make/tools/src/build/tools/cldrconverter/CLDRConverter.java

Print this page




   8  * particular file as subject to the "Classpath" exception as provided
   9  * by Oracle in the LICENSE file that accompanied this code.
  10  *
  11  * This code is distributed in the hope that it will be useful, but WITHOUT
  12  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  13  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  14  * version 2 for more details (a copy is included in the LICENSE file that
  15  * accompanied this code).
  16  *
  17  * You should have received a copy of the GNU General Public License version
  18  * 2 along with this work; if not, write to the Free Software Foundation,
  19  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
  20  *
  21  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
  22  * or visit www.oracle.com if you need additional information or have any
  23  * questions.
  24  */
  25 
  26 package build.tools.cldrconverter;
  27 

  28 import java.io.File;
  29 import java.nio.file.DirectoryStream;
  30 import java.nio.file.FileSystems;
  31 import java.nio.file.Files;
  32 import java.nio.file.Path;
  33 import java.util.*;
  34 import javax.xml.parsers.SAXParser;
  35 import javax.xml.parsers.SAXParserFactory;
  36 
  37 
  38 /**
  39  * Converts locale data from "Locale Data Markup Language" format to
  40  * JRE resource bundle format. LDML is the format used by the Common
  41  * Locale Data Repository maintained by the Unicode Consortium.
  42  */
  43 public class CLDRConverter {
  44 
  45     static final String LDML_DTD_SYSTEM_ID = "http://www.unicode.org/cldr/dtd/2.0/ldml.dtd";
  46     static final String SPPL_LDML_DTD_SYSTEM_ID = "http://www.unicode.org/cldr/dtd/2.0/ldmlSupplemental.dtd";
  47 
  48     private static String CLDR_BASE = "../CLDR/21.0.1/";
  49     static String LOCAL_LDML_DTD;
  50     static String LOCAL_SPPL_LDML_DTD;
  51     private static String SOURCE_FILE_DIR;
  52     private static String SPPL_SOURCE_FILE;
  53     private static String NUMBERING_SOURCE_FILE;
  54     private static String METAZONES_SOURCE_FILE;
  55     static String DESTINATION_DIR = "build/gensrc";
  56 
  57     static final String LOCALE_NAME_PREFIX = "locale.displayname.";
  58     static final String CURRENCY_SYMBOL_PREFIX = "currency.symbol.";
  59     static final String CURRENCY_NAME_PREFIX = "currency.displayname.";
  60     static final String TIMEZONE_ID_PREFIX = "timezone.id.";
  61     static final String TIMEZONE_NAME_PREFIX = "timezone.displayname.";
  62     static final String METAZONE_ID_PREFIX = "metazone.id.";
  63     static final String METAZONE_NAME_PREFIX = "metazone.displayname.";
  64 
  65     private static SupplementDataParseHandler handlerSuppl;
  66     static NumberingSystemsParseHandler handlerNumbering;
  67     static MetaZonesParseHandler handlerMetaZones;
  68     private static BundleGenerator bundleGenerator;
  69 
  70     static int draftType;
  71     private static final String DRAFT_UNCONFIRMED = "unconfirmed";
  72     private static final String DRAFT_PROVISIONAL = "provisional";
  73     private static final String DRAFT_CONTRIBUTED = "contributed";
  74     private static final String DRAFT_APPROVED = "approved";
  75     private static final String DRAFT_TRUE = "true";
  76     private static final String DRAFT_FALSE = "false";
  77     private static final String DRAFT_DEFAULT = DRAFT_APPROVED;
  78     static final Map<String, Integer> DRAFT_MAP = new HashMap<>();
  79 
  80     static {
  81         DRAFT_MAP.put(DRAFT_UNCONFIRMED, 0);
  82         DRAFT_MAP.put(DRAFT_PROVISIONAL, 1);
  83         DRAFT_MAP.put(DRAFT_CONTRIBUTED, 2);


 219         ResourceBundle.Control defCon = ResourceBundle.Control.getControl(ResourceBundle.Control.FORMAT_DEFAULT);
 220         List<Bundle> retList = new ArrayList<>();
 221         Path path = FileSystems.getDefault().getPath(SOURCE_FILE_DIR);
 222         try (DirectoryStream<Path> dirStr = Files.newDirectoryStream(path)) {
 223             for (Path entry : dirStr) {
 224                 String fileName = entry.getFileName().toString();
 225                 if (fileName.endsWith(".xml")) {
 226                     String id = fileName.substring(0, fileName.indexOf('.'));
 227                     Locale cldrLoc = Locale.forLanguageTag(toLanguageTag(id));
 228                     List<Locale> candList = defCon.getCandidateLocales("", cldrLoc);
 229                     StringBuilder sb = new StringBuilder();
 230                     for (Locale loc : candList) {
 231                         if (!loc.equals(Locale.ROOT)) {
 232                             sb.append(toLocaleName(loc.toLanguageTag()));
 233                             sb.append(",");
 234                         }
 235                     }
 236                     if (sb.indexOf("root") == -1) {
 237                         sb.append("root");
 238                     }
 239                     retList.add(new Bundle(id, sb.toString(), null, null));







 240                 }
 241             }
 242         }
 243         return retList;
 244     }
 245 
 246     private static Map<String, Map<String, Object>> cldrBundles = new HashMap<>();
 247 
 248     static Map<String, Object> getCLDRBundle(String id) throws Exception {
 249         Map<String, Object> bundle = cldrBundles.get(id);
 250         if (bundle != null) {
 251             return bundle;
 252         }
 253         SAXParserFactory factory = SAXParserFactory.newInstance();
 254         factory.setValidating(true);
 255         SAXParser parser = factory.newSAXParser();
 256         LDMLParseHandler handler = new LDMLParseHandler(id);
 257         File file = new File(SOURCE_FILE_DIR + File.separator + id + ".xml");
 258         if (!file.exists()) {
 259             // Skip if the file doesn't exist.


 295         // Parse numberingSystems to get digit zero character information.
 296         SAXParserFactory numberingParser = SAXParserFactory.newInstance();
 297         numberingParser.setValidating(true);
 298         SAXParser parserNumbering = numberingParser.newSAXParser();
 299         handlerNumbering = new NumberingSystemsParseHandler();
 300         File fileNumbering = new File(NUMBERING_SOURCE_FILE);
 301         parserNumbering.parse(fileNumbering, handlerNumbering);
 302 
 303         // Parse metaZones to create mappings between Olson tzids and CLDR meta zone names
 304         SAXParserFactory metazonesParser = SAXParserFactory.newInstance();
 305         metazonesParser.setValidating(true);
 306         SAXParser parserMetaZones = metazonesParser.newSAXParser();
 307         handlerMetaZones = new MetaZonesParseHandler();
 308         File fileMetaZones = new File(METAZONES_SOURCE_FILE);
 309         parserNumbering.parse(fileMetaZones, handlerMetaZones);
 310 
 311         // For generating information on supported locales.
 312         Map<String, SortedSet<String>> metaInfo = new HashMap<>();
 313         metaInfo.put("LocaleNames", new TreeSet<String>());
 314         metaInfo.put("CurrencyNames", new TreeSet<String>());

 315         metaInfo.put("CalendarData", new TreeSet<String>());
 316         metaInfo.put("FormatData", new TreeSet<String>());
 317 
 318         for (Bundle bundle : bundles) {
 319             // Get the target map, which contains all the data that should be
 320             // visible for the bundle's locale
 321 
 322             Map<String, Object> targetMap = bundle.getTargetMap();
 323 
 324             EnumSet<Bundle.Type> bundleTypes = bundle.getBundleTypes();
 325 
 326             // Fill in any missing resources in the base bundle from en and en-US data.
 327             // This is because CLDR root.xml is supposed to be language neutral and doesn't
 328             // provide some resource data. Currently, the runtime assumes that there are all
 329             // resources though the parent resource bundle chain.
 330             if (bundle.isRoot()) {
 331                 Map<String, Object> enData = new HashMap<>();
 332                 // Create a superset of en-US and en bundles data in order to
 333                 // fill in any missing resources in the base bundle.
 334                 enData.putAll(Bundle.getBundle("en").getTargetMap());
 335                 enData.putAll(Bundle.getBundle("en_US").getTargetMap());
 336                 for (String key : enData.keySet()) {
 337                     if (!targetMap.containsKey(key)) {
 338                         targetMap.put(key, enData.get(key));
 339                     }
 340                 }
 341                 // Add DateTimePatternChars because CLDR no longer supports localized patterns.
 342                 targetMap.put("DateTimePatternChars", "GyMdkHmsSEDFwWahKzZ");
 343             }
 344 
 345             // Now the map contains just the entries that need to be in the resources bundles.
 346             // Go ahead and generate them.
 347             if (bundleTypes.contains(Bundle.Type.LOCALENAMES)) {
 348                 Map<String, Object> localeNamesMap = extractLocaleNames(targetMap, bundle.getID());
 349                 if (!localeNamesMap.isEmpty() || bundle.isRoot()) {
 350                     metaInfo.get("LocaleNames").add(toLanguageTag(bundle.getID()));
 351                     bundleGenerator.generateBundle("util", "LocaleNames", bundle.getID(), true, localeNamesMap, true);
 352                 }
 353             }
 354             if (bundleTypes.contains(Bundle.Type.CURRENCYNAMES)) {
 355                 Map<String, Object> currencyNamesMap = extractCurrencyNames(targetMap, bundle.getID(), bundle.getCurrencies());
 356                 if (!currencyNamesMap.isEmpty() || bundle.isRoot()) {
 357                     metaInfo.get("CurrencyNames").add(toLanguageTag(bundle.getID()));
 358                     bundleGenerator.generateBundle("util", "CurrencyNames", bundle.getID(), true, currencyNamesMap, true);
 359                 }
 360             }
 361             if (bundleTypes.contains(Bundle.Type.TIMEZONENAMES)) {
 362                 Map<String, Object> zoneNamesMap = extractZoneNames(targetMap, bundle.getID());




 363             }
 364             if (bundleTypes.contains(Bundle.Type.CALENDARDATA)) {
 365                 Map<String, Object> calendarDataMap = extractCalendarData(targetMap, bundle.getID());
 366                 if (!calendarDataMap.isEmpty() || bundle.isRoot()) {
 367                     metaInfo.get("CalendarData").add(toLanguageTag(bundle.getID()));
 368                     bundleGenerator.generateBundle("util", "CalendarData", bundle.getID(), true, calendarDataMap, false);
 369                 }
 370             }
 371             if (bundleTypes.contains(Bundle.Type.FORMATDATA)) {
 372                 Map<String, Object> formatDataMap = extractFormatData(targetMap, bundle.getID());
 373                 // LocaleData.getAvailableLocales depends on having FormatData bundles around
 374                 if (!formatDataMap.isEmpty() || bundle.isRoot()) {
 375                     metaInfo.get("FormatData").add(toLanguageTag(bundle.getID()));
 376                     bundleGenerator.generateBundle("text", "FormatData", bundle.getID(), true, formatDataMap, false);
 377                 }
 378             }

 379             // For testing
 380             SortedSet<String> allLocales = new TreeSet<>();
 381             allLocales.addAll(metaInfo.get("CurrencyNames"));
 382             allLocales.addAll(metaInfo.get("LocaleNames"));
 383             allLocales.addAll(metaInfo.get("CalendarData"));
 384             allLocales.addAll(metaInfo.get("FormatData"));
 385             metaInfo.put("All", allLocales);
 386         }
 387 
 388         bundleGenerator.generateMetaInfo(metaInfo);
 389     }
 390 
 391     /*
 392      * Returns the language portion of the given id.
 393      * If id is "root", "" is returned.
 394      */
 395     static String getLanguageCode(String id) {
 396         int index = id.indexOf('_');
 397         String lang = null;
 398         if (index != -1) {


 414         if (id.indexOf('@') != -1) {
 415             id = id.substring(0, id.indexOf('@'));
 416         }
 417         String[] tokens = id.split("_");
 418         for (int index = 1; index < tokens.length; ++index) {
 419             if (tokens[index].length() == 2
 420                     && Character.isLetter(tokens[index].charAt(0))
 421                     && Character.isLetter(tokens[index].charAt(1))) {
 422                 return tokens[index];
 423             }
 424         }
 425         return null;
 426     }
 427 
 428     private static class KeyComparator implements Comparator<String> {
 429         static KeyComparator INSTANCE = new KeyComparator();
 430 
 431         private KeyComparator() {
 432         }
 433 

 434         public int compare(String o1, String o2) {
 435             int len1 = o1.length();
 436             int len2 = o2.length();
 437             if (!isDigit(o1.charAt(0)) && !isDigit(o2.charAt(0))) {
 438                 // Shorter string comes first unless either starts with a digit.
 439                 if (len1 < len2) {
 440                     return -1;
 441                 }
 442                 if (len1 > len2) {
 443                     return 1;
 444                 }
 445             }
 446             return o1.compareTo(o2);
 447         }
 448 
 449         private boolean isDigit(char c) {
 450             return c >= '0' && c <= '9';
 451         }
 452     }
 453 


 459             }
 460         }
 461         return localeNames;
 462     }
 463 
 464     @SuppressWarnings("AssignmentToForLoopParameter")
 465     private static Map<String, Object> extractCurrencyNames(Map<String, Object> map, String id, String names)
 466             throws Exception {
 467         Map<String, Object> currencyNames = new TreeMap<>(KeyComparator.INSTANCE);
 468         for (String key : map.keySet()) {
 469             if (key.startsWith(CURRENCY_NAME_PREFIX)) {
 470                 currencyNames.put(key.substring(CURRENCY_NAME_PREFIX.length()), map.get(key));
 471             } else if (key.startsWith(CURRENCY_SYMBOL_PREFIX)) {
 472                 currencyNames.put(key.substring(CURRENCY_SYMBOL_PREFIX.length()), map.get(key));
 473             }
 474         }
 475         return currencyNames;
 476     }
 477 
 478     private static Map<String, Object> extractZoneNames(Map<String, Object> map, String id) {
 479         return null;



















 480     }
 481 
 482     private static Map<String, Object> extractCalendarData(Map<String, Object> map, String id) {
 483         Map<String, Object> calendarData = new LinkedHashMap<>();
 484         copyIfPresent(map, "firstDayOfWeek", calendarData);
 485         copyIfPresent(map, "minimalDaysInFirstWeek", calendarData);
 486         return calendarData;
 487     }
 488 
 489     private static Map<String, Object> extractFormatData(Map<String, Object> map, String id) {
 490         Map<String, Object> formatData = new LinkedHashMap<>();
 491         for (CalendarType calendarType : CalendarType.values()) {
 492             String prefix = calendarType.keyElementName();
 493             copyIfPresent(map, prefix + "MonthNames", formatData); // default FORMAT since JDK8
 494             copyIfPresent(map, prefix + "standalone.MonthNames", formatData);
 495             copyIfPresent(map, prefix + "MonthAbbreviations", formatData);
 496             copyIfPresent(map, prefix + "standalone.MonthAbbreviations", formatData);


 497             copyIfPresent(map, prefix + "DayNames", formatData);

 498             copyIfPresent(map, prefix + "DayAbbreviations", formatData);



 499             copyIfPresent(map, prefix + "AmPmMarkers", formatData);


 500             copyIfPresent(map, prefix + "Eras", formatData);
 501             copyIfPresent(map, prefix + "short.Eras", formatData);
 502             copyIfPresent(map, prefix + "TimePatterns", formatData);
 503             copyIfPresent(map, prefix + "DatePatterns", formatData);
 504             copyIfPresent(map, prefix + "DateTimePatterns", formatData);
 505             copyIfPresent(map, prefix + "DateTimePatternChars", formatData);
 506         }
 507 
 508         copyIfPresent(map, "DefaultNumberingSystem", formatData);
 509         String defaultScript = (String) map.get("DefaultNumberingSystem");
 510         @SuppressWarnings("unchecked")
 511         List<String> numberingScripts = (List<String>) map.remove("numberingScripts");
 512         if (numberingScripts != null) {
 513             for (String script : numberingScripts) {
 514                 copyIfPresent(map, script + "." + "NumberElements", formatData);
 515             }
 516         } else {
 517             copyIfPresent(map, "NumberElements", formatData);
 518         }
 519         copyIfPresent(map, "NumberPatterns", formatData);
 520         return formatData;
 521     }


 543 
 544         String specialSaveChars;
 545         if (useJava) {
 546             specialSaveChars = specialSaveCharsJava;
 547         } else {
 548             specialSaveChars = specialSaveCharsProperties;
 549         }
 550         boolean escapeSpace = false;
 551 
 552         int len = theString.length();
 553         StringBuilder outBuffer = new StringBuilder(len * 2);
 554         Formatter formatter = new Formatter(outBuffer, Locale.ROOT);
 555 
 556         for (int x = 0; x < len; x++) {
 557             char aChar = theString.charAt(x);
 558             switch (aChar) {
 559             case ' ':
 560                 if (x == 0 || escapeSpace) {
 561                     outBuffer.append('\\');
 562                 }
 563 
 564                 outBuffer.append(' ');
 565                 break;
 566             case '\\':
 567                 outBuffer.append('\\');
 568                 outBuffer.append('\\');
 569                 break;
 570             case '\t':
 571                 outBuffer.append('\\');
 572                 outBuffer.append('t');
 573                 break;
 574             case '\n':
 575                 outBuffer.append('\\');
 576                 outBuffer.append('n');
 577                 break;
 578             case '\r':
 579                 outBuffer.append('\\');
 580                 outBuffer.append('r');
 581                 break;
 582             case '\f':
 583                 outBuffer.append('\\');
 584                 outBuffer.append('f');
 585                 break;
 586             default:
 587                 if (!USE_UTF8 && ((aChar < 0x0020) || (aChar > 0x007e))) {
 588                     formatter.format("\\u%04x", (int)aChar);
 589                 } else {
 590                     if (specialSaveChars.indexOf(aChar) != -1) {
 591                         outBuffer.append('\\');
 592                     }
 593                     outBuffer.append(aChar);
 594                 }
 595             }
 596         }
 597         return outBuffer.toString();
 598     }
 599 
 600     private static String toLanguageTag(String locName) {
 601         if (locName.indexOf('_') == -1) {
 602             return locName;
 603         }
 604         String tag = locName.replaceAll("_", "-");
 605         Locale loc = Locale.forLanguageTag(tag);
 606         return loc.toLanguageTag();
 607     }


   8  * particular file as subject to the "Classpath" exception as provided
   9  * by Oracle in the LICENSE file that accompanied this code.
  10  *
  11  * This code is distributed in the hope that it will be useful, but WITHOUT
  12  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  13  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  14  * version 2 for more details (a copy is included in the LICENSE file that
  15  * accompanied this code).
  16  *
  17  * You should have received a copy of the GNU General Public License version
  18  * 2 along with this work; if not, write to the Free Software Foundation,
  19  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
  20  *
  21  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
  22  * or visit www.oracle.com if you need additional information or have any
  23  * questions.
  24  */
  25 
  26 package build.tools.cldrconverter;
  27 
  28 import build.tools.cldrconverter.BundleGenerator.BundleType;
  29 import java.io.File;
  30 import java.nio.file.DirectoryStream;
  31 import java.nio.file.FileSystems;
  32 import java.nio.file.Files;
  33 import java.nio.file.Path;
  34 import java.util.*;
  35 import javax.xml.parsers.SAXParser;
  36 import javax.xml.parsers.SAXParserFactory;
  37 
  38 
  39 /**
  40  * Converts locale data from "Locale Data Markup Language" format to
  41  * JRE resource bundle format. LDML is the format used by the Common
  42  * Locale Data Repository maintained by the Unicode Consortium.
  43  */
  44 public class CLDRConverter {
  45 
  46     static final String LDML_DTD_SYSTEM_ID = "http://www.unicode.org/cldr/dtd/2.0/ldml.dtd";
  47     static final String SPPL_LDML_DTD_SYSTEM_ID = "http://www.unicode.org/cldr/dtd/2.0/ldmlSupplemental.dtd";
  48 
  49     private static String CLDR_BASE = "../CLDR/21.0.1/";
  50     static String LOCAL_LDML_DTD;
  51     static String LOCAL_SPPL_LDML_DTD;
  52     private static String SOURCE_FILE_DIR;
  53     private static String SPPL_SOURCE_FILE;
  54     private static String NUMBERING_SOURCE_FILE;
  55     private static String METAZONES_SOURCE_FILE;
  56     static String DESTINATION_DIR = "build/gensrc";
  57 
  58     static final String LOCALE_NAME_PREFIX = "locale.displayname.";
  59     static final String CURRENCY_SYMBOL_PREFIX = "currency.symbol.";
  60     static final String CURRENCY_NAME_PREFIX = "currency.displayname.";
  61     static final String TIMEZONE_ID_PREFIX = "timezone.id.";
  62     static final String ZONE_NAME_PREFIX = "timezone.displayname.";
  63     static final String METAZONE_ID_PREFIX = "metazone.id.";

  64 
  65     private static SupplementDataParseHandler handlerSuppl;
  66     static NumberingSystemsParseHandler handlerNumbering;
  67     static MetaZonesParseHandler handlerMetaZones;
  68     private static BundleGenerator bundleGenerator;
  69 
  70     static int draftType;
  71     private static final String DRAFT_UNCONFIRMED = "unconfirmed";
  72     private static final String DRAFT_PROVISIONAL = "provisional";
  73     private static final String DRAFT_CONTRIBUTED = "contributed";
  74     private static final String DRAFT_APPROVED = "approved";
  75     private static final String DRAFT_TRUE = "true";
  76     private static final String DRAFT_FALSE = "false";
  77     private static final String DRAFT_DEFAULT = DRAFT_APPROVED;
  78     static final Map<String, Integer> DRAFT_MAP = new HashMap<>();
  79 
  80     static {
  81         DRAFT_MAP.put(DRAFT_UNCONFIRMED, 0);
  82         DRAFT_MAP.put(DRAFT_PROVISIONAL, 1);
  83         DRAFT_MAP.put(DRAFT_CONTRIBUTED, 2);


 219         ResourceBundle.Control defCon = ResourceBundle.Control.getControl(ResourceBundle.Control.FORMAT_DEFAULT);
 220         List<Bundle> retList = new ArrayList<>();
 221         Path path = FileSystems.getDefault().getPath(SOURCE_FILE_DIR);
 222         try (DirectoryStream<Path> dirStr = Files.newDirectoryStream(path)) {
 223             for (Path entry : dirStr) {
 224                 String fileName = entry.getFileName().toString();
 225                 if (fileName.endsWith(".xml")) {
 226                     String id = fileName.substring(0, fileName.indexOf('.'));
 227                     Locale cldrLoc = Locale.forLanguageTag(toLanguageTag(id));
 228                     List<Locale> candList = defCon.getCandidateLocales("", cldrLoc);
 229                     StringBuilder sb = new StringBuilder();
 230                     for (Locale loc : candList) {
 231                         if (!loc.equals(Locale.ROOT)) {
 232                             sb.append(toLocaleName(loc.toLanguageTag()));
 233                             sb.append(",");
 234                         }
 235                     }
 236                     if (sb.indexOf("root") == -1) {
 237                         sb.append("root");
 238                     }
 239                     Bundle b = new Bundle(id, sb.toString(), null, null);
 240                     // Insert the bundle for en at the top so that it will get
 241                     // processed first.
 242                     if ("en".equals(id)) {
 243                         retList.add(0, b);
 244                     } else {
 245                         retList.add(b);
 246                     }
 247                 }
 248             }
 249         }
 250         return retList;
 251     }
 252 
 253     private static Map<String, Map<String, Object>> cldrBundles = new HashMap<>();
 254 
 255     static Map<String, Object> getCLDRBundle(String id) throws Exception {
 256         Map<String, Object> bundle = cldrBundles.get(id);
 257         if (bundle != null) {
 258             return bundle;
 259         }
 260         SAXParserFactory factory = SAXParserFactory.newInstance();
 261         factory.setValidating(true);
 262         SAXParser parser = factory.newSAXParser();
 263         LDMLParseHandler handler = new LDMLParseHandler(id);
 264         File file = new File(SOURCE_FILE_DIR + File.separator + id + ".xml");
 265         if (!file.exists()) {
 266             // Skip if the file doesn't exist.


 302         // Parse numberingSystems to get digit zero character information.
 303         SAXParserFactory numberingParser = SAXParserFactory.newInstance();
 304         numberingParser.setValidating(true);
 305         SAXParser parserNumbering = numberingParser.newSAXParser();
 306         handlerNumbering = new NumberingSystemsParseHandler();
 307         File fileNumbering = new File(NUMBERING_SOURCE_FILE);
 308         parserNumbering.parse(fileNumbering, handlerNumbering);
 309 
 310         // Parse metaZones to create mappings between Olson tzids and CLDR meta zone names
 311         SAXParserFactory metazonesParser = SAXParserFactory.newInstance();
 312         metazonesParser.setValidating(true);
 313         SAXParser parserMetaZones = metazonesParser.newSAXParser();
 314         handlerMetaZones = new MetaZonesParseHandler();
 315         File fileMetaZones = new File(METAZONES_SOURCE_FILE);
 316         parserNumbering.parse(fileMetaZones, handlerMetaZones);
 317 
 318         // For generating information on supported locales.
 319         Map<String, SortedSet<String>> metaInfo = new HashMap<>();
 320         metaInfo.put("LocaleNames", new TreeSet<String>());
 321         metaInfo.put("CurrencyNames", new TreeSet<String>());
 322         metaInfo.put("TimeZoneNames", new TreeSet<String>());
 323         metaInfo.put("CalendarData", new TreeSet<String>());
 324         metaInfo.put("FormatData", new TreeSet<String>());
 325 
 326         for (Bundle bundle : bundles) {
 327             // Get the target map, which contains all the data that should be
 328             // visible for the bundle's locale
 329 
 330             Map<String, Object> targetMap = bundle.getTargetMap();
 331 
 332             EnumSet<Bundle.Type> bundleTypes = bundle.getBundleTypes();
 333 
 334             // Fill in any missing resources in the base bundle from en and en-US data.
 335             // This is because CLDR root.xml is supposed to be language neutral and doesn't
 336             // provide some resource data. Currently, the runtime assumes that there are all
 337             // resources though the parent resource bundle chain.
 338             if (bundle.isRoot()) {
 339                 Map<String, Object> enData = new HashMap<>();
 340                 // Create a superset of en-US and en bundles data in order to
 341                 // fill in any missing resources in the base bundle.
 342                 enData.putAll(Bundle.getBundle("en").getTargetMap());
 343                 enData.putAll(Bundle.getBundle("en_US").getTargetMap());
 344                 for (String key : enData.keySet()) {
 345                     if (!targetMap.containsKey(key)) {
 346                         targetMap.put(key, enData.get(key));
 347                     }
 348                 }
 349                 // Add DateTimePatternChars because CLDR no longer supports localized patterns.
 350                 targetMap.put("DateTimePatternChars", "GyMdkHmsSEDFwWahKzZ");
 351             }
 352 
 353             // Now the map contains just the entries that need to be in the resources bundles.
 354             // Go ahead and generate them.
 355             if (bundleTypes.contains(Bundle.Type.LOCALENAMES)) {
 356                 Map<String, Object> localeNamesMap = extractLocaleNames(targetMap, bundle.getID());
 357                 if (!localeNamesMap.isEmpty() || bundle.isRoot()) {
 358                     metaInfo.get("LocaleNames").add(toLanguageTag(bundle.getID()));
 359                     bundleGenerator.generateBundle("util", "LocaleNames", bundle.getID(), true, localeNamesMap, BundleType.OPEN);
 360                 }
 361             }
 362             if (bundleTypes.contains(Bundle.Type.CURRENCYNAMES)) {
 363                 Map<String, Object> currencyNamesMap = extractCurrencyNames(targetMap, bundle.getID(), bundle.getCurrencies());
 364                 if (!currencyNamesMap.isEmpty() || bundle.isRoot()) {
 365                     metaInfo.get("CurrencyNames").add(toLanguageTag(bundle.getID()));
 366                     bundleGenerator.generateBundle("util", "CurrencyNames", bundle.getID(), true, currencyNamesMap, BundleType.OPEN);
 367                 }
 368             }
 369             if (bundleTypes.contains(Bundle.Type.TIMEZONENAMES)) {
 370                 Map<String, Object> zoneNamesMap = extractZoneNames(targetMap, bundle.getID());
 371                 if (!zoneNamesMap.isEmpty() || bundle.isRoot()) {
 372                     metaInfo.get("TimeZoneNames").add(toLanguageTag(bundle.getID()));
 373                     bundleGenerator.generateBundle("util", "TimeZoneNames", bundle.getID(), true, zoneNamesMap, BundleType.TIMEZONE);
 374                 }
 375             }
 376             if (bundleTypes.contains(Bundle.Type.CALENDARDATA)) {
 377                 Map<String, Object> calendarDataMap = extractCalendarData(targetMap, bundle.getID());
 378                 if (!calendarDataMap.isEmpty() || bundle.isRoot()) {
 379                     metaInfo.get("CalendarData").add(toLanguageTag(bundle.getID()));
 380                     bundleGenerator.generateBundle("util", "CalendarData", bundle.getID(), true, calendarDataMap, BundleType.PLAIN);
 381                 }
 382             }
 383             if (bundleTypes.contains(Bundle.Type.FORMATDATA)) {
 384                 Map<String, Object> formatDataMap = extractFormatData(targetMap, bundle.getID());
 385                 // LocaleData.getAvailableLocales depends on having FormatData bundles around
 386                 if (!formatDataMap.isEmpty() || bundle.isRoot()) {
 387                     metaInfo.get("FormatData").add(toLanguageTag(bundle.getID()));
 388                     bundleGenerator.generateBundle("text", "FormatData", bundle.getID(), true, formatDataMap, BundleType.OPEN);
 389                 }
 390             }
 391 
 392             // For testing
 393             SortedSet<String> allLocales = new TreeSet<>();
 394             allLocales.addAll(metaInfo.get("CurrencyNames"));
 395             allLocales.addAll(metaInfo.get("LocaleNames"));
 396             allLocales.addAll(metaInfo.get("CalendarData"));
 397             allLocales.addAll(metaInfo.get("FormatData"));
 398             metaInfo.put("All", allLocales);
 399         }
 400 
 401         bundleGenerator.generateMetaInfo(metaInfo);
 402     }
 403 
 404     /*
 405      * Returns the language portion of the given id.
 406      * If id is "root", "" is returned.
 407      */
 408     static String getLanguageCode(String id) {
 409         int index = id.indexOf('_');
 410         String lang = null;
 411         if (index != -1) {


 427         if (id.indexOf('@') != -1) {
 428             id = id.substring(0, id.indexOf('@'));
 429         }
 430         String[] tokens = id.split("_");
 431         for (int index = 1; index < tokens.length; ++index) {
 432             if (tokens[index].length() == 2
 433                     && Character.isLetter(tokens[index].charAt(0))
 434                     && Character.isLetter(tokens[index].charAt(1))) {
 435                 return tokens[index];
 436             }
 437         }
 438         return null;
 439     }
 440 
 441     private static class KeyComparator implements Comparator<String> {
 442         static KeyComparator INSTANCE = new KeyComparator();
 443 
 444         private KeyComparator() {
 445         }
 446 
 447         @Override
 448         public int compare(String o1, String o2) {
 449             int len1 = o1.length();
 450             int len2 = o2.length();
 451             if (!isDigit(o1.charAt(0)) && !isDigit(o2.charAt(0))) {
 452                 // Shorter string comes first unless either starts with a digit.
 453                 if (len1 < len2) {
 454                     return -1;
 455                 }
 456                 if (len1 > len2) {
 457                     return 1;
 458                 }
 459             }
 460             return o1.compareTo(o2);
 461         }
 462 
 463         private boolean isDigit(char c) {
 464             return c >= '0' && c <= '9';
 465         }
 466     }
 467 


 473             }
 474         }
 475         return localeNames;
 476     }
 477 
 478     @SuppressWarnings("AssignmentToForLoopParameter")
 479     private static Map<String, Object> extractCurrencyNames(Map<String, Object> map, String id, String names)
 480             throws Exception {
 481         Map<String, Object> currencyNames = new TreeMap<>(KeyComparator.INSTANCE);
 482         for (String key : map.keySet()) {
 483             if (key.startsWith(CURRENCY_NAME_PREFIX)) {
 484                 currencyNames.put(key.substring(CURRENCY_NAME_PREFIX.length()), map.get(key));
 485             } else if (key.startsWith(CURRENCY_SYMBOL_PREFIX)) {
 486                 currencyNames.put(key.substring(CURRENCY_SYMBOL_PREFIX.length()), map.get(key));
 487             }
 488         }
 489         return currencyNames;
 490     }
 491 
 492     private static Map<String, Object> extractZoneNames(Map<String, Object> map, String id) {
 493         Map<String, Object> names = new HashMap<>();
 494         for (String tzid : handlerMetaZones.keySet()) {
 495             String tzKey = TIMEZONE_ID_PREFIX + tzid;
 496             Object data = map.get(tzKey);
 497             if (data instanceof String[]) {
 498                 names.put(tzid, data);
 499             } else {
 500                 String meta = handlerMetaZones.get(tzid);
 501                 if (meta != null) {
 502                     String metaKey = METAZONE_ID_PREFIX + meta;
 503                     data = map.get(metaKey);
 504                     if (data instanceof String[]) {
 505                         // Keep the metazone prefix here.
 506                         names.put(metaKey, data);
 507                         names.put(tzid, meta);
 508                     }
 509                 }
 510             }
 511         }
 512         return names;
 513     }
 514 
 515     private static Map<String, Object> extractCalendarData(Map<String, Object> map, String id) {
 516         Map<String, Object> calendarData = new LinkedHashMap<>();
 517         copyIfPresent(map, "firstDayOfWeek", calendarData);
 518         copyIfPresent(map, "minimalDaysInFirstWeek", calendarData);
 519         return calendarData;
 520     }
 521 
 522     private static Map<String, Object> extractFormatData(Map<String, Object> map, String id) {
 523         Map<String, Object> formatData = new LinkedHashMap<>();
 524         for (CalendarType calendarType : CalendarType.values()) {
 525             String prefix = calendarType.keyElementName();
 526             copyIfPresent(map, prefix + "MonthNames", formatData); // default FORMAT since JDK8
 527             copyIfPresent(map, prefix + "standalone.MonthNames", formatData);
 528             copyIfPresent(map, prefix + "MonthAbbreviations", formatData);
 529             copyIfPresent(map, prefix + "standalone.MonthAbbreviations", formatData);
 530             copyIfPresent(map, prefix + "MonthNarrow", formatData);
 531             copyIfPresent(map, prefix + "standalone.MonthNarrows", formatData);
 532             copyIfPresent(map, prefix + "DayNames", formatData);
 533             copyIfPresent(map, prefix + "standalone.DayNames", formatData);
 534             copyIfPresent(map, prefix + "DayAbbreviations", formatData);
 535             copyIfPresent(map, prefix + "standalone.DayAbbreviations", formatData);
 536             copyIfPresent(map, prefix + "DayNarrows", formatData);
 537             copyIfPresent(map, prefix + "standalone.DayNarrows", formatData);
 538             copyIfPresent(map, prefix + "AmPmMarkers", formatData);
 539             copyIfPresent(map, prefix + "narrow.AmPmMarkers", formatData);
 540             copyIfPresent(map, prefix + "long.Eras", formatData);
 541             copyIfPresent(map, prefix + "Eras", formatData);
 542             copyIfPresent(map, prefix + "narrow.Eras", formatData);
 543             copyIfPresent(map, prefix + "TimePatterns", formatData);
 544             copyIfPresent(map, prefix + "DatePatterns", formatData);
 545             copyIfPresent(map, prefix + "DateTimePatterns", formatData);
 546             copyIfPresent(map, prefix + "DateTimePatternChars", formatData);
 547         }
 548 
 549         copyIfPresent(map, "DefaultNumberingSystem", formatData);
 550         String defaultScript = (String) map.get("DefaultNumberingSystem");
 551         @SuppressWarnings("unchecked")
 552         List<String> numberingScripts = (List<String>) map.remove("numberingScripts");
 553         if (numberingScripts != null) {
 554             for (String script : numberingScripts) {
 555                 copyIfPresent(map, script + "." + "NumberElements", formatData);
 556             }
 557         } else {
 558             copyIfPresent(map, "NumberElements", formatData);
 559         }
 560         copyIfPresent(map, "NumberPatterns", formatData);
 561         return formatData;
 562     }


 584 
 585         String specialSaveChars;
 586         if (useJava) {
 587             specialSaveChars = specialSaveCharsJava;
 588         } else {
 589             specialSaveChars = specialSaveCharsProperties;
 590         }
 591         boolean escapeSpace = false;
 592 
 593         int len = theString.length();
 594         StringBuilder outBuffer = new StringBuilder(len * 2);
 595         Formatter formatter = new Formatter(outBuffer, Locale.ROOT);
 596 
 597         for (int x = 0; x < len; x++) {
 598             char aChar = theString.charAt(x);
 599             switch (aChar) {
 600             case ' ':
 601                 if (x == 0 || escapeSpace) {
 602                     outBuffer.append('\\');
 603                 }

 604                 outBuffer.append(' ');
 605                 break;
 606             case '\\':
 607                 outBuffer.append('\\');
 608                 outBuffer.append('\\');
 609                 break;
 610             case '\t':
 611                 outBuffer.append('\\');
 612                 outBuffer.append('t');
 613                 break;
 614             case '\n':
 615                 outBuffer.append('\\');
 616                 outBuffer.append('n');
 617                 break;
 618             case '\r':
 619                 outBuffer.append('\\');
 620                 outBuffer.append('r');
 621                 break;
 622             case '\f':
 623                 outBuffer.append('\\');
 624                 outBuffer.append('f');
 625                 break;
 626             default:
 627                 if (aChar < 0x0020 || (!USE_UTF8 && aChar > 0x007e)) {
 628                     formatter.format("\\u%04x", (int)aChar);
 629                 } else {
 630                     if (specialSaveChars.indexOf(aChar) != -1) {
 631                         outBuffer.append('\\');
 632                     }
 633                     outBuffer.append(aChar);
 634                 }
 635             }
 636         }
 637         return outBuffer.toString();
 638     }
 639 
 640     private static String toLanguageTag(String locName) {
 641         if (locName.indexOf('_') == -1) {
 642             return locName;
 643         }
 644         String tag = locName.replaceAll("_", "-");
 645         Locale loc = Locale.forLanguageTag(tag);
 646         return loc.toLanguageTag();
 647     }