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 * file - find type of a file or files - main program. 30 */ 31 32 #include "file.h" 33 34 #ifndef lint 35 FILE_RCSID("@(#)$File: file.c,v 1.175 2018/03/02 16:11:37 christos Exp $") 36 #endif /* lint */ 37 38 #include "magic.h" 39 40 #include <stdlib.h> 41 #include <unistd.h> 42 #include <string.h> 43 #ifdef RESTORE_TIME 44 # if (__COHERENT__ >= 0x420) 45 # include <sys/utime.h> 46 # else 47 # ifdef USE_UTIMES 48 # include <sys/time.h> 49 # else 50 # include <utime.h> 51 # endif 52 # endif 53 #endif 54 #ifdef HAVE_UNISTD_H 55 #include <unistd.h> /* for read() */ 56 #endif 57 #ifdef HAVE_WCHAR_H 58 #include <wchar.h> 59 #endif 60 61 #if defined(HAVE_GETOPT_H) && defined(HAVE_STRUCT_OPTION) 62 # include <getopt.h> 63 # ifndef HAVE_GETOPT_LONG 64 int getopt_long(int, char * const *, const char *, 65 const struct option *, int *); 66 # endif 67 # else 68 # include "mygetopt.h" 69 #endif 70 71 #ifdef S_IFLNK 72 # define IFLNK_h "h" 73 # define IFLNK_L "L" 74 #else 75 # define IFLNK_h "" 76 # define IFLNK_L "" 77 #endif 78 79 #ifdef HAVE_LIBSECCOMP 80 # define SECCOMP_S "S" 81 #else 82 # define SECCOMP_S "" 83 #endif 84 85 #define FILE_FLAGS "bcCdE" IFLNK_h "ik" IFLNK_L "lNnprs" SECCOMP_S "vzZ0" 86 #define OPTSTRING "bcCde:Ef:F:hiklLm:nNpP:rsSvzZ0" 87 88 # define USAGE \ 89 "Usage: %s [-" FILE_FLAGS "] [--apple] [--extension] [--mime-encoding]\n" \ 90 " [--mime-type] [-e <testname>] [-F <separator>] " \ 91 " [-f <namefile>]\n" \ 92 " [-m <magicfiles>] [-P <parameter=value>] <file> ...\n" \ 93 " %s -C [-m <magicfiles>]\n" \ 94 " %s [--help]\n" 95 96 private int /* Global command-line options */ 97 bflag = 0, /* brief output format */ 98 nopad = 0, /* Don't pad output */ 99 nobuffer = 0, /* Do not buffer stdout */ 100 nulsep = 0; /* Append '\0' to the separator */ 101 102 private const char *separator = ":"; /* Default field separator */ 103 private const struct option long_options[] = { 104 #define OPT_HELP 1 105 #define OPT_APPLE 2 106 #define OPT_EXTENSIONS 3 107 #define OPT_MIME_TYPE 4 108 #define OPT_MIME_ENCODING 5 109 #define OPT(shortname, longname, opt, def, doc) \ 110 {longname, opt, NULL, shortname}, 111 #define OPT_LONGONLY(longname, opt, def, doc, id) \ 112 {longname, opt, NULL, id}, 113 #include "file_opts.h" 114 #undef OPT 115 #undef OPT_LONGONLY 116 {0, 0, NULL, 0} 117 }; 118 119 private const struct { 120 const char *name; 121 int value; 122 } nv[] = { 123 { "apptype", MAGIC_NO_CHECK_APPTYPE }, 124 { "ascii", MAGIC_NO_CHECK_ASCII }, 125 { "cdf", MAGIC_NO_CHECK_CDF }, 126 { "compress", MAGIC_NO_CHECK_COMPRESS }, 127 { "elf", MAGIC_NO_CHECK_ELF }, 128 { "encoding", MAGIC_NO_CHECK_ENCODING }, 129 { "soft", MAGIC_NO_CHECK_SOFT }, 130 { "tar", MAGIC_NO_CHECK_TAR }, 131 { "text", MAGIC_NO_CHECK_TEXT }, /* synonym for ascii */ 132 { "tokens", MAGIC_NO_CHECK_TOKENS }, /* OBSOLETE: ignored for backwards compatibility */ 133 }; 134 135 private struct { 136 const char *name; 137 int tag; 138 size_t value; 139 } pm[] = { 140 { "indir", MAGIC_PARAM_INDIR_MAX, 0 }, 141 { "name", MAGIC_PARAM_NAME_MAX, 0 }, 142 { "elf_phnum", MAGIC_PARAM_ELF_PHNUM_MAX, 0 }, 143 { "elf_shnum", MAGIC_PARAM_ELF_SHNUM_MAX, 0 }, 144 { "elf_notes", MAGIC_PARAM_ELF_NOTES_MAX, 0 }, 145 { "regex", MAGIC_PARAM_REGEX_MAX, 0 }, 146 { "bytes", MAGIC_PARAM_BYTES_MAX, 0 }, 147 }; 148 149 private int posixly; 150 151 #ifdef __dead 152 __dead 153 #endif 154 private void usage(void); 155 private void docprint(const char *, int); 156 #ifdef __dead 157 __dead 158 #endif 159 private void help(void); 160 161 private int unwrap(struct magic_set *, const char *); 162 private int process(struct magic_set *ms, const char *, int); 163 private struct magic_set *load(const char *, int); 164 private void setparam(const char *); 165 private void applyparam(magic_t); 166 167 168 /* 169 * main - parse arguments and handle options 170 */ 171 int 172 main(int argc, char *argv[]) 173 { 174 int c; 175 size_t i; 176 int action = 0, didsomefiles = 0, errflg = 0; 177 int flags = 0, e = 0; 178 #ifdef HAVE_LIBSECCOMP 179 int sandbox = 1; 180 #endif 181 struct magic_set *magic = NULL; 182 int longindex; 183 const char *magicfile = NULL; /* where the magic is */ 184 char *progname; 185 186 /* makes islower etc work for other langs */ 187 #ifdef HAVE_SETLOCALE 188 (void)setlocale(LC_CTYPE, ""); 189 #endif 190 191 #ifdef __EMX__ 192 /* sh-like wildcard expansion! Shouldn't hurt at least ... */ 193 _wildcard(&argc, &argv); 194 #endif 195 196 if ((progname = strrchr(argv[0], '/')) != NULL) 197 progname++; 198 else 199 progname = argv[0]; 200 201 file_setprogname(progname); 202 203 204 #ifdef S_IFLNK 205 posixly = getenv("POSIXLY_CORRECT") != NULL; 206 flags |= posixly ? MAGIC_SYMLINK : 0; 207 #endif 208 while ((c = getopt_long(argc, argv, OPTSTRING, long_options, 209 &longindex)) != -1) 210 switch (c) { 211 case OPT_HELP: 212 help(); 213 break; 214 case OPT_APPLE: 215 flags |= MAGIC_APPLE; 216 break; 217 case OPT_EXTENSIONS: 218 flags |= MAGIC_EXTENSION; 219 break; 220 case OPT_MIME_TYPE: 221 flags |= MAGIC_MIME_TYPE; 222 break; 223 case OPT_MIME_ENCODING: 224 flags |= MAGIC_MIME_ENCODING; 225 break; 226 case '0': 227 nulsep++; 228 break; 229 case 'b': 230 bflag++; 231 break; 232 case 'c': 233 action = FILE_CHECK; 234 break; 235 case 'C': 236 action = FILE_COMPILE; 237 break; 238 case 'd': 239 flags |= MAGIC_DEBUG|MAGIC_CHECK; 240 break; 241 case 'E': 242 flags |= MAGIC_ERROR; 243 break; 244 case 'e': 245 for (i = 0; i < sizeof(nv) / sizeof(nv[0]); i++) 246 if (strcmp(nv[i].name, optarg) == 0) 247 break; 248 249 if (i == sizeof(nv) / sizeof(nv[0])) 250 errflg++; 251 else 252 flags |= nv[i].value; 253 break; 254 255 case 'f': 256 if(action) 257 usage(); 258 if (magic == NULL) 259 if ((magic = load(magicfile, flags)) == NULL) 260 return 1; 261 applyparam(magic); 262 e |= unwrap(magic, optarg); 263 ++didsomefiles; 264 break; 265 case 'F': 266 separator = optarg; 267 break; 268 case 'i': 269 flags |= MAGIC_MIME; 270 break; 271 case 'k': 272 flags |= MAGIC_CONTINUE; 273 break; 274 case 'l': 275 action = FILE_LIST; 276 break; 277 case 'm': 278 magicfile = optarg; 279 break; 280 case 'n': 281 ++nobuffer; 282 break; 283 case 'N': 284 ++nopad; 285 break; 286 #if defined(HAVE_UTIME) || defined(HAVE_UTIMES) 287 case 'p': 288 flags |= MAGIC_PRESERVE_ATIME; 289 break; 290 #endif 291 case 'P': 292 setparam(optarg); 293 break; 294 case 'r': 295 flags |= MAGIC_RAW; 296 break; 297 case 's': 298 flags |= MAGIC_DEVICES; 299 break; 300 #ifdef HAVE_LIBSECCOMP 301 case 'S': 302 sandbox = 0; 303 break; 304 #endif 305 case 'v': 306 if (magicfile == NULL) 307 magicfile = magic_getpath(magicfile, action); 308 (void)fprintf(stdout, "%s-%s\n", file_getprogname(), 309 VERSION); 310 (void)fprintf(stdout, "magic file from %s\n", 311 magicfile); 312 return 0; 313 case 'z': 314 flags |= MAGIC_COMPRESS; 315 break; 316 317 case 'Z': 318 flags |= MAGIC_COMPRESS|MAGIC_COMPRESS_TRANSP; 319 break; 320 #ifdef S_IFLNK 321 case 'L': 322 flags |= MAGIC_SYMLINK; 323 break; 324 case 'h': 325 flags &= ~MAGIC_SYMLINK; 326 break; 327 #endif 328 case '?': 329 default: 330 errflg++; 331 break; 332 } 333 334 if (errflg) { 335 usage(); 336 } 337 if (e) 338 return e; 339 340 #ifdef HAVE_LIBSECCOMP 341 #if 0 342 if (sandbox && enable_sandbox_basic() == -1) 343 #else 344 if (sandbox && enable_sandbox_full() == -1) 345 #endif 346 file_err(EXIT_FAILURE, "SECCOMP initialisation failed"); 347 #endif /* HAVE_LIBSECCOMP */ 348 349 if (MAGIC_VERSION != magic_version()) 350 file_warnx("Compiled magic version [%d] " 351 "does not match with shared library magic version [%d]\n", 352 MAGIC_VERSION, magic_version()); 353 354 switch(action) { 355 case FILE_CHECK: 356 case FILE_COMPILE: 357 case FILE_LIST: 358 /* 359 * Don't try to check/compile ~/.magic unless we explicitly 360 * ask for it. 361 */ 362 magic = magic_open(flags|MAGIC_CHECK); 363 if (magic == NULL) { 364 file_warn("Can't create magic"); 365 return 1; 366 } 367 368 369 switch(action) { 370 case FILE_CHECK: 371 c = magic_check(magic, magicfile); 372 break; 373 case FILE_COMPILE: 374 c = magic_compile(magic, magicfile); 375 break; 376 case FILE_LIST: 377 c = magic_list(magic, magicfile); 378 break; 379 default: 380 abort(); 381 } 382 if (c == -1) { 383 file_warnx("%s", magic_error(magic)); 384 e = 1; 385 goto out; 386 } 387 goto out; 388 default: 389 if (magic == NULL) 390 if ((magic = load(magicfile, flags)) == NULL) 391 return 1; 392 applyparam(magic); 393 } 394 395 if (optind == argc) { 396 if (!didsomefiles) 397 usage(); 398 } 399 else { 400 size_t j, wid, nw; 401 for (wid = 0, j = (size_t)optind; j < (size_t)argc; j++) { 402 nw = file_mbswidth(argv[j]); 403 if (nw > wid) 404 wid = nw; 405 } 406 /* 407 * If bflag is only set twice, set it depending on 408 * number of files [this is undocumented, and subject to change] 409 */ 410 if (bflag == 2) { 411 bflag = optind >= argc - 1; 412 } 413 for (; optind < argc; optind++) 414 e |= process(magic, argv[optind], wid); 415 } 416 417 out: 418 if (magic) 419 magic_close(magic); 420 return e; 421 } 422 423 private void 424 applyparam(magic_t magic) 425 { 426 size_t i; 427 428 for (i = 0; i < __arraycount(pm); i++) { 429 if (pm[i].value == 0) 430 continue; 431 if (magic_setparam(magic, pm[i].tag, &pm[i].value) == -1) 432 file_err(EXIT_FAILURE, "Can't set %s", pm[i].name); 433 } 434 } 435 436 private void 437 setparam(const char *p) 438 { 439 size_t i; 440 char *s; 441 442 if ((s = strchr(p, '=')) == NULL) 443 goto badparm; 444 445 for (i = 0; i < __arraycount(pm); i++) { 446 if (strncmp(p, pm[i].name, s - p) != 0) 447 continue; 448 pm[i].value = atoi(s + 1); 449 return; 450 } 451 badparm: 452 file_errx(EXIT_FAILURE, "Unknown param %s", p); 453 } 454 455 private struct magic_set * 456 /*ARGSUSED*/ 457 load(const char *magicfile, int flags) 458 { 459 struct magic_set *magic = magic_open(flags); 460 const char *e; 461 462 if (magic == NULL) { 463 file_warn("Can't create magic"); 464 return NULL; 465 } 466 if (magic_load(magic, magicfile) == -1) { 467 file_warn("%s", magic_error(magic)); 468 magic_close(magic); 469 return NULL; 470 } 471 if ((e = magic_error(magic)) != NULL) 472 file_warn("%s", e); 473 return magic; 474 } 475 476 /* 477 * unwrap -- read a file of filenames, do each one. 478 */ 479 private int 480 unwrap(struct magic_set *ms, const char *fn) 481 { 482 FILE *f; 483 ssize_t len; 484 char *line = NULL; 485 size_t llen = 0; 486 int wid = 0, cwid; 487 int e = 0; 488 489 if (strcmp("-", fn) == 0) { 490 f = stdin; 491 wid = 1; 492 } else { 493 if ((f = fopen(fn, "r")) == NULL) { 494 file_warn("Cannot open `%s'", fn); 495 return 1; 496 } 497 498 while ((len = getline(&line, &llen, f)) > 0) { 499 if (line[len - 1] == '\n') 500 line[len - 1] = '\0'; 501 cwid = file_mbswidth(line); 502 if (cwid > wid) 503 wid = cwid; 504 } 505 506 rewind(f); 507 } 508 509 while ((len = getline(&line, &llen, f)) > 0) { 510 if (line[len - 1] == '\n') 511 line[len - 1] = '\0'; 512 e |= process(ms, line, wid); 513 if(nobuffer) 514 (void)fflush(stdout); 515 } 516 517 free(line); 518 (void)fclose(f); 519 return e; 520 } 521 522 /* 523 * Called for each input file on the command line (or in a list of files) 524 */ 525 private int 526 process(struct magic_set *ms, const char *inname, int wid) 527 { 528 const char *type, c = nulsep > 1 ? '\0' : '\n'; 529 int std_in = strcmp(inname, "-") == 0; 530 531 if (wid > 0 && !bflag) { 532 (void)printf("%s", std_in ? "/dev/stdin" : inname); 533 if (nulsep) 534 (void)putc('\0', stdout); 535 if (nulsep < 2) { 536 (void)printf("%s", separator); 537 (void)printf("%*s ", 538 (int) (nopad ? 0 : (wid - file_mbswidth(inname))), 539 ""); 540 } 541 } 542 543 type = magic_file(ms, std_in ? NULL : inname); 544 545 if (type == NULL) { 546 (void)printf("ERROR: %s%c", magic_error(ms), c); 547 return 1; 548 } else { 549 (void)printf("%s%c", type, c); 550 return 0; 551 } 552 } 553 554 protected size_t 555 file_mbswidth(const char *s) 556 { 557 #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH) 558 size_t bytesconsumed, old_n, n, width = 0; 559 mbstate_t state; 560 wchar_t nextchar; 561 (void)memset(&state, 0, sizeof(mbstate_t)); 562 old_n = n = strlen(s); 563 564 while (n > 0) { 565 bytesconsumed = mbrtowc(&nextchar, s, n, &state); 566 if (bytesconsumed == (size_t)(-1) || 567 bytesconsumed == (size_t)(-2)) { 568 /* Something went wrong, return something reasonable */ 569 return old_n; 570 } 571 if (s[0] == '\n') { 572 /* 573 * do what strlen() would do, so that caller 574 * is always right 575 */ 576 width++; 577 } else { 578 int w = wcwidth(nextchar); 579 if (w > 0) 580 width += w; 581 } 582 583 s += bytesconsumed, n -= bytesconsumed; 584 } 585 return width; 586 #else 587 return strlen(s); 588 #endif 589 } 590 591 private void 592 usage(void) 593 { 594 const char *pn = file_getprogname(); 595 (void)fprintf(stderr, USAGE, pn, pn, pn); 596 exit(EXIT_FAILURE); 597 } 598 599 private void 600 defprint(int def) 601 { 602 if (!def) 603 return; 604 if (((def & 1) && posixly) || ((def & 2) && !posixly)) 605 fprintf(stdout, " (default)"); 606 fputc('\n', stdout); 607 } 608 609 private void 610 docprint(const char *opts, int def) 611 { 612 size_t i; 613 int comma; 614 char *sp, *p; 615 616 p = strstr(opts, "%o"); 617 if (p == NULL) { 618 fprintf(stdout, "%s", opts); 619 defprint(def); 620 return; 621 } 622 623 for (sp = p - 1; sp > opts && *sp == ' '; sp--) 624 continue; 625 626 fprintf(stdout, "%.*s", (int)(p - opts), opts); 627 628 comma = 0; 629 for (i = 0; i < __arraycount(nv); i++) { 630 fprintf(stdout, "%s%s", comma++ ? ", " : "", nv[i].name); 631 if (i && i % 5 == 0) { 632 fprintf(stdout, ",\n%*s", (int)(p - sp - 1), ""); 633 comma = 0; 634 } 635 } 636 637 fprintf(stdout, "%s", opts + (p - opts) + 2); 638 } 639 640 private void 641 help(void) 642 { 643 (void)fputs( 644 "Usage: file [OPTION...] [FILE...]\n" 645 "Determine type of FILEs.\n" 646 "\n", stdout); 647 #define OPT(shortname, longname, opt, def, doc) \ 648 fprintf(stdout, " -%c, --" longname, shortname), \ 649 docprint(doc, def); 650 #define OPT_LONGONLY(longname, opt, def, doc, id) \ 651 fprintf(stdout, " --" longname), \ 652 docprint(doc, def); 653 #include "file_opts.h" 654 #undef OPT 655 #undef OPT_LONGONLY 656 fprintf(stdout, "\nReport bugs to http://bugs.gw.com/\n"); 657 exit(EXIT_SUCCESS); 658 } 659 660 private const char *file_progname; 661 662 protected void 663 file_setprogname(const char *progname) 664 { 665 file_progname = progname; 666 } 667 668 protected const char * 669 file_getprogname(void) 670 { 671 return file_progname; 672 } 673 674 protected void 675 file_err(int e, const char *fmt, ...) 676 { 677 va_list ap; 678 int se = errno; 679 680 va_start(ap, fmt); 681 fprintf(stderr, "%s: ", file_progname); 682 vfprintf(stderr, fmt, ap); 683 va_end(ap); 684 fprintf(stderr, " (%s)\n", strerror(se)); 685 exit(e); 686 } 687 688 protected void 689 file_errx(int e, const char *fmt, ...) 690 { 691 va_list ap; 692 693 va_start(ap, fmt); 694 fprintf(stderr, "%s: ", file_progname); 695 vfprintf(stderr, fmt, ap); 696 va_end(ap); 697 fprintf(stderr, "\n"); 698 exit(e); 699 } 700 701 protected void 702 file_warn(const char *fmt, ...) 703 { 704 va_list ap; 705 int se = errno; 706 707 va_start(ap, fmt); 708 fprintf(stderr, "%s: ", file_progname); 709 vfprintf(stderr, fmt, ap); 710 va_end(ap); 711 fprintf(stderr, " (%s)\n", strerror(se)); 712 errno = se; 713 } 714 715 protected void 716 file_warnx(const char *fmt, ...) 717 { 718 va_list ap; 719 int se = errno; 720 721 va_start(ap, fmt); 722 fprintf(stderr, "%s: ", file_progname); 723 vfprintf(stderr, fmt, ap); 724 va_end(ap); 725 fprintf(stderr, "\n"); 726 errno = se; 727 } 728