1 /* 2 * Copyright (c) Ian F. Darwin 1986-1995. 3 * Software written by Ian F. Darwin and others; 4 * maintained 1995-present by Christos Zoulas and others. 5 * 6 * Redistribution and use in source and binary forms, with or without 7 * modification, are permitted provided that the following conditions 8 * are met: 9 * 1. Redistributions of source code must retain the above copyright 10 * notice immediately at the beginning of the file, without modification, 11 * this list of conditions, and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR 20 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 26 * SUCH DAMAGE. 27 */ 28 /* 29 * Encoding -- determine the character encoding of a text file. 30 * 31 * Joerg Wunsch <joerg@freebsd.org> wrote the original support for 8-bit 32 * international characters. 33 */ 34 35 #include "file.h" 36 37 #ifndef lint 38 FILE_RCSID("@(#)$File: encoding.c,v 1.3 2009/02/03 20:27:51 christos Exp $") 39 #endif /* lint */ 40 41 #include "magic.h" 42 #include <string.h> 43 #include <memory.h> 44 #include <stdlib.h> 45 46 47 private int looks_ascii(const unsigned char *, size_t, unichar *, size_t *); 48 private int looks_utf8_with_BOM(const unsigned char *, size_t, unichar *, 49 size_t *); 50 private int looks_ucs16(const unsigned char *, size_t, unichar *, size_t *); 51 private int looks_latin1(const unsigned char *, size_t, unichar *, size_t *); 52 private int looks_extended(const unsigned char *, size_t, unichar *, size_t *); 53 private void from_ebcdic(const unsigned char *, size_t, unsigned char *); 54 55 /* 56 * Try to determine whether text is in some character code we can 57 * identify. Each of these tests, if it succeeds, will leave 58 * the text converted into one-unichar-per-character Unicode in 59 * ubuf, and the number of characters converted in ulen. 60 */ 61 protected int 62 file_encoding(struct magic_set *ms, const unsigned char *buf, size_t nbytes, unichar **ubuf, size_t *ulen, const char **code, const char **code_mime, const char **type) 63 { 64 size_t mlen; 65 int rv = 1, ucs_type; 66 unsigned char *nbuf = NULL; 67 68 mlen = (nbytes + 1) * sizeof(nbuf[0]); 69 if ((nbuf = CAST(unsigned char *, calloc((size_t)1, mlen))) == NULL) { 70 file_oomem(ms, mlen); 71 goto done; 72 } 73 mlen = (nbytes + 1) * sizeof((*ubuf)[0]); 74 if ((*ubuf = CAST(unichar *, calloc((size_t)1, mlen))) == NULL) { 75 file_oomem(ms, mlen); 76 goto done; 77 } 78 79 *type = "text"; 80 if (looks_ascii(buf, nbytes, *ubuf, ulen)) { 81 *code = "ASCII"; 82 *code_mime = "us-ascii"; 83 } else if (looks_utf8_with_BOM(buf, nbytes, *ubuf, ulen) > 0) { 84 *code = "UTF-8 Unicode (with BOM)"; 85 *code_mime = "utf-8"; 86 } else if (file_looks_utf8(buf, nbytes, *ubuf, ulen) > 1) { 87 *code = "UTF-8 Unicode"; 88 *code_mime = "utf-8"; 89 } else if ((ucs_type = looks_ucs16(buf, nbytes, *ubuf, ulen)) != 0) { 90 if (ucs_type == 1) { 91 *code = "Little-endian UTF-16 Unicode"; 92 *code_mime = "utf-16le"; 93 } else { 94 *code = "Big-endian UTF-16 Unicode"; 95 *code_mime = "utf-16be"; 96 } 97 } else if (looks_latin1(buf, nbytes, *ubuf, ulen)) { 98 *code = "ISO-8859"; 99 *code_mime = "iso-8859-1"; 100 } else if (looks_extended(buf, nbytes, *ubuf, ulen)) { 101 *code = "Non-ISO extended-ASCII"; 102 *code_mime = "unknown-8bit"; 103 } else { 104 from_ebcdic(buf, nbytes, nbuf); 105 106 if (looks_ascii(nbuf, nbytes, *ubuf, ulen)) { 107 *code = "EBCDIC"; 108 *code_mime = "ebcdic"; 109 } else if (looks_latin1(nbuf, nbytes, *ubuf, ulen)) { 110 *code = "International EBCDIC"; 111 *code_mime = "ebcdic"; 112 } else { /* Doesn't look like text at all */ 113 rv = 0; 114 *type = "binary"; 115 } 116 } 117 118 done: 119 if (nbuf) 120 free(nbuf); 121 122 return rv; 123 } 124 125 /* 126 * This table reflects a particular philosophy about what constitutes 127 * "text," and there is room for disagreement about it. 128 * 129 * Version 3.31 of the file command considered a file to be ASCII if 130 * each of its characters was approved by either the isascii() or 131 * isalpha() function. On most systems, this would mean that any 132 * file consisting only of characters in the range 0x00 ... 0x7F 133 * would be called ASCII text, but many systems might reasonably 134 * consider some characters outside this range to be alphabetic, 135 * so the file command would call such characters ASCII. It might 136 * have been more accurate to call this "considered textual on the 137 * local system" than "ASCII." 138 * 139 * It considered a file to be "International language text" if each 140 * of its characters was either an ASCII printing character (according 141 * to the real ASCII standard, not the above test), a character in 142 * the range 0x80 ... 0xFF, or one of the following control characters: 143 * backspace, tab, line feed, vertical tab, form feed, carriage return, 144 * escape. No attempt was made to determine the language in which files 145 * of this type were written. 146 * 147 * 148 * The table below considers a file to be ASCII if all of its characters 149 * are either ASCII printing characters (again, according to the X3.4 150 * standard, not isascii()) or any of the following controls: bell, 151 * backspace, tab, line feed, form feed, carriage return, esc, nextline. 152 * 153 * I include bell because some programs (particularly shell scripts) 154 * use it literally, even though it is rare in normal text. I exclude 155 * vertical tab because it never seems to be used in real text. I also 156 * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85), 157 * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline 158 * character to. It might be more appropriate to include it in the 8859 159 * set instead of the ASCII set, but it's got to be included in *something* 160 * we recognize or EBCDIC files aren't going to be considered textual. 161 * Some old Unix source files use SO/SI (^N/^O) to shift between Greek 162 * and Latin characters, so these should possibly be allowed. But they 163 * make a real mess on VT100-style displays if they're not paired properly, 164 * so we are probably better off not calling them text. 165 * 166 * A file is considered to be ISO-8859 text if its characters are all 167 * either ASCII, according to the above definition, or printing characters 168 * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF. 169 * 170 * Finally, a file is considered to be international text from some other 171 * character code if its characters are all either ISO-8859 (according to 172 * the above definition) or characters in the range 0x80 ... 0x9F, which 173 * ISO-8859 considers to be control characters but the IBM PC and Macintosh 174 * consider to be printing characters. 175 */ 176 177 #define F 0 /* character never appears in text */ 178 #define T 1 /* character appears in plain ASCII text */ 179 #define I 2 /* character appears in ISO-8859 text */ 180 #define X 3 /* character appears in non-ISO extended ASCII (Mac, IBM PC) */ 181 182 private char text_chars[256] = { 183 /* BEL BS HT LF FF CR */ 184 F, F, F, F, F, F, F, T, T, T, T, F, T, T, F, F, /* 0x0X */ 185 /* ESC */ 186 F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F, /* 0x1X */ 187 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x2X */ 188 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x3X */ 189 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x4X */ 190 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x5X */ 191 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x6X */ 192 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F, /* 0x7X */ 193 /* NEL */ 194 X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X, /* 0x8X */ 195 X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 0x9X */ 196 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xaX */ 197 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xbX */ 198 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xcX */ 199 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xdX */ 200 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xeX */ 201 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I /* 0xfX */ 202 }; 203 204 private int 205 looks_ascii(const unsigned char *buf, size_t nbytes, unichar *ubuf, 206 size_t *ulen) 207 { 208 size_t i; 209 210 *ulen = 0; 211 212 for (i = 0; i < nbytes; i++) { 213 int t = text_chars[buf[i]]; 214 215 if (t != T) 216 return 0; 217 218 ubuf[(*ulen)++] = buf[i]; 219 } 220 221 return 1; 222 } 223 224 private int 225 looks_latin1(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen) 226 { 227 size_t i; 228 229 *ulen = 0; 230 231 for (i = 0; i < nbytes; i++) { 232 int t = text_chars[buf[i]]; 233 234 if (t != T && t != I) 235 return 0; 236 237 ubuf[(*ulen)++] = buf[i]; 238 } 239 240 return 1; 241 } 242 243 private int 244 looks_extended(const unsigned char *buf, size_t nbytes, unichar *ubuf, 245 size_t *ulen) 246 { 247 size_t i; 248 249 *ulen = 0; 250 251 for (i = 0; i < nbytes; i++) { 252 int t = text_chars[buf[i]]; 253 254 if (t != T && t != I && t != X) 255 return 0; 256 257 ubuf[(*ulen)++] = buf[i]; 258 } 259 260 return 1; 261 } 262 263 /* 264 * Decide whether some text looks like UTF-8. Returns: 265 * 266 * -1: invalid UTF-8 267 * 0: uses odd control characters, so doesn't look like text 268 * 1: 7-bit text 269 * 2: definitely UTF-8 text (valid high-bit set bytes) 270 * 271 * If ubuf is non-NULL on entry, text is decoded into ubuf, *ulen; 272 * ubuf must be big enough! 273 */ 274 protected int 275 file_looks_utf8(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen) 276 { 277 size_t i; 278 int n; 279 unichar c; 280 int gotone = 0, ctrl = 0; 281 282 if (ubuf) 283 *ulen = 0; 284 285 for (i = 0; i < nbytes; i++) { 286 if ((buf[i] & 0x80) == 0) { /* 0xxxxxxx is plain ASCII */ 287 /* 288 * Even if the whole file is valid UTF-8 sequences, 289 * still reject it if it uses weird control characters. 290 */ 291 292 if (text_chars[buf[i]] != T) 293 ctrl = 1; 294 295 if (ubuf) 296 ubuf[(*ulen)++] = buf[i]; 297 } else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */ 298 return -1; 299 } else { /* 11xxxxxx begins UTF-8 */ 300 int following; 301 302 if ((buf[i] & 0x20) == 0) { /* 110xxxxx */ 303 c = buf[i] & 0x1f; 304 following = 1; 305 } else if ((buf[i] & 0x10) == 0) { /* 1110xxxx */ 306 c = buf[i] & 0x0f; 307 following = 2; 308 } else if ((buf[i] & 0x08) == 0) { /* 11110xxx */ 309 c = buf[i] & 0x07; 310 following = 3; 311 } else if ((buf[i] & 0x04) == 0) { /* 111110xx */ 312 c = buf[i] & 0x03; 313 following = 4; 314 } else if ((buf[i] & 0x02) == 0) { /* 1111110x */ 315 c = buf[i] & 0x01; 316 following = 5; 317 } else 318 return -1; 319 320 for (n = 0; n < following; n++) { 321 i++; 322 if (i >= nbytes) 323 goto done; 324 325 if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40)) 326 return -1; 327 328 c = (c << 6) + (buf[i] & 0x3f); 329 } 330 331 if (ubuf) 332 ubuf[(*ulen)++] = c; 333 gotone = 1; 334 } 335 } 336 done: 337 return ctrl ? 0 : (gotone ? 2 : 1); 338 } 339 340 /* 341 * Decide whether some text looks like UTF-8 with BOM. If there is no 342 * BOM, return -1; otherwise return the result of looks_utf8 on the 343 * rest of the text. 344 */ 345 private int 346 looks_utf8_with_BOM(const unsigned char *buf, size_t nbytes, unichar *ubuf, 347 size_t *ulen) 348 { 349 if (nbytes > 3 && buf[0] == 0xef && buf[1] == 0xbb && buf[2] == 0xbf) 350 return file_looks_utf8(buf + 3, nbytes - 3, ubuf, ulen); 351 else 352 return -1; 353 } 354 355 private int 356 looks_ucs16(const unsigned char *buf, size_t nbytes, unichar *ubuf, 357 size_t *ulen) 358 { 359 int bigend; 360 size_t i; 361 362 if (nbytes < 2) 363 return 0; 364 365 if (buf[0] == 0xff && buf[1] == 0xfe) 366 bigend = 0; 367 else if (buf[0] == 0xfe && buf[1] == 0xff) 368 bigend = 1; 369 else 370 return 0; 371 372 *ulen = 0; 373 374 for (i = 2; i + 1 < nbytes; i += 2) { 375 /* XXX fix to properly handle chars > 65536 */ 376 377 if (bigend) 378 ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i]; 379 else 380 ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1]; 381 382 if (ubuf[*ulen - 1] == 0xfffe) 383 return 0; 384 if (ubuf[*ulen - 1] < 128 && 385 text_chars[(size_t)ubuf[*ulen - 1]] != T) 386 return 0; 387 } 388 389 return 1 + bigend; 390 } 391 392 #undef F 393 #undef T 394 #undef I 395 #undef X 396 397 /* 398 * This table maps each EBCDIC character to an (8-bit extended) ASCII 399 * character, as specified in the rationale for the dd(1) command in 400 * draft 11.2 (September, 1991) of the POSIX P1003.2 standard. 401 * 402 * Unfortunately it does not seem to correspond exactly to any of the 403 * five variants of EBCDIC documented in IBM's _Enterprise Systems 404 * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh 405 * Edition, July, 1999, pp. I-1 - I-4. 406 * 407 * Fortunately, though, all versions of EBCDIC, including this one, agree 408 * on most of the printing characters that also appear in (7-bit) ASCII. 409 * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all. 410 * 411 * Fortunately too, there is general agreement that codes 0x00 through 412 * 0x3F represent control characters, 0x41 a nonbreaking space, and the 413 * remainder printing characters. 414 * 415 * This is sufficient to allow us to identify EBCDIC text and to distinguish 416 * between old-style and internationalized examples of text. 417 */ 418 419 private unsigned char ebcdic_to_ascii[] = { 420 0, 1, 2, 3, 156, 9, 134, 127, 151, 141, 142, 11, 12, 13, 14, 15, 421 16, 17, 18, 19, 157, 133, 8, 135, 24, 25, 146, 143, 28, 29, 30, 31, 422 128, 129, 130, 131, 132, 10, 23, 27, 136, 137, 138, 139, 140, 5, 6, 7, 423 144, 145, 22, 147, 148, 149, 150, 4, 152, 153, 154, 155, 20, 21, 158, 26, 424 ' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|', 425 '&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~', 426 '-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?', 427 186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"', 428 195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201, 429 202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208, 430 209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215, 431 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231, 432 '{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237, 433 '}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243, 434 '\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249, 435 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255 436 }; 437 438 #ifdef notdef 439 /* 440 * The following EBCDIC-to-ASCII table may relate more closely to reality, 441 * or at least to modern reality. It comes from 442 * 443 * http://ftp.s390.ibm.com/products/oe/bpxqp9.html 444 * 445 * and maps the characters of EBCDIC code page 1047 (the code used for 446 * Unix-derived software on IBM's 390 systems) to the corresponding 447 * characters from ISO 8859-1. 448 * 449 * If this table is used instead of the above one, some of the special 450 * cases for the NEL character can be taken out of the code. 451 */ 452 453 private unsigned char ebcdic_1047_to_8859[] = { 454 0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F, 455 0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F, 456 0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07, 457 0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A, 458 0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C, 459 0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E, 460 0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F, 461 0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22, 462 0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1, 463 0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4, 464 0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE, 465 0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7, 466 0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5, 467 0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF, 468 0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5, 469 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F 470 }; 471 #endif 472 473 /* 474 * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII. 475 */ 476 private void 477 from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out) 478 { 479 size_t i; 480 481 for (i = 0; i < nbytes; i++) { 482 out[i] = ebcdic_to_ascii[buf[i]]; 483 } 484 } 485