saxdriver.java

好东西啊！你看看就知道了

/*
 * Copyright (c) 1999-2000 by David Brownell.  All Rights Reserved.
 *
 * This program is open source software; you may use, copy, modify, and
 * redistribute it under the terms of the LICENSE with which it was
 * originally distributed.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * LICENSE for more details.
 */

//
// Copyright (c) 1998 by Microstar Software Ltd.
// From Microstar's README (the entire original license):
//
// AElfred is free for both commercial and non-commercial use and
// redistribution, provided that Microstar's copyright and disclaimer are
// retained intact.  You are free to modify AElfred for your own use and
// to redistribute AElfred with your modifications, provided that the
// modifications are clearly documented.
//
// This program is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// merchantability or fitness for a particular purpose.  Please use it AT
// YOUR OWN RISK.
//


package org.brownell.xml.aelfred2;

import java.io.InputStreamReader;
import java.io.IOException;
import java.io.Reader;
import java.util.Enumeration;
import java.util.Hashtable;
import java.util.Locale;
import java.util.Stack;
import java.util.Vector;

import org.xml.sax.*;
import org.xml.sax.ext.*;
import org.xml.sax.helpers.NamespaceSupport;

import org.brownell.xml.DefaultHandler;


// $Id: SAXDriver.java,v 1.23 2000/05/29 12:10:24 mojo Exp $

/**
 * An enhanced SAX2 version of Microstar's Ælfred XML parser.
 * The enhancements primarily relate to significant improvements in
 * conformance to the XML specification, and SAX2 support.  Performance
 * has been improved.  However, the Ælfred proprietary APIs are
 * no longer public.  See the package level documentation for more
 * information.
 *
 * <table border="1" width='100%' cellpadding='3' cellspacing='0'>
 * <tr bgcolor='#ccccff'>
 *	<th><font size='+1'>Name</font></th>
 *	<th><font size='+1'>Notes</font></th></tr>
 *
 * <tr><td colspan=2><center><em>Features ... URL prefix is
 * <b>http://xml.org/sax/features/</b></em></center></td></tr>
 *
 * <tr><td>(URL)/external-general-entities</td>
 *	<td>Value is fixed at <em>true</em></td></tr>
 * <tr><td>(URL)/external-parameter-entities</td>
 *	<td>Value is fixed at <em>true</em></td></tr>
 * <tr><td>(URL)/namespace-prefixes</td>
 *	<td>Value defaults to <em>false</em> (but XML 1.0 names are
 *		always reported)</td></tr>
 * <tr><td>(URL)/namespaces</td>
 *	<td>Value defaults to <em>true</em></td></tr>
 * <tr><td>(URL)/string-interning</td>
 *	<td>Value is fixed at <em>true</em></td></tr>
 * <tr><td>(URL)/validation</td>
 *	<td>Value is fixed at <em>false</em></td></tr>
 *
 * <tr><td colspan=2><center><em>Handler Properties ... URL prefix is
 * <b>http://xml.org/sax/properties/</b></em></center></td></tr>
 *
 * <tr><td>(URL)/declaration-handler</td>
 *	<td>A declaration handler may be provided.  Declaration of general
 *	entities is exposed, but not parameter entities; none of the entity
 *	names reported here will begin with "%". </td></tr>
 * <tr><td>(URL)/lexical-handler</td>
 *	<td>A lexical handler may be provided.  Entity boundaries and
 *	comments are not exposed; only CDATA sections and the start/end of
 *	the DTD (the internal subset is not detectible). </td></tr>
 * </table>
 *
 * <p> Note that the declaration handler doesn't suffice for showing all
 * the logical structure
 * of the DTD; it doesn't expose the name of the root element, or the values
 * that are permitted in a NOTATIONS attribute.  (The former is exposed as
 * lexical data, and SAX2 doesn't expose the latter.)
 *
 * <p> Although support for several features and properties is "built in"
 * to this parser, it support all others by storing the assigned values
 * and returning them.
 *
 * <p>This parser currently implements the SAX1 Parser API, but
 * it may not continue to do so in the future.
 *
 * @author Written by David Megginson &lt;[e-mail]&gt;
 *	(version 1.2a from Microstar)
 * @author Updated by David Brownell &lt;[e-mail]&gt;
 * @version $Date: 2000/05/29 12:10:24 $
 * @see org.xml.sax.Parser
 */
final public class SAXDriver
    implements Locator, Attributes, XMLReader, Parser, AttributeList
{
    private final DefaultHandler	base = new DefaultHandler ();
    private XmlParser			parser;

    private EntityResolver		entityResolver = base;
    private ContentHandler		contentHandler = base;
    private DTDHandler			dtdHandler = base;
    private ErrorHandler 		errorHandler = base;
    private DeclHandler			declHandler = base;
    private LexicalHandler		lexicalHandler = base;

    private String			elementName = null;
    private Stack			entityStack = new Stack ();

    private Vector			attributeNames = new Vector ();
    private Vector			attributeNamespaces = new Vector ();
    private Vector			attributeLocalNames = new Vector ();
    private Vector			attributeValues = new Vector ();

    private boolean			namespaces = true;
    private boolean			xmlNames = false;

    private int				attributeCount = 0;
    private boolean			nspending = false;
    private String			nsTemp [] = new String [3];
    private NamespaceSupport		prefixStack = new NamespaceSupport ();

    private Hashtable			features;
    private Hashtable			properties;


    //
    // Constructor.
    //

    /** Constructs a SAX Parser.  */
    public SAXDriver () {}


    //
    // Implementation of org.xml.sax.Parser.
    //

    /**
     * <b>SAX1</b>: Sets the locale used for diagnostics; currently,
     * only locales using the English language are supported.
     * @param locale The locale for which diagnostics will be generated
     */
    public void setLocale (Locale locale)
    throws SAXException
    {
	if ("en".equals (locale.getLanguage ()))
	    return ;

	throw new SAXException ("AElfred only supports English locales.");
    }


    /**
     * <b>SAX2</b>: Returns the object used when resolving external
     * entities during parsing (both general and parameter entities).
     */
    public EntityResolver getEntityResolver ()
    {
	return entityResolver;
    }

    /**
     * <b>SAX1, SAX2</b>: Set the entity resolver for this parser.
     * @param handler The object to receive entity events.
     */
    public void setEntityResolver (EntityResolver resolver)
    {
	if (resolver == null)
	    resolver = base;
	this.entityResolver = resolver;
    }


    /**
     * <b>SAX2</b>: Returns the object used to process declarations related
     * to notations and unparsed entities.
     */
    public DTDHandler getDTDHandler ()
    {
	return dtdHandler;
    }

    /**
     * <b>SAX1, SAX2</b>: Set the DTD handler for this parser.
     * @param handler The object to receive DTD events.
     */
    public void setDTDHandler (DTDHandler handler)
    {
	if (handler == null)
	    handler = base;
	this.dtdHandler = handler;
    }


    /**
     * <b>SAX1</b>: Set the document handler for this parser.  If a
     * content handler was set, this document handler will supplant it.
     * The parser is set to report all XML 1.0 names rather than to
     * filter out "xmlns" attributes (the "namespace-prefixes" feature
     * is set to true).
     *
     * @deprecated SAX2 programs should use the XMLReader interface
     *	and a ContentHandler.
     *
     * @param handler The object to receive document events.
     */
    public void setDocumentHandler (DocumentHandler handler)
    {
	contentHandler = new Adapter (handler);
	xmlNames = true;
    }

    /**
     * <b>SAX2</b>: Returns the object used to report the logical
     * content of an XML document.
     */
    public ContentHandler getContentHandler ()
    {
	return contentHandler;
    }

    /**
     * <b>SAX2</b>: Assigns the object used to report the logical
     * content of an XML document.  If a document handler was set,
     * this content handler will supplant it (but XML 1.0 style name
     * reporting may remain enabled).
     */
    public void setContentHandler (ContentHandler handler)
    {
	if (handler == null)
	    handler = base;
	contentHandler = handler;
    }

    /**
     * <b>SAX1, SAX2</b>: Set the error handler for this parser.
     * @param handler The object to receive error events.
     */
    public void setErrorHandler (ErrorHandler handler)
    {
	if (handler == null)
	    handler = base;
	this.errorHandler = handler;
    }

    /**
     * <b>SAX2</b>: Returns the object used to receive callbacks for XML
     * errors of all levels (fatal, nonfatal, warning); this is never null;
     */
    public ErrorHandler getErrorHandler ()
	{ return errorHandler; }


    /**
     * <b>SAX1, SAX2</b>: Auxiliary API to parse an XML document, used mostly
     * when no URI is available.
     * If you want anything useful to happen, you should set
     * at least one type of handler.
     * @param source The XML input source.  Don't set 'encoding' unless
     *	you know for a fact that it's correct.
     * @see #setEntityResolver
     * @see #setDTDHandler
     * @see #setContentHandler
     * @see #setErrorHandler
     * @exception SAXException The handlers may throw any SAXException,
     *	and the parser normally throws SAXParseException objects.
     * @exception IOException IOExceptions are normally through through
     *	the parser if there are problems reading the source document.
     */
    public void parse (InputSource source)
    throws SAXException, IOException
    {
	synchronized (base) {
	    parser = new XmlParser ();
	    parser.setHandler (this);

	    try {
		String	systemId = source.getSystemId ();

		// duplicate first entry, in case startDocument handler
		// needs to use Locator.getSystemId(), before entities
		// start to get reported by the parser

		if (systemId != null)
		    entityStack.push (systemId);
		else
		    entityStack.push ("illegal:unknown system ID");

		parser.doParse (systemId,
			      source.getPublicId (),
			      source.getCharacterStream (),
			      source.getByteStream (),
			      source.getEncoding ());
	    } catch (SAXException e) {
		throw e;
	    } catch (IOException e) {
		throw e;
	    } catch (RuntimeException e) {
		throw e;
	    } catch (Exception e) {
		throw new SAXException (e.getMessage (), e);
	    } finally {
		contentHandler.endDocument ();
		entityStack.removeAllElements ();
	    }
	}
    }


    /**
     * <b>SAX1, SAX2</b>: Preferred API to parse an XML document, using a
     * system identifier (URI).
     */
    public void parse (String systemId)
    throws SAXException, IOException
    {
	parse (new InputSource (systemId));
    }

    //
    // Implementation of SAX2 "XMLReader" interface
    //
    static final String	FEATURE = "http://xml.org/sax/features/";
    static final String	HANDLER = "http://xml.org/sax/properties/";

    /**
     * <b>SAX2</b>: Tells the value of the specified feature flag.
     *
     * @exception SAXNotRecognizedException thrown if the feature flag
     *	is neither built in, nor yet assigned.
     */
    public boolean getFeature (String featureId)
    throws SAXNotRecognizedException
    {
	if ((FEATURE + "validation").equals (featureId))
	    return false;

	// external entities (both types) are currently always included
	if ((FEATURE + "external-general-entities").equals (featureId)
		|| (FEATURE + "external-parameter-entities")
		.equals (featureId))
	    return true;

	// element/attribute names are as written in document; no mangling
	if ((FEATURE + "namespace-prefixes").equals (featureId))
	    return xmlNames;

	// report element/attribute namespaces?
	if ((FEATURE + "namespaces").equals (featureId))
	    return namespaces;

	// always interns
	if ((FEATURE + "string-interning").equals (featureId))
	    return true;

	if (features != null && features.containsKey (featureId))
	    return ((Boolean)features.get (featureId)).booleanValue ();

	throw new SAXNotRecognizedException (featureId);
    }

    /**
     * <b>SAX2</b>:  Returns the specified property.
     *
     * @exception SAXNotRecognizedException thrown if the property value
     *	is neither built in, nor yet stored.
     */
    public Object getProperty (String propertyId)
    throws SAXNotRecognizedException
    {
	if ((HANDLER + "declaration-handler").equals (propertyId))
	    return declHandler;

	if ((HANDLER + "lexical-handler").equals (propertyId))
	    return lexicalHandler;
	
	if (properties != null && properties.containsKey (propertyId))
	    return properties.get (propertyId);

	// unknown properties
	throw new SAXNotRecognizedException (propertyId);
    }

    /**
     * <b>SAX2</b>:  Sets the state of feature flags in this parser.  Some
     * built-in feature flags are mutable; all flags not built-in are
     * motable.
     */
    public void setFeature (String featureId, boolean state)
    throws SAXNotRecognizedException, SAXNotSupportedException
    {
	boolean	value;
	
	try {
	    // Features with a defined value, we just change it if we can.
	    value = getFeature (featureId);

	    if (state == value)
		return;