1 /*
2 * Copyright (c) 2007, 2013, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation.
8 *
9 * This code is distributed in the hope that it will be useful, but WITHOUT
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
12 * version 2 for more details (a copy is included in the LICENSE file that
13 * accompanied this code).
14 *
15 * You should have received a copy of the GNU General Public License version
16 * 2 along with this work; if not, write to the Free Software Foundation,
17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
18 *
19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
20 * or visit www.oracle.com if you need additional information or have any
21 * questions.
22 */
23 /*
24 * @test
25 * @bug 4052473 4052679 4055602 4066550 4067619 4068012 4068073 4070174 4070452
26 * 4070178 4070450 4070695 4070725 4070795 4071003 4071183 4071782 4072013
27 * 4072388 4072773 4075404 4084356 4087238 4092361 4094033 4094371 4098518
28 * 4099810 4103218 4103220 4103861 4112136 4113638 4113654 4117054 4122468
29 * 4122840 4139860 4156708 4175306 4215747 4209960 4290801 4900884 4942982
30 * 4518811 4945388 4936845 4794068 4461740 4965260 4984277 4826794 5032580
31 * 5102005 5074431 6182685 6208712 6277020 6245766 6351682 6386647 6379382
32 * 6414459 6455680 6498742 6558863 6488119 6547501 6497154 6558856 6481177
33 * 6379214 6485516 6486607 4225362 4494727 6533691 6531591 6531593 6570259
34 * 6509039 6609737 6610748 6645271 6507067 6873931 6450945 6645268 6646611
35 * 6645405 6650730 6910489 6573250 6870908 6585666 6716626 6914413 6916787
36 * 6919624 6998391 7019267 7020960 7025837 7020583 7036905 7066203 7101495
37 * 7003124 7085757 7028073 7171028 7189611 8000983 7195759 8004489 8006509
38 * 7114053 7074882 7040556 8013836 8021121 6192407 6931564 8027695 7090826
39 * 8017142
40 * @summary Verify locale data
41 *
42 */
43
44 /*
45 *
46 * (C) Copyright Taligent, Inc. 1996, 1997 - All Rights Reserved
47 * (C) Copyright IBM Corp. 1996 - 1998 - All Rights Reserved
48 *
49 * Portions copyright (c) 2007 Sun Microsystems, Inc.
50 * All Rights Reserved.
51 *
52 * The original version of this source code and documentation
53 * is copyrighted and owned by Taligent, Inc., a wholly-owned
54 * subsidiary of IBM. These materials are provided under terms
55 * of a License Agreement between Taligent and Sun. This technology
56 * is protected by multiple US and International patents.
57 *
58 * This notice and attribution to Taligent may not be removed.
59 * Taligent is a registered trademark of Taligent, Inc.
60 *
61 * Permission to use, copy, modify, and distribute this software
62 * and its documentation for NON-COMMERCIAL purposes and without
63 * fee is hereby granted provided that this copyright notice
64 * appears in all copies. Please refer to the file "copyright.html"
65 * for further important copyright and licensing information.
66 *
67 * SUN MAKES NO REPRESENTATIONS OR WARRANTIES ABOUT THE SUITABILITY OF
68 * THE SOFTWARE, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED
69 * TO THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
70 * PARTICULAR PURPOSE, OR NON-INFRINGEMENT. SUN SHALL NOT BE LIABLE FOR
71 * ANY DAMAGES SUFFERED BY LICENSEE AS A RESULT OF USING, MODIFYING OR
72 * DISTRIBUTING THIS SOFTWARE OR ITS DERIVATIVES.
73 *
74 */
75
76 /* This test is a generalized test for verifying changes to the locale data.
77 * It is driven by an external file that specifies the particular pieces of locale
78 * data to check. That file is in .properties file format: a series of key/value
79 * pairs delimited by newline characters, with the keys separated from the values
80 * by = signs. The keys are similar in syntax to a Unix pathname, with keys at
81 * successive levels of containment in the resource-data hierarchy separated by
82 * slashes. The file is in ISO 8859-1 encoding, with control characters and
83 * non-ASCII characters denoted with backslash-u escape sequences. The program also allows
84 * blank lines and comment lines to be interspersed with the data. Comment lines
85 * begin with '#'.
86 *
87 * A data file for this test would look something like this:<pre>
88 * FormatData//MonthNames/0=January
89 * FormatData//MonthNames/1=February
90 * LocaleNames//US=United States
91 * LocaleNames//FR=France
92 * FormatData/fr_FR/MonthNames/0=janvier
93 * FormatData/fr_FR/MonthNames/1=f\u00e9vrier
94 * LocaleNames/fr_FR/US=\u00c9tats-Unis
95 * LocaleNames/fr_FR/FR=France</pre>
96 *
97 * You can use language tag with '-' in locale field like this:<pre>
98 * LocaleNames/sr-Latn/SR=Surinam
99 * FormatData/sr-Latn-BA/DayNames/2=utorak</pre>
100 *
101 * The command-line syntax of this test is
102 * <tt>java LocaleDataTest [-w] [{ -s | <filename> }]</tt>
103 *
104 * This program always sends its results to standard output. If -w is not specified,
105 * this program prints out only the differences between the data file and the actual
106 * resource data. If -w is specified, the program prints out every entry, comment,
107 * and blank line from the data file. Where there is a difference between the data
108 * file and the resource data, the data is the data from the resources. This feature
109 * can be used to quickly generate a new data file.
110 *
111 * The user can specify an optional filename or -s. If the user specifies a filename,
112 * the program uses that file as the data file. If the user specifies -s, the program
113 * reads its input from standard input rather than from a file. If the user specifies
114 * neither, the program reads its input from a file called LocaleData in the same
115 * directory the program itself resides in.
116 *
117 * The -nothrow option prevents the program from throwing an exception when it
118 * gets an error. -w implies -nothrow.
119 *
120 * Other command-line options can be specified, but are ignored.
121 *
122 * It's important to note what this test will NOT test. Certain changes to the locale
123 * data are meant to have certain effects on the internationalization frameworks. For
124 * instance, we could ensure round-trip formatting/parsing integrity for the full
125 * date/time format of SimpleDateFormat by making sure that the full date and time
126 * patterns include sufficient data. The test of this is not whether changes were
127 * made to the locale data; it's whether using this data gives round-trip integrity.
128 * Likewise, changing the currency patterns to use \u00a4 instead of local currency
129 * symbols isn't something that can be tested by this test; instead, you want to
130 * actually format currency values and make sure the proper currency symbol was used.
131 *
132 * This test by itself doesn't do an exhaustive comparison of locale data. It is
133 * possible to do this manually, however: Use the GenerateKeyList tool to produce
134 * a complete list of keys for the two versions of the locales you want to compare,
135 * and then diff them. This will flag additions and deletions. Generate a data file
136 * for the base version of the data using the -w option and the output from
137 * GenerateKeyList, and then use the resultant file as the data file when you run
138 * this test against the new version of the data.
139 */
140
141 import java.io.*;
142 import java.text.*;
143 import java.util.Locale;
144 import java.util.ResourceBundle;
145 import java.util.ResourceBundle.Control;
146 import java.util.MissingResourceException;
147
148 public class LocaleDataTest
149 {
150 static final String TEXT_RESOURCES_PACKAGE ="sun.text.resources";
151 static final String UTIL_RESOURCES_PACKAGE ="sun.util.resources";
152
153 public static void main(String[] args) throws Exception {
154
155 // set up our flags and our input and output streams based on the
156 // command-line arguments (exceptions generated here will propagate out
157 // to the environment)
158 BufferedReader in = null;
159 PrintWriter out = null;
160 boolean writeNewFile = false;
161 boolean doThrow = true;
162
163 for (int i = 0; i < args.length; i++) {
164 if (args[i].equals("-w")) {
165 writeNewFile = true;
166 doThrow = false;
167 }
168
169 else if (args[i].equals("-nothrow"))
170 doThrow = false;
171
172 else if (args[i].equals("-s") && in == null)
173 in = new BufferedReader(new EscapeReader(new InputStreamReader(System.in,
174 "ISO8859_1")));
175 else if (!args[i].startsWith("-") && in == null)
176 in = new BufferedReader(new EscapeReader(new InputStreamReader(new
177 FileInputStream(args[i]), "ISO8859_1")));
178 }
179 if (in == null) {
180 File localeData = new File(System.getProperty("test.src", "."), "LocaleData");
181 in = new BufferedReader(new EscapeReader(new InputStreamReader(new
182 FileInputStream(localeData), "ISO8859_1")));
183 }
184 out = new PrintWriter(new EscapeWriter(new OutputStreamWriter(System.out,
185 "ISO8859_1")), true);
186
187 // perform the actual test
188 int errorCount = doTest(in, out, writeNewFile);
189
190 // write out the error count, and throw an exception out into the environment
191 // if there were any errors
192 if (errorCount != 0) {
193 if (!writeNewFile)
194 out.println("Test failed. " + errorCount + " errors.");
195 if (doThrow)
196 throw new Exception("Test failed. " + errorCount + " errors.");
197 }
198 else if (!writeNewFile)
199 out.println("Test passed.");
200
201 in.close();
202 out.close();
203 }
204
205 static int doTest(BufferedReader in, PrintWriter out, boolean writeNewFile)
206 throws Exception {
207 int errorCount = 0;
208
209 String key = null;
210 String expectedValue = null;
211 String line = in.readLine();
212 while (line != null) {
213 if (line.startsWith("#") || line.length() == 0) {
214 if (writeNewFile)
215 out.println(line);
216 }
217
218 else {
219 int index = line.indexOf("=");
220 if (index == -1) {
221 key = line;
222 expectedValue = "";
223 }
224 else {
225 key = line.substring(0, index);
226 if (index + 1 == line.length())
227 expectedValue = "";
228 else
229 expectedValue = line.substring(index + 1);
230 }
231 if (!processLine(key, expectedValue, out, writeNewFile))
232 ++errorCount;
233 }
234 line = in.readLine();
235 }
236 return errorCount;
237 }
238
239 static boolean processLine(String key, String expectedValue, PrintWriter out,
240 boolean writeNewFile) throws Exception {
241 String rbName, localeName, resTag, qualifier;
242 String language = "", country = "", variant = "";
243 int index, oldIndex;
244
245 index = key.indexOf("/");
246 if (index == -1 || index + 1 == key.length())
247 throw new Exception("Malformed input file: no slashes in \"" + key + "\"");
248 rbName = key.substring(0, index);
249
250 oldIndex = index + 1;
251 index = key.indexOf("/", oldIndex);
252 if (index == -1 || index + 1 == key.length())
253 throw new Exception("Malformed input file: \"" + key + "\" is missing locale name");
254 localeName = key.substring(oldIndex, index);
255 boolean use_tag = localeName.indexOf("-") != -1;
256
257 if (use_tag == false && localeName.length() > 0) {
258 language = localeName.substring(0, 2);
259 if (localeName.length() > 3) {
260 country = localeName.substring(3, 5);
261 if (localeName.length() > 5)
262 variant = localeName.substring(6);
263 }
264 }
265
266 oldIndex = index + 1;
267 index = key.indexOf("/", oldIndex);
268 if (index == -1)
269 index = key.length();
270 resTag = key.substring(oldIndex, index);
271
272 // TimeZone name may have "/" in it, for example "Asia/Taipei", so use "Asia\/Taipei in LocaleData.
273 if(resTag.endsWith("\\")) {
274 resTag = resTag.substring(0, resTag.length() - 1);
275 oldIndex = index;
276 index = key.indexOf("/", oldIndex + 1);
277 if (index == -1) index = key.length();
278 resTag += key.substring(oldIndex, index);
279 }
280
281 if (index < key.length() - 1)
282 qualifier = key.substring(index + 1);
283 else
284 qualifier = "";
285
286 String retrievedValue = null;
287 Object resource = null;
288 try {
289 String fullName = null;
290 if (rbName.equals("CalendarData")
291 || rbName.equals("CurrencyNames")
292 || rbName.equals("LocaleNames")
293 || rbName.equals("TimeZoneNames")) {
294 fullName = UTIL_RESOURCES_PACKAGE + "." + rbName;
295 } else {
296 fullName = TEXT_RESOURCES_PACKAGE + "." + rbName;
297 }
298 Locale locale;
299 if (use_tag) {
300 locale = Locale.forLanguageTag(localeName);
301 } else {
302 locale = new Locale(language, country, variant);
303 }
304 ResourceBundle bundle = ResourceBundle.getBundle(fullName,
305 locale,
306 JRELocaleResourceBundleControl.INSTANCE);
307 resource = bundle.getObject(resTag);
308 }
309 catch (MissingResourceException e) {
310 }
311
312 if (resource != null) {
313 if (resource instanceof String) {
314 retrievedValue = (String)resource;
315 }
316 else if (resource instanceof String[]) {
317 int element = Integer.valueOf(qualifier).intValue();
318 String[] stringList = (String[])resource;
319 if (element >= 0 && element < stringList.length)
320 retrievedValue = stringList[element];
321 }
322 else if (resource instanceof String[][]) {
323 String[][] stringArray = (String[][])resource;
324 int slash = qualifier.indexOf("/");
325 if (slash == -1) {
326 for (int i = 0; i < stringArray.length; i++) {
327 if (stringArray[i][0].equals(qualifier))
328 retrievedValue = stringArray[i][1];
329 }
330 }
331 else {
332 int row = Integer.valueOf(qualifier.substring(0, slash)).intValue();
333 int column = Integer.valueOf(qualifier.substring(slash + 1)).intValue();
334 if (row >= 0 && row < stringArray.length && column >= 0 && column <
335 stringArray[row].length)
336 retrievedValue = stringArray[row][column];
337 }
338 }
339 }
340
341 if (retrievedValue == null || !retrievedValue.equals(expectedValue)) {
342 if (retrievedValue == null)
343 retrievedValue = "<MISSING!>";
344
345 if (writeNewFile)
346 out.println(key + "=" + retrievedValue);
347 else {
348 out.println("Mismatch in " + key + ":");
349 out.println(" file = \"" + expectedValue + "\"");
350 out.println(" jvm = \"" + retrievedValue + "\"");
351 }
352 return false;
353 }
354 else {
355 if (writeNewFile)
356 out.println(key + "=" + expectedValue);
357 }
358 return true;
359 }
360
361 private static class JRELocaleResourceBundleControl extends ResourceBundle.Control {
362 static final JRELocaleResourceBundleControl INSTANCE = new JRELocaleResourceBundleControl();
363
364 private JRELocaleResourceBundleControl() {
365 }
366
367 @Override
368 public Locale getFallbackLocale(String baseName, Locale locale) {
369 if (baseName == null || locale == null) {
370 throw new NullPointerException();
371 }
372 return null;
373 }
374
375 private static final String CLDR = ".cldr";
376
377 /**
378 * Changes baseName to its per-language package name and
379 * calls the super class implementation. For example,
380 * if the baseName is "sun.text.resources.FormatData" and locale is ja_JP,
381 * the baseName is changed to "sun.text.resources.ja.FormatData". If
382 * baseName contains "cldr", such as "sun.text.resources.cldr.FormatData",
383 * the name is changed to "sun.text.resources.cldr.jp.FormatData".
384 */
385 @Override
386 public String toBundleName(String baseName, Locale locale) {
387 String newBaseName = baseName;
388 String lang = locale.getLanguage();
389 if (lang.length() > 0) {
390 if (baseName.startsWith(UTIL_RESOURCES_PACKAGE)
391 || baseName.startsWith(TEXT_RESOURCES_PACKAGE)) {
392 // Assume the lengths are the same.
393 if (UTIL_RESOURCES_PACKAGE.length()
394 != TEXT_RESOURCES_PACKAGE.length()) {
395 throw new InternalError("The resources package names have different lengths.");
396 }
397 int index = TEXT_RESOURCES_PACKAGE.length();
398 if (baseName.indexOf(CLDR, index) > 0) {
399 index += CLDR.length();
400 }
401 newBaseName = baseName.substring(0, index + 1) + lang
402 + baseName.substring(index);
403 }
404 }
405 return super.toBundleName(newBaseName, locale);
406 }
407 }
408 }
409
410 class EscapeReader extends FilterReader {
411 public EscapeReader(Reader in) {
412 super(in);
413 }
414
415 public int read() throws IOException {
416 if (buffer != null) {
417 String b = buffer.toString();
418 int result = b.charAt(0);
419 if (b.length() > 1)
420 buffer = new StringBuffer(b.substring(1));
421 else
422 buffer = null;
423 return result;
424 }
425 else {
426 int result = super.read();
427 if (result != '\\')
428 return result;
429 else {
430 buffer = new StringBuffer();
431 result = super.read();
432 buffer.append((char)result);
433 if (result == 'u') {
434 for (int i = 0; i < 4; i++) {
435 result = super.read();
436 if (result == -1)
437 break;
438 buffer.append((char)result);
439 }
440 String number = buffer.toString().substring(1);
441 result = Integer.parseInt(number, 16);
442 buffer = null;
443 return result;
444 }
445 return '\\';
446 }
447 }
448 }
449
450 public int read(char[] cbuf, int start, int len) throws IOException {
451 int p = start;
452 int end = start + len;
453 int c = 0;
454 while (c != -1 && p < end) {
455 c = read();
456 if (c != -1)
457 cbuf[p++] = (char)c;
458 }
459 if (c == -1 && p == start)
460 return -1;
461 else
462 return p - start;
463 }
464
465 private StringBuffer buffer = null;
466 }
467
468 class EscapeWriter extends FilterWriter {
469 public EscapeWriter(Writer out) {
470 super(out);
471 }
472
473 public void write(int c) throws IOException {
474 if ((c >= ' ' && c <= '\u007e') || c == '\r' || c == '\n')
475 super.write(c);
476 else {
477 super.write('\\');
478 super.write('u');
479 String number = Integer.toHexString(c);
480 if (number.length() < 4)
481 number = zeros.substring(0, 4 - number.length()) + number;
482 super.write(number.charAt(0));
483 super.write(number.charAt(1));
484 super.write(number.charAt(2));
485 super.write(number.charAt(3));
486 }
487 }
488
489 public void write(char[] cbuf, int off, int len) throws IOException {
490 int end = off + len;
491 while (off < end)
492 write(cbuf[off++]);
493 }
494
495 public void write(String str, int off, int len) throws IOException {
496 int end = off + len;
497 while (off < end)
498 write(str.charAt(off++));
499 }
500
501 private static String zeros = "0000";
502 }