1 /*
2 * Copyright (c) 2015, 2017, 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
21 package com.sun.org.apache.xml.internal.utils;
22
23 import com.sun.org.apache.xml.internal.dtm.ref.dom2dtm.DOM2DTM.CharacterNodeHandler;
24 import javax.xml.transform.Result;
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 /** Locator object for this TreeWalker */
51 private LocatorImpl m_locator = new LocatorImpl();
52
53 /**
54 * Get the ContentHandler used for the tree walk.
55 *
56 * @return the ContentHandler used for the tree walk
57 */
58 public ContentHandler getContentHandler()
59 {
60 return m_contentHandler;
61 }
62
63 /**
64 * Get the ContentHandler used for the tree walk.
65 *
66 * @return the ContentHandler used for the tree walk
67 */
68 public void setContentHandler(ContentHandler ch)
69 {
70 m_contentHandler = ch;
71 }
72
73 /**
74 * Constructor.
75 * @param contentHandler The implementation of the
76 * @param systemId System identifier for the document.
77 * contentHandler operation (toXMLString, digest, ...)
78 */
79 public TreeWalker(ContentHandler contentHandler, String systemId)
80 {
81 this.m_contentHandler = contentHandler;
82 if (m_contentHandler != null) {
83 m_contentHandler.setDocumentLocator(m_locator);
84 }
85 if (systemId != null) {
86 m_locator.setSystemId(systemId);
87 }
88 }
89
90 /**
91 * Constructor.
92 * @param contentHandler The implementation of the
93 * contentHandler operation (toXMLString, digest, ...)
94 */
95 public TreeWalker(ContentHandler contentHandler)
96 {
97 this(contentHandler, null);
98 }
99
100 /**
101 * Perform a pre-order traversal non-recursive style.
102 *
103 * Note that TreeWalker assumes that the subtree is intended to represent
104 * a complete (though not necessarily well-formed) document and, during a
105 * traversal, startDocument and endDocument will always be issued to the
106 * SAX listener.
107 *
108 * @param pos Node in the tree where to start traversal
109 *
110 * @throws TransformerException
111 */
112 public void traverse(Node pos) throws org.xml.sax.SAXException
113 {
114 this.m_contentHandler.startDocument();
115
116 traverseFragment(pos);
117
118 this.m_contentHandler.endDocument();
119 }
120
121 /**
122 * Perform a pre-order traversal non-recursive style.
123 *
124 * In contrast to the traverse() method this method will not issue
125 * startDocument() and endDocument() events to the SAX listener.
126 *
127 * @param pos Node in the tree where to start traversal
128 *
129 * @throws TransformerException
130 */
131 public void traverseFragment(Node pos) throws org.xml.sax.SAXException
132 {
133 Node top = pos;
134
135 while (null != pos)
136 {
137 startNode(pos);
138
139 Node nextNode = pos.getFirstChild();
140
141 while (null == nextNode)
142 {
143 endNode(pos);
144
145 if (top.equals(pos))
146 break;
147
148 nextNode = pos.getNextSibling();
149
150 if (null == nextNode)
151 {
152 pos = pos.getParentNode();
153
154 if ((null == pos) || (top.equals(pos)))
155 {
156 if (null != pos)
157 endNode(pos);
158
159 nextNode = null;
160
161 break;
162 }
163 }
164 }
165
166 pos = nextNode;
167 }
168 }
169
170 /**
171 * Perform a pre-order traversal non-recursive style.
172
173 * Note that TreeWalker assumes that the subtree is intended to represent
174 * a complete (though not necessarily well-formed) document and, during a
175 * traversal, startDocument and endDocument will always be issued to the
176 * SAX listener.
177 *
178 * @param pos Node in the tree where to start traversal
179 * @param top Node in the tree where to end traversal
180 *
181 * @throws TransformerException
182 */
183 public void traverse(Node pos, Node top) throws org.xml.sax.SAXException
184 {
185
186 this.m_contentHandler.startDocument();
187
188 while (null != pos)
189 {
190 startNode(pos);
191
192 Node nextNode = pos.getFirstChild();
193
194 while (null == nextNode)
195 {
196 endNode(pos);
197
198 if ((null != top) && top.equals(pos))
199 break;
200
201 nextNode = pos.getNextSibling();
202
203 if (null == nextNode)
204 {
205 pos = pos.getParentNode();
206
207 if ((null == pos) || ((null != top) && top.equals(pos)))
208 {
209 nextNode = null;
210
211 break;
212 }
213 }
214 }
215
216 pos = nextNode;
217 }
218 this.m_contentHandler.endDocument();
219 }
220
221 // Flag indicating whether following text to be processed is raw text
222 boolean nextIsRaw = false;
223
224 /**
225 * Optimized dispatch of characters.
226 */
227 private final void dispatachChars(Node node)
228 throws org.xml.sax.SAXException
229 {
230 if(m_contentHandler instanceof CharacterNodeHandler)
231 {
232 ((CharacterNodeHandler)m_contentHandler).characters(node);
233 }
234 else
235 {
236 String data = ((Text) node).getData();
237 this.m_contentHandler.characters(data.toCharArray(), 0, data.length());
238 }
239 }
240
241 /**
242 * Start processing given node
243 *
244 *
245 * @param node Node to process
246 *
247 * @throws org.xml.sax.SAXException
248 */
249 protected void startNode(Node node) throws org.xml.sax.SAXException
250 {
251
252 if (m_contentHandler instanceof NodeConsumer)
253 {
254 ((NodeConsumer) m_contentHandler).setOriginatingNode(node);
255 }
256
257 if (node instanceof Locator)
258 {
259 Locator loc = (Locator)node;
260 m_locator.setColumnNumber(loc.getColumnNumber());
261 m_locator.setLineNumber(loc.getLineNumber());
262 m_locator.setPublicId(loc.getPublicId());
263 m_locator.setSystemId(loc.getSystemId());
264 }
265 else
266 {
267 m_locator.setColumnNumber(0);
268 m_locator.setLineNumber(0);
269 }
270
271 switch (node.getNodeType())
272 {
273 case Node.COMMENT_NODE :
274 {
275 String data = ((Comment) node).getData();
276
277 if (m_contentHandler instanceof LexicalHandler)
278 {
279 LexicalHandler lh = ((LexicalHandler) this.m_contentHandler);
280
281 lh.comment(data.toCharArray(), 0, data.length());
282 }
283 }
284 break;
285 case Node.DOCUMENT_FRAGMENT_NODE :
286
287 // ??;
288 break;
289 case Node.DOCUMENT_NODE :
290
291 break;
292 case Node.ELEMENT_NODE :
293 NamedNodeMap atts = ((Element) node).getAttributes();
294 int nAttrs = atts.getLength();
295
296 for (int i = 0; i < nAttrs; i++)
297 {
298 Node attr = atts.item(i);
299 String attrName = attr.getNodeName();
300
301 if (attrName.equals("xmlns") || attrName.startsWith("xmlns:"))
302 {
303 int index;
304 // Use "" instead of null, as Xerces likes "" for the
305 // name of the default namespace. Fix attributed
306 // to "Steven Murray" <smurray@ebt.com>.
307 String prefix = (index = attrName.indexOf(":")) < 0
308 ? "" : attrName.substring(index + 1);
309
310 this.m_contentHandler.startPrefixMapping(prefix,
311 attr.getNodeValue());
312 }
313
314 }
315
316 String ns = DOM2Helper.getNamespaceOfNode(node);
317 if(null == ns)
318 ns = "";
319 this.m_contentHandler.startElement(ns,
320 DOM2Helper.getLocalNameOfNode(node),
321 node.getNodeName(),
322 new AttList(atts));
323 break;
324 case Node.PROCESSING_INSTRUCTION_NODE :
325 {
326 ProcessingInstruction pi = (ProcessingInstruction) node;
327 String name = pi.getNodeName();
328
329 // String data = pi.getData();
330 if (name.equals("xslt-next-is-raw"))
331 {
332 nextIsRaw = true;
333 }
334 else
335 {
336 this.m_contentHandler.processingInstruction(pi.getNodeName(),
337 pi.getData());
338 }
339 }
340 break;
341 case Node.CDATA_SECTION_NODE :
342 {
343 boolean isLexH = (m_contentHandler instanceof LexicalHandler);
344 LexicalHandler lh = isLexH
345 ? ((LexicalHandler) this.m_contentHandler) : null;
346
347 if (isLexH)
348 {
349 lh.startCDATA();
350 }
351
352 dispatachChars(node);
353
354 {
355 if (isLexH)
356 {
357 lh.endCDATA();
358 }
359 }
360 }
361 break;
362 case Node.TEXT_NODE :
363 {
364 //String data = ((Text) node).getData();
365
366 if (nextIsRaw)
367 {
368 nextIsRaw = false;
369
370 m_contentHandler.processingInstruction(Result.PI_DISABLE_OUTPUT_ESCAPING, "");
371 dispatachChars(node);
372 m_contentHandler.processingInstruction(Result.PI_ENABLE_OUTPUT_ESCAPING, "");
373 }
374 else
375 {
376 dispatachChars(node);
377 }
378 }
379 break;
380 case Node.ENTITY_REFERENCE_NODE :
381 {
382 EntityReference eref = (EntityReference) node;
383
384 if (m_contentHandler instanceof LexicalHandler)
385 {
386 ((LexicalHandler) this.m_contentHandler).startEntity(
387 eref.getNodeName());
388 }
389 else
390 {
391
392 // warning("Can not output entity to a pure SAX ContentHandler");
393 }
394 }
395 break;
396 default :
397 }
398 }
399
400 /**
401 * End processing of given node
402 *
403 *
404 * @param node Node we just finished processing
405 *
406 * @throws org.xml.sax.SAXException
407 */
408 protected void endNode(Node node) throws org.xml.sax.SAXException
409 {
410
411 switch (node.getNodeType())
412 {
413 case Node.DOCUMENT_NODE :
414 break;
415
416 case Node.ELEMENT_NODE :
417 String ns = DOM2Helper.getNamespaceOfNode(node);
418 if(null == ns)
419 ns = "";
420 this.m_contentHandler.endElement(ns,
421 DOM2Helper.getLocalNameOfNode(node),
422 node.getNodeName());
423
424 NamedNodeMap atts = ((Element) node).getAttributes();
425 int nAttrs = atts.getLength();
426
427 for (int i = 0; i < nAttrs; i++)
428 {
429 Node attr = atts.item(i);
430 String attrName = attr.getNodeName();
431
432 if (attrName.equals("xmlns") || attrName.startsWith("xmlns:"))
433 {
434 int index;
435 // Use "" instead of null, as Xerces likes "" for the
436 // name of the default namespace. Fix attributed
437 // to "Steven Murray" <smurray@ebt.com>.
438 String prefix = (index = attrName.indexOf(":")) < 0
439 ? "" : attrName.substring(index + 1);
440
441 this.m_contentHandler.endPrefixMapping(prefix);
442 }
443 }
444 break;
445 case Node.CDATA_SECTION_NODE :
446 break;
447 case Node.ENTITY_REFERENCE_NODE :
448 {
449 EntityReference eref = (EntityReference) node;
450
451 if (m_contentHandler instanceof LexicalHandler)
452 {
453 LexicalHandler lh = ((LexicalHandler) this.m_contentHandler);
454
455 lh.endEntity(eref.getNodeName());
456 }
457 }
458 break;
459 default :
460 }
461 }
462 } //TreeWalker