001 /*
002 * Licensed to the Apache Software Foundation (ASF) under one
003 * or more contributor license agreements. See the NOTICE file
004 * distributed with this work for additional information
005 * regarding copyright ownership. The ASF licenses this file
006 * to you under the Apache License, Version 2.0 (the "License");
007 * you may not use this file except in compliance with the License.
008 * You may obtain a copy of the License at
009 *
010 * http://www.apache.org/licenses/LICENSE-2.0
011 *
012 * Unless required by applicable law or agreed to in writing, software
013 * distributed under the License is distributed on an "AS IS" BASIS,
014 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
015 * See the License for the specific language governing permissions and
016 * limitations under the License.
017 */
018 /*
019 * $Id: TreeWalker.java 468654 2006-10-28 07:09:23Z minchau $
020 */
021 package org.apache.xml.serializer;
022
023 import java.io.File;
024
025 import org.apache.xml.serializer.utils.AttList;
026 import org.apache.xml.serializer.utils.DOM2Helper;
027 import org.w3c.dom.Comment;
028 import org.w3c.dom.Element;
029 import org.w3c.dom.EntityReference;
030 import org.w3c.dom.NamedNodeMap;
031 import org.w3c.dom.Node;
032 import org.w3c.dom.ProcessingInstruction;
033 import org.w3c.dom.Text;
034 import org.xml.sax.ContentHandler;
035 import org.xml.sax.Locator;
036 import org.xml.sax.ext.LexicalHandler;
037 import org.xml.sax.helpers.LocatorImpl;
038
039
040 /**
041 * This class does a pre-order walk of the DOM tree, calling a ContentHandler
042 * interface as it goes.
043 *
044 * This class is a copy of the one in org.apache.xml.utils.
045 * It exists to cut the serializers dependancy on that package.
046 *
047 * @xsl.usage internal
048 */
049
050 public final class TreeWalker
051 {
052
053 /** Local reference to a ContentHandler */
054 final private ContentHandler m_contentHandler;
055 /**
056 * If m_contentHandler is a SerializationHandler, then this is
057 * a reference to the same object.
058 */
059 final private SerializationHandler m_Serializer;
060
061 // ARGHH!! JAXP Uses Xerces without setting the namespace processing to ON!
062 // DOM2Helper m_dh = new DOM2Helper();
063
064 /** DomHelper for this TreeWalker */
065 final protected DOM2Helper m_dh;
066
067 /** Locator object for this TreeWalker */
068 final private LocatorImpl m_locator = new LocatorImpl();
069
070 /**
071 * Get the ContentHandler used for the tree walk.
072 *
073 * @return the ContentHandler used for the tree walk
074 */
075 public ContentHandler getContentHandler()
076 {
077 return m_contentHandler;
078 }
079
080 public TreeWalker(ContentHandler ch) {
081 this(ch,null);
082 }
083 /**
084 * Constructor.
085 * @param contentHandler The implemention of the
086 * contentHandler operation (toXMLString, digest, ...)
087 */
088 public TreeWalker(ContentHandler contentHandler, String systemId)
089 {
090 // Set the content handler
091 m_contentHandler = contentHandler;
092 if (m_contentHandler instanceof SerializationHandler) {
093 m_Serializer = (SerializationHandler) m_contentHandler;
094 }
095 else
096 m_Serializer = null;
097
098 // Set the system ID, if it is given
099 m_contentHandler.setDocumentLocator(m_locator);
100 if (systemId != null)
101 m_locator.setSystemId(systemId);
102 else {
103 try {
104 // Bug see Bugzilla 26741
105 m_locator.setSystemId(System.getProperty("user.dir") + File.separator + "dummy.xsl");
106 }
107 catch (SecurityException se) {// user.dir not accessible from applet
108 }
109 }
110
111 // Set the document locator
112 if (m_contentHandler != null)
113 m_contentHandler.setDocumentLocator(m_locator);
114 try {
115 // Bug see Bugzilla 26741
116 m_locator.setSystemId(System.getProperty("user.dir") + File.separator + "dummy.xsl");
117 }
118 catch (SecurityException se){// user.dir not accessible from applet
119
120 }
121 m_dh = new DOM2Helper();
122 }
123
124 /**
125 * Perform a pre-order traversal non-recursive style.
126 *
127 * Note that TreeWalker assumes that the subtree is intended to represent
128 * a complete (though not necessarily well-formed) document and, during a
129 * traversal, startDocument and endDocument will always be issued to the
130 * SAX listener.
131 *
132 * @param pos Node in the tree where to start traversal
133 *
134 * @throws TransformerException
135 */
136 public void traverse(Node pos) throws org.xml.sax.SAXException
137 {
138
139 this.m_contentHandler.startDocument();
140
141 Node top = pos;
142
143 while (null != pos)
144 {
145 startNode(pos);
146
147 Node nextNode = pos.getFirstChild();
148
149 while (null == nextNode)
150 {
151 endNode(pos);
152
153 if (top.equals(pos))
154 break;
155
156 nextNode = pos.getNextSibling();
157
158 if (null == nextNode)
159 {
160 pos = pos.getParentNode();
161
162 if ((null == pos) || (top.equals(pos)))
163 {
164 if (null != pos)
165 endNode(pos);
166
167 nextNode = null;
168
169 break;
170 }
171 }
172 }
173
174 pos = nextNode;
175 }
176 this.m_contentHandler.endDocument();
177 }
178
179 /**
180 * Perform a pre-order traversal non-recursive style.
181
182 * Note that TreeWalker assumes that the subtree is intended to represent
183 * a complete (though not necessarily well-formed) document and, during a
184 * traversal, startDocument and endDocument will always be issued to the
185 * SAX listener.
186 *
187 * @param pos Node in the tree where to start traversal
188 * @param top Node in the tree where to end traversal
189 *
190 * @throws TransformerException
191 */
192 public void traverse(Node pos, Node top) throws org.xml.sax.SAXException
193 {
194
195 this.m_contentHandler.startDocument();
196
197 while (null != pos)
198 {
199 startNode(pos);
200
201 Node nextNode = pos.getFirstChild();
202
203 while (null == nextNode)
204 {
205 endNode(pos);
206
207 if ((null != top) && top.equals(pos))
208 break;
209
210 nextNode = pos.getNextSibling();
211
212 if (null == nextNode)
213 {
214 pos = pos.getParentNode();
215
216 if ((null == pos) || ((null != top) && top.equals(pos)))
217 {
218 nextNode = null;
219
220 break;
221 }
222 }
223 }
224
225 pos = nextNode;
226 }
227 this.m_contentHandler.endDocument();
228 }
229
230 /** Flag indicating whether following text to be processed is raw text */
231 boolean nextIsRaw = false;
232
233 /**
234 * Optimized dispatch of characters.
235 */
236 private final void dispatachChars(Node node)
237 throws org.xml.sax.SAXException
238 {
239 if(m_Serializer != null)
240 {
241 this.m_Serializer.characters(node);
242 }
243 else
244 {
245 String data = ((Text) node).getData();
246 this.m_contentHandler.characters(data.toCharArray(), 0, data.length());
247 }
248 }
249
250 /**
251 * Start processing given node
252 *
253 *
254 * @param node Node to process
255 *
256 * @throws org.xml.sax.SAXException
257 */
258 protected void startNode(Node node) throws org.xml.sax.SAXException
259 {
260
261 // TODO: <REVIEW>
262 // A Serializer implements ContentHandler, but not NodeConsumer
263 // so drop this reference to NodeConsumer which would otherwise
264 // pull in all sorts of things
265 // if (m_contentHandler instanceof NodeConsumer)
266 // {
267 // ((NodeConsumer) m_contentHandler).setOriginatingNode(node);
268 // }
269 // TODO: </REVIEW>
270
271 if (node instanceof Locator)
272 {
273 Locator loc = (Locator)node;
274 m_locator.setColumnNumber(loc.getColumnNumber());
275 m_locator.setLineNumber(loc.getLineNumber());
276 m_locator.setPublicId(loc.getPublicId());
277 m_locator.setSystemId(loc.getSystemId());
278 }
279 else
280 {
281 m_locator.setColumnNumber(0);
282 m_locator.setLineNumber(0);
283 }
284
285 switch (node.getNodeType())
286 {
287 case Node.COMMENT_NODE :
288 {
289 String data = ((Comment) node).getData();
290
291 if (m_contentHandler instanceof LexicalHandler)
292 {
293 LexicalHandler lh = ((LexicalHandler) this.m_contentHandler);
294
295 lh.comment(data.toCharArray(), 0, data.length());
296 }
297 }
298 break;
299 case Node.DOCUMENT_FRAGMENT_NODE :
300
301 // ??;
302 break;
303 case Node.DOCUMENT_NODE :
304
305 break;
306 case Node.ELEMENT_NODE :
307 Element elem_node = (Element) node;
308 {
309 // Make sure the namespace node
310 // for the element itself is declared
311 // to the ContentHandler
312 String uri = elem_node.getNamespaceURI();
313 if (uri != null) {
314 String prefix = elem_node.getPrefix();
315 if (prefix==null)
316 prefix="";
317 this.m_contentHandler.startPrefixMapping(prefix,uri);
318 }
319 }
320 NamedNodeMap atts = elem_node.getAttributes();
321 int nAttrs = atts.getLength();
322 // System.out.println("TreeWalker#startNode: "+node.getNodeName());
323
324
325 // Make sure the namespace node of
326 // each attribute is declared to the ContentHandler
327 for (int i = 0; i < nAttrs; i++)
328 {
329 final Node attr = atts.item(i);
330 final String attrName = attr.getNodeName();
331 final int colon = attrName.indexOf(':');
332 final String prefix;
333
334 // System.out.println("TreeWalker#startNode: attr["+i+"] = "+attrName+", "+attr.getNodeValue());
335 if (attrName.equals("xmlns") || attrName.startsWith("xmlns:"))
336 {
337 // Use "" instead of null, as Xerces likes "" for the
338 // name of the default namespace. Fix attributed
339 // to "Steven Murray" <smurray@ebt.com>.
340 if (colon < 0)
341 prefix = "";
342 else
343 prefix = attrName.substring(colon + 1);
344
345 this.m_contentHandler.startPrefixMapping(prefix,
346 attr.getNodeValue());
347 }
348 else if (colon > 0) {
349 prefix = attrName.substring(0,colon);
350 String uri = attr.getNamespaceURI();
351 if (uri != null)
352 this.m_contentHandler.startPrefixMapping(prefix,uri);
353 }
354 }
355
356 String ns = m_dh.getNamespaceOfNode(node);
357 if(null == ns)
358 ns = "";
359 this.m_contentHandler.startElement(ns,
360 m_dh.getLocalNameOfNode(node),
361 node.getNodeName(),
362 new AttList(atts, m_dh));
363 break;
364 case Node.PROCESSING_INSTRUCTION_NODE :
365 {
366 ProcessingInstruction pi = (ProcessingInstruction) node;
367 String name = pi.getNodeName();
368
369 // String data = pi.getData();
370 if (name.equals("xslt-next-is-raw"))
371 {
372 nextIsRaw = true;
373 }
374 else
375 {
376 this.m_contentHandler.processingInstruction(pi.getNodeName(),
377 pi.getData());
378 }
379 }
380 break;
381 case Node.CDATA_SECTION_NODE :
382 {
383 boolean isLexH = (m_contentHandler instanceof LexicalHandler);
384 LexicalHandler lh = isLexH
385 ? ((LexicalHandler) this.m_contentHandler) : null;
386
387 if (isLexH)
388 {
389 lh.startCDATA();
390 }
391
392 dispatachChars(node);
393
394 {
395 if (isLexH)
396 {
397 lh.endCDATA();
398 }
399 }
400 }
401 break;
402 case Node.TEXT_NODE :
403 {
404 //String data = ((Text) node).getData();
405
406 if (nextIsRaw)
407 {
408 nextIsRaw = false;
409
410 m_contentHandler.processingInstruction(javax.xml.transform.Result.PI_DISABLE_OUTPUT_ESCAPING, "");
411 dispatachChars(node);
412 m_contentHandler.processingInstruction(javax.xml.transform.Result.PI_ENABLE_OUTPUT_ESCAPING, "");
413 }
414 else
415 {
416 dispatachChars(node);
417 }
418 }
419 break;
420 case Node.ENTITY_REFERENCE_NODE :
421 {
422 EntityReference eref = (EntityReference) node;
423
424 if (m_contentHandler instanceof LexicalHandler)
425 {
426 ((LexicalHandler) this.m_contentHandler).startEntity(
427 eref.getNodeName());
428 }
429 else
430 {
431
432 // warning("Can not output entity to a pure SAX ContentHandler");
433 }
434 }
435 break;
436 default :
437 }
438 }
439
440 /**
441 * End processing of given node
442 *
443 *
444 * @param node Node we just finished processing
445 *
446 * @throws org.xml.sax.SAXException
447 */
448 protected void endNode(Node node) throws org.xml.sax.SAXException
449 {
450
451 switch (node.getNodeType())
452 {
453 case Node.DOCUMENT_NODE :
454 break;
455
456 case Node.ELEMENT_NODE :
457 String ns = m_dh.getNamespaceOfNode(node);
458 if(null == ns)
459 ns = "";
460 this.m_contentHandler.endElement(ns,
461 m_dh.getLocalNameOfNode(node),
462 node.getNodeName());
463
464 if (m_Serializer == null) {
465 // Don't bother with endPrefixMapping calls if the ContentHandler is a
466 // SerializationHandler because SerializationHandler's ignore the
467 // endPrefixMapping() calls anyways. . . . This is an optimization.
468 Element elem_node = (Element) node;
469 NamedNodeMap atts = elem_node.getAttributes();
470 int nAttrs = atts.getLength();
471
472 // do the endPrefixMapping calls in reverse order
473 // of the startPrefixMapping calls
474 for (int i = (nAttrs-1); 0 <= i; i--)
475 {
476 final Node attr = atts.item(i);
477 final String attrName = attr.getNodeName();
478 final int colon = attrName.indexOf(':');
479 final String prefix;
480
481 if (attrName.equals("xmlns") || attrName.startsWith("xmlns:"))
482 {
483 // Use "" instead of null, as Xerces likes "" for the
484 // name of the default namespace. Fix attributed
485 // to "Steven Murray" <smurray@ebt.com>.
486 if (colon < 0)
487 prefix = "";
488 else
489 prefix = attrName.substring(colon + 1);
490
491 this.m_contentHandler.endPrefixMapping(prefix);
492 }
493 else if (colon > 0) {
494 prefix = attrName.substring(0, colon);
495 this.m_contentHandler.endPrefixMapping(prefix);
496 }
497 }
498 {
499 String uri = elem_node.getNamespaceURI();
500 if (uri != null) {
501 String prefix = elem_node.getPrefix();
502 if (prefix==null)
503 prefix="";
504 this.m_contentHandler.endPrefixMapping(prefix);
505 }
506 }
507 }
508 break;
509 case Node.CDATA_SECTION_NODE :
510 break;
511 case Node.ENTITY_REFERENCE_NODE :
512 {
513 EntityReference eref = (EntityReference) node;
514
515 if (m_contentHandler instanceof LexicalHandler)
516 {
517 LexicalHandler lh = ((LexicalHandler) this.m_contentHandler);
518
519 lh.endEntity(eref.getNodeName());
520 }
521 }
522 break;
523 default :
524 }
525 }
526 } //TreeWalker
527