icuSources/test/intltest/transrt.cpp

   1 /*
   2 **********************************************************************
   3 *   Copyright (C) 2000-2004, International Business Machines
   4 *   Corporation and others.  All Rights Reserved.
   5 **********************************************************************
   6 *   Date        Name        Description
   7 *   05/23/00    aliu        Creation.
   8 **********************************************************************
   9 */
  10
  11 #include "unicode/utypes.h"
  12
  13 #if !UCONFIG_NO_TRANSLITERATION
  14
  15 #include "unicode/translit.h"
  16 #include "rbt.h"
  17 #include "unicode/calendar.h"
  18 #include "unicode/uniset.h"
  19 #include "unicode/uchar.h"
  20 #include "unicode/normlzr.h"
  21 #include "unicode/uchar.h"
  22 #include "unicode/parseerr.h"
  23 #include "unicode/usetiter.h"
  24 #include "unicode/putil.h"
  25 #include "unicode/uversion.h"
  26 #include "unicode/locid.h"
  27 #include "unicode/ulocdata.h"
  28 #include "unicode/utf8.h"
  29 #include "putilimp.h"
  30 #include "cmemory.h"
  31 #include "transrt.h"
  32 #include "testutil.h"
  33 #include <string.h>
  34 #include <stdio.h>
  35
  36 #define CASE(id,test) case id:                          \
  37                           name = #test;                 \
  38                           if (exec) {                   \
  39                               logln(#test "---");       \
  40                               logln((UnicodeString)""); \
  41                               UDate t = uprv_getUTCtime(); \
  42                               test();                   \
  43                               t = uprv_getUTCtime() - t; \
  44                               logln((UnicodeString)#test " took " + t/U_MILLIS_PER_DAY + " seconds"); \
  45                           }                             \
  46                           break
  47
  48 #define EXHAUSTIVE(id,test) case id:                            \
  49                               if(quick==FALSE){                 \
  50                                   name = #test;                 \
  51                                   if (exec){                    \
  52                                       logln(#test "---");       \
  53                                       logln((UnicodeString)""); \
  54                                       test();                   \
  55                                   }                             \
  56                               }else{                            \
  57                                 name="";                        \
  58                               }                                 \
  59                               break
  60 void
  61 TransliteratorRoundTripTest::runIndexedTest(int32_t index, UBool exec,
  62                                    const char* &name, char* /*par*/) {
  63     switch (index) {
  64         CASE(0, TestCyrillic);
  65         // CASE(0,TestKana);
  66         CASE(1,TestHiragana);
  67         CASE(2,TestKatakana);
  68         CASE(3,TestJamo);
  69         CASE(4,TestHangul);
  70         CASE(5,TestGreek);
  71         CASE(6,TestGreekUNGEGN);
  72         CASE(7,Testel);
  73         CASE(8,TestDevanagariLatin);
  74         CASE(9,TestInterIndic);
  75         CASE(10, TestHebrew);
  76         CASE(11, TestArabic);
  77         CASE(12, TestHan);
  78         default: name = ""; break;
  79     }
  80 }
  81
  82 //--------------------------------------------------------------------
  83 // Time bomb - allows temporary behavior that expires at a given
  84 //             release
  85 //--------------------------------------------------------------------
  86 /*
  87 static const UVersionInfo ICU_30 = {3,0,0,0};
  88
  89 static UBool isICUVersionAtLeast(const UVersionInfo x) {
  90     UVersionInfo v;
  91     u_getVersion(v);
  92     return (uprv_memcmp(v, x, U_MAX_VERSION_LENGTH) >= 0);
  93 }
  94 */
  95
  96 //--------------------------------------------------------------------
  97 // TransliteratorPointer
  98 //--------------------------------------------------------------------
  99
 100 /**
 101  * A transliterator pointer wrapper that deletes the contained
 102  * pointer automatically when the wrapper goes out of scope.
 103  * Sometimes called a "janitor" or "smart pointer".
 104  */
 105 class TransliteratorPointer {
 106     Transliterator* t;
 107     // disallowed:
 108     TransliteratorPointer(const TransliteratorPointer& rhs);
 109     TransliteratorPointer& operator=(const TransliteratorPointer& rhs);
 110 public:
 111     TransliteratorPointer(Transliterator* adopted) {
 112         t = adopted;
 113     }
 114     ~TransliteratorPointer() {
 115         delete t;
 116     }
 117     inline Transliterator* operator->() { return t; }
 118     inline operator const Transliterator*() const { return t; }
 119     inline operator Transliterator*() { return t; }
 120 };
 121
 122 //--------------------------------------------------------------------
 123 // Legal
 124 //--------------------------------------------------------------------
 125
 126 class Legal {
 127 public:
 128     Legal() {}
 129     virtual ~Legal() {}
 130     virtual UBool is(const UnicodeString& /*sourceString*/) const {return TRUE;}
 131 };
 132
 133 class LegalJamo : public Legal {
 134     // any initial must be followed by a medial (or initial)
 135     // any medial must follow an initial (or medial)
 136     // any final must follow a medial (or final)
 137 public:
 138     LegalJamo() {}
 139     virtual ~LegalJamo() {}
 140     virtual UBool is(const UnicodeString& sourceString) const;
 141             int   getType(UChar c) const;
 142 };
 143
 144 UBool LegalJamo::is(const UnicodeString& sourceString) const {
 145     int t;
 146     UnicodeString decomp;
 147     UErrorCode ec = U_ZERO_ERROR;
 148     Normalizer::decompose(sourceString, FALSE, 0, decomp, ec);
 149     if (U_FAILURE(ec)) {
 150         return FALSE;
 151     }
 152     for (int i = 0; i < decomp.length(); ++i) { // don't worry about surrogates
 153         switch (getType(decomp.charAt(i))) {
 154         case 0: t = getType(decomp.charAt(i+1));
 155                 if (t != 0 && t != 1) { return FALSE; }
 156                 break;
 157         case 1: t = getType(decomp.charAt(i-1));
 158                 if (t != 0 && t != 1) { return FALSE; }
 159                 break;
 160         case 2: t = getType(decomp.charAt(i-1));
 161                 if (t != 1 && t != 2) { return FALSE; }
 162                 break;
 163         }
 164     }
 165     return TRUE;
 166 }
 167
 168 int LegalJamo::getType(UChar c) const {
 169     if (0x1100 <= c && c <= 0x1112)
 170         return 0;
 171     else if (0x1161 <= c && c  <= 0x1175)
 172              return 1;
 173          else if (0x11A8 <= c && c  <= 0x11C2)
 174                   return 2;
 175     return -1; // other
 176 }
 177
 178 class LegalGreek : public Legal {
 179     UBool full;
 180 public:
 181     LegalGreek(UBool _full) { full = _full; }
 182     virtual ~LegalGreek() {}
 183
 184     virtual UBool is(const UnicodeString& sourceString) const;
 185
 186     static UBool isVowel(UChar c);
 187
 188     static UBool isRho(UChar c);
 189 };
 190
 191 UBool LegalGreek::is(const UnicodeString& sourceString) const {
 192     UnicodeString decomp;
 193     UErrorCode ec = U_ZERO_ERROR;
 194     Normalizer::decompose(sourceString, FALSE, 0, decomp, ec);
 195
 196     // modern is simpler: don't care about anything but a grave
 197     if (full == FALSE) {
 198         // A special case which is legal but should be
 199         // excluded from round trip
 200         // if (sourceString == UnicodeString("\\u039C\\u03C0", "")) {
 201         //    return FALSE;
 202         // }
 203         for (int32_t i = 0; i < decomp.length(); ++i) {
 204             UChar c = decomp.charAt(i);
 205             // exclude all the accents
 206             if (c == 0x0313 || c == 0x0314 || c == 0x0300 || c == 0x0302
 207                 || c == 0x0342 || c == 0x0345
 208                 ) return FALSE;
 209         }
 210         return TRUE;
 211     }
 212
 213     // Legal greek has breathing marks IFF there is a vowel or RHO at the start
 214     // IF it has them, it has exactly one.
 215     // IF it starts with a RHO, then the breathing mark must come before the second letter.
 216     // Since there are no surrogates in greek, don't worry about them
 217     UBool firstIsVowel = FALSE;
 218     UBool firstIsRho = FALSE;
 219     UBool noLetterYet = TRUE;
 220     int32_t breathingCount = 0;
 221     int32_t letterCount = 0;
 222     for (int32_t i = 0; i < decomp.length(); ++i) {
 223         UChar c = decomp.charAt(i);
 224         if (u_isalpha(c)) {
 225             ++letterCount;
 226             if (noLetterYet) {
 227                 noLetterYet =  FALSE;
 228                 firstIsVowel = isVowel(c);
 229                 firstIsRho = isRho(c);
 230             }
 231             if (firstIsRho && letterCount == 2 && breathingCount == 0) {
 232                 return FALSE;
 233             }
 234         }
 235         if (c == 0x0313 || c == 0x0314) {
 236             ++breathingCount;
 237         }
 238     }
 239
 240     if (firstIsVowel || firstIsRho) return breathingCount == 1;
 241     return breathingCount == 0;
 242 }
 243
 244 UBool LegalGreek::isVowel(UChar c) {
 245     switch (c) {
 246     case 0x03B1:
 247     case 0x03B5:
 248     case 0x03B7:
 249     case 0x03B9:
 250     case 0x03BF:
 251     case 0x03C5:
 252     case 0x03C9:
 253     case 0x0391:
 254     case 0x0395:
 255     case 0x0397:
 256     case 0x0399:
 257     case 0x039F:
 258     case 0x03A5:
 259     case 0x03A9:
 260         return TRUE;
 261     }
 262     return FALSE;
 263 }
 264
 265 UBool LegalGreek::isRho(UChar c) {
 266     switch (c) {
 267     case 0x03C1:
 268     case 0x03A1:
 269         return TRUE;
 270     }
 271     return FALSE;
 272 }
 273
 274 // AbbreviatedUnicodeSetIterator Interface ---------------------------------------------
 275
 276 class AbbreviatedUnicodeSetIterator : public UnicodeSetIterator {
 277 public :
 278
 279     AbbreviatedUnicodeSetIterator();
 280     virtual ~AbbreviatedUnicodeSetIterator();
 281     void reset(UnicodeSet& set, UBool abb = FALSE, int32_t density = 100);
 282
 283     /**
 284      * ICU "poor man's RTTI", returns a UClassID for the actual class.
 285      */
 286     virtual inline UClassID getDynamicClassID() const { return getStaticClassID(); }
 287
 288     /**
 289      * ICU "poor man's RTTI", returns a UClassID for this class.
 290      */
 291     static inline UClassID getStaticClassID() { return (UClassID)&fgClassID; }
 292
 293 private :
 294     UBool abbreviated;
 295     int32_t perRange;
 296     virtual void loadRange(int32_t range);
 297
 298     /**
 299      * The address of this static class variable serves as this class's ID
 300      * for ICU "poor man's RTTI".
 301      */
 302     static const char fgClassID;
 303 };
 304
 305 // AbbreviatedUnicodeSetIterator Implementation ---------------------------------------
 306
 307 const char AbbreviatedUnicodeSetIterator::fgClassID=0;
 308
 309 AbbreviatedUnicodeSetIterator::AbbreviatedUnicodeSetIterator() :
 310     UnicodeSetIterator(), abbreviated(FALSE) {
 311 }
 312
 313 AbbreviatedUnicodeSetIterator::~AbbreviatedUnicodeSetIterator() {
 314 }
 315
 316 void AbbreviatedUnicodeSetIterator::reset(UnicodeSet& newSet, UBool abb, int32_t density) {
 317     UnicodeSetIterator::reset(newSet);
 318     abbreviated = abb;
 319     perRange = newSet.getRangeCount();
 320     if (perRange != 0) {
 321         perRange = density / perRange;
 322     }
 323 }
 324
 325 void AbbreviatedUnicodeSetIterator::loadRange(int32_t myRange) {
 326     UnicodeSetIterator::loadRange(myRange);
 327     if (abbreviated && (endElement > nextElement + perRange)) {
 328         endElement = nextElement + perRange;
 329     }
 330 }
 331
 332 //--------------------------------------------------------------------
 333 // RTTest Interface
 334 //--------------------------------------------------------------------
 335
 336 class RTTest : public IntlTest {
 337
 338     // PrintWriter out;
 339
 340     UnicodeString transliteratorID;
 341     int32_t errorLimit;
 342     int32_t errorCount;
 343     int32_t pairLimit;
 344     UnicodeSet sourceRange;
 345     UnicodeSet targetRange;
 346     UnicodeSet toSource;
 347     UnicodeSet toTarget;
 348     UnicodeSet roundtripExclusionsSet;
 349     IntlTest* parent;
 350     Legal* legalSource; // NOT owned
 351     UnicodeSet badCharacters;
 352
 353 public:
 354
 355     /*
 356      * create a test for the given script transliterator.
 357      */
 358     RTTest(const UnicodeString& transliteratorIDStr);
 359
 360     virtual ~RTTest();
 361
 362     void setErrorLimit(int32_t limit);
 363
 364     void setPairLimit(int32_t limit);
 365
 366     void test(const UnicodeString& sourceRange,
 367               const UnicodeString& targetRange,
 368               const char* roundtripExclusions,
 369               IntlTest* parent,
 370               UBool     quick,
 371               Legal* adoptedLegal,
 372               int32_t density = 100);
 373
 374 private:
 375
 376     // Added to do better equality check.
 377
 378     static UBool isSame(const UnicodeString& a, const UnicodeString& b);
 379
 380     static UBool isCamel(const UnicodeString& a);
 381
 382     UBool checkIrrelevants(Transliterator *t, const UnicodeString& irrelevants);
 383
 384     void test2(UBool quick, int32_t density);
 385
 386     void logWrongScript(const UnicodeString& label,
 387                         const UnicodeString& from,
 388                         const UnicodeString& to);
 389
 390     void logNotCanonical(const UnicodeString& label,
 391                          const UnicodeString& from,
 392                          const UnicodeString& to,
 393                          const UnicodeString& fromCan,
 394                          const UnicodeString& toCan);
 395
 396     void logFails(const UnicodeString& label);
 397
 398     void logToRulesFails(const UnicodeString& label,
 399                          const UnicodeString& from,
 400                          const UnicodeString& to,
 401                          const UnicodeString& toCan);
 402
 403     void logRoundTripFailure(const UnicodeString& from,
 404                              const UnicodeString& toID,
 405                              const UnicodeString& to,
 406                              const UnicodeString& backID,
 407                              const UnicodeString& back);
 408 };
 409
 410 //--------------------------------------------------------------------
 411 // RTTest Implementation
 412 //--------------------------------------------------------------------
 413
 414 /*
 415  * create a test for the given script transliterator.
 416  */
 417 RTTest::RTTest(const UnicodeString& transliteratorIDStr) {
 418     transliteratorID = transliteratorIDStr;
 419     errorLimit = 500;
 420     errorCount = 0;
 421     pairLimit  = 0x10000;
 422 }
 423
 424 RTTest::~RTTest() {
 425 }
 426
 427 void RTTest::setErrorLimit(int32_t limit) {
 428     errorLimit = limit;
 429 }
 430
 431 void RTTest::setPairLimit(int32_t limit) {
 432     pairLimit = limit;
 433 }
 434
 435 UBool RTTest::isSame(const UnicodeString& a, const UnicodeString& b) {
 436     if (a == b) return TRUE;
 437     if (a.caseCompare(b, U_FOLD_CASE_DEFAULT)==0 && isCamel(a)) return TRUE;
 438     UnicodeString aa, bb;
 439     UErrorCode ec = U_ZERO_ERROR;
 440     Normalizer::decompose(a, FALSE, 0, aa, ec);
 441     Normalizer::decompose(b, FALSE, 0, bb, ec);
 442     if (aa == bb) return TRUE;
 443     if (aa.caseCompare(bb, U_FOLD_CASE_DEFAULT)==0 && isCamel(aa)) return TRUE;
 444     return FALSE;
 445 }
 446
 447 UBool RTTest::isCamel(const UnicodeString& a) {
 448     // see if string is of the form aB; e.g. lower, then upper or title
 449     UChar32 cp;
 450     UBool haveLower = FALSE;
 451     for (int32_t i = 0; i < a.length(); i += UTF_CHAR_LENGTH(cp)) {
 452         cp = a.char32At(i);
 453         int8_t t = u_charType(cp);
 454         switch (t) {
 455         case U_UPPERCASE_LETTER:
 456             if (haveLower) return TRUE;
 457             break;
 458         case U_TITLECASE_LETTER:
 459             if (haveLower) return TRUE;
 460             // drop through, since second letter is lower.
 461         case U_LOWERCASE_LETTER:
 462             haveLower = TRUE;
 463             break;
 464         }
 465     }
 466     return FALSE;
 467 }
 468
 469 void RTTest::test(const UnicodeString& sourceRangeVal,
 470                   const UnicodeString& targetRangeVal,
 471                   const char* roundtripExclusions,
 472                   IntlTest* logVal, UBool quickRt,
 473                   Legal* adoptedLegal,
 474                   int32_t density)
 475 {
 476
 477     UErrorCode status = U_ZERO_ERROR;
 478
 479     this->parent = logVal;
 480     this->legalSource = adoptedLegal;
 481
 482     UnicodeSet neverOk("[:Other:]", status);
 483     UnicodeSet okAnyway("[^[:Letter:]]", status);
 484
 485     if (U_FAILURE(status)) {
 486         parent->errln("FAIL: Initializing UnicodeSet with [:Other:] or [^[:Letter:]]");
 487         return;
 488     }
 489
 490     this->sourceRange.clear();
 491     this->sourceRange.applyPattern(sourceRangeVal, status);
 492     if (U_FAILURE(status)) {
 493         parent->errln("FAIL: UnicodeSet::applyPattern(" +
 494                    sourceRangeVal + ")");
 495         return;
 496     }
 497     this->sourceRange.removeAll(neverOk);
 498
 499     this->targetRange.clear();
 500     this->targetRange.applyPattern(targetRangeVal, status);
 501     if (U_FAILURE(status)) {
 502         parent->errln("FAIL: UnicodeSet::applyPattern(" +
 503                    targetRangeVal + ")");
 504         return;
 505     }
 506     this->targetRange.removeAll(neverOk);
 507
 508     this->toSource.clear();
 509     this->toSource.applyPattern(sourceRangeVal, status);
 510     if (U_FAILURE(status)) {
 511         parent->errln("FAIL: UnicodeSet::applyPattern(" +
 512                    sourceRangeVal + ")");
 513         return;
 514     }
 515     this->toSource.addAll(okAnyway);
 516
 517     this->toTarget.clear();
 518     this->toTarget.applyPattern(targetRangeVal, status);
 519     if (U_FAILURE(status)) {
 520         parent->errln("FAIL: UnicodeSet::applyPattern(" +
 521                    targetRangeVal + ")");
 522         return;
 523     }
 524     this->toTarget.addAll(okAnyway);
 525
 526     this->roundtripExclusionsSet.clear();
 527     if (roundtripExclusions != NULL && strlen(roundtripExclusions) > 0) {
 528         this->roundtripExclusionsSet.applyPattern(roundtripExclusions, status);
 529         if (U_FAILURE(status)) {
 530             parent->errln("FAIL: UnicodeSet::applyPattern(%s)", roundtripExclusions);
 531             return;
 532         }
 533     }
 534
 535     badCharacters.clear();
 536     badCharacters.applyPattern("[:Other:]", status);
 537     if (U_FAILURE(status)) {
 538         parent->errln("FAIL: UnicodeSet::applyPattern([:Other:])");
 539         return;
 540     }
 541
 542     test2(quickRt, density);
 543
 544     if (errorCount > 0) {
 545         char str[100];
 546         int32_t length = transliteratorID.extract(str, 100, NULL, status);
 547         str[length] = 0;
 548         parent->errln("FAIL: %s errors: %d %s", str, errorCount, (errorCount > errorLimit ? " (at least!)" : " ")); // + ", see " + logFileName);
 549     } else {
 550         char str[100];
 551         int32_t length = transliteratorID.extract(str, 100, NULL, status);
 552         str[length] = 0;
 553         parent->logln("%s ok", str);
 554     }
 555 }
 556
 557 UBool RTTest::checkIrrelevants(Transliterator *t,
 558                                const UnicodeString& irrelevants) {
 559     for (int i = 0; i < irrelevants.length(); ++i) {
 560         UChar c = irrelevants.charAt(i);
 561         UnicodeString srcStr(c);
 562         UnicodeString targ = srcStr;
 563         t->transliterate(targ);
 564         if (srcStr == targ) return TRUE;
 565     }
 566     return FALSE;
 567 }
 568
 569 void RTTest::test2(UBool quickRt, int32_t density) {
 570
 571     UnicodeString srcStr, targ, reverse;
 572     UErrorCode status = U_ZERO_ERROR;
 573     UParseError parseError ;
 574     TransliteratorPointer sourceToTarget(
 575         Transliterator::createInstance(transliteratorID, UTRANS_FORWARD, parseError,
 576                                        status));
 577     if ((Transliterator *)sourceToTarget == NULL) {
 578         parent->errln("FAIL: createInstance(" + transliteratorID +
 579                    ") returned NULL. Error: " + u_errorName(status)
 580                    + "\n\tpreContext : " + prettify(parseError.preContext)
 581                    + "\n\tpostContext : " + prettify(parseError.postContext));
 582
 583                 return;
 584     }
 585     TransliteratorPointer targetToSource(sourceToTarget->createInverse(status));
 586     if ((Transliterator *)targetToSource == NULL) {
 587         parent->errln("FAIL: " + transliteratorID +
 588                    ".createInverse() returned NULL. Error:" + u_errorName(status)
 589                    + "\n\tpreContext : " + prettify(parseError.preContext)
 590                    + "\n\tpostContext : " + prettify(parseError.postContext));
 591         return;
 592     }
 593
 594     AbbreviatedUnicodeSetIterator usi;
 595     AbbreviatedUnicodeSetIterator usi2;
 596
 597     parent->logln("Checking that at least one irrelevant character is not NFC'ed");
 598     // string is from NFC_NO in the UCD
 599     UnicodeString irrelevants = CharsToUnicodeString("\\u2000\\u2001\\u2126\\u212A\\u212B\\u2329");
 600
 601     if (checkIrrelevants(sourceToTarget, irrelevants) == FALSE) {
 602         logFails("Source-Target, irrelevants");
 603     }
 604     if (checkIrrelevants(targetToSource, irrelevants) == FALSE) {
 605         logFails("Target-Source, irrelevants");
 606     }
 607
 608     if (!quickRt){
 609       parent->logln("Checking that toRules works");
 610       UnicodeString rules = "";
 611
 612       UParseError parseError;
 613       rules = sourceToTarget->toRules(rules, TRUE);
 614       // parent->logln((UnicodeString)"toRules => " + rules);
 615       TransliteratorPointer sourceToTarget2(Transliterator::createFromRules(
 616                                                        "s2t2", rules,
 617                                                        UTRANS_FORWARD,
 618                                                        parseError, status));
 619       if (U_FAILURE(status)) {
 620           parent->errln("FAIL: createFromRules %s\n", u_errorName(status));
 621           return;
 622       }
 623
 624       rules = targetToSource->toRules(rules, FALSE);
 625       TransliteratorPointer targetToSource2(Transliterator::createFromRules(
 626                                                        "t2s2", rules,
 627                                                        UTRANS_FORWARD,
 628                                                        parseError, status));
 629       if (U_FAILURE(status)) {
 630           parent->errln("FAIL: createFromRules %s\n", u_errorName(status));
 631           return;
 632       }
 633
 634       usi.reset(sourceRange);
 635       for (;;) {
 636           if (!usi.next() || usi.isString()) break;
 637           UChar32 c = usi.getCodepoint();
 638
 639           UnicodeString srcStr((UChar32)c);
 640           UnicodeString targ = srcStr;
 641           sourceToTarget->transliterate(targ);
 642           UnicodeString targ2 = srcStr;
 643           sourceToTarget2->transliterate(targ2);
 644           if (targ != targ2) {
 645               logToRulesFails("Source-Target, toRules", srcStr, targ, targ2);
 646           }
 647       }
 648
 649       usi.reset(targetRange);
 650       for (;;) {
 651           if (!usi.next() || usi.isString()) break;
 652           UChar32 c = usi.getCodepoint();
 653
 654           UnicodeString srcStr((UChar32)c);
 655           UnicodeString targ = srcStr;
 656           targetToSource->transliterate(targ);
 657           UnicodeString targ2 = srcStr;
 658           targetToSource2->transliterate(targ2);
 659           if (targ != targ2) {
 660               logToRulesFails("Target-Source, toRules", srcStr, targ, targ2);
 661           }
 662       }
 663     }
 664
 665     parent->logln("Checking that all source characters convert to target - Singles");
 666
 667     UnicodeSet failSourceTarg;
 668     usi.reset(sourceRange);
 669     for (;;) {
 670         if (!usi.next() || usi.isString()) break;
 671         UChar32 c = usi.getCodepoint();
 672
 673         UnicodeString srcStr((UChar32)c);
 674         UnicodeString targ = srcStr;
 675         sourceToTarget->transliterate(targ);
 676         if (toTarget.containsAll(targ) == FALSE
 677             || badCharacters.containsSome(targ) == TRUE) {
 678             UnicodeString targD;
 679             Normalizer::decompose(targ, FALSE, 0, targD, status);
 680             if (U_FAILURE(status)) {
 681                 parent->errln("FAIL: Internal error during decomposition %s\n", u_errorName(status));
 682                 return;
 683             }
 684             if (toTarget.containsAll(targD) == FALSE ||
 685                 badCharacters.containsSome(targD) == TRUE) {
 686                 logWrongScript("Source-Target", srcStr, targ);
 687                 failSourceTarg.add(c);
 688                 continue;
 689             }
 690         }
 691
 692         UnicodeString cs2;
 693         Normalizer::decompose(srcStr, FALSE, 0, cs2, status);
 694         if (U_FAILURE(status)) {
 695             parent->errln("FAIL: Internal error during decomposition %s\n", u_errorName(status));
 696             return;
 697         }
 698         UnicodeString targ2 = cs2;
 699         sourceToTarget->transliterate(targ2);
 700         if (targ != targ2) {
 701             logNotCanonical("Source-Target", srcStr, targ,cs2, targ2);
 702         }
 703     }
 704
 705     parent->logln("Checking that all source characters convert to target - Doubles");
 706
 707     UnicodeSet sourceRangeMinusFailures(sourceRange);
 708     sourceRangeMinusFailures.removeAll(failSourceTarg);
 709
 710     usi.reset(sourceRangeMinusFailures, quickRt, density);
 711     for (;;) {
 712         if (!usi.next() || usi.isString()) break;
 713         UChar32 c = usi.getCodepoint();
 714
 715         usi2.reset(sourceRangeMinusFailures, quickRt, density);
 716         for (;;) {
 717             if (!usi2.next() || usi2.isString()) break;
 718             UChar32 d = usi2.getCodepoint();
 719
 720             UnicodeString srcStr;
 721             srcStr += (UChar32)c;
 722             srcStr += (UChar32)d;
 723             UnicodeString targ = srcStr;
 724             sourceToTarget->transliterate(targ);
 725             if (toTarget.containsAll(targ) == FALSE ||
 726                 badCharacters.containsSome(targ) == TRUE)
 727             {
 728                 UnicodeString targD;
 729                 Normalizer::decompose(targ, FALSE, 0, targD, status);
 730                 if (U_FAILURE(status)) {
 731                     parent->errln("FAIL: Internal error during decomposition %s\n", u_errorName(status));
 732                     return;
 733                 }
 734                 if (toTarget.containsAll(targD) == FALSE ||
 735                     badCharacters.containsSome(targD) == TRUE) {
 736                     logWrongScript("Source-Target", srcStr, targ);
 737                     continue;
 738                 }
 739             }
 740             UnicodeString cs2;
 741             Normalizer::decompose(srcStr, FALSE, 0, cs2, status);
 742             if (U_FAILURE(status)) {
 743                 parent->errln("FAIL: Internal error during decomposition %s\n", u_errorName(status));
 744                 return;
 745             }
 746             UnicodeString targ2 = cs2;
 747             sourceToTarget->transliterate(targ2);
 748             if (targ != targ2) {
 749                 logNotCanonical("Source-Target", srcStr, targ, cs2,targ2);
 750             }
 751         }
 752     }
 753
 754     parent->logln("Checking that target characters convert to source and back - Singles");
 755
 756     UnicodeSet failTargSource;
 757     UnicodeSet failRound;
 758
 759     usi.reset(targetRange);
 760     for (;;) {
 761         if (!usi.next()) break;
 762
 763         if(usi.isString()){
 764             srcStr = usi.getString();
 765         }else{
 766             srcStr = (UnicodeString)usi.getCodepoint();
 767         }
 768
 769         UChar32 c = srcStr.char32At(0);
 770
 771         targ = srcStr;
 772         targetToSource->transliterate(targ);
 773         reverse = targ;
 774         sourceToTarget->transliterate(reverse);
 775
 776         if (toSource.containsAll(targ) == FALSE ||
 777             badCharacters.containsSome(targ) == TRUE) {
 778             UnicodeString targD;
 779             Normalizer::decompose(targ, FALSE, 0, targD, status);
 780             if (U_FAILURE(status)) {
 781                 parent->errln("FAIL: Internal error during decomposition%s\n", u_errorName(status));
 782                 return;
 783             }
 784             if (toSource.containsAll(targD) == FALSE) {
 785                 logWrongScript("Target-Source", srcStr, targ);
 786                 failTargSource.add(c);
 787                 continue;
 788             }
 789             if (badCharacters.containsSome(targD) == TRUE) {
 790                 logWrongScript("Target-Source*", srcStr, targ);
 791                 failTargSource.add(c);
 792                 continue;
 793             }
 794         }
 795         if (isSame(srcStr, reverse) == FALSE &&
 796             roundtripExclusionsSet.contains(c) == FALSE
 797             && roundtripExclusionsSet.contains(srcStr)==FALSE) {
 798             logRoundTripFailure(srcStr,targetToSource->getID(), targ,sourceToTarget->getID(), reverse);
 799             failRound.add(c);
 800             continue;
 801         }
 802
 803         UnicodeString targ2;
 804         Normalizer::decompose(targ, FALSE, 0, targ2, status);
 805         if (U_FAILURE(status)) {
 806             parent->errln("FAIL: Internal error during decomposition%s\n", u_errorName(status));
 807             return;
 808         }
 809         UnicodeString reverse2 = targ2;
 810         sourceToTarget->transliterate(reverse2);
 811         if (reverse != reverse2) {
 812             logNotCanonical("Target-Source", targ, reverse, targ2, reverse2);
 813         }
 814     }
 815
 816     parent->logln("Checking that target characters convert to source and back - Doubles");
 817     int32_t count = 0;
 818
 819     UnicodeSet targetRangeMinusFailures(targetRange);
 820     targetRangeMinusFailures.removeAll(failTargSource);
 821     targetRangeMinusFailures.removeAll(failRound);
 822
 823     usi.reset(targetRangeMinusFailures, quickRt, density);
 824     UnicodeString targ2;
 825     UnicodeString reverse2;
 826     UnicodeString targD;
 827     for (;;) {
 828         if (!usi.next() || usi.isString()) break;
 829         UChar32 c = usi.getCodepoint();
 830         if (++count > pairLimit) {
 831             //throw new TestTruncated("Test truncated at " + pairLimit + " x 64k pairs");
 832             parent->logln("");
 833             parent->logln((UnicodeString)"Test truncated at " + pairLimit + " x 64k pairs");
 834             return;
 835         }
 836
 837         usi2.reset(targetRangeMinusFailures, quickRt, density);
 838         for (;;) {
 839             if (!usi2.next() || usi2.isString())
 840                 break;
 841             UChar32 d = usi2.getCodepoint();
 842             srcStr.truncate(0);  // empty the variable without construction/destruction
 843             srcStr += c;
 844             srcStr += d;
 845
 846             targ = srcStr;
 847             targetToSource->transliterate(targ);
 848             reverse = targ;
 849             sourceToTarget->transliterate(reverse);
 850
 851             if (toSource.containsAll(targ) == FALSE ||
 852                 badCharacters.containsSome(targ) == TRUE)
 853             {
 854                 targD.truncate(0);  // empty the variable without construction/destruction
 855                 Normalizer::decompose(targ, FALSE, 0, targD, status);
 856                 if (U_FAILURE(status)) {
 857                     parent->errln("FAIL: Internal error during decomposition%s\n",
 858                                u_errorName(status));
 859                     return;
 860                 }
 861                 if (toSource.containsAll(targD) == FALSE
 862                     || badCharacters.containsSome(targD) == TRUE)
 863                 {
 864                     logWrongScript("Target-Source", srcStr, targ);
 865                     continue;
 866                 }
 867             }
 868             if (isSame(srcStr, reverse) == FALSE &&
 869                 roundtripExclusionsSet.contains(c) == FALSE&&
 870                 roundtripExclusionsSet.contains(d) == FALSE &&
 871                 roundtripExclusionsSet.contains(srcStr)== FALSE)
 872             {
 873                 logRoundTripFailure(srcStr,targetToSource->getID(), targ, sourceToTarget->getID(),reverse);
 874                 continue;
 875             }
 876
 877             targ2.truncate(0);  // empty the variable without construction/destruction
 878             Normalizer::decompose(targ, FALSE, 0, targ2, status);
 879             if (U_FAILURE(status)) {
 880                 parent->errln("FAIL: Internal error during decomposition%s\n", u_errorName(status));
 881                 return;
 882             }
 883             reverse2 = targ2;
 884             sourceToTarget->transliterate(reverse2);
 885             if (reverse != reverse2) {
 886                 logNotCanonical("Target-Source", targ,reverse, targ2, reverse2);
 887             }
 888         }
 889     }
 890     parent->logln("");
 891 }
 892
 893 void RTTest::logWrongScript(const UnicodeString& label,
 894                             const UnicodeString& from,
 895                             const UnicodeString& to) {
 896     parent->errln((UnicodeString)"FAIL " +
 897                label + ": " +
 898                from + "(" + TestUtility::hex(from) + ") => " +
 899                to + "(" + TestUtility::hex(to) + ")");
 900     ++errorCount;
 901 }
 902
 903 void RTTest::logNotCanonical(const UnicodeString& label,
 904                              const UnicodeString& from,
 905                              const UnicodeString& to,
 906                              const UnicodeString& fromCan,
 907                              const UnicodeString& toCan) {
 908     parent->errln((UnicodeString)"FAIL (can.equiv)" +
 909                label + ": " +
 910                from + "(" + TestUtility::hex(from) + ") => " +
 911                to + "(" + TestUtility::hex(to) + ")" +
 912                fromCan + "(" + TestUtility::hex(fromCan) + ") => " +
 913                toCan + " (" +
 914                TestUtility::hex(toCan) + ")"
 915                );
 916     ++errorCount;
 917 }
 918
 919 void RTTest::logFails(const UnicodeString& label) {
 920     parent->errln((UnicodeString)"<br>FAIL " + label);
 921     ++errorCount;
 922 }
 923
 924 void RTTest::logToRulesFails(const UnicodeString& label,
 925                              const UnicodeString& from,
 926                              const UnicodeString& to,
 927                              const UnicodeString& otherTo)
 928 {
 929     parent->errln((UnicodeString)"FAIL: " +
 930                label + ": " +
 931                from + "(" + TestUtility::hex(from) + ") => " +
 932                to + "(" + TestUtility::hex(to) + ")" +
 933                "!=" +
 934                otherTo + " (" +
 935                TestUtility::hex(otherTo) + ")"
 936                );
 937     ++errorCount;
 938 }
 939
 940
 941 void RTTest::logRoundTripFailure(const UnicodeString& from,
 942                                  const UnicodeString& toID,
 943                                  const UnicodeString& to,
 944                                  const UnicodeString& backID,
 945                                  const UnicodeString& back) {
 946     if (legalSource->is(from) == FALSE) return; // skip illegals
 947
 948     parent->errln((UnicodeString)"FAIL Roundtrip: " +
 949                from + "(" + TestUtility::hex(from) + ") => " +
 950                to + "(" + TestUtility::hex(to) + ")  "+toID+" => " +
 951                back + "(" + TestUtility::hex(back) + ") "+backID+" => ");
 952     ++errorCount;
 953 }
 954
 955 //--------------------------------------------------------------------
 956 // Specific Tests
 957 //--------------------------------------------------------------------
 958
 959     /*
 960     Note: Unicode 3.2 added new Hiragana/Katakana characters:
 961
 962 3095..3096    ; 3.2 #   [2] HIRAGANA LETTER SMALL KA..HIRAGANA LETTER SMALL KE
 963 309F..30A0    ; 3.2 #   [2] HIRAGANA DIGRAPH YORI..KATAKANA-HIRAGANA DOUBLE HYPHEN
 964 30FF          ; 3.2 #       KATAKANA DIGRAPH KOTO
 965 31F0..31FF    ; 3.2 #  [16] KATAKANA LETTER SMALL KU..KATAKANA LETTER SMALL RO
 966
 967     We will not add them to the rules until they are more supported (e.g. in fonts on Windows)
 968     A bug has been filed to remind us to do this: #1979.
 969     */
 970
 971 static const char KATAKANA[] = "[[[:katakana:][\\u30A1-\\u30FA\\u30FC]]-[\\u30FF\\u31F0-\\u31FF]]";
 972 static const char HIRAGANA[] = "[[[:hiragana:][\\u3040-\\u3094]]-[\\u3095-\\u3096\\u309F-\\u30A0]]";
 973 static const char LENGTH[] = "[\\u30FC]";
 974 static const char HALFWIDTH_KATAKANA[] = "[\\uFF65-\\uFF9D]";
 975 static const char KATAKANA_ITERATION[] = "[\\u30FD\\u30FE]";
 976 static const char HIRAGANA_ITERATION[] = "[\\u309D\\u309E]";
 977 static const int32_t TEMP_MAX=256;
 978
 979 void TransliteratorRoundTripTest::TestKana() {
 980     RTTest test("Katakana-Hiragana");
 981     Legal *legal = new Legal();
 982     char temp[TEMP_MAX];
 983     strcpy(temp, "[");
 984     strcat(temp, HALFWIDTH_KATAKANA);
 985     strcat(temp, LENGTH);
 986     strcat(temp, "]");
 987     test.test(KATAKANA, UnicodeString("[") + HIRAGANA + LENGTH + UnicodeString("]"),
 988               temp,
 989               this, quick, legal);
 990     delete legal;
 991 }
 992
 993 void TransliteratorRoundTripTest::TestHiragana() {
 994     RTTest test("Latin-Hiragana");
 995     Legal *legal = new Legal();
 996     test.test(UnicodeString("[a-zA-Z]", ""),
 997               HIRAGANA,
 998               HIRAGANA_ITERATION, this, quick, legal);
 999     delete legal;
1000 }
1001
1002 void TransliteratorRoundTripTest::TestKatakana() {
1003     RTTest test("Latin-Katakana");
1004     Legal *legal = new Legal();
1005     char temp[TEMP_MAX];
1006     strcpy(temp, "[");
1007     strcat(temp, KATAKANA_ITERATION);
1008     strcat(temp, HALFWIDTH_KATAKANA);
1009     strcat(temp, "]");
1010     test.test(UnicodeString("[a-zA-Z]", ""),
1011               KATAKANA,
1012               temp,
1013               this, quick, legal);
1014     delete legal;
1015 }
1016
1017 void TransliteratorRoundTripTest::TestJamo() {
1018     RTTest t("Latin-Jamo");
1019     Legal *legal = new LegalJamo();
1020     t.test(UnicodeString("[a-zA-Z]", ""),
1021            UnicodeString("[\\u1100-\\u1112 \\u1161-\\u1175 \\u11A8-\\u11C2]",
1022                          ""),
1023            NULL, this, quick, legal);
1024     delete legal;
1025 }
1026
1027 void TransliteratorRoundTripTest::TestHangul() {
1028     RTTest t("Latin-Hangul");
1029     Legal *legal = new Legal();
1030     if (quick) t.setPairLimit(1000);
1031     t.test(UnicodeString("[a-zA-Z]", ""),
1032            UnicodeString("[\\uAC00-\\uD7A4]", ""),
1033            NULL, this, quick, legal, 1);
1034     delete legal;
1035 }
1036
1037
1038 #define ASSERT_SUCCESS(status) {if (U_FAILURE(status)) { \
1039      errln("error at file %s, line %d, status = %s", __FILE__, __LINE__, \
1040          u_errorName(status)); \
1041          return;}}
1042
1043
1044 static void writeStringInU8(FILE *out, const UnicodeString &s) {
1045     int i;
1046     for (i=0; i<s.length(); i=s.moveIndex32(i, 1)) {
1047         UChar32  c = s.char32At(i);
1048         uint8_t  bufForOneChar[10];
1049         UBool    isError = FALSE;
1050         int32_t  destIdx = 0;
1051         U8_APPEND(bufForOneChar, destIdx, sizeof(bufForOneChar), c, isError);
1052         fwrite(bufForOneChar, 1, destIdx, out);
1053     }
1054 }
1055
1056
1057
1058
1059 void TransliteratorRoundTripTest::TestHan() {
1060     UErrorCode  status = U_ZERO_ERROR;
1061
1062     // TODO:  getExemplars() exists only as a C API, taking a USet.
1063     //        The set API we need to use exists only on UnicodeSet, not on USet.
1064     //        Do a hacky cast, knowing that a USet is really a UnicodeSet in
1065     //        the implementation.  Once USet gets the missing API, switch back
1066     //        to using that.
1067     USet       *USetExemplars = NULL;
1068     USetExemplars = uset_open(0, 0);
1069     USetExemplars = ulocdata_getExemplarSet(USetExemplars, "zh", 0, &status);
1070     ASSERT_SUCCESS(status);
1071     UnicodeSet *exemplars = (UnicodeSet *)USetExemplars;
1072
1073     UnicodeString source;
1074     UChar32       c;
1075     int           i;
1076     for (i=0; ;i++) {
1077         // Add all of the Chinese exemplar chars to the string "source".
1078         c = exemplars->charAt(i);
1079         if (c == (UChar32)-1) {
1080             break;
1081         }
1082         source.append(c);
1083     }
1084
1085     // transform with Han translit
1086     Transliterator *hanTL = Transliterator::createInstance("Han-Latin", UTRANS_FORWARD, status);
1087     ASSERT_SUCCESS(status);
1088     UnicodeString target=source;
1089     hanTL->transliterate(target);
1090     // now verify that there are no Han characters left
1091     UnicodeSet allHan("[:han:]", status);
1092     ASSERT_SUCCESS(status);
1093     if (allHan.containsSome(target)) {
1094         errln("file %s, line %d, No Han must be left after Han-Latin transliteration",
1095             __FILE__, __LINE__);
1096     }
1097
1098     // check the pinyin translit
1099     Transliterator *pn = Transliterator::createInstance("Latin-NumericPinyin", UTRANS_FORWARD, status);
1100     ASSERT_SUCCESS(status);
1101     UnicodeString target2 = target;
1102     pn->transliterate(target2);
1103
1104     // verify that there are no marks
1105     Transliterator *nfc = Transliterator::createInstance("nfc", UTRANS_FORWARD, status);
1106     ASSERT_SUCCESS(status);
1107
1108     UnicodeString nfced = target2;
1109     nfc->transliterate(nfced);
1110     UnicodeSet allMarks("[:mark:]", status);
1111     ASSERT_SUCCESS(status);
1112     assertFalse("NumericPinyin must contain no marks", allMarks.containsSome(nfced));
1113
1114     // verify roundtrip
1115     Transliterator *np = pn->createInverse(status);
1116     ASSERT_SUCCESS(status);
1117     UnicodeString target3 = target;
1118     np->transliterate(target3);
1119     UBool roundtripOK = (target3.compare(target) == 0);
1120     assertTrue("NumericPinyin must roundtrip", roundtripOK);
1121     if (!roundtripOK) {
1122         const char *filename = "numeric-pinyin.log.txt";
1123         FILE *out = fopen(filename, "w");
1124         errln("Creating log file %s\n", filename);
1125         fprintf(out, "Pinyin:                ");
1126         writeStringInU8(out, target);
1127         fprintf(out, "\nPinyin-Numeric-Pinyin: ");
1128         writeStringInU8(out, target2);
1129         fprintf(out, "\n");
1130         fclose(out);
1131     }
1132
1133     delete hanTL;
1134     delete pn;
1135     delete nfc;
1136     delete np;
1137     uset_close(USetExemplars);
1138 }
1139
1140
1141 void TransliteratorRoundTripTest::TestGreek() {
1142     // weiv removed the test and the fiter
1143     /*
1144     if (isICUVersionAtLeast(ICU_30)) {
1145         // We temporarily filter against Unicode 3.2, but we only do this
1146         // before version 3.0.
1147         errln("FAIL: TestGreek needs to be updated to remove Unicode 3.2 filter");
1148         return;
1149     } else {
1150         logln("Warning: TestGreek needs to be updated to remove Unicode 3.2 filter");
1151     }
1152     */
1153     RTTest test("Latin-Greek");
1154     LegalGreek *legal = new LegalGreek(TRUE);
1155
1156     test.test(UnicodeString("[a-zA-Z]", ""),
1157         UnicodeString("[\\u003B\\u00B7[[:Greek:]&[:Letter:]]-["
1158             "\\u1D26-\\u1D2A" // L&   [5] GREEK LETTER SMALL CAPITAL GAMMA..GREEK LETTER SMALL CAPITAL PSI
1159             "\\u1D5D-\\u1D61" // Lm   [5] MODIFIER LETTER SMALL BETA..MODIFIER LETTER SMALL CHI
1160             "\\u1D66-\\u1D6A" // L&   [5] GREEK SUBSCRIPT SMALL LETTER BETA..GREEK SUBSCRIPT SMALL LETTER CHI
1161             "\\u03D7-\\u03EF" // \N{GREEK KAI SYMBOL}..\N{COPTIC SMALL LETTER DEI}
1162             "]]",
1163
1164               //UnicodeString("[[\\u003B\\u00B7[:Greek:]-[\\u0374\\u0385\\u1fcd\\u1fce\\u1fdd\\u1fde\\u1fed-\\u1fef\\u1ffd\\u03D7-\\u03EF]]&[:Age=3.2:]]",
1165                             ""),
1166               "[\\u00B5\\u037A\\u03D0-\\u03F5\\u03f9]", /* exclusions */
1167               this, quick, legal, 50);
1168
1169
1170     delete legal;
1171 }
1172
1173
1174 void TransliteratorRoundTripTest::TestGreekUNGEGN() {
1175     // weiv removed the test and the fiter
1176     /*
1177     if (isICUVersionAtLeast(ICU_30)) {
1178         // We temporarily filter against Unicode 3.2, but we only do this
1179         // before version 3.0.
1180         errln("FAIL: TestGreekUNGEGN needs to be updated to remove Unicode 3.2 filter");
1181         return;
1182     } else {
1183         logln("Warning: TestGreekUNGEGN needs to be updated to remove Unicode 3.2 filter");
1184     }
1185     */
1186     RTTest test("Latin-Greek/UNGEGN");
1187     LegalGreek *legal = new LegalGreek(FALSE);
1188
1189     test.test(UnicodeString("[a-zA-Z]", ""),
1190         UnicodeString("[\\u003B\\u00B7[[:Greek:]&[:Letter:]]-["
1191             "\\u1D26-\\u1D2A" // L&   [5] GREEK LETTER SMALL CAPITAL GAMMA..GREEK LETTER SMALL CAPITAL PSI
1192             "\\u1D5D-\\u1D61" // Lm   [5] MODIFIER LETTER SMALL BETA..MODIFIER LETTER SMALL CHI
1193             "\\u1D66-\\u1D6A" // L&   [5] GREEK SUBSCRIPT SMALL LETTER BETA..GREEK SUBSCRIPT SMALL LETTER CHI
1194             "\\u03D7-\\u03EF" // \N{GREEK KAI SYMBOL}..\N{COPTIC SMALL LETTER DEI}
1195             "]]",
1196               //UnicodeString("[[\\u003B\\u00B7[:Greek:]-[\\u0374\\u0385\\u1fce\\u1fde\\u03D7-\\u03EF]]&[:Age=3.2:]]",
1197                             ""),
1198               "[\\u0385\\u00B5\\u037A\\u03D0-\\uFFFF {\\u039C\\u03C0}]", /* roundtrip exclusions */
1199               this, quick, legal);
1200
1201     delete legal;
1202 }
1203
1204 void TransliteratorRoundTripTest::Testel() {
1205     // weiv removed the test and the fiter
1206     /*
1207     if (isICUVersionAtLeast(ICU_30)) {
1208         // We temporarily filter against Unicode 3.2, but we only do this
1209         // before version 3.0.
1210         errln("FAIL: Testel needs to be updated to remove Unicode 3.2 filter");
1211         return;
1212     } else {
1213         logln("Warning: Testel needs to be updated to remove Unicode 3.2 filter");
1214     }
1215     */
1216     RTTest test("Latin-el");
1217     LegalGreek *legal = new LegalGreek(FALSE);
1218
1219     test.test(UnicodeString("[a-zA-Z]", ""),
1220         UnicodeString("[\\u003B\\u00B7[[:Greek:]&[:Letter:]]-["
1221             "\\u1D26-\\u1D2A" // L&   [5] GREEK LETTER SMALL CAPITAL GAMMA..GREEK LETTER SMALL CAPITAL PSI
1222             "\\u1D5D-\\u1D61" // Lm   [5] MODIFIER LETTER SMALL BETA..MODIFIER LETTER SMALL CHI
1223             "\\u1D66-\\u1D6A" // L&   [5] GREEK SUBSCRIPT SMALL LETTER BETA..GREEK SUBSCRIPT SMALL LETTER CHI
1224             "\\u03D7-\\u03EF" // \N{GREEK KAI SYMBOL}..\N{COPTIC SMALL LETTER DEI}
1225             "]]",
1226               //UnicodeString("[[\\u003B\\u00B7[:Greek:]-[\\u0374\\u0385\\u1fce\\u1fde\\u03D7-\\u03EF]]&[:Age=3.2:]]",
1227                             ""),
1228               "[\\u00B5\\u037A\\u03D0-\\uFFFF {\\u039C\\u03C0}]", /* exclusions */
1229               this, quick, legal);
1230
1231
1232     delete legal;
1233 }
1234
1235
1236 void TransliteratorRoundTripTest::TestArabic() {
1237     UnicodeString ARABIC("[\\u060C\\u061B\\u061F\\u0621\\u0627-\\u063A\\u0641-\\u0655\\u0660-\\u066C\\u067E\\u0686\\u0698\\u06A4\\u06AD\\u06AF\\u06CB-\\u06CC\\u06F0-\\u06F9]");
1238     Legal *legal = new Legal();
1239     RTTest test("Latin-Arabic");
1240         test.test("[a-zA-Z\\u02BE\\u02BF\\u207F]", ARABIC, "[a-zA-Z\\u02BE\\u02BF\\u207F]",this, quick, legal); //
1241    delete legal;
1242 }
1243 class LegalHebrew : public Legal {
1244 private:
1245     UnicodeSet FINAL;
1246     UnicodeSet NON_FINAL;
1247     UnicodeSet LETTER;
1248 public:
1249     LegalHebrew(UErrorCode& error);
1250     virtual ~LegalHebrew() {}
1251     virtual UBool is(const UnicodeString& sourceString) const;
1252 };
1253
1254 LegalHebrew::LegalHebrew(UErrorCode& error){
1255     FINAL.applyPattern("[\\u05DA\\u05DD\\u05DF\\u05E3\\u05E5]", error);
1256     NON_FINAL.applyPattern("[\\u05DB\\u05DE\\u05E0\\u05E4\\u05E6]", error);
1257     LETTER.applyPattern("[:letter:]", error);
1258 }
1259 UBool LegalHebrew::is(const UnicodeString& sourceString)const{
1260
1261     if (sourceString.length() == 0) return TRUE;
1262     // don't worry about surrogates.
1263     for (int i = 0; i < sourceString.length(); ++i) {
1264         UChar ch = sourceString.charAt(i);
1265         UChar next = i+1 == sourceString.length() ? 0x0000 : sourceString.charAt(i);
1266         if (FINAL.contains(ch)) {
1267             if (LETTER.contains(next)) return FALSE;
1268         } else if (NON_FINAL.contains(ch)) {
1269             if (!LETTER.contains(next)) return FALSE;
1270         }
1271     }
1272     return TRUE;
1273 }
1274 void TransliteratorRoundTripTest::TestHebrew() {
1275     //long start = System.currentTimeMillis();
1276     UErrorCode error = U_ZERO_ERROR;
1277     LegalHebrew* legal = new LegalHebrew(error);
1278     if(U_FAILURE(error)){
1279         errln("Could not construct LegalHebrew object. Error: %s", u_errorName(error));
1280         return;
1281     }
1282     RTTest test("Latin-Hebrew");
1283         test.test("[a-zA-Z\\u02BC\\u02BB]", "[[:hebrew:]-[\\u05BD\\uFB00-\\uFBFF]]", "[\\u05F0\\u05F1\\u05F2]", this, quick, legal);
1284     //showElapsed(start, "TestHebrew");
1285     delete legal;
1286 }
1287 void TransliteratorRoundTripTest::TestCyrillic() {
1288     RTTest test("Latin-Cyrillic");
1289     Legal *legal = new Legal();
1290
1291     test.test(UnicodeString("[a-zA-Z\\u0110\\u0111\\u02BA\\u02B9]", ""),
1292               UnicodeString("[[\\u0400-\\u045F] & [:Age=3.2:]]", ""), NULL, this, quick,
1293               legal);
1294
1295     delete legal;
1296 }
1297
1298
1299 // Inter-Indic Tests ----------------------------------
1300 class LegalIndic :public Legal{
1301     UnicodeSet vowelSignSet;
1302     UnicodeSet avagraha;
1303     UnicodeSet nukta;
1304     UnicodeSet virama;
1305     UnicodeSet sanskritStressSigns;
1306     UnicodeSet chandrabindu;
1307
1308 public:
1309     LegalIndic(){
1310         UErrorCode status = U_ZERO_ERROR;
1311         vowelSignSet.addAll( UnicodeSet("[\\u0902\\u0903\\u0904\\u093e-\\u094c\\u0962\\u0963]",status));/* Devanagari */
1312         vowelSignSet.addAll( UnicodeSet("[\\u0982\\u0983\\u09be-\\u09cc\\u09e2\\u09e3\\u09D7]",status));/* Bengali */
1313         vowelSignSet.addAll( UnicodeSet("[\\u0a02\\u0a03\\u0a3e-\\u0a4c\\u0a62\\u0a63\\u0a70\\u0a71]",status));/* Gurmukhi */
1314         vowelSignSet.addAll( UnicodeSet("[\\u0a82\\u0a83\\u0abe-\\u0acc\\u0ae2\\u0ae3]",status));/* Gujarati */
1315         vowelSignSet.addAll( UnicodeSet("[\\u0b02\\u0b03\\u0b3e-\\u0b4c\\u0b62\\u0b63\\u0b56\\u0b57]",status));/* Oriya */
1316         vowelSignSet.addAll( UnicodeSet("[\\u0b82\\u0b83\\u0bbe-\\u0bcc\\u0be2\\u0be3\\u0bd7]",status));/* Tamil */
1317         vowelSignSet.addAll( UnicodeSet("[\\u0c02\\u0c03\\u0c3e-\\u0c4c\\u0c62\\u0c63\\u0c55\\u0c56]",status));/* Telugu */
1318         vowelSignSet.addAll( UnicodeSet("[\\u0c82\\u0c83\\u0cbe-\\u0ccc\\u0ce2\\u0ce3\\u0cd5\\u0cd6]",status));/* Kannada */
1319         vowelSignSet.addAll( UnicodeSet("[\\u0d02\\u0d03\\u0d3e-\\u0d4c\\u0d62\\u0d63\\u0d57]",status));/* Malayalam */
1320
1321         avagraha.addAll(UnicodeSet("[\\u093d\\u09bd\\u0abd\\u0b3d\\u0cbd]",status));
1322         nukta.addAll(UnicodeSet("[\\u093c\\u09bc\\u0a3c\\u0abc\\u0b3c\\u0cbc]",status));
1323         virama.addAll(UnicodeSet("[\\u094d\\u09cd\\u0a4d\\u0acd\\u0b4d\\u0bcd\\u0c4d\\u0ccd\\u0d4d]",status));
1324         sanskritStressSigns.addAll(UnicodeSet("[\\u0951\\u0952\\u0953\\u0954]",status));
1325         chandrabindu.addAll(UnicodeSet("[\\u0901\\u0981\\u0A81\\u0b01\\u0c01]",status));
1326
1327     }
1328     virtual UBool is(const UnicodeString& sourceString) const;
1329     virtual ~LegalIndic() {};
1330 };
1331 UBool LegalIndic::is(const UnicodeString& sourceString) const{
1332     int cp=sourceString.charAt(0);
1333
1334     // A vowel sign cannot be the first char
1335     if(vowelSignSet.contains(cp)){
1336         return FALSE;
1337     }else if(avagraha.contains(cp)){
1338         return FALSE;
1339     }else if(virama.contains(cp)){
1340         return FALSE;
1341     }else if(nukta.contains(cp)){
1342         return FALSE;
1343     }else if(sanskritStressSigns.contains(cp)){
1344         return FALSE;
1345     }else if(chandrabindu.contains(cp) &&
1346                 ((sourceString.length()>1) &&
1347                     vowelSignSet.contains(sourceString.charAt(1)))){
1348         return FALSE;
1349     }
1350     return TRUE;
1351 }
1352
1353 static const char latinForIndic[] = "[['.0-9A-Za-z~\\u00C0-\\u00C5\\u00C7-\\u00CF\\u00D1-\\u00D6\\u00D9-\\u00DD"
1354                                    "\\u00E0-\\u00E5\\u00E7-\\u00EF\\u00F1-\\u00F6\\u00F9-\\u00FD\\u00FF-\\u010F"
1355                                    "\\u0112-\\u0125\\u0128-\\u0130\\u0134-\\u0137\\u0139-\\u013E\\u0143-\\u0148"
1356                                    "\\u014C-\\u0151\\u0154-\\u0165\\u0168-\\u017E\\u01A0-\\u01A1\\u01AF-\\u01B0"
1357                                    "\\u01CD-\\u01DC\\u01DE-\\u01E3\\u01E6-\\u01ED\\u01F0\\u01F4-\\u01F5\\u01F8-\\u01FB"
1358                                    "\\u0200-\\u021B\\u021E-\\u021F\\u0226-\\u0233\\u0303-\\u0304\\u0306\\u0314-\\u0315"
1359                                    "\\u0325\\u040E\\u0419\\u0439\\u045E\\u04C1-\\u04C2\\u04D0-\\u04D1\\u04D6-\\u04D7"
1360                                    "\\u04E2-\\u04E3\\u04EE-\\u04EF\\u1E00-\\u1E99\\u1EA0-\\u1EF9\\u1F01\\u1F03\\u1F05"
1361                                    "\\u1F07\\u1F09\\u1F0B\\u1F0D\\u1F0F\\u1F11\\u1F13\\u1F15\\u1F19\\u1F1B\\u1F1D\\u1F21"
1362                                    "\\u1F23\\u1F25\\u1F27\\u1F29\\u1F2B\\u1F2D\\u1F2F\\u1F31\\u1F33\\u1F35\\u1F37\\u1F39"
1363                                    "\\u1F3B\\u1F3D\\u1F3F\\u1F41\\u1F43\\u1F45\\u1F49\\u1F4B\\u1F4D\\u1F51\\u1F53\\u1F55"
1364                                    "\\u1F57\\u1F59\\u1F5B\\u1F5D\\u1F5F\\u1F61\\u1F63\\u1F65\\u1F67\\u1F69\\u1F6B\\u1F6D"
1365                                    "\\u1F6F\\u1F81\\u1F83\\u1F85\\u1F87\\u1F89\\u1F8B\\u1F8D\\u1F8F\\u1F91\\u1F93\\u1F95"
1366                                    "\\u1F97\\u1F99\\u1F9B\\u1F9D\\u1F9F\\u1FA1\\u1FA3\\u1FA5\\u1FA7\\u1FA9\\u1FAB\\u1FAD"
1367                                    "\\u1FAF-\\u1FB1\\u1FB8-\\u1FB9\\u1FD0-\\u1FD1\\u1FD8-\\u1FD9\\u1FE0-\\u1FE1\\u1FE5"
1368                                    "\\u1FE8-\\u1FE9\\u1FEC\\u212A-\\u212B\\uE04D\\uE064]"
1369                                    "-[\\uE000-\\uE080 \\u01E2\\u01E3]& [[:latin:][:mark:]]]";
1370
1371 void TransliteratorRoundTripTest::TestDevanagariLatin() {
1372     {
1373         UErrorCode status = U_ZERO_ERROR;
1374         UParseError parseError;
1375         TransliteratorPointer t1(Transliterator::createInstance("[\\u0964-\\u0965\\u0981-\\u0983\\u0985-\\u098C\\u098F-\\u0990\\u0993-\\u09A8\\u09AA-\\u09B0\\u09B2\\u09B6-\\u09B9\\u09BC\\u09BE-\\u09C4\\u09C7-\\u09C8\\u09CB-\\u09CD\\u09D7\\u09DC-\\u09DD\\u09DF-\\u09E3\\u09E6-\\u09FA];NFD;Bengali-InterIndic;InterIndic-Gujarati;NFC;",UTRANS_FORWARD, parseError, status));
1376         if((Transliterator *)t1 != NULL){
1377             TransliteratorPointer t2(t1->createInverse(status));
1378             if(U_FAILURE(status)){
1379                 errln("FAIL: could not create the Inverse:-( \n");
1380             }
1381         }else {
1382             errln("FAIL: could not create the transliterator. Error: %s\n", u_errorName(status));
1383         }
1384
1385     }
1386     RTTest test("Latin-Devanagari");
1387     Legal *legal = new LegalIndic();
1388
1389     test.test(UnicodeString(latinForIndic, ""),
1390             UnicodeString("[[:Devanagari:][\\u094d][\\u0964\\u0965]]", ""), "[\\u0965\\u0904]", this, quick,
1391             legal, 50);
1392
1393     delete legal;
1394 }
1395
1396 /* Defined this way for HP/UX11CC :-( */
1397 static const int32_t INTER_INDIC_ARRAY_WIDTH = 4;
1398 static const char * const interIndicArray[] = {
1399     "BENGALI-DEVANAGARI", "[:BENGALI:]", "[:Devanagari:]",
1400     "[\\u0904\\u0951-\\u0954\\u0943-\\u0949\\u094a\\u0962\\u0963\\u090D\\u090e\\u0911\\u0912\\u0929\\u0933\\u0934\\u0935\\u093d\\u0950\\u0958\\u0959\\u095a\\u095b\\u095e\\u09f0\\u09f1\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1401
1402     "DEVANAGARI-BENGALI", "[:Devanagari:]", "[:BENGALI:]",
1403     "[\\u0951-\\u0954\\u0951-\\u0954\\u09D7\\u090D\\u090e\\u0911\\u0912\\u0929\\u0933\\u0934\\u0935\\u093d\\u0950\\u0958\\u0959\\u095a\\u095b\\u095e\\u09f0\\u09f1\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1404
1405     "GURMUKHI-DEVANAGARI", "[:GURMUKHI:]", "[:Devanagari:]",
1406     "[\\u0904\\u0901\\u0902\\u0936\\u0933\\u0951-\\u0954\\u0902\\u0903\\u0943-\\u0949\\u094a\\u0962\\u0963\\u090B\\u090C\\u090D\\u090e\\u0911\\u0912\\u0934\\u0937\\u093D\\u0950\\u0960\\u0961\\u0a72\\u0a73\\u0a74]", /*roundtrip exclusions*/
1407
1408     "DEVANAGARI-GURMUKHI", "[:Devanagari:]", "[:GURMUKHI:]",
1409     "[\\u0904\\u0A02\\u0946\\u0A5C\\u0951-\\u0954\\u0A70\\u0A71\\u090B\\u090C\\u090D\\u090e\\u0911\\u0912\\u0934\\u0937\\u093D\\u0950\\u0960\\u0961\\u0a72\\u0a73\\u0a74]", /*roundtrip exclusions*/
1410
1411     "GUJARATI-DEVANAGARI", "[:GUJARATI:]", "[:Devanagari:]",
1412     "[\\u0946\\u094A\\u0962\\u0963\\u0951-\\u0954\\u0961\\u090c\\u090e\\u0912]", /*roundtrip exclusions*/
1413
1414     "DEVANAGARI-GUJARATI", "[:Devanagari:]", "[:GUJARATI:]",
1415     "[\\u0951-\\u0954\\u0961\\u090c\\u090e\\u0912]", /*roundtrip exclusions*/
1416
1417     "ORIYA-DEVANAGARI", "[:ORIYA:]", "[:Devanagari:]",
1418     "[\\u0904\\u0943-\\u094a\\u0962\\u0963\\u0951-\\u0954\\u0950\\u090D\\u090e\\u0912\\u0911\\u0931\\u0935]", /*roundtrip exclusions*/
1419
1420     "DEVANAGARI-ORIYA", "[:Devanagari:]", "[:ORIYA:]",
1421     "[\\u0b5f\\u0b56\\u0b57\\u0b70\\u0b71\\u0950\\u090D\\u090e\\u0912\\u0911\\u0931]", /*roundtrip exclusions*/
1422
1423     "Tamil-DEVANAGARI", "[:tamil:]", "[:Devanagari:]",
1424     "[\\u0901\\u0904\\u093c\\u0943-\\u094a\\u0951-\\u0954\\u0962\\u0963\\u090B\\u090C\\u090D\\u0911\\u0916\\u0917\\u0918\\u091B\\u091D\\u0920\\u0921\\u0922\\u0925\\u0926\\u0927\\u092B\\u092C\\u092D\\u0936\\u093d\\u0950[\\u0958-\\u0961]]", /*roundtrip exclusions*/
1425
1426     "DEVANAGARI-Tamil", "[:Devanagari:]", "[:tamil:]",
1427     "[\\u0bd7]", /*roundtrip exclusions*/
1428
1429     "Telugu-DEVANAGARI", "[:telugu:]", "[:Devanagari:]",
1430     "[\\u0904\\u093c\\u0950\\u0945\\u0949\\u0951-\\u0954\\u0962\\u0963\\u090D\\u0911\\u093d\\u0929\\u0934[\\u0958-\\u095f]]", /*roundtrip exclusions*/
1431
1432     "DEVANAGARI-TELUGU", "[:Devanagari:]", "[:TELUGU:]",
1433     "[\\u0c55\\u0c56\\u0950\\u090D\\u0911\\u093d\\u0929\\u0934[\\u0958-\\u095f]]", /*roundtrip exclusions*/
1434
1435     "KANNADA-DEVANAGARI", "[:KANNADA:]", "[:Devanagari:]",
1436     "[\\u0901\\u0904\\u0946\\u093c\\u0950\\u0945\\u0949\\u0951-\\u0954\\u0962\\u0963\\u0950\\u090D\\u0911\\u093d\\u0929\\u0934[\\u0958-\\u095f]]", /*roundtrip exclusions*/
1437
1438     "DEVANAGARI-KANNADA", "[:Devanagari:]", "[:KANNADA:]",
1439     "[{\\u0cb0\\u0cbc}{\\u0cb3\\u0cbc}\\u0cde\\u0cd5\\u0cd6\\u0950\\u090D\\u0911\\u093d\\u0929\\u0934[\\u0958-\\u095f]]", /*roundtrip exclusions*/
1440
1441     "MALAYALAM-DEVANAGARI", "[:MALAYALAM:]", "[:Devanagari:]",
1442     "[\\u0901\\u0904\\u094a\\u094b\\u094c\\u093c\\u0950\\u0944\\u0945\\u0949\\u0951-\\u0954\\u0962\\u0963\\u090D\\u0911\\u093d\\u0929\\u0934[\\u0958-\\u095f]]", /*roundtrip exclusions*/
1443
1444     "DEVANAGARI-MALAYALAM", "[:Devanagari:]", "[:MALAYALAM:]",
1445     "[\\u0d4c\\u0d57\\u0950\\u090D\\u0911\\u093d\\u0929\\u0934[\\u0958-\\u095f]]", /*roundtrip exclusions*/
1446
1447     "GURMUKHI-BENGALI", "[:GURMUKHI:]", "[:BENGALI:]",
1448     "[\\u0981\\u0982\\u09b6\\u09e2\\u09e3\\u09c3\\u09c4\\u09d7\\u098B\\u098C\\u09B7\\u09E0\\u09E1\\u09F0\\u09F1\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1449
1450     "BENGALI-GURMUKHI", "[:BENGALI:]", "[:GURMUKHI:]",
1451     "[\\u0A02\\u0a5c\\u0a47\\u0a70\\u0a71\\u0A33\\u0A35\\u0A59\\u0A5A\\u0A5B\\u0A5E\\u0A72\\u0A73\\u0A74]", /*roundtrip exclusions*/
1452
1453     "GUJARATI-BENGALI", "[:GUJARATI:]", "[:BENGALI:]",
1454     "[\\u09d7\\u09e2\\u09e3\\u098c\\u09e1\\u09f0\\u09f1\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1455
1456     "BENGALI-GUJARATI", "[:BENGALI:]", "[:GUJARATI:]",
1457     "[\\u0A82\\u0a83\\u0Ac9\\u0Ac5\\u0ac7\\u0A8D\\u0A91\\u0AB3\\u0AB5\\u0ABD\\u0AD0]", /*roundtrip exclusions*/
1458
1459     "ORIYA-BENGALI", "[:ORIYA:]", "[:BENGALI:]",
1460     "[\\u09c4\\u09e2\\u09e3\\u09f0\\u09f1\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1461
1462     "BENGALI-ORIYA", "[:BENGALI:]", "[:ORIYA:]",
1463     "[\\u0b35\\u0b71\\u0b5f\\u0b56\\u0b33\\u0b3d]", /*roundtrip exclusions*/
1464
1465     "Tamil-BENGALI", "[:tamil:]", "[:BENGALI:]",
1466     "[\\u0981\\u09bc\\u09c3\\u09c4\\u09e2\\u09e3\\u09f0\\u09f1\\u098B\\u098C\\u0996\\u0997\\u0998\\u099B\\u099D\\u09A0\\u09A1\\u09A2\\u09A5\\u09A6\\u09A7\\u09AB\\u09AC\\u09AD\\u09B6\\u09DC\\u09DD\\u09DF\\u09E0\\u09E1\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1467
1468     "BENGALI-Tamil", "[:BENGALI:]", "[:tamil:]",
1469     "[\\u0bc6\\u0bc7\\u0bca\\u0B8E\\u0B92\\u0BA9\\u0BB1\\u0BB3\\u0BB4\\u0BB5]", /*roundtrip exclusions*/
1470
1471     "Telugu-BENGALI", "[:telugu:]", "[:BENGALI:]",
1472     "[\\u09e2\\u09e3\\u09bc\\u09d7\\u09f0\\u09f1\\u09dc\\u09dd\\u09df\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1473
1474     "BENGALI-TELUGU", "[:BENGALI:]", "[:TELUGU:]",
1475     "[\\u0c55\\u0c56\\u0c47\\u0c46\\u0c4a\\u0C0E\\u0C12\\u0C31\\u0C33\\u0C35]", /*roundtrip exclusions*/
1476
1477     "KANNADA-BENGALI", "[:KANNADA:]", "[:BENGALI:]",
1478     "[\\u0981\\u09e2\\u09e3\\u09bc\\u09d7\\u09dc\\u09dd\\u09df\\u09f0\\u09f1\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1479
1480     "BENGALI-KANNADA", "[:BENGALI:]", "[:KANNADA:]",
1481     "[{\\u0cb0\\u0cbc}{\\u0cb3\\u0cbc}\\u0cc6\\u0cca\\u0cd5\\u0cd6\\u0cc7\\u0C8E\\u0C92\\u0CB1\\u0cb3\\u0cb5\\u0cde]", /*roundtrip exclusions*/
1482
1483     "MALAYALAM-BENGALI", "[:MALAYALAM:]", "[:BENGALI:]",
1484     "[\\u0981\\u09e2\\u09e3\\u09bc\\u09c4\\u09f0\\u09f1\\u09dc\\u09dd\\u09df\\u09dc\\u09dd\\u09df\\u09f2-\\u09fa]", /*roundtrip exclusions*/
1485
1486     "BENGALI-MALAYALAM", "[:BENGALI:]", "[:MALAYALAM:]",
1487     "[\\u0d46\\u0d4a\\u0d47\\u0d31-\\u0d35\\u0d0e\\u0d12]", /*roundtrip exclusions*/
1488
1489     "GUJARATI-GURMUKHI", "[:GUJARATI:]", "[:GURMUKHI:]",
1490     "[\\u0A02\\u0ab3\\u0ab6\\u0A70\\u0a71\\u0a82\\u0a83\\u0ac3\\u0ac4\\u0ac5\\u0ac9\\u0a5c\\u0a72\\u0a73\\u0a74\\u0a8b\\u0a8d\\u0a91\\u0abd]", /*roundtrip exclusions*/
1491
1492     "GURMUKHI-GUJARATI", "[:GURMUKHI:]", "[:GUJARATI:]",
1493     "[\\u0a5c\\u0A70\\u0a71\\u0a72\\u0a73\\u0a74\\u0a82\\u0a83\\u0a8b\\u0a8c\\u0a8d\\u0a91\\u0ab3\\u0ab6\\u0ab7\\u0abd\\u0ac3\\u0ac4\\u0ac5\\u0ac9\\u0ad0\\u0ae0\\u0ae1]", /*roundtrip exclusions*/
1494
1495     "ORIYA-GURMUKHI", "[:ORIYA:]", "[:GURMUKHI:]",
1496     "[\\u0A01\\u0A02\\u0a5c\\u0a21\\u0a47\\u0a71\\u0b02\\u0b03\\u0b33\\u0b36\\u0b43\\u0b56\\u0b57\\u0B0B\\u0B0C\\u0B37\\u0B3D\\u0B5F\\u0B60\\u0B61\\u0a35\\u0a72\\u0a73\\u0a74]", /*roundtrip exclusions*/
1497
1498     "GURMUKHI-ORIYA", "[:GURMUKHI:]", "[:ORIYA:]",
1499     "[\\u0b01\\u0b02\\u0b03\\u0b33\\u0b36\\u0b43\\u0b56\\u0b57\\u0B0B\\u0B0C\\u0B37\\u0B3D\\u0B5F\\u0B60\\u0B61\\u0b70\\u0b71]", /*roundtrip exclusions*/
1500
1501     "TAMIL-GURMUKHI", "[:TAMIL:]", "[:GURMUKHI:]",
1502     "[\\u0A01\\u0A02\\u0a33\\u0a36\\u0a3c\\u0a70\\u0a71\\u0a47\\u0A16\\u0A17\\u0A18\\u0A1B\\u0A1D\\u0A20\\u0A21\\u0A22\\u0A25\\u0A26\\u0A27\\u0A2B\\u0A2C\\u0A2D\\u0A59\\u0A5A\\u0A5B\\u0A5C\\u0A5E\\u0A72\\u0A73\\u0A74]", /*roundtrip exclusions*/
1503
1504     "GURMUKHI-TAMIL", "[:GURMUKHI:]", "[:TAMIL:]",
1505     "[\\u0b82\\u0bc6\\u0bca\\u0bd7\\u0bb7\\u0bb3\\u0b83\\u0B8E\\u0B92\\u0BA9\\u0BB1\\u0BB4]", /*roundtrip exclusions*/
1506
1507     "TELUGU-GURMUKHI", "[:TELUGU:]", "[:GURMUKHI:]",
1508     "[\\u0A02\\u0a33\\u0a36\\u0a3c\\u0a70\\u0a71\\u0A59\\u0A5A\\u0A5B\\u0A5C\\u0A5E\\u0A72\\u0A73\\u0A74]", /*roundtrip exclusions*/
1509
1510     "GURMUKHI-TELUGU", "[:GURMUKHI:]", "[:TELUGU:]",
1511     "[\\u0c01\\u0c02\\u0c03\\u0c33\\u0c36\\u0c44\\u0c43\\u0c46\\u0c4a\\u0c56\\u0c55\\u0C0B\\u0C0C\\u0C0E\\u0C12\\u0C31\\u0C37\\u0C60\\u0C61]", /*roundtrip exclusions*/
1512
1513     "KANNADA-GURMUKHI", "[:KANNADA:]", "[:GURMUKHI:]",
1514     "[\\u0A01\\u0A02\\u0a33\\u0a36\\u0a3c\\u0a70\\u0a71\\u0A59\\u0A5A\\u0A5B\\u0A5C\\u0A5E\\u0A72\\u0A73\\u0A74]", /*roundtrip exclusions*/
1515
1516     "GURMUKHI-KANNADA", "[:GURMUKHI:]", "[:KANNADA:]",
1517     "[{\\u0cb0\\u0cbc}{\\u0cb3\\u0cbc}\\u0c82\\u0c83\\u0cb3\\u0cb6\\u0cc4\\u0cc3\\u0cc6\\u0cca\\u0cd5\\u0cd6\\u0C8B\\u0C8C\\u0C8E\\u0C92\\u0CB1\\u0CB7\\u0cbd\\u0CE0\\u0CE1\\u0cde]", /*roundtrip exclusions*/
1518
1519     "MALAYALAM-GURMUKHI", "[:MALAYALAM:]", "[:GURMUKHI:]",
1520     "[\\u0A01\\u0A02\\u0a4b\\u0a4c\\u0a33\\u0a36\\u0a3c\\u0a70\\u0a71\\u0A59\\u0A5A\\u0A5B\\u0A5C\\u0A5E\\u0A72\\u0A73\\u0A74]", /*roundtrip exclusions*/
1521
1522     "GURMUKHI-MALAYALAM", "[:GURMUKHI:]", "[:MALAYALAM:]",
1523     "[\\u0d02\\u0d03\\u0d33\\u0d36\\u0d43\\u0d46\\u0d4a\\u0d4c\\u0d57\\u0D0B\\u0D0C\\u0D0E\\u0D12\\u0D31\\u0D34\\u0D37\\u0D60\\u0D61]", /*roundtrip exclusions*/
1524
1525     "GUJARATI-ORIYA", "[:GUJARATI:]", "[:ORIYA:]",
1526     "[\\u0b56\\u0b57\\u0B0C\\u0B5F\\u0B61\\u0b70\\u0b71]", /*roundtrip exclusions*/
1527
1528     "ORIYA-GUJARATI", "[:ORIYA:]", "[:GUJARATI:]",
1529     "[\\u0Ac4\\u0Ac5\\u0Ac9\\u0Ac7\\u0A8D\\u0A91\\u0AB5\\u0Ad0]", /*roundtrip exclusions*/
1530
1531     "TAMIL-GUJARATI", "[:TAMIL:]", "[:GUJARATI:]",
1532     "[\\u0A81\\u0a8c\\u0abc\\u0ac3\\u0Ac4\\u0Ac5\\u0Ac9\\u0Ac7\\u0A8B\\u0A8D\\u0A91\\u0A96\\u0A97\\u0A98\\u0A9B\\u0A9D\\u0AA0\\u0AA1\\u0AA2\\u0AA5\\u0AA6\\u0AA7\\u0AAB\\u0AAC\\u0AAD\\u0AB6\\u0ABD\\u0AD0\\u0AE0\\u0AE1]", /*roundtrip exclusions*/
1533
1534     "GUJARATI-TAMIL", "[:GUJARATI:]", "[:TAMIL:]",
1535     "[\\u0Bc6\\u0Bca\\u0Bd7\\u0B8E\\u0B92\\u0BA9\\u0BB1\\u0BB4]", /*roundtrip exclusions*/
1536
1537     "TELUGU-GUJARATI", "[:TELUGU:]", "[:GUJARATI:]",
1538     "[\\u0abc\\u0Ac5\\u0Ac9\\u0A8D\\u0A91\\u0ABD\\u0Ad0]", /*roundtrip exclusions*/
1539
1540     "GUJARATI-TELUGU", "[:GUJARATI:]", "[:TELUGU:]",
1541     "[\\u0c46\\u0c4a\\u0c55\\u0c56\\u0C0C\\u0C0E\\u0C12\\u0C31\\u0C61]", /*roundtrip exclusions*/
1542
1543     "KANNADA-GUJARATI", "[:KANNADA:]", "[:GUJARATI:]",
1544     "[\\u0A81\\u0abc\\u0Ac5\\u0Ac9\\u0A8D\\u0A91\\u0ABD\\u0Ad0]", /*roundtrip exclusions*/
1545
1546     "GUJARATI-KANNADA", "[:GUJARATI:]", "[:KANNADA:]",
1547     "[{\\u0cb0\\u0cbc}{\\u0cb3\\u0cbc}\\u0cc6\\u0cca\\u0cd5\\u0cd6\\u0C8C\\u0C8E\\u0C92\\u0CB1\\u0CDE\\u0CE1]", /*roundtrip exclusions*/
1548
1549     "MALAYALAM-GUJARATI", "[:MALAYALAM:]", "[:GUJARATI:]",
1550     "[\\u0A81\\u0ac4\\u0acb\\u0acc\\u0abc\\u0Ac5\\u0Ac9\\u0A8D\\u0A91\\u0ABD\\u0Ad0]", /*roundtrip exclusions*/
1551
1552     "GUJARATI-MALAYALAM", "[:GUJARATI:]", "[:MALAYALAM:]",
1553     "[\\u0d46\\u0d4a\\u0d4c\\u0d55\\u0d57\\u0D0C\\u0D0E\\u0D12\\u0D31\\u0D34\\u0D61]", /*roundtrip exclusions*/
1554
1555     "TAMIL-ORIYA", "[:TAMIL:]", "[:ORIYA:]",
1556     "[\\u0B01\\u0b3c\\u0b43\\u0b56\\u0B0B\\u0B0C\\u0B16\\u0B17\\u0B18\\u0B1B\\u0B1D\\u0B20\\u0B21\\u0B22\\u0B25\\u0B26\\u0B27\\u0B2B\\u0B2C\\u0B2D\\u0B36\\u0B3D\\u0B5C\\u0B5D\\u0B5F\\u0B60\\u0B61\\u0b70\\u0b71]", /*roundtrip exclusions*/
1557
1558     "ORIYA-TAMIL", "[:ORIYA:]", "[:TAMIL:]",
1559     "[\\u0bc6\\u0bca\\u0bc7\\u0B8E\\u0B92\\u0BA9\\u0BB1\\u0BB4\\u0BB5]", /*roundtrip exclusions*/
1560
1561     "TELUGU-ORIYA", "[:TELUGU:]", "[:ORIYA:]",
1562     "[\\u0b3c\\u0b57\\u0b56\\u0B3D\\u0B5C\\u0B5D\\u0B5F\\u0b70\\u0b71]", /*roundtrip exclusions*/
1563
1564     "ORIYA-TELUGU", "[:ORIYA:]", "[:TELUGU:]",
1565     "[\\u0c44\\u0c46\\u0c4a\\u0c55\\u0c47\\u0C0E\\u0C12\\u0C31\\u0C35]", /*roundtrip exclusions*/
1566
1567     "KANNADA-ORIYA", "[:KANNADA:]", "[:ORIYA:]",
1568     "[\\u0B01\\u0b3c\\u0b57\\u0B3D\\u0B5C\\u0B5D\\u0B5F\\u0b70\\u0b71]", /*roundtrip exclusions*/
1569
1570     "ORIYA-KANNADA", "[:ORIYA:]", "[:KANNADA:]",
1571     "[{\\u0cb0\\u0cbc}{\\u0cb3\\u0cbc}\\u0cc4\\u0cc6\\u0cca\\u0cd5\\u0cc7\\u0C8E\\u0C92\\u0CB1\\u0CB5\\u0CDE]", /*roundtrip exclusions*/
1572
1573     "MALAYALAM-ORIYA", "[:MALAYALAM:]", "[:ORIYA:]",
1574     "[\\u0B01\\u0b3c\\u0b56\\u0B3D\\u0B5C\\u0B5D\\u0B5F\\u0b70\\u0b71]", /*roundtrip exclusions*/
1575
1576     "ORIYA-MALAYALAM", "[:ORIYA:]", "[:MALAYALAM:]",
1577     "[\\u0D47\\u0D46\\u0D4a\\u0D0E\\u0D12\\u0D31\\u0D34\\u0D35]", /*roundtrip exclusions*/
1578
1579     "TELUGU-TAMIL", "[:TELUGU:]", "[:TAMIL:]",
1580     "[\\u0bd7\\u0ba9\\u0bb4]", /*roundtrip exclusions*/
1581
1582     "TAMIL-TELUGU", "[:TAMIL:]", "[:TELUGU:]",
1583     "[\\u0C01\\u0c43\\u0c44\\u0c46\\u0c47\\u0c55\\u0c56\\u0c66\\u0C0B\\u0C0C\\u0C16\\u0C17\\u0C18\\u0C1B\\u0C1D\\u0C20\\u0C21\\u0C22\\u0C25\\u0C26\\u0C27\\u0C2B\\u0C2C\\u0C2D\\u0C36\\u0C60\\u0C61]", /*roundtrip exclusions*/
1584
1585     "KANNADA-TAMIL", "[:KANNADA:]", "[:TAMIL:]",
1586     "[\\u0bd7\\u0bc6\\u0ba9\\u0bb4]", /*roundtrip exclusions*/
1587
1588     "TAMIL-KANNADA", "[:TAMIL:]", "[:KANNADA:]",
1589     "[\\u0cc3\\u0cc4\\u0cc6\\u0cc7\\u0cd5\\u0cd6\\u0C8B\\u0C8C\\u0C96\\u0C97\\u0C98\\u0C9B\\u0C9D\\u0CA0\\u0CA1\\u0CA2\\u0CA5\\u0CA6\\u0CA7\\u0CAB\\u0CAC\\u0CAD\\u0CB6\\u0cbc\\u0cbd\\u0CDE\\u0CE0\\u0CE1]", /*roundtrip exclusions*/
1590
1591     "MALAYALAM-TAMIL", "[:MALAYALAM:]", "[:TAMIL:]",
1592     "[\\u0ba9]", /*roundtrip exclusions*/
1593
1594     "TAMIL-MALAYALAM", "[:TAMIL:]", "[:MALAYALAM:]",
1595     "[\\u0d43\\u0d12\\u0D0B\\u0D0C\\u0D16\\u0D17\\u0D18\\u0D1B\\u0D1D\\u0D20\\u0D21\\u0D22\\u0D25\\u0D26\\u0D27\\u0D2B\\u0D2C\\u0D2D\\u0D36\\u0D60\\u0D61]", /*roundtrip exclusions*/
1596
1597     "KANNADA-TELUGU", "[:KANNADA:]", "[:TELUGU:]",
1598     "[\\u0C01\\u0c3f\\u0c46\\u0c48\\u0c4a]", /*roundtrip exclusions*/
1599
1600     "TELUGU-KANNADA", "[:TELUGU:]", "[:KANNADA:]",
1601     "[\\u0cc8\\u0cd5\\u0cd6\\u0cbc\\u0cbd\\u0CDE]", /*roundtrip exclusions*/
1602
1603     "MALAYALAM-TELUGU", "[:MALAYALAM:]", "[:TELUGU:]",
1604     "[\\u0C01\\u0c44\\u0c4a\\u0c4c\\u0c4b\\u0c55\\u0c56]", /*roundtrip exclusions*/
1605
1606     "TELUGU-MALAYALAM", "[:TELUGU:]", "[:MALAYALAM:]",
1607     "[\\u0d4c\\u0d57\\u0D34]", /*roundtrip exclusions*/
1608
1609     "MALAYALAM-KANNADA", "[:MALAYALAM:]", "[:KANNADA:]",
1610     "[\\u0cbc\\u0cbd\\u0cc4\\u0cc6\\u0cca\\u0ccc\\u0ccb\\u0cd5\\u0cd6\\u0cDe]", /*roundtrip exclusions*/
1611
1612     "KANNADA-MALAYALAM", "[:KANNADA:]", "[:MALAYALAM:]",
1613     "[\\u0d4c\\u0d57\\u0d46\\u0D34]", /*roundtrip exclusions*/
1614
1615     "Latin-Bengali",latinForIndic, "[[:Bengali:][\\u0964\\u0965]]",
1616     "[\\u0965\\u09f0-\\u09fa]" /*roundtrip exclusions*/ ,
1617
1618     "Latin-Gurmukhi", latinForIndic, "[[:Gurmukhi:][\\u0964\\u0965]]",
1619     "[\\u0a01\\u0965\\u0a02\\u0a72\\u0a73\\u0a74]" /*roundtrip exclusions*/,
1620
1621     "Latin-Gujarati",latinForIndic, "[[:Gujarati:][\\u0964\\u0965]]",
1622     "[\\u0965]" /*roundtrip exclusions*/,
1623
1624     "Latin-Oriya",latinForIndic, "[[:Oriya:][\\u0964\\u0965]]",
1625     "[\\u0965\\u0b70]" /*roundtrip exclusions*/,
1626
1627     "Latin-Tamil",latinForIndic, "[:Tamil:]",
1628     NULL /*roundtrip exclusions*/,
1629
1630     "Latin-Telugu",latinForIndic, "[:Telugu:]",
1631     NULL /*roundtrip exclusions*/,
1632
1633     "Latin-Kannada",latinForIndic, "[:Kannada:]",
1634     NULL /*roundtrip exclusions*/,
1635
1636     "Latin-Malayalam",latinForIndic, "[:Malayalam:]",
1637     NULL /*roundtrip exclusions*/
1638
1639 };
1640
1641 void TransliteratorRoundTripTest::TestDebug(const char* name,const char fromSet[],
1642                                             const char* toSet,const char* exclusions){
1643
1644     RTTest test(name);
1645     Legal *legal = new LegalIndic();
1646     test.test(UnicodeString(fromSet,""),UnicodeString(toSet,""),exclusions,this,quick,legal);
1647 }
1648
1649 void TransliteratorRoundTripTest::TestInterIndic() {
1650     //TestDebug("Latin-Gurmukhi", latinForIndic, "[:Gurmukhi:]","[\\u0965\\u0a02\\u0a72\\u0a73\\u0a74]",TRUE);
1651     int32_t num = (int32_t)(sizeof(interIndicArray)/(INTER_INDIC_ARRAY_WIDTH*sizeof(char*)));
1652     if(quick){
1653         logln("Testing only 5 of %i. Skipping rest (use -e for exhaustive)",num);
1654         num = 5;
1655     }
1656     for(int i = 0; i < num;i++){
1657         RTTest test(interIndicArray[i*INTER_INDIC_ARRAY_WIDTH + 0]);
1658         Legal *legal = new LegalIndic();
1659
1660         test.test(  interIndicArray[i*INTER_INDIC_ARRAY_WIDTH + 1],
1661                     interIndicArray[i*INTER_INDIC_ARRAY_WIDTH + 2],
1662                     interIndicArray[i*INTER_INDIC_ARRAY_WIDTH + 3], // roundtrip exclusions
1663                     this, quick, legal, 50);
1664
1665         delete legal;
1666     }
1667
1668 }
1669
1670 // end indic tests ----------------------------------------------------------
1671
1672 #endif /* #if !UCONFIG_NO_TRANSLITERATION */