CFStringEncodingDatabase.c

   1 /*
   2  * Copyright (c) 2009 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  *  CoreFoundation
  26  *
  27  *  Created by Aki Inoue on 07/12/05.
  28  *  Copyright 2007-2009, Apple Inc. All rights reserved.
  29  *
  30  */
  31
  32 #include "CFInternal.h"
  33 #include <CoreFoundation/CFStringEncodingExt.h>
  34 #include "CFStringEncodingConverterPriv.h"
  35 #include "CFStringEncodingDatabase.h"
  36 #include <stdio.h>
  37
  38 #if DEPLOYMENT_TARGET_WINDOWS
  39 #define strncasecmp_l(a, b, c, d) _strnicmp(a, b, c)
  40 #define snprintf _snprintf
  41 #endif
  42
  43 #define ISO8859CODEPAGE_BASE (28590)
  44
  45 static const uint16_t __CFKnownEncodingList[] = {
  46     kCFStringEncodingMacRoman,
  47     kCFStringEncodingMacJapanese,
  48     kCFStringEncodingMacChineseTrad,
  49     kCFStringEncodingMacKorean,
  50     kCFStringEncodingMacArabic,
  51     kCFStringEncodingMacHebrew,
  52     kCFStringEncodingMacGreek,
  53     kCFStringEncodingMacCyrillic,
  54     kCFStringEncodingMacDevanagari,
  55     kCFStringEncodingMacGurmukhi,
  56     kCFStringEncodingMacGujarati,
  57     kCFStringEncodingMacOriya,
  58     kCFStringEncodingMacBengali,
  59     kCFStringEncodingMacTamil,
  60     kCFStringEncodingMacTelugu,
  61     kCFStringEncodingMacKannada,
  62     kCFStringEncodingMacMalayalam,
  63     kCFStringEncodingMacSinhalese,
  64     kCFStringEncodingMacBurmese,
  65     kCFStringEncodingMacKhmer,
  66     kCFStringEncodingMacThai,
  67     kCFStringEncodingMacLaotian,
  68     kCFStringEncodingMacGeorgian,
  69     kCFStringEncodingMacArmenian,
  70     kCFStringEncodingMacChineseSimp,
  71     kCFStringEncodingMacTibetan,
  72     kCFStringEncodingMacMongolian,
  73     kCFStringEncodingMacEthiopic,
  74     kCFStringEncodingMacCentralEurRoman,
  75     kCFStringEncodingMacVietnamese,
  76     kCFStringEncodingMacSymbol,
  77     kCFStringEncodingMacDingbats,
  78     kCFStringEncodingMacTurkish,
  79     kCFStringEncodingMacCroatian,
  80     kCFStringEncodingMacIcelandic,
  81     kCFStringEncodingMacRomanian,
  82     kCFStringEncodingMacCeltic,
  83     kCFStringEncodingMacGaelic,
  84     kCFStringEncodingMacFarsi,
  85     kCFStringEncodingMacUkrainian,
  86     kCFStringEncodingMacInuit,
  87
  88     kCFStringEncodingDOSLatinUS,
  89     kCFStringEncodingDOSGreek,
  90     kCFStringEncodingDOSBalticRim,
  91     kCFStringEncodingDOSLatin1,
  92     kCFStringEncodingDOSGreek1,
  93     kCFStringEncodingDOSLatin2,
  94     kCFStringEncodingDOSCyrillic,
  95     kCFStringEncodingDOSTurkish,
  96     kCFStringEncodingDOSPortuguese,
  97     kCFStringEncodingDOSIcelandic,
  98     kCFStringEncodingDOSHebrew,
  99     kCFStringEncodingDOSCanadianFrench,
 100     kCFStringEncodingDOSArabic,
 101     kCFStringEncodingDOSNordic,
 102     kCFStringEncodingDOSRussian,
 103     kCFStringEncodingDOSGreek2,
 104     kCFStringEncodingDOSThai,
 105     kCFStringEncodingDOSJapanese,
 106     kCFStringEncodingDOSChineseSimplif,
 107     kCFStringEncodingDOSKorean,
 108     kCFStringEncodingDOSChineseTrad,
 109
 110     kCFStringEncodingWindowsLatin1,
 111     kCFStringEncodingWindowsLatin2,
 112     kCFStringEncodingWindowsCyrillic,
 113     kCFStringEncodingWindowsGreek,
 114     kCFStringEncodingWindowsLatin5,
 115     kCFStringEncodingWindowsHebrew,
 116     kCFStringEncodingWindowsArabic,
 117     kCFStringEncodingWindowsBalticRim,
 118     kCFStringEncodingWindowsVietnamese,
 119     kCFStringEncodingWindowsKoreanJohab,
 120     kCFStringEncodingASCII,
 121
 122     kCFStringEncodingShiftJIS_X0213,
 123     kCFStringEncodingGB_18030_2000,
 124
 125     kCFStringEncodingISO_2022_JP,
 126     kCFStringEncodingISO_2022_JP_2,
 127     kCFStringEncodingISO_2022_JP_1,
 128     kCFStringEncodingISO_2022_JP_3,
 129     kCFStringEncodingISO_2022_CN,
 130     kCFStringEncodingISO_2022_CN_EXT,
 131     kCFStringEncodingISO_2022_KR,
 132     kCFStringEncodingEUC_JP,
 133     kCFStringEncodingEUC_CN,
 134     kCFStringEncodingEUC_TW,
 135     kCFStringEncodingEUC_KR,
 136
 137     kCFStringEncodingShiftJIS,
 138
 139     kCFStringEncodingKOI8_R,
 140
 141     kCFStringEncodingBig5,
 142
 143     kCFStringEncodingMacRomanLatin1,
 144     kCFStringEncodingHZ_GB_2312,
 145     kCFStringEncodingBig5_HKSCS_1999,
 146     kCFStringEncodingVISCII,
 147     kCFStringEncodingKOI8_U,
 148     kCFStringEncodingBig5_E,
 149     kCFStringEncodingUTF7_IMAP,
 150
 151     kCFStringEncodingNextStepLatin,
 152
 153     kCFStringEncodingEBCDIC_CP037
 154 };
 155
 156 // Windows codepage mapping
 157 static const uint16_t __CFWindowsCPList[] = {
 158     10000,
 159     10001,
 160     10002,
 161     10003,
 162     10004,
 163     10005,
 164     10006,
 165     10007,
 166     0,
 167     0,
 168     0,
 169     0,
 170     0,
 171     0,
 172     0,
 173     0,
 174     0,
 175     0,
 176     0,
 177     0,
 178     10021,
 179     0,
 180     0,
 181     0,
 182     10008,
 183     0,
 184     0,
 185     0,
 186     10029,
 187     0,
 188     0,
 189     0,
 190     10081,
 191     10082,
 192     10079,
 193     10010,
 194     0,
 195     0,
 196     0,
 197     10017,
 198     0,
 199
 200     437,
 201     737,
 202     775,
 203     850,
 204     851,
 205     852,
 206     855,
 207     857,
 208     860,
 209     861,
 210     862,
 211     863,
 212     864,
 213     865,
 214     866,
 215     869,
 216     874,
 217     932,
 218     936,
 219     949,
 220     950,
 221
 222     1252,
 223     1250,
 224     1251,
 225     1253,
 226     1254,
 227     1255,
 228     1256,
 229     1257,
 230     1258,
 231     1361,
 232
 233     20127,
 234
 235     0,
 236     54936,
 237
 238     50221, // we prefere this over 50220/50221 since that's what CF coverter generates
 239     0,
 240     0,
 241     0,
 242     50227,
 243     0,
 244     50225,
 245
 246     51932,
 247     51936,
 248     51950,
 249     51949,
 250
 251     0,
 252
 253     20866,
 254
 255     0,
 256
 257     0,
 258     52936,
 259     0,
 260     0,
 261     21866,
 262     0,
 263     0,
 264
 265     0,
 266
 267     37
 268 };
 269
 270 // Canonical name
 271 static const char *__CFCanonicalNameList[] = {
 272     "macintosh",
 273     "japanese",
 274     "trad-chinese",
 275     "korean",
 276     "arabic",
 277     "hebrew",
 278     "greek",
 279     "cyrillic",
 280     "devanagari",
 281     "gurmukhi",
 282     "gujarati",
 283     "oriya",
 284     "bengali",
 285     "tamil",
 286     "telugu",
 287     "kannada",
 288     "malayalam",
 289     "sinhalese",
 290     "burmese",
 291     "khmer",
 292     "thai",
 293     "laotian",
 294     "georgian",
 295     "armenian",
 296     "simp-chinese",
 297     "tibetan",
 298     "mongolian",
 299     "ethiopic",
 300     "centraleurroman",
 301     "vietnamese",
 302     "symbol",
 303     "dingbats",
 304     "turkish",
 305     "croatian",
 306     "icelandic",
 307     "romanian",
 308     "celtic",
 309     "gaelic",
 310     "farsi",
 311     "ukrainian",
 312     "inuit",
 313
 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     NULL,
 328     NULL,
 329     NULL,
 330     NULL,
 331     NULL,
 332     NULL,
 333     NULL,
 334     NULL,
 335
 336     NULL,
 337     NULL,
 338     NULL,
 339     NULL,
 340     NULL,
 341     NULL,
 342     NULL,
 343     NULL,
 344     NULL,
 345     NULL,
 346
 347     "us-ascii",
 348
 349     NULL,
 350     "gb18030",
 351
 352     "iso-2022-jp",
 353     "iso-2022-jp-2",
 354     "iso-2022-jp-1",
 355     "iso-2022-jp-3",
 356     "iso-2022-cn",
 357     "iso-2022-cn-ext",
 358     "iso-2022-kr",
 359     "euc-jp",
 360     "gb2312",
 361     "euc-tw",
 362     "euc-kr",
 363
 364     "shift_jis",
 365
 366     "koi8-r",
 367
 368     "big5",
 369
 370     "roman-latin1",
 371     "hz-gb-2312",
 372     "big5-hkscs",
 373     "viscii",
 374     "koi8-u",
 375     NULL,
 376     "utf7-imap",
 377
 378     "x-nextstep",
 379
 380     "ibm037",
 381 };
 382
 383 static inline CFIndex __CFGetEncodingIndex(CFStringEncoding encoding) {
 384     const uint16_t *head = __CFKnownEncodingList;
 385     const uint16_t *tail = head + ((sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList)) - 1);
 386     const uint16_t *middle;
 387
 388     encoding &= 0x0FFF;
 389     while (head <= tail) {
 390         middle = head + ((tail - head) >> 1);
 391
 392         if (encoding == *middle) {
 393             return middle - __CFKnownEncodingList;
 394         } else if (encoding < *middle) {
 395             tail = middle - 1;
 396         } else {
 397             head = middle + 1;
 398         }
 399     }
 400
 401     return kCFNotFound;
 402 }
 403
 404 __private_extern__ uint16_t __CFStringEncodingGetWindowsCodePage(CFStringEncoding encoding) {
 405     CFStringEncoding encodingBase = encoding & 0x0F00;
 406
 407     if (0x0100 == encodingBase) { // UTF
 408         switch (encoding) {
 409             case kCFStringEncodingUTF7: return 65000;
 410             case kCFStringEncodingUTF8: return 65001;
 411             case kCFStringEncodingUTF16: return 1200;
 412             case kCFStringEncodingUTF16BE: return 1201;
 413             case kCFStringEncodingUTF32: return 65005;
 414             case kCFStringEncodingUTF32BE: return 65006;
 415         }
 416     } else if (0x0200 == encodingBase) { // ISO 8859 range
 417         return ISO8859CODEPAGE_BASE + (encoding & 0xFF);
 418     } else { // others
 419         CFIndex index = __CFGetEncodingIndex(encoding);
 420
 421         if (kCFNotFound != index) return __CFWindowsCPList[index];
 422     }
 423
 424     return 0;
 425 }
 426
 427 __private_extern__ CFStringEncoding __CFStringEncodingGetFromWindowsCodePage(uint16_t codepage) {
 428     switch (codepage) {
 429         case 65001: return kCFStringEncodingUTF8;
 430         case 1200: return kCFStringEncodingUTF16;
 431         case 0: return kCFStringEncodingInvalidId;
 432         case 1201: return kCFStringEncodingUTF16BE;
 433         case 65005: return kCFStringEncodingUTF32;
 434         case 65006: return kCFStringEncodingUTF32BE;
 435         case 65000: return kCFStringEncodingUTF7;
 436     }
 437
 438     if ((codepage > ISO8859CODEPAGE_BASE) && (codepage <= (ISO8859CODEPAGE_BASE + 16))) {
 439         return (codepage - ISO8859CODEPAGE_BASE) + 0x0200;
 440     } else {
 441         static CFMutableDictionaryRef mappingTable = NULL;
 442         static CFSpinLock_t lock = CFSpinLockInit;
 443         uintptr_t value;
 444
 445         __CFSpinLock(&lock);
 446         if (NULL == mappingTable) {
 447             CFIndex index, count = sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList);
 448
 449             mappingTable = CFDictionaryCreateMutable(NULL, 0, NULL, NULL);
 450
 451             for (index = 0;index < count;index++) {
 452                 if (0 != __CFWindowsCPList[index]) CFDictionarySetValue(mappingTable, (const void *)(uintptr_t)__CFWindowsCPList[index], (const void *)(uintptr_t)__CFKnownEncodingList[index]);
 453             }
 454         }
 455         __CFSpinUnlock(&lock);
 456
 457         if (CFDictionaryGetValueIfPresent(mappingTable, (const void *)(uintptr_t)codepage, (const void **)&value)) return (CFStringEncoding)value;
 458     }
 459
 460
 461     return kCFStringEncodingInvalidId;
 462 }
 463
 464 __private_extern__ bool __CFStringEncodingGetCanonicalName(CFStringEncoding encoding, char *buffer, CFIndex bufferSize) {
 465     const char *format = "%s";
 466     const char *name = NULL;
 467     uint32_t value = 0;
 468     CFIndex index;
 469
 470     switch (encoding & 0x0F00) {
 471         case 0x0100: // UTF range
 472             switch (encoding) {
 473                 case kCFStringEncodingUTF7: name = "utf-7"; break;
 474                 case kCFStringEncodingUTF8: name = "utf-8"; break;
 475                 case kCFStringEncodingUTF16: name = "utf-16"; break;
 476                 case kCFStringEncodingUTF16BE: name = "utf-16be"; break;
 477                 case kCFStringEncodingUTF16LE: name = "utf-16le"; break;
 478                 case kCFStringEncodingUTF32: name = "utf-32"; break;
 479                 case kCFStringEncodingUTF32BE: name = "utf-32be"; break;
 480                 case kCFStringEncodingUTF32LE: name = "utf-32le"; break;
 481             }
 482             break;
 483
 484         case 0x0200: // ISO 8859 range
 485             format = "iso-8859-%d";
 486             value = (encoding & 0xFF);
 487             break;
 488
 489         case 0x0400: // DOS code page range
 490         case 0x0500: // Windows code page range
 491             index = __CFGetEncodingIndex(encoding);
 492
 493             if (kCFNotFound != index) {
 494                 value = __CFWindowsCPList[index];
 495                 if (0 != value) format = ((0x0400 == (encoding & 0x0F00)) ? "cp%d" : "windows-%d");
 496             }
 497             break;
 498
 499         default: // others
 500             index = __CFGetEncodingIndex(encoding);
 501
 502             if (kCFNotFound != index) {
 503                 if (((0 == (encoding & 0x0F00)) && (kCFStringEncodingMacRoman != encoding)) || (kCFStringEncodingMacRomanLatin1 == encoding)) format = "x-mac-%s";
 504                 name = (const char *)__CFCanonicalNameList[index];
 505             }
 506             break;
 507     }
 508
 509     if ((0 == value) && (NULL == name)) {
 510         return false;
 511     } else if (0 != value) {
 512         return ((snprintf(buffer, bufferSize, format, value) < bufferSize) ? true : false);
 513     } else {
 514         return ((snprintf(buffer, bufferSize, format, name) < bufferSize) ? true : false);
 515     }
 516 }
 517
 518 #define LENGTH_LIMIT (256)
 519 static Boolean __CFCanonicalNameCompare(const void *value1, const void *value2) { return ((0 == strncasecmp_l((const char *)value1, (const char *)value2, LENGTH_LIMIT, NULL)) ? true : false); }
 520
 521 static CFHashCode __CFCanonicalNameHash(const void *value) {
 522     const char *name = (const char *)value;
 523     CFHashCode code = 0;
 524
 525     while ((0 != *name) && ((name - (const char *)value) < LENGTH_LIMIT)) {
 526         char character = *(name++);
 527
 528         code += (character + (((character >= 'A') && (character <= 'Z')) ? 'a' - 'A' : 0));
 529     }
 530
 531     return code * (name - (const char *)value);
 532 }
 533
 534 __private_extern__ CFStringEncoding __CFStringEncodingGetFromCanonicalName(const char *canonicalName) {
 535     CFStringEncoding encoding;
 536     CFIndex prefixLength;
 537     static CFMutableDictionaryRef mappingTable = NULL;
 538     static CFSpinLock_t lock = CFSpinLockInit;
 539
 540     prefixLength = strlen("iso-8859-");
 541     if (0 == strncasecmp_l(canonicalName, "iso-8859-", prefixLength, NULL)) {// do ISO
 542         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 543
 544         return (((0 == encoding) || (encoding > 16)) ? kCFStringEncodingInvalidId : encoding + 0x0200);
 545     }
 546
 547     prefixLength = strlen("cp");
 548     if (0 == strncasecmp_l(canonicalName, "cp", prefixLength, NULL)) {// do DOS
 549         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 550
 551         return __CFStringEncodingGetFromWindowsCodePage(encoding);
 552     }
 553
 554     prefixLength = strlen("windows-");
 555     if (0 == strncasecmp_l(canonicalName, "windows-", prefixLength, NULL)) {// do DOS
 556         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 557
 558         return __CFStringEncodingGetFromWindowsCodePage(encoding);
 559     }
 560
 561     __CFSpinLock(&lock);
 562     if (NULL == mappingTable) {
 563         CFIndex index, count = sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList);
 564
 565         CFDictionaryKeyCallBacks keys = {
 566             0, NULL, NULL, NULL, &__CFCanonicalNameCompare, &__CFCanonicalNameHash
 567         };
 568
 569         mappingTable = CFDictionaryCreateMutable(NULL, 0, &keys, NULL);
 570
 571         // Add UTFs
 572         CFDictionarySetValue(mappingTable, "utf-7", (const void *)kCFStringEncodingUTF7);
 573         CFDictionarySetValue(mappingTable, "utf-8", (const void *)kCFStringEncodingUTF8);
 574         CFDictionarySetValue(mappingTable, "utf-16", (const void *)kCFStringEncodingUTF16);
 575         CFDictionarySetValue(mappingTable, "utf-16be", (const void *)kCFStringEncodingUTF16BE);
 576         CFDictionarySetValue(mappingTable, "utf-16le", (const void *)kCFStringEncodingUTF16LE);
 577         CFDictionarySetValue(mappingTable, "utf-32", (const void *)kCFStringEncodingUTF32);
 578         CFDictionarySetValue(mappingTable, "utf-32be", (const void *)kCFStringEncodingUTF32BE);
 579         CFDictionarySetValue(mappingTable, "utf-32le", (const void *)kCFStringEncodingUTF32LE);
 580
 581         for (index = 0;index < count;index++) {
 582             if (NULL != __CFCanonicalNameList[index]) CFDictionarySetValue(mappingTable, (const void *)(uintptr_t)__CFCanonicalNameList[index], (const void *)(uintptr_t)__CFKnownEncodingList[index]);
 583         }
 584     }
 585     __CFSpinUnlock(&lock);
 586
 587     if (0 == strncasecmp_l(canonicalName, "macintosh", sizeof("macintosh") - 1, NULL)) return kCFStringEncodingMacRoman;
 588
 589
 590     prefixLength = strlen("x-mac-");
 591     encoding = (CFStringEncoding)(CFIndex)CFDictionaryGetValue(mappingTable, canonicalName + ((0 == strncasecmp_l(canonicalName, "x-mac-", prefixLength, NULL)) ? prefixLength : 0));
 592
 593     return ((0 == encoding) ? kCFStringEncodingInvalidId : encoding);
 594 }
 595 #undef LENGTH_LIMIT
 596
 597 #if DEPLOYMENT_TARGET_MACOSX
 598 // This list indexes from DOS range
 599 static uint16_t __CFISO8859SimilarScriptList[] = {
 600     kCFStringEncodingMacRoman,
 601     kCFStringEncodingMacCentralEurRoman,
 602     kCFStringEncodingMacRoman,
 603     kCFStringEncodingMacCentralEurRoman,
 604     kCFStringEncodingMacCyrillic,
 605     kCFStringEncodingMacArabic,
 606     kCFStringEncodingMacGreek,
 607     kCFStringEncodingMacHebrew,
 608     kCFStringEncodingMacTurkish,
 609     kCFStringEncodingMacInuit,
 610     kCFStringEncodingMacThai,
 611     kCFStringEncodingMacRoman,
 612     kCFStringEncodingMacCentralEurRoman,
 613     kCFStringEncodingMacCeltic,
 614     kCFStringEncodingMacRoman,
 615     kCFStringEncodingMacRomanian};
 616
 617 static uint16_t __CFOtherSimilarScriptList[] = {
 618     kCFStringEncodingMacRoman,
 619     kCFStringEncodingMacGreek,
 620     kCFStringEncodingMacCentralEurRoman,
 621     kCFStringEncodingMacRoman,
 622     kCFStringEncodingMacGreek,
 623     kCFStringEncodingMacCentralEurRoman,
 624     kCFStringEncodingMacCyrillic,
 625     kCFStringEncodingMacTurkish,
 626     kCFStringEncodingMacRoman,
 627     kCFStringEncodingMacIcelandic,
 628     kCFStringEncodingMacHebrew,
 629     kCFStringEncodingMacRoman,
 630     kCFStringEncodingMacArabic,
 631     kCFStringEncodingMacInuit,
 632     kCFStringEncodingMacCyrillic,
 633     kCFStringEncodingMacGreek,
 634     kCFStringEncodingMacThai,
 635     kCFStringEncodingMacJapanese,
 636     kCFStringEncodingMacChineseSimp,
 637     kCFStringEncodingMacKorean,
 638     kCFStringEncodingMacChineseTrad,
 639
 640     kCFStringEncodingMacRoman,
 641     kCFStringEncodingMacCentralEurRoman,
 642     kCFStringEncodingMacCyrillic,
 643     kCFStringEncodingMacGreek,
 644     kCFStringEncodingMacTurkish,
 645     kCFStringEncodingMacHebrew,
 646     kCFStringEncodingMacArabic,
 647     kCFStringEncodingMacCentralEurRoman,
 648     kCFStringEncodingMacVietnamese,
 649     kCFStringEncodingMacKorean,
 650
 651     kCFStringEncodingMacRoman,
 652
 653     kCFStringEncodingMacJapanese,
 654     kCFStringEncodingMacChineseSimp,
 655
 656     kCFStringEncodingMacJapanese,
 657     kCFStringEncodingMacJapanese,
 658     kCFStringEncodingMacJapanese,
 659     kCFStringEncodingMacJapanese,
 660     kCFStringEncodingMacChineseSimp,
 661     kCFStringEncodingMacChineseSimp,
 662     kCFStringEncodingMacKorean,
 663     kCFStringEncodingMacJapanese,
 664     kCFStringEncodingMacChineseSimp,
 665     kCFStringEncodingMacChineseTrad,
 666     kCFStringEncodingMacKorean,
 667
 668     kCFStringEncodingMacJapanese,
 669
 670     kCFStringEncodingMacCyrillic,
 671
 672     kCFStringEncodingMacChineseTrad,
 673
 674     kCFStringEncodingMacRoman,
 675     kCFStringEncodingMacChineseSimp,
 676     kCFStringEncodingMacChineseTrad,
 677     kCFStringEncodingMacVietnamese,
 678     kCFStringEncodingMacUkrainian,
 679     kCFStringEncodingMacChineseTrad,
 680     kCFStringEncodingMacRoman,
 681
 682     kCFStringEncodingMacRoman,
 683
 684     kCFStringEncodingMacRoman
 685 };
 686
 687 static const char *__CFISONameList[] = {
 688     "Western (ISO Latin 1)",
 689     "Central European (ISO Latin 2)",
 690     "Western (ISO Latin 3)",
 691     "Central European (ISO Latin 4)",
 692     "Cyrillic (ISO 8859-5)",
 693     "Arabic (ISO 8859-6)",
 694     "Greek (ISO 8859-7)",
 695     "Hebrew (ISO 8859-8)",
 696     "Turkish (ISO Latin 5)",
 697     "Nordic (ISO Latin 6)",
 698     "Thai (ISO 8859-11)",
 699     NULL,
 700     "Baltic (ISO Latin 7)",
 701     "Celtic (ISO Latin 8)",
 702     "Western (ISO Latin 9)",
 703     "Romanian (ISO Latin 10)",
 704 };
 705
 706 static const char *__CFOtherNameList[] = {
 707     "Western (Mac OS Roman)",
 708     "Japanese (Mac OS)",
 709     "Traditional Chinese (Mac OS)",
 710     "Korean (Mac OS)",
 711     "Arabic (Mac OS)",
 712     "Hebrew (Mac OS)",
 713     "Greek (Mac OS)",
 714     "Cyrillic (Mac OS)",
 715     "Devanagari (Mac OS)",
 716     "Gurmukhi (Mac OS)",
 717     "Gujarati (Mac OS)",
 718     "Oriya (Mac OS)",
 719     "Bengali (Mac OS)",
 720     "Tamil (Mac OS)",
 721     "Telugu (Mac OS)",
 722     "Kannada (Mac OS)",
 723     "Malayalam (Mac OS)",
 724     "Sinhalese (Mac OS)",
 725     "Burmese (Mac OS)",
 726     "Khmer (Mac OS)",
 727     "Thai (Mac OS)",
 728     "Laotian (Mac OS)",
 729     "Georgian (Mac OS)",
 730     "Armenian (Mac OS)",
 731     "Simplified Chinese (Mac OS)",
 732     "Tibetan (Mac OS)",
 733     "Mongolian (Mac OS)",
 734     "Ethiopic (Mac OS)",
 735     "Central European (Mac OS)",
 736     "Vietnamese (Mac OS)",
 737     "Symbol (Mac OS)",
 738     "Dingbats (Mac OS)",
 739     "Turkish (Mac OS)",
 740     "Croatian (Mac OS)",
 741     "Icelandic (Mac OS)",
 742     "Romanian (Mac OS)",
 743     "Celtic (Mac OS)",
 744     "Gaelic (Mac OS)",
 745     "Farsi (Mac OS)",
 746     "Cyrillic (Mac OS Ukrainian)",
 747     "Inuit (Mac OS)",
 748     "Latin-US (DOS)",
 749     "Greek (DOS)",
 750     "Baltic (DOS)",
 751     "Western (DOS Latin 1)",
 752     "Greek (DOS Greek 1)",
 753     "Central European (DOS Latin 2)",
 754     "Cyrillic (DOS)",
 755     "Turkish (DOS)",
 756     "Portuguese (DOS)",
 757     "Icelandic (DOS)",
 758     "Hebrew (DOS)",
 759     "Canadian French (DOS)",
 760     "Arabic (DOS)",
 761     "Nordic (DOS)",
 762     "Russian (DOS)",
 763     "Greek (DOS Greek 2)",
 764     "Thai (Windows, DOS)",
 765     "Japanese (Windows, DOS)",
 766     "Simplified Chinese (Windows, DOS)",
 767     "Korean (Windows, DOS)",
 768     "Traditional Chinese (Windows, DOS)",
 769     "Western (Windows Latin 1)",
 770     "Central European (Windows Latin 2)",
 771     "Cyrillic (Windows)",
 772     "Greek (Windows)",
 773     "Turkish (Windows Latin 5)",
 774     "Hebrew (Windows)",
 775     "Arabic (Windows)",
 776     "Baltic (Windows)",
 777     "Vietnamese (Windows)",
 778     "Korean (Windows Johab)",
 779     "Western (ASCII)",
 780     "Japanese (Shift JIS X0213)",
 781     "Chinese (GB 18030)",
 782     "Japanese (ISO 2022-JP)",
 783     "Japanese (ISO 2022-JP-2)",
 784     "Japanese (ISO 2022-JP-1)",
 785     "Japanese (ISO 2022-JP-3)",
 786     "Chinese (ISO 2022-CN)",
 787     "Chinese (ISO 2022-CN-EXT)",
 788     "Korean (ISO 2022-KR)",
 789     "Japanese (EUC)",
 790     "Simplified Chinese (GB 2312)",
 791     "Traditional Chinese (EUC)",
 792     "Korean (EUC)",
 793     "Japanese (Shift JIS)",
 794     "Cyrillic (KOI8-R)",
 795     "Traditional Chinese (Big 5)",
 796     "Western (Mac Mail)",
 797     "Simplified Chinese (HZ GB 2312)",
 798     "Traditional Chinese (Big 5 HKSCS)",
 799     NULL,
 800     "Ukrainian (KOI8-U)",
 801     "Traditional Chinese (Big 5-E)",
 802     NULL,
 803     "Western (NextStep)",
 804     "Western (EBCDIC Latin 1)",
 805 };
 806 #endif /* DEPLOYMENT_TARGET_MACOSX */
 807
 808 __private_extern__ CFStringEncoding __CFStringEncodingGetMostCompatibleMacScript(CFStringEncoding encoding) {
 809 #if DEPLOYMENT_TARGET_MACOSX
 810     switch (encoding & 0x0F00) {
 811         case 0: return encoding & 0xFF; break; // Mac scripts
 812
 813         case 0x0100: return kCFStringEncodingUnicode; break; // Unicode
 814
 815         case 0x200: // ISO 8859
 816             return (((encoding & 0xFF) <= (sizeof(__CFISO8859SimilarScriptList) / sizeof(*__CFISO8859SimilarScriptList))) ? __CFISO8859SimilarScriptList[(encoding & 0xFF) - 1] : kCFStringEncodingInvalidId);
 817             break;
 818
 819         default: {
 820             CFIndex index = __CFGetEncodingIndex(encoding);
 821
 822             if (kCFNotFound != index) {
 823                 index -= __CFGetEncodingIndex(kCFStringEncodingDOSLatinUS);
 824                 return __CFOtherSimilarScriptList[index];
 825             }
 826         }
 827     }
 828 #endif /* DEPLOYMENT_TARGET_MACOSX */
 829
 830     return kCFStringEncodingInvalidId;
 831 }
 832
 833 __private_extern__ const char *__CFStringEncodingGetName(CFStringEncoding encoding) {
 834     switch (encoding) {
 835         case kCFStringEncodingUTF8: return "Unicode (UTF-8)"; break;
 836         case kCFStringEncodingUTF16: return "Unicode (UTF-16)"; break;
 837         case kCFStringEncodingUTF16BE: return "Unicode (UTF-16BE)"; break;
 838         case kCFStringEncodingUTF16LE: return "Unicode (UTF-16LE)"; break;
 839         case kCFStringEncodingUTF32: return "Unicode (UTF-32)"; break;
 840         case kCFStringEncodingUTF32BE: return "Unicode (UTF-32BE)"; break;
 841         case kCFStringEncodingUTF32LE: return "Unicode (UTF-32LE)"; break;
 842         case kCFStringEncodingNonLossyASCII: return "Non-lossy ASCII"; break;
 843         case kCFStringEncodingUTF7: return "Unicode (UTF-7)"; break;
 844     }
 845
 846 #if DEPLOYMENT_TARGET_MACOSX
 847     if (0x0200 == (encoding & 0x0F00)) {
 848         encoding &= 0x00FF;
 849
 850         if (encoding <= (sizeof(__CFISONameList) / sizeof(*__CFISONameList))) return __CFISONameList[encoding - 1];
 851     } else {
 852         CFIndex index = __CFGetEncodingIndex(encoding);
 853
 854         if (kCFNotFound != index) return __CFOtherNameList[index];
 855     }
 856 #endif /* DEPLOYMENT_TARGET_MACOSX */
 857
 858     return NULL;
 859 }