iconvdata/euc-jisx0213.c

   1 /* Conversion from and to EUC-JISX0213.
   2    Copyright (C) 2002-2023 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the License, or (at your option) any later version.
   9
  10    The GNU C Library is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with the GNU C Library; if not, see
  17    <https://www.gnu.org/licenses/>.  */
  18
  19 #include <dlfcn.h>
  20 #include <stdint.h>
  21 #include <gconv.h>
  22
  23 /* The structure of EUC-JISX0213 is as follows:
  24
  25    0x00..0x7F: ASCII
  26
  27    0x8E{A1..FE}: JISX0201 Katakana, with prefix 0x8E, offset by +0x80.
  28
  29    0x8F{A1..FE}{A1..FE}: JISX0213 plane 2, with prefix 0x8F, offset by +0x8080.
  30
  31    0x{A1..FE}{A1..FE}: JISX0213 plane 1, offset by +0x8080.
  32
  33    Note that some JISX0213 characters are not contained in Unicode 3.2
  34    and are therefore best represented as sequences of Unicode characters.
  35 */
  36
  37 #include "jisx0213.h"
  38
  39 /* Definitions used in the body of the `gconv' function.  */
  40 #define CHARSET_NAME            "EUC-JISX0213//"
  41 #define FROM_LOOP               from_euc_jisx0213
  42 #define TO_LOOP                 to_euc_jisx0213
  43 #define DEFINE_INIT             1
  44 #define DEFINE_FINI             1
  45 #define ONE_DIRECTION           0
  46 #define FROM_LOOP_MIN_NEEDED_FROM       1
  47 #define FROM_LOOP_MAX_NEEDED_FROM       3
  48 #define FROM_LOOP_MIN_NEEDED_TO         4
  49 #define FROM_LOOP_MAX_NEEDED_TO         8
  50 #define TO_LOOP_MIN_NEEDED_FROM         4
  51 #define TO_LOOP_MAX_NEEDED_FROM         4
  52 #define TO_LOOP_MIN_NEEDED_TO           1
  53 #define TO_LOOP_MAX_NEEDED_TO           3
  54 #define PREPARE_LOOP \
  55   int saved_state;                                                            \
  56   int *statep = &data->__statep->__count;
  57 #define EXTRA_LOOP_ARGS         , statep
  58
  59
  60 /* Since we might have to reset input pointer we must be able to save
  61    and restore the state.  */
  62 #define SAVE_RESET_STATE(Save) \
  63   if (Save)                                                                   \
  64     saved_state = *statep;                                                    \
  65   else                                                                        \
  66     *statep = saved_state
  67
  68
  69 /* During EUC-JISX0213 to UCS-4 conversion, the COUNT element of the state
  70    contains the last UCS-4 character, shifted by 3 bits.
  71    During UCS-4 to EUC-JISX0213 conversion, the COUNT element of the state
  72    contains the last two bytes to be output, shifted by 3 bits.  */
  73
  74 /* Since this is a stateful encoding we have to provide code which resets
  75    the output state to the initial state.  This has to be done during the
  76    flushing.  */
  77 #define EMIT_SHIFT_TO_INIT \
  78   if (data->__statep->__count != 0)                                           \
  79     {                                                                         \
  80       if (FROM_DIRECTION)                                                     \
  81         {                                                                     \
  82           if (__glibc_likely (outbuf + 4 <= outend))                          \
  83             {                                                                 \
  84               /* Write out the last character.  */                            \
  85               *((uint32_t *) outbuf) = data->__statep->__count >> 3;          \
  86               outbuf += sizeof (uint32_t);                                    \
  87               data->__statep->__count = 0;                                    \
  88             }                                                                 \
  89           else                                                                \
  90             /* We don't have enough room in the output buffer.  */            \
  91             status = __GCONV_FULL_OUTPUT;                                     \
  92         }                                                                     \
  93       else                                                                    \
  94         {                                                                     \
  95           if (__glibc_likely (outbuf + 2 <= outend))                          \
  96             {                                                                 \
  97               /* Write out the last character.  */                            \
  98               uint32_t lasttwo = data->__statep->__count >> 3;                \
  99               *outbuf++ = (lasttwo >> 8) & 0xff;                              \
 100               *outbuf++ = lasttwo & 0xff;                                     \
 101               data->__statep->__count = 0;                                    \
 102             }                                                                 \
 103           else                                                                \
 104             /* We don't have enough room in the output buffer.  */            \
 105             status = __GCONV_FULL_OUTPUT;                                     \
 106         }                                                                     \
 107     }
 108
 109
 110 /* First define the conversion function from EUC-JISX0213 to UCS-4.  */
 111 #define MIN_NEEDED_INPUT        FROM_LOOP_MIN_NEEDED_FROM
 112 #define MAX_NEEDED_INPUT        FROM_LOOP_MAX_NEEDED_FROM
 113 #define MIN_NEEDED_OUTPUT       FROM_LOOP_MIN_NEEDED_TO
 114 #define MAX_NEEDED_OUTPUT       FROM_LOOP_MAX_NEEDED_TO
 115 #define LOOPFCT                 FROM_LOOP
 116 #define BODY \
 117   {                                                                           \
 118     uint32_t ch;                                                              \
 119                                                                               \
 120     /* Determine whether there is a buffered character pending.  */           \
 121     ch = *statep >> 3;                                                        \
 122     if (__glibc_likely (ch == 0))                                             \
 123       {                                                                       \
 124         /* No - so look at the next input byte.  */                           \
 125         ch = *inptr;                                                          \
 126                                                                               \
 127         if (ch < 0x80)                                                        \
 128           /* Plain ASCII character.  */                                       \
 129           ++inptr;                                                            \
 130         else if ((ch >= 0xa1 && ch <= 0xfe) || ch == 0x8e || ch == 0x8f)      \
 131           {                                                                   \
 132             /* Two or three byte character.  */                               \
 133             uint32_t ch2;                                                     \
 134                                                                               \
 135             if (__glibc_unlikely (inptr + 1 >= inend))                        \
 136               {                                                               \
 137                 /* The second byte is not available.  */                      \
 138                 result = __GCONV_INCOMPLETE_INPUT;                            \
 139                 break;                                                        \
 140               }                                                               \
 141                                                                               \
 142             ch2 = inptr[1];                                                   \
 143                                                                               \
 144             /* The second byte must be >= 0xa1 and <= 0xfe.  */               \
 145             if (__glibc_unlikely (ch2 < 0xa1 || ch2 > 0xfe))                  \
 146               {                                                               \
 147                 /* This is an illegal character.  */                          \
 148                 STANDARD_FROM_LOOP_ERR_HANDLER (1);                           \
 149               }                                                               \
 150                                                                               \
 151             if (ch == 0x8e)                                                   \
 152               {                                                               \
 153                 /* Half-width katakana.  */                                   \
 154                 if (__glibc_unlikely (ch2 > 0xdf))                            \
 155                   STANDARD_FROM_LOOP_ERR_HANDLER (1);                         \
 156                                                                               \
 157                 ch = ch2 + 0xfec0;                                            \
 158                 inptr += 2;                                                   \
 159               }                                                               \
 160             else                                                              \
 161               {                                                               \
 162                 const unsigned char *endp;                                    \
 163                                                                               \
 164                 if (ch == 0x8f)                                               \
 165                   {                                                           \
 166                     /* JISX 0213 plane 2.  */                                 \
 167                     uint32_t ch3;                                             \
 168                                                                               \
 169                     if (__glibc_unlikely (inptr + 2 >= inend))                \
 170                       {                                                       \
 171                         /* The third byte is not available.  */               \
 172                         result = __GCONV_INCOMPLETE_INPUT;                    \
 173                         break;                                                \
 174                       }                                                       \
 175                                                                               \
 176                     ch3 = inptr[2];                                           \
 177                     endp = inptr + 3;                                         \
 178                                                                               \
 179                     ch = jisx0213_to_ucs4 (0x200 - 0x80 + ch2, ch3 ^ 0x80);   \
 180                   }                                                           \
 181                 else                                                          \
 182                   {                                                           \
 183                     /* JISX 0213 plane 1.  */                                 \
 184                     endp = inptr + 2;                                         \
 185                                                                               \
 186                     ch = jisx0213_to_ucs4 (0x100 - 0x80 + ch, ch2 ^ 0x80);    \
 187                   }                                                           \
 188                                                                               \
 189                 if (ch == 0)                                                  \
 190                   /* This is an illegal character.  */                        \
 191                   STANDARD_FROM_LOOP_ERR_HANDLER (1);                         \
 192                                                                               \
 193                 inptr = endp;                                                 \
 194                                                                               \
 195                 if (ch < 0x80)                                                \
 196                   {                                                           \
 197                     /* It's a combining character.  */                        \
 198                     uint32_t u1 = __jisx0213_to_ucs_combining[ch - 1][0];     \
 199                     uint32_t u2 = __jisx0213_to_ucs_combining[ch - 1][1];     \
 200                                                                               \
 201                     put32 (outptr, u1);                                       \
 202                     outptr += 4;                                              \
 203                                                                               \
 204                     /* See whether we have room for two characters.  */       \
 205                     if (outptr + 4 <= outend)                                 \
 206                       {                                                       \
 207                         put32 (outptr, u2);                                   \
 208                         outptr += 4;                                          \
 209                         continue;                                             \
 210                       }                                                       \
 211                                                                               \
 212                     /* Otherwise store only the first character now, and      \
 213                        put the second one into the queue.  */                 \
 214                     *statep = u2 << 3;                                        \
 215                     /* Tell the caller why we terminate the loop.  */         \
 216                     result = __GCONV_FULL_OUTPUT;                             \
 217                     break;                                                    \
 218                   }                                                           \
 219               }                                                               \
 220           }                                                                   \
 221         else                                                                  \
 222           {                                                                   \
 223             /* This is illegal.  */                                           \
 224             STANDARD_FROM_LOOP_ERR_HANDLER (1);                               \
 225           }                                                                   \
 226       }                                                                       \
 227                                                                               \
 228     put32 (outptr, ch);                                                       \
 229     outptr += 4;                                                              \
 230   }
 231 #define LOOP_NEED_FLAGS
 232 #define EXTRA_LOOP_DECLS        , int *statep
 233 #define ONEBYTE_BODY \
 234   {                                                                           \
 235     if (c < 0x80)                                                             \
 236       return c;                                                               \
 237     else                                                                      \
 238       return WEOF;                                                            \
 239   }
 240 #include <iconv/loop.c>
 241
 242
 243 /* Next, define the other direction, from UCS-4 to EUC-JISX0213.  */
 244
 245 /* Composition tables for each of the relevant combining characters.  */
 246 static const struct
 247 {
 248   uint16_t base;
 249   uint16_t composed;
 250 } comp_table_data[] =
 251 {
 252 #define COMP_TABLE_IDX_02E5 0
 253 #define COMP_TABLE_LEN_02E5 1
 254   { 0xabe4, 0xabe5 }, /* 0x12B65 = 0x12B64 U+02E5 */
 255 #define COMP_TABLE_IDX_02E9 (COMP_TABLE_IDX_02E5 + COMP_TABLE_LEN_02E5)
 256 #define COMP_TABLE_LEN_02E9 1
 257   { 0xabe0, 0xabe6 }, /* 0x12B66 = 0x12B60 U+02E9 */
 258 #define COMP_TABLE_IDX_0300 (COMP_TABLE_IDX_02E9 + COMP_TABLE_LEN_02E9)
 259 #define COMP_TABLE_LEN_0300 5
 260   { 0xa9dc, 0xabc4 }, /* 0x12B44 = 0x1295C U+0300 */
 261   { 0xabb8, 0xabc8 }, /* 0x12B48 = 0x12B38 U+0300 */
 262   { 0xabb7, 0xabca }, /* 0x12B4A = 0x12B37 U+0300 */
 263   { 0xabb0, 0xabcc }, /* 0x12B4C = 0x12B30 U+0300 */
 264   { 0xabc3, 0xabce }, /* 0x12B4E = 0x12B43 U+0300 */
 265 #define COMP_TABLE_IDX_0301 (COMP_TABLE_IDX_0300 + COMP_TABLE_LEN_0300)
 266 #define COMP_TABLE_LEN_0301 4
 267   { 0xabb8, 0xabc9 }, /* 0x12B49 = 0x12B38 U+0301 */
 268   { 0xabb7, 0xabcb }, /* 0x12B4B = 0x12B37 U+0301 */
 269   { 0xabb0, 0xabcd }, /* 0x12B4D = 0x12B30 U+0301 */
 270   { 0xabc3, 0xabcf }, /* 0x12B4F = 0x12B43 U+0301 */
 271 #define COMP_TABLE_IDX_309A (COMP_TABLE_IDX_0301 + COMP_TABLE_LEN_0301)
 272 #define COMP_TABLE_LEN_309A 14
 273   { 0xa4ab, 0xa4f7 }, /* 0x12477 = 0x1242B U+309A */
 274   { 0xa4ad, 0xa4f8 }, /* 0x12478 = 0x1242D U+309A */
 275   { 0xa4af, 0xa4f9 }, /* 0x12479 = 0x1242F U+309A */
 276   { 0xa4b1, 0xa4fa }, /* 0x1247A = 0x12431 U+309A */
 277   { 0xa4b3, 0xa4fb }, /* 0x1247B = 0x12433 U+309A */
 278   { 0xa5ab, 0xa5f7 }, /* 0x12577 = 0x1252B U+309A */
 279   { 0xa5ad, 0xa5f8 }, /* 0x12578 = 0x1252D U+309A */
 280   { 0xa5af, 0xa5f9 }, /* 0x12579 = 0x1252F U+309A */
 281   { 0xa5b1, 0xa5fa }, /* 0x1257A = 0x12531 U+309A */
 282   { 0xa5b3, 0xa5fb }, /* 0x1257B = 0x12533 U+309A */
 283   { 0xa5bb, 0xa5fc }, /* 0x1257C = 0x1253B U+309A */
 284   { 0xa5c4, 0xa5fd }, /* 0x1257D = 0x12544 U+309A */
 285   { 0xa5c8, 0xa5fe }, /* 0x1257E = 0x12548 U+309A */
 286   { 0xa6f5, 0xa6f8 }, /* 0x12678 = 0x12675 U+309A */
 287 };
 288
 289 #define MIN_NEEDED_INPUT        TO_LOOP_MIN_NEEDED_FROM
 290 #define MAX_NEEDED_INPUT        TO_LOOP_MAX_NEEDED_FROM
 291 #define MIN_NEEDED_OUTPUT       TO_LOOP_MIN_NEEDED_TO
 292 #define MAX_NEEDED_OUTPUT       TO_LOOP_MAX_NEEDED_TO
 293 #define LOOPFCT                 TO_LOOP
 294 #define BODY \
 295   {                                                                           \
 296     uint32_t ch = get32 (inptr);                                              \
 297                                                                               \
 298     if ((*statep >> 3) != 0)                                                  \
 299       {                                                                       \
 300         /* Attempt to combine the last character with this one.  */           \
 301         uint16_t lasttwo = *statep >> 3;                                      \
 302         unsigned int idx;                                                     \
 303         unsigned int len;                                                     \
 304                                                                               \
 305         if (ch == 0x02e5)                                                     \
 306           idx = COMP_TABLE_IDX_02E5, len = COMP_TABLE_LEN_02E5;               \
 307         else if (ch == 0x02e9)                                                \
 308           idx = COMP_TABLE_IDX_02E9, len = COMP_TABLE_LEN_02E9;               \
 309         else if (ch == 0x0300)                                                \
 310           idx = COMP_TABLE_IDX_0300, len = COMP_TABLE_LEN_0300;               \
 311         else if (ch == 0x0301)                                                \
 312           idx = COMP_TABLE_IDX_0301, len = COMP_TABLE_LEN_0301;               \
 313         else if (ch == 0x309a)                                                \
 314           idx = COMP_TABLE_IDX_309A, len = COMP_TABLE_LEN_309A;               \
 315         else                                                                  \
 316           goto not_combining;                                                 \
 317                                                                               \
 318         do                                                                    \
 319           if (comp_table_data[idx].base == lasttwo)                           \
 320             break;                                                            \
 321         while (++idx, --len > 0);                                             \
 322                                                                               \
 323         if (len > 0)                                                          \
 324           {                                                                   \
 325             /* Output the combined character.  */                             \
 326             if (__glibc_unlikely (outptr + 1 >= outend))                      \
 327               {                                                               \
 328                 result = __GCONV_FULL_OUTPUT;                                 \
 329                 break;                                                        \
 330               }                                                               \
 331             lasttwo = comp_table_data[idx].composed;                          \
 332             *outptr++ = (lasttwo >> 8) & 0xff;                                \
 333             *outptr++ = lasttwo & 0xff;                                       \
 334             *statep = 0;                                                      \
 335             inptr += 4;                                                       \
 336             continue;                                                         \
 337           }                                                                   \
 338                                                                               \
 339       not_combining:                                                          \
 340         /* Output the buffered character.  */                                 \
 341         if (__glibc_unlikely (outptr + 1 >= outend))                          \
 342           {                                                                   \
 343             result = __GCONV_FULL_OUTPUT;                                     \
 344             break;                                                            \
 345           }                                                                   \
 346         *outptr++ = (lasttwo >> 8) & 0xff;                                    \
 347         *outptr++ = lasttwo & 0xff;                                           \
 348         *statep = 0;                                                          \
 349         continue;                                                             \
 350       }                                                                       \
 351                                                                               \
 352     if (ch < 0x80)                                                            \
 353       /* Plain ASCII character.  */                                           \
 354       *outptr++ = ch;                                                         \
 355     else if (ch >= 0xff61 && ch <= 0xff9f)                                    \
 356       {                                                                       \
 357         /* Half-width katakana.  */                                           \
 358         if (__glibc_unlikely (outptr + 1 >= outend))                          \
 359           {                                                                   \
 360             result = __GCONV_FULL_OUTPUT;                                     \
 361             break;                                                            \
 362           }                                                                   \
 363         *outptr++ = 0x8e;                                                     \
 364         *outptr++ = ch - 0xfec0;                                              \
 365       }                                                                       \
 366     else                                                                      \
 367       {                                                                       \
 368         uint32_t jch = ucs4_to_jisx0213 (ch);                                 \
 369         if (jch == 0)                                                         \
 370           {                                                                   \
 371             UNICODE_TAG_HANDLER (ch, 4);                                      \
 372                                                                               \
 373             /* Illegal character.  */                                         \
 374             STANDARD_TO_LOOP_ERR_HANDLER (4);                                 \
 375           }                                                                   \
 376                                                                               \
 377         if (jch & 0x0080)                                                     \
 378           {                                                                   \
 379             /* A possible match in comp_table_data.  We have to buffer it.  */\
 380                                                                               \
 381             /* We know it's a JISX 0213 plane 1 character.  */                \
 382             assert ((jch & 0x8000) == 0);                                     \
 383                                                                               \
 384             *statep = (jch | 0x8080) << 3;                                    \
 385             inptr += 4;                                                       \
 386             continue;                                                         \
 387           }                                                                   \
 388                                                                               \
 389         if (jch & 0x8000)                                                     \
 390           {                                                                   \
 391             /* JISX 0213 plane 2.  */                                         \
 392             if (__glibc_unlikely (outptr + 2 >= outend))                      \
 393               {                                                               \
 394                 result = __GCONV_FULL_OUTPUT;                                 \
 395                 break;                                                        \
 396               }                                                               \
 397             *outptr++ = 0x8f;                                                 \
 398           }                                                                   \
 399         else                                                                  \
 400           {                                                                   \
 401             /* JISX 0213 plane 1.  */                                         \
 402             if (__glibc_unlikely (outptr + 1 >= outend))                      \
 403               {                                                               \
 404                 result = __GCONV_FULL_OUTPUT;                                 \
 405                 break;                                                        \
 406               }                                                               \
 407           }                                                                   \
 408         *outptr++ = (jch >> 8) | 0x80;                                        \
 409         *outptr++ = (jch & 0xff) | 0x80;                                      \
 410       }                                                                       \
 411                                                                               \
 412     inptr += 4;                                                               \
 413   }
 414 #define LOOP_NEED_FLAGS
 415 #define EXTRA_LOOP_DECLS        , int *statep
 416 #include <iconv/loop.c>
 417
 418
 419 /* Now define the toplevel functions.  */
 420 #include <iconv/skeleton.c>