1 /*
2 * Copyright (c) 2015, Oracle and/or its affiliates. All rights reserved.
3 */
4 /*
5 * Licensed to the Apache Software Foundation (ASF) under one or more
6 * contributor license agreements. See the NOTICE file distributed with
7 * this work for additional information regarding copyright ownership.
8 * The ASF licenses this file to You under the Apache License, Version 2.0
9 * (the "License"); you may not use this file except in compliance with
10 * the License. You may obtain a copy of the License at
11 *
12 * http://www.apache.org/licenses/LICENSE-2.0
13 *
14 * Unless required by applicable law or agreed to in writing, software
15 * distributed under the License is distributed on an "AS IS" BASIS,
16 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
17 * See the License for the specific language governing permissions and
18 * limitations under the License.
19 */
20 package com.sun.org.apache.xml.internal.utils;
21
22 import com.sun.org.apache.xalan.internal.utils.SecuritySupport;
23 import java.io.File;
24
25 import org.w3c.dom.Comment;
26 import org.w3c.dom.Element;
27 import org.w3c.dom.EntityReference;
28 import org.w3c.dom.NamedNodeMap;
29 import org.w3c.dom.Node;
30 import org.w3c.dom.ProcessingInstruction;
31 import org.w3c.dom.Text;
32
33 import org.xml.sax.ContentHandler;
34 import org.xml.sax.Locator;
35 import org.xml.sax.ext.LexicalHandler;
36 import org.xml.sax.helpers.LocatorImpl;
37
38 /**
39 * This class does a pre-order walk of the DOM tree, calling a ContentHandler
40 * interface as it goes.
41 * @xsl.usage advanced
42 */
43
44 public class TreeWalker
45 {
46
47 /** Local reference to a ContentHandler */
48 private ContentHandler m_contentHandler = null;
49
50 // ARGHH!! JAXP Uses Xerces without setting the namespace processing to ON!
51 // DOM2Helper m_dh = new DOM2Helper();
52
53 /** DomHelper for this TreeWalker */
54 protected DOMHelper m_dh;
55
56 /** Locator object for this TreeWalker */
57 private LocatorImpl m_locator = new LocatorImpl();
58
59 /**
60 * Get the ContentHandler used for the tree walk.
61 *
62 * @return the ContentHandler used for the tree walk
63 */
64 public ContentHandler getContentHandler()
65 {
66 return m_contentHandler;
67 }
68
69 /**
70 * Get the ContentHandler used for the tree walk.
71 *
72 * @return the ContentHandler used for the tree walk
73 */
74 public void setContentHandler(ContentHandler ch)
75 {
76 m_contentHandler = ch;
77 }
78
79 /**
80 * Constructor.
81 * @param contentHandler The implementation of the
82 * @param systemId System identifier for the document.
83 * contentHandler operation (toXMLString, digest, ...)
84 */
85 public TreeWalker(ContentHandler contentHandler, DOMHelper dh, String systemId)
86 {
87 this.m_contentHandler = contentHandler;
88 m_contentHandler.setDocumentLocator(m_locator);
89 if (systemId != null) {
90 m_locator.setSystemId(systemId);
91 }
92 m_dh = dh;
93 }
94
95 /**
96 * Constructor.
97 * @param contentHandler The implementation of the
98 * contentHandler operation (toXMLString, digest, ...)
99 */
100 public TreeWalker(ContentHandler contentHandler, DOMHelper dh)
101 {
102 this.m_contentHandler = contentHandler;
103 m_contentHandler.setDocumentLocator(m_locator);
104 m_dh = dh;
105 }
106
107 /**
108 * Constructor.
109 * @param contentHandler The implementation of the
110 * contentHandler operation (toXMLString, digest, ...)
111 */
112 public TreeWalker(ContentHandler contentHandler)
113 {
114 this.m_contentHandler = contentHandler;
115 if (m_contentHandler != null) {
116 m_contentHandler.setDocumentLocator(m_locator);
117 }
118 m_dh = new DOM2Helper();
119 }
120
121 /**
122 * Perform a pre-order traversal non-recursive style.
123 *
124 * Note that TreeWalker assumes that the subtree is intended to represent
125 * a complete (though not necessarily well-formed) document and, during a
126 * traversal, startDocument and endDocument will always be issued to the
127 * SAX listener.
128 *
129 * @param pos Node in the tree where to start traversal
130 *
131 * @throws TransformerException
132 */
133 public void traverse(Node pos) throws org.xml.sax.SAXException
134 {
135 this.m_contentHandler.startDocument();
136
137 traverseFragment(pos);
138
139 this.m_contentHandler.endDocument();
140 }
141
142 /**
143 * Perform a pre-order traversal non-recursive style.
144 *
145 * In contrast to the traverse() method this method will not issue
146 * startDocument() and endDocument() events to the SAX listener.
147 *
148 * @param pos Node in the tree where to start traversal
149 *
150 * @throws TransformerException
151 */
152 public void traverseFragment(Node pos) throws org.xml.sax.SAXException
153 {
154 Node top = pos;
155
156 while (null != pos)
157 {
158 startNode(pos);
159
160 Node nextNode = pos.getFirstChild();
161
162 while (null == nextNode)
163 {
164 endNode(pos);
165
166 if (top.equals(pos))
167 break;
168
169 nextNode = pos.getNextSibling();
170
171 if (null == nextNode)
172 {
173 pos = pos.getParentNode();
174
175 if ((null == pos) || (top.equals(pos)))
176 {
177 if (null != pos)
178 endNode(pos);
179
180 nextNode = null;
181
182 break;
183 }
184 }
185 }
186
187 pos = nextNode;
188 }
189 }
190
191 /**
192 * Perform a pre-order traversal non-recursive style.
193
194 * Note that TreeWalker assumes that the subtree is intended to represent
195 * a complete (though not necessarily well-formed) document and, during a
196 * traversal, startDocument and endDocument will always be issued to the
197 * SAX listener.
198 *
199 * @param pos Node in the tree where to start traversal
200 * @param top Node in the tree where to end traversal
201 *
202 * @throws TransformerException
203 */
204 public void traverse(Node pos, Node top) throws org.xml.sax.SAXException
205 {
206
207 this.m_contentHandler.startDocument();
208
209 while (null != pos)
210 {
211 startNode(pos);
212
213 Node nextNode = pos.getFirstChild();
214
215 while (null == nextNode)
216 {
217 endNode(pos);
218
219 if ((null != top) && top.equals(pos))
220 break;
221
222 nextNode = pos.getNextSibling();
223
224 if (null == nextNode)
225 {
226 pos = pos.getParentNode();
227
228 if ((null == pos) || ((null != top) && top.equals(pos)))
229 {
230 nextNode = null;
231
232 break;
233 }
234 }
235 }
236
237 pos = nextNode;
238 }
239 this.m_contentHandler.endDocument();
240 }
241
242 /** Flag indicating whether following text to be processed is raw text */
243 boolean nextIsRaw = false;
244
245 /**
246 * Optimized dispatch of characters.
247 */
248 private final void dispatachChars(Node node)
249 throws org.xml.sax.SAXException
250 {
251 if(m_contentHandler instanceof com.sun.org.apache.xml.internal.dtm.ref.dom2dtm.DOM2DTM.CharacterNodeHandler)
252 {
253 ((com.sun.org.apache.xml.internal.dtm.ref.dom2dtm.DOM2DTM.CharacterNodeHandler)m_contentHandler).characters(node);
254 }
255 else
256 {
257 String data = ((Text) node).getData();
258 this.m_contentHandler.characters(data.toCharArray(), 0, data.length());
259 }
260 }
261
262 /**
263 * Start processing given node
264 *
265 *
266 * @param node Node to process
267 *
268 * @throws org.xml.sax.SAXException
269 */
270 protected void startNode(Node node) throws org.xml.sax.SAXException
271 {
272
273 if (m_contentHandler instanceof NodeConsumer)
274 {
275 ((NodeConsumer) m_contentHandler).setOriginatingNode(node);
276 }
277
278 if (node instanceof Locator)
279 {
280 Locator loc = (Locator)node;
281 m_locator.setColumnNumber(loc.getColumnNumber());
282 m_locator.setLineNumber(loc.getLineNumber());
283 m_locator.setPublicId(loc.getPublicId());
284 m_locator.setSystemId(loc.getSystemId());
285 }
286 else
287 {
288 m_locator.setColumnNumber(0);
289 m_locator.setLineNumber(0);
290 }
291
292 switch (node.getNodeType())
293 {
294 case Node.COMMENT_NODE :
295 {
296 String data = ((Comment) node).getData();
297
298 if (m_contentHandler instanceof LexicalHandler)
299 {
300 LexicalHandler lh = ((LexicalHandler) this.m_contentHandler);
301
302 lh.comment(data.toCharArray(), 0, data.length());
303 }
304 }
305 break;
306 case Node.DOCUMENT_FRAGMENT_NODE :
307
308 // ??;
309 break;
310 case Node.DOCUMENT_NODE :
311
312 break;
313 case Node.ELEMENT_NODE :
314 NamedNodeMap atts = ((Element) node).getAttributes();
315 int nAttrs = atts.getLength();
316 // System.out.println("TreeWalker#startNode: "+node.getNodeName());
317
318 for (int i = 0; i < nAttrs; i++)
319 {
320 Node attr = atts.item(i);
321 String attrName = attr.getNodeName();
322
323 // System.out.println("TreeWalker#startNode: attr["+i+"] = "+attrName+", "+attr.getNodeValue());
324 if (attrName.equals("xmlns") || attrName.startsWith("xmlns:"))
325 {
326 // System.out.println("TreeWalker#startNode: attr["+i+"] = "+attrName+", "+attr.getNodeValue());
327 int index;
328 // Use "" instead of null, as Xerces likes "" for the
329 // name of the default namespace. Fix attributed
330 // to "Steven Murray" <smurray@ebt.com>.
331 String prefix = (index = attrName.indexOf(":")) < 0
332 ? "" : attrName.substring(index + 1);
333
334 this.m_contentHandler.startPrefixMapping(prefix,
335 attr.getNodeValue());
336 }
337
338 }
339
340 // System.out.println("m_dh.getNamespaceOfNode(node): "+m_dh.getNamespaceOfNode(node));
341 // System.out.println("m_dh.getLocalNameOfNode(node): "+m_dh.getLocalNameOfNode(node));
342 String ns = m_dh.getNamespaceOfNode(node);
343 if(null == ns)
344 ns = "";
345 this.m_contentHandler.startElement(ns,
346 m_dh.getLocalNameOfNode(node),
347 node.getNodeName(),
348 new AttList(atts, m_dh));
349 break;
350 case Node.PROCESSING_INSTRUCTION_NODE :
351 {
352 ProcessingInstruction pi = (ProcessingInstruction) node;
353 String name = pi.getNodeName();
354
355 // String data = pi.getData();
356 if (name.equals("xslt-next-is-raw"))
357 {
358 nextIsRaw = true;
359 }
360 else
361 {
362 this.m_contentHandler.processingInstruction(pi.getNodeName(),
363 pi.getData());
364 }
365 }
366 break;
367 case Node.CDATA_SECTION_NODE :
368 {
369 boolean isLexH = (m_contentHandler instanceof LexicalHandler);
370 LexicalHandler lh = isLexH
371 ? ((LexicalHandler) this.m_contentHandler) : null;
372
373 if (isLexH)
374 {
375 lh.startCDATA();
376 }
377
378 dispatachChars(node);
379
380 {
381 if (isLexH)
382 {
383 lh.endCDATA();
384 }
385 }
386 }
387 break;
388 case Node.TEXT_NODE :
389 {
390 //String data = ((Text) node).getData();
391
392 if (nextIsRaw)
393 {
394 nextIsRaw = false;
395
396 m_contentHandler.processingInstruction(javax.xml.transform.Result.PI_DISABLE_OUTPUT_ESCAPING, "");
397 dispatachChars(node);
398 m_contentHandler.processingInstruction(javax.xml.transform.Result.PI_ENABLE_OUTPUT_ESCAPING, "");
399 }
400 else
401 {
402 dispatachChars(node);
403 }
404 }
405 break;
406 case Node.ENTITY_REFERENCE_NODE :
407 {
408 EntityReference eref = (EntityReference) node;
409
410 if (m_contentHandler instanceof LexicalHandler)
411 {
412 ((LexicalHandler) this.m_contentHandler).startEntity(
413 eref.getNodeName());
414 }
415 else
416 {
417
418 // warning("Can not output entity to a pure SAX ContentHandler");
419 }
420 }
421 break;
422 default :
423 }
424 }
425
426 /**
427 * End processing of given node
428 *
429 *
430 * @param node Node we just finished processing
431 *
432 * @throws org.xml.sax.SAXException
433 */
434 protected void endNode(Node node) throws org.xml.sax.SAXException
435 {
436
437 switch (node.getNodeType())
438 {
439 case Node.DOCUMENT_NODE :
440 break;
441
442 case Node.ELEMENT_NODE :
443 String ns = m_dh.getNamespaceOfNode(node);
444 if(null == ns)
445 ns = "";
446 this.m_contentHandler.endElement(ns,
447 m_dh.getLocalNameOfNode(node),
448 node.getNodeName());
449
450 NamedNodeMap atts = ((Element) node).getAttributes();
451 int nAttrs = atts.getLength();
452
453 for (int i = 0; i < nAttrs; i++)
454 {
455 Node attr = atts.item(i);
456 String attrName = attr.getNodeName();
457
458 if (attrName.equals("xmlns") || attrName.startsWith("xmlns:"))
459 {
460 int index;
461 // Use "" instead of null, as Xerces likes "" for the
462 // name of the default namespace. Fix attributed
463 // to "Steven Murray" <smurray@ebt.com>.
464 String prefix = (index = attrName.indexOf(":")) < 0
465 ? "" : attrName.substring(index + 1);
466
467 this.m_contentHandler.endPrefixMapping(prefix);
468 }
469 }
470 break;
471 case Node.CDATA_SECTION_NODE :
472 break;
473 case Node.ENTITY_REFERENCE_NODE :
474 {
475 EntityReference eref = (EntityReference) node;
476
477 if (m_contentHandler instanceof LexicalHandler)
478 {
479 LexicalHandler lh = ((LexicalHandler) this.m_contentHandler);
480
481 lh.endEntity(eref.getNodeName());
482 }
483 }
484 break;
485 default :
486 }
487 }
488 } //TreeWalker