1 /* $NetBSD: encoding.c,v 1.13 2026/06/10 20:54:16 christos Exp $ */ 2 3 /* 4 * Copyright (c) Ian F. Darwin 1986-1995. 5 * Software written by Ian F. Darwin and others; 6 * maintained 1995-present by Christos Zoulas and others. 7 * 8 * Redistribution and use in source and binary forms, with or without 9 * modification, are permitted provided that the following conditions 10 * are met: 11 * 1. Redistributions of source code must retain the above copyright 12 * notice immediately at the beginning of the file, without modification, 13 * this list of conditions, and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 19 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 20 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 21 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR 22 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 23 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 24 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 25 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 28 * SUCH DAMAGE. 29 */ 30 /* 31 * Encoding -- determine the character encoding of a text file. 32 * 33 * Joerg Wunsch <joerg (at) freebsd.org> wrote the original support for 8-bit 34 * international characters. 35 */ 36 37 #include "file.h" 38 39 #ifndef lint 40 #if 0 41 FILE_RCSID("@(#)$File: encoding.c,v 1.45 2026/06/02 17:19:02 christos Exp $") 42 #else 43 __RCSID("$NetBSD: encoding.c,v 1.13 2026/06/10 20:54:16 christos Exp $"); 44 #endif 45 #endif /* lint */ 46 47 #include "magic.h" 48 #include <string.h> 49 #include <stdlib.h> 50 51 52 file_private int looks_ascii(const unsigned char *, size_t, file_unichar_t *, 53 size_t *); 54 file_private int looks_utf8_with_BOM(const unsigned char *, size_t, file_unichar_t *, 55 size_t *); 56 file_private int looks_utf7(const unsigned char *, size_t, file_unichar_t *, 57 size_t *); 58 file_private int looks_ucs16(const unsigned char *, size_t, file_unichar_t *, 59 size_t *); 60 file_private int looks_ucs32(const unsigned char *, size_t, file_unichar_t *, 61 size_t *); 62 file_private int looks_latin1(const unsigned char *, size_t, file_unichar_t *, 63 size_t *); 64 file_private int looks_extended(const unsigned char *, size_t, file_unichar_t *, 65 size_t *); 66 file_private void from_ebcdic(const unsigned char *, size_t, unsigned char *); 67 68 #ifdef DEBUG_ENCODING 69 #define DPRINTF(a) printf a 70 #else 71 #define DPRINTF(a) 72 #endif 73 74 /* 75 * Try to determine whether text is in some character code we can 76 * identify. Each of these tests, if it succeeds, will leave 77 * the text converted into one-file_unichar_t-per-character Unicode in 78 * ubuf, and the number of characters converted in ulen. 79 */ 80 file_protected int 81 file_encoding(struct magic_set *ms, const struct buffer *b, 82 file_unichar_t **ubuf, size_t *ulen, const char **code, 83 const char **code_mime, const char **type) 84 { 85 const unsigned char *buf = CAST(const unsigned char *, b->fbuf); 86 size_t nbytes = b->flen; 87 size_t mlen; 88 int rv = 1, ucs_type; 89 file_unichar_t *udefbuf; 90 size_t udeflen; 91 92 if (buf == NULL) 93 return 0; 94 if (ubuf == NULL) 95 ubuf = &udefbuf; 96 if (ulen == NULL) 97 ulen = &udeflen; 98 99 *type = "text"; 100 *ulen = 0; 101 *code = "unknown"; 102 *code_mime = "binary"; 103 104 if (nbytes > ms->encoding_max) 105 nbytes = ms->encoding_max; 106 107 mlen = (nbytes + 1) * sizeof((*ubuf)[0]); 108 *ubuf = CAST(file_unichar_t *, calloc(CAST(size_t, 1), mlen)); 109 if (*ubuf == NULL) { 110 file_oomem(ms, mlen); 111 goto done; 112 } 113 if (looks_ascii(buf, nbytes, *ubuf, ulen)) { 114 if (looks_utf7(buf, nbytes, *ubuf, ulen) > 0) { 115 DPRINTF(("utf-7 %" SIZE_T_FORMAT "u\n", *ulen)); 116 *code = "Unicode text, UTF-7"; 117 *code_mime = "utf-7"; 118 } else { 119 DPRINTF(("ascii %" SIZE_T_FORMAT "u\n", *ulen)); 120 *code = "ASCII"; 121 *code_mime = "us-ascii"; 122 } 123 } else if (looks_utf8_with_BOM(buf, nbytes, *ubuf, ulen) > 0) { 124 DPRINTF(("utf8/bom %" SIZE_T_FORMAT "u\n", *ulen)); 125 *code = "Unicode text, UTF-8 (with BOM)"; 126 *code_mime = "utf-8"; 127 } else if (file_looks_utf8(buf, nbytes, *ubuf, ulen) > 1) { 128 DPRINTF(("utf8 %" SIZE_T_FORMAT "u\n", *ulen)); 129 *code = "Unicode text, UTF-8"; 130 *code_mime = "utf-8"; 131 } else if ((ucs_type = looks_ucs32(buf, nbytes, *ubuf, ulen)) != 0) { 132 if (ucs_type == 1) { 133 *code = "Unicode text, UTF-32, little-endian"; 134 *code_mime = "utf-32le"; 135 } else { 136 *code = "Unicode text, UTF-32, big-endian"; 137 *code_mime = "utf-32be"; 138 } 139 DPRINTF(("ucs32 %" SIZE_T_FORMAT "u\n", *ulen)); 140 } else if ((ucs_type = looks_ucs16(buf, nbytes, *ubuf, ulen)) != 0) { 141 if (ucs_type == 1) { 142 *code = "Unicode text, UTF-16, little-endian"; 143 *code_mime = "utf-16le"; 144 } else { 145 *code = "Unicode text, UTF-16, big-endian"; 146 *code_mime = "utf-16be"; 147 } 148 DPRINTF(("ucs16 %" SIZE_T_FORMAT "u\n", *ulen)); 149 } else if (looks_latin1(buf, nbytes, *ubuf, ulen)) { 150 DPRINTF(("latin1 %" SIZE_T_FORMAT "u\n", *ulen)); 151 *code = "ISO-8859"; 152 *code_mime = "iso-8859-1"; 153 } else if (looks_extended(buf, nbytes, *ubuf, ulen)) { 154 DPRINTF(("extended %" SIZE_T_FORMAT "u\n", *ulen)); 155 *code = "Non-ISO extended-ASCII"; 156 *code_mime = "unknown-8bit"; 157 } else { 158 unsigned char *nbuf; 159 160 mlen = (nbytes + 1) * sizeof(nbuf[0]); 161 if ((nbuf = CAST(unsigned char *, malloc(mlen))) == NULL) { 162 file_oomem(ms, mlen); 163 goto done; 164 } 165 from_ebcdic(buf, nbytes, nbuf); 166 167 if (looks_ascii(nbuf, nbytes, *ubuf, ulen)) { 168 DPRINTF(("ebcdic %" SIZE_T_FORMAT "u\n", *ulen)); 169 *code = "EBCDIC"; 170 *code_mime = "ebcdic"; 171 } else if (looks_latin1(nbuf, nbytes, *ubuf, ulen)) { 172 DPRINTF(("ebcdic/international %" SIZE_T_FORMAT "u\n", 173 *ulen)); 174 *code = "International EBCDIC"; 175 *code_mime = "ebcdic"; 176 } else { /* Doesn't look like text at all */ 177 DPRINTF(("binary\n")); 178 rv = 0; 179 *type = "binary"; 180 } 181 free(nbuf); 182 } 183 184 done: 185 if (ubuf == &udefbuf) 186 free(udefbuf); 187 188 return rv; 189 } 190 191 /* 192 * This table reflects a particular philosophy about what constitutes 193 * "text," and there is room for disagreement about it. 194 * 195 * Version 3.31 of the file command considered a file to be ASCII if 196 * each of its characters was approved by either the isascii() or 197 * isalpha() function. On most systems, this would mean that any 198 * file consisting only of characters in the range 0x00 ... 0x7F 199 * would be called ASCII text, but many systems might reasonably 200 * consider some characters outside this range to be alphabetic, 201 * so the file command would call such characters ASCII. It might 202 * have been more accurate to call this "considered textual on the 203 * local system" than "ASCII." 204 * 205 * It considered a file to be "International language text" if each 206 * of its characters was either an ASCII printing character (according 207 * to the real ASCII standard, not the above test), a character in 208 * the range 0x80 ... 0xFF, or one of the following control characters: 209 * backspace, tab, line feed, vertical tab, form feed, carriage return, 210 * escape. No attempt was made to determine the language in which files 211 * of this type were written. 212 * 213 * 214 * The table below considers a file to be ASCII if all of its characters 215 * are either ASCII printing characters (again, according to the X3.4 216 * standard, not isascii()) or any of the following controls: bell, 217 * backspace, tab, line feed, form feed, carriage return, esc, nextline. 218 * 219 * I include bell because some programs (particularly shell scripts) 220 * use it literally, even though it is rare in normal text. I exclude 221 * vertical tab because it never seems to be used in real text. I also 222 * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85), 223 * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline 224 * character to. It might be more appropriate to include it in the 8859 225 * set instead of the ASCII set, but it's got to be included in *something* 226 * we recognize or EBCDIC files aren't going to be considered textual. 227 * Some old Unix source files use SO/SI (^N/^O) to shift between Greek 228 * and Latin characters, so these should possibly be allowed. But they 229 * make a real mess on VT100-style displays if they're not paired properly, 230 * so we are probably better off not calling them text. 231 * 232 * A file is considered to be ISO-8859 text if its characters are all 233 * either ASCII, according to the above definition, or printing characters 234 * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF. 235 * 236 * Finally, a file is considered to be international text from some other 237 * character code if its characters are all either ISO-8859 (according to 238 * the above definition) or characters in the range 0x80 ... 0x9F, which 239 * ISO-8859 considers to be control characters but the IBM PC and Macintosh 240 * consider to be printing characters. 241 */ 242 243 #define F 0 /* character never appears in text */ 244 #define T 1 /* character appears in plain ASCII text */ 245 #define I 2 /* character appears in ISO-8859 text */ 246 #define X 3 /* character appears in non-ISO extended ASCII (Mac, IBM PC) */ 247 248 /* 249 * SUB (substitute character ^Z) was used as EOF in DOS and early Windows 250 * NEL (next line 0x85) is considered in ECMAScript as whitespace 251 */ 252 file_private char text_chars[256] = { 253 /* BEL BS HT LF VT FF CR */ 254 F, F, F, F, F, F, F, T, T, T, T, T, T, T, F, F, /* 0x0X */ 255 /* SUB ESC */ 256 F, F, F, F, F, F, F, F, F, F, T, T, F, F, F, F, /* 0x1X */ 257 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x2X */ 258 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x3X */ 259 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x4X */ 260 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x5X */ 261 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x6X */ 262 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F, /* 0x7X */ 263 /* NEL */ 264 X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X, /* 0x8X */ 265 X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 0x9X */ 266 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xaX */ 267 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xbX */ 268 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xcX */ 269 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xdX */ 270 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xeX */ 271 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I /* 0xfX */ 272 }; 273 274 #define LOOKS(NAME, COND) \ 275 file_private int \ 276 looks_ ## NAME(const unsigned char *buf, size_t nbytes, file_unichar_t *ubuf, \ 277 size_t *ulen) \ 278 { \ 279 size_t i; \ 280 \ 281 *ulen = 0; \ 282 \ 283 for (i = 0; i < nbytes; i++) { \ 284 int t = text_chars[buf[i]]; \ 285 \ 286 if (COND) \ 287 return 0; \ 288 \ 289 ubuf[(*ulen)++] = buf[i]; \ 290 } \ 291 return 1; \ 292 } 293 294 LOOKS(ascii, t != T) 295 LOOKS(latin1, t != T && t != I) 296 LOOKS(extended, t != T && t != I && t != X) 297 298 /* 299 * Decide whether some text looks like UTF-8. Returns: 300 * 301 * -1: invalid UTF-8 302 * 0: uses odd control characters, so doesn't look like text 303 * 1: 7-bit text 304 * 2: definitely UTF-8 text (valid high-bit set bytes) 305 * 306 * If ubuf is non-NULL on entry, text is decoded into ubuf, *ulen; 307 * ubuf must be big enough! 308 */ 309 310 // from: https://golang.org/src/unicode/utf8/utf8.go 311 312 #define XX 0xF1 // invalid: size 1 313 #define AS 0xF0 // ASCII: size 1 314 #define S1 0x02 // accept 0, size 2 315 #define S2 0x13 // accept 1, size 3 316 #define S3 0x03 // accept 0, size 3 317 #define S4 0x23 // accept 2, size 3 318 #define S5 0x34 // accept 3, size 4 319 #define S6 0x04 // accept 0, size 4 320 #define S7 0x44 // accept 4, size 4 321 322 #define LOCB 0x80 323 #define HICB 0xBF 324 325 // first is information about the first byte in a UTF-8 sequence. 326 static const uint8_t first[] = { 327 // 1 2 3 4 5 6 7 8 9 A B C D E F 328 AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, // 0x00-0x0F 329 AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, // 0x10-0x1F 330 AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, // 0x20-0x2F 331 AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, // 0x30-0x3F 332 AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, // 0x40-0x4F 333 AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, // 0x50-0x5F 334 AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, // 0x60-0x6F 335 AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, AS, // 0x70-0x7F 336 // 1 2 3 4 5 6 7 8 9 A B C D E F 337 XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, // 0x80-0x8F 338 XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, // 0x90-0x9F 339 XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, // 0xA0-0xAF 340 XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, // 0xB0-0xBF 341 XX, XX, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, // 0xC0-0xCF 342 S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, S1, // 0xD0-0xDF 343 S2, S3, S3, S3, S3, S3, S3, S3, S3, S3, S3, S3, S3, S4, S3, S3, // 0xE0-0xEF 344 S5, S6, S6, S6, S7, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, // 0xF0-0xFF 345 }; 346 347 // acceptRange gives the range of valid values for the second byte in a UTF-8 348 // sequence. 349 static struct accept_range { 350 uint8_t lo; // lowest value for second byte. 351 uint8_t hi; // highest value for second byte. 352 } accept_ranges[16] = { 353 // acceptRanges has size 16 to avoid bounds checks in the code that uses it. 354 { LOCB, HICB }, 355 { 0xA0, HICB }, 356 { LOCB, 0x9F }, 357 { 0x90, HICB }, 358 { LOCB, 0x8F }, 359 }; 360 361 file_protected int 362 file_looks_utf8(const unsigned char *buf, size_t nbytes, file_unichar_t *ubuf, 363 size_t *ulen) 364 { 365 size_t i; 366 int n; 367 file_unichar_t c; 368 int gotone = 0, ctrl = 0; 369 370 if (ubuf) 371 *ulen = 0; 372 373 for (i = 0; i < nbytes; i++) { 374 if ((buf[i] & 0x80) == 0) { /* 0xxxxxxx is plain ASCII */ 375 /* 376 * Even if the whole file is valid UTF-8 sequences, 377 * still reject it if it uses weird control characters. 378 */ 379 380 if (text_chars[buf[i]] != T) 381 ctrl = 1; 382 383 if (ubuf) 384 ubuf[(*ulen)++] = buf[i]; 385 } else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */ 386 return -1; 387 } else { /* 11xxxxxx begins UTF-8 */ 388 int following; 389 uint8_t x = first[buf[i]]; 390 const struct accept_range *ar = 391 &accept_ranges[(unsigned int)x >> 4]; 392 if (x == XX) 393 return -1; 394 395 if ((buf[i] & 0x20) == 0) { /* 110xxxxx */ 396 c = buf[i] & 0x1f; 397 following = 1; 398 } else if ((buf[i] & 0x10) == 0) { /* 1110xxxx */ 399 c = buf[i] & 0x0f; 400 following = 2; 401 } else if ((buf[i] & 0x08) == 0) { /* 11110xxx */ 402 c = buf[i] & 0x07; 403 following = 3; 404 } else if ((buf[i] & 0x04) == 0) { /* 111110xx */ 405 c = buf[i] & 0x03; 406 following = 4; 407 } else if ((buf[i] & 0x02) == 0) { /* 1111110x */ 408 c = buf[i] & 0x01; 409 following = 5; 410 } else 411 return -1; 412 413 for (n = 0; n < following; n++) { 414 i++; 415 if (i >= nbytes) 416 goto done; 417 418 if (n == 0 && 419 (buf[i] < ar->lo || buf[i] > ar->hi)) 420 return -1; 421 422 if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40)) 423 return -1; 424 425 c = (c << 6) + (buf[i] & 0x3f); 426 } 427 428 if (ubuf) 429 ubuf[(*ulen)++] = c; 430 gotone = 1; 431 } 432 } 433 done: 434 return ctrl ? 0 : (gotone ? 2 : 1); 435 } 436 437 /* 438 * Decide whether some text looks like UTF-8 with BOM. If there is no 439 * BOM, return -1; otherwise return the result of looks_utf8 on the 440 * rest of the text. 441 */ 442 file_private int 443 looks_utf8_with_BOM(const unsigned char *buf, size_t nbytes, 444 file_unichar_t *ubuf, size_t *ulen) 445 { 446 if (nbytes > 3 && buf[0] == 0xef && buf[1] == 0xbb && buf[2] == 0xbf) 447 return file_looks_utf8(buf + 3, nbytes - 3, ubuf, ulen); 448 else 449 return -1; 450 } 451 452 file_private int 453 looks_utf7(const unsigned char *buf, size_t nbytes, file_unichar_t *ubuf, 454 size_t *ulen) 455 { 456 if (nbytes > 4 && buf[0] == '+' && buf[1] == '/' && buf[2] == 'v') 457 switch (buf[3]) { 458 case '8': 459 case '9': 460 case '+': 461 case '/': 462 if (ubuf) 463 *ulen = 0; 464 return 1; 465 default: 466 return -1; 467 } 468 else 469 return -1; 470 } 471 472 #define UCS16_NOCHAR(c) ((c) >= 0xfdd0 && (c) <= 0xfdef) 473 #define UCS16_HISURR(c) ((c) >= 0xd800 && (c) <= 0xdbff) 474 #define UCS16_LOSURR(c) ((c) >= 0xdc00 && (c) <= 0xdfff) 475 476 file_private int 477 looks_ucs16(const unsigned char *bf, size_t nbytes, file_unichar_t *ubf, 478 size_t *ulen) 479 { 480 int bigend; 481 uint32_t hi; 482 size_t i; 483 484 if (nbytes < 2) 485 return 0; 486 487 if (bf[0] == 0xff && bf[1] == 0xfe) 488 bigend = 0; 489 else if (bf[0] == 0xfe && bf[1] == 0xff) 490 bigend = 1; 491 else 492 return 0; 493 494 *ulen = 0; 495 hi = 0; 496 497 for (i = 2; i + 1 < nbytes; i += 2) { 498 uint32_t uc; 499 500 if (bigend) 501 uc = CAST(uint32_t, 502 bf[i + 1] | (CAST(file_unichar_t, bf[i]) << 8)); 503 else 504 uc = CAST(uint32_t, 505 bf[i] | (CAST(file_unichar_t, bf[i + 1]) << 8)); 506 507 uc &= 0xffff; 508 509 switch (uc) { 510 case 0xfffe: 511 case 0xffff: 512 return 0; 513 default: 514 if (UCS16_NOCHAR(uc)) 515 return 0; 516 break; 517 } 518 if (hi) { 519 if (!UCS16_LOSURR(uc)) 520 return 0; 521 uc = 0x10000 + 0x400 * (hi - 1) + (uc - 0xdc00); 522 hi = 0; 523 } 524 if (uc < 128 && text_chars[CAST(size_t, uc)] != T) 525 return 0; 526 ubf[(*ulen)++] = uc; 527 if (UCS16_HISURR(uc)) 528 hi = uc - 0xd800 + 1; 529 if (UCS16_LOSURR(uc)) 530 return 0; 531 } 532 533 return 1 + bigend; 534 } 535 536 file_private int 537 looks_ucs32(const unsigned char *bf, size_t nbytes, file_unichar_t *ubf, 538 size_t *ulen) 539 { 540 int bigend; 541 size_t i; 542 543 if (nbytes < 4) 544 return 0; 545 546 if (bf[0] == 0xff && bf[1] == 0xfe && bf[2] == 0 && bf[3] == 0) 547 bigend = 0; 548 else if (bf[0] == 0 && bf[1] == 0 && bf[2] == 0xfe && bf[3] == 0xff) 549 bigend = 1; 550 else 551 return 0; 552 553 *ulen = 0; 554 555 for (i = 4; i + 3 < nbytes; i += 4) { 556 /* XXX fix to properly handle chars > 65536 */ 557 558 if (bigend) 559 ubf[(*ulen)++] = CAST(file_unichar_t, bf[i + 3]) 560 | (CAST(file_unichar_t, bf[i + 2]) << 8) 561 | (CAST(file_unichar_t, bf[i + 1]) << 16) 562 | (CAST(file_unichar_t, bf[i]) << 24); 563 else 564 ubf[(*ulen)++] = CAST(file_unichar_t, bf[i + 0]) 565 | (CAST(file_unichar_t, bf[i + 1]) << 8) 566 | (CAST(file_unichar_t, bf[i + 2]) << 16) 567 | (CAST(file_unichar_t, bf[i + 3]) << 24); 568 569 if (ubf[*ulen - 1] == 0xfffe) 570 return 0; 571 if (ubf[*ulen - 1] < 128 && 572 text_chars[CAST(size_t, ubf[*ulen - 1])] != T) 573 return 0; 574 } 575 576 return 1 + bigend; 577 } 578 #undef F 579 #undef T 580 #undef I 581 #undef X 582 583 /* 584 * This table maps each EBCDIC character to an (8-bit extended) ASCII 585 * character, as specified in the rationale for the dd(1) command in 586 * draft 11.2 (September, 1991) of the POSIX P1003.2 standard. 587 * 588 * Unfortunately it does not seem to correspond exactly to any of the 589 * five variants of EBCDIC documented in IBM's _Enterprise Systems 590 * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh 591 * Edition, July, 1999, pp. I-1 - I-4. 592 * 593 * Fortunately, though, all versions of EBCDIC, including this one, agree 594 * on most of the printing characters that also appear in (7-bit) ASCII. 595 * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all. 596 * 597 * Fortunately too, there is general agreement that codes 0x00 through 598 * 0x3F represent control characters, 0x41 a nonbreaking space, and the 599 * remainder printing characters. 600 * 601 * This is sufficient to allow us to identify EBCDIC text and to distinguish 602 * between old-style and internationalized examples of text. 603 */ 604 605 file_private unsigned char ebcdic_to_ascii[] = { 606 0, 1, 2, 3, 156, 9, 134, 127, 151, 141, 142, 11, 12, 13, 14, 15, 607 16, 17, 18, 19, 157, 133, 8, 135, 24, 25, 146, 143, 28, 29, 30, 31, 608 128, 129, 130, 131, 132, 10, 23, 27, 136, 137, 138, 139, 140, 5, 6, 7, 609 144, 145, 22, 147, 148, 149, 150, 4, 152, 153, 154, 155, 20, 21, 158, 26, 610 ' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|', 611 '&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~', 612 '-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?', 613 186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"', 614 195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201, 615 202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208, 616 209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215, 617 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231, 618 '{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237, 619 '}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243, 620 '\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249, 621 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255 622 }; 623 624 #ifdef notdef 625 /* 626 * The following EBCDIC-to-ASCII table may relate more closely to reality, 627 * or at least to modern reality. It comes from 628 * 629 * http://ftp.s390.ibm.com/products/oe/bpxqp9.html 630 * 631 * and maps the characters of EBCDIC code page 1047 (the code used for 632 * Unix-derived software on IBM's 390 systems) to the corresponding 633 * characters from ISO 8859-1. 634 * 635 * If this table is used instead of the above one, some of the special 636 * cases for the NEL character can be taken out of the code. 637 */ 638 639 file_private unsigned char ebcdic_1047_to_8859[] = { 640 0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F, 641 0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F, 642 0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07, 643 0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A, 644 0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C, 645 0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E, 646 0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F, 647 0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22, 648 0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1, 649 0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4, 650 0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE, 651 0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7, 652 0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5, 653 0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF, 654 0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5, 655 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F 656 }; 657 #endif 658 659 /* 660 * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII. 661 */ 662 file_private void 663 from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out) 664 { 665 size_t i; 666 667 for (i = 0; i < nbytes; i++) { 668 out[i] = ebcdic_to_ascii[buf[i]]; 669 } 670 } 671
Indexes created Sun Sep 13 00:25:22 UTC 2026