release/src/router/libiconv/lib/gb18030ext.h

   1 /*
   2  * Copyright (C) 1999-2001, 2005, 2011 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., 51 Franklin Street,
  18  * Fifth Floor, Boston, MA 02110-1301, 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 int gb18030ext_2uni_pagefe[96] = {
  31   /* 0xfe */
  32    0xfffd,  0xfffd,  0xfffd,  0xfffd,  0xfffd,  0xfffd,  0xfffd,  0xfffd,
  33    0xfffd,  0xfffd,  0xfffd,  0xfffd,  0xfffd,  0xfffd,  0xfffd,  0xfffd,
  34    0x2e81, 0x20087, 0x20089, 0x200cc,  0x2e84,  0x3473,  0x3447,  0x2e88,
  35    0x2e8b,  0x9fb4,  0x359e,  0x361a,  0x360e,  0x2e8c,  0x2e97,  0x396e,
  36    0x3918,  0x9fb5,  0x39cf,  0x39df,  0x3a73,  0x39d0,  0x9fb6,  0x9fb7,
  37    0x3b4e,  0x3c6e,  0x3ce0,  0x2ea7, 0x215d7,  0x9fb8,  0x2eaa,  0x4056,
  38    0x415f,  0x2eae,  0x4337,  0x2eb3,  0x2eb6,  0x2eb7, 0x2298f,  0x43b1,
  39    0x43ac,  0x2ebb,  0x43dd,  0x44d6,  0x4661,  0x464c,  0x9fb9,  0x4723,
  40    0x4729,  0x477c,  0x478d,  0x2eca,  0x4947,  0x497a,  0x497d,  0x4982,
  41    0x4983,  0x4985,  0x4986,  0x499f,  0x499b,  0x49b7,  0x49b6,  0x9fba,
  42   0x241fe,  0x4ca3,  0x4c9f,  0x4ca0,  0x4ca1,  0x4c77,  0x4ca2,  0x4d13,
  43    0x4d14,  0x4d15,  0x4d16,  0x4d17,  0x4d18,  0x4d19,  0x4dae,  0x9fbb,
  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 >= 0xa4 && c1 <= 0xa9) || (c1 == 0xd7) || (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 int wc = 0xfffd;
  56         switch (c1) {
  57           case 0xa2:
  58             if (i >= 6376 && i <= 6381) /* 0xA2AB..0xA2B0 */
  59               wc = 0xe766 + (i - 6376);
  60             else if (i == 6432) /* 0xA2E3 */
  61               wc = 0x20ac;
  62             else if (i == 6433) /* 0xA2E4 */
  63               wc = 0xe76d;
  64             else if (i >= 6444 && i <= 6445) /* 0xA2EF..0xA2F0 */
  65               wc = 0xe76e + (i - 6444);
  66             else if (i >= 6458 && i <= 6459) /* 0xA2FD..0xA2FE */
  67               wc = 0xe770 + (i - 6458);
  68             break;
  69           case 0xa4:
  70             if (i >= 6829 && i <= 6839) /* 0xA4F4..0xA4FE */
  71               wc = 0xe772 + (i - 6829);
  72             break;
  73           case 0xa5:
  74             if (i >= 7022 && i <= 7029) /* 0xA5F7..0xA5FE */
  75               wc = 0xe77d + (i - 7022);
  76             break;
  77           case 0xa6:
  78             if (i >= 7150 && i <= 7157) /* 0xA6B9..0xA6C0 */
  79               wc = 0xe785 + (i - 7150);
  80             else if (i >= 7183 && i <= 7184) /* 0xA6DA..0xA6DB */
  81               wc = 0xfe12 - (i - 7183);
  82             else if (i >= 7182 && i <= 7190) /* 0xA6D9..0xA6DF */
  83               wc = 0xfe10 + (i - 7182);
  84             else if (i >= 7201 && i <= 7202) /* 0xA6EC..0xA6ED */
  85               wc = 0xfe17 + (i - 7201);
  86             else if (i == 7208) /* 0xA6F3 */
  87               wc = 0xfe19;
  88             else if (i >= 7211 && i <= 7219) /* 0xA6F6..0xA6FE */
  89               wc = 0xe797 + (i - 7211);
  90             break;
  91           case 0xa7:
  92             if (i >= 7349 && i <= 7363) /* 0xA7C2..0xA7D0 */
  93               wc = 0xe7a0 + (i - 7349);
  94             else if (i >= 7397 && i <= 7409) /* 0xA7F2..0xA7FE */
  95               wc = 0xe7af + (i - 7397);
  96             break;
  97           case 0xa8:
  98             if (i >= 7495 && i <= 7505) /* 0xA896..0xA8A0 */
  99               wc = 0xe7bc + (i - 7495);
 100             else if (i == 7533) /* 0xA8BC */
 101               wc = 0x1e3f;
 102             else if (i == 7536) /* 0xA8BF */
 103               wc = 0x01f9;
 104             else if (i >= 7538 && i <= 7541) /* 0xA8C1..0xA8C4 */
 105               wc = 0xe7c9 + (i - 7538);
 106             else if (i >= 7579 && i <= 7599) /* 0xA8EA..0xA8FE */
 107               wc = 0xe7cd + (i - 7579);
 108             break;
 109           case 0xa9:
 110             if (i == 7624) /* 0xA958 */
 111               wc = 0xe7e2;
 112             else if (i == 7627) /* 0xA95B */
 113               wc = 0xe7e3;
 114             else if (i >= 7629 && i <= 7631) /* 0xA95D..0xA95F */
 115               wc = 0xe7e4 + (i - 7629);
 116             else if (i >= 7672 && i < 7685) /* 0xA989..0xA995 */
 117               wc = gb18030ext_2uni_pagea9[i-7672];
 118             else if (i >= 7686 && i <= 7698) /* 0xA997..0xA9A3 */
 119               wc = 0xe7f4 + (i - 7686);
 120             else if (i >= 7775 && i <= 7789) /* 0xA9F0..0xA9FE */
 121               wc = 0xe801 + (i - 7775);
 122             break;
 123           case 0xd7:
 124             if (i >= 16525 && i <= 16529) /* 0xD7FA..0xD7FE */
 125               wc = 0xe810 + (i - 16525);
 126             break;
 127           case 0xfe:
 128             if (i < 23846)
 129               wc = gb18030ext_2uni_pagefe[i-23750];
 130             break;
 131           default:
 132             break;
 133         }
 134         if (wc != 0xfffd) {
 135           *pwc = (ucs4_t) wc;
 136           return 2;
 137         }
 138       }
 139       return RET_ILSEQ;
 140     }
 141     return RET_TOOFEW(0);
 142   }
 143   return RET_ILSEQ;
 144 }
 145
 146 static const unsigned short gb18030ext_page2e[80] = {
 147   0x0000, 0xfe50, 0x0000, 0x0000, 0xfe54, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
 148   0xfe57, 0x0000, 0x0000, 0xfe58, 0xfe5d, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
 149   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe5e, /*0x90-0x97*/
 150   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x98-0x9f*/
 151   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe6b, /*0xa0-0xa7*/
 152   0x0000, 0x0000, 0xfe6e, 0x0000, 0x0000, 0x0000, 0xfe71, 0x0000, /*0xa8-0xaf*/
 153   0x0000, 0x0000, 0x0000, 0xfe73, 0x0000, 0x0000, 0xfe74, 0xfe75, /*0xb0-0xb7*/
 154   0x0000, 0x0000, 0x0000, 0xfe79, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
 155   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc0-0xc7*/
 156   0x0000, 0x0000, 0xfe84, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc8-0xcf*/
 157 };
 158 static const unsigned short gb18030ext_page2f[16] = {
 159   0xa98a, 0xa98b, 0xa98c, 0xa98d, 0xa98e, 0xa98f, 0xa990, 0xa991, /*0xf0-0xf7*/
 160   0xa992, 0xa993, 0xa994, 0xa995, 0x0000, 0x0000, 0x0000, 0x0000, /*0xf8-0xff*/
 161 };
 162 static const unsigned short gb18030ext_page34[56] = {
 163   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe56, /*0x40-0x47*/
 164   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
 165   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
 166   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
 167   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
 168   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x68-0x6f*/
 169   0x0000, 0x0000, 0x0000, 0xfe55, 0x0000, 0x0000, 0x0000, 0x0000, /*0x70-0x77*/
 170 };
 171 static const unsigned short gb18030ext_page36[24] = {
 172   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe5c, 0x0000, /*0x08-0x0f*/
 173   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x10-0x17*/
 174   0x0000, 0x0000, 0xfe5b, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
 175 };
 176 static const unsigned short gb18030ext_page39[24] = {
 177   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe62, /*0xc8-0xcf*/
 178   0xfe65, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xd0-0xd7*/
 179   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe63, /*0xd8-0xdf*/
 180 };
 181 static const unsigned short gb18030ext_page43[56] = {
 182   0x0000, 0x0000, 0x0000, 0x0000, 0xfe78, 0x0000, 0x0000, 0x0000, /*0xa8-0xaf*/
 183   0x0000, 0xfe77, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb0-0xb7*/
 184   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
 185   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc0-0xc7*/
 186   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc8-0xcf*/
 187   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xd0-0xd7*/
 188   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe7a, 0x0000, 0x0000, /*0xd8-0xdf*/
 189 };
 190 static const unsigned short gb18030ext_page46[32] = {
 191   0x0000, 0x0000, 0x0000, 0x0000, 0xfe7d, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
 192   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
 193   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
 194   0x0000, 0xfe7c, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
 195 };
 196 static const unsigned short gb18030ext_page47_1[16] = {
 197   0x0000, 0x0000, 0x0000, 0xfe80, 0x0000, 0x0000, 0x0000, 0x0000, /*0x20-0x27*/
 198   0x0000, 0xfe81, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x28-0x2f*/
 199 };
 200 static const unsigned short gb18030ext_page47_2[24] = {
 201   0x0000, 0x0000, 0x0000, 0x0000, 0xfe82, 0x0000, 0x0000, 0x0000, /*0x78-0x7f*/
 202   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
 203   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe83, 0x0000, 0x0000, /*0x88-0x8f*/
 204 };
 205 static const unsigned short gb18030ext_page49[120] = {
 206   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe85, /*0x40-0x47*/
 207   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
 208   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
 209   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
 210   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
 211   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x68-0x6f*/
 212   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x70-0x77*/
 213   0x0000, 0x0000, 0xfe86, 0x0000, 0x0000, 0xfe87, 0x0000, 0x0000, /*0x78-0x7f*/
 214   0x0000, 0x0000, 0xfe88, 0xfe89, 0x0000, 0xfe8a, 0xfe8b, 0x0000, /*0x80-0x87*/
 215   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
 216   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x90-0x97*/
 217   0x0000, 0x0000, 0x0000, 0xfe8d, 0x0000, 0x0000, 0x0000, 0xfe8c, /*0x98-0x9f*/
 218   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa0-0xa7*/
 219   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa8-0xaf*/
 220   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe8f, 0xfe8e, /*0xb0-0xb7*/
 221 };
 222 static const unsigned short gb18030ext_page4c[56] = {
 223   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe96, /*0x70-0x77*/
 224   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x78-0x7f*/
 225   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
 226   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
 227   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x90-0x97*/
 228   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe93, /*0x98-0x9f*/
 229   0xfe94, 0xfe95, 0xfe97, 0xfe92, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa0-0xa7*/
 230 };
 231 static const unsigned short gb18030ext_page4d[16] = {
 232   0x0000, 0x0000, 0x0000, 0xfe98, 0xfe99, 0xfe9a, 0xfe9b, 0xfe9c, /*0x10-0x17*/
 233   0xfe9d, 0xfe9e, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
 234 };
 235 static const unsigned short gb18030ext_page9f[16] = {
 236   0x0000, 0x0000, 0x0000, 0x0000, 0xfe59, 0xfe61, 0xfe66, 0xfe67, /*0xb0-0xb7*/
 237   0xfe6d, 0xfe7e, 0xfe90, 0xfea0, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
 238 };
 239 static const unsigned short gb18030ext_pagefe[16] = {
 240   0xa6d9, 0xa6db, 0xa6da, 0xa6dc, 0xa6dd, 0xa6de, 0xa6df, 0xa6ec, /*0x10-0x17*/
 241   0xa6ed, 0xa6f3, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
 242 };
 243
 244 static int
 245 gb18030ext_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n)
 246 {
 247   if (n >= 2) {
 248     unsigned short c = 0;
 249     if (wc == 0x01f9)
 250       c = 0xa8bf;
 251     else if (wc == 0x1e3f)
 252       c = 0xa8bc;
 253     else if (wc == 0x20ac)
 254       c = 0xa2e3;
 255     else if (wc >= 0x2e80 && wc < 0x2ed0)
 256       c = gb18030ext_page2e[wc-0x2e80];
 257     else if (wc >= 0x2ff0 && wc < 0x3000)
 258       c = gb18030ext_page2f[wc-0x2ff0];
 259     else if (wc == 0x303e)
 260       c = 0xa989;
 261     else if (wc >= 0x3440 && wc < 0x3478)
 262       c = gb18030ext_page34[wc-0x3440];
 263     else if (wc == 0x359e)
 264       c = 0xfe5a;
 265     else if (wc >= 0x3608 && wc < 0x3620)
 266       c = gb18030ext_page36[wc-0x3608];
 267     else if (wc == 0x3918)
 268       c = 0xfe60;
 269     else if (wc == 0x396e)
 270       c = 0xfe5f;
 271     else if (wc >= 0x39c8 && wc < 0x39e0)
 272       c = gb18030ext_page39[wc-0x39c8];
 273     else if (wc == 0x3a73)
 274       c = 0xfe64;
 275     else if (wc == 0x3b4e)
 276       c = 0xfe68;
 277     else if (wc == 0x3c6e)
 278       c = 0xfe69;
 279     else if (wc == 0x3ce0)
 280       c = 0xfe6a;
 281     else if (wc == 0x4056)
 282       c = 0xfe6f;
 283     else if (wc == 0x415f)
 284       c = 0xfe70;
 285     else if (wc == 0x4337)
 286       c = 0xfe72;
 287     else if (wc >= 0x43a8 && wc < 0x43e0)
 288       c = gb18030ext_page43[wc-0x43a8];
 289     else if (wc == 0x44d6)
 290       c = 0xfe7b;
 291     else if (wc >= 0x4648 && wc < 0x4668)
 292       c = gb18030ext_page46[wc-0x4648];
 293     else if (wc >= 0x4720 && wc < 0x4730)
 294       c = gb18030ext_page47_1[wc-0x4720];
 295     else if (wc >= 0x4778 && wc < 0x4790)
 296       c = gb18030ext_page47_2[wc-0x4778];
 297     else if (wc >= 0x4940 && wc < 0x49b8)
 298       c = gb18030ext_page49[wc-0x4940];
 299     else if (wc >= 0x4c70 && wc < 0x4ca8)
 300       c = gb18030ext_page4c[wc-0x4c70];
 301     else if (wc >= 0x4d10 && wc < 0x4d20)
 302       c = gb18030ext_page4d[wc-0x4d10];
 303     else if (wc == 0x4dae)
 304       c = 0xfe9f;
 305     else if (wc >= 0x9fb4 && wc < 0x9fbc)
 306       c = gb18030ext_page9f[wc-0x9fb0];
 307     else if (wc >= 0xfe10 && wc < 0xfe1a)
 308       c = gb18030ext_pagefe[wc-0xfe10];
 309     else if (wc == 0x20087)
 310       c = 0xfe51;
 311     else if (wc == 0x20089)
 312       c = 0xfe52;
 313     else if (wc == 0x200cc)
 314       c = 0xfe53;
 315     else if (wc == 0x215d7)
 316       c = 0xfe6c;
 317     else if (wc == 0x2298f)
 318       c = 0xfe76;
 319     else if (wc == 0x241fe)
 320       c = 0xfe91;
 321     if (c != 0) {
 322       r[0] = (c >> 8); r[1] = (c & 0xff);
 323       return 2;
 324     }
 325     return RET_ILUNI;
 326   }
 327   return RET_TOOSMALL;
 328 }