1 /*
2 * Licensed to the Apache Software Foundation (ASF) under one or more
3 * contributor license agreements. See the NOTICE file distributed with
4 * this work for additional information regarding copyright ownership.
5 * The ASF licenses this file to You under the Apache License, Version 2.0
6 * (the "License"); you may not use this file except in compliance with
7 * the License. You may obtain a copy of the License at
8 *
9 * http://www.apache.org/licenses/LICENSE-2.0
10 *
11 * Unless required by applicable law or agreed to in writing, software
12 * distributed under the License is distributed on an "AS IS" BASIS,
13 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 * See the License for the specific language governing permissions and
15 * limitations under the License.
16 */
17
18 package com.sun.org.apache.xml.internal.resolver.tools;
19
20 import java.io.IOException;
21 import java.io.InputStream;
22 import java.net.MalformedURLException;
23 import java.net.URL;
24
25 import org.xml.sax.InputSource;
26 import org.xml.sax.SAXException;
27 import org.xml.sax.XMLReader;
28 import org.xml.sax.Attributes;
29 import org.xml.sax.helpers.XMLFilterImpl;
30
31 import com.sun.org.apache.xml.internal.resolver.Catalog;
32 import com.sun.org.apache.xml.internal.resolver.CatalogManager;
33
34 import com.sun.org.apache.xml.internal.resolver.helpers.FileURL;
35
36 /**
37 * A SAX XMLFilter that performs catalog-based entity resolution.
38 *
39 * <p>This class implements a SAX XMLFilter that performs entity resolution
40 * using the CatalogResolver. The actual, underlying parser is obtained
41 * from a SAXParserFactory.</p>
42 * </p>
43 *
44 * @see CatalogResolver
45 * @see org.xml.sax.XMLFilter
46 *
47 * @author Norman Walsh
48 * <a href="mailto:Norman.Walsh@Sun.COM">Norman.Walsh@Sun.COM</a>
49 *
50 * @version 1.0
51 */
52 public class ResolvingXMLFilter extends XMLFilterImpl {
53 /**
54 * Suppress explanatory message?
55 *
56 * @see #parse(InputSource)
57 */
58 private static boolean suppressExplanation = false;
59
60 /** The manager for the underlying resolver. */
61 CatalogManager catalogManager = CatalogManager.getStaticManager();
62
63 /** The underlying catalog resolver. */
64 private CatalogResolver catalogResolver = null;
65
66 /** A separate resolver for oasis-xml-pi catalogs. */
67 private CatalogResolver piCatalogResolver = null;
68
69 /** Are we in the prolog? Is an oasis-xml-catalog PI valid now? */
70 private boolean allowXMLCatalogPI = false;
71
72 /** The base URI of the input document, if known. */
73 private URL baseURL = null;
74
75 /** Construct an empty XML Filter with no parent. */
76 public ResolvingXMLFilter() {
77 super();
78 catalogResolver = new CatalogResolver(catalogManager);
79 }
80
81 /** Construct an XML filter with the specified parent. */
82 public ResolvingXMLFilter(XMLReader parent) {
83 super(parent);
84 catalogResolver = new CatalogResolver(catalogManager);
85 }
86
87 /** Construct an XML filter with the specified parent. */
88 public ResolvingXMLFilter(CatalogManager manager) {
89 super();
90 catalogManager = manager;
91 catalogResolver = new CatalogResolver(catalogManager);
92 }
93
94 /** Construct an XML filter with the specified parent. */
95 public ResolvingXMLFilter(XMLReader parent, CatalogManager manager) {
96 super(parent);
97 catalogManager = manager;
98 catalogResolver = new CatalogResolver(catalogManager);
99 }
100
101 /**
102 * Provide accessto the underlying Catalog.
103 */
104 public Catalog getCatalog() {
105 return catalogResolver.getCatalog();
106 }
107
108 /**
109 * SAX XMLReader API.
110 *
111 * <p>Note that the JAXP 1.1ea2 parser crashes with an InternalError if
112 * it encounters a system identifier that appears to be a relative URI
113 * that begins with a slash. For example, the declaration:</p>
114 *
115 * <pre>
116 * <!DOCTYPE book SYSTEM "/path/to/dtd/on/my/system/docbookx.dtd">
117 * </pre>
118 *
119 * <p>would cause such an error. As a convenience, this method catches
120 * that error and prints an explanation. (Unfortunately, it's not possible
121 * to identify the particular system identifier that causes the problem.)
122 * </p>
123 *
124 * <p>The underlying error is forwarded after printing the explanatory
125 * message. The message is only every printed once and if
126 * <code>suppressExplanation</code> is set to <code>false</code> before
127 * parsing, it will never be printed.</p>
128 */
129 public void parse(InputSource input)
130 throws IOException, SAXException {
131 allowXMLCatalogPI = true;
132
133 setupBaseURI(input.getSystemId());
134
135 try {
136 super.parse(input);
137 } catch (InternalError ie) {
138 explain(input.getSystemId());
139 throw ie;
140 }
141 }
142
143 /** SAX XMLReader API.
144 *
145 * @see #parse(InputSource)
146 */
147 public void parse(String systemId)
148 throws IOException, SAXException {
149 allowXMLCatalogPI = true;
150
151 setupBaseURI(systemId);
152
153 try {
154 super.parse(systemId);
155 } catch (InternalError ie) {
156 explain(systemId);
157 throw ie;
158 }
159 }
160
161 /**
162 * Implements the <code>resolveEntity</code> method
163 * for the SAX interface, using an underlying CatalogResolver
164 * to do the real work.
165 */
166 public InputSource resolveEntity (String publicId, String systemId) {
167 allowXMLCatalogPI = false;
168 String resolved = catalogResolver.getResolvedEntity(publicId, systemId);
169
170 if (resolved == null && piCatalogResolver != null) {
171 resolved = piCatalogResolver.getResolvedEntity(publicId, systemId);
172 }
173
174 if (resolved != null) {
175 try {
176 InputSource iSource = new InputSource(resolved);
177 iSource.setPublicId(publicId);
178
179 // Ideally this method would not attempt to open the
180 // InputStream, but there is a bug (in Xerces, at least)
181 // that causes the parser to mistakenly open the wrong
182 // system identifier if the returned InputSource does
183 // not have a byteStream.
184 //
185 // It could be argued that we still shouldn't do this here,
186 // but since the purpose of calling the entityResolver is
187 // almost certainly to open the input stream, it seems to
188 // do little harm.
189 //
190 URL url = new URL(resolved);
191 InputStream iStream = url.openStream();
192 iSource.setByteStream(iStream);
193
194 return iSource;
195 } catch (Exception e) {
196 catalogManager.debug.message(1,
197 "Failed to create InputSource ("
198 + e.toString()
199 + ")", resolved);
200 return null;
201 }
202 } else {
203 return null;
204 }
205 }
206
207 /** SAX DTDHandler API.
208 *
209 * <p>Captured here only to detect the end of the prolog so that
210 * we can ignore subsequent oasis-xml-catalog PIs. Otherwise
211 * the events are just passed through.</p>
212 */
213 public void notationDecl (String name, String publicId, String systemId)
214 throws SAXException {
215 allowXMLCatalogPI = false;
216 super.notationDecl(name,publicId,systemId);
217 }
218
219 /** SAX DTDHandler API.
220 *
221 * <p>Captured here only to detect the end of the prolog so that
222 * we can ignore subsequent oasis-xml-catalog PIs. Otherwise
223 * the events are just passed through.</p>
224 */
225 public void unparsedEntityDecl (String name,
226 String publicId,
227 String systemId,
228 String notationName)
229 throws SAXException {
230 allowXMLCatalogPI = false;
231 super.unparsedEntityDecl (name, publicId, systemId, notationName);
232 }
233
234 /** SAX ContentHandler API.
235 *
236 * <p>Captured here only to detect the end of the prolog so that
237 * we can ignore subsequent oasis-xml-catalog PIs. Otherwise
238 * the events are just passed through.</p>
239 */
240 public void startElement (String uri, String localName, String qName,
241 Attributes atts)
242 throws SAXException {
243 allowXMLCatalogPI = false;
244 super.startElement(uri,localName,qName,atts);
245 }
246
247 /** SAX ContentHandler API.
248 *
249 * <p>Detect and use the oasis-xml-catalog PI if it occurs.</p>
250 */
251 public void processingInstruction(String target, String pidata)
252 throws SAXException {
253 if (target.equals("oasis-xml-catalog")) {
254 URL catalog = null;
255 String data = pidata;
256
257 int pos = data.indexOf("catalog=");
258 if (pos >= 0) {
259 data = data.substring(pos+8);
260 if (data.length() > 1) {
261 String quote = data.substring(0,1);
262 data = data.substring(1);
263 pos = data.indexOf(quote);
264 if (pos >= 0) {
265 data = data.substring(0, pos);
266 try {
267 if (baseURL != null) {
268 catalog = new URL(baseURL, data);
269 } else {
270 catalog = new URL(data);
271 }
272 } catch (MalformedURLException mue) {
273 // nevermind
274 }
275 }
276 }
277 }
278
279 if (allowXMLCatalogPI) {
280 if (catalogManager.getAllowOasisXMLCatalogPI()) {
281 catalogManager.debug.message(4,"oasis-xml-catalog PI", pidata);
282
283 if (catalog != null) {
284 try {
285 catalogManager.debug.message(4,"oasis-xml-catalog", catalog.toString());
286
287 if (piCatalogResolver == null) {
288 piCatalogResolver = new CatalogResolver(true);
289 }
290
291 piCatalogResolver.getCatalog().parseCatalog(catalog.toString());
292 } catch (Exception e) {
293 catalogManager.debug.message(3, "Exception parsing oasis-xml-catalog: "
294 + catalog.toString());
295 }
296 } else {
297 catalogManager.debug.message(3, "PI oasis-xml-catalog unparseable: " + pidata);
298 }
299 } else {
300 catalogManager.debug.message(4,"PI oasis-xml-catalog ignored: " + pidata);
301 }
302 } else {
303 catalogManager.debug.message(3, "PI oasis-xml-catalog occurred in an invalid place: "
304 + pidata);
305 }
306 } else {
307 super.processingInstruction(target, pidata);
308 }
309 }
310
311 /** Save the base URI of the document being parsed. */
312 private void setupBaseURI(String systemId) {
313 URL cwd = null;
314
315 try {
316 cwd = FileURL.makeURL("basename");
317 } catch (MalformedURLException mue) {
318 cwd = null;
319 }
320
321 try {
322 baseURL = new URL(systemId);
323 } catch (MalformedURLException mue) {
324 if (cwd != null) {
325 try {
326 baseURL = new URL(cwd, systemId);
327 } catch (MalformedURLException mue2) {
328 // give up
329 baseURL = null;
330 }
331 } else {
332 // give up
333 baseURL = null;
334 }
335 }
336 }
337
338 /** Provide one possible explanation for an InternalError. */
339 private void explain(String systemId) {
340 if (!suppressExplanation) {
341 System.out.println("XMLReader probably encountered bad URI in " + systemId);
342 System.out.println("For example, replace '/some/uri' with 'file:/some/uri'.");
343 }
344 suppressExplanation = true;
345 }
346 }