iconvdata/cns11643.h

   1 /* Access functions for CNS 11643 handling.
   2    Copyright (C) 1998,1999,2000,2001,2002,2003,2007
   3    Free Software Foundation, Inc.
   4    This file is part of the GNU C Library.
   5    Contributed by Ulrich Drepper <drepper@cygnus.com>, 1998.
   6
   7    The GNU C Library is free software; you can redistribute it and/or
   8    modify it under the terms of the GNU Lesser General Public
   9    License as published by the Free Software Foundation; either
  10    version 2.1 of the License, or (at your option) any later version.
  11
  12    The GNU C Library is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15    Lesser General Public License for more details.
  16
  17    You should have received a copy of the GNU Lesser General Public
  18    License along with the GNU C Library; if not, write to the Free
  19    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  20    02111-1307 USA.  */
  21
  22 #include <stdint.h>
  23
  24
  25 /* Table for CNS 11643, plane 1 to UCS4 conversion.  */
  26 extern const uint16_t __cns11643l1_to_ucs4_tab[];
  27 /* Table for CNS 11643, plane 2 to UCS4 conversion.  */
  28 extern const uint16_t __cns11643l2_to_ucs4_tab[];
  29 /* Table for CNS 11643, plane 3 to UCS4 conversion.  */
  30 extern const uint32_t __cns11643l3_to_ucs4_tab[];
  31 /* Table for CNS 11643, plane 4 to UCS4 conversion.  */
  32 extern const uint32_t __cns11643l4_to_ucs4_tab[];
  33 /* Table for CNS 11643, plane 5 to UCS4 conversion.  */
  34 extern const uint32_t __cns11643l5_to_ucs4_tab[];
  35 /* Table for CNS 11643, plane 6 to UCS4 conversion.  */
  36 extern const uint32_t __cns11643l6_to_ucs4_tab[];
  37 /* Table for CNS 11643, plane 7 to UCS4 conversion.  */
  38 extern const uint32_t __cns11643l7_to_ucs4_tab[];
  39 /* Table for CNS 11643, plane 15 (old) to UCS4 conversion.  */
  40 extern const uint32_t __cns11643l15_to_ucs4_tab[];
  41
  42
  43 static inline uint32_t
  44 __attribute ((always_inline))
  45 cns11643_to_ucs4 (const unsigned char **s, size_t avail, unsigned char offset)
  46 {
  47   unsigned char ch = *(*s);
  48   unsigned char ch2;
  49   unsigned char ch3;
  50   uint32_t result;
  51   int idx;
  52
  53   if (ch < offset || (ch - offset) <= 0x20 || (ch - offset) > 0x30)
  54     return __UNKNOWN_10646_CHAR;
  55
  56   if (avail < 3)
  57     return 0;
  58
  59   ch2 = (*s)[1];
  60   if ((ch2 - offset) <= 0x20 || (ch2 - offset) >= 0x7f)
  61     return __UNKNOWN_10646_CHAR;
  62
  63   ch3 = (*s)[2];
  64   if ((ch3 - offset) <= 0x20 || (ch3 - offset) >= 0x7f)
  65     return __UNKNOWN_10646_CHAR;
  66
  67   idx = (ch2 - 0x21 - offset) * 94 + (ch3 - 0x21 - offset);
  68
  69   switch (ch - 0x20 - offset)
  70     {
  71     case 1:
  72       if (idx > 0x21f2)
  73         return __UNKNOWN_10646_CHAR;
  74       result = __cns11643l1_to_ucs4_tab[idx];
  75       break;
  76     case 2:
  77       if (idx > 0x1de1)
  78         return __UNKNOWN_10646_CHAR;
  79       result = __cns11643l2_to_ucs4_tab[idx];
  80       break;
  81     case 3:
  82       if (idx > 0x19bd)
  83         return __UNKNOWN_10646_CHAR;
  84       result = __cns11643l3_to_ucs4_tab[idx];
  85       break;
  86     case 4:
  87       if (idx > 0x1c81)
  88         return __UNKNOWN_10646_CHAR;
  89       result = __cns11643l4_to_ucs4_tab[idx];
  90       break;
  91     case 5:
  92       if (idx > 0x219a)
  93         return __UNKNOWN_10646_CHAR;
  94       result = __cns11643l5_to_ucs4_tab[idx];
  95       break;
  96     case 6:
  97       if (idx > 0x18f3)
  98         return __UNKNOWN_10646_CHAR;
  99       result = __cns11643l6_to_ucs4_tab[idx];
 100       break;
 101     case 7:
 102       if (idx > 0x198a)
 103         return __UNKNOWN_10646_CHAR;
 104       result = __cns11643l7_to_ucs4_tab[idx];
 105       break;
 106     case 15:
 107       if (idx > 0x1c00)
 108         return __UNKNOWN_10646_CHAR;
 109       result = __cns11643l15_to_ucs4_tab[idx];
 110       break;
 111     default:
 112       return __UNKNOWN_10646_CHAR;
 113     }
 114
 115   if (result != L'\0')
 116     (*s) += 3;
 117   else
 118     result = __UNKNOWN_10646_CHAR;
 119
 120   return result;
 121 }
 122
 123
 124 /* Tables for the UCS4 -> CNS conversion.  */
 125 extern const char __cns11643l1_from_ucs4_tab1[][2];
 126 extern const char __cns11643l1_from_ucs4_tab2[][2];
 127 extern const char __cns11643l1_from_ucs4_tab3[][2];
 128 extern const char __cns11643l1_from_ucs4_tab4[][2];
 129 extern const char __cns11643l1_from_ucs4_tab5[][2];
 130 extern const char __cns11643l1_from_ucs4_tab6[][2];
 131 extern const char __cns11643l1_from_ucs4_tab7[][2];
 132 extern const char __cns11643l1_from_ucs4_tab8[][2];
 133 extern const char __cns11643l1_from_ucs4_tab9[][2];
 134 extern const char __cns11643l1_from_ucs4_tab10[][2];
 135 extern const char __cns11643l1_from_ucs4_tab11[][2];
 136 extern const char __cns11643l1_from_ucs4_tab12[][2];
 137 extern const char __cns11643l1_from_ucs4_tab13[][2];
 138 extern const char __cns11643l1_from_ucs4_tab14[][2];
 139 extern const char __cns11643_from_ucs4p0_tab[][3];
 140 extern const char __cns11643_from_ucs4p2_tab[][3];
 141 extern const char __cns11643_from_ucs4p2c_tab[][3];
 142
 143
 144 static inline size_t
 145 __attribute ((always_inline))
 146 ucs4_to_cns11643 (uint32_t wch, unsigned char *s, size_t avail)
 147 {
 148   unsigned int ch = (unsigned int) wch;
 149   char buf[2];
 150   const char *cp = buf;
 151   size_t needed = 2;
 152
 153   switch (ch)
 154     {
 155     case 0xa7 ... 0xf7:
 156       cp = __cns11643l1_from_ucs4_tab1[ch - 0xa7];
 157       break;
 158     case 0x2c7 ... 0x2d9:
 159       cp = __cns11643l1_from_ucs4_tab2[ch - 0x2c7];
 160       break;
 161     case 0x391 ... 0x3c9:
 162       cp = __cns11643l1_from_ucs4_tab3[ch - 0x391];
 163       break;
 164     case 0x2013 ... 0x203e:
 165       cp = __cns11643l1_from_ucs4_tab4[ch - 0x2013];
 166       break;
 167     case 0x2103:
 168       cp = "\x22\x6a";
 169       break;
 170     case 0x2105:
 171       cp = "\x22\x22";
 172       break;
 173     case 0x2109:
 174       cp = "\x22\x6b";
 175       break;
 176     case 0x2160 ... 0x2169:
 177       buf[0] = '\x24';
 178       buf[1] = '\x2b' + (ch - 0x2160);
 179       break;
 180     case 0x2170 ... 0x2179:
 181       buf[0] = '\x26';
 182       buf[1] = '\x35' + (ch - 0x2170);
 183       break;
 184     case 0x2190 ... 0x2199:
 185       cp = __cns11643l1_from_ucs4_tab5[ch - 0x2190];
 186       break;
 187     case 0x2215 ... 0x2267:
 188       cp = __cns11643l1_from_ucs4_tab6[ch - 0x2215];
 189       break;
 190     case 0x22a5:
 191       cp = "\x22\x47";
 192       break;
 193     case 0x22bf:
 194       cp = "\x22\x4a";
 195       break;
 196     case 0x2400 ... 0x2421:
 197       cp = __cns11643l1_from_ucs4_tab7[ch - 0x2400];
 198       break;
 199     case 0x2460 ... 0x247d:
 200       cp = __cns11643l1_from_ucs4_tab8[ch - 0x2460];
 201       break;
 202     case 0x2500 ... 0x2642:
 203       cp = __cns11643l1_from_ucs4_tab9[ch - 0x2500];
 204       break;
 205     case 0x3000 ... 0x3029:
 206       cp = __cns11643l1_from_ucs4_tab10[ch - 0x3000];
 207       break;
 208     case 0x30fb:
 209       cp = "\x21\x26";
 210       break;
 211     case 0x3105 ... 0x3129:
 212       buf[0] = '\x25';
 213       buf[1] = '\x47' + (ch - 0x3105);
 214       break;
 215     case 0x32a3:
 216       cp = "\x22\x21";
 217       break;
 218     case 0x338e ... 0x33d5:
 219       cp = __cns11643l1_from_ucs4_tab11[ch - 0x338e];
 220       break;
 221     case 0x4e00 ... 0x9f9c:
 222       cp = __cns11643l1_from_ucs4_tab12[ch - 0x4e00];
 223       if (cp[0] != '\0')
 224         break;
 225       /* FALLTHROUGH.  Let's try the other planes.  */
 226     case 0x3400 ... 0x4dff:
 227     case 0x9f9d ... 0x9fa5:
 228       /* Let's try the other planes.  */
 229       needed = 3;
 230       cp = __cns11643_from_ucs4p0_tab[ch - 0x3400];
 231       break;
 232     case 0xfa28:
 233       needed = 3;
 234       cp = "\x0f\x58\x4c";
 235       break;
 236     case 0xfe30 ... 0xfe6b:
 237       cp = __cns11643l1_from_ucs4_tab13[ch - 0xfe30];
 238       break;
 239     case 0xff01 ... 0xff5d:
 240       cp = __cns11643l1_from_ucs4_tab14[ch - 0xff01];
 241       break;
 242     case 0xffe0:
 243       cp = "\x22\x66";
 244       break;
 245     case 0xffe1:
 246       cp = "\x22\x67";
 247       break;
 248     case 0xffe5:
 249       cp = "\x22\x64";
 250       break;
 251     case 0x20000 ... 0x2a6d6:
 252       needed = 3;
 253       cp = __cns11643_from_ucs4p2_tab[ch - 0x20000];
 254       break;
 255     case 0x2f800 ... 0x2fa1d:
 256       needed = 3;
 257       cp = __cns11643_from_ucs4p2c_tab[ch - 0x2f800];
 258       break;
 259     default:
 260       return __UNKNOWN_10646_CHAR;
 261     }
 262
 263   if (cp[0] == '\0')
 264     return __UNKNOWN_10646_CHAR;
 265
 266   if (avail < needed)
 267     return 0;
 268
 269   s[0] = cp[0];
 270   s[1] = cp[1];
 271   if (needed == 3)
 272     s[2] = cp[2];
 273
 274   return needed;
 275 }