1
|
/**
|
2
|
* '$RCSfile$'
|
3
|
* Purpose: A Class that implements org.xml.sax.EntityResolver interface
|
4
|
* for resolving external entities
|
5
|
* Copyright: 2000 Regents of the University of California and the
|
6
|
* National Center for Ecological Analysis and Synthesis
|
7
|
* Authors: Jivka Bojilova, Matt Jones
|
8
|
*
|
9
|
* '$Author: leinfelder $'
|
10
|
* '$Date: 2011-11-04 14:45:59 -0700 (Fri, 04 Nov 2011) $'
|
11
|
* '$Revision: 6606 $'
|
12
|
*
|
13
|
* This program is free software; you can redistribute it and/or modify
|
14
|
* it under the terms of the GNU General Public License as published by
|
15
|
* the Free Software Foundation; either version 2 of the License, or
|
16
|
* (at your option) any later version.
|
17
|
*
|
18
|
* This program is distributed in the hope that it will be useful,
|
19
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
20
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
21
|
* GNU General Public License for more details.
|
22
|
*
|
23
|
* You should have received a copy of the GNU General Public License
|
24
|
* along with this program; if not, write to the Free Software
|
25
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
26
|
*/
|
27
|
|
28
|
package edu.ucsb.nceas.metacat;
|
29
|
|
30
|
import org.apache.log4j.Logger;
|
31
|
import org.xml.sax.*;
|
32
|
import org.xml.sax.helpers.DefaultHandler;
|
33
|
|
34
|
import edu.ucsb.nceas.metacat.accesscontrol.AccessControlList;
|
35
|
import edu.ucsb.nceas.metacat.database.DBConnection;
|
36
|
import edu.ucsb.nceas.metacat.database.DBConnectionPool;
|
37
|
import edu.ucsb.nceas.metacat.util.SystemUtil;
|
38
|
import edu.ucsb.nceas.utilities.PropertyNotFoundException;
|
39
|
|
40
|
import java.sql.*;
|
41
|
import java.io.File;
|
42
|
import java.io.Reader;
|
43
|
import java.io.BufferedReader;
|
44
|
import java.io.BufferedInputStream;
|
45
|
import java.io.FileWriter;
|
46
|
import java.io.BufferedWriter;
|
47
|
import java.io.InputStream;
|
48
|
import java.io.IOException;
|
49
|
import java.net.URL;
|
50
|
import java.net.MalformedURLException;
|
51
|
|
52
|
/**
|
53
|
* A database aware Class implementing EntityResolver interface for the SAX
|
54
|
* parser to call when processing the XML stream and intercepting any
|
55
|
* external entities (including the external DTD subset and external
|
56
|
* parameter entities, if any) before including them.
|
57
|
*/
|
58
|
public class DBEntityResolver implements EntityResolver
|
59
|
{
|
60
|
private DBConnection connection = null;
|
61
|
private DefaultHandler handler = null;
|
62
|
private String docname = null;
|
63
|
private String doctype = null;
|
64
|
private String systemid = null;
|
65
|
private Reader dtdtext = null;
|
66
|
private static Logger logMetacat = Logger.getLogger(DBEntityResolver.class);
|
67
|
|
68
|
/**
|
69
|
* Construct an instance of the DBEntityResolver class
|
70
|
*
|
71
|
* @param conn the JDBC connection to which information is written
|
72
|
*/
|
73
|
public DBEntityResolver(DBConnection conn)
|
74
|
{
|
75
|
this.connection= conn;
|
76
|
}
|
77
|
/**
|
78
|
* Construct an instance of the DBEntityResolver class
|
79
|
*
|
80
|
* @param conn the JDBC connection to which information is written
|
81
|
* @param handler the SAX handler to determine parsing context
|
82
|
* @param dtd Reader of new dtd to be uploaded on server's file system
|
83
|
*/
|
84
|
public DBEntityResolver(DBConnection conn, DefaultHandler handler, Reader dtd)
|
85
|
{
|
86
|
this.connection = conn;
|
87
|
this.handler = handler;
|
88
|
this.dtdtext = dtd;
|
89
|
}
|
90
|
|
91
|
/**
|
92
|
* The Parser call this method before opening any external entity
|
93
|
* except the top-level document entity (including the external DTD subset,
|
94
|
* external entities referenced within the DTD, and external entities
|
95
|
* referenced within the document element)
|
96
|
*/
|
97
|
public InputSource resolveEntity (String publicId, String systemId)
|
98
|
throws SAXException
|
99
|
{
|
100
|
logMetacat.debug("DBEntityResolver.resolveEntity - in resolveEntity");
|
101
|
String dbSystemID;
|
102
|
String doctype = null;
|
103
|
|
104
|
// Won't have a handler under all cases
|
105
|
if ( handler != null ) {
|
106
|
if ( handler instanceof DBSAXHandler ) {
|
107
|
DBSAXHandler dhandler = null;
|
108
|
dhandler = (DBSAXHandler)handler;
|
109
|
if ( dhandler.processingDTD() ) {
|
110
|
|
111
|
// public ID is doctype
|
112
|
if (publicId != null) {
|
113
|
doctype = publicId;
|
114
|
logMetacat.debug("DBEntityResolver.resolveEntity - in get type from publicId and doctype is: "
|
115
|
+ doctype);
|
116
|
// assume public ID (doctype) is docname
|
117
|
} else if (systemId != null) {
|
118
|
doctype = dhandler.getDocname();
|
119
|
}
|
120
|
}
|
121
|
} else if ( handler instanceof AccessControlList ) {
|
122
|
AccessControlList ahandler = null;
|
123
|
ahandler = (AccessControlList)handler;
|
124
|
//if ( ahandler.processingDTD() ) {
|
125
|
// public ID is doctype
|
126
|
if (publicId != null) {
|
127
|
doctype = publicId;
|
128
|
// assume public ID (doctype) is docname
|
129
|
} else if (systemId != null) {
|
130
|
doctype = ahandler.getDocname();
|
131
|
}
|
132
|
//}
|
133
|
}
|
134
|
}
|
135
|
|
136
|
// get System ID for doctype
|
137
|
if (doctype != null) {
|
138
|
// look at db XML Catalog for System ID
|
139
|
logMetacat.info("DBEntityResolver.resolveEntity - get systemId from doctype: " + doctype);
|
140
|
dbSystemID = getDTDSystemID(doctype);
|
141
|
logMetacat.info("DBEntityResolver.resolveEntity - The Systemid is: " + dbSystemID);
|
142
|
// check that it is accessible on our system before getting too far
|
143
|
try {
|
144
|
InputStream in = checkURLConnection(dbSystemID);
|
145
|
} catch (SAXException se) {
|
146
|
// after an upgrade, the dtd will not exist on disk, but it is in xml catalog. The db system id may be pointing
|
147
|
// back at this system Try and download it from the original system id and see if we still have a problem
|
148
|
// checking the URL connection.
|
149
|
logMetacat.warn("DBEntityResolver.resolveEntity - Problem when checking URL Connection: " + se.getMessage());
|
150
|
logMetacat.warn("DBEntityResolver.resolveEntity - Probably, dtd for doc type " + doctype + " existed in xml catalog, but not on disk. Uploading from: " + systemId);
|
151
|
InputStream istream = checkURLConnection(systemId);
|
152
|
uploadDTDFromURL(istream, systemId);
|
153
|
try {
|
154
|
Thread.currentThread().sleep(6000);
|
155
|
checkURLConnection(dbSystemID);
|
156
|
} catch (Exception e2) {
|
157
|
logMetacat.error("DBEntityResolver.resolveEntity - still could not find dtd for doc type " + doctype + " at "
|
158
|
+ dbSystemID + " : " + e2.getMessage());
|
159
|
dbSystemID = null;
|
160
|
}
|
161
|
}
|
162
|
boolean doctypeIsInDB = true;
|
163
|
// no System ID found in db XML Catalog
|
164
|
if (dbSystemID == null) {
|
165
|
doctypeIsInDB = false;
|
166
|
// use the provided System ID
|
167
|
if (systemId != null) {
|
168
|
dbSystemID = systemId;
|
169
|
}
|
170
|
logMetacat.info("DBEntityResolver.resolveEntity - If above Systemid is null, then get "
|
171
|
+ "system id from file: " + dbSystemID);
|
172
|
}
|
173
|
// there are dtd text provided; try to upload on Metacat
|
174
|
if ( dtdtext != null ) {
|
175
|
dbSystemID = uploadDTD(dbSystemID);
|
176
|
}
|
177
|
|
178
|
// open URLConnection to check first
|
179
|
InputStream istream = checkURLConnection(dbSystemID);
|
180
|
|
181
|
// need to register System ID in db XML Catalog if not yet
|
182
|
if ( !doctypeIsInDB ) {
|
183
|
// new DTD from outside URL location; try to upload on Metacat
|
184
|
if ( dtdtext == null ) {
|
185
|
dbSystemID = uploadDTDFromURL(istream, dbSystemID);
|
186
|
}
|
187
|
registerDTD(doctype, dbSystemID);
|
188
|
}
|
189
|
// return a byte-input stream for use
|
190
|
InputSource is = new InputSource(dbSystemID);
|
191
|
|
192
|
// close and open URLConnection again
|
193
|
try {
|
194
|
istream.close();
|
195
|
} catch (IOException e) {
|
196
|
throw new SAXException
|
197
|
("DBEntityResolver.resolveEntity - I/O issue when resolving entity: " + e.getMessage());
|
198
|
}
|
199
|
istream = checkURLConnection(dbSystemID);
|
200
|
is.setByteStream(istream);
|
201
|
return is;
|
202
|
} else {
|
203
|
// use provided systemId for the other cases
|
204
|
logMetacat.info("DBEntityResolver.resolveEntity - doctype is null and using system id from file");
|
205
|
InputStream istream = checkURLConnection(systemId);
|
206
|
return null;
|
207
|
|
208
|
}
|
209
|
|
210
|
}
|
211
|
|
212
|
/**
|
213
|
* Look at db XML Catalog to get System ID (if any) for @doctype.
|
214
|
* Return null if there are no System ID found for @doctype
|
215
|
*/
|
216
|
public static String getDTDSystemID( String doctype )
|
217
|
throws SAXException
|
218
|
{
|
219
|
String systemid = null;
|
220
|
PreparedStatement pstmt = null;
|
221
|
DBConnection conn = null;
|
222
|
int serialNumber = -1;
|
223
|
try {
|
224
|
//check out DBConnection
|
225
|
conn=DBConnectionPool.getDBConnection("DBEntityResolver.getDTDSystemID");
|
226
|
serialNumber=conn.getCheckOutSerialNumber();
|
227
|
|
228
|
String sql = "SELECT system_id FROM xml_catalog " +
|
229
|
"WHERE entry_type = 'DTD' AND public_id = ?";
|
230
|
|
231
|
pstmt = conn.prepareStatement(sql);
|
232
|
pstmt.setString(1, doctype);
|
233
|
|
234
|
pstmt.execute();
|
235
|
ResultSet rs = pstmt.getResultSet();
|
236
|
boolean tableHasRows = rs.next();
|
237
|
if (tableHasRows) {
|
238
|
systemid = rs.getString(1);
|
239
|
// system id may not have server url on front. Add it if not.
|
240
|
if (!systemid.startsWith("http://")) {
|
241
|
systemid = SystemUtil.getContextURL() + systemid;
|
242
|
}
|
243
|
}
|
244
|
pstmt.close();
|
245
|
} catch (SQLException e) {
|
246
|
throw new SAXException
|
247
|
("DBEntityResolver.getDTDSystemID - SQL error when getting DTD system ID: " + e.getMessage());
|
248
|
} catch (PropertyNotFoundException pnfe) {
|
249
|
throw new SAXException
|
250
|
("DBEntityResolver.getDTDSystemID - Property error when getting DTD system ID: " + pnfe.getMessage());
|
251
|
}
|
252
|
finally
|
253
|
{
|
254
|
try
|
255
|
{
|
256
|
pstmt.close();
|
257
|
}//try
|
258
|
catch (SQLException sqlE)
|
259
|
{
|
260
|
logMetacat.error("DBEntityResolver.getDTDSystemId - SQL error: " + sqlE.getMessage());
|
261
|
}//catch
|
262
|
finally
|
263
|
{
|
264
|
DBConnectionPool.returnDBConnection(conn, serialNumber);
|
265
|
}//finally
|
266
|
}//finally
|
267
|
|
268
|
// return the selected System ID
|
269
|
return systemid;
|
270
|
}
|
271
|
|
272
|
/**
|
273
|
* Register new DTD identified by @systemId in Metacat XML Catalog
|
274
|
* . make a reference with @systemId for @doctype in Metacat DB
|
275
|
*/
|
276
|
private void registerDTD ( String doctype, String systemId )
|
277
|
throws SAXException
|
278
|
{
|
279
|
String existingSystemId = getDTDSystemID(doctype);
|
280
|
if (existingSystemId != null && existingSystemId.equals(systemId)) {
|
281
|
logMetacat.warn("DBEntityResolver.registerDTD - doctype/systemId already registered in DB: " + doctype);
|
282
|
return;
|
283
|
}
|
284
|
//DBConnection conn = null;
|
285
|
//int serialNumber = -1;
|
286
|
PreparedStatement pstmt = null;
|
287
|
// make a reference in db catalog table with @systemId for @doctype
|
288
|
try {
|
289
|
//check out DBConnection
|
290
|
//conn=DBConnectionPool.getDBConnection("DBEntityResolver.registerDTD");
|
291
|
//serialNumber=conn.getCheckOutSerialNumber();
|
292
|
|
293
|
|
294
|
pstmt = connection.prepareStatement(
|
295
|
"INSERT INTO xml_catalog " +
|
296
|
"(entry_type, public_id, system_id) " +
|
297
|
"VALUES ('DTD', ?, ?)");
|
298
|
// Increase usage count
|
299
|
connection.increaseUsageCount(1);
|
300
|
// Bind the values to the query
|
301
|
pstmt.setString(1, doctype);
|
302
|
pstmt.setString(2, systemId);
|
303
|
// Do the insertion
|
304
|
pstmt.execute();
|
305
|
int updateCnt = pstmt.getUpdateCount();
|
306
|
logMetacat.debug("DBEntityReolver.registerDTD - DTDs registered: " + updateCnt);
|
307
|
pstmt.close();
|
308
|
} catch (SQLException e) {
|
309
|
throw new SAXException
|
310
|
("DBEntityResolver.registerDTD - SQL issue when registering DTD: " + e.getMessage());
|
311
|
}
|
312
|
finally
|
313
|
{
|
314
|
try
|
315
|
{
|
316
|
pstmt.close();
|
317
|
}//try
|
318
|
catch (SQLException sqlE)
|
319
|
{
|
320
|
logMetacat.error("DBEntityResolver.registerDTD - SQL error: " + sqlE.getMessage());
|
321
|
}//catch
|
322
|
//DBConnectionPool.returnDBConnection(conn, serialNumber);
|
323
|
}//finally
|
324
|
|
325
|
}
|
326
|
|
327
|
/**
|
328
|
* Upload new DTD text identified by
|
329
|
*
|
330
|
* @systemId to Metacat file system
|
331
|
*/
|
332
|
private String uploadDTD(String systemId) throws SAXException {
|
333
|
String dtdPath = null;
|
334
|
String dtdURL = null;
|
335
|
try {
|
336
|
dtdPath = SystemUtil.getContextDir() + "/dtd/";
|
337
|
dtdURL = SystemUtil.getContextURL() + "/dtd/";
|
338
|
} catch (PropertyNotFoundException pnfe) {
|
339
|
throw new SAXException("DBEntityResolver.uploadDTD: " + pnfe.getMessage());
|
340
|
}
|
341
|
|
342
|
// get filename from systemId
|
343
|
String filename = systemId;
|
344
|
int slash = Math.max(filename.lastIndexOf('/'), filename.lastIndexOf('\\'));
|
345
|
if (slash > -1) {
|
346
|
filename = filename.substring(slash + 1);
|
347
|
}
|
348
|
|
349
|
// writing dtd text on Metacat file system as filename
|
350
|
try {
|
351
|
// create a buffering character-input stream
|
352
|
// that uses a default-sized input buffer
|
353
|
BufferedReader in = new BufferedReader(dtdtext);
|
354
|
|
355
|
// open file writer to write the input into it
|
356
|
// String dtdPath = "/opt/tomcat/webapps/bojilova/dtd/";
|
357
|
File f = new File(dtdPath, filename);
|
358
|
synchronized (f) {
|
359
|
try {
|
360
|
if (f.exists()) {
|
361
|
throw new IOException("File already exist: "
|
362
|
+ f.getCanonicalFile());
|
363
|
// if ( f.exists() && !f.canWrite() ) {
|
364
|
// throw new IOException("Not writable: " +
|
365
|
// f.getCanonicalFile());
|
366
|
}
|
367
|
} catch (SecurityException se) {
|
368
|
// if a security manager exists,
|
369
|
// its checkRead method is called for f.exist()
|
370
|
// or checkWrite method is called for f.canWrite()
|
371
|
throw se;
|
372
|
}
|
373
|
// create a buffered character-output stream
|
374
|
// that uses a default-sized output buffer
|
375
|
FileWriter fw = new FileWriter(f);
|
376
|
BufferedWriter out = new BufferedWriter(fw);
|
377
|
|
378
|
// read the input and write into the file writer
|
379
|
String inputLine;
|
380
|
while ((inputLine = in.readLine()) != null) {
|
381
|
out.write(inputLine, 0, inputLine.length());
|
382
|
out.newLine(); // instead of out.write('\r\n');
|
383
|
}
|
384
|
|
385
|
// the input and the output streams must be closed
|
386
|
in.close();
|
387
|
out.flush();
|
388
|
out.close();
|
389
|
fw.close();
|
390
|
} // end of synchronized
|
391
|
} catch (MalformedURLException e) {
|
392
|
throw new SAXException("DBEntityResolver.uploadDTD() - Malformed URL when uploading DTD: " + e.getMessage());
|
393
|
} catch (IOException e) {
|
394
|
throw new SAXException("DBEntityResolver.uploadDTD - I/O issue when uploading DTD: " + e.getMessage());
|
395
|
} catch (SecurityException e) {
|
396
|
throw new SAXException("DBEntityResolver.uploadDTD() - Security issue when uploading DTD: " + e.getMessage());
|
397
|
}
|
398
|
|
399
|
// String dtdURL = "http://dev.nceas.ucsb.edu/bojilova/dtd/";
|
400
|
return dtdURL + filename;
|
401
|
}
|
402
|
|
403
|
|
404
|
/**
|
405
|
* Upload new DTD located at outside URL to Metacat file system
|
406
|
*/
|
407
|
private String uploadDTDFromURL(InputStream istream, String systemId)
|
408
|
throws SAXException {
|
409
|
String dtdPath = null;
|
410
|
String dtdURL = null;
|
411
|
try {
|
412
|
dtdPath = SystemUtil.getContextDir() + "/dtd/";
|
413
|
dtdURL = SystemUtil.getContextURL() + "/dtd/";
|
414
|
} catch (PropertyNotFoundException pnfe) {
|
415
|
throw new SAXException("DBEntityResolver.uploadDTDFromURL - Property issue when uploading DTD from URL: "
|
416
|
+ pnfe.getMessage());
|
417
|
}
|
418
|
|
419
|
// get filename from systemId
|
420
|
String filename = systemId;
|
421
|
int slash = Math.max(filename.lastIndexOf('/'), filename.lastIndexOf('\\'));
|
422
|
if (slash > -1) {
|
423
|
filename = filename.substring(slash + 1);
|
424
|
}
|
425
|
|
426
|
// writing dtd text on Metacat file system as filename
|
427
|
try {
|
428
|
// create a buffering character-input stream
|
429
|
// that uses a default-sized input buffer
|
430
|
BufferedInputStream in = new BufferedInputStream(istream);
|
431
|
|
432
|
// open file writer to write the input into it
|
433
|
//String dtdPath = "/opt/tomcat/webapps/bojilova/dtd/";
|
434
|
File f = new File(dtdPath, filename);
|
435
|
synchronized (f) {
|
436
|
try {
|
437
|
if (f.exists()) {
|
438
|
logMetacat.warn("DBEntityResolver.uploadDTDFromURL - File already exists: " + f.getCanonicalFile());
|
439
|
//return dtdURL + filename;
|
440
|
//throw new IOException("File already exist: "
|
441
|
// + f.getCanonicalFile());
|
442
|
//if ( f.exists() && !f.canWrite() ) {
|
443
|
// throw new IOException("Not writable: " + f.getCanonicalFile());
|
444
|
}
|
445
|
} catch (SecurityException se) {
|
446
|
// if a security manager exists,
|
447
|
// its checkRead method is called for f.exist()
|
448
|
// or checkWrite method is called for f.canWrite()
|
449
|
throw se;
|
450
|
}
|
451
|
// create a buffered character-output stream
|
452
|
// that uses a default-sized output buffer
|
453
|
FileWriter fw = new FileWriter(f);
|
454
|
BufferedWriter out = new BufferedWriter(fw);
|
455
|
|
456
|
// read the input and write into the file writer
|
457
|
int inputByte;
|
458
|
while ((inputByte = in.read()) != -1) {
|
459
|
out.write(inputByte);
|
460
|
//out.newLine(); //instead of out.write('\r\n');
|
461
|
}
|
462
|
|
463
|
// the input and the output streams must be closed
|
464
|
in.close();
|
465
|
out.flush();
|
466
|
out.close();
|
467
|
fw.close();
|
468
|
} // end of synchronized
|
469
|
} catch (MalformedURLException e) {
|
470
|
throw new SAXException("DBEntityResolver.uploadDTDFromURL - Malformed URL when uploading DTD from URL: "
|
471
|
+ e.getMessage());
|
472
|
} catch (IOException e) {
|
473
|
throw new SAXException("DBEntityResolver.uploadDTDFromURL - I/O issue when uploading DTD from URL: "
|
474
|
+ e.getMessage());
|
475
|
} catch (SecurityException e) {
|
476
|
throw new SAXException("DBEntityResolver.uploadDTDFromURL - Security issue when uploading DTD from URL: "
|
477
|
+ e.getMessage());
|
478
|
}
|
479
|
|
480
|
//String dtdURL = "http://dev.nceas.ucsb.edu/bojilova/dtd/";
|
481
|
return dtdURL + filename;
|
482
|
}
|
483
|
|
484
|
/**
|
485
|
* Check URL Connection for @systemId, and return an InputStream
|
486
|
* that can be used to read from the systemId URL. The parser ends
|
487
|
* up using this via the InputSource to read the DTD.
|
488
|
*
|
489
|
* @param systemId a URI (in practice URL) to be checked and opened
|
490
|
*/
|
491
|
public static InputStream checkURLConnection(String systemId) throws SAXException {
|
492
|
try {
|
493
|
return (new URL(systemId).openStream());
|
494
|
|
495
|
} catch (MalformedURLException e) {
|
496
|
throw new SAXException("DBEntityResolver.checkURLConnection - Malformed URL when checking URL Connection: "
|
497
|
+ e.getMessage());
|
498
|
} catch (IOException e) {
|
499
|
throw new SAXException("DBEntityResolver.checkURLConnection - I/O issue when checking URL Connection: "
|
500
|
+ e.getMessage());
|
501
|
}
|
502
|
}
|
503
|
}
|