HTMLBuilder


1   /*
2    * Copyright 1999,2000,2004 The Apache Software Foundation.
3    * 
4    * Licensed under the Apache License, Version 2.0 (the "License");
5    * you may not use this file except in compliance with the License.
6    * You may obtain a copy of the License at
7    * 
8    *      http://www.apache.org/licenses/LICENSE-2.0
9    * 
10   * Unless required by applicable law or agreed to in writing, software
11   * distributed under the License is distributed on an "AS IS" BASIS,
12   * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13   * See the License for the specific language governing permissions and
14   * limitations under the License.
15   */
16  package org.apache.html.dom;
17  
18  
19  import java.util.Vector  ;
20  
21  import org.apache.xerces.dom.ElementImpl;
22  import org.apache.xerces.dom.ProcessingInstructionImpl;
23  import org.apache.xerces.dom.TextImpl;
24  import org.w3c.dom.Node  ;
25  import org.w3c.dom.html.HTMLDocument;
26  import org.xml.sax.AttributeList  ;
27  import org.xml.sax.DocumentHandler  ;
28  import org.xml.sax.Locator  ;
29  import org.xml.sax.SAXException  ;
30  
31  
32  /**
33   * This is a SAX document handler that is used to build an HTML document.
34   * It can build a document from any SAX parser, but is specifically tuned
35   * for working with the OpenXML HTML parser.
36   * 
37   * 
38   * @version $Revision: 1.7 $ $Date: 2004/02/24 23:34:00 $
39   * @author <a HREF="mailto:arkin@openxml.org">Assaf Arkin</a>
40   */
41  public class HTMLBuilder
42      implements DocumentHandler  
43  {
44  
45  
46      /**
47       * The document that is being built.
48       */
49      protected HTMLDocumentImpl    _document;
50      
51      
52      /**
53       * The current node in the document into which elements, text and
54       * other nodes will be inserted. This starts as the document iself
55       * and reflects each element that is currently being parsed.
56       */
57      protected ElementImpl        _current;
58      
59      /**
60       * A reference to the current locator, this is generally the parser
61       * itself. The locator is used to locate errors and identify the
62       * source locations of elements.
63       */
64      private Locator           _locator;
65  
66  
67      /**
68       * Applies only to whitespace appearing between element tags in element content,
69       * as per the SAX definition, and true by default.
70       */
71      private boolean         _ignoreWhitespace = true;
72  
73  
74      /**
75       * Indicates whether finished building a document. If so, can start building
76       * another document. Must be initially true to get the first document processed.
77       */
78      private boolean         _done = true;
79  
80  
81      /**    
82       * The document is only created the same time as the document element, however, certain
83       * nodes may precede the document element (comment and PI), and they are accumulated
84       * in this vector.
85       */
86      protected Vector           _preRootNodes;
87  
88      
89      public void startDocument()
90          throws SAXException  
91      {
92          if ( ! _done )
93          throw new SAXException  ( "HTM001 State error: startDocument fired twice on one builder." );
94      _document = null;
95      _done = false;
96      }
97  
98  
99      public void endDocument()
100         throws SAXException  
101     {
102         if ( _document == null )
103             throw new SAXException  ( "HTM002 State error: document never started or missing document element." );
104     if ( _current != null )
105         throw new SAXException  ( "HTM003 State error: document ended before end of document element." );
106         _current = null;
107     _done = true;
108     }
109 
110 
111     public synchronized void startElement( String   tagName, AttributeList   attrList )
112         throws SAXException  
113     {
114         ElementImpl elem;
115         int         i;
116         
117     if ( tagName == null )
118         throw new SAXException  ( "HTM004 Argument 'tagName' is null." );
119 
120     // If this is the root element, this is the time to create a new document,
121     // because only know we know the document element name and namespace URI.
122     if ( _document == null )
123     {
124         // No need to create the element explicitly.
125         _document = new HTMLDocumentImpl();
126         elem = (ElementImpl) _document.getDocumentElement();
127         _current = elem;
128         if ( _current == null )
129         throw new SAXException  ( "HTM005 State error: Document.getDocumentElement returns null." );
130 
131         // Insert nodes (comment and PI) that appear before the root element.
132         if ( _preRootNodes != null )
133         {
134         for ( i = _preRootNodes.size() ; i-- > 0 ; )
135             _document.insertBefore( (Node  ) _preRootNodes.elementAt( i ), elem );
136         _preRootNodes = null;
137         }
138          
139     }
140     else
141     {
142         // This is a state error, indicates that document has been parsed in full,
143         // or that there are two root elements.
144         if ( _current == null )
145         throw new SAXException  ( "HTM006 State error: startElement called after end of document element." );
146         elem = (ElementImpl) _document.createElement( tagName );
147         _current.appendChild( elem );
148         _current = elem;
149     }
150 
151     // Add the attributes (specified and not-specified) to this element.
152         if ( attrList != null )
153         {
154             for ( i = 0 ; i < attrList.getLength() ; ++ i )
155                 elem.setAttribute( attrList.getName( i ), attrList.getValue( i ) );
156         }
157     }
158 
159     
160     public void endElement( String   tagName )
161         throws SAXException  
162     {
163         if ( _current == null )
164             throw new SAXException  ( "HTM007 State error: endElement called with no current node." );
165     if ( ! _current.getNodeName().equalsIgnoreCase( tagName ))
166         throw new SAXException  ( "HTM008 State error: mismatch in closing tag name " + tagName + "\n" + tagName);
167 
168     // Move up to the parent element. When you reach the top (closing the root element).
169     // the parent is document and current is null.
170     if ( _current.getParentNode() == _current.getOwnerDocument() )
171         _current = null;
172     else
173         _current = (ElementImpl) _current.getParentNode();
174     }
175 
176 
177     public void characters( String   text )
178         throws SAXException  
179     {
180     if ( _current == null )
181             throw new SAXException  ( "HTM009 State error: character data found outside of root element." );
182     _current.appendChild( new TextImpl( _document, text ) );
183     }
184 
185     
186     public void characters( char[] text, int start, int length )
187         throws SAXException  
188     {
189     if ( _current == null )
190             throw new SAXException  ( "HTM010 State error: character data found outside of root element." );
191     _current.appendChild( new TextImpl( _document, new String  ( text, start, length ) ) );
192     }
193     
194     
195     public void ignorableWhitespace( char[] text, int start, int length )
196         throws SAXException  
197     {
198         Node      node;
199         
200         if ( ! _ignoreWhitespace )
201         _current.appendChild( new TextImpl( _document, new String  ( text, start, length ) ) );
202      }
203     
204     
205     public void processingInstruction( String   target, String   instruction )
206         throws SAXException  
207     {
208         Node      node;
209         
210     // Processing instruction may appear before the document element (in fact, before the
211     // document has been created, or after the document element has been closed.
212         if ( _current == null && _document == null )
213     {
214         if ( _preRootNodes == null )
215         _preRootNodes = new Vector  ();
216         _preRootNodes.addElement( new ProcessingInstructionImpl( null, target, instruction ) );
217     }
218     else
219         if ( _current == null && _document != null )
220         _document.appendChild( new ProcessingInstructionImpl( _document, target, instruction ) );
221     else
222         _current.appendChild( new ProcessingInstructionImpl( _document, target, instruction ) );
223     }
224     
225     
226     public HTMLDocument getHTMLDocument()
227     {
228         return (HTMLDocument) _document;
229     }
230 
231     
232     public void setDocumentLocator( Locator   locator )
233     {
234         _locator = locator;
235     }
236 
237 
238 }
239
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags