icuSources/tools/dumpce/dumpce.cpp

   1 /********************************************************************
   2  * COPYRIGHT:
   3  * Copyright (C) 2001-2003 IBM, Inc.   All Rights Reserved.
   4  *
   5  ********************************************************************/
   6 /********************************************************************************
   7 *
   8 * File dumpce.cpp
   9 *
  10 * Modification History:
  11 * Name          Date           Description
  12 * synwee        May 31 2001    Creation
  13 *
  14 *********************************************************************************
  15 */
  16
  17 /**
  18 * This program outputs the collation elements used for a requested tailoring.
  19 *
  20 * Usage:
  21 *     dumpce options... please check main function.
  22 */
  23 #include <unicode/utypes.h>
  24 #include <unicode/ucol.h>
  25 #include <unicode/uloc.h>
  26 #include <unicode/ucoleitr.h>
  27 #include <unicode/uchar.h>
  28 #include <unicode/uscript.h>
  29 #include <unicode/utf16.h>
  30 #include <unicode/putil.h>
  31 #include <unicode/ustring.h>
  32 #include <stdio.h>
  33 #include <stdlib.h>
  34 #include <string.h>
  35 #include <time.h>
  36 #include "ucol_tok.h"
  37 #include "cstring.h"
  38 #include "uoptions.h"
  39 #include "ucol_imp.h"
  40 #include <unicode/ures.h>
  41 #include <unicode/uniset.h>
  42 #include <unicode/usetiter.h>
  43
  44 /**
  45 * Command line option variables.
  46 * These global variables are set according to the options specified on the
  47 * command line by the user.
  48 */
  49 static UOption options[]={
  50     /* 00 */ UOPTION_HELP_H,
  51     /* 01 */ UOPTION_HELP_QUESTION_MARK,
  52     /* 02 */ {"locale",        NULL, NULL, NULL, 'l', UOPT_REQUIRES_ARG, 0},
  53     /* 03 */ {"serialize",     NULL, NULL, NULL, 'z', UOPT_NO_ARG, 0},
  54         /* 04 */ UOPTION_DESTDIR,
  55     /* 05 */ UOPTION_SOURCEDIR,
  56     /* 06 */ {"attribute",     NULL, NULL, NULL, 'a', UOPT_REQUIRES_ARG, 0},
  57     /* 07 */ {"rule",          NULL, NULL, NULL, 'r', UOPT_REQUIRES_ARG, 0},
  58     /* 08 */ {"normalization", NULL, NULL, NULL, 'n', UOPT_REQUIRES_ARG, 0},
  59     /* 09 */ {"scripts",       NULL, NULL, NULL, 't', UOPT_NO_ARG, 0},
  60     /* 10 */ {"reducehan",     NULL, NULL, NULL, 'e', UOPT_NO_ARG, 0},
  61         /* 11 */ UOPTION_VERBOSE,
  62     /* 12 */ {"wholescripts",      NULL, NULL, NULL, 'W', UOPT_NO_ARG, 0}
  63 };
  64
  65 /**
  66 * Collator used in this program
  67 */
  68 static UCollator *COLLATOR_;
  69 /**
  70 * Output strea, used in this program
  71 */
  72 static FILE *OUTPUT_;
  73
  74 static UColAttributeValue ATTRIBUTE_[UCOL_ATTRIBUTE_COUNT] = {
  75     UCOL_DEFAULT, UCOL_DEFAULT, UCOL_DEFAULT, UCOL_DEFAULT, UCOL_DEFAULT,
  76     UCOL_DEFAULT, UCOL_DEFAULT, UCOL_DEFAULT,
  77 };
  78
  79 typedef struct {
  80     int   value;
  81     char *name;
  82 } EnumNameValuePair;
  83
  84 static const EnumNameValuePair ATTRIBUTE_NAME_[] = {
  85     {UCOL_FRENCH_COLLATION, "UCOL_FRENCH_COLLATION"},
  86     {UCOL_ALTERNATE_HANDLING, "UCOL_ALTERNATE_HANDLING"},
  87     {UCOL_CASE_FIRST, "UCOL_CASE_FIRST"},
  88     {UCOL_CASE_LEVEL, "UCOL_CASE_LEVEL"},
  89     {UCOL_NORMALIZATION_MODE,
  90         "UCOL_NORMALIZATION_MODE|UCOL_DECOMPOSITION_MODE"},
  91     {UCOL_STRENGTH, "UCOL_STRENGTH"},
  92         {UCOL_HIRAGANA_QUATERNARY_MODE, "UCOL_HIRAGANA_QUATERNARY_MODE"},
  93     {UCOL_NUMERIC_COLLATION, "UCOL_NUMERIC_COLLATION"},
  94     NULL
  95 };
  96
  97 static const EnumNameValuePair ATTRIBUTE_VALUE_[] = {
  98     {UCOL_PRIMARY, "UCOL_PRIMARY"},
  99     {UCOL_SECONDARY, "UCOL_SECONDARY"},
 100     {UCOL_TERTIARY, "UCOL_TERTIARY|UCOL_DEFAULT_STRENGTH"},
 101     {UCOL_QUATERNARY, "UCOL_QUATERNARY"},
 102     {UCOL_IDENTICAL, "UCOL_IDENTICAL"},
 103     {UCOL_OFF, "UCOL_OFF"},
 104     {UCOL_ON, "UCOL_ON"},
 105     {UCOL_SHIFTED, "UCOL_SHIFTED"},
 106     {UCOL_NON_IGNORABLE, "UCOL_NON_IGNORABLE"},
 107     {UCOL_LOWER_FIRST, "UCOL_LOWER_FIRST"},
 108     {UCOL_UPPER_FIRST, "UCOL_UPPER_FIRST"},
 109     NULL
 110 };
 111
 112 typedef struct {
 113     UChar ch[32];
 114     int   count; // number of codepoint
 115     UBool tailored;
 116 } ScriptElement;
 117
 118 /**
 119 * Writes the hexadecimal of a null-terminated array of codepoints into a
 120 * file
 121 * @param f UFILE instance to store
 122 * @param c codepoints array
 123 */
 124 void serialize(FILE *f, const UChar *c)
 125 {
 126     UChar cp = *(c ++);
 127
 128     fprintf(f, " %04x", cp);
 129
 130     while (*c != 0) {
 131         cp = *(c ++);
 132         fprintf(f, " %04x", cp);
 133     }
 134 }
 135
 136 /**
 137 * Writes the hexadecimal of a non-null-terminated array of codepoints into a
 138 * file
 139 * @param f UFILE instance to store
 140 * @param c codepoints array
 141 * @param l codepoints array length
 142 */
 143 void serialize(FILE *f, const UChar *c, int l)
 144 {
 145     int   count = 1;
 146     UChar cp    = *(c ++);
 147
 148     fprintf(f, " %04x", cp);
 149
 150     while (count < l) {
 151         cp = *(c ++);
 152         fprintf(f, " %04x", cp);
 153         count ++;
 154     }
 155 }
 156
 157 /**
 158 * Sets the iterator to the argument string and outputs the collation elements.
 159 * @param f file output stream
 160 * @param iter collation element iterator
 161 */
 162 void serialize(FILE *f, UCollationElements *iter) {
 163     UChar   *codepoint = iter->iteratordata_.string;
 164     // unlikely that sortkeys will be over this size
 165     uint8_t  sortkey[64];
 166     uint8_t *psortkey = sortkey;
 167     int      sortkeylength = 0;
 168
 169     if (iter->iteratordata_.flags & UCOL_ITER_HASLEN) {
 170         serialize(f, codepoint, iter->iteratordata_.endp - codepoint);
 171         sortkeylength = ucol_getSortKey(iter->iteratordata_.coll, codepoint,
 172                         iter->iteratordata_.endp - codepoint, sortkey, 64);
 173     }
 174     else {
 175         serialize(f, codepoint);
 176         sortkeylength = ucol_getSortKey(iter->iteratordata_.coll, codepoint,
 177                                         -1, sortkey, 64);
 178     }
 179     if (options[11].doesOccur) {
 180         serialize(stdout, codepoint);
 181         fprintf(stdout, "\n");
 182     }
 183
 184     fprintf(f, "; ");
 185
 186     UErrorCode error = U_ZERO_ERROR;
 187     uint32_t ce = ucol_next(iter, &error);
 188     if (U_FAILURE(error)) {
 189         fprintf(f, "Error retrieving collation elements\n");
 190         return;
 191     }
 192
 193     while (TRUE) {
 194         fprintf(f, "[");
 195         if (UCOL_PRIMARYORDER(ce) != 0) {
 196             fprintf(f, "%04x", UCOL_PRIMARYORDER(ce));
 197         }
 198         fprintf(f, ",");
 199         if (UCOL_SECONDARYORDER(ce) != 0) {
 200             fprintf(f, " %02x", UCOL_SECONDARYORDER(ce));
 201         }
 202         fprintf(f, ",");
 203         if (UCOL_TERTIARYORDER(ce) != 0) {
 204             fprintf(f, " %02x", UCOL_TERTIARYORDER(ce));
 205         }
 206         fprintf(f, "] ");
 207
 208         ce = ucol_next(iter, &error);
 209         if (ce == UCOL_NULLORDER) {
 210             break;
 211         }
 212         if (U_FAILURE(error)) {
 213             fprintf(stdout, "Error retrieving collation elements");
 214             return;
 215         }
 216     }
 217
 218     if (sortkeylength > 64) {
 219         fprintf(f, "Sortkey exceeds pre-allocated size");
 220     }
 221
 222     fprintf(f, "[");
 223     while (TRUE) {
 224         fprintf(f, "%02x", *psortkey);
 225         psortkey ++;
 226         if ((*psortkey) == 0) {
 227             break;
 228         }
 229         fprintf(f, " ");
 230     }
 231     fprintf(f, "]\n");
 232 }
 233
 234 /**
 235 * Serializes the contraction within the given argument rule
 236 * @param f file output stream
 237 * @param r rule
 238 * @param rlen rule length
 239 * @param contractionsonly flag to indicate if only contractions are to be
 240 *                         output or all collation elements
 241 * @param iter iterator to iterate over collation elements
 242 */
 243 void serialize(FILE *f, UChar *rule, int rlen, UBool contractiononly,
 244                UCollationElements *iter) {
 245     const UChar           *current  = NULL;
 246           uint32_t         strength = 0;
 247           uint32_t         chOffset = 0;
 248           uint32_t         chLen    = 0;
 249           uint32_t         exOffset = 0;
 250           uint32_t         exLen    = 0;
 251           uint32_t         prefixOffset = 0;
 252           uint32_t         prefixLen    = 0;
 253           uint8_t          specs    = 0;
 254           UBool            rstart   = TRUE;
 255           UColTokenParser  src;
 256           UColOptionSet    opts;
 257           UParseError      parseError;
 258           UErrorCode       error    = U_ZERO_ERROR;
 259
 260     src.opts = &opts;
 261
 262     src.source       = rule;
 263         src.current = rule;
 264     src.end          = rule + rlen;
 265     src.extraCurrent = src.end;
 266     src.extraEnd     = src.end + UCOL_TOK_EXTRA_RULE_SPACE_SIZE;
 267
 268
 269     while ((current = ucol_tok_parseNextToken(&src, rstart, &parseError,
 270                                               &error)) != NULL) {
 271       chOffset = src.parsedToken.charsOffset;
 272       chLen = src.parsedToken.charsLen;
 273         // contractions handled here
 274         if (!contractiononly || chLen > 1) {
 275             ucol_setText(iter, rule + chOffset, chLen, &error);
 276             if (U_FAILURE(error)) {
 277                 fprintf(stdout, "Error setting text in iterator\n");
 278                 return;
 279             }
 280             serialize(f, iter);
 281         }
 282         rstart = FALSE;
 283     }
 284 }
 285
 286 /**
 287 * Prints the attribute values in the argument collator into the output stream
 288 * @param collator
 289 */
 290 void outputAttribute(UCollator *collator, UErrorCode *error)
 291 {
 292     UColAttribute attribute = UCOL_FRENCH_COLLATION;
 293     while (attribute < UCOL_ATTRIBUTE_COUNT) {
 294         int count = 0;
 295         while (TRUE) {
 296             // getting attribute name
 297             if (ATTRIBUTE_NAME_[count].value == attribute) {
 298                 fprintf(OUTPUT_, "%s = ", ATTRIBUTE_NAME_[count].name);
 299                 break;
 300             }
 301             count ++;
 302         }
 303         count = 0;
 304         int attributeval = ucol_getAttribute(collator, attribute, error);
 305         if (U_FAILURE(*error)) {
 306             fprintf(stdout, "Failure in reading collator attribute\n");
 307             return;
 308         }
 309         while (TRUE) {
 310             // getting attribute value
 311             if (ATTRIBUTE_VALUE_[count].value == attributeval) {
 312                 fprintf(OUTPUT_, "%s\n", ATTRIBUTE_VALUE_[count].name);
 313                 break;
 314             }
 315             count ++;
 316         }
 317         attribute = (UColAttribute)(attribute + 1);
 318     }
 319 }
 320
 321 /**
 322 * Prints the normalization mode in the argument collator into the output stream
 323 * @param collator
 324 */
 325 void outputNormalization(UCollator *collator)
 326 {
 327         UErrorCode status = U_ZERO_ERROR;
 328     int normmode = ucol_getAttribute(collator, UCOL_NORMALIZATION_MODE, &status);
 329     int count = 0;
 330     while (TRUE) {
 331         // getting attribute name
 332         if (ATTRIBUTE_VALUE_[count].value == normmode) {
 333             break;
 334         }
 335         count ++;
 336     }
 337     fprintf(OUTPUT_, "NORMALIZATION MODE = %s\n",
 338             ATTRIBUTE_VALUE_[count].name);
 339 }
 340
 341 /**
 342 * Output the collation element belonging to the locale into a file
 343 * @param locale string
 344 * @param fullrules flag to indicate if only tailored collation elements are to
 345 *        be output or all collation elements
 346 */
 347 void serialize(const char *locale, UBool tailoredonly) {
 348     UErrorCode  error              = U_ZERO_ERROR;
 349     UChar       str[128];
 350     int         strlen = 0;
 351
 352     fprintf(OUTPUT_, "# This file contains the serialized collation elements\n");
 353     fprintf(OUTPUT_, "# as of the collation version indicated below.\n");
 354     fprintf(OUTPUT_, "# Data format: xxxx xxxx..; [yyyy, yy, yy] [yyyy, yy, yy] ... [yyyy, yy, yy] [zz zz..\n");
 355     fprintf(OUTPUT_, "#              where xxxx are codepoints in hexadecimals,\n");
 356     fprintf(OUTPUT_, "#              yyyyyyyy are the corresponding\n");
 357     fprintf(OUTPUT_, "#              collation elements in hexadecimals\n");
 358     fprintf(OUTPUT_, "#              and zz are the sortkey values in hexadecimals\n");
 359
 360     fprintf(OUTPUT_, "\n# Collator information\n");
 361
 362     fprintf(OUTPUT_, "\nLocale: %s\n", locale);
 363     fprintf(stdout, "Locale: %s\n", locale);
 364     UVersionInfo version;
 365     ucol_getVersion(COLLATOR_, version);
 366     fprintf(OUTPUT_, "Version number: %d.%d.%d.%d\n",
 367                       version[0], version[1], version[2], version[3]);
 368     outputAttribute(COLLATOR_, &error);
 369     outputNormalization(COLLATOR_);
 370
 371     UCollationElements *iter = ucol_openElements(COLLATOR_, str, strlen,
 372                                                  &error);
 373     if (U_FAILURE(error)) {
 374         fprintf(stdout, "Error creating iterator\n");
 375         return;
 376     }
 377
 378     if (!tailoredonly) {
 379         fprintf(OUTPUT_, "\n# Range of unicode characters\n\n");
 380         UChar32     codepoint          = 0;
 381         while (codepoint <= UCHAR_MAX_VALUE) {
 382             if (u_isdefined(codepoint)) {
 383                 strlen = 0;
 384                 UTF16_APPEND_CHAR_UNSAFE(str, strlen, codepoint);
 385                 str[strlen] = 0;
 386                 ucol_setText(iter, str, strlen, &error);
 387                 if (U_FAILURE(error)) {
 388                     fprintf(stdout, "Error setting text in iterator\n");
 389                     return;
 390                 }
 391                 serialize(OUTPUT_, iter);
 392             }
 393             codepoint ++;
 394         }
 395     }
 396
 397     UChar    ucarules[0x10000];
 398     UChar   *rules;
 399     int32_t  rulelength = 0;
 400     rules      = ucarules;
 401
 402     if (tailoredonly) {
 403               int32_t  rulelength = 0;
 404         const UChar   *temp = ucol_getRules(COLLATOR_, &rulelength);
 405         if (rulelength + UCOL_TOK_EXTRA_RULE_SPACE_SIZE > 0x10000) {
 406             rules = (UChar *)malloc(sizeof(UChar) *
 407                                 (rulelength + UCOL_TOK_EXTRA_RULE_SPACE_SIZE));
 408         }
 409         memcpy(rules, temp, rulelength * sizeof(UChar));
 410         rules[rulelength] = 0;
 411         fprintf(OUTPUT_, "\n# Tailorings\n\n");
 412         serialize(OUTPUT_, rules, rulelength, FALSE, iter);
 413         if (rules != ucarules) {
 414             free(rules);
 415         }
 416     }
 417     else {
 418         rulelength = ucol_getRulesEx(COLLATOR_, UCOL_FULL_RULES, ucarules,
 419                                      0x10000);
 420         if (rulelength + UCOL_TOK_EXTRA_RULE_SPACE_SIZE > 0x10000) {
 421             rules = (UChar *)malloc(sizeof(UChar) *
 422                                 (rulelength + UCOL_TOK_EXTRA_RULE_SPACE_SIZE));
 423             rulelength = ucol_getRulesEx(COLLATOR_, UCOL_FULL_RULES, rules,
 424                                          rulelength);
 425         }
 426         fprintf(OUTPUT_, "\n# Contractions\n\n");
 427         serialize(OUTPUT_, rules, rulelength, TRUE, iter);
 428         if (rules != ucarules) {
 429             free(rules);
 430         }
 431     }
 432
 433     ucol_closeElements(iter);
 434 }
 435
 436 /**
 437 * Sets the collator with the attribute values
 438 * @param collator
 439 * @param error status
 440 */
 441 void setAttributes(UCollator *collator, UErrorCode *error)
 442 {
 443     int count = 0;
 444     while (count < UCOL_ATTRIBUTE_COUNT) {
 445         if (ATTRIBUTE_[count] != UCOL_DEFAULT) {
 446             ucol_setAttribute(collator, (UColAttribute)count,
 447                               ATTRIBUTE_[count], error);
 448             if (U_FAILURE(*error)) {
 449                 return;
 450             }
 451         }
 452         count ++;
 453     }
 454 }
 455
 456 /**
 457 * Appends directory path with an ending seperator if necessary.
 458 * @param path with enough space to append one seperator
 459 * @return new directory path length
 460 */
 461 int appendDirSeparator(char *dir)
 462 {
 463     int dirlength = strlen(dir);
 464     char dirending = dir[dirlength - 1];
 465     if (dirending != U_FILE_SEP_CHAR) {
 466         dir[dirlength] = U_FILE_SEP_CHAR;
 467         dir[dirlength + 1] = 0;
 468         return dirlength + 1;
 469     }
 470     return dirlength;
 471 }
 472
 473 /**
 474 * Output the collation element into a file
 475 */
 476 void serialize() {
 477     char filename[128];
 478     int  dirlength = 0;
 479
 480     if (options[4].doesOccur) {
 481         strcpy(filename, options[4].value);
 482         dirlength = appendDirSeparator(filename);
 483     }
 484
 485     if (options[2].doesOccur) {
 486         const char    *locale      = (char *)options[2].value;
 487               int32_t  localeindex = 0;
 488
 489         if (strcmp(locale, "all") == 0) {
 490             if (options[4].doesOccur) {
 491                 strcat(filename, "UCA.txt");
 492                 OUTPUT_ = fopen(filename, "w");
 493                 if (OUTPUT_ == NULL) {
 494                     fprintf(stdout, "Cannot open file:%s\n", filename);
 495                     return;
 496                 }
 497             }
 498             fprintf(stdout, "UCA\n");
 499             UErrorCode error = U_ZERO_ERROR;
 500             COLLATOR_ = ucol_open("en_US", &error);
 501             if (U_FAILURE(error)) {
 502                 fprintf(stdout, "Collator creation failed:");
 503                 fprintf(stdout, u_errorName(error));
 504                 goto CLOSEUCA;
 505                 return;
 506             }
 507             setAttributes(COLLATOR_, &error);
 508             if (U_FAILURE(error)) {
 509                 fprintf(stdout, "Collator attribute setting failed:");
 510                 fprintf(stdout, u_errorName(error));
 511                 goto CLOSEUCA;
 512                 return;
 513             }
 514
 515             serialize("UCA", FALSE);
 516 CLOSEUCA :
 517             if (options[4].doesOccur) {
 518                 filename[dirlength] = 0;
 519                 fclose(OUTPUT_);
 520             }
 521             ucol_close(COLLATOR_);
 522             localeindex = ucol_countAvailable() - 1;
 523             fprintf(stdout, "Number of locales: %d\n", localeindex + 1);
 524             locale      = ucol_getAvailable(localeindex);
 525         }
 526
 527         while (TRUE) {
 528             UErrorCode error = U_ZERO_ERROR;
 529             COLLATOR_ = ucol_open(locale, &error);
 530             if (U_FAILURE(error)) {
 531                 fprintf(stdout, "Collator creation failed:");
 532                 fprintf(stdout, u_errorName(error));
 533                 goto CLOSETAILOR;
 534                 return;
 535             }
 536             setAttributes(COLLATOR_, &error);
 537             if (U_FAILURE(error)) {
 538                 fprintf(stdout, "Collator attribute setting failed:");
 539                 fprintf(stdout, u_errorName(error));
 540                 goto CLOSETAILOR;
 541                 return;
 542             }
 543
 544             if (options[4].doesOccur) {
 545                 strcat(filename, locale);
 546                 strcat(filename, ".txt");
 547                 OUTPUT_ = fopen(filename, "w");
 548                 if (OUTPUT_ == NULL) {
 549                     fprintf(stdout, "Cannot open file:%s\n", filename);
 550                     return;
 551                 }
 552             }
 553
 554             if (options[3].doesOccur) {
 555                 serialize(locale, TRUE);
 556             }
 557
 558             ucol_close(COLLATOR_);
 559
 560 CLOSETAILOR :
 561             if (options[4].doesOccur) {
 562                 filename[dirlength] = 0;
 563                 fclose(OUTPUT_);
 564             }
 565
 566             localeindex --;
 567             if (localeindex < 0) {
 568                 break;
 569             }
 570             locale = ucol_getAvailable(localeindex);
 571         }
 572     }
 573
 574     if (options[7].doesOccur) {
 575         char inputfilename[128];
 576         // rules are to be used
 577         if (options[5].doesOccur) {
 578             strcpy(inputfilename, options[5].value);
 579             appendDirSeparator(inputfilename);
 580         }
 581         strcat(inputfilename, options[7].value);
 582         FILE *input = fopen(inputfilename, "r");
 583         if (input == NULL) {
 584             fprintf(stdout, "Cannot open file:%s\n", filename);
 585             return;
 586         }
 587
 588         char   s[1024];
 589         UChar  rule[1024];
 590         UChar *prule = rule;
 591         int    size = 1024;
 592         // synwee TODO: make this part dynamic
 593         while (fscanf(input, "%[^\n]s", s) != EOF) {
 594             size -= u_unescape(s, prule, size);
 595             prule = prule + u_strlen(prule);
 596         }
 597         fclose(input);
 598
 599         if (options[4].doesOccur) {
 600             strcat(filename, "Rules.txt");
 601             OUTPUT_ = fopen(filename, "w");
 602             if (OUTPUT_ == NULL) {
 603                 fprintf(stdout, "Cannot open file:%s\n", filename);
 604                 return;
 605             }
 606         }
 607
 608         fprintf(stdout, "Rules\n");
 609         UErrorCode  error = U_ZERO_ERROR;
 610         UParseError parseError;
 611         COLLATOR_ = ucol_openRules(rule, u_strlen(rule), UCOL_DEFAULT,
 612                                    UCOL_DEFAULT_STRENGTH, &parseError, &error);
 613         if (U_FAILURE(error)) {
 614             fprintf(stdout, "Collator creation failed:");
 615             fprintf(stdout, u_errorName(error));
 616             goto CLOSERULES;
 617             return;
 618         }
 619         setAttributes(COLLATOR_, &error);
 620         if (U_FAILURE(error)) {
 621             fprintf(stdout, "Collator attribute setting failed:");
 622             fprintf(stdout, u_errorName(error));
 623             goto CLOSERULES;
 624             return;
 625         }
 626
 627         serialize("Rule-based", TRUE);
 628         ucol_close(COLLATOR_);
 629
 630 CLOSERULES :
 631         if (options[4].doesOccur) {
 632             filename[dirlength] = 0;
 633             fclose(OUTPUT_);
 634         }
 635     }
 636 }
 637
 638 /**
 639 * Parse for enum values.
 640 * Note this only works for positive enum values.
 641 * @param enumarray array containing names of the enum values in string and
 642 *        their corresponding value.
 643 *        declared enum value.
 644 * @param str string to be parsed
 645 * @return corresponding integer enum value or -1 if value is not found.
 646 */
 647 int parseEnums(const EnumNameValuePair enumarray[], const char *str)
 648 {
 649     const char *enumname = enumarray[0].name;
 650     int result = atoi(str);
 651     if (result == 0 && str[0] != '0') {
 652         while (strcmp(enumname, str) != 0) {
 653             // checking for multiple enum names sharing the same values
 654             enumname = strstr(enumname, str);
 655             if (enumname != NULL) {
 656                 int size = strchr(enumname, '|') - enumname;
 657                 if (size < 0) {
 658                     size = strlen(enumname);
 659                 }
 660                 if (size == (int)strlen(str)) {
 661                     return enumarray[result].value;
 662                 }
 663             }
 664             result ++;
 665             if (&(enumarray[result]) == NULL) {
 666                 return -1;
 667             }
 668             enumname = enumarray[result].name;
 669         }
 670     }
 671     return -1;
 672 }
 673
 674 /**
 675 * Parser for attribute name value pair
 676 */
 677 void parseAttributes() {
 678     char str[32];
 679     const char *pname = options[6].value;
 680     const char *pend  = options[6].value + strlen(options[6].value);
 681     const char *pvalue;
 682
 683     while (pname < pend) {
 684         pvalue = strchr(pname, '=');
 685         if (pvalue == NULL) {
 686             fprintf(stdout,
 687                     "No matching value found for attribute argument %s\n",
 688                     pname);
 689             return;
 690         }
 691         int count = pvalue - pname;
 692         strncpy(str, pname, count);
 693         str[count] = 0;
 694
 695         int name = parseEnums(ATTRIBUTE_NAME_, str);
 696         if (name == -1) {
 697             fprintf(stdout, "Attribute name not found: %s\n", str);
 698             return;
 699         }
 700
 701         pvalue ++;
 702         // getting corresponding enum value
 703         pname = strchr(pvalue, ',');
 704         if (pname == NULL) {
 705             pname = pend;
 706         }
 707         count = pname - pvalue;
 708         strncpy(str, pvalue, count);
 709         str[count] = 0;
 710         int value = parseEnums(ATTRIBUTE_VALUE_, str);
 711         if (value == -1) {
 712             fprintf(stdout, "Attribute value not found: %s\n", str);
 713             return;
 714         }
 715         ATTRIBUTE_[name] = (UColAttributeValue)value;
 716         pname ++;
 717     }
 718 }
 719
 720 /**
 721 * Checks if the locale argument is a base language
 722 * @param locale to be checked
 723 * @return TRUE if it is a base language
 724 */
 725 inline UBool checkLocaleForLanguage(const char *locale)
 726 {
 727     return strlen(locale) <= 2;
 728 }
 729
 730 /**
 731 * Converts a UChar array into its string form "xxxx xxxx"
 732 * @param ch array of UChar characters
 733 * @param count number of UChar characters
 734 */
 735 void outputUChar(UChar ch[], int count)
 736 {
 737     for (int i = 0; i < count; i ++) {
 738         fprintf(OUTPUT_, "%04X ", ch[i]);
 739     }
 740 }
 741
 742 /**
 743 * If it is a primary difference returns -1 or 1.
 744 * If it is a secondary difference returns -2 or 2.
 745 * If it is a tertiary difference returns -3 or 3.
 746 * If equals returns 0.
 747 */
 748 int compareSortKey(const void *elem1, const void *elem2)
 749 {
 750     // compare the 2 script element sort key
 751     UChar     *ch1   = ((ScriptElement *)elem1)->ch;
 752     UChar     *ch2   = ((ScriptElement *)elem2)->ch;
 753     int        size1 = ((ScriptElement *)elem1)->count;
 754     int        size2 = ((ScriptElement *)elem2)->count;
 755     UErrorCode error = U_ZERO_ERROR;
 756
 757     ucol_setStrength(COLLATOR_, UCOL_PRIMARY);
 758     int result = ucol_strcoll(COLLATOR_, ch1, size1, ch2, size2);
 759     if (result == 0) {
 760         ucol_setStrength(COLLATOR_, UCOL_SECONDARY);
 761         result = ucol_strcoll(COLLATOR_, ch1, size1, ch2, size2);
 762         if (result == 0) {
 763             ucol_setStrength(COLLATOR_, UCOL_TERTIARY);
 764             result = ucol_strcoll(COLLATOR_, ch1, size1, ch2, size2);
 765             if (result < 0) {
 766                 return -3;
 767             }
 768             if (result > 0) {
 769                 return 3;
 770             }
 771         }
 772         if (result < 0) {
 773             return -2;
 774         }
 775         if (result > 0) {
 776             return 2;
 777         }
 778     }
 779     return result;
 780 }
 781
 782 /**
 783 * Output serialized script elements
 784 * @param element the element to output
 785 * @param compare the comparison with the previous element
 786 * @param expansion flags TRUE if element has an expansion
 787 */
 788 void outputScriptElem(ScriptElement &element, int compare, UBool expansion)
 789 {
 790     switch (compare) {
 791     case 0:
 792         if (expansion) {
 793             fprintf(OUTPUT_, "<tr><td class='eq' title='[");
 794         }
 795         else {
 796             fprintf(OUTPUT_, "<tr><td class='q' title='[");
 797         }
 798         break;
 799     case -1:
 800         if (expansion) {
 801             fprintf(OUTPUT_, "<tr><td class='ep' title='[");
 802         }
 803         else {
 804             fprintf(OUTPUT_, "<tr><td class='p' title='[");
 805         }
 806         break;
 807     case -2:
 808         if (expansion) {
 809             fprintf(OUTPUT_, "<tr><td class='es' title='[");
 810         }
 811         else {
 812             fprintf(OUTPUT_, "<tr><td class='s' title='[");
 813         }
 814         break;
 815     default:
 816         if (expansion) {
 817             fprintf(OUTPUT_, "<tr><td class='et' title='[");
 818         }
 819         else {
 820             fprintf(OUTPUT_, "<tr><td class='t' title='[");
 821         }
 822     }
 823
 824     uint8_t sortkey[32];
 825     ucol_setStrength(COLLATOR_, UCOL_TERTIARY);
 826     ucol_getSortKey(COLLATOR_, element.ch, element.count, sortkey, 32);
 827     int i = 0;
 828     while (sortkey[i] != 0) {
 829         if (sortkey[i] == 1) {
 830             fprintf(OUTPUT_, " | ");
 831         }
 832         else {
 833             fprintf(OUTPUT_, "%02x", sortkey[i]);
 834         }
 835
 836         i ++;
 837     }
 838
 839     fprintf(OUTPUT_, "]'>");
 840
 841     UErrorCode error = U_ZERO_ERROR;
 842     char       utf8[64];
 843     UChar      nfc[32];
 844     int32_t    length = unorm_normalize(element.ch, element.count, UNORM_NFC, 0, nfc,
 845                                         32, &error);
 846     if (U_FAILURE(error)) {
 847         fprintf(stdout, "Error normalizing contractions to NFC\n");
 848     }
 849     u_strToUTF8(utf8, 64, &length, nfc, length, &error);
 850     if (U_FAILURE(error)) {
 851         fprintf(stdout, "Error converting UChar to utf8\n");
 852         return;
 853     }
 854
 855     fprintf(OUTPUT_, "%s<br>", utf8);
 856     fprintf(OUTPUT_, "<tt>");
 857     outputUChar(element.ch, element.count);
 858
 859     if (compare == 0) {
 860         fprintf(OUTPUT_, "</tt></td><td>&nbsp;</td><td>&nbsp;</td><td>&nbsp;</td><td>Q</td><td>");
 861     }
 862     else if (compare == -1) {
 863         fprintf(OUTPUT_, "</tt></td><td>P</td><td>&nbsp;</td><td>&nbsp;</td><td>&nbsp;</td><td>");
 864     }
 865     else if (compare == -2) {
 866         fprintf(OUTPUT_, "</tt></td><td>&nbsp;</td><td>S</td><td>&nbsp;</td><td>&nbsp;</td><td>");
 867     }
 868     else if (compare == -3) {
 869         fprintf(OUTPUT_, "</tt></td><td>&nbsp;</td><td>&nbsp;</td><td>T</td><td>&nbsp;</td><td>");
 870     }
 871
 872     i = 0;
 873     while (i < element.count) {
 874         char    str[128];
 875         UChar32 codepoint;
 876         UTF_NEXT_CHAR(element.ch, i, element.count, codepoint);
 877         int32_t temp = u_charName(codepoint, U_UNICODE_CHAR_NAME, str, 128,
 878                                       &error);
 879         if (U_FAILURE(error)) {
 880             fprintf(stdout, "Error getting character name\n");
 881             return;
 882         }
 883         if (element.tailored) {
 884             fprintf(OUTPUT_, "<b>");
 885         }
 886         fprintf(OUTPUT_, "%s", str);
 887         if (element.tailored) {
 888             fprintf(OUTPUT_, " *</b>");
 889         }
 890         if (i < element.count) {
 891             fprintf(OUTPUT_, "<br>\n");
 892         }
 893     }
 894
 895     fprintf(OUTPUT_, "</td></tr>\n");
 896 }
 897
 898 /**
 899 * Checks if codepoint belongs to scripts
 900 * @param script list
 901 * @param scriptcount number of scripts
 902 * @param codepoint to test
 903 * @return TRUE if codepoint belongs to scripts
 904 */
 905 UBool checkInScripts(UScriptCode script[], int scriptcount,
 906                      UChar32 codepoint)
 907 {
 908     UErrorCode error = U_ZERO_ERROR;
 909     for (int i = 0; i < scriptcount; i ++) {
 910         if (script[i] == USCRIPT_HAN && options[10].doesOccur) {
 911             if ((codepoint >= 0x2E80 && codepoint <= 0x2EE4) ||
 912                 (codepoint >= 0x2A672 && codepoint <= 0x2A6D6)) {
 913                 // reduce han
 914                 return TRUE;
 915             }
 916         }
 917         else if (uscript_getScript(codepoint, &error) == script[i]) {
 918             return TRUE;
 919         }
 920         if (U_FAILURE(error)) {
 921             fprintf(stdout, "Error checking character in scripts\n");
 922             return FALSE;
 923         }
 924     }
 925     return FALSE;
 926 }
 927
 928 /**
 929 * Checks if the set of codepoints belongs to the script
 930 * @param script list
 931 * @param scriptcount number of scripts
 932 * @param scriptelem
 933 * @return TRUE if all codepoints belongs to the script
 934 */
 935 inline UBool checkInScripts(UScriptCode script[], int scriptcount,
 936                            ScriptElement scriptelem)
 937 {
 938     int i = 0;
 939     while (i < scriptelem.count) {
 940         UChar32     codepoint;
 941         UTF_NEXT_CHAR(scriptelem.ch, i, scriptelem.count, codepoint);
 942         UErrorCode  error = U_ZERO_ERROR;
 943         if (checkInScripts(script, scriptcount, codepoint)) {
 944             return TRUE;
 945         }
 946     }
 947     return FALSE;
 948 }
 949
 950 /**
 951 * Gets the script elements and contractions belonging to the script
 952 * @param elems output list
 953 * @param locale locale
 954 * @return number of script elements
 955 * Add by Richard
 956 */
 957 int getScriptElementsFromExemplars(ScriptElement scriptelem[], const char* locale) {
 958         UErrorCode error = U_ZERO_ERROR;
 959     UChar32    codepoint = 0;
 960
 961         UResourceBundle* ures = ures_open(NULL, locale, &error);
 962         if (U_FAILURE(error)) {
 963                 fprintf(stdout, "Can not find resource bundle for locale: %s\n", locale);
 964         return -1;
 965         }
 966         int32_t length;
 967         const UChar* exemplarChars = ures_getStringByKey(ures, "ExemplarCharacters", &length, &error);
 968
 969         if (U_FAILURE(error)) {
 970                 fprintf(stdout, "Can not find ExemplarCharacters in resource bundle\n");
 971         return -1;
 972         }
 973
 974         UChar* upperChars = new UChar[length*2];
 975         if (upperChars == 0) {
 976                 fprintf(stdout, "Memory error\n");
 977         return -1;
 978         }
 979
 980         int32_t destLength = u_strToUpper(upperChars, length*2, exemplarChars, -1, locale, &error);
 981         if (U_FAILURE(error)) {
 982                 fprintf(stdout, "Error when u_strToUpper() \n");
 983         return -1;
 984         }
 985
 986         UChar* pattern = new UChar[length + destLength + 10];
 987         UChar left[2] = {0x005b, 0x0};
 988         UChar right[2] = {0x005d, 0x0};
 989         pattern = u_strcpy(pattern, left);
 990         pattern = u_strcat(pattern, exemplarChars);
 991         pattern = u_strcat(pattern, upperChars);
 992         pattern = u_strcat(pattern, right);
 993
 994         UnicodeSet * uniset = new UnicodeSet(UnicodeString(pattern), error);
 995         if (U_FAILURE(error)) {
 996                 fprintf(stdout, "Can not open USet \n");
 997         return -1;
 998         }
 999
1000         UnicodeSetIterator* usetiter = new UnicodeSetIterator(*uniset);
1001
1002         int32_t count = 0;
1003
1004         while (usetiter -> next()) {
1005                 if (usetiter -> isString()) {
1006                         UnicodeString strItem = usetiter -> getString();
1007
1008                         scriptelem[count].count = 0;
1009                         for (int i = 0; i < strItem.length(); i++) {
1010                                 codepoint = strItem.char32At(i);
1011                                 UTF16_APPEND_CHAR_UNSAFE(scriptelem[count].ch,
1012                                                                                 scriptelem[count].count, codepoint);
1013                                 scriptelem[count].tailored = FALSE;
1014                         }
1015                 } else {
1016                         codepoint = usetiter -> getCodepoint();
1017                         scriptelem[count].count = 0;
1018                         UTF16_APPEND_CHAR_UNSAFE(scriptelem[count].ch,
1019                                                      scriptelem[count].count, codepoint);
1020                         scriptelem[count].tailored = FALSE;
1021                 }
1022
1023                 count++;
1024         }
1025
1026         return count;
1027 }
1028
1029 /**
1030 * Gets the script elements and contractions belonging to the script
1031 * @param script list
1032 * @param scriptcount number of scripts
1033 * @param elems output list
1034 * @return number of script elements
1035 */
1036 int getScriptElements(UScriptCode script[], int scriptcount,
1037                       ScriptElement scriptelem[])
1038 {
1039     UErrorCode error = U_ZERO_ERROR;
1040     UChar32    codepoint = 0;
1041     int        count     = 0;
1042     while (codepoint <= UCHAR_MAX_VALUE) {
1043         if (checkInScripts(script, scriptcount, codepoint)) {
1044             scriptelem[count].count = 0;
1045             UTF16_APPEND_CHAR_UNSAFE(scriptelem[count].ch,
1046                                      scriptelem[count].count, codepoint);
1047             scriptelem[count].tailored = FALSE;
1048             count ++;
1049         }
1050         if (U_FAILURE(error)) {
1051             fprintf(stdout, "Error determining codepoint in script\n");
1052             return -1;
1053         }
1054         codepoint ++;
1055     }
1056
1057     const UChar           *current  = NULL;
1058           uint32_t         strength = 0;
1059           uint32_t         chOffset = 0;
1060           uint32_t         chLen    = 0;
1061           uint32_t         exOffset = 0;
1062           uint32_t         exLen    = 0;
1063           uint32_t         prefixOffset = 0;
1064           uint32_t         prefixLen    = 0;
1065           uint8_t          specs    = 0;
1066           UBool            rstart   = TRUE;
1067           UColTokenParser  src;
1068           UColOptionSet    opts;
1069           UParseError      parseError;
1070
1071     int32_t  rulelength = ucol_getRulesEx(COLLATOR_, UCOL_FULL_RULES, NULL, 0);
1072     src.source       = (UChar *)malloc(sizeof(UChar) *
1073                                 (rulelength + UCOL_TOK_EXTRA_RULE_SPACE_SIZE));
1074     rulelength = ucol_getRulesEx(COLLATOR_, UCOL_FULL_RULES, src.source,
1075                                  rulelength);
1076     src.current      = src.source;
1077     src.end          = src.source + rulelength;
1078     src.extraCurrent = src.end;
1079     src.extraEnd     = src.end + UCOL_TOK_EXTRA_RULE_SPACE_SIZE;
1080     src.opts         = &opts;
1081
1082         /*
1083         ucol_tok_parseNextToken(&src, &strength, &chOffset,
1084                                               &chLen, &exOffset, &exLen,
1085                                               &prefixOffset, &prefixLen,
1086                                               &specs, rstart, &parseError,
1087                                               &error)
1088     */
1089     while ((current = ucol_tok_parseNextToken(&src, rstart, &parseError,
1090                                               &error)) != NULL) {
1091         // contractions handled here
1092         if (chLen > 1) {
1093             u_strncpy(scriptelem[count].ch, src.source + chOffset, chLen);
1094             scriptelem[count].count = chLen;
1095             if (checkInScripts(script, scriptcount, scriptelem[count])) {
1096                 scriptelem[count].tailored     = FALSE;
1097                 count ++;
1098             }
1099         }
1100         rstart = FALSE;
1101     }
1102     if (U_FAILURE(error)) {
1103         fprintf(stdout, "Error parsing rules: %s\n", u_errorName(error));
1104     }
1105         // rule might have been reallocated, so delete this instead
1106     free(src.source);
1107     return count;
1108 }
1109
1110 int compareCodepoints(const void *elem1, const void *elem2)
1111 {
1112     UChar *ch1 = ((ScriptElement *)elem1)->ch; // key
1113     UChar *ch2 = ((ScriptElement *)elem2)->ch;
1114     ch1[((ScriptElement *)elem1)->count] = 0;
1115     ch2[((ScriptElement *)elem2)->count] = 0;
1116
1117     // compare the 2 codepoints
1118     return u_strcmp(ch1, ch2);
1119 }
1120
1121 UBool hasSubNFD(ScriptElement &se, ScriptElement &key)
1122 {
1123     UChar *ch1 = se.ch;
1124     UChar *ch2 = key.ch; // key
1125     ch1[se.count] = 0;
1126     ch2[key.count] = 0;
1127
1128     // compare the 2 codepoints
1129     if (u_strstr(ch1, ch2) != NULL) {
1130         return TRUE;
1131     }
1132
1133     // check the decomposition
1134     UChar      norm[32];
1135     UErrorCode error = U_ZERO_ERROR;
1136     int        size  = unorm_normalize(ch1, se.count, UNORM_NFD, 0, norm, 32,
1137                                        &error);
1138     if (U_FAILURE(error)) {
1139         fprintf(stdout, "Error normalizing\n");
1140     }
1141     if (u_strstr(norm, ch2) != NULL) {
1142         return TRUE;
1143     }
1144     return FALSE;
1145 }
1146
1147 /**
1148 * Marks tailored elements
1149 * @param script list
1150 * @param scriptcount number of scripts
1151 * @param scriptelem script element list
1152 * @param scriptelemlength size of the script element list
1153 */
1154 void markTailored(UScriptCode script[], int scriptcount,
1155                   ScriptElement scriptelem[], int scriptelemlength)
1156 {
1157           int32_t  rulelength;
1158     const UChar   *rule = ucol_getRules(COLLATOR_, &rulelength);
1159
1160     const UChar           *current  = NULL;
1161           uint32_t         strength = 0;
1162           uint32_t         chOffset = 0;
1163           uint32_t         chLen    = 0;
1164           uint32_t         exOffset = 0;
1165           uint32_t         exLen    = 0;
1166           uint32_t         prefixOffset = 0;
1167           uint32_t         prefixLen    = 0;
1168           uint8_t          specs    = 0;
1169           UBool            rstart   = TRUE;
1170           UColTokenParser  src;
1171           UColOptionSet    opts;
1172           UParseError      parseError;
1173
1174     src.opts         = &opts;
1175     src.source       = (UChar *)malloc(
1176                (rulelength + UCOL_TOK_EXTRA_RULE_SPACE_SIZE) * sizeof(UChar));
1177     memcpy(src.source, rule, rulelength * sizeof(UChar));
1178         src.current      = src.source;
1179     src.end          = (UChar *)src.source + rulelength;
1180     src.extraCurrent = src.end;
1181     src.extraEnd     = src.end + UCOL_TOK_EXTRA_RULE_SPACE_SIZE;
1182
1183     UErrorCode    error = U_ZERO_ERROR;
1184
1185     while ((current = ucol_tok_parseNextToken(&src, rstart, &parseError,
1186                                               &error)) != NULL) {
1187         if (chLen >= 1 && strength != UCOL_TOK_RESET) {
1188             // skipping the reset characters and non useful stuff.
1189             ScriptElement se;
1190             u_strncpy(se.ch, src.source + chOffset, chLen);
1191             se.count = chLen;
1192
1193             if (checkInScripts(script, scriptcount, se)) {
1194                 /*
1195                 ScriptElement *tse = (ScriptElement *)bsearch(&se, scriptelem,
1196                                                               scriptelemlength,
1197                                                          sizeof(ScriptElement),
1198                                                          compareCodepoints);
1199                 */
1200                 for (int i = 0; i < scriptelemlength; i ++) {
1201                     if (!scriptelem[i].tailored &&
1202                         hasSubNFD(scriptelem[i], se)) {
1203                         scriptelem[i].tailored = TRUE;
1204                     }
1205                 }
1206             }
1207         }
1208         rstart = FALSE;
1209     }
1210     free(src.source);
1211     if (U_FAILURE(error)) {
1212         fprintf(stdout, "Error parsing rules\n");
1213     }
1214 }
1215
1216 /**
1217 * Checks if the collation iterator has more than 1 collation element
1218 * @parem coleiter collation element iterator
1219 * @return TRUE if collation iterator has more than 1 collation element
1220 */
1221 UBool hasExpansions(UCollationElements *coleiter)
1222 {
1223     UErrorCode error = U_ZERO_ERROR;
1224     int32_t    ce    = ucol_next(coleiter, &error);
1225     int        count = 0;
1226
1227     if (U_FAILURE(error)) {
1228         fprintf(stdout, "Error getting next collation element\n");
1229     }
1230     while (ce != UCOL_NULLORDER) {
1231         if ((UCOL_PRIMARYORDER(ce) != 0) && !isContinuation(ce)) {
1232             count ++;
1233             if (count == 2) {
1234                 return TRUE;
1235             }
1236         }
1237         ce = ucol_next(coleiter, &error);
1238         if (U_FAILURE(error)) {
1239             fprintf(stdout, "Error getting next collation element\n");
1240         }
1241     }
1242     return FALSE;
1243 }
1244
1245 /**
1246 * Prints the footer for index.html
1247 * @param file output file
1248 */
1249 void outputHTMLFooter()
1250 {
1251     fprintf(OUTPUT_, "</table>\n");
1252     fprintf(OUTPUT_, "</body>\n");
1253     fprintf(OUTPUT_, "</html>\n");
1254 }
1255
1256 /**
1257 * Serialize the codepoints from start to end into an html file.
1258 * Arranging them into ascending collation order.
1259 * @param script code list
1260 * @param scriptcount number of scripts
1261 */
1262 //void serializeScripts(UScriptCode script[], int scriptcount)
1263 //Richard
1264 void serializeScripts(UScriptCode script[], int scriptcount, const char* locale = NULL)
1265 {
1266     UErrorCode  error  = U_ZERO_ERROR;
1267
1268     ScriptElement *scriptelem =
1269                      (ScriptElement *)malloc(sizeof(ScriptElement) * 0x20000);
1270     if (scriptelem == NULL) {
1271         fprintf(stdout, "Memory error\n");
1272         return;
1273     }
1274     int count = 0;
1275     if(locale) {
1276       count = getScriptElementsFromExemplars(scriptelem, locale);
1277     } else {
1278       count = getScriptElements(script, scriptcount, scriptelem);
1279     }
1280
1281     // Sort script elements using Quicksort algorithm:
1282     qsort(scriptelem, count, sizeof(ScriptElement), compareCodepoints);
1283     markTailored(script, scriptcount, scriptelem, count);
1284     // Sort script elements using Quicksort algorithm:
1285     qsort(scriptelem, count, sizeof(ScriptElement), compareSortKey);
1286
1287     UCollationElements* coleiter = ucol_openElements(COLLATOR_,
1288                                                      scriptelem[0].ch,
1289                                                      scriptelem[0].count,
1290                                                      &error);
1291     if (U_FAILURE(error)) {
1292         fprintf(stdout, "Error creating collation element iterator\n");
1293         return;
1294     }
1295
1296     outputScriptElem(scriptelem[0], -1, hasExpansions(coleiter));
1297     for (int i = 0; i < count - 1; i ++) {
1298         ucol_setText(coleiter, scriptelem[i + 1].ch, scriptelem[i + 1].count,
1299                      &error);
1300         if (U_FAILURE(error)) {
1301             fprintf(stdout, "Error setting text in collation element iterator\n");
1302             return;
1303         }
1304         outputScriptElem(scriptelem[i + 1],
1305                          compareSortKey(scriptelem + i, scriptelem + i + 1),
1306                          hasExpansions(coleiter));
1307     }
1308     free(scriptelem);
1309     outputHTMLFooter();
1310 }
1311
1312 /**
1313 * Prints the header for the html
1314 * @param locale name
1315 * @param script
1316 * @param scriptcount number of scripts
1317 */
1318 void outputHTMLHeader(const char *locale, UScriptCode script[],
1319                       int scriptcount)
1320 {
1321     fprintf(OUTPUT_, "<html>\n");
1322     fprintf(OUTPUT_, "<head>\n");
1323     fprintf(OUTPUT_, "<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\">\n");
1324     fprintf(OUTPUT_, "<meta http-equiv=\"Content-Language\" content=\"en-us\">\n");
1325     fprintf(OUTPUT_, "<link rel=\"stylesheet\" href=\"charts.css\" type=\"text/css\">\n");
1326     fprintf(OUTPUT_, "<title>ICU Collation charts</title>\n");
1327     fprintf(OUTPUT_, "<base target=\"main\">\n");
1328     fprintf(OUTPUT_, "</head>\n");
1329
1330     fprintf(OUTPUT_, "<body bgcolor=#FFFFFF>\n");
1331     fprintf(OUTPUT_, "<!--\n");
1332     fprintf(OUTPUT_, "This file contains sorted characters in ascending order according to the locale stated\n");
1333     fprintf(OUTPUT_, "If the character is in red, it is tailored in the collation rules.\n");
1334     fprintf(OUTPUT_, "Background colours have certain meanings:\n");
1335     fprintf(OUTPUT_, "White - equals the previous character\n");
1336     fprintf(OUTPUT_, "dark blue - primary greater than the previous character\n");
1337     fprintf(OUTPUT_, "blue - secondary greater than the previous character\n");
1338     fprintf(OUTPUT_, "light blue - tertiary greater than the previous character\n");
1339     fprintf(OUTPUT_, "--!>\n");
1340
1341     fprintf(OUTPUT_, "<table border=0>\n");
1342     UChar      displayname[64];
1343     UErrorCode error = U_ZERO_ERROR;
1344     int32_t size = uloc_getDisplayName(locale, "en_US", displayname, 64, &error);
1345     char       utf8displayname[128];
1346     if (U_FAILURE(error)) {
1347         utf8displayname[0] = 0;
1348     }
1349     else {
1350         int32_t utf8size = 0;
1351         u_strToUTF8(utf8displayname, 128, &utf8size, displayname, size, &error);
1352     }
1353
1354     fprintf(OUTPUT_, "<tr><th>Locale</th><td class='noborder'>%s</td></tr>\n", utf8displayname);
1355     fprintf(OUTPUT_, "<tr><th>Script(s)</th>");
1356     fprintf(OUTPUT_, "<td class='noborder'>");
1357     for (int i = 0; i < scriptcount; i ++) {
1358         fprintf(OUTPUT_, "%s", uscript_getName(script[i]));
1359         if (i + 1 != scriptcount) {
1360             fprintf(OUTPUT_, ", ");
1361         }
1362     }
1363     fprintf(OUTPUT_, "</td></tr>\n");
1364
1365     fprintf(OUTPUT_, "<tr><th>Rules</th><td class='noborder'><a href=http://oss.software.ibm.com/cvs/icu/~checkout~/icu/source/data/locales/%s.txt>%s.txt</a></td></tr>\n", locale, locale);
1366
1367     UVersionInfo version;
1368     ucol_getVersion(COLLATOR_, version);
1369     fprintf(OUTPUT_, "<tr><th>Collator version</th><td class='noborder'>%d.%d.%d.%d</td></tr>\n",
1370                       version[0], version[1], version[2], version[3]);
1371
1372     UColAttribute attr = UCOL_FRENCH_COLLATION;
1373     while (attr < UCOL_ATTRIBUTE_COUNT) {
1374         UColAttributeValue value = ucol_getAttribute(COLLATOR_, attr, &error);
1375         if (U_FAILURE(error)) {
1376             fprintf(stdout, "Error getting attribute\n");
1377             return;
1378         }
1379         if (value != UCOL_DEFAULT) {
1380             if (attr == UCOL_FRENCH_COLLATION && value != UCOL_OFF) {
1381                 fprintf(OUTPUT_, "<tr><th>French Collation</th><td class='noborder'>on, code %d</td></tr>\n", value);
1382             }
1383             if (attr == UCOL_ALTERNATE_HANDLING && value != UCOL_NON_IGNORABLE) {
1384                 fprintf(OUTPUT_, "<tr><th>Alternate Handling</th><td class='noborder'>shifted, code%d</td></tr>\n", value);
1385             }
1386             if (attr == UCOL_CASE_FIRST && value != UCOL_OFF) {
1387                 fprintf(OUTPUT_, "<tr><th>Case First</th><td class='noborder'>on, code %d</td></tr>\n", value);
1388             }
1389             if (attr == UCOL_CASE_LEVEL && value != UCOL_OFF) {
1390                 fprintf(OUTPUT_, "<tr><th>Case Level</th><td class='noborder'>on, code %d</td></tr>\n", value);
1391             }
1392             if (attr == UCOL_NORMALIZATION_MODE && value != UCOL_OFF) {
1393                 fprintf(OUTPUT_, "<tr><th>Normalization</th><td class='noborder'>on, code %d</td></tr>\n", value);
1394             }
1395             if (attr == UCOL_STRENGTH && value != UCOL_TERTIARY) {
1396                 fprintf(OUTPUT_, "<tr><th>Strength</th><td class='noborder'>code %d</td></tr>\n", value);
1397             }
1398             if (attr == UCOL_HIRAGANA_QUATERNARY_MODE && value != UCOL_OFF) {
1399                 fprintf(OUTPUT_, "<tr><th>Hiragana Quaternary</th><td class='noborder'>on, code %d</td></tr>\n", value);
1400             }
1401         }
1402         attr = (UColAttribute)(attr + 1);
1403     }
1404
1405     // Get UNIX-style time and display as number and string.
1406     time_t ltime;
1407     time( &ltime );
1408     fprintf(OUTPUT_, "<tr><th>Date Generated</th><td class='noborder'>%s</td></tr>", ctime(&ltime));
1409
1410     fprintf(OUTPUT_, "</table>\n");
1411
1412     fprintf(OUTPUT_, "<p><a href=help.html>How to read the table</a><br>\n");
1413     fprintf(OUTPUT_, "<a href=http://www.jtcsv.com/cgi-bin/icu-bugs/ target=new>Submit a bug</a></p>\n");
1414     fprintf(OUTPUT_, "\n<table>\n");
1415     fprintf(OUTPUT_, "\n<tr><th>Codepoint</th><th>P</th><th>S</th><th>T</th><th>Q</th><th>Name</th></tr>\n");
1416 }
1417
1418 /**
1419 * Prints the header for index.html
1420 * @param file output file
1421 */
1422 void outputListHTMLHeader(FILE *file)
1423 {
1424     fprintf(file, "<html>\n");
1425     fprintf(file, "<head>\n");
1426     fprintf(file, "<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\">\n");
1427     fprintf(file, "<meta http-equiv=\"Content-Language\" content=\"en-us\">\n");
1428     fprintf(file, "<title>ICU Collation Charts</title>\n");
1429     fprintf(file, "<base target=\"main\">\n");
1430     fprintf(file, "</head>\n");
1431     fprintf(file, "<body bgcolor=#FFFFFF>\n");
1432     fprintf(file, "<h2 align=center>ICU Collation Charts</h2>\n");
1433     fprintf(file, "<p align=center>\n");
1434     fprintf(file, "<a href=http://www.unicode.org/charts/collation/ target=new>UCA Charts</a><br>");
1435 }
1436
1437 /**
1438 * Prints the footer for index.html
1439 * @param file output file
1440 */
1441 void outputListHTMLFooter(FILE *file)
1442 {
1443     fprintf(file, "</p>\n");
1444         fprintf(file, "<center><image src=http://oss.software.ibm.com/icu/images/w24.gif></center>\n");
1445     fprintf(file, "</body>\n");
1446     fprintf(file, "</html>\n");
1447 }
1448
1449 /**
1450 * Gets all scripts and serialize their codepoints into an html file.
1451 */
1452 void serializeScripts() {
1453     char filename[128];
1454     int  dirlength = 0;
1455
1456     if (options[4].doesOccur) {
1457         strcpy(filename, options[4].value);
1458         dirlength = appendDirSeparator(filename);
1459     } else {
1460       filename[0] = 0;
1461     }
1462
1463     const char    *locale;
1464           int32_t  localelist = 0;
1465           int32_t  localesize;
1466
1467     localesize = ucol_countAvailable();
1468     locale     = ucol_getAvailable(localelist);
1469
1470     strcat(filename, "list.html");
1471     FILE *list = fopen(filename, "w");
1472     filename[dirlength] = 0;
1473     if (list == NULL) {
1474         fprintf(stdout, "Cannot open file: %s\n", filename);
1475         return;
1476     }
1477
1478     outputListHTMLHeader(list);
1479     fprintf(list, "<blockquote>\n");
1480     while (TRUE) {
1481         UErrorCode error = U_ZERO_ERROR;
1482         COLLATOR_ = ucol_open(locale, &error);
1483         if (U_FAILURE(error)) {
1484             fprintf(stdout, "Collator creation failed:");
1485             fprintf(stdout, u_errorName(error));
1486             return;
1487         }
1488         if ((error != U_USING_FALLBACK_WARNING && // not tailored
1489             error != U_USING_DEFAULT_WARNING) ||
1490             checkLocaleForLanguage(locale)) {
1491             fprintf(list, "<a href=%s.html>%s</a> ", locale, locale);
1492                 setAttributes(COLLATOR_, &error);
1493             if (U_FAILURE(error)) {
1494                fprintf(stdout, "Collator attribute setting failed:");
1495                fprintf(stdout, u_errorName(error));
1496                return;
1497             }
1498
1499             UScriptCode scriptcode[32];
1500             uint32_t scriptcount = uscript_getCode(locale, scriptcode, 32,
1501                                                    &error);
1502             if (U_FAILURE(error)) {
1503                 fprintf(stdout, "Error getting lcale scripts\n");
1504                 return;
1505             }
1506
1507             strcat(filename, locale);
1508             strcat(filename, ".html");
1509             OUTPUT_ = fopen(filename, "w");
1510             if (OUTPUT_ == NULL) {
1511                 fprintf(stdout, "Cannot open file:%s\n", filename);
1512                 return;
1513             }
1514             outputHTMLHeader(locale, scriptcode, scriptcount);
1515                         fprintf(stdout, "%s\n", locale);
1516
1517             if(options[12].doesOccur) {
1518               // use whole scripts
1519                           serializeScripts(scriptcode, scriptcount);
1520             } else {
1521               // use exemplar chars
1522               serializeScripts(scriptcode, scriptcount, locale);
1523             }
1524             fclose(OUTPUT_);
1525         }
1526         ucol_close(COLLATOR_);
1527
1528         filename[dirlength] = 0;
1529         localelist ++;
1530         if (localelist == localesize) {
1531             break;
1532         }
1533         locale = ucol_getAvailable(localelist);
1534     }
1535     fprintf(list, "<br><a href=help.html>help</a><br>");
1536     fprintf(list, "</blockquote>\n");
1537     outputListHTMLFooter(list);
1538     fclose(list);
1539 }
1540
1541 /**
1542 * Main   --  process command line, read in and pre-process the test file,
1543 *            call other functions to do the actual tests.
1544 */
1545 int main(int argc, char *argv[]) {
1546
1547     argc = u_parseArgs(argc, argv, sizeof(options)/sizeof(options[0]),
1548                        options);
1549
1550     // error handling, printing usage message
1551     if (argc < 0) {
1552         fprintf(stdout, "error in command line argument: ");
1553         fprintf(stdout, argv[-argc]);
1554         fprintf(stdout, "\n");
1555     }
1556     if (argc < 0 || options[0].doesOccur || options[1].doesOccur) {
1557         fprintf(stdout, "Usage: dumpce options...\n"
1558                         "--help\n"
1559                         "    Display this message.\n"
1560                         "--locale name|all\n"
1561                         "    ICU locale to use. Default is en_US\n"
1562                         "--serialize\n"
1563                         "    Serializes the collation elements in -locale or all locales available and outputs them into --outputdir/locale_ce.txt\n"
1564                         "--destdir dir_name\n"
1565                         "    Path for outputing the serialized collation elements. Defaults to stdout if no defined\n"
1566                         "--sourcedir dir_name\n"
1567                         "    Path for the input rule file for collation\n"
1568                         "--attribute name=value,name=value...\n"
1569                         "    Pairs of attribute names and values for setting\n"
1570                         "--rule filename\n"
1571                         "    Name of file containing the collation rules.\n"
1572                         "--normalizaton mode\n"
1573                         "    UNormalizationMode mode to be used.\n"
1574                         "--scripts\n"
1575                         "    Codepoints from all scripts are sorted and serialized.\n"
1576                         "--reducehan\n"
1577                         "    Only 200 Han script characters will be displayed with the use of --scripts.\n"
1578                         "--wholescripts\n"
1579                         "    Show collation order for whole scripts instead of just for exemplar characters of a locale\n\n");
1580
1581         fprintf(stdout, "Example to generate *.txt files : dumpce --serialize --locale af --destdir /temp --attribute UCOL_STRENGTH=UCOL_DEFAULT_STRENGTH,4=17\n\n");
1582         fprintf(stdout, "Example to generate *.html files for oss web display: dumpce --scripts --destdir /temp --reducehan\n");
1583         return argc < 0 ? U_ILLEGAL_ARGUMENT_ERROR : U_ZERO_ERROR;
1584     }
1585
1586     OUTPUT_ = stdout;
1587     if (options[6].doesOccur) {
1588         fprintf(stdout, "attributes %s\n", options[6].value);
1589         parseAttributes();
1590     }
1591     if (options[3].doesOccur) {
1592         serialize();
1593     }
1594     if (options[9].doesOccur) {
1595         serializeScripts();
1596     }
1597     return 0;
1598 }