sal/textenc/tables.cxx

   1 /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
   2 /*************************************************************************
   3  *
   4  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5  *
   6  * Copyright 2000, 2010 Oracle and/or its affiliates.
   7  *
   8  * OpenOffice.org - a multi-platform office productivity suite
   9  *
  10  * This file is part of OpenOffice.org.
  11  *
  12  * OpenOffice.org is free software: you can redistribute it and/or modify
  13  * it under the terms of the GNU Lesser General Public License version 3
  14  * only, as published by the Free Software Foundation.
  15  *
  16  * OpenOffice.org is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19  * GNU Lesser General Public License version 3 for more details
  20  * (a copy is included in the LICENSE file that accompanied this code).
  21  *
  22  * You should have received a copy of the GNU Lesser General Public License
  23  * version 3 along with OpenOffice.org.  If not, see
  24  * <http://www.openoffice.org/license.html>
  25  * for a copy of the LGPLv3 License.
  26  *
  27  ************************************************************************/
  28
  29 #include "sal/config.h"
  30
  31 #include <cassert>
  32 #include <cstddef>
  33
  34 #include "rtl/textenc.h"
  35 #include "sal/types.h"
  36
  37 #define NOTABUNI_START 0xFF
  38 #define NOTABUNI_END 0x00
  39
  40 #define NOTABCHAR_START 0xFFFF
  41 #define NOTABCHAR_END 0x0000
  42
  43 #define SAME8090UNI_START 0x80
  44 #define SAME8090UNI_END 0x9F
  45 static sal_uInt16 const aImpl8090SameToUniTab[SAME8090UNI_END
  46                                                   - SAME8090UNI_START
  47                                                   + 1]
  48 = { 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, /* 0x80 */
  49     0x0088, 0x0089, 0x008A, 0x008B, 0x008C, 0x008D, 0x008E, 0x008F,
  50     0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, /* 0x90 */
  51     0x0098, 0x0099, 0x009A, 0x009B, 0x009C, 0x009D, 0x009E, 0x009F };
  52
  53 #define SAME8090CHAR_START 0x0080
  54 #define SAME8090CHAR_END 0x009F
  55 static sal_uChar const aImpl8090SameToCharTab[SAME8090CHAR_END
  56                                                   - SAME8090CHAR_START
  57                                                   + 1]
  58     = { 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 0x0080 */
  59         0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x8D, 0x8E, 0x8F,
  60         0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 0x0090 */
  61         0x98, 0x99, 0x9A, 0x9B, 0x9C, 0x9D, 0x9E, 0x9F };
  62
  63 #define SAMEA0FFCHAR_START 0x00A0
  64 #define SAMEA0FFCHAR_END 0x00FF
  65 static sal_uChar const aImplA0FFSameToCharTab[SAMEA0FFCHAR_END
  66                                                   - SAMEA0FFCHAR_START
  67                                                   + 1]
  68     = { 0xA0, 0xA1, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, /* 0x00A0 */
  69         0xA8, 0xA9, 0xAA, 0xAB, 0xAC, 0xAD, 0xAE, 0xAF,
  70         0xB0, 0xB1, 0xB2, 0xB3, 0xB4, 0xB5, 0xB6, 0xB7, /* 0x00B0 */
  71         0xB8, 0xB9, 0xBA, 0xBB, 0xBC, 0xBD, 0xBE, 0xBF,
  72         0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, /* 0x00C0 */
  73         0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF,
  74         0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, /* 0x00D0 */
  75         0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF,
  76         0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, /* 0x00E0 */
  77         0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF,
  78         0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, /* 0x00F0 */
  79         0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF };
  80
  81 static sal_uInt16 const aImplDoubleByteIdentifierTab[1] = { 0 };
  82
  83 #include "tcvtarb1.tab"
  84 #include "tcvteas1.tab"
  85 #include "tcvtest1.tab"
  86 #include "tcvtjp1.tab"
  87 #include "tcvtjp2.tab"
  88 #include "tcvtjp3.tab"
  89 #include "tcvtjp4.tab"
  90 #include "tcvtjp5.tab"
  91 #include "tcvtjp6.tab"
  92 #include "tcvtkr1.tab"
  93 #include "tcvtkr2.tab"
  94 #include "tcvtkr4.tab"
  95 #include "tcvtkr5.tab"
  96 #include "tcvtkr6.tab"
  97 #include "tcvtlat1.tab"
  98 #include "tcvtscn1.tab"
  99 #include "tcvtscn2.tab"
 100 #include "tcvtscn3.tab"
 101 #include "tcvtscn4.tab"
 102 #include "tcvtscn5.tab"
 103 #include "tcvtscn6.tab"
 104 #include "tcvtsym1.tab"
 105 #include "tcvttcn1.tab"
 106 #include "tcvttcn2.tab"
 107 #include "tcvttcn6.tab"
 108 #include "tcvtuni1.tab"
 109 #include "convertiscii.tab"
 110
 111 #include "convertbig5hkscs.tab"
 112 #include "converteuctw.tab"
 113 #include "convertgb18030.tab"
 114 #include "convertiso2022cn.tab"
 115 #include "convertiso2022jp.tab"
 116 #include "convertiso2022kr.tab"
 117 #include "convertadobe.tab"
 118
 119 extern "C" SAL_DLLPUBLIC_EXPORT ImplTextEncodingData const *
 120     sal_getFullTextEncodingData( rtl_TextEncoding nEncoding )
 121 {
 122     assert(
 123         nEncoding != RTL_TEXTENCODING_ASCII_US &&
 124         nEncoding != RTL_TEXTENCODING_ISO_8859_1 &&
 125         nEncoding != RTL_TEXTENCODING_JAVA_UTF8 &&
 126         nEncoding != RTL_TEXTENCODING_MS_1252 &&
 127         nEncoding != RTL_TEXTENCODING_UTF8);
 128         // handled by Impl_getTextEncodingData
 129     static ImplTextEncodingData const * const aData[]
 130         = { NULL, /* DONTKNOW */
 131             NULL, /* MS_1252, see above */
 132             &aImplAPPLEROMANTextEncodingData, /* APPLE_ROMAN */
 133             &aImplIBM437TextEncodingData, /* IBM_437 */
 134             &aImplIBM850TextEncodingData, /* IBM_850 */
 135             &aImplIBM860TextEncodingData, /* IBM_860 */
 136             &aImplIBM861TextEncodingData, /* IBM_861 */
 137             &aImplIBM863TextEncodingData, /* IBM_863 */
 138             &aImplIBM865TextEncodingData, /* IBM_865 */
 139             NULL, /* reserved (SYSTEM) */
 140             &aImplSYMBOLTextEncodingData, /* SYMBOL */
 141             NULL, /* ASCII_US, see above */
 142             NULL, /* ISO_8859_1, see above */
 143             &aImplISO88592TextEncodingData, /* ISO_8859_2 */
 144             &aImplISO88593TextEncodingData, /* ISO_8859_3 */
 145             &aImplISO88594TextEncodingData, /* ISO_8859_4 */
 146             &aImplISO88595TextEncodingData, /* ISO_8859_5 */
 147             &aImplISO88596TextEncodingData, /* ISO_8859_6 */
 148             &aImplISO88597TextEncodingData, /* ISO_8859_7 */
 149             &aImplISO88598TextEncodingData, /* ISO_8859_8 */
 150             &aImplISO88599TextEncodingData, /* ISO_8859_9 */
 151             &aImplISO885914TextEncodingData, /* ISO_8859_14 */
 152             &aImplISO885915TextEncodingData, /* ISO_8859_15 */
 153             &aImplIBM737TextEncodingData, /* IBM_737 */
 154             &aImplIBM775TextEncodingData, /* IBM_775 */
 155             &aImplIBM852TextEncodingData, /* IBM_852 */
 156             &aImplIBM855TextEncodingData, /* IBM_855 */
 157             &aImplIBM857TextEncodingData, /* IBM_857 */
 158             &aImplIBM862TextEncodingData, /* IBM_862 */
 159             &aImplIBM864TextEncodingData, /* IBM_864 */
 160             &aImplIBM866TextEncodingData, /* IBM_866 */
 161             &aImplIBM869TextEncodingData, /* IBM_869 */
 162             &aImplMS874TextEncodingData, /* MS_874 */
 163             &aImplMS1250TextEncodingData, /* MS_1250 */
 164             &aImplMS1251TextEncodingData, /* MS_1251 */
 165             &aImplMS1253TextEncodingData, /* MS_1253 */
 166             &aImplMS1254TextEncodingData, /* MS_1254 */
 167             &aImplMS1255TextEncodingData, /* MS_1255 */
 168             &aImplMS1256TextEncodingData, /* MS_1256 */
 169             &aImplMS1257TextEncodingData, /* MS_1257 */
 170             &aImplMS1258TextEncodingData, /* MS_1258 */
 171             NULL, /* TODO! APPLE_ARABIC */
 172             &aImplAPPLECENTEUROTextEncodingData, /* APPLE_CENTEURO */
 173             &aImplAPPLECROATIANTextEncodingData, /* APPLE_CROATIAN */
 174             &aImplAPPLECYRILLICTextEncodingData, /* APPLE_CYRILLIC */
 175             NULL, /* TODO! APPLE_DEVANAGARI */
 176             NULL, /* TODO! APPLE_FARSI */
 177             &aImplAPPLEGREEKTextEncodingData, /* APPLE_GREEK */
 178             NULL, /* TODO! APPLE_GUJARATI */
 179             NULL, /* TODO! APPLE_GURMUKHI */
 180             NULL, /* TODO! APPLE_HEBREW */
 181             &aImplAPPLEICELANDTextEncodingData, /* APPLE_ICELAND */
 182             &aImplAPPLEROMANIANTextEncodingData, /* APPLE_ROMANIAN */
 183             NULL, /* TODO! APPLE_THAI */
 184             &aImplAPPLETURKISHTextEncodingData, /* APPLE_TURKISH */
 185             &aImplAPPLEUKRAINIANTextEncodingData, /* APPLE_UKRAINIAN */
 186             &aImplAPPLECHINSIMPTextEncodingData, /* APPLE_CHINSIMP */
 187             &aImplAPPLECHINTRADTextEncodingData, /* APPLE_CHINTRAD */
 188             &aImplAPPLEJAPANESETextEncodingData, /* APPLE_JAPANESE */
 189             &aImplAPPLEKOREANTextEncodingData, /* APPLE_KOREAN */
 190             &aImplMS932TextEncodingData, /* MS_932 */
 191             &aImplMS936TextEncodingData, /* MS_936 */
 192             &aImplMS949TextEncodingData, /* MS_949 */
 193             &aImplMS950TextEncodingData, /* MS_950 */
 194             &aImplSJISTextEncodingData, /* SHIFT_JIS */
 195             &aImplGB2312TextEncodingData, /* GB_2312 */
 196             &aImplGBT12345TextEncodingData, /* GBT_12345 */
 197             &aImplGBKTextEncodingData, /* GBK */
 198             &aImplBIG5TextEncodingData, /* BIG5 */
 199             &aImplEUCJPTextEncodingData, /* EUC_JP */
 200             &aImplEUCCNTextEncodingData, /* EUC_CN */
 201             &aImplEucTwTextEncodingData, /* EUC_TW */
 202             &aImplIso2022JpTextEncodingData, /* ISO_2022_JP */
 203             &aImplIso2022CnTextEncodingData, /* ISO_2022_CN */
 204             &aImplKOI8RTextEncodingData, /* KOI8_R */
 205             &aImplUTF7TextEncodingData, /* UTF7 */
 206             NULL, /* UTF8, see above */
 207             &aImplISO885910TextEncodingData, /* ISO_8859_10 */
 208             &aImplISO885913TextEncodingData, /* ISO_8859_13 */
 209             &aImplEUCKRTextEncodingData, /* EUC_KR */
 210             &aImplIso2022KrTextEncodingData, /* ISO_2022_KR */
 211             &aImplJISX0201TextEncodingData, /* JIS_X_0201 */
 212             &aImplJISX0208TextEncodingData, /* JIS_X_0208 */
 213             &aImplJISX0212TextEncodingData, /* JIS_X_0212 */
 214             &aImplMS1361TextEncodingData, /* MS_1361 */
 215             &aImplGb18030TextEncodingData, /* GB_18030 */
 216             &aImplBig5HkscsTextEncodingData, /* BIG5_HKSCS */
 217             &aImplTis620TextEncodingData, /* TIS_620 */
 218             &aImplKoi8UTextEncodingData, /* KOI8_U */
 219             &aImplIsciiDevanagariTextEncodingData, /* ISCII_DEVANAGARI */
 220             NULL, /* JAVA_UTF8, see above */
 221             &adobeStandardEncodingData, /* ADOBE_STANDARD */
 222             &adobeSymbolEncodingData, /* ADOBE_SYMBOL */
 223             &aImplPT154TextEncodingData }; /* PT154 */
 224     return
 225         nEncoding < SAL_N_ELEMENTS(aData) ? aData[nEncoding] : NULL;
 226 }
 227
 228 /* vim:set shiftwidth=4 softtabstop=4 expandtab: */