CFStringEncodingDatabase.c

   1 /*
   2  * Copyright (c) 2012 Apple Inc. All rights reserved.
   3  *
   4  * @APPLE_LICENSE_HEADER_START@
   5  *
   6  * This file contains Original Code and/or Modifications of Original Code
   7  * as defined in and that are subject to the Apple Public Source License
   8  * Version 2.0 (the 'License'). You may not use this file except in
   9  * compliance with the License. Please obtain a copy of the License at
  10  * http://www.opensource.apple.com/apsl/ and read it before using this
  11  * file.
  12  *
  13  * The Original Code and all software distributed under the License are
  14  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  15  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  16  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  17  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  18  * Please see the License for the specific language governing rights and
  19  * limitations under the License.
  20  *
  21  * @APPLE_LICENSE_HEADER_END@
  22  */
  23
  24 /*      CFStringEncodingDatabase.c
  25         Copyright (c) 2005-2012, Apple Inc. All rights reserved.
  26         Responsibility: Aki Inoue
  27 */
  28
  29 #include "CFInternal.h"
  30 #include <CoreFoundation/CFStringEncodingExt.h>
  31 #include "CFStringEncodingConverterPriv.h"
  32 #include "CFStringEncodingDatabase.h"
  33 #include <stdio.h>
  34
  35 #define ISO8859CODEPAGE_BASE (28590)
  36
  37 static const uint16_t __CFKnownEncodingList[] = {
  38     kCFStringEncodingMacRoman,
  39     kCFStringEncodingMacJapanese,
  40     kCFStringEncodingMacChineseTrad,
  41     kCFStringEncodingMacKorean,
  42     kCFStringEncodingMacArabic,
  43     kCFStringEncodingMacHebrew,
  44     kCFStringEncodingMacGreek,
  45     kCFStringEncodingMacCyrillic,
  46     kCFStringEncodingMacDevanagari,
  47     kCFStringEncodingMacGurmukhi,
  48     kCFStringEncodingMacGujarati,
  49     kCFStringEncodingMacOriya,
  50     kCFStringEncodingMacBengali,
  51     kCFStringEncodingMacTamil,
  52     kCFStringEncodingMacTelugu,
  53     kCFStringEncodingMacKannada,
  54     kCFStringEncodingMacMalayalam,
  55     kCFStringEncodingMacSinhalese,
  56     kCFStringEncodingMacBurmese,
  57     kCFStringEncodingMacKhmer,
  58     kCFStringEncodingMacThai,
  59     kCFStringEncodingMacLaotian,
  60     kCFStringEncodingMacGeorgian,
  61     kCFStringEncodingMacArmenian,
  62     kCFStringEncodingMacChineseSimp,
  63     kCFStringEncodingMacTibetan,
  64     kCFStringEncodingMacMongolian,
  65     kCFStringEncodingMacEthiopic,
  66     kCFStringEncodingMacCentralEurRoman,
  67     kCFStringEncodingMacVietnamese,
  68     kCFStringEncodingMacSymbol,
  69     kCFStringEncodingMacDingbats,
  70     kCFStringEncodingMacTurkish,
  71     kCFStringEncodingMacCroatian,
  72     kCFStringEncodingMacIcelandic,
  73     kCFStringEncodingMacRomanian,
  74     kCFStringEncodingMacCeltic,
  75     kCFStringEncodingMacGaelic,
  76     kCFStringEncodingMacFarsi,
  77     kCFStringEncodingMacUkrainian,
  78     kCFStringEncodingMacInuit,
  79
  80     kCFStringEncodingDOSLatinUS,
  81     kCFStringEncodingDOSGreek,
  82     kCFStringEncodingDOSBalticRim,
  83     kCFStringEncodingDOSLatin1,
  84     kCFStringEncodingDOSGreek1,
  85     kCFStringEncodingDOSLatin2,
  86     kCFStringEncodingDOSCyrillic,
  87     kCFStringEncodingDOSTurkish,
  88     kCFStringEncodingDOSPortuguese,
  89     kCFStringEncodingDOSIcelandic,
  90     kCFStringEncodingDOSHebrew,
  91     kCFStringEncodingDOSCanadianFrench,
  92     kCFStringEncodingDOSArabic,
  93     kCFStringEncodingDOSNordic,
  94     kCFStringEncodingDOSRussian,
  95     kCFStringEncodingDOSGreek2,
  96     kCFStringEncodingDOSThai,
  97     kCFStringEncodingDOSJapanese,
  98     kCFStringEncodingDOSChineseSimplif,
  99     kCFStringEncodingDOSKorean,
 100     kCFStringEncodingDOSChineseTrad,
 101
 102     kCFStringEncodingWindowsLatin1,
 103     kCFStringEncodingWindowsLatin2,
 104     kCFStringEncodingWindowsCyrillic,
 105     kCFStringEncodingWindowsGreek,
 106     kCFStringEncodingWindowsLatin5,
 107     kCFStringEncodingWindowsHebrew,
 108     kCFStringEncodingWindowsArabic,
 109     kCFStringEncodingWindowsBalticRim,
 110     kCFStringEncodingWindowsVietnamese,
 111     kCFStringEncodingWindowsKoreanJohab,
 112     kCFStringEncodingASCII,
 113
 114     kCFStringEncodingShiftJIS_X0213,
 115     kCFStringEncodingGB_18030_2000,
 116
 117     kCFStringEncodingISO_2022_JP,
 118     kCFStringEncodingISO_2022_JP_2,
 119     kCFStringEncodingISO_2022_JP_1,
 120     kCFStringEncodingISO_2022_JP_3,
 121     kCFStringEncodingISO_2022_CN,
 122     kCFStringEncodingISO_2022_CN_EXT,
 123     kCFStringEncodingISO_2022_KR,
 124     kCFStringEncodingEUC_JP,
 125     kCFStringEncodingEUC_CN,
 126     kCFStringEncodingEUC_TW,
 127     kCFStringEncodingEUC_KR,
 128
 129     kCFStringEncodingShiftJIS,
 130
 131     kCFStringEncodingKOI8_R,
 132
 133     kCFStringEncodingBig5,
 134
 135     kCFStringEncodingMacRomanLatin1,
 136     kCFStringEncodingHZ_GB_2312,
 137     kCFStringEncodingBig5_HKSCS_1999,
 138     kCFStringEncodingVISCII,
 139     kCFStringEncodingKOI8_U,
 140     kCFStringEncodingBig5_E,
 141     kCFStringEncodingUTF7_IMAP,
 142
 143     kCFStringEncodingNextStepLatin,
 144
 145     kCFStringEncodingEBCDIC_CP037
 146 };
 147
 148 // Windows codepage mapping
 149 static const uint16_t __CFWindowsCPList[] = {
 150     10000,
 151     10001,
 152     10002,
 153     10003,
 154     10004,
 155     10005,
 156     10006,
 157     10007,
 158     0,
 159     0,
 160     0,
 161     0,
 162     0,
 163     0,
 164     0,
 165     0,
 166     0,
 167     0,
 168     0,
 169     0,
 170     10021,
 171     0,
 172     0,
 173     0,
 174     10008,
 175     0,
 176     0,
 177     0,
 178     10029,
 179     0,
 180     0,
 181     0,
 182     10081,
 183     10082,
 184     10079,
 185     10010,
 186     0,
 187     0,
 188     0,
 189     10017,
 190     0,
 191
 192     437,
 193     737,
 194     775,
 195     850,
 196     851,
 197     852,
 198     855,
 199     857,
 200     860,
 201     861,
 202     862,
 203     863,
 204     864,
 205     865,
 206     866,
 207     869,
 208     874,
 209     932,
 210     936,
 211     949,
 212     950,
 213
 214     1252,
 215     1250,
 216     1251,
 217     1253,
 218     1254,
 219     1255,
 220     1256,
 221     1257,
 222     1258,
 223     1361,
 224
 225     20127,
 226
 227     0,
 228     54936,
 229
 230     50221, // we prefere this over 50220/50221 since that's what CF coverter generates
 231     0,
 232     0,
 233     0,
 234     50227,
 235     0,
 236     50225,
 237
 238     51932,
 239     51936,
 240     51950,
 241     51949,
 242
 243     0,
 244
 245     20866,
 246
 247     0,
 248
 249     0,
 250     52936,
 251     0,
 252     0,
 253     21866,
 254     0,
 255     0,
 256
 257     0,
 258
 259     37
 260 };
 261
 262 // Canonical name
 263 static const char *__CFCanonicalNameList[] = {
 264     "macintosh",
 265     "japanese",
 266     "trad-chinese",
 267     "korean",
 268     "arabic",
 269     "hebrew",
 270     "greek",
 271     "cyrillic",
 272     "devanagari",
 273     "gurmukhi",
 274     "gujarati",
 275     "oriya",
 276     "bengali",
 277     "tamil",
 278     "telugu",
 279     "kannada",
 280     "malayalam",
 281     "sinhalese",
 282     "burmese",
 283     "khmer",
 284     "thai",
 285     "laotian",
 286     "georgian",
 287     "armenian",
 288     "simp-chinese",
 289     "tibetan",
 290     "mongolian",
 291     "ethiopic",
 292     "centraleurroman",
 293     "vietnamese",
 294     "symbol",
 295     "dingbats",
 296     "turkish",
 297     "croatian",
 298     "icelandic",
 299     "romanian",
 300     "celtic",
 301     "gaelic",
 302     "farsi",
 303     "ukrainian",
 304     "inuit",
 305
 306     NULL,
 307     NULL,
 308     NULL,
 309     NULL,
 310     NULL,
 311     NULL,
 312     NULL,
 313     NULL,
 314     NULL,
 315     NULL,
 316     NULL,
 317     NULL,
 318     NULL,
 319     NULL,
 320     NULL,
 321     NULL,
 322     NULL,
 323     NULL,
 324     NULL,
 325     NULL,
 326     NULL,
 327
 328     NULL,
 329     NULL,
 330     NULL,
 331     NULL,
 332     NULL,
 333     NULL,
 334     NULL,
 335     NULL,
 336     NULL,
 337     NULL,
 338
 339     "us-ascii",
 340
 341     NULL,
 342     "gb18030",
 343
 344     "iso-2022-jp",
 345     "iso-2022-jp-2",
 346     "iso-2022-jp-1",
 347     "iso-2022-jp-3",
 348     "iso-2022-cn",
 349     "iso-2022-cn-ext",
 350     "iso-2022-kr",
 351     "euc-jp",
 352     "gb2312",
 353     "euc-tw",
 354     "euc-kr",
 355
 356     "shift_jis",
 357
 358     "koi8-r",
 359
 360     "big5",
 361
 362     "roman-latin1",
 363     "hz-gb-2312",
 364     "big5-hkscs",
 365     "viscii",
 366     "koi8-u",
 367     NULL,
 368     "utf7-imap",
 369
 370     "x-nextstep",
 371
 372     "ibm037",
 373 };
 374
 375 static inline CFIndex __CFGetEncodingIndex(CFStringEncoding encoding) {
 376     const uint16_t *head = __CFKnownEncodingList;
 377     const uint16_t *tail = head + ((sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList)) - 1);
 378     const uint16_t *middle;
 379
 380     encoding &= 0x0FFF;
 381     while (head <= tail) {
 382         middle = head + ((tail - head) >> 1);
 383
 384         if (encoding == *middle) {
 385             return middle - __CFKnownEncodingList;
 386         } else if (encoding < *middle) {
 387             tail = middle - 1;
 388         } else {
 389             head = middle + 1;
 390         }
 391     }
 392
 393     return kCFNotFound;
 394 }
 395
 396 __private_extern__ uint16_t __CFStringEncodingGetWindowsCodePage(CFStringEncoding encoding) {
 397     CFStringEncoding encodingBase = encoding & 0x0F00;
 398
 399     if (0x0100 == encodingBase) { // UTF
 400         switch (encoding) {
 401             case kCFStringEncodingUTF7: return 65000;
 402             case kCFStringEncodingUTF8: return 65001;
 403             case kCFStringEncodingUTF16: return 1200;
 404             case kCFStringEncodingUTF16BE: return 1201;
 405             case kCFStringEncodingUTF32: return 65005;
 406             case kCFStringEncodingUTF32BE: return 65006;
 407         }
 408     } else if (0x0200 == encodingBase) { // ISO 8859 range
 409         return ISO8859CODEPAGE_BASE + (encoding & 0xFF);
 410     } else { // others
 411         CFIndex index = __CFGetEncodingIndex(encoding);
 412
 413         if (kCFNotFound != index) return __CFWindowsCPList[index];
 414     }
 415
 416     return 0;
 417 }
 418
 419 __private_extern__ CFStringEncoding __CFStringEncodingGetFromWindowsCodePage(uint16_t codepage) {
 420     switch (codepage) {
 421         case 65001: return kCFStringEncodingUTF8;
 422         case 1200: return kCFStringEncodingUTF16;
 423         case 0: return kCFStringEncodingInvalidId;
 424         case 1201: return kCFStringEncodingUTF16BE;
 425         case 65005: return kCFStringEncodingUTF32;
 426         case 65006: return kCFStringEncodingUTF32BE;
 427         case 65000: return kCFStringEncodingUTF7;
 428     }
 429
 430     if ((codepage > ISO8859CODEPAGE_BASE) && (codepage <= (ISO8859CODEPAGE_BASE + 16))) {
 431         return (codepage - ISO8859CODEPAGE_BASE) + 0x0200;
 432     } else {
 433         static CFMutableDictionaryRef mappingTable = NULL;
 434         static CFSpinLock_t lock = CFSpinLockInit;
 435         uintptr_t value;
 436
 437         __CFSpinLock(&lock);
 438         if (NULL == mappingTable) {
 439             CFIndex index, count = sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList);
 440
 441             mappingTable = CFDictionaryCreateMutable(NULL, 0, NULL, NULL);
 442
 443             for (index = 0;index < count;index++) {
 444                 if (0 != __CFWindowsCPList[index]) CFDictionarySetValue(mappingTable, (const void *)(uintptr_t)__CFWindowsCPList[index], (const void *)(uintptr_t)__CFKnownEncodingList[index]);
 445             }
 446         }
 447         __CFSpinUnlock(&lock);
 448
 449         if (CFDictionaryGetValueIfPresent(mappingTable, (const void *)(uintptr_t)codepage, (const void **)&value)) return (CFStringEncoding)value;
 450     }
 451
 452
 453     return kCFStringEncodingInvalidId;
 454 }
 455
 456 __private_extern__ bool __CFStringEncodingGetCanonicalName(CFStringEncoding encoding, char *buffer, CFIndex bufferSize) {
 457     const char *format = "%s";
 458     const char *name = NULL;
 459     uint32_t value = 0;
 460     CFIndex index;
 461
 462     switch (encoding & 0x0F00) {
 463         case 0x0100: // UTF range
 464             switch (encoding) {
 465                 case kCFStringEncodingUTF7: name = "utf-7"; break;
 466                 case kCFStringEncodingUTF8: name = "utf-8"; break;
 467                 case kCFStringEncodingUTF16: name = "utf-16"; break;
 468                 case kCFStringEncodingUTF16BE: name = "utf-16be"; break;
 469                 case kCFStringEncodingUTF16LE: name = "utf-16le"; break;
 470                 case kCFStringEncodingUTF32: name = "utf-32"; break;
 471                 case kCFStringEncodingUTF32BE: name = "utf-32be"; break;
 472                 case kCFStringEncodingUTF32LE: name = "utf-32le"; break;
 473             }
 474             break;
 475
 476         case 0x0200: // ISO 8859 range
 477             format = "iso-8859-%d";
 478             value = (encoding & 0xFF);
 479             break;
 480
 481         case 0x0400: // DOS code page range
 482         case 0x0500: // Windows code page range
 483             index = __CFGetEncodingIndex(encoding);
 484
 485             if (kCFNotFound != index) {
 486                 value = __CFWindowsCPList[index];
 487                 if (0 != value) format = ((0x0400 == (encoding & 0x0F00)) ? "cp%d" : "windows-%d");
 488             }
 489             break;
 490
 491         default: // others
 492             index = __CFGetEncodingIndex(encoding);
 493
 494             if (kCFNotFound != index) {
 495                 if (((0 == (encoding & 0x0F00)) && (kCFStringEncodingMacRoman != encoding)) || (kCFStringEncodingMacRomanLatin1 == encoding)) format = "x-mac-%s";
 496                 name = (const char *)__CFCanonicalNameList[index];
 497             }
 498             break;
 499     }
 500
 501     if ((0 == value) && (NULL == name)) {
 502         return false;
 503     } else if (0 != value) {
 504         return ((snprintf(buffer, bufferSize, format, value) < bufferSize) ? true : false);
 505     } else {
 506         return ((snprintf(buffer, bufferSize, format, name) < bufferSize) ? true : false);
 507     }
 508 }
 509
 510 #define LENGTH_LIMIT (256)
 511 static Boolean __CFCanonicalNameCompare(const void *value1, const void *value2) { return ((0 == strncasecmp_l((const char *)value1, (const char *)value2, LENGTH_LIMIT, NULL)) ? true : false); }
 512
 513 static CFHashCode __CFCanonicalNameHash(const void *value) {
 514     const char *name = (const char *)value;
 515     CFHashCode code = 0;
 516
 517     while ((0 != *name) && ((name - (const char *)value) < LENGTH_LIMIT)) {
 518         char character = *(name++);
 519
 520         code += (character + (((character >= 'A') && (character <= 'Z')) ? 'a' - 'A' : 0));
 521     }
 522
 523     return code * (name - (const char *)value);
 524 }
 525
 526 __private_extern__ CFStringEncoding __CFStringEncodingGetFromCanonicalName(const char *canonicalName) {
 527     CFStringEncoding encoding;
 528     CFIndex prefixLength;
 529     static CFMutableDictionaryRef mappingTable = NULL;
 530     static CFSpinLock_t lock = CFSpinLockInit;
 531
 532     prefixLength = strlen("iso-8859-");
 533     if (0 == strncasecmp_l(canonicalName, "iso-8859-", prefixLength, NULL)) {// do ISO
 534         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 535
 536         return (((0 == encoding) || (encoding > 16)) ? kCFStringEncodingInvalidId : encoding + 0x0200);
 537     }
 538
 539     prefixLength = strlen("cp");
 540     if (0 == strncasecmp_l(canonicalName, "cp", prefixLength, NULL)) {// do DOS
 541         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 542
 543         return __CFStringEncodingGetFromWindowsCodePage(encoding);
 544     }
 545
 546     prefixLength = strlen("windows-");
 547     if (0 == strncasecmp_l(canonicalName, "windows-", prefixLength, NULL)) {// do DOS
 548         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 549
 550         return __CFStringEncodingGetFromWindowsCodePage(encoding);
 551     }
 552
 553     __CFSpinLock(&lock);
 554     if (NULL == mappingTable) {
 555         CFIndex index, count = sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList);
 556
 557         CFDictionaryKeyCallBacks keys = {
 558             0, NULL, NULL, NULL, &__CFCanonicalNameCompare, &__CFCanonicalNameHash
 559         };
 560
 561         mappingTable = CFDictionaryCreateMutable(NULL, 0, &keys, NULL);
 562
 563         // Add UTFs
 564         CFDictionarySetValue(mappingTable, "utf-7", (const void *)kCFStringEncodingUTF7);
 565         CFDictionarySetValue(mappingTable, "utf-8", (const void *)kCFStringEncodingUTF8);
 566         CFDictionarySetValue(mappingTable, "utf-16", (const void *)kCFStringEncodingUTF16);
 567         CFDictionarySetValue(mappingTable, "utf-16be", (const void *)kCFStringEncodingUTF16BE);
 568         CFDictionarySetValue(mappingTable, "utf-16le", (const void *)kCFStringEncodingUTF16LE);
 569         CFDictionarySetValue(mappingTable, "utf-32", (const void *)kCFStringEncodingUTF32);
 570         CFDictionarySetValue(mappingTable, "utf-32be", (const void *)kCFStringEncodingUTF32BE);
 571         CFDictionarySetValue(mappingTable, "utf-32le", (const void *)kCFStringEncodingUTF32LE);
 572
 573         for (index = 0;index < count;index++) {
 574             if (NULL != __CFCanonicalNameList[index]) CFDictionarySetValue(mappingTable, (const void *)(uintptr_t)__CFCanonicalNameList[index], (const void *)(uintptr_t)__CFKnownEncodingList[index]);
 575         }
 576     }
 577     __CFSpinUnlock(&lock);
 578
 579     if (0 == strncasecmp_l(canonicalName, "macintosh", sizeof("macintosh") - 1, NULL)) return kCFStringEncodingMacRoman;
 580
 581
 582     prefixLength = strlen("x-mac-");
 583     encoding = (CFStringEncoding)(CFIndex)CFDictionaryGetValue(mappingTable, canonicalName + ((0 == strncasecmp_l(canonicalName, "x-mac-", prefixLength, NULL)) ? prefixLength : 0));
 584
 585     return ((0 == encoding) ? kCFStringEncodingInvalidId : encoding);
 586 }
 587 #undef LENGTH_LIMIT
 588
 589 #if DEPLOYMENT_TARGET_MACOSX
 590 // This list indexes from DOS range
 591 static uint16_t __CFISO8859SimilarScriptList[] = {
 592     kCFStringEncodingMacRoman,
 593     kCFStringEncodingMacCentralEurRoman,
 594     kCFStringEncodingMacRoman,
 595     kCFStringEncodingMacCentralEurRoman,
 596     kCFStringEncodingMacCyrillic,
 597     kCFStringEncodingMacArabic,
 598     kCFStringEncodingMacGreek,
 599     kCFStringEncodingMacHebrew,
 600     kCFStringEncodingMacTurkish,
 601     kCFStringEncodingMacInuit,
 602     kCFStringEncodingMacThai,
 603     kCFStringEncodingMacRoman,
 604     kCFStringEncodingMacCentralEurRoman,
 605     kCFStringEncodingMacCeltic,
 606     kCFStringEncodingMacRoman,
 607     kCFStringEncodingMacRomanian};
 608
 609 static uint16_t __CFOtherSimilarScriptList[] = {
 610     kCFStringEncodingMacRoman,
 611     kCFStringEncodingMacGreek,
 612     kCFStringEncodingMacCentralEurRoman,
 613     kCFStringEncodingMacRoman,
 614     kCFStringEncodingMacGreek,
 615     kCFStringEncodingMacCentralEurRoman,
 616     kCFStringEncodingMacCyrillic,
 617     kCFStringEncodingMacTurkish,
 618     kCFStringEncodingMacRoman,
 619     kCFStringEncodingMacIcelandic,
 620     kCFStringEncodingMacHebrew,
 621     kCFStringEncodingMacRoman,
 622     kCFStringEncodingMacArabic,
 623     kCFStringEncodingMacInuit,
 624     kCFStringEncodingMacCyrillic,
 625     kCFStringEncodingMacGreek,
 626     kCFStringEncodingMacThai,
 627     kCFStringEncodingMacJapanese,
 628     kCFStringEncodingMacChineseSimp,
 629     kCFStringEncodingMacKorean,
 630     kCFStringEncodingMacChineseTrad,
 631
 632     kCFStringEncodingMacRoman,
 633     kCFStringEncodingMacCentralEurRoman,
 634     kCFStringEncodingMacCyrillic,
 635     kCFStringEncodingMacGreek,
 636     kCFStringEncodingMacTurkish,
 637     kCFStringEncodingMacHebrew,
 638     kCFStringEncodingMacArabic,
 639     kCFStringEncodingMacCentralEurRoman,
 640     kCFStringEncodingMacVietnamese,
 641     kCFStringEncodingMacKorean,
 642
 643     kCFStringEncodingMacRoman,
 644
 645     kCFStringEncodingMacJapanese,
 646     kCFStringEncodingMacChineseSimp,
 647
 648     kCFStringEncodingMacJapanese,
 649     kCFStringEncodingMacJapanese,
 650     kCFStringEncodingMacJapanese,
 651     kCFStringEncodingMacJapanese,
 652     kCFStringEncodingMacChineseSimp,
 653     kCFStringEncodingMacChineseSimp,
 654     kCFStringEncodingMacKorean,
 655     kCFStringEncodingMacJapanese,
 656     kCFStringEncodingMacChineseSimp,
 657     kCFStringEncodingMacChineseTrad,
 658     kCFStringEncodingMacKorean,
 659
 660     kCFStringEncodingMacJapanese,
 661
 662     kCFStringEncodingMacCyrillic,
 663
 664     kCFStringEncodingMacChineseTrad,
 665
 666     kCFStringEncodingMacRoman,
 667     kCFStringEncodingMacChineseSimp,
 668     kCFStringEncodingMacChineseTrad,
 669     kCFStringEncodingMacVietnamese,
 670     kCFStringEncodingMacUkrainian,
 671     kCFStringEncodingMacChineseTrad,
 672     kCFStringEncodingMacRoman,
 673
 674     kCFStringEncodingMacRoman,
 675
 676     kCFStringEncodingMacRoman
 677 };
 678
 679 static const char *__CFISONameList[] = {
 680     "Western (ISO Latin 1)",
 681     "Central European (ISO Latin 2)",
 682     "Western (ISO Latin 3)",
 683     "Central European (ISO Latin 4)",
 684     "Cyrillic (ISO 8859-5)",
 685     "Arabic (ISO 8859-6)",
 686     "Greek (ISO 8859-7)",
 687     "Hebrew (ISO 8859-8)",
 688     "Turkish (ISO Latin 5)",
 689     "Nordic (ISO Latin 6)",
 690     "Thai (ISO 8859-11)",
 691     NULL,
 692     "Baltic (ISO Latin 7)",
 693     "Celtic (ISO Latin 8)",
 694     "Western (ISO Latin 9)",
 695     "Romanian (ISO Latin 10)",
 696 };
 697
 698 static const char *__CFOtherNameList[] = {
 699     "Western (Mac OS Roman)",
 700     "Japanese (Mac OS)",
 701     "Traditional Chinese (Mac OS)",
 702     "Korean (Mac OS)",
 703     "Arabic (Mac OS)",
 704     "Hebrew (Mac OS)",
 705     "Greek (Mac OS)",
 706     "Cyrillic (Mac OS)",
 707     "Devanagari (Mac OS)",
 708     "Gurmukhi (Mac OS)",
 709     "Gujarati (Mac OS)",
 710     "Oriya (Mac OS)",
 711     "Bengali (Mac OS)",
 712     "Tamil (Mac OS)",
 713     "Telugu (Mac OS)",
 714     "Kannada (Mac OS)",
 715     "Malayalam (Mac OS)",
 716     "Sinhalese (Mac OS)",
 717     "Burmese (Mac OS)",
 718     "Khmer (Mac OS)",
 719     "Thai (Mac OS)",
 720     "Laotian (Mac OS)",
 721     "Georgian (Mac OS)",
 722     "Armenian (Mac OS)",
 723     "Simplified Chinese (Mac OS)",
 724     "Tibetan (Mac OS)",
 725     "Mongolian (Mac OS)",
 726     "Ethiopic (Mac OS)",
 727     "Central European (Mac OS)",
 728     "Vietnamese (Mac OS)",
 729     "Symbol (Mac OS)",
 730     "Dingbats (Mac OS)",
 731     "Turkish (Mac OS)",
 732     "Croatian (Mac OS)",
 733     "Icelandic (Mac OS)",
 734     "Romanian (Mac OS)",
 735     "Celtic (Mac OS)",
 736     "Gaelic (Mac OS)",
 737     "Farsi (Mac OS)",
 738     "Cyrillic (Mac OS Ukrainian)",
 739     "Inuit (Mac OS)",
 740     "Latin-US (DOS)",
 741     "Greek (DOS)",
 742     "Baltic (DOS)",
 743     "Western (DOS Latin 1)",
 744     "Greek (DOS Greek 1)",
 745     "Central European (DOS Latin 2)",
 746     "Cyrillic (DOS)",
 747     "Turkish (DOS)",
 748     "Portuguese (DOS)",
 749     "Icelandic (DOS)",
 750     "Hebrew (DOS)",
 751     "Canadian French (DOS)",
 752     "Arabic (DOS)",
 753     "Nordic (DOS)",
 754     "Russian (DOS)",
 755     "Greek (DOS Greek 2)",
 756     "Thai (Windows, DOS)",
 757     "Japanese (Windows, DOS)",
 758     "Simplified Chinese (Windows, DOS)",
 759     "Korean (Windows, DOS)",
 760     "Traditional Chinese (Windows, DOS)",
 761     "Western (Windows Latin 1)",
 762     "Central European (Windows Latin 2)",
 763     "Cyrillic (Windows)",
 764     "Greek (Windows)",
 765     "Turkish (Windows Latin 5)",
 766     "Hebrew (Windows)",
 767     "Arabic (Windows)",
 768     "Baltic (Windows)",
 769     "Vietnamese (Windows)",
 770     "Korean (Windows Johab)",
 771     "Western (ASCII)",
 772     "Japanese (Shift JIS X0213)",
 773     "Chinese (GB 18030)",
 774     "Japanese (ISO 2022-JP)",
 775     "Japanese (ISO 2022-JP-2)",
 776     "Japanese (ISO 2022-JP-1)",
 777     "Japanese (ISO 2022-JP-3)",
 778     "Chinese (ISO 2022-CN)",
 779     "Chinese (ISO 2022-CN-EXT)",
 780     "Korean (ISO 2022-KR)",
 781     "Japanese (EUC)",
 782     "Simplified Chinese (GB 2312)",
 783     "Traditional Chinese (EUC)",
 784     "Korean (EUC)",
 785     "Japanese (Shift JIS)",
 786     "Cyrillic (KOI8-R)",
 787     "Traditional Chinese (Big 5)",
 788     "Western (Mac Mail)",
 789     "Simplified Chinese (HZ GB 2312)",
 790     "Traditional Chinese (Big 5 HKSCS)",
 791     NULL,
 792     "Ukrainian (KOI8-U)",
 793     "Traditional Chinese (Big 5-E)",
 794     NULL,
 795     "Western (NextStep)",
 796     "Western (EBCDIC Latin 1)",
 797 };
 798 #endif /* DEPLOYMENT_TARGET_MACOSX */
 799
 800 __private_extern__ CFStringEncoding __CFStringEncodingGetMostCompatibleMacScript(CFStringEncoding encoding) {
 801 #if DEPLOYMENT_TARGET_MACOSX
 802     switch (encoding & 0x0F00) {
 803         case 0: return encoding & 0xFF; break; // Mac scripts
 804
 805         case 0x0100: return kCFStringEncodingUnicode; break; // Unicode
 806
 807         case 0x200: // ISO 8859
 808             return (((encoding & 0xFF) <= (sizeof(__CFISO8859SimilarScriptList) / sizeof(*__CFISO8859SimilarScriptList))) ? __CFISO8859SimilarScriptList[(encoding & 0xFF) - 1] : kCFStringEncodingInvalidId);
 809             break;
 810
 811         default: {
 812             CFIndex index = __CFGetEncodingIndex(encoding);
 813
 814             if (kCFNotFound != index) {
 815                 index -= __CFGetEncodingIndex(kCFStringEncodingDOSLatinUS);
 816                 return __CFOtherSimilarScriptList[index];
 817             }
 818         }
 819     }
 820 #endif /* DEPLOYMENT_TARGET_MACOSX */
 821
 822     return kCFStringEncodingInvalidId;
 823 }
 824
 825 __private_extern__ const char *__CFStringEncodingGetName(CFStringEncoding encoding) {
 826     switch (encoding) {
 827         case kCFStringEncodingUTF8: return "Unicode (UTF-8)"; break;
 828         case kCFStringEncodingUTF16: return "Unicode (UTF-16)"; break;
 829         case kCFStringEncodingUTF16BE: return "Unicode (UTF-16BE)"; break;
 830         case kCFStringEncodingUTF16LE: return "Unicode (UTF-16LE)"; break;
 831         case kCFStringEncodingUTF32: return "Unicode (UTF-32)"; break;
 832         case kCFStringEncodingUTF32BE: return "Unicode (UTF-32BE)"; break;
 833         case kCFStringEncodingUTF32LE: return "Unicode (UTF-32LE)"; break;
 834         case kCFStringEncodingNonLossyASCII: return "Non-lossy ASCII"; break;
 835         case kCFStringEncodingUTF7: return "Unicode (UTF-7)"; break;
 836     }
 837
 838 #if DEPLOYMENT_TARGET_MACOSX
 839     if (0x0200 == (encoding & 0x0F00)) {
 840         encoding &= 0x00FF;
 841
 842         if (encoding <= (sizeof(__CFISONameList) / sizeof(*__CFISONameList))) return __CFISONameList[encoding - 1];
 843     } else {
 844         CFIndex index = __CFGetEncodingIndex(encoding);
 845
 846         if (kCFNotFound != index) return __CFOtherNameList[index];
 847     }
 848 #endif /* DEPLOYMENT_TARGET_MACOSX */
 849
 850     return NULL;
 851 }