1/* 2 * Copyright (C) 1999-2004, 2008 Free Software Foundation, Inc. 3 * This file is part of the GNU LIBICONV Library. 4 * 5 * The GNU LIBICONV Library is free software; you can redistribute it 6 * and/or modify it under the terms of the GNU Library General Public 7 * License as published by the Free Software Foundation; either version 2 8 * of the License, or (at your option) any later version. 9 * 10 * The GNU LIBICONV Library is distributed in the hope that it will be 11 * useful, but WITHOUT ANY WARRANTY; without even the implied warranty of 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 13 * Library General Public License for more details. 14 * 15 * You should have received a copy of the GNU Library General Public 16 * License along with the GNU LIBICONV Library; see the file COPYING.LIB. 17 * If not, write to the Free Software Foundation, Inc., 51 Franklin Street, 18 * Fifth Floor, Boston, MA 02110-1301, USA. 19 */ 20 21/* 22 * ISO-2022-JP-3 23 */ 24 25#include "jisx0213.h" 26 27#define ESC 0x1b 28 29/* 30 * The state is composed of one of the following values 31 */ 32#define STATE_ASCII 0 /* Esc ( B */ 33#define STATE_JISX0201ROMAN 1 /* Esc ( J */ 34#define STATE_JISX0201KATAKANA 2 /* Esc ( I */ 35#define STATE_JISX0208 3 /* Esc $ @ or Esc $ B */ 36#define STATE_JISX02131 4 /* Esc $ ( O or Esc $ ( Q*/ 37#define STATE_JISX02132 5 /* Esc $ ( P */ 38 39/* 40 * In the ISO-2022-JP-3 to UCS-4 direction, the state also holds the last 41 * character to be output, shifted by 3 bits. 42 */ 43 44static int 45iso2022_jp3_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n) 46{ 47 ucs4_t last_wc = conv->istate >> 3; 48 if (last_wc) { 49 /* Output the buffered character. */ 50 conv->istate &= 7; 51 *pwc = last_wc; 52 return 0; /* Don't advance the input pointer. */ 53 } else { 54 state_t state = conv->istate; 55 int count = 0; 56 unsigned char c; 57 for (;;) { 58 c = *s; 59 if (c == ESC) { 60 if (n < count+3) 61 goto none; 62 if (s[1] == '(') { 63 if (s[2] == 'B') { 64 state = STATE_ASCII; 65 s += 3; count += 3; 66 if (n < count+1) 67 goto none; 68 continue; 69 } 70 if (s[2] == 'J') { 71 state = STATE_JISX0201ROMAN; 72 s += 3; count += 3; 73 if (n < count+1) 74 goto none; 75 continue; 76 } 77 if (s[2] == 'I') { 78 state = STATE_JISX0201KATAKANA; 79 s += 3; count += 3; 80 if (n < count+1) 81 goto none; 82 continue; 83 } 84 goto ilseq; 85 } 86 if (s[1] == '$') { 87 if (s[2] == '@' || s[2] == 'B') { 88 /* We don't distinguish JIS X 0208-1978 and JIS X 0208-1983. */ 89 state = STATE_JISX0208; 90 s += 3; count += 3; 91 if (n < count+1) 92 goto none; 93 continue; 94 } 95 if (s[2] == '(') { 96 if (n < count+4) 97 goto none; 98 if (s[3] == 'O' || s[3] == 'Q') { 99 state = STATE_JISX02131; 100 s += 4; count += 4; 101 if (n < count+1) 102 goto none; 103 continue; 104 } 105 if (s[3] == 'P') { 106 state = STATE_JISX02132; 107 s += 4; count += 4; 108 if (n < count+1) 109 goto none; 110 continue; 111 } 112 } 113 goto ilseq; 114 } 115 goto ilseq; 116 } 117 break; 118 } 119 switch (state) { 120 case STATE_ASCII: 121 if (c < 0x80) { 122 int ret = ascii_mbtowc(conv,pwc,s,1); 123 if (ret == RET_ILSEQ) 124 goto ilseq; 125 if (ret != 1) abort(); 126 conv->istate = state; 127 return count+1; 128 } else 129 goto ilseq; 130 case STATE_JISX0201ROMAN: 131 if (c < 0x80) { 132 int ret = jisx0201_mbtowc(conv,pwc,s,1); 133 if (ret == RET_ILSEQ) 134 goto ilseq; 135 if (ret != 1) abort(); 136 conv->istate = state; 137 return count+1; 138 } else 139 goto ilseq; 140 case STATE_JISX0201KATAKANA: 141 if (c < 0x80) { 142 unsigned char buf = c+0x80; 143 int ret = jisx0201_mbtowc(conv,pwc,&buf,1); 144 if (ret == RET_ILSEQ) 145 goto ilseq; 146 if (ret != 1) abort(); 147 conv->istate = state; 148 return count+1; 149 } else 150 goto ilseq; 151 case STATE_JISX0208: 152 if (n < count+2) 153 goto none; 154 if (s[0] < 0x80 && s[1] < 0x80) { 155 int ret = jisx0208_mbtowc(conv,pwc,s,2); 156 if (ret == RET_ILSEQ) 157 goto ilseq; 158 if (ret != 2) abort(); 159 conv->istate = state; 160 return count+2; 161 } else 162 goto ilseq; 163 case STATE_JISX02131: 164 case STATE_JISX02132: 165 if (n < count+2) 166 goto none; 167 if (s[0] < 0x80 && s[1] < 0x80) { 168 ucs4_t wc = jisx0213_to_ucs4(((state-STATE_JISX02131+1)<<8)+s[0],s[1]); 169 if (wc) { 170 if (wc < 0x80) { 171 /* It's a combining character. */ 172 ucs4_t wc1 = jisx0213_to_ucs_combining[wc - 1][0]; 173 ucs4_t wc2 = jisx0213_to_ucs_combining[wc - 1][1]; 174 /* We cannot output two Unicode characters at once. So, 175 output the first character and buffer the second one. */ 176 *pwc = wc1; 177 conv->istate = (wc2 << 3) | state; 178 } else { 179 *pwc = wc; 180 conv->istate = state; 181 } 182 return count+2; 183 } 184 } 185 goto ilseq; 186 default: abort(); 187 } 188 none: 189 conv->istate = state; 190 return RET_TOOFEW(count); 191 192 ilseq: 193 conv->istate = state; 194 return RET_SHIFT_ILSEQ(count); 195 } 196} 197 198static int 199iso2022_jp3_flushwc (conv_t conv, ucs4_t *pwc) 200{ 201 ucs4_t last_wc = conv->istate >> 3; 202 if (last_wc) { 203 /* Output the buffered character. */ 204 conv->istate &= 7; 205 *pwc = last_wc; 206 return 1; 207 } else 208 return 0; 209} 210 211/* 212 * In the UCS-4 to ISO-2022-JP-3 direction, the state also holds the last two 213 * bytes to be output, shifted by 3 bits, and the STATE_xxxxx value that was 214 * effective before this buffered character, shifted by 19 bits. 215 */ 216 217/* Composition tables for each of the relevant combining characters. */ 218static const struct { unsigned short base; unsigned short composed; } iso2022_jp3_comp_table_data[] = { 219#define iso2022_jp3_comp_table02e5_idx 0 220#define iso2022_jp3_comp_table02e5_len 1 221 { 0x2b64, 0x2b65 }, /* 0x12B65 = 0x12B64 U+02E5 */ 222#define iso2022_jp3_comp_table02e9_idx (iso2022_jp3_comp_table02e5_idx+iso2022_jp3_comp_table02e5_len) 223#define iso2022_jp3_comp_table02e9_len 1 224 { 0x2b60, 0x2b66 }, /* 0x12B66 = 0x12B60 U+02E9 */ 225#define iso2022_jp3_comp_table0300_idx (iso2022_jp3_comp_table02e9_idx+iso2022_jp3_comp_table02e9_len) 226#define iso2022_jp3_comp_table0300_len 5 227 { 0x295c, 0x2b44 }, /* 0x12B44 = 0x1295C U+0300 */ 228 { 0x2b38, 0x2b48 }, /* 0x12B48 = 0x12B38 U+0300 */ 229 { 0x2b37, 0x2b4a }, /* 0x12B4A = 0x12B37 U+0300 */ 230 { 0x2b30, 0x2b4c }, /* 0x12B4C = 0x12B30 U+0300 */ 231 { 0x2b43, 0x2b4e }, /* 0x12B4E = 0x12B43 U+0300 */ 232#define iso2022_jp3_comp_table0301_idx (iso2022_jp3_comp_table0300_idx+iso2022_jp3_comp_table0300_len) 233#define iso2022_jp3_comp_table0301_len 4 234 { 0x2b38, 0x2b49 }, /* 0x12B49 = 0x12B38 U+0301 */ 235 { 0x2b37, 0x2b4b }, /* 0x12B4B = 0x12B37 U+0301 */ 236 { 0x2b30, 0x2b4d }, /* 0x12B4D = 0x12B30 U+0301 */ 237 { 0x2b43, 0x2b4f }, /* 0x12B4F = 0x12B43 U+0301 */ 238#define iso2022_jp3_comp_table309a_idx (iso2022_jp3_comp_table0301_idx+iso2022_jp3_comp_table0301_len) 239#define iso2022_jp3_comp_table309a_len 14 240 { 0x242b, 0x2477 }, /* 0x12477 = 0x1242B U+309A */ 241 { 0x242d, 0x2478 }, /* 0x12478 = 0x1242D U+309A */ 242 { 0x242f, 0x2479 }, /* 0x12479 = 0x1242F U+309A */ 243 { 0x2431, 0x247a }, /* 0x1247A = 0x12431 U+309A */ 244 { 0x2433, 0x247b }, /* 0x1247B = 0x12433 U+309A */ 245 { 0x252b, 0x2577 }, /* 0x12577 = 0x1252B U+309A */ 246 { 0x252d, 0x2578 }, /* 0x12578 = 0x1252D U+309A */ 247 { 0x252f, 0x2579 }, /* 0x12579 = 0x1252F U+309A */ 248 { 0x2531, 0x257a }, /* 0x1257A = 0x12531 U+309A */ 249 { 0x2533, 0x257b }, /* 0x1257B = 0x12533 U+309A */ 250 { 0x253b, 0x257c }, /* 0x1257C = 0x1253B U+309A */ 251 { 0x2544, 0x257d }, /* 0x1257D = 0x12544 U+309A */ 252 { 0x2548, 0x257e }, /* 0x1257E = 0x12548 U+309A */ 253 { 0x2675, 0x2678 }, /* 0x12678 = 0x12675 U+309A */ 254}; 255 256#define SPLIT_STATE \ 257 unsigned short lasttwo = state >> 3; state_t prevstate = state >> 19; state &= 7 258#define COMBINE_STATE \ 259 state |= (prevstate << 19) | (lasttwo << 3) 260#define COMBINE_STATE_NO_LASTTWO \ 261 /* assume lasttwo == 0, then prevstate is ignored */ 262 263static int 264iso2022_jp3_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n) 265{ 266 int count = 0; 267 unsigned char buf[2]; 268 unsigned short jch; 269 int ret; 270 state_t state = conv->ostate; 271 SPLIT_STATE; 272 273 if (lasttwo) { 274 /* Attempt to combine the last character with this one. */ 275 unsigned int idx; 276 unsigned int len; 277 278 if (wc == 0x02e5) 279 idx = iso2022_jp3_comp_table02e5_idx, 280 len = iso2022_jp3_comp_table02e5_len; 281 else if (wc == 0x02e9) 282 idx = iso2022_jp3_comp_table02e9_idx, 283 len = iso2022_jp3_comp_table02e9_len; 284 else if (wc == 0x0300) 285 idx = iso2022_jp3_comp_table0300_idx, 286 len = iso2022_jp3_comp_table0300_len; 287 else if (wc == 0x0301) 288 idx = iso2022_jp3_comp_table0301_idx, 289 len = iso2022_jp3_comp_table0301_len; 290 else if (wc == 0x309a) 291 idx = iso2022_jp3_comp_table309a_idx, 292 len = iso2022_jp3_comp_table309a_len; 293 else 294 goto not_combining; 295 296 do 297 if (iso2022_jp3_comp_table_data[idx].base == lasttwo) 298 break; 299 while (++idx, --len > 0); 300 301 if (len > 0) { 302 /* Output the combined character. */ 303 /* We know the combined character is in JISX0213 plane 1, but 304 the buffered character may have been in JISX0208 or in 305 JISX0213 plane 1. */ 306 count = (state != STATE_JISX02131 ? 4 : 0) + 2; 307 if (n < count) 308 return RET_TOOSMALL; 309 if (state != STATE_JISX02131) { 310 r[0] = ESC; 311 r[1] = '$'; 312 r[2] = '('; 313 r[3] = 'Q'; 314 r += 4; 315 state = STATE_JISX02131; 316 } 317 lasttwo = iso2022_jp3_comp_table_data[idx].composed; 318 r[0] = (lasttwo >> 8) & 0xff; 319 r[1] = lasttwo & 0xff; 320 COMBINE_STATE_NO_LASTTWO; 321 conv->ostate = state; 322 return count; 323 } 324 325 not_combining: 326 /* Output the buffered character. */ 327 /* We know it is in JISX0208 or in JISX0213 plane 1. */ 328 count = (prevstate != state ? 3 : 0) + 2; 329 if (n < count) 330 return RET_TOOSMALL; 331 if (prevstate != state) { 332 if (state != STATE_JISX0208) abort(); 333 r[0] = ESC; 334 r[1] = '$'; 335 r[2] = 'B'; 336 r += 3; 337 } 338 r[0] = (lasttwo >> 8) & 0xff; 339 r[1] = lasttwo & 0xff; 340 r += 2; 341 } 342 343 /* Try ASCII. */ 344 ret = ascii_wctomb(conv,buf,wc,1); 345 if (ret != RET_ILUNI) { 346 if (ret != 1) abort(); 347 if (buf[0] < 0x80) { 348 count += (state == STATE_ASCII ? 1 : 4); 349 if (n < count) 350 return RET_TOOSMALL; 351 if (state != STATE_ASCII) { 352 r[0] = ESC; 353 r[1] = '('; 354 r[2] = 'B'; 355 r += 3; 356 state = STATE_ASCII; 357 } 358 r[0] = buf[0]; 359 COMBINE_STATE_NO_LASTTWO; 360 conv->ostate = state; 361 return count; 362 } 363 } 364 365 /* Try JIS X 0201-1976 Roman. */ 366 ret = jisx0201_wctomb(conv,buf,wc,1); 367 if (ret != RET_ILUNI) { 368 if (ret != 1) abort(); 369 if (buf[0] < 0x80) { 370 count += (state == STATE_JISX0201ROMAN ? 1 : 4); 371 if (n < count) 372 return RET_TOOSMALL; 373 if (state != STATE_JISX0201ROMAN) { 374 r[0] = ESC; 375 r[1] = '('; 376 r[2] = 'J'; 377 r += 3; 378 state = STATE_JISX0201ROMAN; 379 } 380 r[0] = buf[0]; 381 COMBINE_STATE_NO_LASTTWO; 382 conv->ostate = state; 383 return count; 384 } 385 } 386 387 jch = ucs4_to_jisx0213(wc); 388 389 /* Try JIS X 0208-1990 in place of JIS X 0208-1978 and JIS X 0208-1983. */ 390 ret = jisx0208_wctomb(conv,buf,wc,2); 391 if (ret != RET_ILUNI) { 392 if (ret != 2) abort(); 393 if (buf[0] < 0x80 && buf[1] < 0x80) { 394 if (jch & 0x0080) { 395 /* A possible match in comp_table_data. Buffer it. */ 396 prevstate = state; 397 lasttwo = jch & 0x7f7f; 398 state = STATE_JISX0208; 399 COMBINE_STATE; 400 conv->ostate = state; 401 return count; 402 } else { 403 count += (state == STATE_JISX0208 ? 2 : 5); 404 if (n < count) 405 return RET_TOOSMALL; 406 if (state != STATE_JISX0208) { 407 r[0] = ESC; 408 r[1] = '$'; 409 r[2] = 'B'; 410 r += 3; 411 state = STATE_JISX0208; 412 } 413 r[0] = buf[0]; 414 r[1] = buf[1]; 415 COMBINE_STATE_NO_LASTTWO; 416 conv->ostate = state; 417 return count; 418 } 419 } 420 } 421 422 /* Try JISX 0213 plane 1 and JISX 0213 plane 2. */ 423 if (jch != 0) { 424 if (jch & 0x8000) { 425 /* JISX 0213 plane 2. */ 426 if (state != STATE_JISX02132) { 427 count += 4; 428 if (n < count) 429 return RET_TOOSMALL; 430 r[0] = ESC; 431 r[1] = '$'; 432 r[2] = '('; 433 r[3] = 'P'; 434 r += 4; 435 state = STATE_JISX02132; 436 } 437 } else { 438 /* JISX 0213 plane 1. */ 439 if (state != STATE_JISX02131) { 440 count += 4; 441 if (n < count) 442 return RET_TOOSMALL; 443 r[0] = ESC; 444 r[1] = '$'; 445 r[2] = '('; 446 r[3] = 'Q'; 447 r += 4; 448 state = STATE_JISX02131; 449 } 450 } 451 if (jch & 0x0080) { 452 /* A possible match in comp_table_data. We have to buffer it. */ 453 /* We know it's a JISX 0213 plane 1 character. */ 454 if (jch & 0x8000) abort(); 455 prevstate = state; 456 lasttwo = jch & 0x7f7f; 457 COMBINE_STATE; 458 conv->ostate = state; 459 return count; 460 } 461 count += 2; 462 if (n < count) 463 return RET_TOOSMALL; 464 r[0] = (jch >> 8) & 0x7f; 465 r[1] = jch & 0x7f; 466 COMBINE_STATE_NO_LASTTWO; 467 conv->ostate = state; 468 return count; 469 } 470 471 /* Try JIS X 0201-1976 Katakana. This is not officially part of 472 ISO-2022-JP-3. Therefore we try it after all other attempts. */ 473 ret = jisx0201_wctomb(conv,buf,wc,1); 474 if (ret != RET_ILUNI) { 475 if (ret != 1) abort(); 476 if (buf[0] >= 0x80) { 477 count += (state == STATE_JISX0201KATAKANA ? 1 : 4); 478 if (n < count) 479 return RET_TOOSMALL; 480 if (state != STATE_JISX0201KATAKANA) { 481 r[0] = ESC; 482 r[1] = '('; 483 r[2] = 'I'; 484 r += 3; 485 state = STATE_JISX0201KATAKANA; 486 } 487 r[0] = buf[0]-0x80; 488 COMBINE_STATE_NO_LASTTWO; 489 conv->ostate = state; 490 return count; 491 } 492 } 493 494 return RET_ILUNI; 495} 496 497static int 498iso2022_jp3_reset (conv_t conv, unsigned char *r, int n) 499{ 500 state_t state = conv->ostate; 501 SPLIT_STATE; 502 { 503 int count = 504 (lasttwo ? (prevstate != state ? 3 : 0) + 2 : 0) 505 + (state != STATE_ASCII ? 3 : 0); 506 if (n < count) 507 return RET_TOOSMALL; 508 if (lasttwo) { 509 if (prevstate != state) { 510 if (state != STATE_JISX0208) abort(); 511 r[0] = ESC; 512 r[1] = '$'; 513 r[2] = 'B'; 514 r += 3; 515 } 516 r[0] = (lasttwo >> 8) & 0xff; 517 r[1] = lasttwo & 0xff; 518 r += 2; 519 } 520 if (state != STATE_ASCII) { 521 r[0] = ESC; 522 r[1] = '('; 523 r[2] = 'B'; 524 } 525 /* conv->ostate = 0; will be done by the caller */ 526 return count; 527 } 528} 529 530#undef COMBINE_STATE_NO_LASTTWO 531#undef COMBINE_STATE 532#undef SPLIT_STATE 533#undef STATE_JISX02132 534#undef STATE_JISX02131 535#undef STATE_JISX0208 536#undef STATE_JISX0201KATAKANA 537#undef STATE_JISX0201ROMAN 538#undef STATE_ASCII 539