lib/gb18030ext.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  * GB18030 two-byte extension
  23  */
  24
  25 static const unsigned short gb18030ext_2uni_pagea9[13] = {
  26   /* 0xa9 */
  27   0x303e, 0x2ff0, 0x2ff1, 0x2ff2, 0x2ff3, 0x2ff4, 0x2ff5, 0x2ff6,
  28   0x2ff7, 0x2ff8, 0x2ff9, 0x2ffa, 0x2ffb,
  29 };
  30 static const unsigned short gb18030ext_2uni_pagefe[95] = {
  31   /* 0xfe */
  32   0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
  33   0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
  34   0x2e81, 0xfffd, 0xfffd, 0xfffd, 0x2e84, 0x3473, 0x3447, 0x2e88,
  35   0x2e8b, 0xfffd, 0x359e, 0x361a, 0x360e, 0x2e8c, 0x2e97, 0x396e,
  36   0x3918, 0xfffd, 0x39cf, 0x39df, 0x3a73, 0x39d0, 0xfffd, 0xfffd,
  37   0x3b4e, 0x3c6e, 0x3ce0, 0x2ea7, 0xfffd, 0xfffd, 0x2eaa, 0x4056,
  38   0x415f, 0x2eae, 0x4337, 0x2eb3, 0x2eb6, 0x2eb7, 0xfffd, 0x43b1,
  39   0x43ac, 0x2ebb, 0x43dd, 0x44d6, 0x4661, 0x464c, 0xfffd, 0x4723,
  40   0x4729, 0x477c, 0x478d, 0x2eca, 0x4947, 0x497a, 0x497d, 0x4982,
  41   0x4983, 0x4985, 0x4986, 0x499f, 0x499b, 0x49b7, 0x49b6, 0xfffd,
  42   0xfffd, 0x4ca3, 0x4c9f, 0x4ca0, 0x4ca1, 0x4c77, 0x4ca2, 0x4d13,
  43   0x4d14, 0x4d15, 0x4d16, 0x4d17, 0x4d18, 0x4d19, 0x4dae,
  44 };
  45
  46 static int
  47 gb18030ext_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n)
  48 {
  49   unsigned char c1 = s[0];
  50   if ((c1 == 0xa2) || (c1 >= 0xa8 && c1 <= 0xa9) || (c1 == 0xfe)) {
  51     if (n >= 2) {
  52       unsigned char c2 = s[1];
  53       if ((c2 >= 0x40 && c2 < 0x7f) || (c2 >= 0x80 && c2 < 0xff)) {
  54         unsigned int i = 190 * (c1 - 0x81) + (c2 - (c2 >= 0x80 ? 0x41 : 0x40));
  55         unsigned short wc = 0xfffd;
  56         if (i < 7410) {
  57           if (i == 6432)
  58             wc = 0x20ac;
  59         } else if (i < 23750) {
  60           if (i == 7536)
  61             wc = 0x01f9;
  62           else if (i >= 7672 && i < 7685)
  63             wc = gb18030ext_2uni_pagea9[i-7672];
  64         } else {
  65           if (i < 23845)
  66             wc = gb18030ext_2uni_pagefe[i-23750];
  67         }
  68         if (wc != 0xfffd) {
  69           *pwc = (ucs4_t) wc;
  70           return 2;
  71         }
  72       }
  73       return RET_ILSEQ;
  74     }
  75     return RET_TOOFEW(0);
  76   }
  77   return RET_ILSEQ;
  78 }
  79
  80 static const unsigned short gb18030ext_page2e[80] = {
  81   0x0000, 0xfe50, 0x0000, 0x0000, 0xfe54, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
  82   0xfe57, 0x0000, 0x0000, 0xfe58, 0xfe5d, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
  83   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe5e, /*0x90-0x97*/
  84   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x98-0x9f*/
  85   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe6b, /*0xa0-0xa7*/
  86   0x0000, 0x0000, 0xfe6e, 0x0000, 0x0000, 0x0000, 0xfe71, 0x0000, /*0xa8-0xaf*/
  87   0x0000, 0x0000, 0x0000, 0xfe73, 0x0000, 0x0000, 0xfe74, 0xfe75, /*0xb0-0xb7*/
  88   0x0000, 0x0000, 0x0000, 0xfe79, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
  89   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc0-0xc7*/
  90   0x0000, 0x0000, 0xfe84, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc8-0xcf*/
  91 };
  92 static const unsigned short gb18030ext_page2f[16] = {
  93   0xa98a, 0xa98b, 0xa98c, 0xa98d, 0xa98e, 0xa98f, 0xa990, 0xa991, /*0xf0-0xf7*/
  94   0xa992, 0xa993, 0xa994, 0xa995, 0x0000, 0x0000, 0x0000, 0x0000, /*0xf8-0xff*/
  95 };
  96 static const unsigned short gb18030ext_page34[56] = {
  97   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe56, /*0x40-0x47*/
  98   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
  99   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
 100   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
 101   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
 102   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x68-0x6f*/
 103   0x0000, 0x0000, 0x0000, 0xfe55, 0x0000, 0x0000, 0x0000, 0x0000, /*0x70-0x77*/
 104 };
 105 static const unsigned short gb18030ext_page36[24] = {
 106   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe5c, 0x0000, /*0x08-0x0f*/
 107   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x10-0x17*/
 108   0x0000, 0x0000, 0xfe5b, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
 109 };
 110 static const unsigned short gb18030ext_page39[24] = {
 111   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe62, /*0xc8-0xcf*/
 112   0xfe65, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xd0-0xd7*/
 113   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe63, /*0xd8-0xdf*/
 114 };
 115 static const unsigned short gb18030ext_page43[56] = {
 116   0x0000, 0x0000, 0x0000, 0x0000, 0xfe78, 0x0000, 0x0000, 0x0000, /*0xa8-0xaf*/
 117   0x0000, 0xfe77, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb0-0xb7*/
 118   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
 119   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc0-0xc7*/
 120   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc8-0xcf*/
 121   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xd0-0xd7*/
 122   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe7a, 0x0000, 0x0000, /*0xd8-0xdf*/
 123 };
 124 static const unsigned short gb18030ext_page46[32] = {
 125   0x0000, 0x0000, 0x0000, 0x0000, 0xfe7d, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
 126   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
 127   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
 128   0x0000, 0xfe7c, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
 129 };
 130 static const unsigned short gb18030ext_page47_1[16] = {
 131   0x0000, 0x0000, 0x0000, 0xfe80, 0x0000, 0x0000, 0x0000, 0x0000, /*0x20-0x27*/
 132   0x0000, 0xfe81, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x28-0x2f*/
 133 };
 134 static const unsigned short gb18030ext_page47_2[24] = {
 135   0x0000, 0x0000, 0x0000, 0x0000, 0xfe82, 0x0000, 0x0000, 0x0000, /*0x78-0x7f*/
 136   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
 137   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe83, 0x0000, 0x0000, /*0x88-0x8f*/
 138 };
 139 static const unsigned short gb18030ext_page49[120] = {
 140   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe85, /*0x40-0x47*/
 141   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
 142   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
 143   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
 144   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
 145   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x68-0x6f*/
 146   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x70-0x77*/
 147   0x0000, 0x0000, 0xfe86, 0x0000, 0x0000, 0xfe87, 0x0000, 0x0000, /*0x78-0x7f*/
 148   0x0000, 0x0000, 0xfe88, 0xfe89, 0x0000, 0xfe8a, 0xfe8b, 0x0000, /*0x80-0x87*/
 149   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
 150   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x90-0x97*/
 151   0x0000, 0x0000, 0x0000, 0xfe8d, 0x0000, 0x0000, 0x0000, 0xfe8c, /*0x98-0x9f*/
 152   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa0-0xa7*/
 153   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa8-0xaf*/
 154   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe8f, 0xfe8e, /*0xb0-0xb7*/
 155 };
 156 static const unsigned short gb18030ext_page4c[56] = {
 157   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe96, /*0x70-0x77*/
 158   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x78-0x7f*/
 159   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
 160   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
 161   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x90-0x97*/
 162   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe93, /*0x98-0x9f*/
 163   0xfe94, 0xfe95, 0xfe97, 0xfe92, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa0-0xa7*/
 164 };
 165 static const unsigned short gb18030ext_page4d[16] = {
 166   0x0000, 0x0000, 0x0000, 0xfe98, 0xfe99, 0xfe9a, 0xfe9b, 0xfe9c, /*0x10-0x17*/
 167   0xfe9d, 0xfe9e, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
 168 };
 169
 170 static int
 171 gb18030ext_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n)
 172 {
 173   if (n >= 2) {
 174     unsigned short c = 0;
 175     if (wc == 0x01f9)
 176       c = 0xa8bf;
 177     else if (wc == 0x20ac)
 178       c = 0xa2e3;
 179     else if (wc >= 0x2e80 && wc < 0x2ed0)
 180       c = gb18030ext_page2e[wc-0x2e80];
 181     else if (wc >= 0x2ff0 && wc < 0x3000)
 182       c = gb18030ext_page2f[wc-0x2ff0];
 183     else if (wc == 0x303e)
 184       c = 0xa989;
 185     else if (wc >= 0x3440 && wc < 0x3478)
 186       c = gb18030ext_page34[wc-0x3440];
 187     else if (wc == 0x359e)
 188       c = 0xfe5a;
 189     else if (wc >= 0x3608 && wc < 0x3620)
 190       c = gb18030ext_page36[wc-0x3608];
 191     else if (wc == 0x3918)
 192       c = 0xfe60;
 193     else if (wc == 0x396e)
 194       c = 0xfe5f;
 195     else if (wc >= 0x39c8 && wc < 0x39e0)
 196       c = gb18030ext_page39[wc-0x39c8];
 197     else if (wc == 0x3a73)
 198       c = 0xfe64;
 199     else if (wc == 0x3b4e)
 200       c = 0xfe68;
 201     else if (wc == 0x3c6e)
 202       c = 0xfe69;
 203     else if (wc == 0x3ce0)
 204       c = 0xfe6a;
 205     else if (wc == 0x4056)
 206       c = 0xfe6f;
 207     else if (wc == 0x415f)
 208       c = 0xfe70;
 209     else if (wc == 0x4337)
 210       c = 0xfe72;
 211     else if (wc >= 0x43a8 && wc < 0x43e0)
 212       c = gb18030ext_page43[wc-0x43a8];
 213     else if (wc == 0x44d6)
 214       c = 0xfe7b;
 215     else if (wc >= 0x4648 && wc < 0x4668)
 216       c = gb18030ext_page46[wc-0x4648];
 217     else if (wc >= 0x4720 && wc < 0x4730)
 218       c = gb18030ext_page47_1[wc-0x4720];
 219     else if (wc >= 0x4778 && wc < 0x4790)
 220       c = gb18030ext_page47_2[wc-0x4778];
 221     else if (wc >= 0x4940 && wc < 0x49b8)
 222       c = gb18030ext_page49[wc-0x4940];
 223     else if (wc >= 0x4c70 && wc < 0x4ca8)
 224       c = gb18030ext_page4c[wc-0x4c70];
 225     else if (wc >= 0x4d10 && wc < 0x4d20)
 226       c = gb18030ext_page4d[wc-0x4d10];
 227     else if (wc == 0x4dae)
 228       c = 0xfe9f;
 229     if (c != 0) {
 230       r[0] = (c >> 8); r[1] = (c & 0xff);
 231       return 2;
 232     }
 233     return RET_ILSEQ;
 234   }
 235   return RET_TOOSMALL;
 236 }