lib/cp950.h

   1 /*
   2  * Copyright (C) 1999-2001 Free Software Foundation, Inc.
   3  * This file is part of the GNU LIBICONV Library.
   4  *
   5  * The GNU LIBICONV Library is free software; you can redistribute it
   6  * and/or modify it under the terms of the GNU Library General Public
   7  * License as published by the Free Software Foundation; either version 2
   8  * of the License, or (at your option) any later version.
   9  *
  10  * The GNU LIBICONV Library is distributed in the hope that it will be
  11  * useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * Library General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Library General Public
  16  * License along with the GNU LIBICONV Library; see the file COPYING.LIB.
  17  * If not, write to the Free Software Foundation, Inc., 59 Temple Place -
  18  * Suite 330, Boston, MA 02111-1307, USA.
  19  */
  20
  21 /*
  22  * CP950
  23  */
  24
  25 /*
  26  * Microsoft CP950 is a slightly extended and slightly modified version of
  27  * BIG5. The differences between the EASTASIA/OTHER/BIG5.TXT and
  28  * VENDORS/MICSFT/WINDOWS/CP950.TXT tables found on ftp.unicode.org are
  29  * as follows:
  30  *
  31  * 1. Some characters in the BIG5 range are defined differently:
  32  *
  33  *     code   BIG5.TXT                       CP950.TXT
  34  *    0xA145  0x2022 # BULLET                0x2027 # HYPHENATION POINT
  35  *    0xA14E  0xFF64 # HALFWIDTH IDEOGRAPHIC COMMA
  36  *                                           0xFE51 # SMALL IDEOGRAPHIC COMMA
  37  *    0xA15A    ---                          0x2574 # BOX DRAWINGS LIGHT LEFT
  38  *    0xA1C2  0x203E # OVERLINE              0x00AF # MACRON
  39  *    0xA1C3    ---                          0xFFE3 # FULLWIDTH MACRON
  40  *    0xA1C5    ---                          0x02CD # MODIFIER LETTER LOW MACRON
  41  *    0xA1E3  0x223C # TILDE OPERATOR        0xFF5E # FULLWIDTH TILDE
  42  *    0xA1F2  0x2641 # EARTH                 0x2295 # CIRCLED PLUS
  43  *    0xA1F3  0x2609 # SUN                   0x2299 # CIRCLED DOT OPERATOR
  44  *    0xA1FE    ---                          0xFF0F # FULLWIDTH SOLIDUS
  45  *    0xA240    ---                          0xFF3C # FULLWIDTH REVERSE SOLIDUS
  46  *    0xA241  0xFF0F # FULLWIDTH SOLIDUS     0x2215 # DIVISION SLASH
  47  *    0xA242  0xFF3C # FULLWIDTH REVERSE SOLIDUS
  48  *                                           0xFE68 # SMALL REVERSE SOLIDUS
  49  *    0xA244  0x00A5 # YEN SIGN              0xFFE5 # FULLWIDTH YEN SIGN
  50  *    0xA246  0x00A2 # CENT SIGN             0xFFE0 # FULLWIDTH CENT SIGN
  51  *    0xA247  0x00A3 # POUND SIGN            0xFFE1 # FULLWIDTH POUND SIGN
  52  *    0xA2CC    ---                          0x5341
  53  *    0xA2CE    ---                          0x5345
  54  *
  55  * 2. A small new row. See cp950ext.h.
  56  *
  57  * 3. CP950.TXT is lacking the range 0xC6A1..0xC7FC (Hiragana, Katakana,
  58  *    Cyrillic, circled digits, parenthesized digits).
  59  *
  60  *    We implement this omission, because said range is marked "uncertain"
  61  *    in the unicode.org BIG5 table.
  62  */
  63
  64 static const unsigned short cp950_2uni_pagea1[314] = {
  65   /* 0xa1 */
  66   0x3000, 0xff0c, 0x3001, 0x3002, 0xff0e, 0x2027, 0xff1b, 0xff1a,
  67   0xff1f, 0xff01, 0xfe30, 0x2026, 0x2025, 0xfe50, 0xfe51, 0xfe52,
  68   0x00b7, 0xfe54, 0xfe55, 0xfe56, 0xfe57, 0xff5c, 0x2013, 0xfe31,
  69   0x2014, 0xfe33, 0x2574, 0xfe34, 0xfe4f, 0xff08, 0xff09, 0xfe35,
  70   0xfe36, 0xff5b, 0xff5d, 0xfe37, 0xfe38, 0x3014, 0x3015, 0xfe39,
  71   0xfe3a, 0x3010, 0x3011, 0xfe3b, 0xfe3c, 0x300a, 0x300b, 0xfe3d,
  72   0xfe3e, 0x3008, 0x3009, 0xfe3f, 0xfe40, 0x300c, 0x300d, 0xfe41,
  73   0xfe42, 0x300e, 0x300f, 0xfe43, 0xfe44, 0xfe59, 0xfe5a, 0xfe5b,
  74   0xfe5c, 0xfe5d, 0xfe5e, 0x2018, 0x2019, 0x201c, 0x201d, 0x301d,
  75   0x301e, 0x2035, 0x2032, 0xff03, 0xff06, 0xff0a, 0x203b, 0x00a7,
  76   0x3003, 0x25cb, 0x25cf, 0x25b3, 0x25b2, 0x25ce, 0x2606, 0x2605,
  77   0x25c7, 0x25c6, 0x25a1, 0x25a0, 0x25bd, 0x25bc, 0x32a3, 0x2105,
  78   0x00af, 0xffe3, 0xff3f, 0x02cd, 0xfe49, 0xfe4a, 0xfe4d, 0xfe4e,
  79   0xfe4b, 0xfe4c, 0xfe5f, 0xfe60, 0xfe61, 0xff0b, 0xff0d, 0x00d7,
  80   0x00f7, 0x00b1, 0x221a, 0xff1c, 0xff1e, 0xff1d, 0x2266, 0x2267,
  81   0x2260, 0x221e, 0x2252, 0x2261, 0xfe62, 0xfe63, 0xfe64, 0xfe65,
  82   0xfe66, 0xff5e, 0x2229, 0x222a, 0x22a5, 0x2220, 0x221f, 0x22bf,
  83   0x33d2, 0x33d1, 0x222b, 0x222e, 0x2235, 0x2234, 0x2640, 0x2642,
  84   0x2295, 0x2299, 0x2191, 0x2193, 0x2190, 0x2192, 0x2196, 0x2197,
  85   0x2199, 0x2198, 0x2225, 0x2223, 0xff0f,
  86   /* 0xa2 */
  87   0xff3c, 0x2215, 0xfe68, 0xff04, 0xffe5, 0x3012, 0xffe0, 0xffe1,
  88   0xff05, 0xff20, 0x2103, 0x2109, 0xfe69, 0xfe6a, 0xfe6b, 0x33d5,
  89   0x339c, 0x339d, 0x339e, 0x33ce, 0x33a1, 0x338e, 0x338f, 0x33c4,
  90   0x00b0, 0x5159, 0x515b, 0x515e, 0x515d, 0x5161, 0x5163, 0x55e7,
  91   0x74e9, 0x7cce, 0x2581, 0x2582, 0x2583, 0x2584, 0x2585, 0x2586,
  92   0x2587, 0x2588, 0x258f, 0x258e, 0x258d, 0x258c, 0x258b, 0x258a,
  93   0x2589, 0x253c, 0x2534, 0x252c, 0x2524, 0x251c, 0x2594, 0x2500,
  94   0x2502, 0x2595, 0x250c, 0x2510, 0x2514, 0x2518, 0x256d, 0x256e,
  95   0x2570, 0x256f, 0x2550, 0x255e, 0x256a, 0x2561, 0x25e2, 0x25e3,
  96   0x25e5, 0x25e4, 0x2571, 0x2572, 0x2573, 0xff10, 0xff11, 0xff12,
  97   0xff13, 0xff14, 0xff15, 0xff16, 0xff17, 0xff18, 0xff19, 0x2160,
  98   0x2161, 0x2162, 0x2163, 0x2164, 0x2165, 0x2166, 0x2167, 0x2168,
  99   0x2169, 0x3021, 0x3022, 0x3023, 0x3024, 0x3025, 0x3026, 0x3027,
 100   0x3028, 0x3029, 0x5341, 0x5344, 0x5345, 0xff21, 0xff22, 0xff23,
 101   0xff24, 0xff25, 0xff26, 0xff27, 0xff28, 0xff29, 0xff2a, 0xff2b,
 102   0xff2c, 0xff2d, 0xff2e, 0xff2f, 0xff30, 0xff31, 0xff32, 0xff33,
 103   0xff34, 0xff35, 0xff36, 0xff37, 0xff38, 0xff39, 0xff3a, 0xff41,
 104   0xff42, 0xff43, 0xff44, 0xff45, 0xff46, 0xff47, 0xff48, 0xff49,
 105   0xff4a, 0xff4b, 0xff4c, 0xff4d, 0xff4e, 0xff4f, 0xff50, 0xff51,
 106   0xff52, 0xff53, 0xff54, 0xff55, 0xff56,
 107 };
 108
 109 #include "cp950ext.h"
 110
 111 static int
 112 cp950_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n)
 113 {
 114   unsigned char c = *s;
 115   /* Code set 0 (ASCII) */
 116   if (c < 0x80)
 117     return ascii_mbtowc(conv,pwc,s,n);
 118   /* Code set 1 (BIG5 extended) */
 119   if (c >= 0xa1 && c < 0xff) {
 120     if (n < 2)
 121       return RET_TOOFEW(0);
 122     {
 123       unsigned char c2 = s[1];
 124       if ((c2 >= 0x40 && c2 < 0x7f) || (c2 >= 0xa1 && c2 < 0xff)) {
 125         if (c < 0xa3) {
 126           unsigned int i = 157 * (c - 0xa1) + (c2 - (c2 >= 0xa1 ? 0x62 : 0x40));
 127           unsigned short wc = cp950_2uni_pagea1[i];
 128           if (wc != 0xfffd) {
 129             *pwc = (ucs4_t) wc;
 130             return 2;
 131           }
 132         }
 133         if (!((c == 0xc6 && c2 >= 0xa1) || c == 0xc7)) {
 134           int ret = big5_mbtowc(conv,pwc,s,2);
 135           if (ret != RET_ILSEQ)
 136             return ret;
 137         }
 138       }
 139     }
 140     if (c == 0xf9) {
 141       int ret = cp950ext_mbtowc(conv,pwc,s,2);
 142       if (ret != RET_ILSEQ)
 143         return ret;
 144     }
 145   }
 146   return RET_ILSEQ;
 147 }
 148
 149 static int
 150 cp950_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n)
 151 {
 152   unsigned char buf[2];
 153   int ret;
 154
 155   /* Code set 0 (ASCII) */
 156   ret = ascii_wctomb(conv,r,wc,n);
 157   if (ret != RET_ILUNI)
 158     return ret;
 159
 160   /* Code set 1 (BIG5 extended) */
 161   switch (wc >> 8) {
 162     case 0x00:
 163       if (wc == 0x00af) { buf[0] = 0xa1; buf[1] = 0xc2; ret = 2; break; }
 164       if (wc == 0x00a2 || wc == 0x00a3 || wc == 0x00a4)
 165         return RET_ILUNI;
 166       break;
 167     case 0x02:
 168       if (wc == 0x02cd) { buf[0] = 0xa1; buf[1] = 0xc5; ret = 2; break; }
 169       break;
 170     case 0x20:
 171       if (wc == 0x2027) { buf[0] = 0xa1; buf[1] = 0x45; ret = 2; break; }
 172       if (wc == 0x2022 || wc == 0x203e)
 173         return RET_ILUNI;
 174       break;
 175     case 0x22:
 176       if (wc == 0x2215) { buf[0] = 0xa2; buf[1] = 0x41; ret = 2; break; }
 177       if (wc == 0x2295) { buf[0] = 0xa1; buf[1] = 0xf2; ret = 2; break; }
 178       if (wc == 0x2299) { buf[0] = 0xa1; buf[1] = 0xf3; ret = 2; break; }
 179       if (wc == 0x223c)
 180         return RET_ILUNI;
 181       break;
 182     case 0x25:
 183       if (wc == 0x2574) { buf[0] = 0xa1; buf[1] = 0x5a; ret = 2; break; }
 184       break;
 185     case 0x26:
 186       if (wc == 0x2609 || wc == 0x2641)
 187         return RET_ILUNI;
 188       break;
 189     case 0xfe:
 190       if (wc == 0xfe51) { buf[0] = 0xa1; buf[1] = 0x4e; ret = 2; break; }
 191       if (wc == 0xfe68) { buf[0] = 0xa2; buf[1] = 0x42; ret = 2; break; }
 192       break;
 193     case 0xff:
 194       if (wc == 0xff0f) { buf[0] = 0xa1; buf[1] = 0xfe; ret = 2; break; }
 195       if (wc == 0xff3c) { buf[0] = 0xa2; buf[1] = 0x40; ret = 2; break; }
 196       if (wc == 0xff5e) { buf[0] = 0xa1; buf[1] = 0xe3; ret = 2; break; }
 197       if (wc == 0xffe0) { buf[0] = 0xa2; buf[1] = 0x46; ret = 2; break; }
 198       if (wc == 0xffe1) { buf[0] = 0xa2; buf[1] = 0x47; ret = 2; break; }
 199       if (wc == 0xffe3) { buf[0] = 0xa1; buf[1] = 0xc3; ret = 2; break; }
 200       if (wc == 0xffe5) { buf[0] = 0xa2; buf[1] = 0x44; ret = 2; break; }
 201       if (wc == 0xff64)
 202         return RET_ILUNI;
 203       break;
 204   }
 205   if (ret == RET_ILUNI)
 206     ret = big5_wctomb(conv,buf,wc,2);
 207   if (ret != RET_ILUNI) {
 208     if (ret != 2) abort();
 209     if (!((buf[0] == 0xc6 && buf[1] >= 0xa1) || buf[0] == 0xc7)) {
 210       if (n < 2)
 211         return RET_TOOSMALL;
 212       r[0] = buf[0];
 213       r[1] = buf[1];
 214       return 2;
 215     }
 216   }
 217   ret = cp950ext_wctomb(conv,buf,wc,2);
 218   if (ret != RET_ILUNI) {
 219     if (ret != 2) abort();
 220     if (n < 2)
 221       return RET_TOOSMALL;
 222     r[0] = buf[0];
 223     r[1] = buf[1];
 224     return 2;
 225   }
 226
 227   return RET_ILUNI;
 228 }