HTMLBuilder


1   /*
2    * The Apache Software License, Version 1.1
3    *
4    *
5    * Copyright (c) 1999,2000 The Apache Software Foundation.  All rights 
6    * reserved.
7    *
8    * Redistribution and use in source and binary forms, with or without
9    * modification, are permitted provided that the following conditions
10   * are met:
11   *
12   * 1. Redistributions of source code must retain the above copyright
13   *    notice, this list of conditions and the following disclaimer. 
14   *
15   * 2. Redistributions in binary form must reproduce the above copyright
16   *    notice, this list of conditions and the following disclaimer in
17   *    the documentation and/or other materials provided with the
18   *    distribution.
19   *
20   * 3. The end-user documentation included with the redistribution,
21   *    if any, must include the following acknowledgment:  
22   *       "This product includes software developed by the
23   *        Apache Software Foundation (http://www.apache.org/)."
24   *    Alternately, this acknowledgment may appear in the software itself,
25   *    if and wherever such third-party acknowledgments normally appear.
26   *
27   * 4. The names "Xerces" and "Apache Software Foundation" must
28   *    not be used to endorse or promote products derived from this
29   *    software without prior written permission. For written 
30   *    permission, please contact apache@apache.org.
31   *
32   * 5. Products derived from this software may not be called "Apache",
33   *    nor may "Apache" appear in their name, without prior written
34   *    permission of the Apache Software Foundation.
35   *
36   * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED
37   * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
38   * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
39   * DISCLAIMED.  IN NO EVENT SHALL THE APACHE SOFTWARE FOUNDATION OR
40   * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
41   * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
42   * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
43   * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
44   * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
45   * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
46   * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
47   * SUCH DAMAGE.
48   * ====================================================================
49   *
50   * This software consists of voluntary contributions made by many
51   * individuals on behalf of the Apache Software Foundation and was
52   * originally based on software copyright (c) 1999, International
53   * Business Machines, Inc., http://www.apache.org.  For more
54   * information on the Apache Software Foundation, please see
55   * <http://www.apache.org/>.
56   */
57  package org.enhydra.apache.html.dom;
58  
59  
60  import java.util.Vector  ;
61  
62  import org.enhydra.apache.xerces.dom.ElementImpl;
63  import org.enhydra.apache.xerces.dom.ProcessingInstructionImpl;
64  import org.enhydra.apache.xerces.dom.TextImpl;
65  import org.w3c.dom.Node  ;
66  import org.w3c.dom.html.HTMLDocument;
67  import org.xml.sax.AttributeList  ;
68  import org.xml.sax.DocumentHandler  ;
69  import org.xml.sax.Locator  ;
70  import org.xml.sax.SAXException  ;
71  
72  
73  /**
74   * This is a SAX document handler that is used to build an HTML document.
75   * It can build a document from any SAX parser, but is specifically tuned
76   * for working with the OpenXML HTML parser.
77   * 
78   * 
79   * @version $Revision: 1.2 $ $Date: 2005/01/26 08:28:44 $
80   * @author <a HREF="mailto:arkin@openxml.org">Assaf Arkin</a>
81   */
82  public class HTMLBuilder
83      implements DocumentHandler  
84  {
85  
86  
87      /**
88       * The document that is being built.
89       */
90      protected HTMLDocumentImpl    _document;
91      
92      
93      /**
94       * The current node in the document into which elements, text and
95       * other nodes will be inserted. This starts as the document iself
96       * and reflects each element that is currently being parsed.
97       */
98      protected ElementImpl        _current;
99      
100     /**
101      * A reference to the current locator, this is generally the parser
102      * itself. The locator is used to locate errors and identify the
103      * source locations of elements.
104      */
105     private Locator           _locator;
106 
107 
108     /**
109      * Applies only to whitespace appearing between element tags in element content,
110      * as per the SAX definition, and true by default.
111      */
112     private boolean         _ignoreWhitespace = true;
113 
114 
115     /**
116      * Indicates whether finished building a document. If so, can start building
117      * another document. Must be initially true to get the first document processed.
118      */
119     private boolean         _done = true;
120 
121 
122     /**    
123      * The document is only created the same time as the document element, however, certain
124      * nodes may precede the document element (comment and PI), and they are accumulated
125      * in this vector.
126      */
127     protected Vector           _preRootNodes;
128 
129     
130     public void startDocument()
131         throws SAXException  
132     {
133         if ( ! _done )
134         throw new SAXException  ( "HTM001 State error: startDocument fired twice on one builder." );
135     _document = null;
136     _done = false;
137     }
138 
139 
140     public void endDocument()
141         throws SAXException  
142     {
143         if ( _document == null )
144             throw new SAXException  ( "HTM002 State error: document never started or missing document element." );
145     if ( _current != null )
146         throw new SAXException  ( "HTM003 State error: document ended before end of document element." );
147         _current = null;
148     _done = true;
149     }
150 
151 
152     public synchronized void startElement( String   tagName, AttributeList   attrList )
153         throws SAXException  
154     {
155         ElementImpl elem;
156         int         i;
157         
158     if ( tagName == null )
159         throw new SAXException  ( "HTM004 Argument 'tagName' is null." );
160 
161     // If this is the root element, this is the time to create a new document,
162     // because only know we know the document element name and namespace URI.
163     if ( _document == null )
164     {
165         // No need to create the element explicitly.
166         _document = new HTMLDocumentImpl();
167         elem = (ElementImpl) _document.getDocumentElement();
168         _current = elem;
169         if ( _current == null )
170         throw new SAXException  ( "HTM005 State error: Document.getDocumentElement returns null." );
171 
172         // Insert nodes (comment and PI) that appear before the root element.
173         if ( _preRootNodes != null )
174         {
175         for ( i = _preRootNodes.size() ; i-- > 0 ; )
176             _document.insertBefore( (Node  ) _preRootNodes.elementAt( i ), elem );
177         _preRootNodes = null;
178         }
179          
180     }
181     else
182     {
183         // This is a state error, indicates that document has been parsed in full,
184         // or that there are two root elements.
185         if ( _current == null )
186         throw new SAXException  ( "HTM006 State error: startElement called after end of document element." );
187         elem = (ElementImpl) _document.createElement( tagName );
188         _current.appendChild( elem );
189         _current = elem;
190     }
191 
192     // Add the attributes (specified and not-specified) to this element.
193         if ( attrList != null )
194         {
195             for ( i = 0 ; i < attrList.getLength() ; ++ i )
196                 elem.setAttribute( attrList.getName( i ), attrList.getValue( i ) );
197         }
198     }
199 
200     
201     public void endElement( String   tagName )
202         throws SAXException  
203     {
204 
205         if ( _current == null )
206             throw new SAXException  ( "HTM007 State error: endElement called with no current node." );
207     if ( ! _current.getNodeName().equals( tagName ) )
208         throw new SAXException  ( "HTM008 State error: mismatch in closing tag name " + tagName + "\n" + tagName);
209 
210     // Move up to the parent element. When you reach the top (closing the root element).
211     // the parent is document and current is null.
212     if ( _current.getParentNode() == _current.getOwnerDocument() )
213         _current = null;
214     else
215         _current = (ElementImpl) _current.getParentNode();
216     }
217 
218 
219     public void characters( String   text )
220         throws SAXException  
221     {
222     if ( _current == null )
223             throw new SAXException  ( "HTM009 State error: character data found outside of root element." );
224     _current.appendChild( new TextImpl( _document, text ) );
225     }
226 
227     
228     public void characters( char[] text, int start, int length )
229         throws SAXException  
230     {
231     if ( _current == null )
232             throw new SAXException  ( "HTM010 State error: character data found outside of root element." );
233     _current.appendChild( new TextImpl( _document, new String  ( text, start, length ) ) );
234     }
235     
236     
237     public void ignorableWhitespace( char[] text, int start, int length )
238         throws SAXException  
239     {
240         Node      node;
241         
242         if ( ! _ignoreWhitespace )
243         _current.appendChild( new TextImpl( _document, new String  ( text, start, length ) ) );
244      }
245     
246     
247     public void processingInstruction( String   target, String   instruction )
248         throws SAXException  
249     {
250         Node      node;
251         
252     // Processing instruction may appear before the document element (in fact, before the
253     // document has been created, or after the document element has been closed.
254         if ( _current == null && _document == null )
255     {
256         if ( _preRootNodes == null )
257         _preRootNodes = new Vector  ();
258         _preRootNodes.addElement( new ProcessingInstructionImpl( null, target, instruction ) );
259     }
260     else
261         if ( _current == null && _document != null )
262         _document.appendChild( new ProcessingInstructionImpl( _document, target, instruction ) );
263     else
264         _current.appendChild( new ProcessingInstructionImpl( _document, target, instruction ) );
265     }
266     
267     
268     public HTMLDocument getHTMLDocument()
269     {
270         return (HTMLDocument) _document;
271     }
272 
273     
274     public void setDocumentLocator( Locator   locator )
275     {
276         _locator = locator;
277     }
278 
279 
280 }
281
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags