <xref linkend="htmltable"/> should be
        considered similar.</para>

        <example id="htmltable">
          <title>Example for
          <literal>RecursiveElementNameAndTextQualifier</literal>
          (the comments are not part of the example)</title>
          <programlisting language="HTML"><![CDATA[
<!-- control -->
<table>
  <tr>            <!-- control row 1 -->
    <td>foo</td>
  </tr>
  <tr>            <!-- control row 2 -->
    <td>bar</td>
  </tr>
</table>

<!-- test -->
<table>
  <tr>            <!-- test row 1 -->
    <td>bar</td>
  </tr>
  <tr>            <!-- test row 2 -->
    <td>foo</td>
  </tr>
</table>
]]></programlisting></example>

        <para>At first glance
        <literal>ElementNameAndTextQualifier</literal> should work but
        it doesn't.  When <literal>DifferenceEngine</literal>
        processed the children of <literal>table</literal> it would
        compare control row 1 to test row 1 since both
        <literal>tr</literal> elements have the same name and both
        have no textual content at all.</para>

        <para>What is needed in this case is an
        <literal>ElementQualifier</literal> that looks at the element's
        name, as well as the name of the first child element and the
        text nested into that first child element.  This is what
        <literal>RecursiveElementNameAndTextQualifier</literal>
        does.</para>

        <para><literal>RecursiveElementNameAndTextQualifier</literal>
        ignores whitespace between the elements leading up to the
        nested text.</para>
      </section>

      <section id="MultiLevelElementNameAndTextQualifier">
        <title><literal>org.custommonkey.xmlunit.examples.MultiLevelElementNameAndTextQualifier</literal></title>

        <para>
          <literal>MultiLevelElementNameAndTextQualifier</literal> has
          in a way been the predecessor
          of <xref linkend="RecursiveElementNameAndTextQualifier"><literal>RecursiveElementNameAndTextQualifier</literal></xref>.
          It also matches element names and those of nested child
          elements until it finds matches, but
          unlike <literal>RecursiveElementNameAndTextQualifier</literal>,
          you must
          tell <literal>MultiLevelElementNameAndTextQualifier</literal>
          at which nesting level it should expect the nested text.
        </para>

        <para>
          <literal>MultiLevelElementNameAndTextQualifier</literal>'s
          constructor expects a single argument which is the nesting
          level of the expected text.  If you use an argument of 1,
          <literal>MultiLevelElementNameAndTextQualifier</literal> is
          identical to <literal>ElementNameAndTextQualifier</literal>.
          In <xref linkend="htmltable"/> a value of 2 would be
          needed.</para>

        <para>By default
          <literal>MultiLevelElementNameAndTextQualifier</literal>
          will not ignore whitespace between the elements leading up
          to the nested text.  If your piece of XML contains this sort
          of whitespace (like <xref linkend="htmltable"/> which
          contains a newline and several space characters between
          <literal>&lt;tr&gt;</literal> and
          <literal>&lt;td&gt;</literal>) you can either instruct
          XMLUnit to ignore whitespace completely (see
          <xref linkend="Whitespace Handling"/>) or use the two-arg
          constructor of
          <literal>MultiLevelElementNameAndTextQualifier</literal>
          introduced with XMLUnit 1.2 and set the
          <literal>ignoreEmptyTexts</literal> argument to
          true.</para>

        <para>In
          general <literal>RecursiveElementNameAndTextQualifier</literal>
          requires less knowledge upfront and its whitespace-handling
          is more intuitive.</para>
      </section>

    </section>

    <section id="Diff">
      <title><literal>Diff</literal> and
      <literal>DetailedDiff</literal></title>

      <para><literal>Diff</literal> and
      <literal>DetailedDiff</literal> provide simplified access to
      <literal>DifferenceEngine</literal> by implementing the
      <literal>ComparisonController</literal> and
      <literal>DifferenceListener</literal> interfaces themselves.
      They cover the two most common use cases for comparing two
      pieces of XML: checking whether the pieces are different (this
      is what <literal>Diff</literal> does) and finding all
      differences between them (this is what
      <literal>DetailedDiff</literal> does).</para>

      <para><literal>DetailedDiff</literal> is a subclass of
      <literal>Diff</literal> and can only be constructed by creating
      a <literal>Diff</literal> instance first.</para>

      <para>The major difference between them is their implementation
      of the <literal>ComparisonController</literal> interface:
      <literal>DetailedDiff</literal> will never stop the comparison
      since it wants to collect all differences.
      <literal>Diff</literal> in turn will halt the comparison as soon
      as the first <literal>Difference</literal> is found that is not
      recoverable.  In addition <literal>DetailedDiff</literal>
      collects all <literal>Difference</literal>s in a list and
      provides access to it.</para>

      <para>By default <literal>Diff</literal> will consider two
      pieces of XML as identical if no differences have been found at
      all, similar if all differences that have been found have been
      recoverable (see <xref linkend="docleveldiff"/> to <xref
      linkend="otherdiff"/>) and different as soon as any
      non-recoverable difference has been found.</para>

      <para>It is possible to specify a
      <literal>DifferenceListener</literal> to <literal>Diff</literal>
      using the <literal>overrideDifferenceListener</literal> method.
      In this case each <literal>Difference</literal> will be
      evaluated by the passed in
      <literal>DifferenceListener</literal>.  By returning
      <literal>RETURN_IGNORE_DIFFERENCE_NODES_IDENTICAL</literal> the
      custom listener can make <literal>Diff</literal> ignore the
      difference completely.  Likewise any
      <literal>Difference</literal> for which the custom listener
      returns
      <literal>RETURN_IGNORE_DIFFERENCE_NODES_SIMILAR</literal> will
      be treated as if the <literal>Difference</literal> was
      recoverable.</para>

      <para>There are several overloads of the <literal>Diff</literal>
      constructor that allow you to specify your piece of XML in many
      ways.  There are overloads that accept additional
      <literal>DifferenceEngine</literal> and
      <literal>ElementQualifier</literal> arguments.  Passing in a
      <literal>DifferenceEngine</literal> of your own is the only way
      to use a <literal>ComparisonController</literal> other than
      <literal>Diff</literal>.</para>

      <para>Note that <literal>Diff</literal> and
      <literal>DetailedDiff</literal> use
      <literal>ElementNameQualifier</literal> as their default
      <literal>ElementQualifier</literal>.  This is different from
      <literal>DifferenceEngine</literal> which defaults to no
      <literal>ElementQualifier</literal> at all.</para>

      <para>To use a custom <literal>ElementQualifier</literal> you
      can also use the <literal>overrideElementQualifier</literal>
      method.  Use this with an argument of <literal>null</literal> to
      unset the default <literal>ElementQualifier</literal> as
      well.</para>

      <para>To compare two pieces of XML you'd create a
      <literal>Diff</literal> instance from those two pieces and
      invoke <literal>identical</literal> to check that there have
      been no differences at all and <literal>similar</literal> to
      check that any difference, if any, has been recoverable.  If the
      pieces are identical they are also similar.  Likewise if they
      are not similar they can't be identical either.</para>

      <example>
        <title>Comparing Two Pieces of XML Using
        <literal>Diff</literal></title>
        <programlisting language="Java"><![CDATA[
Diff d = new Diff("<a><b/><c/></a>", "<a><c/><b/></a>");
assertFalse(d.identical()); // CHILD_NODELIST_SEQUENCE Difference
assertTrue(d.similar());
]]></programlisting></example>

      <para>The result of the comparison is cached in
      <literal>Diff</literal>, repeated invocations of
      <literal>identical</literal> or <literal>similar</literal> will
      not reevaluate the pieces of XML.</para>

      <para>Note: calling <literal>toString</literal> on an instance
      of <literal>Diff</literal> or <literal>DetailedDiff</literal>
      will perform the comparision and cache its result immediately.
      If you change the <literal>DifferenceListener</literal> or
      <literal>ElementQualifier</literal> after calling
      <literal>toString</literal> it won't have any effect.</para>

      <para><literal>DetailedDiff</literal> provides only a single
      constructor that expects a <literal>Diff</literal> as argument.
      Don't use <literal>DetailedDiff</literal> if all you need to
      know is whether two pieces of XML are identical/similar - use
      <literal>Diff</literal> directly since its short-cut
      <literal>ComparisonController</literal> implementation will save
      time in this case.</para>

      <example>
        <title>Finding All Differences Using
        <literal>DetailedDiff</literal></title>
        <programlisting language="Java"><![CDATA[
Diff d = new Diff("<a><b/><c/></a>", "<a><c/><b/></a>");
DetailedDiff dd = new DetailedDiff(d);
dd.overrideElementQualifier(null);
assertFalse(dd.similar());
List l = dd.getAllDifferences();
assertEquals(2, l.size()); // expected <b/> but was <c/> and vice versa
]]></programlisting></example>

    </section>

    <section id="MatchTracker">
      <title><literal>MatchTracker</literal></title>

      <para>Sometimes you might be interested in any sort of
      comparison result and want to get notified of successful matches
      as well.  Maybe you want to provide feedback on the amount of
      differences and similarities between two documents, for
      example.</para>

      <para>The interface <literal>MatchTracker</literal> can be
      implemented to get notified on each and every successful match,
      note that there may be a lot more comparisons going on than you
      might expect and that your callback gets notified a lot.</para>

      <example>
        <title>The <literal>MatchTracker</literal> interface</title>
        <programlisting language="Java"><![CDATA[
package org.custommonkey.xmlunit;

/**
 * Listener for callbacks from a {@link DifferenceEngine#compare
 * DifferenceEngine comparison} that is notified on each and every
 * comparision that resulted in a match.
 */
public interface MatchTracker {
    /**
     * Receive notification that 2 match.
     * @param match a Difference instance as defined in {@link
     * DifferenceConstants DifferenceConstants} describing the test
     * that matched and containing the detail of the nodes that have
     * been compared
     */
    void matchFound(Difference difference);
}
]]></programlisting></example>

      <para>Despite its name the <literal>Difference</literal>
        instance passed into the <literal>matchFound</literal> method
        really describes a match and not a difference.  You can expect
        that the <literal>getValue</literal> method on both the
        control and the test <literal>NodeDetail</literal> will be
        equal.</para>

      <para><literal>DifferenceEngine</literal> provides a constructor
        overload that allows you to pass in
        a <literal>MatchTracker</literal> instance and also provides
        a <literal>setMatchTracker</literal>
        method.  <literal>Diff</literal>
        and <literal>DetailedDiff</literal>
        provide <literal>overrideMatchTracker</literal> methods that
        fill the same purpose.</para>

      <para>Note that your <literal>MatchTracker</literal> won't
        receive any callbacks once the
        configured <literal>ComparisonController</literal> has decided
        that <literal>DifferenceEngine</literal> should halt the
        comparison.</para>

    </section>

    <section id="Comparing: JUnit 3">
      <title>JUnit 3.x Convenience Methods</title>

      <para><literal>XMLAssert</literal> and
      <literal>XMLTestCase</literal> contain quite a few overloads of
      methods for comparing two pieces of XML.</para>

      <para>The method's names use the word <literal>Equal</literal>
      to mean the same as <literal>similar</literal> in the
      <literal>Diff</literal> class (or throughout this guide).  So
      <literal>assertXMLEqual</literal> will assert that only
      recoverable differences have been encountered where
      <literal>assertXMLNotEqual</literal> asserts that some
      differences have been non-recoverable.
      <literal>assertXMLIdentical</literal> asserts that there haven't
      been any differences at all while
      <literal>assertXMLNotIdentical</literal> asserts that there have
      been differences (recoverable or not).</para>

      <para>Most of the overloads of <literal>assertXMLEqual</literal>
      just provide different means to specify the pieces of XML as
      <literal>String</literal>s, <literal>InputSource</literal>s,
      <literal>Reader</literal>s<footnote><para>See <xref
      linkend="Providing Input to XMLUnit"/> for some advice on choosing your input
      format.</para></footnote> or <literal>Document</literal>s.  For each
      method there is a version that takes an additional
      <literal>err</literal> argument which is used to create the
      message if the assertion fails.</para>

      <para>If you don't need any control over the
      <literal>ElementQualifier</literal> or
      <literal>DifferenceListener</literal> used by
      <literal>Diff</literal> these methods will save some boilerplate
      code.  If <literal>CONTROL</literal> and <literal>TEST</literal>
      are pieces of XML represented as one of the supported inputs
      then</para>

      <programlisting language="Java"><![CDATA[
Diff d = new Diff(CONTROL, TEST);
assertTrue("expected pieces to be similar, " + d.toString(),
           d.similar());
]]></programlisting>

      <para>and</para>

      <programlisting language="Java"><![CDATA[
assertXMLEqual("expected pieces to be similar", CONTROL, TEST);
]]></programlisting>

      <para>are equivalent.</para>

      <para>If you need more control over the <literal>Diff</literal>
      instance there is a version of <literal>assertXMLEqual</literal>
      (and <literal>assertXMLIdentical</literal>) that accepts a
      <literal>Diff</literal> instance as its argument as well as a
      <literal>boolean</literal> indicating whether you expect the
      <literal>Diff</literal> to be <literal>similar</literal>
      (<literal>identical</literal>) or not.</para>

      <para><literal>XMLTestCase</literal> contains a couple of
      <literal>compareXML</literal> methods that really are only
      shortcuts to <literal>Diff</literal>'s constructors.</para>

      <para>There is no way to use <literal>DifferenceEngine</literal>
      or <literal>DetailedDiff</literal> directly via the convenience
      methods.</para>

    </section>

    <section id="Comparing: Configuration">
      <title>Configuration Options</title>

      <para>Unless you are using <literal>Document</literal> or
      <literal>DOMSource</literal> overrides when specifying your
      pieces of XML, XMLUnit will use the configured XML parsers (see
      <xref linkend="JAXP"/>) and <literal>EntityResolver</literal>s
      (see <xref linkend="EntityResolver"/>).  There are configuration
      options to use different settings for the control and test
      pieces of XML.</para>

      <para>In addition some of the other configuration settings may
      lead to XMLUnit using the configured XSLT transformer (see <xref
      linkend="JAXP"/>) under the covers.</para>

      <section id="Whitespace Handling">
        <title>Whitespace Handling</title>

        <para>Two different configuration options affect how XMLUnit
        treats whitespace in comparisons:</para>

        <itemizedlist>

          <listitem>Element Content Whitespace (see <xref
          linkend="Basic: Element Content Whitespace"/>)

            <para>If XMLUnit has been configured to ignore element
            content whitespace it will trim any text nodes found by
            the parser.  This means that there won't appear to be any
            textual content in element <literal>&lt;foo&gt;</literal>
            for the following example.  If you don't set
            <literal>XMLUnit.setIgnoreWhitespace</literal> there would
            be textual content consisting of a new line
            character.</para>

            <programlisting language="XML"><![CDATA[
<foo>
</foo>
]]></programlisting>

            <para>At the same time the following two
            <literal>&lt;foo&gt;</literal> elements will be considered
            identical if the option has been enabled, though.</para>

            <programlisting language="XML"><![CDATA[
<foo>bar</foo>
<foo> bar </foo>
]]></programlisting>

            <para>When this option is set to <literal>true</literal>,
            <literal>Diff</literal> will use the XSLT transformer
            under the covers.</para>
          </listitem>

          <listitem>"Normalizing" Whitespace

            <para>If you set
            <literal>XMLUnit.setNormalizeWhitespace</literal> to true
            then XMLUnit will replace any kind of whitespace found in
            character content with a SPACE character and collapse
            consecutive whitespace characters to a single SPACE.  It
            will also trim the resulting character content on both
            ends.</para>

            <para>The following two <literal>&lt;foo&gt;</literal>
            elements will be considered identical if the option has
            been set:</para>

            <programlisting language="XML"><![CDATA[
<foo>bar baz</foo>
<foo> bar
            baz</foo>
]]></programlisting>

            <para>Note that this is not related to "normalizing" the
            document as a whole (see <xref
            linkend="Normalizing Documents"/>).</para>

          </listitem> 

        </itemizedlist>
      </section>

      <section id="Normalizing Documents">
        <title>"Normalizing" <literal>Document</literal>s</title>

        <para>"Normalize" in this context corresponds to the
        <literal>normalize</literal> method in DOM's
        <literal>Document</literal> class.  It is the process of
        merging adjacent <literal>Text</literal> nodes and is not
        related to "normalizing whitespace" as described in the
        previous section.</para>

        <para>Usually you don't need to care about this option since
        the XML parser is required to normalize the
        <literal>Document</literal> when creating it.  The only reason
        you may want to change the option via
        <literal>XMLUnit.setNormalize</literal> is that your
        <literal>Document</literal> instances have not been created by
        an XML parser but rather been put together in memory using the
        DOM API directly.</para>
      </section>

      <section id="Ignoring Comments">
        <title>Ignoring Comments</title>

        <para>Using <literal>XMLUnit.setIgnoreComments</literal> you
        can make XMLUnit's difference engine ignore comments
        completely.</para>

        <para>When this option is set to <literal>true</literal>,
        <literal>Diff</literal> will use the XSLT transformer under
        the covers.</para>
      </section>

      <section id="Treating CDATA Sections and Text Nodes Alike">
        <title>Treating CDATA Sections and Text Nodes Alike</title>

        <para>It is not always necessary to know whether a text has
        been put into a CDATA section or not.  Using
        <literal>XMLUnit.setIgnoreDiffBetweenTextAndCDATA</literal>
        you can make XMLUnit consider the following two pieces of XML
        identical:</para>

        <programlisting language="XML"><![CDATA[
<foo>&lt;bar&gt;</foo>
]]></programlisting>
        <programlisting language="XML">
&lt;foo>&lt;![CDATA[&lt;bar&gt;]]&gt;&lt;/foo>
</programlisting>
      </section>

      <section id="Entity Reference Expansion">
        <title>Entity Reference Expansion</title>

        <para>Normally the XML parser will expand character references
        to their Unicode equivalents but for more complex entity
        definitions the parser may expand them or not.
        Using <literal>XMLUnit.setExpandEntityReferences</literal> you
        can control the parser's setting.</para>

      </section>

      <section id="Comparison of Unmatched Elements">
        <title>Comparison of Unmatched Elements</title>

        <para>When XMLUnit cannot match a control Element to a test
          Element (the configured ElementQualifier - see
          <xref linkend="ElementQualifier"/> - doesn't return true for
          any of the test Elements) it will try to compare it against
          the first unmatched test Element (if there is one).
          Starting with XMLUnit 1.3 one can
          use <literal>XMLUnit.setCompareUnmatched</literal> to
          disable this behavior and
          generate <literal>CHILD_NODE_NOT_FOUND</literal> differences
          instead.</para>

        <para>If the control document is
        <programlisting language="XML"><![CDATA[
<root>
  <a/>
</root>
]]></programlisting>
        and the test document is
        <programlisting language="XML"><![CDATA[
<root>
  <b/>
</root>
]]></programlisting>

        the default setting will create a
        single <literal>ELEMENT_TAG_NAME</literal> Difference
        ("expected a but found b").
        Setting <literal>XMLUnit.setCompareUnmatched</literal> to
        false will create two Differences of
        type <literal>CHILD_NODE_NOT_FOUND</literal> (one for "a" and
        one for "b") instead.</para>
      </section>

    </section>
  </section>

  <section id="Validating XML Documents">
    <title>Validating XML Documents</title>

    <section id="The Validator Class">
      <title>The <literal>Validator</literal> Class</title>

      <para>The <literal>Validator</literal> class encapsulates
      XMLUnit's validation support.  It will use the
      <literal>SAXParser</literal> configured in XMLUnit (see <xref
      linkend="JAXP"/>).</para>

      <para>The piece of XML to validate is specified in the
      constructor.  The constructors using more than a single argument
      are only relevant if you want to validate against a DTD and need
      to provide the location of the DTD itself - for details see the
      next section.</para>

      <para>By default, <literal>Validator</literal> will validate
      against a DTD, but it is possible to validate against a (or
      multiple) Schema(s) as well.  Schema validation requires an XML
      parser that supports it, of course.</para>

      <section id="DTD Validation">
        <title>DTD Validation</title>
      </section>

      <para>Validating against a DTD is straight forward if the piece
      of XML contains a <literal>DOCTYPE</literal> declaration with a
      <literal>SYSTEM</literal> identifier that can be resolved at
      validation time.  Simply create a <literal>Validator</literal>
      object using one of the single argument constructors.</para>

      <example>
        <title>Validating Against the DTD Defined in
          <literal>DOCTYPE</literal></title>
        <programlisting language="Java"><![CDATA[
InputSource is = new InputSource(new FileInputStream(myXmlDocument));
Validator v = new Validator(is);
boolean isValid = v.isValid();
]]></programlisting></example>

      <para>If the piece of XML doesn't contain any
      <literal>DOCTYPE</literal> declaration at all or it contains a
      <literal>DOCTYPE</literal> but you want to validate against a
      different DTD, you'd use one of the three argument versions of
      <literal>Validator</literal>'s constructors.  In this case the
      <literal>publicId</literal> argument becomes the
      <literal>PUBLIC</literal> and <literal>systemId</literal> the
      <literal>SYSTEM</literal> identifier of the
      <literal>DOCTYPE</literal> that is implicitly added to the piece
      of XML.  Any existing <literal>DOCTYPE</literal> will be
      removed.  The <literal>systemId</literal> should be a URL that
      can be resolved by your parser.</para>

      <example>
        <title>Validating a Piece of XML that doesn't Contain a
        <literal>DOCTYPE</literal></title>
        <programlisting language="Java"><![CDATA[
InputSource is = new InputSource(new FileInputStream(myXmlDocument));
Validator v = new Validator(is,
                            (new File(myDTD)).toURI().toURL().toString(),
                            myPublicId);
boolean isValid = v.isValid();
]]></programlisting></example>

      <para>If the piece of XML already has the correct
      <literal>DOCTYPE</literal> declaration but the declaration
      either doesn't specify a <literal>SYSTEM</literal> identifier at
      all or you want the <literal>SYSTEM</literal> identifier to
      resolve to a different location you have two options:</para>

      <itemizedlist>
        <listitem>Use one of the two argument constructors and specify
          the alternative URL as
          <literal>systemId</literal>.

      <example>
        <title>Validating Against a Local DTD</title>
        <programlisting language="Java"><![CDATA[
InputSource is = new InputSource(new FileInputStream(myXmlDocument));
Validator v = new Validator(is,
                            (new File(myDTD)).toURI().toURL().toString());
boolean isValid = v.isValid();
]]></programlisting></example>
        </listitem>

        <listitem>Use a custom <literal>EntityResolver</literal> via
          <literal>XMLUnit.setControlEntityResolver</literal> together
          with one of the single argument constructor overloads of
          Validator.

          <para>This approach would allow you to use an OASIS
          catalog<footnote><para><ulink
          url="http://www.oasis-open.org/committees/download.php/14809/xml-catalogs.html">http://www.oasis-open.org/committees/download.php/14809/xml-catalogs.html</ulink></para></footnote>
          in conjunction with the Apache XML Resolver
          library<footnote><para><ulink
          url="http://xml.apache.org/commons/components/resolver/index.html">http://xml.apache.org/commons/components/resolver/index.html</ulink></para></footnote>
          to resolve the DTD location as well as the location of any
          other entity in your piece of XML, for example.</para>

      <example>
        <title>Validating Against a DTD Using Apache's XML Resolver and
        an XML Catalog</title>
        <programlisting language="Java"><![CDATA[
InputSource is = new InputSource(new FileInputStream(myXmlDocument));
XMLUnit.setControlEntityResolver(new CatalogResolver());
Validator v = new Validator(is);
boolean isValid = v.isValid();
]]></programlisting>

        <programlisting><![CDATA[
#CatalogManager.properties

verbosity=1
relative-catalogs=yes
catalogs=/some/path/to/catalog
prefer=public
static-catalog=yes
catalog-class-name=org.apache.xml.resolver.Resolver
]]></programlisting>

        <programlisting language="XML"><![CDATA[
<!-- catalog file -->

<catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog">
  <public publicId="-//Some//DTD V 1.1//EN"
          uri="mydtd.dtd"/>
</catalog>
]]></programlisting>

</example>

        </listitem>
      </itemizedlist>

      <section id="XML Schema Validation">
        <title>XML Schema Validation</title>
      </section>

      <para>In order to validate against the XML Schema language
      Schema validation has to be enabled via the
      <literal>useXMLSchema</literal> method of
      <literal>Validator</literal>.</para>

      <para>By default the parser will try to resolve the location of
      Schema definition files via a <literal>schemaLocation</literal>
      attribute if it is present in the piece of XML or it will try to
      open the Schema's URI as an URL and read from it.</para>

      <para>The <literal>setJAXP12SchemaSource</literal> method of
      <literal>Validator</literal> allows you to override this
      behavior as long as the parser supports the
      <literal>http://java.sun.com/xml/jaxp/properties/schemaSource</literal>
      property in the way described in "JAXP 1.2 Approved
      CHANGES"<footnote><para><ulink
      url="http://java.sun.com/webservices/jaxp/change-requests-11.html">http://java.sun.com/webservices/jaxp/change-requests-11.html</ulink></para></footnote>.</para>

      <para><literal>setJAXP12SchemaSource</literal>'s argument can be
      one of</para>

      <itemizedlist>
        <listitem>A <literal>String</literal> which contains an
          URI.</listitem>
        <listitem>An <literal>InputStream</literal> the Schema can be
          read from.</listitem>
        <listitem>An <literal>InputSource</literal> the Schema can be
          read from.</listitem>
        <listitem>A <literal>File</literal> the Schema can be
          read from.</listitem>
        <listitem>An array containing any of the above.</listitem>
      </itemizedlist>

      <para>If the property has been set using a
      <literal>String</literal>, the <literal>Validator</literal>
      class will provide its <literal>systemId</literal> as specified
      in the constructor when asked to resolve it.  You must only use
      the single argument constructors if you want to avoid this
      behavior.  If no <literal>systemId</literal> has been specified,
      the configured <literal>EntityResolver</literal> may still be
      used.</para>

      <example id="schema-jaxp12">
        <title>Validating Against a Local XML Schema</title>
        <programlisting language="Java"><![CDATA[
InputSource is = new InputSource(new FileInputStream(myXmlDocument));
Validator v = new Validator(is);
v.useXMLSchema(true);
v.setJAXP12SchemaSource(new File(myXmlSchemaFile));
boolean isValid = v.isValid();
]]></programlisting></example>

    </section>

    <section id="Validation: JUnit 3">
      <title>JUnit 3.x Convenience Methods</title>

      <para>Both <literal>XMLAssert</literal> and
      <literal>XMLTestCase</literal> provide an
      <literal>assertXMLValid(Validator)</literal> method that will
      fail if <literal>Validator</literal>'s
      <literal>isValid</literal> method returns
      <literal>false</literal>.</para>

      <para>In addition several overloads of the
      <literal>assertXMLValid</literal> method are provided that
      directly correspond to similar overloads of
      <literal>Validator</literal>'s constructor.  These overloads
      don't support XML Schema validation at all.</para>

      <para><literal>Validator</literal> itself provides an
      <literal>assertIsValid</literal> method that will throw an
      <literal>AssertionFailedError</literal> if validation
      fails.</para>

      <para>Neither method provides any control over the message of
      the <literal>AssertionFailedError</literal> in case of a
      failure.</para>
    </section>

    <section id="Validation: Configuration">
      <title>Configuration Options</title>

      <itemizedlist>
        <listitem><literal>Validator</literal> uses a SAX parser
        created by the configured SAX parser factory (see <xref
        linkend="JAXP"/>).</listitem>

        <listitem>It will use the "control"
        <literal>EntityResolver</literal> if one has been specified
        (see <xref linkend="EntityResolver"/>).</listitem>

        <listitem>The location of a DTD can be specified via
        <literal>Validator</literal>'s <literal>systemId</literal>
        constructor argument or a custom EntityResolver (see <xref
        linkend="DTD Validation"/>).</listitem>

        <listitem>XML Schema validation is enabled via
        <literal>Validator.useXMLSchema(true)</literal>.</listitem>

        <listitem>The location(s) of XML Schema document(s) can be
        specified via
        <literal>Validator.setJAXP12SchemaSource</literal> (see <xref
        linkend="XML Schema Validation"/>).</listitem>
      </itemizedlist>
    </section>

    <section id="JAXP 1.3 Validation">
      <title>JAXP 1.3 Validation</title>

      <para>JAXP 1.3 - shipping with Java5 or better and available as
      a separate product for earlier Java VMs - introduces a new
      package <ulink url="https://jaxp-sources.dev.java.net/nonav/docs/api/"><literal>javax.xml.validation</literal></ulink>
      designed for validations of snippets of XML against different
      schema languages.  Any compliant implementation must support the
      W3C XML Schema language, but other languages
      like <ulink url="http://www.relaxng.org/">RELAX NG</ulink> or
      <ulink url="http://www.schematron.com/">Schematron</ulink> may
      be supported as well.</para>

      <para>The
      class <literal>org.custommonkey.xmlunit.jaxp13.Validator</literal>
      can be used to validate a piece of XML against a schema
      definition but also to validate the schema definition itself.
      By default <literal>Validator</literal> will assume your
      definition uses the W3C XML Schema language, but it provides a
      constructor that can be used to specify a different language via
      an URL supported by the <literal>SchemaFactory</literal> class.
      Alternatively you can specify the schema factory itself.</para>

      <para>The schema definition itself can be given via
      <literal>Source</literal> elements, just like the pieces of XML
      to validate are specified as <literal>Source</literal> as
      well.</para>

      <para>Note the <literal>Validator</literal> class
      of <literal>javax.xml.validation</literal> will ignore all
      <literal>xsi:namespaceLocation</literal> and
      <literal>xsi:noNamespaceLocation</literal> attributes of the XML
      document you want to validate if you specify at least one schema
      source.</para>

      <para>The following example
      uses <literal>org.custommonkey.xmlunit.jaxp13.Validator</literal>
      to perform the same type of validation shown in
      <xref linkend="schema-jaxp12"/>.</para>

      <example id="schema-jaxp13">
        <title>Validating Against a Local XML Schema</title>
        <programlisting language="Java"><![CDATA[
Validator v = new Validator();
v.addSchemaSource(new StreamSource(new File(myXmlSchemaFile)));
StreamSource is = new StreamSource(new File(myXmlDocument));
boolean isValid = v.isInstanceValid(is);
]]></programlisting></example>

      <para>Validating a schema definition is shown in the next
      example.</para>

      <example>
        <title>Validating an XML Schema Definition</title>
        <programlisting language="Java"><![CDATA[
Validator v = new Validator();
v.addSchemaSource(new StreamSource(new File(myXmlSchemaFile)));
boolean isValid = v.isSchemaValid();
]]></programlisting></example>

      <para>There is no explicit JUnit 3 support
      for <literal>org.custommonkey.xmlunit.jaxp13.Validator</literal>.</para>

    </section>

  </section>

  <section id="XPath Tests">
    <title>XPath Tests</title>

    <section id="XPath Engines">
      <title>XPath Engines</title>

      <para>Central to XMLUnit's XPath support is the
      <literal>XpathEngine</literal> interface which consists of only
      three methods:</para>

      <programlisting language="Java"><![CDATA[
    /**
     * Execute the specified xpath syntax <code>select</code> expression
     * on the specified document and return the list of nodes (could have
     * length zero) that match
     * @param select
     * @param document
     * @return list of matching nodes
     */
    NodeList getMatchingNodes(String select, Document document)
        throws XpathException;
    
    /**
     * Evaluate the result of executing the specified XPath syntax
     * <code>select</code> expression on the specified document
     * @param select
     * @param document
     * @return evaluated result
     */
    String evaluate(String select, Document document)
        throws XpathException;

    /**
     * Establish a namespace context.
     */
    void setNamespaceContext(NamespaceContext ctx);
]]></programlisting>


      <para>The first two methods expect an XPath expression that
      selects content from the DOM document that is the second
      argument.  The result of the selection can be either a DOM
      <literal>NodeList</literal> or a <literal>String</literal>.  The
      later form tries to flatten the result, the value is said to be
      "String-ified".</para>

      <para>The third method is part of XMLUnit's support for XML
      Namespaces in XPath expressions.  See <xref linkend="Using XML Namespaces in XPath Selectors"/>
      for more details.</para>

      <para>There are two implementations of the interface,
      <literal>org.custommonkey.xmlunit.SimpleXpathEngine</literal>
      and
      <literal>org.custommonkey.xmlunit.jaxp13.Jaxp13XpathEngine</literal>.
      The first implementation is the only one available in XMLUnit
      1.0 and uses the <link linked="JAXP">configured</link> JAXP XSLT
      transformer.  The second is new to XMLUnit 1.1 and only
      available if JAXP 1.3 or later is supported, which should be the
      case for Java 5 and later.</para>

      <para><literal>XpathException</literal> is an
      <literal>Exception</literal> that will be thrown for invalid
      XPath expressions or other problems with the underlying XPath
      engine.  It will typically wrap a
      <literal>javax.xml.xpath.XPathExpressionException</literal> in
      the <literal>Jaxp13XpathEngine</literal> case or a
      <literal>javax.xml.transform.TransformerException</literal> when
      <literal>SimpleXpathEngine</literal> is used.</para>

      <para>The <literal>XMLUnit.newXpathEngine</literal> method will
      first try to create an instance of
      <literal>Jaxp13XpathEngine</literal> and fall back to
      <literal>SimpleXpathEngine</literal> if JAXP 1.3 is not
      supported.</para>

      <para>One example of using the XPath support is included inside
      it <literal>org.custommonkey.xmlunit.examples</literal> package.
      It asserts that the string-ified form of an XPath selection
      matches a regular expression.  The code needed for this
      is:</para>

      <example>
        <title>Matching an XPath Selection Against a Regular
          Expression</title>
        <programlisting language="Java"><![CDATA[
    XpathEngine engine = XMLUnit.newXpathEngine();
    String value = engine.evaluate(xpath, doc);
    Assert.assertTrue(message, value.matches(regex));
]]></programlisting></example>

    </section>
    
    <section id="Using XML Namespaces in XPath Selectors">
      <title>Using XML Namespaces in XPath Selectors</title>

      <para>Starting with XMLUnit 1.1 XML Namespaces are supported for
      XPath queries.</para>

      <para>The <literal>NamespaceContext</literal> interface provides
      a mapping from prefix to namespace URI and is used by the XPath
      engine.  XPath selections then use the mapping's prefixes where
      needed.  Note that a prefix used in the document under test and
      a prefix given as part of the
      <literal>NamespaceContext</literal> are not related at all; the
      context only applies to the XPath expression, the prefix used in
      the document is ignored completely.</para>

      <para>Right now XMLUnit provides only a single implementation of
      the <literal>NamespaceContext</literal> interface:
      <literal>SimpleNamespaceContext</literal>.  This implementation
      expects a <literal>java.util.Map</literal> as its constructor
      argument.  The <literal>Map</literal> must contain
      (<literal>String</literal>) prefixes as keys and
      (<literal>String</literal>) namespace URIs as values.</para>

      <para>Note there is nothing like a default namespace in XPath
      selectors.  If you are using namespaces in your XPath, all
      namespaces need a prefix (of length greater than zero).  This
      is independent of the prefixes used in your document.</para>

      <para>The following example is taken from XMLUnit's own tests.
      It demonstrates that the namespace prefix of the document under
      test is irrelevant and shows how to set up the namespace
      context.</para>

      <example>
        <title>Using Namespaces in XPath Tests</title>
        <programlisting language="Java"><![CDATA[
    String testDoc = "<t:test xmlns:t=\"urn:foo\"><t:bar/></t:test>";
    Document d = XMLUnit.buildControlDocument(testDoc);
    HashMap m = new HashMap();
    m.put("foo", "urn:foo");

    NamespaceContext ctx = new SimpleNamespaceContext(m);
    XpathEngine engine = XMLUnit.newXpathEngine();
    engine.setNamespaceContext(ctx);

    NodeList l = engine.getMatchingNodes("//foo:bar", d);
    assertEquals(1, l.getLength());
    assertEquals(Node.ELEMENT_NODE, l.item(0).getNodeType());
]]></programlisting></example>

      <para>In some cases the "stringified" value of an XPath
      evaluation is a qualified name - a string that encodes a
      namespace URI together with a local name.  There are two common
      formats for such qualified names, one used by Java5's
      <literal>QName</literal> in the format
      <literal>{NS-URI}LOCAL-NAME</literal> and one using
      <literal>PREFIX:LOCAL-NAME</literal>.  Starting with XMLUnit 1.6
      a new <literal>QualifiedName</literal> class can parse either
      representation.  The <literal>assertXpathEvaluatesTo</literal>
      overloads where the expected value is a
      <literal>QualifiedName</literal> try to parse the stringified
      value in either format - using the documents namespace context
      when parsing the actual value.</para>

      <para>It is possible to set a global
      <literal>NamespaceContext</literal>, see <xref
      linkend="XPath: Configuration"/> for details.</para>
    </section>

    <section id="XPath: JUnit 3">
      <title>JUnit 3.x Convenience Methods</title>

      <para><literal>XMLTestCase</literal> and
      <literal>XMLAssert</literal> provide several overloads for the
      following common types of assertions:</para>

      <itemizedlist>

        <listitem>Two XPath expression should return the same DOM
        <literal>NodeList</literal> as result:
        <literal>assertXpathsEqual</literal>.  There are methods that
        use two different expressions on the same document and others
        that compare expressions selecting from two different
        documents.

          <para>The <literal>NodeList</literal>s are wrapped into a
          surrogate root XML element and the resulting DOM
          <literal>Document</literal>s are compared using
          <literal>Diff.similar()</literal>.</para>
        </listitem>

        <listitem>The opposite of the above, the expressions should
        yield different results:
        <literal>assertXpathsNotEqual</literal>.</listitem>

        <listitem>Two XPath expression should return the same
        "String-ified" result:
        <literal>assertXpathValuesEqual</literal>.  There are methods
        that use two different expressions on the same document and
        others that compare expressions selecting from two different
        documents.</listitem>

        <listitem>The opposite of the above, the expressions should
        yield different results:
        <literal>assertXpathValuesNotEqual</literal>.</listitem>

        <listitem>The XPath expression should return an expected value
        when "String-ified" or interpreted as qualified name:
        <literal>assertXpathEvaluatesTo</literal>.</listitem>

        <listitem>The <literal>NodeList</literal> selected by an XPath
        expression is not empty:
        <literal>assertXpathExists</literal>.</listitem>

        <listitem>The <literal>NodeList</literal> selected by an XPath
        expression is empty:
        <literal>assertXpathNotExists</literal>.</listitem>
      </itemizedlist>

      <para>Neither method provides any control over the message of
      the <literal>AssertionFailedError</literal> in case of a
      failure.</para>
    </section>

    <section id="XPath: Configuration">
      <title>Configuration Options</title>

      <para>When using <literal>XpathEngine</literal> directly you are
      responsible for creating the DOM document yourself.  If you use
      the convenience methods of <literal>XMLTestCase</literal> or
      <literal>XMLAssert</literal> you have several options to specify
      the input; XMLUnit will use the control or test parser that has
      been configured (see <xref linkend="JAXP"/>) to create a DOM
      document from the given piece of XML in that case - using the
      configured <literal>EntityResolver</literal>(s) (see <xref
      linkend="EntityResolver"/>) if any.</para>

      <para>If JAXP 1.3 is not available,
      <literal>SimpleXpathEngine</literal> will use the configured
      JAXP XSLT transformer (see <xref linkend="JAXP"/>) under the
      covers.</para>

      <para>When using JAXP 1.3 you can chose the actual
      <literal>XPathFactory</literal> implementation using
      <literal>XMLUnit.setXPathFactory</literal>.</para>

      <para>It is possible to establish a global
      <literal>NamespaceContext</literal> with the help of the
      <literal>XMLUnit.setXpathNamespaceContext</literal> method.  Any
      <literal>XpathEngine</literal> created by
      <literal>XMLUnit.newXpathEngine</literal> will automatically use
      the given context.  Note that the JUnit 3 convenience methods
      use <literal>XMLUnit.newXpathEngine</literal> implicitly and
      will thus use the configured
      <literal>NamespaceContext</literal>.</para>

    </section>
  </section>

  <section id="DOM Tree Walking">
    <title>DOM Tree Walking</title>

    <para>Sometimes it is easier to test a piece of XML's validity by
    traversing the whole document node by node and test each node
    individually.  Maybe there is no control XML to validate against
    or the expected value of an element's content has to be
    calculated.  There may be several reasons.</para>

    <para>XMLUnit supports this approach of testing via the
    <literal>NodeTest</literal> class.  In order to use it, you need a
    DOM implementation that generates <literal>Document</literal>
    instances that implement the optional
    <literal>org.w3c.traversal.DocumentTraversal</literal> interface,
    which is not part of JAXP's standardized DOM support.</para>

    <section id="DocumentTraversal">
      <title><literal>DocumentTraversal</literal></title>

      <para>As of the release of XMLUnit 1.1 the
      <literal>Document</literal> instances created by most parsers
      implement <literal>DocumentTraversal</literal>, this includes
      but is not limited to Apache Xerces, the parser shipping with
      Sun's JDK 5 and later or GNU JAXP.  One notable exception is
      Apache Crimson, which also means the parser shipping with Sun's
      JDK 1.4 does not support traversal; you need to specify a
      different parser when using JDK 1.4 (see <xref
      linkend="JAXP"/>).</para>

      <para>You can test whether your XML parser supports
      <literal>DocumentTraversal</literal> by invoking
      <literal>org.w3c.dom.DOMImplementation</literal>'s
      <literal>hasFeature</literal> method with the feature
      <literal>"Traversal"</literal>.</para>
    </section>

    <section id="NodeTest">
      <title><literal>NodeTest</literal></title>

      <para>The <literal>NodeTest</literal> is instantiated with a
      piece of XML to traverse.  It offers two
      <literal>performTest</literal> methods:</para>

      <programlisting language="Java"><![CDATA[
    /**
     * Does this NodeTest pass using the specified NodeTester instance?
     * @param tester
     * @param singleNodeType note <code>Node.ATTRIBUTE_NODE</code> is not
     *  exposed by the DocumentTraversal node iterator unless the root node
     *  is itself an attribute - so a NodeTester that needs to test attributes
     *  should obtain those attributes from <code>Node.ELEMENT_NODE</code>
     *  nodes
     * @exception NodeTestException if test fails
     */
    public void performTest(NodeTester tester, short singleNodeType);

    /**
     * Does this NodeTest pass using the specified NodeTester instance?
     * @param tester
     * @param nodeTypes note <code>Node.ATTRIBUTE_NODE</code> is not
     *  exposed by the DocumentTraversal node iterator unless the root node
     *  is itself an attribute - so a NodeTester that needs to test attributes
     *  should obtain those attributes from <code>Node.ELEMENT_NODE</code>
     *  nodes instead
     * @exception NodeTestException if test fails
     */
    public void performTest(NodeTester tester, short[] nodeTypes);
]]></programlisting>

      <para><literal>NodeTester</literal> is the class testing each
      node and is described in the next section.</para>

      <para>The second argument limits the tests on DOM
      <literal>Node</literal>s of (a) specific type(s).
      <literal>Node</literal> types are specified via the
      <literal>static</literal> fields of the <literal>Node</literal>
      class.  Any <literal>Node</literal> of a type not specified as
      the second argument to <literal>performTest</literal> will be
      ignored.</para>

      <para>Unfortunately XML attributes are not exposed as
      <literal>Node</literal>s during traversal.  If you need access
      to attributes you must add <literal>Node.ELEMENT_NODE</literal>
      to the second argument of <literal>performTest</literal> and
      access the attributes from their parent
      <literal>Element</literal>.</para>

      <example id="nodetest-attributes">
        <title>Accessing Attributes in a
          <literal>NodeTest</literal></title>
        <programlisting language="Java"><![CDATA[
    ...
    NodeTest nt = new NodeTest(myXML);
    NodeTester tester = new MyNodeTester();
    nt.performTest(tester, Node.ELEMENT_NODE);
    ...

class MyNodeTester implements NodeTester {
    public void testNode(Node aNode, NodeTest test) {
        Element anElement = (Element) aNode;
        Attr attributeToTest = anElement.getAttributeNode(ATTRIBUTE_NAME);
        ...
    }
    ...
}
]]></programlisting></example>

      <para>Any entities that appear as part of the
      <literal>Document</literal> are expanded before the traversal
      starts.</para>

    </section>

    <section id="NodeTester">
      <title>NodeTester</title>

      <para>Implementations of the <literal>NodeTester</literal>
      interface are responsible for the actual test:</para>

      <programlisting language="Java"><![CDATA[
    /**
     * Validate a single Node
     * @param aNode
     * @param forTest
     * @exception NodeTestException if the node fails the test
     */
    void testNode(Node aNode, NodeTest forTest) throws NodeTestException ;

    /**
     * Validate that the Nodes passed one-by-one to the <code>testNode</code>
     * method were all the Nodes expected.
     * @param forTest
     * @exception NodeTestException if this instance was expecting more nodes
     */
    void noMoreNodes(NodeTest forTest) throws NodeTestException ;
]]></programlisting>

      <para><literal>NodeTest</literal> invokes
      <literal>testNode</literal> for each <literal>Node</literal> as
      soon as it is reached on the traversal.  This means
      <literal>NodeTester</literal> "sees" the
      <literal>Node</literal>s in the same order they appear within
      the tree.</para>

      <para><literal>noMoreNodes</literal> is invoked when the
      traversal is finished.  It will also be invoked if the tree didn't
      contain any matched <literal>Node</literal>s at all.</para>

      <para>Implementations of <literal>NodeTester</literal> are
      expected to throw a <literal>NodeTestException</literal> if the
      current not doesn't match the test's expectations or more nodes
      have been expected when <literal>noMoreNodes</literal> is
      called.</para>

      <para>XMLUnit ships with two implementations of
      <literal>NodeTest</literal> that are described in the following
      to sections.</para>

      <section id="AbstractNodeTester">
        <title><literal>AbstractNodeTester</literal></title>

        <para><literal>AbstractNodeTester</literal> implements
        <literal>testNode</literal> by testing the passed in
        <literal>Node</literal> for its type and delegating to one of
        the more specific <literal>test...</literal> Methods it adds.
        By default the new <literal>test...</literal> methods all
        throw a <literal>NodeTestException</literal> because of an
        unexpected <literal>Node</literal>.</para>

        <para>It further implements <literal>noMoreNodes</literal>
        with an empty method - i.e. it does nothing.</para>

        <para>If you are only testing for specific types of
        <literal>Node</literal> it may be more convenient to subclass
        <literal>AbstractNodeTester</literal>.  For example <xref
        linkend="nodetest-attributes"/> could be re-written as:</para>

        <example>
          <title>Accessing Attributes in a
            <literal>NodeTest</literal> -
            <literal>AbstractNodeTester</literal> version</title>
          <programlisting language="Java"><![CDATA[
    ...
    NodeTest nt = new NodeTest(myXML);
    NodeTester tester = new AbstractNodeTester() {
        public void testElement(Element element) throws NodeTestException {
            Attr attributeToTest = element.getAttributeNode(ATTRIBUTE_NAME);
            ...
        }
    };
    nt.performTest(tester, Node.ELEMENT_NODE);
    ...
]]></programlisting></example>

        <para>Note that even though
        <literal>AbstractNodeTester</literal> contains a
        <literal>testAttribute</literal> method it will never be
        called by default and you still need to access attributes via
        their parent elements.</para>

        <para>Note also that the root of the test is the document's
        root element, so any <literal>Node</literal>s preceding the
        document's root <literal>Element</literal> won't be visited
        either.  For this reason the
        <literal>testDocumentType</literal>,
        <literal>testEntity</literal> and
        <literal>testNotation</literal> methods are probably never
        called either.</para>

        <para>Finally, all entity references have been expanded before
        the traversal started.  <literal>EntityReference</literal>s
        will have been replaced by their replacement text if it is
        available, which means <literal>testEntityReference</literal>
        will not be called for them either.  Instead the replacement
        text will show up as (part of) a <literal>Text</literal> node
        or as <literal>Element</literal> node, depending on the
        entity's definition.</para>
      </section>

      <section id="CountingNodeTester">
        <title><literal>CountingNodeTester</literal></title>

        <para><literal>org.custommonkey.xmlunit.examples.CountingNodeTester</literal>
        is a simple example <literal>NodeTester</literal> that asserts
        that a given number of <literal>Node</literal>s have been
        traversed.  It will throw a
        <literal>NodeTestException</literal> when
        <literal>noMoreNodes</literal> is called before the expected
        number of <literal>Node</literal>s has been visited or the
        actual number of nodes exceeded the expected count.</para>
      </section>
    </section>

    <section id="DOM Tree: JUnit 3">
      <title>JUnit 3.x Convenience Methods</title>

      <para><literal>XMLAssert</literal> and
      <literal>XMLTestCase</literal> contain overloads of
      <literal>assertNodeTestPasses</literal> methods.</para>

      <para>The most general form of it expects you to create a
      <literal>NodeTest</literal> instance yourself and lets you
      specify whether you expect the test to fail or to pass.</para>

      <para>The other two overloads create a
      <literal>NodeTest</literal> instance from either
      <literal>String</literal> or a SAX
      <literal>InputSource</literal> and are specialized for the case
      where you are only interested in a single <literal>Node</literal>
      type and expect the test to pass.</para>

      <para>Neither method provides any control over the message of
      the <literal>AssertionFailedError</literal> in case of a
      failure.</para>
    </section>

    <section id="DOM Tree: Configuration">
      <title>Configuration Options</title>

      <para>The only configurable option for
      <literal>NodeTest</literal> is the XML parser used if the piece
      of XML is not specified as a <literal>Document</literal> or
      <literal>DocumentTraversal</literal>.
      <literal>NodeTest</literal> will use the "control" parser that
      has been configured - see <xref linkend="JAXP"/> for
      details.</para>

      <para>It will also use the <literal>EntityResolver</literal>
      configured for the control parser if one has been set - see
      <xref linkend="EntityResolver"/>.</para>
    </section>

  </section>

  <appendix id="changes">
    <title>Changes</title>

    <section id="Changes 1.1">
      <title>Changes from XMLUnit 1.0 to 1.1</title>

      <para>XMLUnit 1.1's main focus was to add two features that have
        been asked for repeatedly:</para>

      <itemizedlist>
        <listitem>Support for XML Namespaces in XPath
          processing</listitem>

        <listitem>Support for XML Schema validation.</listitem>
      </itemizedlist>

      <para>In addition some JAXP features that have been added after
      the release of XMLUnit 1.0 are now supported - most notably
      XPath support - and all reported bugs and feature requests have
      been addressed.</para>

      <section id="Breaking Changes 1.1">
        <title>Breaking Changes</title>

        <itemizedlist>
          <listitem>
            <literal>XMLTestCase</literal> is now abstract.  You
            probably have never created instances of this class
            without subclassing it, but if you did, your code will now
            break.  You will most likely want to look at the
            <literal>XMLAssert</literal> class.
          </listitem>

          <listitem>
            <para>All methods that have been deprecated in XMLUnit 1.0
            have been removed.</para>
          </listitem>

          <listitem>
            <para>All methods that had been declared to throw
            <literal>TransformerConfigurationException</literal> or
            <literal>ParserConfigurationException</literal> now no
            longer declare it.  Exceptions of these types cannot be
            recovered from anyway, so XMLUnit will now wrap them in a
            <literal>org.custommonkey.xmlunit.exceptions.ConfigurationException</literal>
            which is an unchecked exception.</para>

            <para>This change doesn't have a big impact on your tests,
            but if you tried to catch these exceptions they will now
            bypass your catch blocks.</para>
          </listitem>

          <listitem>
            <para>A new type of <literal>Difference</literal>
            (<literal>CHILD_NODE_NOT_FOUND_ID</literal>) has been
            added.  It will be raised for the excess children if the
            control element has more child nodes than the test element
            - or vice versa.</para>

            <para>Prior to XMLUnit 1.1 a <literal>Difference</literal>
            of either <literal>ELEMENT_TAG_NAME_ID</literal> or
            <literal>NODE_TYPE_ID</literal> would have been raised if
            the control element had more children.  The excess
            children were compared to the very first child node of the
            test element.  Excess children of the test element were
            not reported at all.</para>
          </listitem>

          <listitem>
            <para>The <literal>schemaLocation</literal> and
            <literal>noNamespaceSchemaLocation</literal> attributes of
            the XMLSchema-Instance Namespace are now treated in a
            different way from "normal" attributes.  They will be
            flagged as new kinds of <literal>Difference</literal> that
            is recoverable.</para>

            <para>This means that two pieces of XML that were
            different in XMLUnit 1.0 because they differed in one of
            the two attributes will be similar in XMLUnit 1.1.</para>
          </listitem>

          <listitem>
            <para>When comparing two elements that differ on
            attributes the comparison is now symmetric.</para>

            <para>In XMLUnit 1.0 if an attribute was present on the
            test but not the control element this wasn't flagged as a
            <literal>Difference</literal>; in XMLUnit 1.1 it
            is.</para>

            <para>In most practical cases this doesn't cause any
            problems since the two elements either have a different
            number of attributes or there are attributes in the
            control element that are missing in the test element - so
            the pieces of XML have been flagged as different before as
            well.  If you are using <literal>DetailedDiff</literal>
            this change may lead to more detected
            <literal>Difference</literal>s, though.</para>
          </listitem>
        </itemizedlist>
      </section>

      <section id="New Features 1.1">
        <title>New Features</title>

        <itemizedlist>
          <listitem>XMLUnit 1.0 shipped with rudimentary support for
          XML Schema validation (it worked with Apache Xerces-J but no
          other parsers).  XMLUnit 1.1 supports Schema validation for
          any JAXP compliant XML parser (that supports Schema itself).
          You can also tell XMLUnit where to look for the XML Schema
          definitions.  See <xref linkend="XML Schema Validation"/> for
          details.</listitem>

          <listitem>XPath support has undergone significant changes,
          see <xref linkend="XPath Tests"/> for more details.  In particular
          XMLUnit will now use <literal>javax.xml.xpath</literal> if
          it is available (which also helps to avoid the buggy XSLTC
          version that is the default transformer engine in Java 5)
          and supports XML namespaces.</listitem>

          <listitem>Several new configuration options have been added,
            see <xref linkend="Comparing: Configuration"/>.
            <itemizedlist>
              <listitem>Treat CDATA sections and Texts alike.  <ulink
              url="http://sourceforge.net/tracker/index.php?func=detail&amp;aid=1262148&amp;group_id=23187&amp;atid=377768">Issue
              1262148</ulink>.</listitem>

              <listitem>Ignore differences in Text whitespace.  <ulink
              url="http://sourceforge.net/tracker/index.php?func=detail&amp;aid=754812&amp;group_id=23187&amp;atid=377771">Issue
              754812</ulink>.</listitem>

              <listitem>Ignore comments completely.  <ulink
              url="http://sourceforge.net/tracker/index.php?func=detail&amp;aid=707255&amp;group_id=23187&amp;atid=377770">Issue
              707255</ulink>.</listitem>

              <listitem>Ignore the order of attributes.</listitem>
            </itemizedlist>
          </listitem>

          <listitem>It is now possible to provide a custom
          <literal>org.xml.sax.EntityResolver</literal> for control
          and test parsers.</listitem>

          <listitem>It is now possible to provide a custom
          <literal>javax.xml.transform.URIResolver</literal> for
          transformations.</listitem>

          <listitem>New overloads have been added that allow
          <literal>org.xml.sax.InputSource</literal> to be used as a
          "piece of XML" in many classes.</listitem>

          <listitem><literal>Validator</literal> will now use the
          custom <literal>EntityResolver</literal>  <link
          linkend="EntityResolver">configured</link> for the "control"
          parser as a fallback.</listitem>

          <listitem>
            <para>A new package
              <literal>org.custommonkey.xmlunit.examples</literal> has
              been added that showcases some of XMLUnit's abilities.
              It currently contains two classes:</para>

            <orderedlist>
              <listitem>
                <literal>MultiLevelElementNameAndTextQualifier</literal>
                see <xref
                linkend="MultiLevelElementNameAndTextQualifier"/> for
                a description.</listitem>

              <listitem><literal>XPathRegexAssert</literal> that
              provides a JUnit 3.x like
              <literal>assertXPathMatches</literal> method to verify
              that the string-ified value of an XPath match matches a
              given regular expression (requires JDK 1.4 or
              above).</listitem>
            </orderedlist>
          </listitem>
        </itemizedlist>
      </section>

      <section id="Bugfixes 1.1">
        <title>Important Bug Fixes</title>

        <itemizedlist>
          <listitem><literal>ElementNameAndAttributeQualifier</literal>
          would throw an <literal>NullPointerException</literal> if
          the control piece of XML contained attributes that were
          missing in the test piece of XML.  <ulink
          url="http://sourceforge.net/tracker/index.php?func=detail&amp;aid=952920&amp;group_id=23187&amp;atid=377768">Issue
          952920</ulink>.</listitem>

          <listitem>
            <literal>XMLTestCase.assertXMLNotEqual(String, Reader,
            Reader)</literal> delegated to
            <literal>assertXMLEqual</literal> instead of
            <literal>assertXMLNotEqual</literal> internally, negating
            the assertion's logic.  <ulink
            url="http://sourceforge.net/tracker/index.php?func=detail&amp;aid=956372&amp;group_id=23187&amp;atid=377768">Issue
            956372</ulink>.</listitem>

          <listitem><literal>XMLTestCase.assertXMLIdentical(Diff,
          boolean)</literal> delegated to
          <literal>assertXMLEqual</literal>, weakening the
          assertion.</listitem>

          <listitem>Under certain circumstances the reported XPath
          expressions for nodes that showed differences were wrong.
          XMLUnit could lose the root element or erroneously append an
          extra attribute name.  Issues <ulink
          url="http://sourceforge.net/tracker/index.php?func=detail&amp;aid=1047364&amp;group_id=23187&amp;atid=377768">1047364</ulink>
          and <ulink url="http://sourceforge.net/tracker/index.php?func=detail&amp;aid=1027863&amp;group_id=23187&amp;atid=377770">1027863</ulink>.</listitem>

          <listitem>
            <literal>TolerantSaxParser</literal>'s logic in
            <literal>characters</literal> was broken and could cause
            <literal>StringIndexOutOfBoundsException</literal>s.
            <ulink
            url="http://sourceforge.net/tracker/index.php?func=detail&amp;aid=1150234&amp;group_id=23187&amp;atid=377768">Issue 1150234</ulink>.</listitem>
        </itemizedlist>
      </section>
    </section>

    <section id="Changes 1.2">
      <title>Changes from XMLUnit 1.1 to 1.2</title>

      <section id="Breaking Changes 1.2">
        <title>Breaking Changes</title>

        <itemizedlist>
          <listitem>
            If XMLUnit detects that it cannot match a certain node
            (i.e. it encounters
            a <literal>CHILD_NODE_NOT_FOUND</literal>
            kind of difference) the XPath for the "missing" node will
            be null.  It used to be some random XPath of a different node.
          </listitem>

          <listitem>
            <literal>XMLUnit.setIgnoreDiffBetweenTextAndCDATA</literal>
            now also
            sets <literal>DocumentBuilderFactory.setCoalescing</literal>.
            This has been done so that whitespace differences can be
            resolved according to the corresponding flags even in the
            presence of CDATA
            sections.  <ulink href="https://sourceforge.net/tracker/index.php?func=detail&amp;aid=1903923&amp;group_id=23187&amp;atid=377768">Issue
            1903923</ulink>.</listitem>

          <listitem>
            Two protected methods
            in <literal>SimpleXPathEngine</literal> (which you
            shouldn't extend anyway) have added XpathException to
            their throws list.
          </listitem>
        </itemizedlist>
      </section>

      <section id="New Features 1.2">
        <title>New Features</title>

        <itemizedlist>
          <listitem>The <literal>SAXParserFactory</literal> used
          by <literal>Validator</literal> can now be configured
          completely.  <ulink href="https://sourceforge.net/tracker/index.php?func=detail&amp;aid=1903928&amp;group_id=23187&amp;atid=377771">Issue
          1903928</ulink>.</listitem>

          <listitem>A new
          class <literal>org.custommonkey.xmlunit.jaxp13.Validator</literal>
          can be used to validate schema definitions and schema
          instances using the <literal>javax.xml.validation</literal>
          package of JAXP 1.3.  Depending on your JAXP implementation
          this may allow you to validate documents against schema
          definitions written in RELAX NG or other schema languages in
          addition to W3C XML Schema.  See
          <xref linkend="JAXP 1.3 Validation"/> for details.</listitem>

          <listitem><literal>DifferenceListener</literal> can now
          "upgrade" recoverable differences to non-recoverable by
          returning <literal>RETURN_UPGRADE_DIFFERENCE_NODES_DIFFERENT</literal>
          in the <literal>differenceFound</literal>
          method.  <ulink url="https://sourceforge.net/tracker/index.php?func=detail&amp;aid=1854284&amp;group_id=23187&amp;atid=377771">Issue
          1854284</ulink>.</listitem>

          <listitem>A new callback
          interface <literal>MatchTracker</literal> is now notified on
          successful matches of Nodes.  For more details see
          <xref linkend="MatchTracker"/>.  <ulink url="https://sourceforge.net/tracker/index.php?func=detail&amp;aid=1860491&amp;group_id=23187&amp;atid=377771">Issue
          1860491</ulink>.</listitem>

          <listitem>It is now possible to have more control over
          whether the parser expand entity references or not by
          using <literal>XMLUnit.setExpandEntityReferences</literal>,
          see <xref linkend="Entity Reference
          Expansion"/>.  <ulink href="https://sourceforge.net/tracker/index.php?func=detail&amp;aid=1877458&amp;group_id=23187&amp;atid=377771">Issue
          1877458</ulink>.</listitem>

          <listitem>New examples have been added:
            <itemizedlist>
              <listitem><literal>RecursiveElementNameAndTextQualifier</literal>
                - a more flexible <literal>ElementQualifier</literal>
                that fills the same need as
                <literal>MultiLevelElementNameAndTextQualifier</literal>
                 See
                 <xref linkend="RecursiveElementNameAndTextQualifier"/>
                 for more details.</listitem>

              <listitem><literal>CaseInsensitiveDifferenceListener</literal>
                a - <literal>DifferenceListener</literal> that ignores
                case when comparing texts.</listitem>

              <listitem><literal>FloatingPointTolerantDifferenceListener</literal>
                a - <literal>DifferenceListener</literal> that tries
                to parse texts as floating point numbers and compares
                them using a configurable tolerance.</listitem>
            </itemizedlist>
          </listitem>
        </itemizedlist>
      </section>

      <section id="Bugfixes 1.2">
        <title>Important Bug Fixes</title>

        <itemizedlist>
          <listitem>
            If XMLUnit couldn't match nodes (i.e. it encountered
            a <literal>CHILD_NODE_NOT_FOUND</literal> kind of
            difference), the XPath expressions of the node details
            have been random.  <ulink
            url="https://sourceforge.net/tracker/index.php?func=detail&amp;aid=1860681&amp;group_id=23187&amp;atid=377768">Issue 1860681</ulink>.
          </listitem>
        </itemizedlist>
      </section>
    </section>

    <section id="Changes 1.3">
      <title>Changes from XMLUnit 1.2 to 1.3</title>

      <section id="Breaking Changes 1.3">
        <title>Breaking Changes</title>

        <!--itemizedlist>
        </itemizedlist-->
      </section>

      <section id="New Features 1.3">
        <title>New Features</title>

        <itemizedlist>
          <listitem>
            If XMLUnit doesn't find a matching Element for a control
            Element, it will match it against the first unmatched test
            Element (if there is one) instead of creating
            a <literal>CHILD_NODE_NOT_FOUND</literal> Difference.
            There now is a new configuration
            option <literal>compareUnmatched</literal> in
            the <literal>XMLUnit</literal> class that can be used to
            turn off this behavior - as a result
            two <literal>CHILD_NODE_NOT_FOUND</literal> Differences
            (one for the unmatched control Element and one for an
            unmatched test Element) will be created instead of a
            single Difference comparing the two likely unrelated
            nodes.  See <xref linkend="Comparison of Unmatched
            Elements"/>.
            <ulink
               url="https://sourceforge.net/tracker/?func=detail&amp;aid=2758280&amp;group_id=23187&amp;atid=377768">Issue 2758280</ulink>.
          </listitem>
        </itemizedlist>
      </section>

      <section id="Bugfixes 1.3">
        <title>Important Bug Fixes</title>

        <itemizedlist>
          <listitem>
            If XMLUnit couldn't match attributes (i.e. it encountered
            a <literal>ATTR_NAME_NOT_FOUND_ID</literal> kind of
            difference), the XPath expressions of the node details
            have been random.  <ulink
            url="https://sourceforge.net/tracker/index.php?func=detail&amp;aid=2386807&amp;group_id=23187&amp;atid=377768">Issue 2386807</ulink>.
          </listitem>
          <listitem>
            In some cases XMLUnit matched test nodes to multiple
            control nodes and then created a "missing child"
            difference for remaining test nodes even though they would
            have been valid targets for control node matches as well.
            <ulink url="https://sourceforge.net/tracker/?func=detail&amp;aid=2807167&amp;group_id=23187&amp;atid=377768">Issue 2807167</ulink>.
          </listitem>
        </itemizedlist>
      </section>
    </section>

    <section id="Changes 1.4">
      <title>Changes from XMLUnit 1.3 to 1.4</title>

      <section id="Breaking Changes 1.4">
        <title>Breaking Changes</title>

        <!--itemizedlist>
        </itemizedlist-->
      </section>

      <section id="New Features 1.4">
        <title>New Features</title>

        <itemizedlist>
          <listitem>
            xsi:type attributes now have their value interpreted as a
            QName and will compare as identical if their namespace URI
            and local names match even if they use different
            prefixes. <ulink url="https://sourceforge.net/tracker/?func=detail&amp;aid=3602981&amp;group_id=23187&amp;atid=377771">Issue 3602981</ulink>
          </listitem>
        </itemizedlist>
      </section>

      <section id="Bugfixes 1.4">
        <title>Important Bug Fixes</title>

        <itemizedlist>
          <listitem>
            <literal>XMLTestCase</literal>'s and <literal>XMLAssert</literal>'s
            <literal>assertXpathsEqual</literal> methods threw an
            exception when at least one XPath matched an attribute.  <ulink
            url="https://sourceforge.net/tracker/?func=detail&amp;aid=3290264&amp;group_id=23187&amp;atid=377768">Issue 377768</ulink>.
          </listitem>

          <listitem>
            <literal>FloatingPointTolerantDifferenceListener</literal>
            expected numbers to differ by less than the given
            tolerance rather than "less or equal" than as the docs said.
            <ulink
                href="https://sourceforge.net/tracker/index.php?func=detail&amp;aid=3593368&amp;group_id=23187&amp;atid=377768">Issue
            3593368</ulink>
          </listitem>
        </itemizedlist>
      </section>
    </section>

    <section id="Changes 1.5">
      <title>Changes from XMLUnit 1.4 to 1.5</title>

      <section id="Breaking Changes 1.5">
        <title>Breaking Changes</title>

        <itemizedlist>
          <listitem>
            If one node in the comparison has children while the other
            one has not, XMLUnit 1.5 will signal a
            <literal>CHILD_NODELIST_LENGTH</literal> difference and
            <literal>CHILD_NODE_NOT_FOUND</literal>
            differences for each child node of the node that has
            children in addition to a <literal>HAS_CHILD_NODES</literal> difference.
            <ulink
            href="https://sourceforge.net/p/xmlunit/bugs/60/">Issue
            60</ulink>
          </listitem>
        </itemizedlist>
      </section>

      <section id="New Features 1.5">
        <title>New Features</title>

        <itemizedlist>
          <listitem>
          </listitem>
        </itemizedlist>
      </section>

      <section id="Bugfixes 1.5">
        <title>Important Bug Fixes</title>

        <itemizedlist>
          <listitem>
            <literal>RecursiveElementNameAndTextQualifier</literal> had some indices
            reversed leading to wrong results in some cases.
            <ulink
            href="https://sourceforge.net/p/xmlunit/bugs/62/">Issue
            62</ulink>
          </listitem>
        </itemizedlist>
      </section>
    </section>

    <section id="Changes 1.6">
      <title>Changes from XMLUnit 1.5 to 1.6</title>

      <section id="Breaking Changes 1.6">
        <title>Breaking Changes</title>

        <itemizedlist>
          <listitem>
            In cases of <literal>ATTR_NAME_NOT_FOUND</literal> and
            <literal>CHILD_NODE_NOT_FOUND</literal> differences the
            value used to be the local name of the missing attribute
            or node.  It will now be a Java5-QName-like
            <literal>{NS-URI}LOCAL-NAME</literal> string if the
            attribute or node belonged to an XML namespace.
            <ulink
            href="https://sourceforge.net/p/xmlunit/bugs/65/">Issue
            65</ulink>
          </listitem>
        </itemizedlist>
      </section>

      <section id="New Features 1.6">
        <title>New Features</title>

        <itemizedlist>
          <listitem>
            New <literal>assertXpathEvaluatesTo</literal>
            overloads in <literal>XMLAssert</literal> and a new
            <literal>QualifiedName</literal> class
            can be used to assert the stringified result of an XPath
            expression is actually a qualified name.
            <ulink
            href="https://sourceforge.net/p/xmlunit/feature-requests/25/">Feature
            Request 25</ulink>
          </listitem>
        </itemizedlist>
      </section>

      <section id="Bugfixes 1.6">
        <title>Important Bug Fixes</title>

        <itemizedlist>
          <listitem>
            The JAXP 1.3 based validator ignored
            <literal>xsi:namespaceLocation</literal> and
            <literal>xsi:noNamespaceLocation</literal> attributes.
            They will now be used if you don't specify any sources at
            all, but are still ignored if you
            specify any schema sources - since this is the way
            <literal>javax.xml.validation</literal> works.
            <ulink
            href="https://sourceforge.net/p/xmlunit/bugs/64/">Issue
            64</ulink>
          </listitem>
          <listitem>
            When an attribute cannot be found (a
            <literal>ATTR_NAME_NOT_FOUND</literal> difference) the
            XPath on the side where the attribute exists will now
            point to the attribute itself rather than its owning element.
            <ulink
            href="https://sourceforge.net/p/xmlunit/feature-requests/33/">Feature
            Request 33</ulink>
          </listitem>
        </itemizedlist>
      </section>
    </section>

  </appendix>

</article>