views:

318

answers:

1

At work, we have just migrated an old web-app from struts 1.1 to 1.2.9 (hopefully first steps to moving to 1.3), but we are now having problems with the commons digester. Struts 1.2.9 uses commons-digester 1.6.

When we try to parse one of our XML files we get the exception:

org.xml.sax.SAXParseException: Attribute "" bound to namespace "null" was already specified for element "metric".
    at org.apache.xerces.util.ErrorHandlerWrapper.createSAXParseException(ErrorHandlerWrapper.java:232)
    at org.apache.xerces.util.ErrorHandlerWrapper.fatalError(ErrorHandlerWrapper.java:213)
    at org.apache.xerces.impl.XMLErrorReporter.reportError(XMLErrorReporter.java:385)
    at org.apache.xerces.impl.XMLErrorReporter.reportError(XMLErrorReporter.java:315)
    at org.apache.xerces.impl.dtd.XMLNSDTDValidator.startNamespaceScope(XMLNSDTDValidator.java:242)
    at org.apache.xerces.impl.dtd.XMLDTDValidator.handleStartElement(XMLDTDValidator.java:1980)
    at org.apache.xerces.impl.dtd.XMLDTDValidator.startElement(XMLDTDValidator.java:802)
    at org.apache.xerces.impl.XMLNSDocumentScannerImpl.scanStartElement(XMLNSDocumentScannerImpl.java:313)
    at org.apache.xerces.impl.XMLNSDocumentScannerImpl$NSContentDispatcher.scanRootElementHook(XMLNSDocumentScannerImpl.java:610)
    at org.apache.xerces.impl.XMLDocumentFragmentScannerImpl$FragmentContentDispatcher.dispatch(XMLDocumentFragmentScannerImpl.java:1608)
    at org.apache.xerces.impl.XMLDocumentFragmentScannerImpl.scanDocument(XMLDocumentFragmentScannerImpl.java:346)
    at org.apache.xerces.parsers.DTDConfiguration.parse(DTDConfiguration.java:529)
    at org.apache.xerces.parsers.DTDConfiguration.parse(DTDConfiguration.java:585)
    at org.apache.xerces.parsers.XMLParser.parse(XMLParser.java:152)
    at org.apache.xerces.parsers.AbstractSAXParser.parse(AbstractSAXParser.java:1142)
    at org.apache.commons.digester.Digester.parse(Digester.java:1572)
    at com.foo.ctms.framework.metrics.parser.MetricsXMLParser$InternalDigester.parse(MetricsXMLParser.java:54)
    at com.foo.ctms.framework.metrics.parser.MetricsXMLParser.parse(MetricsXMLParser.java:40)

In investigating this issue I have tried to get a simplest possible case and this is what I currently have:

package com.foo.ctms.framework.metrics.parser;

import org.apache.commons.digester.Digester;
import org.xml.sax.InputSource;
import org.xml.sax.SAXException;

import java.io.IOException;
import java.net.URL;

/**
 * Class that provides methods for parsing metrics definitions defined via XML and populating an object graph of JavaBeans
 * representing the definition.
 * @version $Revision: 41470 $
 */
public final class MetricsXMLParser {
  /**
   * The set of public identifiers, and corresponding resource names, for the versions of the configuration file DTD that we know
   * about.  The key is the name of the resource as in the XMl file, and the value is the location of the resource with respect to
   * the <code>ClassLoader</code> that this code in running in.
   */
  private static final String registrations[] = {"-//Foo Inc.//DTD Portal Metrics 1.0//EN", "metrics.dtd"};

  private MetricsXMLParser() {
  }

  /**
   * Parses a metric definition specified as an <code>InputStream</code>.
   * @param url The metrics definition to parse. Must not be <code>null</code>.
   * @throws IOException  if an I/O error occured while attempting to parse
   * @throws SAXException if an XML parsing error occured
   */
  public static MetricDefinition parse(URL url)
          throws IOException, SAXException {
    InternalDigester digester = new InternalDigester();
    return digester.parse(url);
  }

  private static final class InternalDigester {
    private final Digester digester;

    /**
     * Parses a metric definition specified as an <code>InputStream</code>.
     * @param input The metrics definition to parse. Must not be <code>null</code>.
     * @throws IOException  if an I/O error occured while attempting to parse
     * @throws SAXException if an XML parsing error occured
     */
    public MetricDefinition parse(URL input)
            throws IOException, SAXException {
      return (MetricDefinition)digester.parse(new InputSource(input.toString()));
    }

    private InternalDigester() {
      digester = new Digester();
      digester.setValidating(true);
      for (int i = 0; i < MetricsXMLParser.registrations.length; i += 2) {
        URL url = getClass().getResource(MetricsXMLParser.registrations[i + 1]);
        if (url != null) {
          digester.register(MetricsXMLParser.registrations[i], url.toString());
        }
      }

      digester.addObjectCreate("metric", MetricDefinition.class);
      digester.addSetProperties("metric");
    }
  }
}

which is given the XML:

<?xml version='1.0' encoding='windows-1252'?>
<!DOCTYPE metric PUBLIC "-//Foo Inc.//DTD Portal Metrics 1.0//EN" "metrics.dtd">

<metric name="metricsConfig" defaultView="trials">

</metric>

The DTD is currently down to:

<!-- A metric element is the document root -->
<!ELEMENT metric ANY>

<!-- A metric has a name and a default view.  The default view must
     exactly match the name of one of the nested views -->
<!ATTLIST metric
    name CDATA #REQUIRED
    defaultView CDATA #IMPLIED
>

Does anyone know what I am doing wrong?

If I remove the defaultView attribute I do not get the error.


As per sfussenegger's suggestion I've now tried the following (non-digester) code:

try {
  SAXParserFactory factory = SAXParserFactory.newInstance();
  factory.setNamespaceAware(false);
  factory.setValidating(true);
  SAXParser parser = factory.newSAXParser();
  XMLReader reader = parser.getXMLReader();
  reader.parse(new InputSource(url.toString()));
} catch (Exception e) {
  e.printStackTrace();
}

and couldn't reproduce the problem, but adding the following (which commons-digester also does in XercesParser) prior to using my factory gives the same exception:

factory.setFeature("http://apache.org/xml/features/validation/dynamic", true);
factory.setFeature("http://apache.org/xml/features/validation/schema", true);

In the end we have decided to try out a more modern version of xerces (2.7.1) and that seems to work.

A: 

Problem appears to have been caused by an incompatibility between commons-digester 1.6 and xerces (call in XercesParser says 2.2), moving to a more recent version of xerces (2.7.1) has fixed the problem.

We chose this version of the library because other web applications we have are already using it.

Steve Bosman