grep.c revision 220422
1/* $NetBSD: grep.c,v 1.4 2011/02/16 01:31:33 joerg Exp $ */ 2/* $FreeBSD: head/usr.bin/grep/grep.c 220422 2011-04-07 13:03:35Z gabor $ */ 3/* $OpenBSD: grep.c,v 1.42 2010/07/02 22:18:03 tedu Exp $ */ 4 5/*- 6 * Copyright (c) 1999 James Howard and Dag-Erling Co��dan Sm��rgrav 7 * Copyright (C) 2008-2009 Gabor Kovesdan <gabor@FreeBSD.org> 8 * All rights reserved. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32#include <sys/cdefs.h> 33__FBSDID("$FreeBSD: head/usr.bin/grep/grep.c 220422 2011-04-07 13:03:35Z gabor $"); 34 35#include <sys/stat.h> 36#include <sys/types.h> 37 38#include <ctype.h> 39#include <err.h> 40#include <errno.h> 41#include <getopt.h> 42#include <limits.h> 43#include <libgen.h> 44#include <locale.h> 45#include <stdbool.h> 46#include <stdio.h> 47#include <stdlib.h> 48#include <string.h> 49#include <unistd.h> 50 51#include "grep.h" 52 53#ifndef WITHOUT_NLS 54#include <nl_types.h> 55nl_catd catalog; 56#endif 57 58/* 59 * Default messags to use when NLS is disabled or no catalogue 60 * is found. 61 */ 62const char *errstr[] = { 63 "", 64/* 1*/ "(standard input)", 65/* 2*/ "cannot read bzip2 compressed file", 66/* 3*/ "unknown %s option", 67/* 4*/ "usage: %s [-abcDEFGHhIiJLlmnOoPqRSsUVvwxZ] [-A num] [-B num] [-C[num]]\n", 68/* 5*/ "\t[-e pattern] [-f file] [--binary-files=value] [--color=when]\n", 69/* 6*/ "\t[--context[=num]] [--directories=action] [--label] [--line-buffered]\n", 70/* 7*/ "\t[--null] [pattern] [file ...]\n", 71/* 8*/ "Binary file %s matches\n", 72/* 9*/ "%s (BSD grep) %s\n", 73}; 74 75/* Flags passed to regcomp() and regexec() */ 76int cflags = 0; 77int eflags = REG_STARTEND; 78 79/* Shortcut for matching all cases like empty regex */ 80bool matchall; 81 82/* Searching patterns */ 83unsigned int patterns, pattern_sz; 84char **pattern; 85regex_t *r_pattern; 86fastgrep_t *fg_pattern; 87 88/* Filename exclusion/inclusion patterns */ 89unsigned int fpatterns, fpattern_sz; 90unsigned int dpatterns, dpattern_sz; 91struct epat *dpattern, *fpattern; 92 93/* For regex errors */ 94char re_error[RE_ERROR_BUF + 1]; 95 96/* Command-line flags */ 97unsigned long long Aflag; /* -A x: print x lines trailing each match */ 98unsigned long long Bflag; /* -B x: print x lines leading each match */ 99bool Hflag; /* -H: always print file name */ 100bool Lflag; /* -L: only show names of files with no matches */ 101bool bflag; /* -b: show block numbers for each match */ 102bool cflag; /* -c: only show a count of matching lines */ 103bool hflag; /* -h: don't print filename headers */ 104bool iflag; /* -i: ignore case */ 105bool lflag; /* -l: only show names of files with matches */ 106bool mflag; /* -m x: stop reading the files after x matches */ 107unsigned long long mcount; /* count for -m */ 108bool nflag; /* -n: show line numbers in front of matching lines */ 109bool oflag; /* -o: print only matching part */ 110bool qflag; /* -q: quiet mode (don't output anything) */ 111bool sflag; /* -s: silent mode (ignore errors) */ 112bool vflag; /* -v: only show non-matching lines */ 113bool wflag; /* -w: pattern must start and end on word boundaries */ 114bool xflag; /* -x: pattern must match entire line */ 115bool lbflag; /* --line-buffered */ 116bool nullflag; /* --null */ 117char *label; /* --label */ 118const char *color; /* --color */ 119int grepbehave = GREP_BASIC; /* -EFGP: type of the regex */ 120int binbehave = BINFILE_BIN; /* -aIU: handling of binary files */ 121int filebehave = FILE_STDIO; /* -JZ: normal, gzip or bzip2 file */ 122int devbehave = DEV_READ; /* -D: handling of devices */ 123int dirbehave = DIR_READ; /* -dRr: handling of directories */ 124int linkbehave = LINK_READ; /* -OpS: handling of symlinks */ 125 126bool dexclude, dinclude; /* --exclude-dir and --include-dir */ 127bool fexclude, finclude; /* --exclude and --include */ 128 129enum { 130 BIN_OPT = CHAR_MAX + 1, 131 COLOR_OPT, 132 HELP_OPT, 133 MMAP_OPT, 134 LINEBUF_OPT, 135 LABEL_OPT, 136 NULL_OPT, 137 R_EXCLUDE_OPT, 138 R_INCLUDE_OPT, 139 R_DEXCLUDE_OPT, 140 R_DINCLUDE_OPT 141}; 142 143static inline const char *init_color(const char *); 144 145/* Housekeeping */ 146bool first = true; /* flag whether we are processing the first match */ 147bool prev; /* flag whether or not the previous line matched */ 148int tail; /* lines left to print */ 149bool notfound; /* file not found */ 150 151extern char *__progname; 152 153/* 154 * Prints usage information and returns 2. 155 */ 156static void 157usage(void) 158{ 159 fprintf(stderr, getstr(4), __progname); 160 fprintf(stderr, "%s", getstr(5)); 161 fprintf(stderr, "%s", getstr(5)); 162 fprintf(stderr, "%s", getstr(6)); 163 fprintf(stderr, "%s", getstr(7)); 164 exit(2); 165} 166 167static const char *optstr = "0123456789A:B:C:D:EFGHIJLOPSRUVZabcd:e:f:hilm:nopqrsuvwxy"; 168 169struct option long_options[] = 170{ 171 {"binary-files", required_argument, NULL, BIN_OPT}, 172 {"help", no_argument, NULL, HELP_OPT}, 173 {"mmap", no_argument, NULL, MMAP_OPT}, 174 {"line-buffered", no_argument, NULL, LINEBUF_OPT}, 175 {"label", required_argument, NULL, LABEL_OPT}, 176 {"null", no_argument, NULL, NULL_OPT}, 177 {"color", optional_argument, NULL, COLOR_OPT}, 178 {"colour", optional_argument, NULL, COLOR_OPT}, 179 {"exclude", required_argument, NULL, R_EXCLUDE_OPT}, 180 {"include", required_argument, NULL, R_INCLUDE_OPT}, 181 {"exclude-dir", required_argument, NULL, R_DEXCLUDE_OPT}, 182 {"include-dir", required_argument, NULL, R_DINCLUDE_OPT}, 183 {"after-context", required_argument, NULL, 'A'}, 184 {"text", no_argument, NULL, 'a'}, 185 {"before-context", required_argument, NULL, 'B'}, 186 {"byte-offset", no_argument, NULL, 'b'}, 187 {"context", optional_argument, NULL, 'C'}, 188 {"count", no_argument, NULL, 'c'}, 189 {"devices", required_argument, NULL, 'D'}, 190 {"directories", required_argument, NULL, 'd'}, 191 {"extended-regexp", no_argument, NULL, 'E'}, 192 {"regexp", required_argument, NULL, 'e'}, 193 {"fixed-strings", no_argument, NULL, 'F'}, 194 {"file", required_argument, NULL, 'f'}, 195 {"basic-regexp", no_argument, NULL, 'G'}, 196 {"no-filename", no_argument, NULL, 'h'}, 197 {"with-filename", no_argument, NULL, 'H'}, 198 {"ignore-case", no_argument, NULL, 'i'}, 199 {"bz2decompress", no_argument, NULL, 'J'}, 200 {"files-with-matches", no_argument, NULL, 'l'}, 201 {"files-without-match", no_argument, NULL, 'L'}, 202 {"max-count", required_argument, NULL, 'm'}, 203 {"line-number", no_argument, NULL, 'n'}, 204 {"only-matching", no_argument, NULL, 'o'}, 205 {"quiet", no_argument, NULL, 'q'}, 206 {"silent", no_argument, NULL, 'q'}, 207 {"recursive", no_argument, NULL, 'r'}, 208 {"no-messages", no_argument, NULL, 's'}, 209 {"binary", no_argument, NULL, 'U'}, 210 {"unix-byte-offsets", no_argument, NULL, 'u'}, 211 {"invert-match", no_argument, NULL, 'v'}, 212 {"version", no_argument, NULL, 'V'}, 213 {"word-regexp", no_argument, NULL, 'w'}, 214 {"line-regexp", no_argument, NULL, 'x'}, 215 {"decompress", no_argument, NULL, 'Z'}, 216 {NULL, no_argument, NULL, 0} 217}; 218 219/* 220 * Adds a searching pattern to the internal array. 221 */ 222static void 223add_pattern(char *pat, size_t len) 224{ 225 226 /* Check if we can do a shortcut */ 227 if (len == 0 || matchall) { 228 matchall = true; 229 return; 230 } 231 /* Increase size if necessary */ 232 if (patterns == pattern_sz) { 233 pattern_sz *= 2; 234 pattern = grep_realloc(pattern, ++pattern_sz * 235 sizeof(*pattern)); 236 } 237 if (len > 0 && pat[len - 1] == '\n') 238 --len; 239 /* pat may not be NUL-terminated */ 240 pattern[patterns] = grep_malloc(len + 1); 241 memcpy(pattern[patterns], pat, len); 242 pattern[patterns][len] = '\0'; 243 ++patterns; 244} 245 246/* 247 * Adds a file include/exclude pattern to the internal array. 248 */ 249static void 250add_fpattern(const char *pat, int mode) 251{ 252 253 /* Increase size if necessary */ 254 if (fpatterns == fpattern_sz) { 255 fpattern_sz *= 2; 256 fpattern = grep_realloc(fpattern, ++fpattern_sz * 257 sizeof(struct epat)); 258 } 259 fpattern[fpatterns].pat = grep_strdup(pat); 260 fpattern[fpatterns].mode = mode; 261 ++fpatterns; 262} 263 264/* 265 * Adds a directory include/exclude pattern to the internal array. 266 */ 267static void 268add_dpattern(const char *pat, int mode) 269{ 270 271 /* Increase size if necessary */ 272 if (dpatterns == dpattern_sz) { 273 dpattern_sz *= 2; 274 dpattern = grep_realloc(dpattern, ++dpattern_sz * 275 sizeof(struct epat)); 276 } 277 dpattern[dpatterns].pat = grep_strdup(pat); 278 dpattern[dpatterns].mode = mode; 279 ++dpatterns; 280} 281 282/* 283 * Reads searching patterns from a file and adds them with add_pattern(). 284 */ 285static void 286read_patterns(const char *fn) 287{ 288 FILE *f; 289 char *line; 290 size_t len; 291 292 if ((f = fopen(fn, "r")) == NULL) 293 err(2, "%s", fn); 294 while ((line = fgetln(f, &len)) != NULL) 295 add_pattern(line, *line == '\n' ? 0 : len); 296 if (ferror(f)) 297 err(2, "%s", fn); 298 fclose(f); 299} 300 301static inline const char * 302init_color(const char *d) 303{ 304 char *c; 305 306 c = getenv("GREP_COLOR"); 307 return (c != NULL ? c : d); 308} 309 310int 311main(int argc, char *argv[]) 312{ 313 char **aargv, **eargv, *eopts; 314 char *ep; 315 unsigned long long l; 316 unsigned int aargc, eargc, i; 317 int c, lastc, needpattern, newarg, prevoptind; 318 319 setlocale(LC_ALL, ""); 320 321#ifndef WITHOUT_NLS 322 catalog = catopen("grep", NL_CAT_LOCALE); 323#endif 324 325 /* Check what is the program name of the binary. In this 326 way we can have all the funcionalities in one binary 327 without the need of scripting and using ugly hacks. */ 328 switch (__progname[0]) { 329 case 'e': 330 grepbehave = GREP_EXTENDED; 331 break; 332 case 'f': 333 grepbehave = GREP_FIXED; 334 break; 335 case 'g': 336 grepbehave = GREP_BASIC; 337 break; 338 case 'z': 339 filebehave = FILE_GZIP; 340 switch(__progname[1]) { 341 case 'e': 342 grepbehave = GREP_EXTENDED; 343 break; 344 case 'f': 345 grepbehave = GREP_FIXED; 346 break; 347 case 'g': 348 grepbehave = GREP_BASIC; 349 break; 350 } 351 break; 352 } 353 354 lastc = '\0'; 355 newarg = 1; 356 prevoptind = 1; 357 needpattern = 1; 358 359 eopts = getenv("GREP_OPTIONS"); 360 361 /* support for extra arguments in GREP_OPTIONS */ 362 eargc = 0; 363 if (eopts != NULL) { 364 char *str; 365 366 /* make an estimation of how many extra arguments we have */ 367 for (unsigned int j = 0; j < strlen(eopts); j++) 368 if (eopts[j] == ' ') 369 eargc++; 370 371 eargv = (char **)grep_malloc(sizeof(char *) * (eargc + 1)); 372 373 eargc = 0; 374 /* parse extra arguments */ 375 while ((str = strsep(&eopts, " ")) != NULL) 376 eargv[eargc++] = grep_strdup(str); 377 378 aargv = (char **)grep_calloc(eargc + argc + 1, 379 sizeof(char *)); 380 381 aargv[0] = argv[0]; 382 for (i = 0; i < eargc; i++) 383 aargv[i + 1] = eargv[i]; 384 for (int j = 1; j < argc; j++, i++) 385 aargv[i + 1] = argv[j]; 386 387 aargc = eargc + argc; 388 } else { 389 aargv = argv; 390 aargc = argc; 391 } 392 393 while (((c = getopt_long(aargc, aargv, optstr, long_options, NULL)) != 394 -1)) { 395 switch (c) { 396 case '0': case '1': case '2': case '3': case '4': 397 case '5': case '6': case '7': case '8': case '9': 398 if (newarg || !isdigit(lastc)) 399 Aflag = 0; 400 else if (Aflag > LLONG_MAX / 10) { 401 errno = ERANGE; 402 err(2, NULL); 403 } 404 Aflag = Bflag = (Aflag * 10) + (c - '0'); 405 break; 406 case 'C': 407 if (optarg == NULL) { 408 Aflag = Bflag = 2; 409 break; 410 } 411 /* FALLTHROUGH */ 412 case 'A': 413 /* FALLTHROUGH */ 414 case 'B': 415 errno = 0; 416 l = strtoull(optarg, &ep, 10); 417 if (((errno == ERANGE) && (l == ULLONG_MAX)) || 418 ((errno == EINVAL) && (l == 0))) 419 err(2, NULL); 420 else if (ep[0] != '\0') { 421 errno = EINVAL; 422 err(2, NULL); 423 } 424 if (c == 'A') 425 Aflag = l; 426 else if (c == 'B') 427 Bflag = l; 428 else 429 Aflag = Bflag = l; 430 break; 431 case 'a': 432 binbehave = BINFILE_TEXT; 433 break; 434 case 'b': 435 bflag = true; 436 break; 437 case 'c': 438 cflag = true; 439 break; 440 case 'D': 441 if (strcasecmp(optarg, "skip") == 0) 442 devbehave = DEV_SKIP; 443 else if (strcasecmp(optarg, "read") == 0) 444 devbehave = DEV_READ; 445 else 446 errx(2, getstr(3), "--devices"); 447 break; 448 case 'd': 449 if (strcasecmp("recurse", optarg) == 0) { 450 Hflag = true; 451 dirbehave = DIR_RECURSE; 452 } else if (strcasecmp("skip", optarg) == 0) 453 dirbehave = DIR_SKIP; 454 else if (strcasecmp("read", optarg) == 0) 455 dirbehave = DIR_READ; 456 else 457 errx(2, getstr(3), "--directories"); 458 break; 459 case 'E': 460 grepbehave = GREP_EXTENDED; 461 break; 462 case 'e': 463 add_pattern(optarg, strlen(optarg)); 464 needpattern = 0; 465 break; 466 case 'F': 467 grepbehave = GREP_FIXED; 468 break; 469 case 'f': 470 read_patterns(optarg); 471 needpattern = 0; 472 break; 473 case 'G': 474 grepbehave = GREP_BASIC; 475 break; 476 case 'H': 477 Hflag = true; 478 break; 479 case 'h': 480 Hflag = false; 481 hflag = true; 482 break; 483 case 'I': 484 binbehave = BINFILE_SKIP; 485 break; 486 case 'i': 487 case 'y': 488 iflag = true; 489 cflags |= REG_ICASE; 490 break; 491 case 'J': 492 filebehave = FILE_BZIP; 493 break; 494 case 'L': 495 lflag = false; 496 Lflag = true; 497 break; 498 case 'l': 499 Lflag = false; 500 lflag = true; 501 break; 502 case 'm': 503 mflag = true; 504 errno = 0; 505 mcount = strtoull(optarg, &ep, 10); 506 if (((errno == ERANGE) && (mcount == ULLONG_MAX)) || 507 ((errno == EINVAL) && (mcount == 0))) 508 err(2, NULL); 509 else if (ep[0] != '\0') { 510 errno = EINVAL; 511 err(2, NULL); 512 } 513 break; 514 case 'n': 515 nflag = true; 516 break; 517 case 'O': 518 linkbehave = LINK_EXPLICIT; 519 break; 520 case 'o': 521 oflag = true; 522 break; 523 case 'p': 524 linkbehave = LINK_SKIP; 525 break; 526 case 'q': 527 qflag = true; 528 break; 529 case 'S': 530 linkbehave = LINK_READ; 531 break; 532 case 'R': 533 case 'r': 534 dirbehave = DIR_RECURSE; 535 Hflag = true; 536 break; 537 case 's': 538 sflag = true; 539 break; 540 case 'U': 541 binbehave = BINFILE_BIN; 542 break; 543 case 'u': 544 case MMAP_OPT: 545 /* noop, compatibility */ 546 break; 547 case 'V': 548 printf(getstr(9), __progname, VERSION); 549 exit(0); 550 case 'v': 551 vflag = true; 552 break; 553 case 'w': 554 wflag = true; 555 break; 556 case 'x': 557 xflag = true; 558 break; 559 case 'Z': 560 filebehave = FILE_GZIP; 561 break; 562 case BIN_OPT: 563 if (strcasecmp("binary", optarg) == 0) 564 binbehave = BINFILE_BIN; 565 else if (strcasecmp("without-match", optarg) == 0) 566 binbehave = BINFILE_SKIP; 567 else if (strcasecmp("text", optarg) == 0) 568 binbehave = BINFILE_TEXT; 569 else 570 errx(2, getstr(3), "--binary-files"); 571 break; 572 case COLOR_OPT: 573 color = NULL; 574 if (optarg == NULL || strcasecmp("auto", optarg) == 0 || 575 strcasecmp("tty", optarg) == 0 || 576 strcasecmp("if-tty", optarg) == 0) { 577 char *term; 578 579 term = getenv("TERM"); 580 if (isatty(STDOUT_FILENO) && term != NULL && 581 strcasecmp(term, "dumb") != 0) 582 color = init_color("01;31"); 583 } else if (strcasecmp("always", optarg) == 0 || 584 strcasecmp("yes", optarg) == 0 || 585 strcasecmp("force", optarg) == 0) { 586 color = init_color("01;31"); 587 } else if (strcasecmp("never", optarg) != 0 && 588 strcasecmp("none", optarg) != 0 && 589 strcasecmp("no", optarg) != 0) 590 errx(2, getstr(3), "--color"); 591 break; 592 case LABEL_OPT: 593 label = optarg; 594 break; 595 case LINEBUF_OPT: 596 lbflag = true; 597 break; 598 case NULL_OPT: 599 nullflag = true; 600 break; 601 case R_INCLUDE_OPT: 602 finclude = true; 603 add_fpattern(optarg, INCL_PAT); 604 break; 605 case R_EXCLUDE_OPT: 606 fexclude = true; 607 add_fpattern(optarg, EXCL_PAT); 608 break; 609 case R_DINCLUDE_OPT: 610 dinclude = true; 611 add_dpattern(optarg, INCL_PAT); 612 break; 613 case R_DEXCLUDE_OPT: 614 dexclude = true; 615 add_dpattern(optarg, EXCL_PAT); 616 break; 617 case HELP_OPT: 618 default: 619 usage(); 620 } 621 lastc = c; 622 newarg = optind != prevoptind; 623 prevoptind = optind; 624 } 625 aargc -= optind; 626 aargv += optind; 627 628 /* Fail if we don't have any pattern */ 629 if (aargc == 0 && needpattern) 630 usage(); 631 632 /* Process patterns from command line */ 633 if (aargc != 0 && needpattern) { 634 add_pattern(*aargv, strlen(*aargv)); 635 --aargc; 636 ++aargv; 637 } 638 639 switch (grepbehave) { 640 case GREP_FIXED: 641 case GREP_BASIC: 642 break; 643 case GREP_EXTENDED: 644 cflags |= REG_EXTENDED; 645 break; 646 default: 647 /* NOTREACHED */ 648 usage(); 649 } 650 651 fg_pattern = grep_calloc(patterns, sizeof(*fg_pattern)); 652 r_pattern = grep_calloc(patterns, sizeof(*r_pattern)); 653/* 654 * XXX: fgrepcomp() and fastcomp() are workarounds for regexec() performance. 655 * Optimizations should be done there. 656 */ 657 /* Check if cheating is allowed (always is for fgrep). */ 658 if (grepbehave == GREP_FIXED) { 659 for (i = 0; i < patterns; ++i) 660 fgrepcomp(&fg_pattern[i], pattern[i]); 661 } else { 662 for (i = 0; i < patterns; ++i) { 663 if (fastcomp(&fg_pattern[i], pattern[i])) { 664 /* Fall back to full regex library */ 665 c = regcomp(&r_pattern[i], pattern[i], cflags); 666 if (c != 0) { 667 regerror(c, &r_pattern[i], re_error, 668 RE_ERROR_BUF); 669 errx(2, "%s", re_error); 670 } 671 } 672 } 673 } 674 675 if (lbflag) 676 setlinebuf(stdout); 677 678 if ((aargc == 0 || aargc == 1) && !Hflag) 679 hflag = true; 680 681 if (aargc == 0) 682 exit(!procfile("-")); 683 684 if (dirbehave == DIR_RECURSE) 685 c = grep_tree(aargv); 686 else 687 for (c = 0; aargc--; ++aargv) { 688 if ((finclude || fexclude) && !file_matching(*aargv)) 689 continue; 690 c+= procfile(*aargv); 691 } 692 693#ifndef WITHOUT_NLS 694 catclose(catalog); 695#endif 696 697 /* Find out the correct return value according to the 698 results and the command line option. */ 699 exit(c ? (notfound ? (qflag ? 0 : 2) : 0) : (notfound ? 2 : 1)); 700} 701