1 /*
   2  * Copyright (c) 2007, 2018, 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 8008577 8013836 8021121 6192407 6931564 8027695
  39  *      8017142 8037343 8055222 8042126 8074791 8075173 8080774 8129361 8134916
  40  *      8145136 8145952 8164784 8037111 8081643 7037368 8178872 8185841 8190918
  41  *      8187946 8195478 8181157 8179071 8193552 8202026 8204269 8202537 8208746
  42  *      8209775
  43  * @summary Verify locale data
  44  * @modules java.base/sun.util.resources
  45  * @modules jdk.localedata
  46  * @run main LocaleDataTest
  47  * @run main LocaleDataTest -cldr
  48  *
  49  */
  50 
  51 /*
  52  *
  53  * (C) Copyright Taligent, Inc. 1996, 1997 - All Rights Reserved
  54  * (C) Copyright IBM Corp. 1996 - 1998 - All Rights Reserved
  55  *
  56  * Portions copyright (c) 2007 Sun Microsystems, Inc.
  57  * All Rights Reserved.
  58  *
  59  * The original version of this source code and documentation
  60  * is copyrighted and owned by Taligent, Inc., a wholly-owned
  61  * subsidiary of IBM. These materials are provided under terms
  62  * of a License Agreement between Taligent and Sun. This technology
  63  * is protected by multiple US and International patents.
  64  *
  65  * This notice and attribution to Taligent may not be removed.
  66  * Taligent is a registered trademark of Taligent, Inc.
  67  *
  68  * Permission to use, copy, modify, and distribute this software
  69  * and its documentation for NON-COMMERCIAL purposes and without
  70  * fee is hereby granted provided that this copyright notice
  71  * appears in all copies. Please refer to the file "copyright.html"
  72  * for further important copyright and licensing information.
  73  *
  74  * SUN MAKES NO REPRESENTATIONS OR WARRANTIES ABOUT THE SUITABILITY OF
  75  * THE SOFTWARE, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED
  76  * TO THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
  77  * PARTICULAR PURPOSE, OR NON-INFRINGEMENT. SUN SHALL NOT BE LIABLE FOR
  78  * ANY DAMAGES SUFFERED BY LICENSEE AS A RESULT OF USING, MODIFYING OR
  79  * DISTRIBUTING THIS SOFTWARE OR ITS DERIVATIVES.
  80  *
  81  */
  82 
  83 /*    This test is a generalized test for verifying changes to the locale data.
  84  *    It is driven by an external file that specifies the particular pieces of locale
  85  *    data to check.  That file is in .properties file format: a series of key/value
  86  *    pairs delimited by newline characters, with the keys separated from the values
  87  *    by = signs.  The keys are similar in syntax to a Unix pathname, with keys at
  88  *    successive levels of containment in the resource-data hierarchy separated by
  89  *    slashes.  The file is in ISO 8859-1 encoding, with control characters and
  90  *    non-ASCII characters denoted with backslash-u escape sequences.  The program also allows
  91  *    blank lines and comment lines to be interspersed with the data.  Comment lines
  92  *    begin with '#'.
  93  *
  94  *    A data file for this test would look something like this:<pre>
  95  *        FormatData//MonthNames/0=January
  96  *        FormatData//MonthNames/1=February
  97  *        LocaleNames//US=United States
  98  *        LocaleNames//FR=France
  99  *        FormatData/fr_FR/MonthNames/0=janvier
 100  *        FormatData/fr_FR/MonthNames/1=f\u00e9vrier
 101  *        LocaleNames/fr_FR/US=\u00c9tats-Unis
 102  *        LocaleNames/fr_FR/FR=France</pre>
 103  *
 104  *    Second field which designates locale is in the form of:
 105  *    1) Legacy locale notation using '_' as a locale component(language/country/variant) separator.
 106  *    language is a mandatory component. country and variant are optional, however,
 107  *    variant cannot exist without country. So for example, while "ja"/"ja_JP"/"ja_JP_JP" are valid,
 108  *    "_JP"/"ja__JP" are invalid.
 109  *
 110  *    2) BCP47 language tag notation in which we can specify language tag with '-' as a subtag
 111  *       separator. Language tag can be specified with '-' in locale field like this:
 112  *       <pre>LocaleNames/sr-Latn/SR=Surinam
 113  *        FormatData/sr-Latn-BA/DayNames/2=utorak</pre>
 114  *
 115  *    The command-line syntax of this test is
 116  *        <tt>java LocaleDataTest [-w] [{ -s | <filename> }] [-cldr]</tt>
 117  *
 118  *    This program always sends its results to standard output.   If -w is not specified,
 119  *    this program prints out only the differences between the data file and the actual
 120  *    resource data.  If -w is specified, the program prints out every entry, comment,
 121  *    and blank line from the data file.  Where there is a difference between the data
 122  *    file and the resource data, the data is the data from the resources.  This feature
 123  *    can be used to quickly generate a new data file.
 124  *
 125  *    The user can specify an optional filename or -s.  If the user specifies a filename,
 126  *    the program uses that file as the data file.  If the user specifies -s, the program
 127  *    reads its input from standard input rather than from a file.  If the user specifies
 128  *    neither, the program reads its input from a file called LocaleData in the same
 129  *    directory the program itself resides in.
 130  *
 131  *    The -nothrow option prevents the program from throwing an exception when it
 132  *    gets an error.  -w implies -nothrow.
 133  *
 134  *    -cldr option specifies to test CLDR locale data. The default data file name for this
 135  *    option is "LocaleData.cldr".
 136  *
 137  *    Other command-line options can be specified, but are ignored.
 138  *
 139  *    It's important to note what this test will NOT test.  Certain changes to the locale
 140  *    data are meant to have certain effects on the internationalization frameworks.  For
 141  *    instance, we could ensure round-trip formatting/parsing integrity for the full
 142  *    date/time format of SimpleDateFormat by making sure that the full date and time
 143  *    patterns include sufficient data.  The test of this is not whether changes were
 144  *    made to the locale data; it's whether using this data gives round-trip integrity.
 145  *    Likewise, changing the currency patterns to use \u00a4 instead of local currency
 146  *    symbols isn't something that can be tested by this test; instead, you want to
 147  *    actually format currency values and make sure the proper currency symbol was used.
 148  *
 149  *    This test by itself doesn't do an exhaustive comparison of locale data.  It is
 150  *    possible to do this manually, however:  Use the GenerateKeyList tool to produce
 151  *    a complete list of keys for the two versions of the locales you want to compare,
 152  *    and then diff them.  This will flag additions and deletions.  Generate a data file
 153  *    for the base version of the data using the -w option and the output from
 154  *    GenerateKeyList, and then use the resultant file as the data file when you run
 155  *    this test against the new version of the data.
 156  */
 157 
 158 import java.io.BufferedReader;
 159 import java.io.File;
 160 import java.io.FileInputStream;
 161 import java.io.FilterReader;
 162 import java.io.FilterWriter;
 163 import java.io.IOException;
 164 import java.io.InputStreamReader;
 165 import java.io.OutputStreamWriter;
 166 import java.io.PrintWriter;
 167 import java.io.Reader;
 168 import java.io.Writer;
 169 import java.util.Locale;
 170 import java.util.MissingResourceException;
 171 import java.util.ResourceBundle;
 172 import sun.util.resources.LocaleData;
 173 
 174 public class LocaleDataTest
 175 {
 176     static final String TEXT_RESOURCES_PACKAGE ="sun.text.resources";
 177     static final String UTIL_RESOURCES_PACKAGE ="sun.util.resources";
 178     static final String DEFAULT_DATAFILE ="LocaleData";
 179     static String cldrSuffix = "";
 180 
 181     public static void main(String[] args) throws Exception {
 182 
 183         // set up our flags and our input and output streams based on the
 184         // command-line arguments (exceptions generated here will propagate out
 185         // to the environment)
 186         BufferedReader in = null;
 187         PrintWriter out = null;
 188         boolean writeNewFile = false;
 189         boolean doThrow = true;
 190 
 191         for (int i = 0; i < args.length; i++) {
 192             if (args[i].equals("-w")) {
 193                 writeNewFile = true;
 194                 doThrow = false;
 195             }
 196 
 197             else if (args[i].equals("-nothrow"))
 198                 doThrow = false;
 199 
 200             else if (args[i].equals("-cldr")) {
 201                 cldrSuffix = ".cldr";
 202             }
 203 
 204             else if (args[i].equals("-s") && in == null)
 205                 in = new BufferedReader(new EscapeReader(new InputStreamReader(System.in,
 206                                 "ISO8859_1")));
 207             else if (!args[i].startsWith("-") && in == null)
 208                 in = new BufferedReader(new EscapeReader(new InputStreamReader(new
 209                                 FileInputStream(args[i]), "ISO8859_1")));
 210         }
 211         if (in == null) {
 212             File localeData = new File(System.getProperty("test.src", "."), DEFAULT_DATAFILE + cldrSuffix);
 213             in = new BufferedReader(new EscapeReader(new InputStreamReader(new
 214                             FileInputStream(localeData), "ISO8859_1")));
 215         }
 216         out = new PrintWriter(new EscapeWriter(new OutputStreamWriter(System.out,
 217                         "ISO8859_1")), true);
 218 
 219         // perform the actual test
 220         int errorCount = doTest(in, out, writeNewFile);
 221 
 222         // write out the error count, and throw an exception out into the environment
 223         // if there were any errors
 224         if (errorCount != 0) {
 225             if (!writeNewFile)
 226                 out.println("Test failed.  " + errorCount + " errors.");
 227             if (doThrow)
 228                 throw new Exception("Test failed.  " + errorCount + " errors.");
 229         }
 230         else if (!writeNewFile)
 231             out.println("Test passed.");
 232 
 233         in.close();
 234         out.close();
 235     }
 236 
 237     static int doTest(BufferedReader in, PrintWriter out, boolean writeNewFile)
 238                     throws Exception {
 239         int errorCount = 0;
 240 
 241         String key = null;
 242         String expectedValue = null;
 243         String line = in.readLine();
 244         while (line != null) {
 245             if (line.startsWith("#") || line.length() == 0) {
 246                 if (writeNewFile)
 247                     out.println(line);
 248             }
 249 
 250             else {
 251                 int index  = line.indexOf("=");
 252                 if (index == -1) {
 253                     key = line;
 254                     expectedValue = "";
 255                 }
 256                 else {
 257                     key = line.substring(0, index);
 258                     if (index + 1 == line.length())
 259                         expectedValue = "";
 260                     else
 261                         expectedValue = line.substring(index + 1);
 262                 }
 263                 if (!processLine(key, expectedValue, out, writeNewFile))
 264                     ++errorCount;
 265             }
 266             line = in.readLine();
 267         }
 268         return errorCount;
 269     }
 270 
 271     static boolean processLine(String key, String expectedValue, PrintWriter out,
 272                     boolean writeNewFile) throws Exception {
 273         String rbName, localeName, resTag, qualifier;
 274         String language = "", country = "", variant = "";
 275         int index, oldIndex;
 276 
 277         index = key.indexOf("/");
 278         if (index == -1 || index + 1 == key.length())
 279             throw new Exception("Malformed input file: no slashes in \"" + key + "\"");
 280         rbName = key.substring(0, index);
 281 
 282         oldIndex = index + 1;
 283         index = key.indexOf("/", oldIndex);
 284         if (index == -1 || index + 1 == key.length())
 285             throw new Exception("Malformed input file: \"" + key + "\" is missing locale name");
 286         localeName = key.substring(oldIndex, index);
 287         boolean use_tag = localeName.indexOf("-") != -1;
 288         if (use_tag == false && localeName.length() > 0) {
 289             String[] locDetails = localeName.split("_");
 290             switch (locDetails.length) {
 291                 case 1:
 292                     language = locDetails[0];
 293                     break;
 294                 case 2:
 295                     language = locDetails[0];
 296                     country = locDetails[1];
 297                     break;
 298                 case 3:
 299                     language = locDetails[0];
 300                     country = locDetails[1];
 301                     variant = locDetails[2];
 302                     break;
 303                 default:
 304                     throw new Exception("locale not specified properly " + locDetails);
 305             }
 306         }
 307         oldIndex = index + 1;
 308         index = key.indexOf("/", oldIndex);
 309         if (index == -1)
 310             index = key.length();
 311         resTag = key.substring(oldIndex, index);
 312 
 313         // TimeZone name may have "/" in it, for example "Asia/Taipei", so use "Asia\/Taipei in LocaleData.
 314         if(resTag.endsWith("\\")) {
 315             resTag = resTag.substring(0, resTag.length() - 1);
 316             oldIndex = index;
 317             index = key.indexOf("/", oldIndex + 1);
 318             if (index == -1) index = key.length();
 319             resTag += key.substring(oldIndex, index);
 320         }
 321 
 322         if (index < key.length() - 1)
 323             qualifier = key.substring(index + 1);
 324         else
 325             qualifier = "";
 326 
 327         String retrievedValue = null;
 328         Object resource = null;
 329         try {
 330             String fullName = null;
 331             if (rbName.equals("CalendarData")
 332                     || rbName.equals("CurrencyNames")
 333                     || rbName.equals("LocaleNames")
 334                     || rbName.equals("TimeZoneNames")) {
 335                 fullName = UTIL_RESOURCES_PACKAGE + cldrSuffix + "." + rbName;
 336             } else {
 337                 fullName = TEXT_RESOURCES_PACKAGE + cldrSuffix + "." + rbName;
 338             }
 339             Locale locale;
 340             if (use_tag) {
 341                 locale = Locale.forLanguageTag(localeName);
 342             } else {
 343                 locale = new Locale(language, country, variant);
 344             }
 345             ResourceBundle bundle = LocaleData.getBundle(fullName, locale);
 346             resource = bundle.getObject(resTag);
 347         }
 348         catch (MissingResourceException e) {
 349         }
 350 
 351         if (resource != null) {
 352             if (resource instanceof String) {
 353                 retrievedValue = (String)resource;
 354             }
 355             else if (resource instanceof String[]) {
 356                 int element = Integer.valueOf(qualifier).intValue();
 357                 String[] stringList = (String[])resource;
 358                 if (element >= 0 && element < stringList.length)
 359                     retrievedValue = stringList[element];
 360             }
 361             else if (resource instanceof String[][]) {
 362                 String[][] stringArray = (String[][])resource;
 363                 int slash = qualifier.indexOf("/");
 364                 if (slash == -1) {
 365                     for (int i = 0; i < stringArray.length; i++) {
 366                         if (stringArray[i][0].equals(qualifier))
 367                             retrievedValue = stringArray[i][1];
 368                     }
 369                 }
 370                 else {
 371                     int row = Integer.valueOf(qualifier.substring(0, slash)).intValue();
 372                     int column = Integer.valueOf(qualifier.substring(slash + 1)).intValue();
 373                     if (row >= 0 && row < stringArray.length && column >= 0 && column <
 374                                     stringArray[row].length)
 375                         retrievedValue = stringArray[row][column];
 376                 }
 377             }
 378         }
 379 
 380         if (retrievedValue == null || !retrievedValue.equals(expectedValue)) {
 381             if (retrievedValue == null)
 382                 retrievedValue = "<MISSING!>";
 383 
 384             if (writeNewFile)
 385                 out.println(key + "=" + retrievedValue);
 386             else {
 387                 out.println("Mismatch in " + key + ":");
 388                 out.println("  file = \"" + expectedValue + "\"");
 389                 out.println("   jvm = \"" + retrievedValue + "\"");
 390             }
 391             return false;
 392         }
 393         else {
 394             if (writeNewFile)
 395                 out.println(key + "=" + expectedValue);
 396         }
 397         return true;
 398     }
 399 }
 400 
 401 class EscapeReader extends FilterReader {
 402     public EscapeReader(Reader in) {
 403         super(in);
 404     }
 405 
 406     public int read() throws IOException {
 407         if (buffer != null) {
 408             String b = buffer.toString();
 409             int result = b.charAt(0);
 410             if (b.length() > 1)
 411                 buffer = new StringBuffer(b.substring(1));
 412             else
 413                 buffer = null;
 414             return result;
 415         }
 416         else {
 417             int result = super.read();
 418             if (result != '\\')
 419                 return result;
 420             else {
 421                 buffer = new StringBuffer();
 422                 result = super.read();
 423                 buffer.append((char)result);
 424                 if (result == 'u') {
 425                     for (int i = 0; i < 4; i++) {
 426                         result = super.read();
 427                         if (result == -1)
 428                             break;
 429                         buffer.append((char)result);
 430                     }
 431                     String number = buffer.toString().substring(1);
 432                     result = Integer.parseInt(number, 16);
 433                     buffer = null;
 434                     return result;
 435                 }
 436                 return '\\';
 437             }
 438         }
 439     }
 440 
 441     public int read(char[] cbuf, int start, int len) throws IOException {
 442         int p = start;
 443         int end = start + len;
 444         int c = 0;
 445         while (c != -1 && p < end) {
 446             c = read();
 447             if (c != -1)
 448                 cbuf[p++] = (char)c;
 449         }
 450         if (c == -1 && p == start)
 451             return -1;
 452         else
 453             return p - start;
 454     }
 455 
 456     private StringBuffer buffer = null;
 457 }
 458 
 459 class EscapeWriter extends FilterWriter {
 460     public EscapeWriter(Writer out) {
 461         super(out);
 462     }
 463 
 464     public void write(int c) throws IOException {
 465         if ((c >= ' ' && c <= '\u007e') || c == '\r' || c == '\n')
 466             super.write(c);
 467         else {
 468             super.write('\\');
 469             super.write('u');
 470             String number = Integer.toHexString(c);
 471             if (number.length() < 4)
 472                 number = zeros.substring(0, 4 - number.length()) + number;
 473             super.write(number.charAt(0));
 474             super.write(number.charAt(1));
 475             super.write(number.charAt(2));
 476             super.write(number.charAt(3));
 477         }
 478     }
 479 
 480     public void write(char[] cbuf, int off, int len) throws IOException {
 481         int end = off + len;
 482         while (off < end)
 483             write(cbuf[off++]);
 484     }
 485 
 486     public void write(String str, int off, int len) throws IOException {
 487         int end = off + len;
 488         while (off < end)
 489             write(str.charAt(off++));
 490     }
 491 
 492     private static String zeros = "0000";
 493 }