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.serializer;
21
22 import com.sun.org.apache.xalan.internal.utils.SecuritySupport;
23 import java.io.File;
24
25 import com.sun.org.apache.xml.internal.serializer.utils.AttList;
26 import com.sun.org.apache.xml.internal.serializer.utils.DOM2Helper;
27 import org.w3c.dom.Comment;
28 import org.w3c.dom.Element;
29 import org.w3c.dom.EntityReference;
30 import org.w3c.dom.NamedNodeMap;
31 import org.w3c.dom.Node;
32 import org.w3c.dom.ProcessingInstruction;
33 import org.w3c.dom.Text;
34
35 import org.xml.sax.ContentHandler;
36 import org.xml.sax.Locator;
37 import org.xml.sax.ext.LexicalHandler;
38 import org.xml.sax.helpers.LocatorImpl;
39
40 /**
41 * This class does a pre-order walk of the DOM tree, calling a ContentHandler
42 * interface as it goes.
43 *
44 * This class is a copy of the one in com.sun.org.apache.xml.internal.utils.
45 * It exists to cut the serializers dependancy on that package.
46 *
47 * @xsl.usage internal
48 */
49
50 public final class TreeWalker
51 {
52
53 /** Local reference to a ContentHandler */
54 final private ContentHandler m_contentHandler;
55 /**
56 * If m_contentHandler is a SerializationHandler, then this is
57 * a reference to the same object.
58 */
59 final private SerializationHandler m_Serializer;
60
61 // ARGHH!! JAXP Uses Xerces without setting the namespace processing to ON!
62 // DOM2Helper m_dh = new DOM2Helper();
63
64 /** DomHelper for this TreeWalker */
65 final protected DOM2Helper m_dh;
66
67 /** Locator object for this TreeWalker */
68 final private LocatorImpl m_locator = new LocatorImpl();
69
70 /**
71 * Get the ContentHandler used for the tree walk.
72 *
73 * @return the ContentHandler used for the tree walk
74 */
75 public ContentHandler getContentHandler()
76 {
77 return m_contentHandler;
78 }
79
80 public TreeWalker(ContentHandler ch) {
81 this(ch,null);
82 }
83 /**
84 * Constructor.
85 * @param contentHandler The implemention of the
86 * contentHandler operation (toXMLString, digest, ...)
87 */
88 public TreeWalker(ContentHandler contentHandler, String systemId)
89 {
90 // Set the content handler
91 m_contentHandler = contentHandler;
92 if (m_contentHandler instanceof SerializationHandler) {
93 m_Serializer = (SerializationHandler) m_contentHandler;
94 } else {
95 m_Serializer = null;
96 }
97 // Set the system ID, if it is given
98 m_contentHandler.setDocumentLocator(m_locator);
99 if (systemId != null) {
100 m_locator.setSystemId(systemId);
101 }
102
103 m_dh = new DOM2Helper();
104 }
105
106 /**
107 * Perform a pre-order traversal non-recursive style.
108 *
109 * Note that TreeWalker assumes that the subtree is intended to represent
110 * a complete (though not necessarily well-formed) document and, during a
111 * traversal, startDocument and endDocument will always be issued to the
112 * SAX listener.
113 *
114 * @param pos Node in the tree where to start traversal
115 *
116 * @throws TransformerException
117 */
118 public void traverse(Node pos) throws org.xml.sax.SAXException
119 {
120
121 this.m_contentHandler.startDocument();
122
123 Node top = pos;
124
125 while (null != pos)
126 {
127 startNode(pos);
128
129 Node nextNode = pos.getFirstChild();
130
131 while (null == nextNode)
132 {
133 endNode(pos);
134
135 if (top.equals(pos))
136 break;
137
138 nextNode = pos.getNextSibling();
139
140 if (null == nextNode)
141 {
142 pos = pos.getParentNode();
143
144 if ((null == pos) || (top.equals(pos)))
145 {
146 if (null != pos)
147 endNode(pos);
148
149 nextNode = null;
150
151 break;
152 }
153 }
154 }
155
156 pos = nextNode;
157 }
158 this.m_contentHandler.endDocument();
159 }
160
161 /**
162 * Perform a pre-order traversal non-recursive style.
163
164 * Note that TreeWalker assumes that the subtree is intended to represent
165 * a complete (though not necessarily well-formed) document and, during a
166 * traversal, startDocument and endDocument will always be issued to the
167 * SAX listener.
168 *
169 * @param pos Node in the tree where to start traversal
170 * @param top Node in the tree where to end traversal
171 *
172 * @throws TransformerException
173 */
174 public void traverse(Node pos, Node top) throws org.xml.sax.SAXException
175 {
176
177 this.m_contentHandler.startDocument();
178
179 while (null != pos)
180 {
181 startNode(pos);
182
183 Node nextNode = pos.getFirstChild();
184
185 while (null == nextNode)
186 {
187 endNode(pos);
188
189 if ((null != top) && top.equals(pos))
190 break;
191
192 nextNode = pos.getNextSibling();
193
194 if (null == nextNode)
195 {
196 pos = pos.getParentNode();
197
198 if ((null == pos) || ((null != top) && top.equals(pos)))
199 {
200 nextNode = null;
201
202 break;
203 }
204 }
205 }
206
207 pos = nextNode;
208 }
209 this.m_contentHandler.endDocument();
210 }
211
212 /** Flag indicating whether following text to be processed is raw text */
213 boolean nextIsRaw = false;
214
215 /**
216 * Optimized dispatch of characters.
217 */
218 private final void dispatachChars(Node node)
219 throws org.xml.sax.SAXException
220 {
221 if(m_Serializer != null)
222 {
223 this.m_Serializer.characters(node);
224 }
225 else
226 {
227 String data = ((Text) node).getData();
228 this.m_contentHandler.characters(data.toCharArray(), 0, data.length());
229 }
230 }
231
232 /**
233 * Start processing given node
234 *
235 *
236 * @param node Node to process
237 *
238 * @throws org.xml.sax.SAXException
239 */
240 protected void startNode(Node node) throws org.xml.sax.SAXException
241 {
242
243 // TODO: <REVIEW>
244 // A Serializer implements ContentHandler, but not NodeConsumer
245 // so drop this reference to NodeConsumer which would otherwise
246 // pull in all sorts of things
247 // if (m_contentHandler instanceof NodeConsumer)
248 // {
249 // ((NodeConsumer) m_contentHandler).setOriginatingNode(node);
250 // }
251 // TODO: </REVIEW>
252
253 if (node instanceof Locator)
254 {
255 Locator loc = (Locator)node;
256 m_locator.setColumnNumber(loc.getColumnNumber());
257 m_locator.setLineNumber(loc.getLineNumber());
258 m_locator.setPublicId(loc.getPublicId());
259 m_locator.setSystemId(loc.getSystemId());
260 }
261 else
262 {
263 m_locator.setColumnNumber(0);
264 m_locator.setLineNumber(0);
265 }
266
267 switch (node.getNodeType())
268 {
269 case Node.COMMENT_NODE :
270 {
271 String data = ((Comment) node).getData();
272
273 if (m_contentHandler instanceof LexicalHandler)
274 {
275 LexicalHandler lh = ((LexicalHandler) this.m_contentHandler);
276
277 lh.comment(data.toCharArray(), 0, data.length());
278 }
279 }
280 break;
281 case Node.DOCUMENT_FRAGMENT_NODE :
282
283 // ??;
284 break;
285 case Node.DOCUMENT_NODE :
286
287 break;
288 case Node.ELEMENT_NODE :
289 Element elem_node = (Element) node;
290 {
291 // Make sure the namespace node
292 // for the element itself is declared
293 // to the ContentHandler
294 String uri = elem_node.getNamespaceURI();
295 if (uri != null) {
296 String prefix = elem_node.getPrefix();
297 if (prefix==null)
298 prefix="";
299 this.m_contentHandler.startPrefixMapping(prefix,uri);
300 }
301 }
302 NamedNodeMap atts = elem_node.getAttributes();
303 int nAttrs = atts.getLength();
304 // System.out.println("TreeWalker#startNode: "+node.getNodeName());
305
306
307 // Make sure the namespace node of
308 // each attribute is declared to the ContentHandler
309 for (int i = 0; i < nAttrs; i++)
310 {
311 final Node attr = atts.item(i);
312 final String attrName = attr.getNodeName();
313 final int colon = attrName.indexOf(':');
314 final String prefix;
315
316 // System.out.println("TreeWalker#startNode: attr["+i+"] = "+attrName+", "+attr.getNodeValue());
317 if (attrName.equals("xmlns") || attrName.startsWith("xmlns:"))
318 {
319 // Use "" instead of null, as Xerces likes "" for the
320 // name of the default namespace. Fix attributed
321 // to "Steven Murray" <smurray@ebt.com>.
322 if (colon < 0)
323 prefix = "";
324 else
325 prefix = attrName.substring(colon + 1);
326
327 this.m_contentHandler.startPrefixMapping(prefix,
328 attr.getNodeValue());
329 }
330 else if (colon > 0) {
331 prefix = attrName.substring(0,colon);
332 String uri = attr.getNamespaceURI();
333 if (uri != null)
334 this.m_contentHandler.startPrefixMapping(prefix,uri);
335 }
336 }
337
338 String ns = m_dh.getNamespaceOfNode(node);
339 if(null == ns)
340 ns = "";
341 this.m_contentHandler.startElement(ns,
342 m_dh.getLocalNameOfNode(node),
343 node.getNodeName(),
344 new AttList(atts, m_dh));
345 break;
346 case Node.PROCESSING_INSTRUCTION_NODE :
347 {
348 ProcessingInstruction pi = (ProcessingInstruction) node;
349 String name = pi.getNodeName();
350
351 // String data = pi.getData();
352 if (name.equals("xslt-next-is-raw"))
353 {
354 nextIsRaw = true;
355 }
356 else
357 {
358 this.m_contentHandler.processingInstruction(pi.getNodeName(),
359 pi.getData());
360 }
361 }
362 break;
363 case Node.CDATA_SECTION_NODE :
364 {
365 boolean isLexH = (m_contentHandler instanceof LexicalHandler);
366 LexicalHandler lh = isLexH
367 ? ((LexicalHandler) this.m_contentHandler) : null;
368
369 if (isLexH)
370 {
371 lh.startCDATA();
372 }
373
374 dispatachChars(node);
375
376 {
377 if (isLexH)
378 {
379 lh.endCDATA();
380 }
381 }
382 }
383 break;
384 case Node.TEXT_NODE :
385 {
386 //String data = ((Text) node).getData();
387
388 if (nextIsRaw)
389 {
390 nextIsRaw = false;
391
392 m_contentHandler.processingInstruction(javax.xml.transform.Result.PI_DISABLE_OUTPUT_ESCAPING, "");
393 dispatachChars(node);
394 m_contentHandler.processingInstruction(javax.xml.transform.Result.PI_ENABLE_OUTPUT_ESCAPING, "");
395 }
396 else
397 {
398 dispatachChars(node);
399 }
400 }
401 break;
402 case Node.ENTITY_REFERENCE_NODE :
403 {
404 EntityReference eref = (EntityReference) node;
405
406 if (m_contentHandler instanceof LexicalHandler)
407 {
408 ((LexicalHandler) this.m_contentHandler).startEntity(
409 eref.getNodeName());
410 }
411 else
412 {
413
414 // warning("Can not output entity to a pure SAX ContentHandler");
415 }
416 }
417 break;
418 default :
419 }
420 }
421
422 /**
423 * End processing of given node
424 *
425 *
426 * @param node Node we just finished processing
427 *
428 * @throws org.xml.sax.SAXException
429 */
430 protected void endNode(Node node) throws org.xml.sax.SAXException
431 {
432
433 switch (node.getNodeType())
434 {
435 case Node.DOCUMENT_NODE :
436 break;
437
438 case Node.ELEMENT_NODE :
439 String ns = m_dh.getNamespaceOfNode(node);
440 if(null == ns)
441 ns = "";
442 this.m_contentHandler.endElement(ns,
443 m_dh.getLocalNameOfNode(node),
444 node.getNodeName());
445
446 if (m_Serializer == null) {
447 // Don't bother with endPrefixMapping calls if the ContentHandler is a
448 // SerializationHandler because SerializationHandler's ignore the
449 // endPrefixMapping() calls anyways. . . . This is an optimization.
450 Element elem_node = (Element) node;
451 NamedNodeMap atts = elem_node.getAttributes();
452 int nAttrs = atts.getLength();
453
454 // do the endPrefixMapping calls in reverse order
455 // of the startPrefixMapping calls
456 for (int i = (nAttrs-1); 0 <= i; i--)
457 {
458 final Node attr = atts.item(i);
459 final String attrName = attr.getNodeName();
460 final int colon = attrName.indexOf(':');
461 final String prefix;
462
463 if (attrName.equals("xmlns") || attrName.startsWith("xmlns:"))
464 {
465 // Use "" instead of null, as Xerces likes "" for the
466 // name of the default namespace. Fix attributed
467 // to "Steven Murray" <smurray@ebt.com>.
468 if (colon < 0)
469 prefix = "";
470 else
471 prefix = attrName.substring(colon + 1);
472
473 this.m_contentHandler.endPrefixMapping(prefix);
474 }
475 else if (colon > 0) {
476 prefix = attrName.substring(0, colon);
477 this.m_contentHandler.endPrefixMapping(prefix);
478 }
479 }
480 {
481 String uri = elem_node.getNamespaceURI();
482 if (uri != null) {
483 String prefix = elem_node.getPrefix();
484 if (prefix==null)
485 prefix="";
486 this.m_contentHandler.endPrefixMapping(prefix);
487 }
488 }
489 }
490 break;
491 case Node.CDATA_SECTION_NODE :
492 break;
493 case Node.ENTITY_REFERENCE_NODE :
494 {
495 EntityReference eref = (EntityReference) node;
496
497 if (m_contentHandler instanceof LexicalHandler)
498 {
499 LexicalHandler lh = ((LexicalHandler) this.m_contentHandler);
500
501 lh.endEntity(eref.getNodeName());
502 }
503 }
504 break;
505 default :
506 }
507 }
508 } //TreeWalker