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 /*
  25  *  CFStringEncodingDatabase.c
  26  *  CoreFoundation
  27  *
  28  *  Created by Aki Inoue on 07/12/05.
  29  *  Copyright 2007-2009, Apple Inc. All rights reserved.
  30  *
  31  */
  32
  33 #include "CFInternal.h"
  34 #include <CoreFoundation/CFStringEncodingExt.h>
  35 #include "CFStringEncodingConverterPriv.h"
  36 #include "CFStringEncodingDatabase.h"
  37 #include <stdio.h>
  38
  39 #if DEPLOYMENT_TARGET_WINDOWS
  40 #define strncasecmp_l(a, b, c, d) _strnicmp(a, b, c)
  41 #define snprintf _snprintf
  42 #endif
  43
  44 #define ISO8859CODEPAGE_BASE (28590)
  45
  46 static const uint16_t __CFKnownEncodingList[] = {
  47     kCFStringEncodingMacRoman,
  48     kCFStringEncodingMacJapanese,
  49     kCFStringEncodingMacChineseTrad,
  50     kCFStringEncodingMacKorean,
  51     kCFStringEncodingMacArabic,
  52     kCFStringEncodingMacHebrew,
  53     kCFStringEncodingMacGreek,
  54     kCFStringEncodingMacCyrillic,
  55     kCFStringEncodingMacDevanagari,
  56     kCFStringEncodingMacGurmukhi,
  57     kCFStringEncodingMacGujarati,
  58     kCFStringEncodingMacOriya,
  59     kCFStringEncodingMacBengali,
  60     kCFStringEncodingMacTamil,
  61     kCFStringEncodingMacTelugu,
  62     kCFStringEncodingMacKannada,
  63     kCFStringEncodingMacMalayalam,
  64     kCFStringEncodingMacSinhalese,
  65     kCFStringEncodingMacBurmese,
  66     kCFStringEncodingMacKhmer,
  67     kCFStringEncodingMacThai,
  68     kCFStringEncodingMacLaotian,
  69     kCFStringEncodingMacGeorgian,
  70     kCFStringEncodingMacArmenian,
  71     kCFStringEncodingMacChineseSimp,
  72     kCFStringEncodingMacTibetan,
  73     kCFStringEncodingMacMongolian,
  74     kCFStringEncodingMacEthiopic,
  75     kCFStringEncodingMacCentralEurRoman,
  76     kCFStringEncodingMacVietnamese,
  77     kCFStringEncodingMacSymbol,
  78     kCFStringEncodingMacDingbats,
  79     kCFStringEncodingMacTurkish,
  80     kCFStringEncodingMacCroatian,
  81     kCFStringEncodingMacIcelandic,
  82     kCFStringEncodingMacRomanian,
  83     kCFStringEncodingMacCeltic,
  84     kCFStringEncodingMacGaelic,
  85     kCFStringEncodingMacFarsi,
  86     kCFStringEncodingMacUkrainian,
  87     kCFStringEncodingMacInuit,
  88
  89     kCFStringEncodingDOSLatinUS,
  90     kCFStringEncodingDOSGreek,
  91     kCFStringEncodingDOSBalticRim,
  92     kCFStringEncodingDOSLatin1,
  93     kCFStringEncodingDOSGreek1,
  94     kCFStringEncodingDOSLatin2,
  95     kCFStringEncodingDOSCyrillic,
  96     kCFStringEncodingDOSTurkish,
  97     kCFStringEncodingDOSPortuguese,
  98     kCFStringEncodingDOSIcelandic,
  99     kCFStringEncodingDOSHebrew,
 100     kCFStringEncodingDOSCanadianFrench,
 101     kCFStringEncodingDOSArabic,
 102     kCFStringEncodingDOSNordic,
 103     kCFStringEncodingDOSRussian,
 104     kCFStringEncodingDOSGreek2,
 105     kCFStringEncodingDOSThai,
 106     kCFStringEncodingDOSJapanese,
 107     kCFStringEncodingDOSChineseSimplif,
 108     kCFStringEncodingDOSKorean,
 109     kCFStringEncodingDOSChineseTrad,
 110
 111     kCFStringEncodingWindowsLatin1,
 112     kCFStringEncodingWindowsLatin2,
 113     kCFStringEncodingWindowsCyrillic,
 114     kCFStringEncodingWindowsGreek,
 115     kCFStringEncodingWindowsLatin5,
 116     kCFStringEncodingWindowsHebrew,
 117     kCFStringEncodingWindowsArabic,
 118     kCFStringEncodingWindowsBalticRim,
 119     kCFStringEncodingWindowsVietnamese,
 120     kCFStringEncodingWindowsKoreanJohab,
 121     kCFStringEncodingASCII,
 122
 123     kCFStringEncodingShiftJIS_X0213,
 124     kCFStringEncodingGB_18030_2000,
 125
 126     kCFStringEncodingISO_2022_JP,
 127     kCFStringEncodingISO_2022_JP_2,
 128     kCFStringEncodingISO_2022_JP_1,
 129     kCFStringEncodingISO_2022_JP_3,
 130     kCFStringEncodingISO_2022_CN,
 131     kCFStringEncodingISO_2022_CN_EXT,
 132     kCFStringEncodingISO_2022_KR,
 133     kCFStringEncodingEUC_JP,
 134     kCFStringEncodingEUC_CN,
 135     kCFStringEncodingEUC_TW,
 136     kCFStringEncodingEUC_KR,
 137
 138     kCFStringEncodingShiftJIS,
 139
 140     kCFStringEncodingKOI8_R,
 141
 142     kCFStringEncodingBig5,
 143
 144     kCFStringEncodingMacRomanLatin1,
 145     kCFStringEncodingHZ_GB_2312,
 146     kCFStringEncodingBig5_HKSCS_1999,
 147     kCFStringEncodingVISCII,
 148     kCFStringEncodingKOI8_U,
 149     kCFStringEncodingBig5_E,
 150     kCFStringEncodingUTF7_IMAP,
 151
 152     kCFStringEncodingNextStepLatin,
 153
 154     kCFStringEncodingEBCDIC_CP037
 155 };
 156
 157 // Windows codepage mapping
 158 static const uint16_t __CFWindowsCPList[] = {
 159     10000,
 160     10001,
 161     10002,
 162     10003,
 163     10004,
 164     10005,
 165     10006,
 166     10007,
 167     0,
 168     0,
 169     0,
 170     0,
 171     0,
 172     0,
 173     0,
 174     0,
 175     0,
 176     0,
 177     0,
 178     0,
 179     10021,
 180     0,
 181     0,
 182     0,
 183     10008,
 184     0,
 185     0,
 186     0,
 187     10029,
 188     0,
 189     0,
 190     0,
 191     10081,
 192     10082,
 193     10079,
 194     10010,
 195     0,
 196     0,
 197     0,
 198     10017,
 199     0,
 200
 201     437,
 202     737,
 203     775,
 204     850,
 205     851,
 206     852,
 207     855,
 208     857,
 209     860,
 210     861,
 211     862,
 212     863,
 213     864,
 214     865,
 215     866,
 216     869,
 217     874,
 218     932,
 219     936,
 220     949,
 221     950,
 222
 223     1252,
 224     1250,
 225     1251,
 226     1253,
 227     1254,
 228     1255,
 229     1256,
 230     1257,
 231     1258,
 232     1361,
 233
 234     20127,
 235
 236     0,
 237     54936,
 238
 239     50221, // we prefere this over 50220/50221 since that's what CF coverter generates
 240     0,
 241     0,
 242     0,
 243     50227,
 244     0,
 245     50225,
 246
 247     51932,
 248     51936,
 249     51950,
 250     51949,
 251
 252     0,
 253
 254     20866,
 255
 256     0,
 257
 258     0,
 259     52936,
 260     0,
 261     0,
 262     21866,
 263     0,
 264     0,
 265
 266     0,
 267
 268     37
 269 };
 270
 271 // Canonical name
 272 static const char *__CFCanonicalNameList[] = {
 273     "macintosh",
 274     "japanese",
 275     "trad-chinese",
 276     "korean",
 277     "arabic",
 278     "hebrew",
 279     "greek",
 280     "cyrillic",
 281     "devanagari",
 282     "gurmukhi",
 283     "gujarati",
 284     "oriya",
 285     "bengali",
 286     "tamil",
 287     "telugu",
 288     "kannada",
 289     "malayalam",
 290     "sinhalese",
 291     "burmese",
 292     "khmer",
 293     "thai",
 294     "laotian",
 295     "georgian",
 296     "armenian",
 297     "simp-chinese",
 298     "tibetan",
 299     "mongolian",
 300     "ethiopic",
 301     "centraleurroman",
 302     "vietnamese",
 303     "symbol",
 304     "dingbats",
 305     "turkish",
 306     "croatian",
 307     "icelandic",
 308     "romanian",
 309     "celtic",
 310     "gaelic",
 311     "farsi",
 312     "ukrainian",
 313     "inuit",
 314
 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     NULL,
 336
 337     NULL,
 338     NULL,
 339     NULL,
 340     NULL,
 341     NULL,
 342     NULL,
 343     NULL,
 344     NULL,
 345     NULL,
 346     NULL,
 347
 348     "us-ascii",
 349
 350     NULL,
 351     "gb18030",
 352
 353     "iso-2022-jp",
 354     "iso-2022-jp-2",
 355     "iso-2022-jp-1",
 356     "iso-2022-jp-3",
 357     "iso-2022-cn",
 358     "iso-2022-cn-ext",
 359     "iso-2022-kr",
 360     "euc-jp",
 361     "gb2312",
 362     "euc-tw",
 363     "euc-kr",
 364
 365     "shift_jis",
 366
 367     "koi8-r",
 368
 369     "big5",
 370
 371     "roman-latin1",
 372     "hz-gb-2312",
 373     "big5-hkscs",
 374     "viscii",
 375     "koi8-u",
 376     NULL,
 377     "utf7-imap",
 378
 379     "x-nextstep",
 380
 381     "ibm037",
 382 };
 383
 384 static inline CFIndex __CFGetEncodingIndex(CFStringEncoding encoding) {
 385     const uint16_t *head = __CFKnownEncodingList;
 386     const uint16_t *tail = head + ((sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList)) - 1);
 387     const uint16_t *middle;
 388
 389     encoding &= 0x0FFF;
 390     while (head <= tail) {
 391         middle = head + ((tail - head) >> 1);
 392
 393         if (encoding == *middle) {
 394             return middle - __CFKnownEncodingList;
 395         } else if (encoding < *middle) {
 396             tail = middle - 1;
 397         } else {
 398             head = middle + 1;
 399         }
 400     }
 401
 402     return kCFNotFound;
 403 }
 404
 405 __private_extern__ uint16_t __CFStringEncodingGetWindowsCodePage(CFStringEncoding encoding) {
 406     CFStringEncoding encodingBase = encoding & 0x0F00;
 407
 408     if (0x0100 == encodingBase) { // UTF
 409         switch (encoding) {
 410             case kCFStringEncodingUTF7: return 65000;
 411             case kCFStringEncodingUTF8: return 65001;
 412             case kCFStringEncodingUTF16: return 1200;
 413             case kCFStringEncodingUTF16BE: return 1201;
 414             case kCFStringEncodingUTF32: return 65005;
 415             case kCFStringEncodingUTF32BE: return 65006;
 416         }
 417     } else if (0x0200 == encodingBase) { // ISO 8859 range
 418         return ISO8859CODEPAGE_BASE + (encoding & 0xFF);
 419     } else { // others
 420         CFIndex index = __CFGetEncodingIndex(encoding);
 421
 422         if (kCFNotFound != index) return __CFWindowsCPList[index];
 423     }
 424
 425     return 0;
 426 }
 427
 428 __private_extern__ CFStringEncoding __CFStringEncodingGetFromWindowsCodePage(uint16_t codepage) {
 429     switch (codepage) {
 430         case 65001: return kCFStringEncodingUTF8;
 431         case 1200: return kCFStringEncodingUTF16;
 432         case 0: return kCFStringEncodingInvalidId;
 433         case 1201: return kCFStringEncodingUTF16BE;
 434         case 65005: return kCFStringEncodingUTF32;
 435         case 65006: return kCFStringEncodingUTF32BE;
 436         case 65000: return kCFStringEncodingUTF7;
 437     }
 438
 439     if ((codepage > ISO8859CODEPAGE_BASE) && (codepage <= (ISO8859CODEPAGE_BASE + 16))) {
 440         return (codepage - ISO8859CODEPAGE_BASE) + 0x0200;
 441     } else {
 442         static CFMutableDictionaryRef mappingTable = NULL;
 443         static CFSpinLock_t lock = CFSpinLockInit;
 444         uintptr_t value;
 445
 446         __CFSpinLock(&lock);
 447         if (NULL == mappingTable) {
 448             CFIndex index, count = sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList);
 449
 450             mappingTable = CFDictionaryCreateMutable(NULL, 0, NULL, NULL);
 451
 452             for (index = 0;index < count;index++) {
 453                 if (0 != __CFWindowsCPList[index]) CFDictionarySetValue(mappingTable, (const void *)(uintptr_t)__CFWindowsCPList[index], (const void *)(uintptr_t)__CFKnownEncodingList[index]);
 454             }
 455         }
 456         __CFSpinUnlock(&lock);
 457
 458         if (CFDictionaryGetValueIfPresent(mappingTable, (const void *)(uintptr_t)codepage, (const void **)&value)) return (CFStringEncoding)value;
 459     }
 460
 461
 462     return kCFStringEncodingInvalidId;
 463 }
 464
 465 __private_extern__ bool __CFStringEncodingGetCanonicalName(CFStringEncoding encoding, char *buffer, CFIndex bufferSize) {
 466     const char *format = "%s";
 467     const char *name = NULL;
 468     uint32_t value = 0;
 469     CFIndex index;
 470
 471     switch (encoding & 0x0F00) {
 472         case 0x0100: // UTF range
 473             switch (encoding) {
 474                 case kCFStringEncodingUTF7: name = "utf-7"; break;
 475                 case kCFStringEncodingUTF8: name = "utf-8"; break;
 476                 case kCFStringEncodingUTF16: name = "utf-16"; break;
 477                 case kCFStringEncodingUTF16BE: name = "utf-16be"; break;
 478                 case kCFStringEncodingUTF16LE: name = "utf-16le"; break;
 479                 case kCFStringEncodingUTF32: name = "utf-32"; break;
 480                 case kCFStringEncodingUTF32BE: name = "utf-32be"; break;
 481                 case kCFStringEncodingUTF32LE: name = "utf-32le"; break;
 482             }
 483             break;
 484
 485         case 0x0200: // ISO 8859 range
 486             format = "iso-8859-%d";
 487             value = (encoding & 0xFF);
 488             break;
 489
 490         case 0x0400: // DOS code page range
 491         case 0x0500: // Windows code page range
 492             index = __CFGetEncodingIndex(encoding);
 493
 494             if (kCFNotFound != index) {
 495                 value = __CFWindowsCPList[index];
 496                 if (0 != value) format = ((0x0400 == (encoding & 0x0F00)) ? "cp%d" : "windows-%d");
 497             }
 498             break;
 499
 500         default: // others
 501             index = __CFGetEncodingIndex(encoding);
 502
 503             if (kCFNotFound != index) {
 504                 if (((0 == (encoding & 0x0F00)) && (kCFStringEncodingMacRoman != encoding)) || (kCFStringEncodingMacRomanLatin1 == encoding)) format = "x-mac-%s";
 505                 name = (const char *)__CFCanonicalNameList[index];
 506             }
 507             break;
 508     }
 509
 510     if ((0 == value) && (NULL == name)) {
 511         return false;
 512     } else if (0 != value) {
 513         return ((snprintf(buffer, bufferSize, format, value) < bufferSize) ? true : false);
 514     } else {
 515         return ((snprintf(buffer, bufferSize, format, name) < bufferSize) ? true : false);
 516     }
 517 }
 518
 519 #define LENGTH_LIMIT (256)
 520 static Boolean __CFCanonicalNameCompare(const void *value1, const void *value2) { return ((0 == strncasecmp_l((const char *)value1, (const char *)value2, LENGTH_LIMIT, NULL)) ? true : false); }
 521
 522 static CFHashCode __CFCanonicalNameHash(const void *value) {
 523     const char *name = (const char *)value;
 524     CFHashCode code = 0;
 525
 526     while ((0 != *name) && ((name - (const char *)value) < LENGTH_LIMIT)) {
 527         char character = *(name++);
 528
 529         code += (character + (((character >= 'A') && (character <= 'Z')) ? 'a' - 'A' : 0));
 530     }
 531
 532     return code * (name - (const char *)value);
 533 }
 534
 535 __private_extern__ CFStringEncoding __CFStringEncodingGetFromCanonicalName(const char *canonicalName) {
 536     CFStringEncoding encoding;
 537     CFIndex prefixLength;
 538     static CFMutableDictionaryRef mappingTable = NULL;
 539     static CFSpinLock_t lock = CFSpinLockInit;
 540
 541     prefixLength = strlen("iso-8859-");
 542     if (0 == strncasecmp_l(canonicalName, "iso-8859-", prefixLength, NULL)) {// do ISO
 543         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 544
 545         return (((0 == encoding) || (encoding > 16)) ? kCFStringEncodingInvalidId : encoding + 0x0200);
 546     }
 547
 548     prefixLength = strlen("cp");
 549     if (0 == strncasecmp_l(canonicalName, "cp", prefixLength, NULL)) {// do DOS
 550         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 551
 552         return __CFStringEncodingGetFromWindowsCodePage(encoding);
 553     }
 554
 555     prefixLength = strlen("windows-");
 556     if (0 == strncasecmp_l(canonicalName, "windows-", prefixLength, NULL)) {// do DOS
 557         encoding = strtol(canonicalName + prefixLength, NULL, 10);
 558
 559         return __CFStringEncodingGetFromWindowsCodePage(encoding);
 560     }
 561
 562     __CFSpinLock(&lock);
 563     if (NULL == mappingTable) {
 564         CFIndex index, count = sizeof(__CFKnownEncodingList) / sizeof(*__CFKnownEncodingList);
 565
 566         CFDictionaryKeyCallBacks keys = {
 567             0, NULL, NULL, NULL, &__CFCanonicalNameCompare, &__CFCanonicalNameHash
 568         };
 569
 570         mappingTable = CFDictionaryCreateMutable(NULL, 0, &keys, NULL);
 571
 572         // Add UTFs
 573         CFDictionarySetValue(mappingTable, "utf-7", (const void *)kCFStringEncodingUTF7);
 574         CFDictionarySetValue(mappingTable, "utf-8", (const void *)kCFStringEncodingUTF8);
 575         CFDictionarySetValue(mappingTable, "utf-16", (const void *)kCFStringEncodingUTF16);
 576         CFDictionarySetValue(mappingTable, "utf-16be", (const void *)kCFStringEncodingUTF16BE);
 577         CFDictionarySetValue(mappingTable, "utf-16le", (const void *)kCFStringEncodingUTF16LE);
 578         CFDictionarySetValue(mappingTable, "utf-32", (const void *)kCFStringEncodingUTF32);
 579         CFDictionarySetValue(mappingTable, "utf-32be", (const void *)kCFStringEncodingUTF32BE);
 580         CFDictionarySetValue(mappingTable, "utf-32le", (const void *)kCFStringEncodingUTF32LE);
 581
 582         for (index = 0;index < count;index++) {
 583             if (NULL != __CFCanonicalNameList[index]) CFDictionarySetValue(mappingTable, (const void *)(uintptr_t)__CFCanonicalNameList[index], (const void *)(uintptr_t)__CFKnownEncodingList[index]);
 584         }
 585     }
 586     __CFSpinUnlock(&lock);
 587
 588     if (0 == strncasecmp_l(canonicalName, "macintosh", sizeof("macintosh") - 1, NULL)) return kCFStringEncodingMacRoman;
 589
 590
 591     prefixLength = strlen("x-mac-");
 592     encoding = (CFStringEncoding)(CFIndex)CFDictionaryGetValue(mappingTable, canonicalName + ((0 == strncasecmp_l(canonicalName, "x-mac-", prefixLength, NULL)) ? prefixLength : 0));
 593
 594     return ((0 == encoding) ? kCFStringEncodingInvalidId : encoding);
 595 }
 596 #undef LENGTH_LIMIT
 597
 598 #if DEPLOYMENT_TARGET_MACOSX
 599 // This list indexes from DOS range
 600 static uint16_t __CFISO8859SimilarScriptList[] = {
 601     kCFStringEncodingMacRoman,
 602     kCFStringEncodingMacCentralEurRoman,
 603     kCFStringEncodingMacRoman,
 604     kCFStringEncodingMacCentralEurRoman,
 605     kCFStringEncodingMacCyrillic,
 606     kCFStringEncodingMacArabic,
 607     kCFStringEncodingMacGreek,
 608     kCFStringEncodingMacHebrew,
 609     kCFStringEncodingMacTurkish,
 610     kCFStringEncodingMacInuit,
 611     kCFStringEncodingMacThai,
 612     kCFStringEncodingMacRoman,
 613     kCFStringEncodingMacCentralEurRoman,
 614     kCFStringEncodingMacCeltic,
 615     kCFStringEncodingMacRoman,
 616     kCFStringEncodingMacRomanian};
 617
 618 static uint16_t __CFOtherSimilarScriptList[] = {
 619     kCFStringEncodingMacRoman,
 620     kCFStringEncodingMacGreek,
 621     kCFStringEncodingMacCentralEurRoman,
 622     kCFStringEncodingMacRoman,
 623     kCFStringEncodingMacGreek,
 624     kCFStringEncodingMacCentralEurRoman,
 625     kCFStringEncodingMacCyrillic,
 626     kCFStringEncodingMacTurkish,
 627     kCFStringEncodingMacRoman,
 628     kCFStringEncodingMacIcelandic,
 629     kCFStringEncodingMacHebrew,
 630     kCFStringEncodingMacRoman,
 631     kCFStringEncodingMacArabic,
 632     kCFStringEncodingMacInuit,
 633     kCFStringEncodingMacCyrillic,
 634     kCFStringEncodingMacGreek,
 635     kCFStringEncodingMacThai,
 636     kCFStringEncodingMacJapanese,
 637     kCFStringEncodingMacChineseSimp,
 638     kCFStringEncodingMacKorean,
 639     kCFStringEncodingMacChineseTrad,
 640
 641     kCFStringEncodingMacRoman,
 642     kCFStringEncodingMacCentralEurRoman,
 643     kCFStringEncodingMacCyrillic,
 644     kCFStringEncodingMacGreek,
 645     kCFStringEncodingMacTurkish,
 646     kCFStringEncodingMacHebrew,
 647     kCFStringEncodingMacArabic,
 648     kCFStringEncodingMacCentralEurRoman,
 649     kCFStringEncodingMacVietnamese,
 650     kCFStringEncodingMacKorean,
 651
 652     kCFStringEncodingMacRoman,
 653
 654     kCFStringEncodingMacJapanese,
 655     kCFStringEncodingMacChineseSimp,
 656
 657     kCFStringEncodingMacJapanese,
 658     kCFStringEncodingMacJapanese,
 659     kCFStringEncodingMacJapanese,
 660     kCFStringEncodingMacJapanese,
 661     kCFStringEncodingMacChineseSimp,
 662     kCFStringEncodingMacChineseSimp,
 663     kCFStringEncodingMacKorean,
 664     kCFStringEncodingMacJapanese,
 665     kCFStringEncodingMacChineseSimp,
 666     kCFStringEncodingMacChineseTrad,
 667     kCFStringEncodingMacKorean,
 668
 669     kCFStringEncodingMacJapanese,
 670
 671     kCFStringEncodingMacCyrillic,
 672
 673     kCFStringEncodingMacChineseTrad,
 674
 675     kCFStringEncodingMacRoman,
 676     kCFStringEncodingMacChineseSimp,
 677     kCFStringEncodingMacChineseTrad,
 678     kCFStringEncodingMacVietnamese,
 679     kCFStringEncodingMacUkrainian,
 680     kCFStringEncodingMacChineseTrad,
 681     kCFStringEncodingMacRoman,
 682
 683     kCFStringEncodingMacRoman,
 684
 685     kCFStringEncodingMacRoman
 686 };
 687
 688 static const char *__CFISONameList[] = {
 689     "Western (ISO Latin 1)",
 690     "Central European (ISO Latin 2)",
 691     "Western (ISO Latin 3)",
 692     "Central European (ISO Latin 4)",
 693     "Cyrillic (ISO 8859-5)",
 694     "Arabic (ISO 8859-6)",
 695     "Greek (ISO 8859-7)",
 696     "Hebrew (ISO 8859-8)",
 697     "Turkish (ISO Latin 5)",
 698     "Nordic (ISO Latin 6)",
 699     "Thai (ISO 8859-11)",
 700     NULL,
 701     "Baltic (ISO Latin 7)",
 702     "Celtic (ISO Latin 8)",
 703     "Western (ISO Latin 9)",
 704     "Romanian (ISO Latin 10)",
 705 };
 706
 707 static const char *__CFOtherNameList[] = {
 708     "Western (Mac OS Roman)",
 709     "Japanese (Mac OS)",
 710     "Traditional Chinese (Mac OS)",
 711     "Korean (Mac OS)",
 712     "Arabic (Mac OS)",
 713     "Hebrew (Mac OS)",
 714     "Greek (Mac OS)",
 715     "Cyrillic (Mac OS)",
 716     "Devanagari (Mac OS)",
 717     "Gurmukhi (Mac OS)",
 718     "Gujarati (Mac OS)",
 719     "Oriya (Mac OS)",
 720     "Bengali (Mac OS)",
 721     "Tamil (Mac OS)",
 722     "Telugu (Mac OS)",
 723     "Kannada (Mac OS)",
 724     "Malayalam (Mac OS)",
 725     "Sinhalese (Mac OS)",
 726     "Burmese (Mac OS)",
 727     "Khmer (Mac OS)",
 728     "Thai (Mac OS)",
 729     "Laotian (Mac OS)",
 730     "Georgian (Mac OS)",
 731     "Armenian (Mac OS)",
 732     "Simplified Chinese (Mac OS)",
 733     "Tibetan (Mac OS)",
 734     "Mongolian (Mac OS)",
 735     "Ethiopic (Mac OS)",
 736     "Central European (Mac OS)",
 737     "Vietnamese (Mac OS)",
 738     "Symbol (Mac OS)",
 739     "Dingbats (Mac OS)",
 740     "Turkish (Mac OS)",
 741     "Croatian (Mac OS)",
 742     "Icelandic (Mac OS)",
 743     "Romanian (Mac OS)",
 744     "Celtic (Mac OS)",
 745     "Gaelic (Mac OS)",
 746     "Farsi (Mac OS)",
 747     "Cyrillic (Mac OS Ukrainian)",
 748     "Inuit (Mac OS)",
 749     "Latin-US (DOS)",
 750     "Greek (DOS)",
 751     "Baltic (DOS)",
 752     "Western (DOS Latin 1)",
 753     "Greek (DOS Greek 1)",
 754     "Central European (DOS Latin 2)",
 755     "Cyrillic (DOS)",
 756     "Turkish (DOS)",
 757     "Portuguese (DOS)",
 758     "Icelandic (DOS)",
 759     "Hebrew (DOS)",
 760     "Canadian French (DOS)",
 761     "Arabic (DOS)",
 762     "Nordic (DOS)",
 763     "Russian (DOS)",
 764     "Greek (DOS Greek 2)",
 765     "Thai (Windows, DOS)",
 766     "Japanese (Windows, DOS)",
 767     "Simplified Chinese (Windows, DOS)",
 768     "Korean (Windows, DOS)",
 769     "Traditional Chinese (Windows, DOS)",
 770     "Western (Windows Latin 1)",
 771     "Central European (Windows Latin 2)",
 772     "Cyrillic (Windows)",
 773     "Greek (Windows)",
 774     "Turkish (Windows Latin 5)",
 775     "Hebrew (Windows)",
 776     "Arabic (Windows)",
 777     "Baltic (Windows)",
 778     "Vietnamese (Windows)",
 779     "Korean (Windows Johab)",
 780     "Western (ASCII)",
 781     "Japanese (Shift JIS X0213)",
 782     "Chinese (GB 18030)",
 783     "Japanese (ISO 2022-JP)",
 784     "Japanese (ISO 2022-JP-2)",
 785     "Japanese (ISO 2022-JP-1)",
 786     "Japanese (ISO 2022-JP-3)",
 787     "Chinese (ISO 2022-CN)",
 788     "Chinese (ISO 2022-CN-EXT)",
 789     "Korean (ISO 2022-KR)",
 790     "Japanese (EUC)",
 791     "Simplified Chinese (GB 2312)",
 792     "Traditional Chinese (EUC)",
 793     "Korean (EUC)",
 794     "Japanese (Shift JIS)",
 795     "Cyrillic (KOI8-R)",
 796     "Traditional Chinese (Big 5)",
 797     "Western (Mac Mail)",
 798     "Simplified Chinese (HZ GB 2312)",
 799     "Traditional Chinese (Big 5 HKSCS)",
 800     NULL,
 801     "Ukrainian (KOI8-U)",
 802     "Traditional Chinese (Big 5-E)",
 803     NULL,
 804     "Western (NextStep)",
 805     "Western (EBCDIC Latin 1)",
 806 };
 807 #endif /* DEPLOYMENT_TARGET_MACOSX */
 808
 809 __private_extern__ CFStringEncoding __CFStringEncodingGetMostCompatibleMacScript(CFStringEncoding encoding) {
 810 #if DEPLOYMENT_TARGET_MACOSX
 811     switch (encoding & 0x0F00) {
 812         case 0: return encoding & 0xFF; break; // Mac scripts
 813
 814         case 0x0100: return kCFStringEncodingUnicode; break; // Unicode
 815
 816         case 0x200: // ISO 8859
 817             return (((encoding & 0xFF) <= (sizeof(__CFISO8859SimilarScriptList) / sizeof(*__CFISO8859SimilarScriptList))) ? __CFISO8859SimilarScriptList[(encoding & 0xFF) - 1] : kCFStringEncodingInvalidId);
 818             break;
 819
 820         default: {
 821             CFIndex index = __CFGetEncodingIndex(encoding);
 822
 823             if (kCFNotFound != index) {
 824                 index -= __CFGetEncodingIndex(kCFStringEncodingDOSLatinUS);
 825                 return __CFOtherSimilarScriptList[index];
 826             }
 827         }
 828     }
 829 #endif /* DEPLOYMENT_TARGET_MACOSX */
 830
 831     return kCFStringEncodingInvalidId;
 832 }
 833
 834 __private_extern__ const char *__CFStringEncodingGetName(CFStringEncoding encoding) {
 835     switch (encoding) {
 836         case kCFStringEncodingUTF8: return "Unicode (UTF-8)"; break;
 837         case kCFStringEncodingUTF16: return "Unicode (UTF-16)"; break;
 838         case kCFStringEncodingUTF16BE: return "Unicode (UTF-16BE)"; break;
 839         case kCFStringEncodingUTF16LE: return "Unicode (UTF-16LE)"; break;
 840         case kCFStringEncodingUTF32: return "Unicode (UTF-32)"; break;
 841         case kCFStringEncodingUTF32BE: return "Unicode (UTF-32BE)"; break;
 842         case kCFStringEncodingUTF32LE: return "Unicode (UTF-32LE)"; break;
 843         case kCFStringEncodingNonLossyASCII: return "Non-lossy ASCII"; break;
 844         case kCFStringEncodingUTF7: return "Unicode (UTF-7)"; break;
 845     }
 846
 847 #if DEPLOYMENT_TARGET_MACOSX
 848     if (0x0200 == (encoding & 0x0F00)) {
 849         encoding &= 0x00FF;
 850
 851         if (encoding <= (sizeof(__CFISONameList) / sizeof(*__CFISONameList))) return __CFISONameList[encoding - 1];
 852     } else {
 853         CFIndex index = __CFGetEncodingIndex(encoding);
 854
 855         if (kCFNotFound != index) return __CFOtherNameList[index];
 856     }
 857 #endif /* DEPLOYMENT_TARGET_MACOSX */
 858
 859     return NULL;
 860 }