iconvdata/euc-jisx0213.c

   1 /* Conversion from and to EUC-JISX0213.
   2    Copyright (C) 2002-2017 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4    Contributed by Bruno Haible <bruno@clisp.org>, 2002.
   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, see
  18    <http://www.gnu.org/licenses/>.  */
  19
  20 #include <dlfcn.h>
  21 #include <stdint.h>
  22 #include <gconv.h>
  23
  24 /* The structure of EUC-JISX0213 is as follows:
  25
  26    0x00..0x7F: ASCII
  27
  28    0x8E{A1..FE}: JISX0201 Katakana, with prefix 0x8E, offset by +0x80.
  29
  30    0x8F{A1..FE}{A1..FE}: JISX0213 plane 2, with prefix 0x8F, offset by +0x8080.
  31
  32    0x{A1..FE}{A1..FE}: JISX0213 plane 1, offset by +0x8080.
  33
  34    Note that some JISX0213 characters are not contained in Unicode 3.2
  35    and are therefore best represented as sequences of Unicode characters.
  36 */
  37
  38 #include "jisx0213.h"
  39
  40 /* Definitions used in the body of the `gconv' function.  */
  41 #define CHARSET_NAME            "EUC-JISX0213//"
  42 #define FROM_LOOP               from_euc_jisx0213
  43 #define TO_LOOP                 to_euc_jisx0213
  44 #define DEFINE_INIT             1
  45 #define DEFINE_FINI             1
  46 #define ONE_DIRECTION           0
  47 #define FROM_LOOP_MIN_NEEDED_FROM       1
  48 #define FROM_LOOP_MAX_NEEDED_FROM       3
  49 #define FROM_LOOP_MIN_NEEDED_TO         4
  50 #define FROM_LOOP_MAX_NEEDED_TO         8
  51 #define TO_LOOP_MIN_NEEDED_FROM         4
  52 #define TO_LOOP_MAX_NEEDED_FROM         4
  53 #define TO_LOOP_MIN_NEEDED_TO           1
  54 #define TO_LOOP_MAX_NEEDED_TO           3
  55 #define PREPARE_LOOP \
  56   int saved_state;                                                            \
  57   int *statep = &data->__statep->__count;
  58 #define EXTRA_LOOP_ARGS         , statep
  59
  60
  61 /* Since we might have to reset input pointer we must be able to save
  62    and restore the state.  */
  63 #define SAVE_RESET_STATE(Save) \
  64   if (Save)                                                                   \
  65     saved_state = *statep;                                                    \
  66   else                                                                        \
  67     *statep = saved_state
  68
  69
  70 /* During EUC-JISX0213 to UCS-4 conversion, the COUNT element of the state
  71    contains the last UCS-4 character, shifted by 3 bits.
  72    During UCS-4 to EUC-JISX0213 conversion, the COUNT element of the state
  73    contains the last two bytes to be output, shifted by 3 bits.  */
  74
  75 /* Since this is a stateful encoding we have to provide code which resets
  76    the output state to the initial state.  This has to be done during the
  77    flushing.  */
  78 #define EMIT_SHIFT_TO_INIT \
  79   if (data->__statep->__count != 0)                                           \
  80     {                                                                         \
  81       if (FROM_DIRECTION)                                                     \
  82         {                                                                     \
  83           if (__glibc_likely (outbuf + 4 <= outend))                          \
  84             {                                                                 \
  85               /* Write out the last character.  */                            \
  86               *((uint32_t *) outbuf) = data->__statep->__count >> 3;          \
  87               outbuf += sizeof (uint32_t);                                    \
  88               data->__statep->__count = 0;                                    \
  89             }                                                                 \
  90           else                                                                \
  91             /* We don't have enough room in the output buffer.  */            \
  92             status = __GCONV_FULL_OUTPUT;                                     \
  93         }                                                                     \
  94       else                                                                    \
  95         {                                                                     \
  96           if (__glibc_likely (outbuf + 2 <= outend))                          \
  97             {                                                                 \
  98               /* Write out the last character.  */                            \
  99               uint32_t lasttwo = data->__statep->__count >> 3;                \
 100               *outbuf++ = (lasttwo >> 8) & 0xff;                              \
 101               *outbuf++ = lasttwo & 0xff;                                     \
 102               data->__statep->__count = 0;                                    \
 103             }                                                                 \
 104           else                                                                \
 105             /* We don't have enough room in the output buffer.  */            \
 106             status = __GCONV_FULL_OUTPUT;                                     \
 107         }                                                                     \
 108     }
 109
 110
 111 /* First define the conversion function from EUC-JISX0213 to UCS-4.  */
 112 #define MIN_NEEDED_INPUT        FROM_LOOP_MIN_NEEDED_FROM
 113 #define MAX_NEEDED_INPUT        FROM_LOOP_MAX_NEEDED_FROM
 114 #define MIN_NEEDED_OUTPUT       FROM_LOOP_MIN_NEEDED_TO
 115 #define MAX_NEEDED_OUTPUT       FROM_LOOP_MAX_NEEDED_TO
 116 #define LOOPFCT                 FROM_LOOP
 117 #define BODY \
 118   {                                                                           \
 119     uint32_t ch;                                                              \
 120                                                                               \
 121     /* Determine whether there is a buffered character pending.  */           \
 122     ch = *statep >> 3;                                                        \
 123     if (__glibc_likely (ch == 0))                                             \
 124       {                                                                       \
 125         /* No - so look at the next input byte.  */                           \
 126         ch = *inptr;                                                          \
 127                                                                               \
 128         if (ch < 0x80)                                                        \
 129           /* Plain ASCII character.  */                                       \
 130           ++inptr;                                                            \
 131         else if ((ch >= 0xa1 && ch <= 0xfe) || ch == 0x8e || ch == 0x8f)      \
 132           {                                                                   \
 133             /* Two or three byte character.  */                               \
 134             uint32_t ch2;                                                     \
 135                                                                               \
 136             if (__glibc_unlikely (inptr + 1 >= inend))                        \
 137               {                                                               \
 138                 /* The second byte is not available.  */                      \
 139                 result = __GCONV_INCOMPLETE_INPUT;                            \
 140                 break;                                                        \
 141               }                                                               \
 142                                                                               \
 143             ch2 = inptr[1];                                                   \
 144                                                                               \
 145             /* The second byte must be >= 0xa1 and <= 0xfe.  */               \
 146             if (__glibc_unlikely (ch2 < 0xa1 || ch2 > 0xfe))                  \
 147               {                                                               \
 148                 /* This is an illegal character.  */                          \
 149                 STANDARD_FROM_LOOP_ERR_HANDLER (1);                           \
 150               }                                                               \
 151                                                                               \
 152             if (ch == 0x8e)                                                   \
 153               {                                                               \
 154                 /* Half-width katakana.  */                                   \
 155                 if (__glibc_unlikely (ch2 > 0xdf))                            \
 156                   STANDARD_FROM_LOOP_ERR_HANDLER (1);                         \
 157                                                                               \
 158                 ch = ch2 + 0xfec0;                                            \
 159                 inptr += 2;                                                   \
 160               }                                                               \
 161             else                                                              \
 162               {                                                               \
 163                 const unsigned char *endp;                                    \
 164                                                                               \
 165                 if (ch == 0x8f)                                               \
 166                   {                                                           \
 167                     /* JISX 0213 plane 2.  */                                 \
 168                     uint32_t ch3;                                             \
 169                                                                               \
 170                     if (__glibc_unlikely (inptr + 2 >= inend))                \
 171                       {                                                       \
 172                         /* The third byte is not available.  */               \
 173                         result = __GCONV_INCOMPLETE_INPUT;                    \
 174                         break;                                                \
 175                       }                                                       \
 176                                                                               \
 177                     ch3 = inptr[2];                                           \
 178                     endp = inptr + 3;                                         \
 179                                                                               \
 180                     ch = jisx0213_to_ucs4 (0x200 - 0x80 + ch2, ch3 ^ 0x80);   \
 181                   }                                                           \
 182                 else                                                          \
 183                   {                                                           \
 184                     /* JISX 0213 plane 1.  */                                 \
 185                     endp = inptr + 2;                                         \
 186                                                                               \
 187                     ch = jisx0213_to_ucs4 (0x100 - 0x80 + ch, ch2 ^ 0x80);    \
 188                   }                                                           \
 189                                                                               \
 190                 if (ch == 0)                                                  \
 191                   /* This is an illegal character.  */                        \
 192                   STANDARD_FROM_LOOP_ERR_HANDLER (1);                         \
 193                                                                               \
 194                 inptr = endp;                                                 \
 195                                                                               \
 196                 if (ch < 0x80)                                                \
 197                   {                                                           \
 198                     /* It's a combining character.  */                        \
 199                     uint32_t u1 = __jisx0213_to_ucs_combining[ch - 1][0];     \
 200                     uint32_t u2 = __jisx0213_to_ucs_combining[ch - 1][1];     \
 201                                                                               \
 202                     put32 (outptr, u1);                                       \
 203                     outptr += 4;                                              \
 204                                                                               \
 205                     /* See whether we have room for two characters.  */       \
 206                     if (outptr + 4 <= outend)                                 \
 207                       {                                                       \
 208                         put32 (outptr, u2);                                   \
 209                         outptr += 4;                                          \
 210                         continue;                                             \
 211                       }                                                       \
 212                                                                               \
 213                     /* Otherwise store only the first character now, and      \
 214                        put the second one into the queue.  */                 \
 215                     *statep = u2 << 3;                                        \
 216                     /* Tell the caller why we terminate the loop.  */         \
 217                     result = __GCONV_FULL_OUTPUT;                             \
 218                     break;                                                    \
 219                   }                                                           \
 220               }                                                               \
 221           }                                                                   \
 222         else                                                                  \
 223           {                                                                   \
 224             /* This is illegal.  */                                           \
 225             STANDARD_FROM_LOOP_ERR_HANDLER (1);                               \
 226           }                                                                   \
 227       }                                                                       \
 228                                                                               \
 229     put32 (outptr, ch);                                                       \
 230     outptr += 4;                                                              \
 231   }
 232 #define LOOP_NEED_FLAGS
 233 #define EXTRA_LOOP_DECLS        , int *statep
 234 #define ONEBYTE_BODY \
 235   {                                                                           \
 236     if (c < 0x80)                                                             \
 237       return c;                                                               \
 238     else                                                                      \
 239       return WEOF;                                                            \
 240   }
 241 #include <iconv/loop.c>
 242
 243
 244 /* Next, define the other direction, from UCS-4 to EUC-JISX0213.  */
 245
 246 /* Composition tables for each of the relevant combining characters.  */
 247 static const struct
 248 {
 249   uint16_t base;
 250   uint16_t composed;
 251 } comp_table_data[] =
 252 {
 253 #define COMP_TABLE_IDX_02E5 0
 254 #define COMP_TABLE_LEN_02E5 1
 255   { 0xabe4, 0xabe5 }, /* 0x12B65 = 0x12B64 U+02E5 */
 256 #define COMP_TABLE_IDX_02E9 (COMP_TABLE_IDX_02E5 + COMP_TABLE_LEN_02E5)
 257 #define COMP_TABLE_LEN_02E9 1
 258   { 0xabe0, 0xabe6 }, /* 0x12B66 = 0x12B60 U+02E9 */
 259 #define COMP_TABLE_IDX_0300 (COMP_TABLE_IDX_02E9 + COMP_TABLE_LEN_02E9)
 260 #define COMP_TABLE_LEN_0300 5
 261   { 0xa9dc, 0xabc4 }, /* 0x12B44 = 0x1295C U+0300 */
 262   { 0xabb8, 0xabc8 }, /* 0x12B48 = 0x12B38 U+0300 */
 263   { 0xabb7, 0xabca }, /* 0x12B4A = 0x12B37 U+0300 */
 264   { 0xabb0, 0xabcc }, /* 0x12B4C = 0x12B30 U+0300 */
 265   { 0xabc3, 0xabce }, /* 0x12B4E = 0x12B43 U+0300 */
 266 #define COMP_TABLE_IDX_0301 (COMP_TABLE_IDX_0300 + COMP_TABLE_LEN_0300)
 267 #define COMP_TABLE_LEN_0301 4
 268   { 0xabb8, 0xabc9 }, /* 0x12B49 = 0x12B38 U+0301 */
 269   { 0xabb7, 0xabcb }, /* 0x12B4B = 0x12B37 U+0301 */
 270   { 0xabb0, 0xabcd }, /* 0x12B4D = 0x12B30 U+0301 */
 271   { 0xabc3, 0xabcf }, /* 0x12B4F = 0x12B43 U+0301 */
 272 #define COMP_TABLE_IDX_309A (COMP_TABLE_IDX_0301 + COMP_TABLE_LEN_0301)
 273 #define COMP_TABLE_LEN_309A 14
 274   { 0xa4ab, 0xa4f7 }, /* 0x12477 = 0x1242B U+309A */
 275   { 0xa4ad, 0xa4f8 }, /* 0x12478 = 0x1242D U+309A */
 276   { 0xa4af, 0xa4f9 }, /* 0x12479 = 0x1242F U+309A */
 277   { 0xa4b1, 0xa4fa }, /* 0x1247A = 0x12431 U+309A */
 278   { 0xa4b3, 0xa4fb }, /* 0x1247B = 0x12433 U+309A */
 279   { 0xa5ab, 0xa5f7 }, /* 0x12577 = 0x1252B U+309A */
 280   { 0xa5ad, 0xa5f8 }, /* 0x12578 = 0x1252D U+309A */
 281   { 0xa5af, 0xa5f9 }, /* 0x12579 = 0x1252F U+309A */
 282   { 0xa5b1, 0xa5fa }, /* 0x1257A = 0x12531 U+309A */
 283   { 0xa5b3, 0xa5fb }, /* 0x1257B = 0x12533 U+309A */
 284   { 0xa5bb, 0xa5fc }, /* 0x1257C = 0x1253B U+309A */
 285   { 0xa5c4, 0xa5fd }, /* 0x1257D = 0x12544 U+309A */
 286   { 0xa5c8, 0xa5fe }, /* 0x1257E = 0x12548 U+309A */
 287   { 0xa6f5, 0xa6f8 }, /* 0x12678 = 0x12675 U+309A */
 288 };
 289
 290 #define MIN_NEEDED_INPUT        TO_LOOP_MIN_NEEDED_FROM
 291 #define MAX_NEEDED_INPUT        TO_LOOP_MAX_NEEDED_FROM
 292 #define MIN_NEEDED_OUTPUT       TO_LOOP_MIN_NEEDED_TO
 293 #define MAX_NEEDED_OUTPUT       TO_LOOP_MAX_NEEDED_TO
 294 #define LOOPFCT                 TO_LOOP
 295 #define BODY \
 296   {                                                                           \
 297     uint32_t ch = get32 (inptr);                                              \
 298                                                                               \
 299     if ((*statep >> 3) != 0)                                                  \
 300       {                                                                       \
 301         /* Attempt to combine the last character with this one.  */           \
 302         uint16_t lasttwo = *statep >> 3;                                      \
 303         unsigned int idx;                                                     \
 304         unsigned int len;                                                     \
 305                                                                               \
 306         if (ch == 0x02e5)                                                     \
 307           idx = COMP_TABLE_IDX_02E5, len = COMP_TABLE_LEN_02E5;               \
 308         else if (ch == 0x02e9)                                                \
 309           idx = COMP_TABLE_IDX_02E9, len = COMP_TABLE_LEN_02E9;               \
 310         else if (ch == 0x0300)                                                \
 311           idx = COMP_TABLE_IDX_0300, len = COMP_TABLE_LEN_0300;               \
 312         else if (ch == 0x0301)                                                \
 313           idx = COMP_TABLE_IDX_0301, len = COMP_TABLE_LEN_0301;               \
 314         else if (ch == 0x309a)                                                \
 315           idx = COMP_TABLE_IDX_309A, len = COMP_TABLE_LEN_309A;               \
 316         else                                                                  \
 317           goto not_combining;                                                 \
 318                                                                               \
 319         do                                                                    \
 320           if (comp_table_data[idx].base == lasttwo)                           \
 321             break;                                                            \
 322         while (++idx, --len > 0);                                             \
 323                                                                               \
 324         if (len > 0)                                                          \
 325           {                                                                   \
 326             /* Output the combined character.  */                             \
 327             if (__glibc_unlikely (outptr + 1 >= outend))                      \
 328               {                                                               \
 329                 result = __GCONV_FULL_OUTPUT;                                 \
 330                 break;                                                        \
 331               }                                                               \
 332             lasttwo = comp_table_data[idx].composed;                          \
 333             *outptr++ = (lasttwo >> 8) & 0xff;                                \
 334             *outptr++ = lasttwo & 0xff;                                       \
 335             *statep = 0;                                                      \
 336             inptr += 4;                                                       \
 337             continue;                                                         \
 338           }                                                                   \
 339                                                                               \
 340       not_combining:                                                          \
 341         /* Output the buffered character.  */                                 \
 342         if (__glibc_unlikely (outptr + 1 >= outend))                          \
 343           {                                                                   \
 344             result = __GCONV_FULL_OUTPUT;                                     \
 345             break;                                                            \
 346           }                                                                   \
 347         *outptr++ = (lasttwo >> 8) & 0xff;                                    \
 348         *outptr++ = lasttwo & 0xff;                                           \
 349         *statep = 0;                                                          \
 350         continue;                                                             \
 351       }                                                                       \
 352                                                                               \
 353     if (ch < 0x80)                                                            \
 354       /* Plain ASCII character.  */                                           \
 355       *outptr++ = ch;                                                         \
 356     else if (ch >= 0xff61 && ch <= 0xff9f)                                    \
 357       {                                                                       \
 358         /* Half-width katakana.  */                                           \
 359         if (__glibc_unlikely (outptr + 1 >= outend))                          \
 360           {                                                                   \
 361             result = __GCONV_FULL_OUTPUT;                                     \
 362             break;                                                            \
 363           }                                                                   \
 364         *outptr++ = 0x8e;                                                     \
 365         *outptr++ = ch - 0xfec0;                                              \
 366       }                                                                       \
 367     else                                                                      \
 368       {                                                                       \
 369         uint32_t jch = ucs4_to_jisx0213 (ch);                                 \
 370         if (jch == 0)                                                         \
 371           {                                                                   \
 372             UNICODE_TAG_HANDLER (ch, 4);                                      \
 373                                                                               \
 374             /* Illegal character.  */                                         \
 375             STANDARD_TO_LOOP_ERR_HANDLER (4);                                 \
 376           }                                                                   \
 377                                                                               \
 378         if (jch & 0x0080)                                                     \
 379           {                                                                   \
 380             /* A possible match in comp_table_data.  We have to buffer it.  */\
 381                                                                               \
 382             /* We know it's a JISX 0213 plane 1 character.  */                \
 383             assert ((jch & 0x8000) == 0);                                     \
 384                                                                               \
 385             *statep = (jch | 0x8080) << 3;                                    \
 386             inptr += 4;                                                       \
 387             continue;                                                         \
 388           }                                                                   \
 389                                                                               \
 390         if (jch & 0x8000)                                                     \
 391           {                                                                   \
 392             /* JISX 0213 plane 2.  */                                         \
 393             if (__glibc_unlikely (outptr + 2 >= outend))                      \
 394               {                                                               \
 395                 result = __GCONV_FULL_OUTPUT;                                 \
 396                 break;                                                        \
 397               }                                                               \
 398             *outptr++ = 0x8f;                                                 \
 399           }                                                                   \
 400         else                                                                  \
 401           {                                                                   \
 402             /* JISX 0213 plane 1.  */                                         \
 403             if (__glibc_unlikely (outptr + 1 >= outend))                      \
 404               {                                                               \
 405                 result = __GCONV_FULL_OUTPUT;                                 \
 406                 break;                                                        \
 407               }                                                               \
 408           }                                                                   \
 409         *outptr++ = (jch >> 8) | 0x80;                                        \
 410         *outptr++ = (jch & 0xff) | 0x80;                                      \
 411       }                                                                       \
 412                                                                               \
 413     inptr += 4;                                                               \
 414   }
 415 #define LOOP_NEED_FLAGS
 416 #define EXTRA_LOOP_DECLS        , int *statep
 417 #include <iconv/loop.c>
 418
 419
 420 /* Now define the toplevel functions.  */
 421 #include <iconv/skeleton.c>