1 /* $NetBSD: grep.c,v 1.6 2011/04/18 03:48:23 joerg Exp $ */
3 /* $OpenBSD: grep.c,v 1.42 2010/07/02 22:18:03 tedu Exp $ */
6 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
8 * Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav
9 * Copyright (C) 2008-2009 Gabor Kovesdan <gabor@FreeBSD.org>
10 * All rights reserved.
12 * Redistribution and use in source and binary forms, with or without
13 * modification, are permitted provided that the following conditions
15 * 1. Redistributions of source code must retain the above copyright
16 * notice, this list of conditions and the following disclaimer.
17 * 2. Redistributions in binary form must reproduce the above copyright
18 * notice, this list of conditions and the following disclaimer in the
19 * documentation and/or other materials provided with the distribution.
21 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 #include <sys/cdefs.h>
35 __FBSDID("$FreeBSD$");
38 #include <sys/types.h>
56 const char *errstr[] = {
58 /* 1*/ "(standard input)",
59 /* 2*/ "unknown %s option",
60 /* 3*/ "usage: %s [-abcDEFGHhIiLlmnOoPqRSsUVvwxz] [-A num] [-B num] [-C[num]]\n",
61 /* 4*/ "\t[-e pattern] [-f file] [--binary-files=value] [--color=when]\n",
62 /* 5*/ "\t[--context[=num]] [--directories=action] [--label] [--line-buffered]\n",
63 /* 6*/ "\t[--null] [pattern] [file ...]\n",
64 /* 7*/ "Binary file %s matches\n",
65 /* 8*/ "%s (BSD grep) %s\n",
66 /* 9*/ "%s (BSD grep, GNU compatible) %s\n",
69 /* Flags passed to regcomp() and regexec() */
70 int cflags = REG_NOSUB | REG_NEWLINE;
71 int eflags = REG_STARTEND;
75 /* Searching patterns */
76 unsigned int patterns;
77 static unsigned int pattern_sz;
81 /* Filename exclusion/inclusion patterns */
82 unsigned int fpatterns, dpatterns;
83 static unsigned int fpattern_sz, dpattern_sz;
84 struct epat *dpattern, *fpattern;
86 /* For regex errors */
87 char re_error[RE_ERROR_BUF + 1];
89 /* Command-line flags */
90 long long Aflag; /* -A x: print x lines trailing each match */
91 long long Bflag; /* -B x: print x lines leading each match */
92 bool Hflag; /* -H: always print file name */
93 bool Lflag; /* -L: only show names of files with no matches */
94 bool bflag; /* -b: show block numbers for each match */
95 bool cflag; /* -c: only show a count of matching lines */
96 bool hflag; /* -h: don't print filename headers */
97 bool iflag; /* -i: ignore case */
98 bool lflag; /* -l: only show names of files with matches */
99 bool mflag; /* -m x: stop reading the files after x matches */
100 long long mcount; /* count for -m */
101 long long mlimit; /* requested value for -m */
102 char fileeol; /* indicator for eol */
103 bool nflag; /* -n: show line numbers in front of matching lines */
104 bool oflag; /* -o: print only matching part */
105 bool qflag; /* -q: quiet mode (don't output anything) */
106 bool sflag; /* -s: silent mode (ignore errors) */
107 bool vflag; /* -v: only show non-matching lines */
108 bool wflag; /* -w: pattern must start and end on word boundaries */
109 bool xflag; /* -x: pattern must match entire line */
110 bool lbflag; /* --line-buffered */
111 bool nullflag; /* --null */
112 char *label; /* --label */
113 const char *color; /* --color */
114 int grepbehave = GREP_BASIC; /* -EFGP: type of the regex */
115 int binbehave = BINFILE_BIN; /* -aIU: handling of binary files */
116 int filebehave = FILE_STDIO;
117 int devbehave = DEV_READ; /* -D: handling of devices */
118 int dirbehave = DIR_READ; /* -dRr: handling of directories */
119 int linkbehave = LINK_READ; /* -OpS: handling of symlinks */
121 bool dexclude, dinclude; /* --exclude-dir and --include-dir */
122 bool fexclude, finclude; /* --exclude and --include */
125 BIN_OPT = CHAR_MAX + 1,
138 static inline const char *init_color(const char *);
141 bool file_err; /* file reading error */
144 * Prints usage information and returns 2.
149 fprintf(stderr, errstr[3], getprogname());
150 fprintf(stderr, "%s", errstr[4]);
151 fprintf(stderr, "%s", errstr[5]);
152 fprintf(stderr, "%s", errstr[6]);
156 static const char *optstr = "0123456789A:B:C:D:EFGHILOPSRUVabcd:e:f:hilm:nopqrsuvwxyz";
158 static const struct option long_options[] =
160 {"binary-files", required_argument, NULL, BIN_OPT},
161 {"help", no_argument, NULL, HELP_OPT},
162 {"mmap", no_argument, NULL, MMAP_OPT},
163 {"line-buffered", no_argument, NULL, LINEBUF_OPT},
164 {"label", required_argument, NULL, LABEL_OPT},
165 {"null", no_argument, NULL, NULL_OPT},
166 {"color", optional_argument, NULL, COLOR_OPT},
167 {"colour", optional_argument, NULL, COLOR_OPT},
168 {"exclude", required_argument, NULL, R_EXCLUDE_OPT},
169 {"include", required_argument, NULL, R_INCLUDE_OPT},
170 {"exclude-dir", required_argument, NULL, R_DEXCLUDE_OPT},
171 {"include-dir", required_argument, NULL, R_DINCLUDE_OPT},
172 {"after-context", required_argument, NULL, 'A'},
173 {"text", no_argument, NULL, 'a'},
174 {"before-context", required_argument, NULL, 'B'},
175 {"byte-offset", no_argument, NULL, 'b'},
176 {"context", optional_argument, NULL, 'C'},
177 {"count", no_argument, NULL, 'c'},
178 {"devices", required_argument, NULL, 'D'},
179 {"directories", required_argument, NULL, 'd'},
180 {"extended-regexp", no_argument, NULL, 'E'},
181 {"regexp", required_argument, NULL, 'e'},
182 {"fixed-strings", no_argument, NULL, 'F'},
183 {"file", required_argument, NULL, 'f'},
184 {"basic-regexp", no_argument, NULL, 'G'},
185 {"no-filename", no_argument, NULL, 'h'},
186 {"with-filename", no_argument, NULL, 'H'},
187 {"ignore-case", no_argument, NULL, 'i'},
188 {"files-with-matches", no_argument, NULL, 'l'},
189 {"files-without-match", no_argument, NULL, 'L'},
190 {"max-count", required_argument, NULL, 'm'},
191 {"line-number", no_argument, NULL, 'n'},
192 {"only-matching", no_argument, NULL, 'o'},
193 {"quiet", no_argument, NULL, 'q'},
194 {"silent", no_argument, NULL, 'q'},
195 {"recursive", no_argument, NULL, 'r'},
196 {"no-messages", no_argument, NULL, 's'},
197 {"binary", no_argument, NULL, 'U'},
198 {"unix-byte-offsets", no_argument, NULL, 'u'},
199 {"invert-match", no_argument, NULL, 'v'},
200 {"version", no_argument, NULL, 'V'},
201 {"word-regexp", no_argument, NULL, 'w'},
202 {"line-regexp", no_argument, NULL, 'x'},
203 {"null-data", no_argument, NULL, 'z'},
204 {NULL, no_argument, NULL, 0}
208 * Adds a searching pattern to the internal array.
211 add_pattern(char *pat, size_t len)
214 /* Check if we can do a shortcut */
219 /* Increase size if necessary */
220 if (patterns == pattern_sz) {
222 pattern = grep_realloc(pattern, ++pattern_sz *
225 if (len > 0 && pat[len - 1] == '\n')
227 /* pat may not be NUL-terminated */
228 pattern[patterns].pat = grep_malloc(len + 1);
229 memcpy(pattern[patterns].pat, pat, len);
230 pattern[patterns].len = len;
231 pattern[patterns].pat[len] = '\0';
236 * Adds a file include/exclude pattern to the internal array.
239 add_fpattern(const char *pat, int mode)
242 /* Increase size if necessary */
243 if (fpatterns == fpattern_sz) {
245 fpattern = grep_realloc(fpattern, ++fpattern_sz *
246 sizeof(struct epat));
248 fpattern[fpatterns].pat = grep_strdup(pat);
249 fpattern[fpatterns].mode = mode;
254 * Adds a directory include/exclude pattern to the internal array.
257 add_dpattern(const char *pat, int mode)
260 /* Increase size if necessary */
261 if (dpatterns == dpattern_sz) {
263 dpattern = grep_realloc(dpattern, ++dpattern_sz *
264 sizeof(struct epat));
266 dpattern[dpatterns].pat = grep_strdup(pat);
267 dpattern[dpatterns].mode = mode;
272 * Reads searching patterns from a file and adds them with add_pattern().
275 read_patterns(const char *fn)
283 if (strcmp(fn, "-") == 0)
285 else if ((f = fopen(fn, "r")) == NULL)
287 if ((fstat(fileno(f), &st) == -1) || (S_ISDIR(st.st_mode))) {
293 while ((rlen = getline(&line, &len, f)) != -1) {
296 add_pattern(line, line[0] == '\n' ? 0 : (size_t)rlen);
302 if (strcmp(fn, "-") != 0)
306 static inline const char *
307 init_color(const char *d)
311 c = getenv("GREP_COLOR");
312 return (c != NULL && c[0] != '\0' ? c : d);
316 main(int argc, char *argv[])
318 char **aargv, **eargv, *eopts;
322 unsigned int aargc, eargc, i;
323 int c, lastc, needpattern, newarg, prevoptind;
326 setlocale(LC_ALL, "");
329 * Check how we've bene invoked to determine the behavior we should
330 * exhibit. In this way we can have all the functionalities in one
331 * binary without the need of scripting and using ugly hacks.
336 grepbehave = GREP_EXTENDED;
339 grepbehave = GREP_FIXED;
342 dirbehave = DIR_RECURSE;
353 eopts = getenv("GREP_OPTIONS");
355 /* support for extra arguments in GREP_OPTIONS */
357 if (eopts != NULL && eopts[0] != '\0') {
360 /* make an estimation of how many extra arguments we have */
361 for (unsigned int j = 0; j < strlen(eopts); j++)
365 eargv = (char **)grep_malloc(sizeof(char *) * (eargc + 1));
368 /* parse extra arguments */
369 while ((str = strsep(&eopts, " ")) != NULL)
371 eargv[eargc++] = grep_strdup(str);
373 aargv = (char **)grep_calloc(eargc + argc + 1,
377 for (i = 0; i < eargc; i++)
378 aargv[i + 1] = eargv[i];
379 for (int j = 1; j < argc; j++, i++)
380 aargv[i + 1] = argv[j];
382 aargc = eargc + argc;
388 while (((c = getopt_long(aargc, aargv, optstr, long_options, NULL)) !=
391 case '0': case '1': case '2': case '3': case '4':
392 case '5': case '6': case '7': case '8': case '9':
393 if (newarg || !isdigit(lastc))
395 else if (Aflag > LLONG_MAX / 10 - 1) {
400 Aflag = Bflag = (Aflag * 10) + (c - '0');
403 if (optarg == NULL) {
412 l = strtoll(optarg, &ep, 10);
413 if (errno == ERANGE || errno == EINVAL)
415 else if (ep[0] != '\0') {
420 err(2, "context argument must be non-negative");
431 binbehave = BINFILE_TEXT;
440 if (strcasecmp(optarg, "skip") == 0)
441 devbehave = DEV_SKIP;
442 else if (strcasecmp(optarg, "read") == 0)
443 devbehave = DEV_READ;
445 errx(2, errstr[2], "--devices");
448 if (strcasecmp("recurse", optarg) == 0) {
450 dirbehave = DIR_RECURSE;
451 } else if (strcasecmp("skip", optarg) == 0)
452 dirbehave = DIR_SKIP;
453 else if (strcasecmp("read", optarg) == 0)
454 dirbehave = DIR_READ;
456 errx(2, errstr[2], "--directories");
459 grepbehave = GREP_EXTENDED;
464 char *string = optarg;
466 while ((token = strsep(&string, "\n")) != NULL)
467 add_pattern(token, strlen(token));
472 grepbehave = GREP_FIXED;
475 read_patterns(optarg);
479 grepbehave = GREP_BASIC;
489 binbehave = BINFILE_SKIP;
507 mlimit = mcount = strtoll(optarg, &ep, 10);
508 if (((errno == ERANGE) && (mcount == LLONG_MAX)) ||
509 ((errno == EINVAL) && (mcount == 0)))
511 else if (ep[0] != '\0') {
520 linkbehave = LINK_EXPLICIT;
524 cflags &= ~REG_NOSUB;
527 linkbehave = LINK_SKIP;
533 linkbehave = LINK_READ;
537 dirbehave = DIR_RECURSE;
544 binbehave = BINFILE_BIN;
548 filebehave = FILE_MMAP;
551 #ifdef WITH_GNU_COMPAT
552 printf(errstr[9], getprogname(), VERSION);
554 printf(errstr[8], getprogname(), VERSION);
562 cflags &= ~REG_NOSUB;
566 cflags &= ~REG_NOSUB;
572 if (strcasecmp("binary", optarg) == 0)
573 binbehave = BINFILE_BIN;
574 else if (strcasecmp("without-match", optarg) == 0)
575 binbehave = BINFILE_SKIP;
576 else if (strcasecmp("text", optarg) == 0)
577 binbehave = BINFILE_TEXT;
579 errx(2, errstr[2], "--binary-files");
583 if (optarg == NULL || strcasecmp("auto", optarg) == 0 ||
584 strcasecmp("tty", optarg) == 0 ||
585 strcasecmp("if-tty", optarg) == 0) {
588 term = getenv("TERM");
589 if (isatty(STDOUT_FILENO) && term != NULL &&
590 strcasecmp(term, "dumb") != 0)
591 color = init_color("01;31");
592 } else if (strcasecmp("always", optarg) == 0 ||
593 strcasecmp("yes", optarg) == 0 ||
594 strcasecmp("force", optarg) == 0) {
595 color = init_color("01;31");
596 } else if (strcasecmp("never", optarg) != 0 &&
597 strcasecmp("none", optarg) != 0 &&
598 strcasecmp("no", optarg) != 0)
599 errx(2, errstr[2], "--color");
600 cflags &= ~REG_NOSUB;
613 add_fpattern(optarg, INCL_PAT);
617 add_fpattern(optarg, EXCL_PAT);
621 add_dpattern(optarg, INCL_PAT);
625 add_dpattern(optarg, EXCL_PAT);
632 newarg = optind != prevoptind;
638 /* xflag takes precedence, don't confuse the matching bits. */
642 /* Fail if we don't have any pattern */
643 if (aargc == 0 && needpattern)
646 /* Process patterns from command line */
647 if (aargc != 0 && needpattern) {
649 char *string = *aargv;
651 while ((token = strsep(&string, "\n")) != NULL)
652 add_pattern(token, strlen(token));
657 switch (grepbehave) {
662 * regex(3) implementations that support fixed-string searches generally
663 * define either REG_NOSPEC or REG_LITERAL. Set the appropriate flag
664 * here. If neither are defined, GREP_FIXED later implies that the
665 * internal literal matcher should be used. Other cflags that have
666 * the same interpretation as REG_NOSPEC and REG_LITERAL should be
667 * similarly added here, and grep.h should be amended to take this into
668 * consideration when defining WITH_INTERNAL_NOSPEC.
670 #if defined(REG_NOSPEC)
671 cflags |= REG_NOSPEC;
672 #elif defined(REG_LITERAL)
673 cflags |= REG_LITERAL;
677 cflags |= REG_EXTENDED;
684 r_pattern = grep_calloc(patterns, sizeof(*r_pattern));
686 #ifdef WITH_INTERNAL_NOSPEC
687 if (grepbehave != GREP_FIXED) {
691 /* Check if cheating is allowed (always is for fgrep). */
692 for (i = 0; i < patterns; ++i) {
693 c = regcomp(&r_pattern[i], pattern[i].pat, cflags);
695 regerror(c, &r_pattern[i], re_error,
697 errx(2, "%s", re_error);
705 if ((aargc == 0 || aargc == 1) && !Hflag)
710 if (aargc == 0 && dirbehave != DIR_RECURSE)
711 exit(!procfile("-"));
713 if (dirbehave == DIR_RECURSE)
714 matched = grep_tree(aargv);
716 for (matched = false; aargc--; ++aargv) {
717 if ((finclude || fexclude) && !file_matching(*aargv))
719 if (procfile(*aargv))
727 * Calculate the correct return value according to the
728 * results and the command line option.
730 exit(matched ? (file_err ? (qflag ? 0 : 2) : 0) : (file_err ? 2 : 1));