iconvdata/tcvn5712-1.c

   1 /* Conversion to and from TCVN5712-1.
   2    Copyright (C) 2001, 2002 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4    Contributed by Ulrich Drepper <drepper@redhat.com>, 2001.
   5
   6    The GNU C Library is free software; you can redistribute it and/or
   7    modify it under the terms of the GNU Lesser General Public
   8    License as published by the Free Software Foundation; either
   9    version 2.1 of the License, or (at your option) any later version.
  10
  11    The GNU C Library is distributed in the hope that it will be useful,
  12    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14    Lesser General Public License for more details.
  15
  16    You should have received a copy of the GNU Lesser General Public
  17    License along with the GNU C Library; if not, write to the Free
  18    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  19    02111-1307 USA.  */
  20
  21 #include <dlfcn.h>
  22 #include <stdint.h>
  23 #include <stdlib.h>
  24
  25 #define NELEMS(arr) (sizeof (arr) / sizeof (arr[0]))
  26
  27 /* Definitions used in the body of the `gconv' function.  */
  28 #define CHARSET_NAME            "TCVN5712-1//"
  29 #define FROM_LOOP               from_tcvn5712_1
  30 #define TO_LOOP                 to_tcvn5712_1
  31 #define DEFINE_INIT             1
  32 #define DEFINE_FINI             1
  33 #define MIN_NEEDED_FROM         1
  34 #define MAX_NEEDED_FROM         2
  35 #define MIN_NEEDED_TO           4
  36 #define MAX_NEEDED_TO           8
  37 #define PREPARE_LOOP \
  38   int saved_state;                                                            \
  39   int *statep = &data->__statep->__count;
  40 #define EXTRA_LOOP_ARGS         , statep
  41
  42
  43 /* Since we might have to reset input pointer we must be able to save
  44    and restore the state.  */
  45 #define SAVE_RESET_STATE(Save) \
  46   if (Save)                                                                   \
  47     saved_state = *statep;                                                    \
  48   else                                                                        \
  49     *statep = saved_state
  50
  51
  52 /* During TCVN5712-1 to UCS4 conversion, the COUNT element of the state
  53    contains the last UCS4 character, shifted by 3 bits.  */
  54
  55
  56 /* Since this is a stateful encoding we have to provide code which resets
  57    the output state to the initial state.  This has to be done during the
  58    flushing.  */
  59 #define EMIT_SHIFT_TO_INIT \
  60   if (data->__statep->__count != 0)                                           \
  61     {                                                                         \
  62       if (FROM_DIRECTION)                                                     \
  63         {                                                                     \
  64           if (__builtin_expect (outbuf + 4 <= outend, 1))                     \
  65             {                                                                 \
  66               /* Write out the last character.  */                            \
  67               *((uint32_t *) outbuf)++ = data->__statep->__count >> 3;        \
  68               data->__statep->__count = 0;                                    \
  69             }                                                                 \
  70           else                                                                \
  71             /* We don't have enough room in the output buffer.  */            \
  72             status = __GCONV_FULL_OUTPUT;                                     \
  73         }                                                                     \
  74       else                                                                    \
  75         /* We don't use shift states in the TO_DIRECTION.  */                 \
  76         data->__statep->__count = 0;                                          \
  77     }
  78
  79
  80 static const uint16_t map_from_tcvn_low[0x18] =
  81   {
  82     0x0000, 0x00da, 0x1ee4, 0x0003, 0x1eea, 0x1eec, 0x1eee, 0x0007,
  83     0x0008, 0x0009, 0x000a, 0x000b, 0x000c, 0x000d, 0x000e, 0x000f,
  84     0x0010, 0x1ee8, 0x1ef0, 0x1ef2, 0x1ef6, 0x1ef8, 0x00dd, 0x1ef4
  85   };
  86
  87 static const uint16_t map_from_tcvn_high[0x80] =
  88   {
  89     0x00c0, 0x1ea2, 0x00c3, 0x00c1, 0x1ea0, 0x1eb6, 0x1eac, 0x00c8,
  90     0x1eba, 0x1ebc, 0x00c9, 0x1eb8, 0x1ec6, 0x00cc, 0x1ec8, 0x0128,
  91     0x00cd, 0x1eca, 0x00d2, 0x1ece, 0x00d5, 0x00d3, 0x1ecc, 0x1ed8,
  92     0x1edc, 0x1ede, 0x1ee0, 0x1eda, 0x1ee2, 0x00d9, 0x1ee6, 0x0168,
  93     0x00a0, 0x0102, 0x00c2, 0x00ca, 0x00d4, 0x01a0, 0x01af, 0x0110,
  94     0x0103, 0x00e2, 0x00ea, 0x00f4, 0x01a1, 0x01b0, 0x0111, 0x1eb0,
  95     0x0300, 0x0309, 0x0303, 0x0301, 0x0323, 0x00e0, 0x1ea3, 0x00e3,
  96     0x00e1, 0x1ea1, 0x1eb2, 0x1eb1, 0x1eb3, 0x1eb5, 0x1eaf, 0x1eb4,
  97     0x1eae, 0x1ea6, 0x1ea8, 0x1eaa, 0x1ea4, 0x1ec0, 0x1eb7, 0x1ea7,
  98     0x1ea9, 0x1eab, 0x1ea5, 0x1ead, 0x00e8, 0x1ec2, 0x1ebb, 0x1ebd,
  99     0x00e9, 0x1eb9, 0x1ec1, 0x1ec3, 0x1ec5, 0x1ebf, 0x1ec7, 0x00ec,
 100     0x1ec9, 0x1ec4, 0x1ebe, 0x1ed2, 0x0129, 0x00ed, 0x1ecb, 0x00f2,
 101     0x1ed4, 0x1ecf, 0x00f5, 0x00f3, 0x1ecd, 0x1ed3, 0x1ed5, 0x1ed7,
 102     0x1ed1, 0x1ed9, 0x1edd, 0x1edf, 0x1ee1, 0x1edb, 0x1ee3, 0x00f9,
 103     0x1ed6, 0x1ee7, 0x0169, 0x00fa, 0x1ee5, 0x1eeb, 0x1eed, 0x1eef,
 104     0x1ee9, 0x1ef1, 0x1ef3, 0x1ef7, 0x1ef9, 0x00fd, 0x1ef5, 0x1ed0
 105   };
 106
 107
 108 /* TCVN5712-1 contains five combining characters:
 109    0x0300, 0x0301, 0x0303, 0x0309, 0x0323.  */
 110
 111 /* Composition tables for each of the relevant combining characters.  */
 112 static const struct
 113 {
 114   uint16_t base;
 115   uint16_t composed;
 116 } comp_table_data[] =
 117   {
 118 #define COMP_TABLE_IDX_0300 0
 119 #define COMP_TABLE_LEN_0300 28
 120     { 0x0041, 0x00C0 },
 121     { 0x0045, 0x00C8 },
 122     { 0x0049, 0x00CC },
 123     { 0x004E, 0x01F8 },
 124     { 0x004F, 0x00D2 },
 125     { 0x0055, 0x00D9 },
 126     { 0x0057, 0x1E80 },
 127     { 0x0059, 0x1EF2 },
 128     { 0x0061, 0x00E0 },
 129     { 0x0065, 0x00E8 },
 130     { 0x0069, 0x00EC },
 131     { 0x006E, 0x01F9 },
 132     { 0x006F, 0x00F2 },
 133     { 0x0075, 0x00F9 },
 134     { 0x0077, 0x1E81 },
 135     { 0x0079, 0x1EF3 },
 136   /*{ 0x00A8, 0x1FED },*/
 137     { 0x00C2, 0x1EA6 },
 138     { 0x00CA, 0x1EC0 },
 139     { 0x00D4, 0x1ED2 },
 140   /*{ 0x00DC, 0x01DB },*/
 141     { 0x00E2, 0x1EA7 },
 142     { 0x00EA, 0x1EC1 },
 143     { 0x00F4, 0x1ED3 },
 144   /*{ 0x00FC, 0x01DC },*/
 145     { 0x0102, 0x1EB0 },
 146     { 0x0103, 0x1EB1 },
 147   /*{ 0x0112, 0x1E14 },*/
 148   /*{ 0x0113, 0x1E15 },*/
 149   /*{ 0x014C, 0x1E50 },*/
 150   /*{ 0x014D, 0x1E51 },*/
 151     { 0x01A0, 0x1EDC },
 152     { 0x01A1, 0x1EDD },
 153     { 0x01AF, 0x1EEA },
 154     { 0x01B0, 0x1EEB },
 155 #define COMP_TABLE_IDX_0301 (COMP_TABLE_IDX_0300 + COMP_TABLE_LEN_0300)
 156 #define COMP_TABLE_LEN_0301 51
 157     { 0x0041, 0x00C1 },
 158     { 0x0043, 0x0106 },
 159     { 0x0045, 0x00C9 },
 160     { 0x0047, 0x01F4 },
 161     { 0x0049, 0x00CD },
 162     { 0x004B, 0x1E30 },
 163     { 0x004C, 0x0139 },
 164     { 0x004D, 0x1E3E },
 165     { 0x004E, 0x0143 },
 166     { 0x004F, 0x00D3 },
 167     { 0x0050, 0x1E54 },
 168     { 0x0052, 0x0154 },
 169     { 0x0053, 0x015A },
 170     { 0x0055, 0x00DA },
 171     { 0x0057, 0x1E82 },
 172     { 0x0059, 0x00DD },
 173     { 0x005A, 0x0179 },
 174     { 0x0061, 0x00E1 },
 175     { 0x0063, 0x0107 },
 176     { 0x0065, 0x00E9 },
 177     { 0x0067, 0x01F5 },
 178     { 0x0069, 0x00ED },
 179     { 0x006B, 0x1E31 },
 180     { 0x006C, 0x013A },
 181     { 0x006D, 0x1E3F },
 182     { 0x006E, 0x0144 },
 183     { 0x006F, 0x00F3 },
 184     { 0x0070, 0x1E55 },
 185     { 0x0072, 0x0155 },
 186     { 0x0073, 0x015B },
 187     { 0x0075, 0x00FA },
 188     { 0x0077, 0x1E83 },
 189     { 0x0079, 0x00FD },
 190     { 0x007A, 0x017A },
 191     { 0x00A5, 0x0385 },
 192   /*{ 0x00A8, 0x1FEE },*/
 193     { 0x00C2, 0x1EA4 },
 194   /*{ 0x00C5, 0x01FA },*/
 195   /*{ 0x00C6, 0x01FC },*/
 196   /*{ 0x00C7, 0x1E08 },*/
 197     { 0x00CA, 0x1EBE },
 198   /*{ 0x00CF, 0x1E2E },*/
 199     { 0x00D4, 0x1ED0 },
 200     { 0x00D5, 0x1E4C },
 201   /*{ 0x00D8, 0x01FE },*/
 202   /*{ 0x00DC, 0x01D7 },*/
 203     { 0x00E2, 0x1EA5 },
 204   /*{ 0x00E5, 0x01FB },*/
 205   /*{ 0x00E6, 0x01FD },*/
 206   /*{ 0x00E7, 0x1E09 },*/
 207     { 0x00EA, 0x1EBF },
 208   /*{ 0x00EF, 0x1E2F },*/
 209     { 0x00F4, 0x1ED1 },
 210     { 0x00F5, 0x1E4D },
 211   /*{ 0x00F8, 0x01FF },*/
 212   /*{ 0x00FC, 0x01D8 },*/
 213     { 0x0102, 0x1EAE },
 214     { 0x0103, 0x1EAF },
 215   /*{ 0x0112, 0x1E16 },*/
 216   /*{ 0x0113, 0x1E17 },*/
 217   /*{ 0x014C, 0x1E52 },*/
 218   /*{ 0x014D, 0x1E53 },*/
 219     { 0x0168, 0x1E78 },
 220     { 0x0169, 0x1E79 },
 221     { 0x01A0, 0x1EDA },
 222     { 0x01A1, 0x1EDB },
 223     { 0x01AF, 0x1EE8 },
 224     { 0x01B0, 0x1EE9 },
 225 #define COMP_TABLE_IDX_0303 (COMP_TABLE_IDX_0301 + COMP_TABLE_LEN_0301)
 226 #define COMP_TABLE_LEN_0303 34
 227     { 0x0041, 0x00C3 },
 228     { 0x0045, 0x1EBC },
 229     { 0x0049, 0x0128 },
 230     { 0x004E, 0x00D1 },
 231     { 0x004F, 0x00D5 },
 232     { 0x0055, 0x0168 },
 233     { 0x0056, 0x1E7C },
 234     { 0x0059, 0x1EF8 },
 235     { 0x0061, 0x00E3 },
 236     { 0x0065, 0x1EBD },
 237     { 0x0069, 0x0129 },
 238     { 0x006E, 0x00F1 },
 239     { 0x006F, 0x00F5 },
 240     { 0x0075, 0x0169 },
 241     { 0x0076, 0x1E7D },
 242     { 0x0079, 0x1EF9 },
 243     { 0x00C2, 0x1EAA },
 244     { 0x00CA, 0x1EC4 },
 245     { 0x00D3, 0x1E4C },
 246     { 0x00D4, 0x1ED6 },
 247     { 0x00D6, 0x1E4E },
 248     { 0x00DA, 0x1E78 },
 249     { 0x00E2, 0x1EAB },
 250     { 0x00EA, 0x1EC5 },
 251     { 0x00F3, 0x1E4D },
 252     { 0x00F4, 0x1ED7 },
 253     { 0x00F6, 0x1E4F },
 254     { 0x00FA, 0x1E79 },
 255     { 0x0102, 0x1EB4 },
 256     { 0x0103, 0x1EB5 },
 257     { 0x01A0, 0x1EE0 },
 258     { 0x01A1, 0x1EE1 },
 259     { 0x01AF, 0x1EEE },
 260     { 0x01B0, 0x1EEF },
 261 #define COMP_TABLE_IDX_0309 (COMP_TABLE_IDX_0303 + COMP_TABLE_LEN_0303)
 262 #define COMP_TABLE_LEN_0309 24
 263     { 0x0041, 0x1EA2 },
 264     { 0x0045, 0x1EBA },
 265     { 0x0049, 0x1EC8 },
 266     { 0x004F, 0x1ECE },
 267     { 0x0055, 0x1EE6 },
 268     { 0x0059, 0x1EF6 },
 269     { 0x0061, 0x1EA3 },
 270     { 0x0065, 0x1EBB },
 271     { 0x0069, 0x1EC9 },
 272     { 0x006F, 0x1ECF },
 273     { 0x0075, 0x1EE7 },
 274     { 0x0079, 0x1EF7 },
 275     { 0x00C2, 0x1EA8 },
 276     { 0x00CA, 0x1EC2 },
 277     { 0x00D4, 0x1ED4 },
 278     { 0x00E2, 0x1EA9 },
 279     { 0x00EA, 0x1EC3 },
 280     { 0x00F4, 0x1ED5 },
 281     { 0x0102, 0x1EB2 },
 282     { 0x0103, 0x1EB3 },
 283     { 0x01A0, 0x1EDE },
 284     { 0x01A1, 0x1EDF },
 285     { 0x01AF, 0x1EEC },
 286     { 0x01B0, 0x1EED },
 287 #define COMP_TABLE_IDX_0323 (COMP_TABLE_IDX_0309 + COMP_TABLE_LEN_0309)
 288 #define COMP_TABLE_LEN_0323 50
 289     { 0x0041, 0x1EA0 },
 290     { 0x0042, 0x1E04 },
 291     { 0x0044, 0x1E0C },
 292     { 0x0045, 0x1EB8 },
 293     { 0x0048, 0x1E24 },
 294     { 0x0049, 0x1ECA },
 295     { 0x004B, 0x1E32 },
 296     { 0x004C, 0x1E36 },
 297     { 0x004D, 0x1E42 },
 298     { 0x004E, 0x1E46 },
 299     { 0x004F, 0x1ECC },
 300     { 0x0052, 0x1E5A },
 301     { 0x0053, 0x1E62 },
 302     { 0x0054, 0x1E6C },
 303     { 0x0055, 0x1EE4 },
 304     { 0x0056, 0x1E7E },
 305     { 0x0057, 0x1E88 },
 306     { 0x0059, 0x1EF4 },
 307     { 0x005A, 0x1E92 },
 308     { 0x0061, 0x1EA1 },
 309     { 0x0062, 0x1E05 },
 310     { 0x0064, 0x1E0D },
 311     { 0x0065, 0x1EB9 },
 312     { 0x0068, 0x1E25 },
 313     { 0x0069, 0x1ECB },
 314     { 0x006B, 0x1E33 },
 315     { 0x006C, 0x1E37 },
 316     { 0x006D, 0x1E43 },
 317     { 0x006E, 0x1E47 },
 318     { 0x006F, 0x1ECD },
 319     { 0x0072, 0x1E5B },
 320     { 0x0073, 0x1E63 },
 321     { 0x0074, 0x1E6D },
 322     { 0x0075, 0x1EE5 },
 323     { 0x0076, 0x1E7F },
 324     { 0x0077, 0x1E89 },
 325     { 0x0079, 0x1EF5 },
 326     { 0x007A, 0x1E93 },
 327     { 0x00C2, 0x1EAC },
 328     { 0x00CA, 0x1EC6 },
 329     { 0x00D4, 0x1ED8 },
 330     { 0x00E2, 0x1EAD },
 331     { 0x00EA, 0x1EC7 },
 332     { 0x00F4, 0x1ED9 },
 333     { 0x0102, 0x1EB6 },
 334     { 0x0103, 0x1EB7 },
 335     { 0x01A0, 0x1EE2 },
 336     { 0x01A1, 0x1EE3 },
 337     { 0x01AF, 0x1EF0 },
 338     { 0x01B0, 0x1EF1 },
 339 #define COMP_TABLE_IDX_END (COMP_TABLE_IDX_0323 + COMP_TABLE_LEN_0323)
 340   };
 341 /* Compile-time verification of table size.  */
 342 typedef int verify1[(NELEMS (comp_table_data) == COMP_TABLE_IDX_END) - 1];
 343
 344 static const struct
 345 {
 346   unsigned int idx;
 347   unsigned int len;
 348 } comp_table[5] =
 349   {
 350     { COMP_TABLE_IDX_0300, COMP_TABLE_LEN_0300 },
 351     { COMP_TABLE_IDX_0301, COMP_TABLE_LEN_0301 },
 352     { COMP_TABLE_IDX_0303, COMP_TABLE_LEN_0303 },
 353     { COMP_TABLE_IDX_0309, COMP_TABLE_LEN_0309 },
 354     { COMP_TABLE_IDX_0323, COMP_TABLE_LEN_0323 },
 355   };
 356
 357
 358 /* First define the conversion function from TCVN5712-1 to UCS4.  */
 359 #define MIN_NEEDED_INPUT        MIN_NEEDED_FROM
 360 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_TO
 361 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_TO
 362 #define LOOPFCT                 FROM_LOOP
 363 #define BODY \
 364   {                                                                           \
 365     uint32_t ch = *inptr;                                                     \
 366     uint32_t last_ch;                                                         \
 367     int must_buffer_ch;                                                       \
 368                                                                               \
 369     if (ch < 0x18)                                                            \
 370       ch = map_from_tcvn_low[ch];                                             \
 371     else if (ch >= 0x80)                                                      \
 372       ch = map_from_tcvn_high[ch - 0x80];                                     \
 373                                                                               \
 374     /* Determine whether there is a buffered character pending.  */           \
 375     last_ch = *statep >> 3;                                                   \
 376                                                                               \
 377     /* We have to buffer ch if it is a possible match in comp_table_data.  */ \
 378     must_buffer_ch = (ch >= 0x0041 && ch <= 0x01b0);                          \
 379                                                                               \
 380     if (last_ch)                                                              \
 381       {                                                                       \
 382         if (ch >= 0x0300 && ch < 0x0340)                                      \
 383           {                                                                   \
 384             /* See whether last_ch and ch can be combined.  */                \
 385             unsigned int i;                                                   \
 386             unsigned int i1;                                                  \
 387             unsigned int i2;                                                  \
 388                                                                               \
 389             switch (ch)                                                       \
 390               {                                                               \
 391               case 0x0300:                                                    \
 392                 i = 0;                                                        \
 393                 break;                                                        \
 394               case 0x0301:                                                    \
 395                 i = 1;                                                        \
 396                 break;                                                        \
 397               case 0x0303:                                                    \
 398                 i = 2;                                                        \
 399                 break;                                                        \
 400               case 0x0309:                                                    \
 401                 i = 3;                                                        \
 402                 break;                                                        \
 403               case 0x0323:                                                    \
 404                 i = 4;                                                        \
 405                 break;                                                        \
 406               default:                                                        \
 407                 abort ();                                                     \
 408               }                                                               \
 409                                                                               \
 410             i1 = comp_table[i].idx;                                           \
 411             i2 = i1 + comp_table[i].len - 1;                                  \
 412                                                                               \
 413             if (last_ch >= comp_table_data[i1].base                           \
 414                 && last_ch <= comp_table_data[i2].base)                       \
 415               {                                                               \
 416                 for (;;)                                                      \
 417                   {                                                           \
 418                     i = (i1 + i2) >> 1;                                       \
 419                     if (last_ch == comp_table_data[i].base)                   \
 420                       break;                                                  \
 421                     if (last_ch < comp_table_data[i].base)                    \
 422                       {                                                       \
 423                         if (i1 == i)                                          \
 424                           goto not_combining;                                 \
 425                         i2 = i;                                               \
 426                       }                                                       \
 427                     else                                                      \
 428                       {                                                       \
 429                         if (i1 != i)                                          \
 430                           i1 = i;                                             \
 431                         else                                                  \
 432                           {                                                   \
 433                             i = i2;                                           \
 434                             if (last_ch == comp_table_data[i].base)           \
 435                               break;                                          \
 436                             goto not_combining;                               \
 437                           }                                                   \
 438                       }                                                       \
 439                   }                                                           \
 440                 last_ch = comp_table_data[i].composed;                        \
 441                 /* Output the combined character.  */                         \
 442                 put32 (outptr, last_ch);                                      \
 443                 outptr += 4;                                                  \
 444                 *statep = 0;                                                  \
 445                 ++inptr;                                                      \
 446                 continue;                                                     \
 447               }                                                               \
 448           }                                                                   \
 449                                                                               \
 450       not_combining:                                                          \
 451         /* Output the buffered character.  */                                 \
 452         put32 (outptr, last_ch);                                              \
 453         outptr += 4;                                                          \
 454         *statep = 0;                                                          \
 455                                                                               \
 456         /* If we don't have enough room to output ch as well, then deal       \
 457            with it in another round.  */                                      \
 458         if (!must_buffer_ch && __builtin_expect (outptr + 4 > outend, 0))     \
 459           continue;                                                           \
 460       }                                                                       \
 461                                                                               \
 462     if (must_buffer_ch)                                                       \
 463       *statep = ch << 3;                                                      \
 464     else                                                                      \
 465       {                                                                       \
 466         put32 (outptr, ch);                                                   \
 467         outptr += 4;                                                          \
 468       }                                                                       \
 469     ++inptr;                                                                  \
 470   }
 471 #define EXTRA_LOOP_DECLS        , int *statep
 472 #include <iconv/loop.c>
 473
 474
 475 /* Next, define the conversion function from UCS4 to CP1258.  */
 476
 477 static const unsigned char from_ucs4[] =
 478   {
 479 #define FROM_IDX_00 0
 480     0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x07, /* 0x0001-0x0007 */
 481     0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 0x0008-0x000f */
 482     0x10,                                           /* 0x0010-0x0010 */
 483
 484 #define FROM_IDX_01 (FROM_IDX_00 + 16)
 485     0x80, 0x83, 0xa2, 0x82, 0x00, 0x00, 0x00, 0x00, /* 0x00c0-0x00c7 */
 486     0x87, 0x8a, 0xa3, 0x00, 0x8d, 0x90, 0x00, 0x00, /* 0x00c8-0x00cf */
 487     0x00, 0x00, 0x92, 0x95, 0xa4, 0x94, 0x00, 0x00, /* 0x00d0-0x00d7 */
 488     0x00, 0x9d, 0x01, 0x00, 0x00, 0x16, 0x00, 0x00, /* 0x00d8-0x00df */
 489     0xb5, 0xb8, 0xa9, 0xb7, 0x00, 0x00, 0x00, 0x00, /* 0x00e0-0x00e7 */
 490     0xcc, 0xd0, 0xaa, 0x00, 0xd7, 0xdd, 0x00, 0x00, /* 0x00e8-0x00ef */
 491     0x00, 0x00, 0xdf, 0xe3, 0xab, 0xe2, 0x00, 0x00, /* 0x00f0-0x00f7 */
 492     0x00, 0xef, 0xf3, 0x00, 0x00, 0xfd, 0x00, 0x00, /* 0x00f8-0x00ff */
 493     0x00, 0x00, 0xa1, 0xa8, 0x00, 0x00, 0x00, 0x00, /* 0x0100-0x0107 */
 494     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0108-0x010f */
 495     0xa7, 0xae, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0110-0x0117 */
 496     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0118-0x011f */
 497     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0120-0x0127 */
 498     0x8f, 0xdc,                                     /* 0x0128-0x0129 */
 499
 500 #define FROM_IDX_02 (FROM_IDX_01 + 106)
 501     0x9f, 0xf2,                                     /* 0x0168-0x0169 */
 502
 503 #define FROM_IDX_03 (FROM_IDX_02 + 2)
 504     0xa5, 0xac, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x01a0-0x01a7 */
 505     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xa6, /* 0x01a8-0x01af */
 506     0xad,                                           /* 0x01b0-0x01b0 */
 507
 508 #define FROM_IDX_04 (FROM_IDX_03 + 17)
 509     0xb0, 0xb3, 0x00, 0xb2, 0x00, 0x00, 0x00, 0x00, /* 0x0300-0x0307 */
 510     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0308-0x030f */
 511     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0310-0x0317 */
 512     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0318-0x031f */
 513     0x00, 0x00, 0x00, 0xb4,                         /* 0x0320-0x0323 */
 514
 515 #define FROM_IDX_05 (FROM_IDX_04 + 36)
 516     0x84, 0xb9, 0x81, 0xb6, 0xc4, 0xca, 0xc1, 0xc7, /* 0x1ea0-0x1ea7 */
 517     0xc2, 0xc8, 0xc3, 0xc9, 0x86, 0xcb, 0xc0, 0xbe, /* 0x1ea8-0x1eaf */
 518     0xaf, 0xbb, 0xba, 0xbc, 0xbf, 0xbd, 0x85, 0xc6, /* 0x1eb0-0x1eb7 */
 519     0x8b, 0xd1, 0x88, 0xce, 0x89, 0xcf, 0xda, 0xd5, /* 0x1eb8-0x1ebf */
 520     0xc5, 0xd2, 0xcd, 0xd3, 0xd9, 0xd4, 0x8c, 0xd6, /* 0x1ec0-0x1ec7 */
 521     0x8e, 0xd8, 0x91, 0xde, 0x96, 0xe4, 0x93, 0xe1, /* 0x1ec8-0x1ecf */
 522     0xff, 0xe8, 0xdb, 0xe5, 0xe0, 0xe6, 0xf0, 0xe7, /* 0x1ed0-0x1ed7 */
 523     0x97, 0xe9, 0x9b, 0xed, 0x98, 0xea, 0x99, 0xeb, /* 0x1ed8-0x1edf */
 524     0x9a, 0xec, 0x9c, 0xee, 0x02, 0xf4, 0x9e, 0xf1, /* 0x1ee0-0x1ee7 */
 525     0x11, 0xf8, 0x04, 0xf5, 0x05, 0xf6, 0x06, 0xf7, /* 0x1ee8-0x1eef */
 526     0x12, 0xf9, 0x13, 0xfa, 0x17, 0xfe, 0x14, 0xfb, /* 0x1ef0-0x1ef7 */
 527     0x15, 0xfc                                      /* 0x1ef8-0x1ef9 */
 528 #define FROM_IDX_END (FROM_IDX_05 + 90)
 529   };
 530 /* Compile-time verification of table size.  */
 531 typedef int verify2[(NELEMS (from_ucs4) == FROM_IDX_END) - 1];
 532
 533 /* Decomposition table for the relevant Unicode characters. */
 534 static const struct
 535 {
 536   uint16_t composed;
 537   uint32_t base:8;
 538   uint32_t comb1:8;
 539 } decomp_table[] =
 540   {
 541     { 0x00d1, 0x4e, 0xb2 },
 542     { 0x00f1, 0x6e, 0xb2 },
 543     { 0x0106, 0x43, 0xb3 },
 544     { 0x0107, 0x63, 0xb3 },
 545     { 0x0139, 0x4c, 0xb3 },
 546     { 0x013a, 0x6c, 0xb3 },
 547     { 0x0143, 0x4e, 0xb3 },
 548     { 0x0144, 0x6e, 0xb3 },
 549     { 0x0154, 0x52, 0xb3 },
 550     { 0x0155, 0x72, 0xb3 },
 551     { 0x015a, 0x53, 0xb3 },
 552     { 0x015b, 0x73, 0xb3 },
 553     { 0x0179, 0x5a, 0xb3 },
 554     { 0x017a, 0x7a, 0xb3 },
 555     { 0x01f4, 0x47, 0xb3 },
 556     { 0x01f5, 0x67, 0xb3 },
 557     { 0x01f8, 0x4e, 0xb0 },
 558     { 0x01f9, 0x6e, 0xb0 },
 559     { 0x1e04, 0x42, 0xb4 },
 560     { 0x1e05, 0x62, 0xb4 },
 561     { 0x1e0c, 0x44, 0xb4 },
 562     { 0x1e0d, 0x64, 0xb4 },
 563     { 0x1e24, 0x48, 0xb4 },
 564     { 0x1e25, 0x68, 0xb4 },
 565     { 0x1e30, 0x4b, 0xb3 },
 566     { 0x1e31, 0x6b, 0xb3 },
 567     { 0x1e32, 0x4b, 0xb4 },
 568     { 0x1e33, 0x6b, 0xb4 },
 569     { 0x1e36, 0x4c, 0xb4 },
 570     { 0x1e37, 0x6c, 0xb4 },
 571     { 0x1e3e, 0x4d, 0xb3 },
 572     { 0x1e3f, 0x6d, 0xb3 },
 573     { 0x1e42, 0x4d, 0xb4 },
 574     { 0x1e43, 0x6d, 0xb4 },
 575     { 0x1e46, 0x4e, 0xb4 },
 576     { 0x1e47, 0x6e, 0xb4 },
 577     { 0x1e54, 0x50, 0xb3 },
 578     { 0x1e55, 0x70, 0xb3 },
 579     { 0x1e5a, 0x52, 0xb4 },
 580     { 0x1e5b, 0x72, 0xb4 },
 581     { 0x1e62, 0x53, 0xb4 },
 582     { 0x1e63, 0x73, 0xb4 },
 583     { 0x1e6c, 0x54, 0xb4 },
 584     { 0x1e6d, 0x74, 0xb4 },
 585     { 0x1e7c, 0x56, 0xb2 },
 586     { 0x1e7d, 0x76, 0xb2 },
 587     { 0x1e7e, 0x56, 0xb4 },
 588     { 0x1e7f, 0x76, 0xb4 },
 589     { 0x1e80, 0x57, 0xb0 },
 590     { 0x1e81, 0x77, 0xb0 },
 591     { 0x1e82, 0x57, 0xb3 },
 592     { 0x1e83, 0x77, 0xb3 },
 593     { 0x1e88, 0x57, 0xb4 },
 594     { 0x1e89, 0x77, 0xb4 },
 595     { 0x1e92, 0x5a, 0xb4 },
 596     { 0x1e93, 0x7a, 0xb4 },
 597   };
 598
 599
 600 /* Next, define the other direction.  */
 601 #define MIN_NEEDED_INPUT        MIN_NEEDED_TO
 602 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_FROM
 603 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_FROM
 604 #define LOOPFCT                 TO_LOOP
 605 #define BODY \
 606   {                                                                           \
 607     uint32_t ch = get32 (inptr);                                              \
 608                                                                               \
 609     if (ch == 0x00 || (ch >= 0x18 && ch < 0x80) || ch == 0xa0)                \
 610       {                                                                       \
 611         *outptr++ = ch;                                                       \
 612         inptr += 4;                                                           \
 613       }                                                                       \
 614     else                                                                      \
 615       {                                                                       \
 616         unsigned char res;                                                    \
 617                                                                               \
 618         if (ch <= 0x0010)                                                     \
 619           res = from_ucs4[ch - 0x0001 + FROM_IDX_00];                         \
 620         else if (ch >= 0x00c0 && ch <= 0x0129)                                \
 621           res = from_ucs4[ch - 0x00c0 + FROM_IDX_01];                         \
 622         else if (ch >= 0x0168 && ch <= 0x0169)                                \
 623           res = from_ucs4[ch - 0x0168 + FROM_IDX_02];                         \
 624         else if (ch >= 0x01a0 && ch <= 0x01b0)                                \
 625           res = from_ucs4[ch - 0x01a0 + FROM_IDX_03];                         \
 626         else if (ch >= 0x0300 && ch <= 0x0323)                                \
 627           res = from_ucs4[ch - 0x0300 + FROM_IDX_04];                         \
 628         else if (ch >= 0x1ea0 && ch <= 0x1ef9)                                \
 629           res = from_ucs4[ch - 0x1ea0 + FROM_IDX_05];                         \
 630         else                                                                  \
 631           {                                                                   \
 632             UNICODE_TAG_HANDLER (ch, 4);                                      \
 633             res = 0;                                                          \
 634           }                                                                   \
 635                                                                               \
 636         if (__builtin_expect (res != 0, 1))                                   \
 637           {                                                                   \
 638             *outptr++ = res;                                                  \
 639             inptr += 4;                                                       \
 640           }                                                                   \
 641         else                                                                  \
 642           {                                                                   \
 643             /* Try canonical decomposition.  */                               \
 644             unsigned int i1;                                                  \
 645             unsigned int i2;                                                  \
 646                                                                               \
 647             i1 = 0;                                                           \
 648             i2 = sizeof (decomp_table) / sizeof (decomp_table[0]) - 1;        \
 649             if (ch >= decomp_table[i1].composed                               \
 650                 && ch <= decomp_table[i2].composed)                           \
 651               {                                                               \
 652                 unsigned int i;                                               \
 653                                                                               \
 654                 for (;;)                                                      \
 655                   {                                                           \
 656                     i = (i1 + i2) >> 1;                                       \
 657                     if (ch == decomp_table[i].composed)                       \
 658                       break;                                                  \
 659                     if (ch < decomp_table[i].composed)                        \
 660                       {                                                       \
 661                         if (i1 == i)                                          \
 662                           goto failed;                                        \
 663                         i2 = i;                                               \
 664                       }                                                       \
 665                     else                                                      \
 666                       {                                                       \
 667                         if (i1 != i)                                          \
 668                           i1 = i;                                             \
 669                         else                                                  \
 670                           {                                                   \
 671                             i = i2;                                           \
 672                             if (ch == decomp_table[i].composed)               \
 673                               break;                                          \
 674                             goto failed;                                      \
 675                           }                                                   \
 676                       }                                                       \
 677                   }                                                           \
 678                                                                               \
 679                 /* See whether we have room for two bytes.  */                \
 680                 if (__builtin_expect (outptr + 1 >= outend, 0))               \
 681                   {                                                           \
 682                     result = __GCONV_FULL_OUTPUT;                             \
 683                     break;                                                    \
 684                   }                                                           \
 685                                                                               \
 686                 /* Found a canonical decomposition.  */                       \
 687                 *outptr++ = decomp_table[i].base;                             \
 688                 *outptr++ = decomp_table[i].comb1;                            \
 689                 inptr += 4;                                                   \
 690                 continue;                                                     \
 691               }                                                               \
 692                                                                               \
 693           failed:                                                             \
 694             /* This is an illegal character.  */                              \
 695             STANDARD_ERR_HANDLER (4);                                         \
 696           }                                                                   \
 697       }                                                                       \
 698   }
 699 #define LOOP_NEED_FLAGS
 700 #define EXTRA_LOOP_DECLS        , int *statep
 701 #include <iconv/loop.c>
 702
 703
 704 /* Now define the toplevel functions.  */
 705 #include <iconv/skeleton.c>