xref: /netbsd-src/usr.bin/grep/grep.c (revision d20841bb642898112fe68f0ad3f7b26dddf56f07)
1 /*	$NetBSD: grep.c,v 1.1.1.2 2004/01/02 15:00:29 cjep Exp $	*/
2 
3 /*-
4  * Copyright (c) 1999 James Howard and Dag-Erling Co�dan Sm�rgrav
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, 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
20  * FOR 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 
30 
31 
32 #include <sys/cdefs.h>
33 #ifndef lint
34 __RCSID("$NetBSD: grep.c,v 1.1.1.2 2004/01/02 15:00:29 cjep Exp $");
35 #endif /* not lint */
36 
37 #include <sys/types.h>
38 #include <sys/stat.h>
39 
40 #include <err.h>
41 #include <errno.h>
42 #include <getopt.h>
43 #include <limits.h>
44 #include <regex.h>
45 #include <stdio.h>
46 #include <stdlib.h>
47 #include <string.h>
48 #include <unistd.h>
49 
50 #include "grep.h"
51 
52 /*
53  * Upper bound of number of digits to represent an int in decimal
54  *  2^8n <= 10^3n. Allow a terminator.
55  */
56 #define MAX_BUF_DIGITS	(sizeof(int) * 3) + 1
57 
58 /* Flags passed to regcomp() and regexec() */
59 int cflags = REG_BASIC;
60 int eflags = REG_STARTEND;
61 
62 int matchall;	/* shortcut */
63 int patterns, pattern_sz;
64 char **pattern;
65 regex_t	*r_pattern;
66 
67 /* For regex errors  */
68 char re_error[RE_ERROR_BUF + 1];
69 
70 /* Command-line flags */
71 int Aflag;		/* -A x: print x lines trailing each match */
72 int Bflag;		/* -B x: print x lines leading each match */
73 int Eflag;		/* -E: interpret pattern as extended regexp */
74 int Fflag;		/* -F: interpret pattern as list of fixed strings */
75 int Gflag;		/* -G: interpret pattern as basic regexp */
76 int Hflag;		/* -H: Always print filenames */
77 int Lflag;		/* -L: only show names of files with no matches */
78 /*int Pflag;		*//* -P: if -r, no symlinks are followed */
79 /*int Sflag;		*//* -S: if -r, follow all symlinks */
80 int bflag;		/* -b: show block numbers for each match */
81 int cflag;		/* -c: only show a count of matching lines */
82 int hflag;		/* -h: Never print filenames. -H overrides */
83 int lflag;		/* -l: only show names of files with matches */
84 int mflag;		/* -m: specify maximum line matches (per file) */
85 int nflag;		/* -n: show line numbers in front of matching lines */
86 int oflag;		/* -o: only print out matches */
87 int qflag;		/* -q: quiet mode (don't output anything) */
88 int sflag;		/* -s: silent mode (ignore errors) */
89 int vflag;		/* -v: only show non-matching lines */
90 int wflag;		/* -w: pattern must start and end on word boundaries */
91 int xflag;		/* -x: pattern must match entire line */
92 
93 int colours = 0;	/* Attempt to use terminal colours */
94 const char *grep_colour = "01;32"; /* Default colour string, green */
95 char *uc;
96 
97 /* Characters to print after filenames */
98 char fn_endchar = '\n';
99 char fn_colonchar = ':';
100 char fn_dashchar = '-';
101 char line_endchar = '\n';	/* End of line character */
102 
103 int maxcount = 0;		/* Maximum line matches per file */
104 int output_filenames = 0;
105 
106 /* Argv[0] flags */
107 int zgrep;		/* If we are invoked as zgrep */
108 
109 int binbehave = BIN_FILE_BIN;
110 int dirbehave = GREP_READ;
111 int devbehave = GREP_READ;
112 /*int linkbehave = LINK_FOLLOW;*/
113 char *stdin_label;
114 
115 enum {
116 	BIN_OPT = CHAR_MAX + 1,
117 	HELP_OPT,
118 	LABEL_OPT,
119 	MMAP_OPT,
120 	LINK_OPT,
121 	COLOUR_OPT
122 };
123 
124 /* Housekeeping */
125 int first;		/* flag whether or not this is our first match */
126 int tail;		/* lines left to print */
127 
128 static void
129 usage(void)
130 {
131 	fprintf(stderr, "usage: %s %s %s\n",
132 		getprogname(),
133 		"[-[ABC] num] [-EFGHILVZabcdhilnoqrsvwxz]",
134 		"[-D action] [-d action] [-e pattern] [-f file]");
135 	exit(2);
136 }
137 
138 static char *optstr = "0123456789A:B:C:D:EFGHILUVZabcd:e:f:hilm:noqrsuvwxyz";
139 
140 struct option long_options[] =
141 {
142 	{"binary-files",	required_argument, NULL, BIN_OPT},
143 	{"help",		no_argument, 	   NULL, HELP_OPT},
144 	{"label",		required_argument, NULL, LABEL_OPT},
145 	{"mmap",		no_argument, 	   NULL, MMAP_OPT},
146 /*	{"links",		required_argument, NULL, LINK_OPT},*/
147 	{"after-context",       required_argument, NULL, 'A'},
148 	{"before-context",      required_argument, NULL, 'B'},
149 	{"color",		optional_argument, NULL, COLOUR_OPT},
150 	{"colour",		optional_argument, NULL, COLOUR_OPT},
151 	{"context",             optional_argument, NULL, 'C'},
152 	{"devices",		required_argument, NULL, 'D'},
153 	{"extended-regexp",     no_argument,       NULL, 'E'},
154 	{"fixed-strings",       no_argument,       NULL, 'F'},
155 	{"fixed-regexp",	no_argument,       NULL, 'F'},
156 	{"basic-regexp",        no_argument,       NULL, 'G'},
157 	{"with-filename",	no_argument,	   NULL, 'H'},
158 	{"files-without-match", no_argument,       NULL, 'L'},
159 	{"binary",              no_argument,       NULL, 'U'},
160 	{"version",             no_argument,       NULL, 'V'},
161 	{"null",		no_argument,	   NULL, 'Z'},
162 	{"text",                no_argument,       NULL, 'a'},
163 	{"byte-offset",         no_argument,       NULL, 'b'},
164 	{"count",               no_argument,       NULL, 'c'},
165 	{"directories",		required_argument, NULL, 'd'},
166 	{"regexp",              required_argument, NULL, 'e'},
167 	{"file",                required_argument, NULL, 'f'},
168 	{"no-filename",         no_argument,       NULL, 'h'},
169 	{"ignore-case",         no_argument,       NULL, 'i'},
170 	{"files-with-matches",  no_argument,       NULL, 'l'},
171 	{"max-count",		required_argument, NULL, 'm'},
172 	{"line-number",         no_argument,       NULL, 'n'},
173 	{"only-matching",	no_argument,	   NULL, 'o'},
174 	{"quiet",               no_argument,       NULL, 'q'},
175 	{"silent",              no_argument,       NULL, 'q'},
176 	{"recursive",           no_argument,       NULL, 'r'},
177 	{"no-messages",         no_argument,       NULL, 's'},
178 	{"unix-byte-offsets",   no_argument,       NULL, 'u'},
179 	{"invert-match",        no_argument,       NULL, 'v'},
180 	{"word-regexp",         no_argument,       NULL, 'w'},
181 	{"line-regexp",         no_argument,       NULL, 'x'},
182 	{"null-data",		no_argument,	   NULL, 'z'},
183 
184 	{NULL,                  no_argument,       NULL, 0}
185 };
186 
187 static void
188 add_pattern(char *pat, size_t len)
189 {
190 	if (len == 0 || matchall) {
191 		matchall = 1;
192 		return;
193 	}
194 	if (patterns == pattern_sz) {
195 		pattern_sz *= 2;
196 		pattern = grep_realloc(pattern, ++pattern_sz * sizeof(*pattern));
197 	}
198 	if (pat[len - 1] == '\n')
199 		--len;
200 	pattern[patterns] = grep_malloc(len + 1);
201 	strncpy(pattern[patterns], pat, len);
202 	pattern[patterns][len] = '\0';
203 	++patterns;
204 }
205 
206 static void
207 read_patterns(char *fn)
208 {
209 	FILE *f;
210 	char *line;
211 	size_t len;
212 	int nl;
213 
214 	if ((f = fopen(fn, "r")) == NULL)
215 		err(2, "%s", fn);
216 	nl = 0;
217 	while ((line = fgetln(f, &len)) != NULL) {
218 		if (*line == '\n') {
219 			++nl;
220 			continue;
221 		}
222 		if (nl) {
223 			matchall = 1;
224 			break;
225 		}
226 		nl = 0;
227 		add_pattern(line, len);
228 	}
229 	if (ferror(f))
230 		err(2, "%s", fn);
231 	fclose(f);
232 }
233 
234 static int
235 check_context_arg(char const *str) {
236 	char *ep;
237 	long lval;
238 
239 	errno = 0;
240 	lval = strtol(str, &ep, 10);
241 
242 	if (str[0] == '\0' || *ep != '\0')
243 		errx(2, "Invalid context argument");
244 
245 	if ((errno == ERANGE && (lval == LONG_MAX || lval == LONG_MIN)) ||
246  	    (lval > INT_MAX || lval < INT_MIN))
247 		errx(2, "Context argument out of range");
248 
249 	return lval;
250 
251 }
252 
253 static int
254 grep_getopt(int argc, char *const *argv)
255 {
256 	int c, ptr;
257 	char buffer[MAX_BUF_DIGITS];
258 
259 	ptr = 0;
260 	while (c = getopt_long(argc, argv, optstr, long_options,
261 		(int *)NULL), '0' <= c &&
262 		c <= '9' && ptr < MAX_BUF_DIGITS) {
263 
264 		/* Avoid leading zeros */
265 		if (ptr != 0 || (ptr == 0 && c != '0'))
266 			buffer[ptr++] = c;
267 	}
268 
269 	if (ptr >= MAX_BUF_DIGITS)
270 		errx(2, "Context argument out of range");
271 
272 	if (ptr) {
273 		buffer[ptr] = '\0';	/* We now have a string of digits */
274 		Aflag = Bflag = check_context_arg(buffer);
275 	}
276 
277 	return c;
278 }
279 
280 int
281 main(int argc, char *argv[])
282 {
283 	const char *progname;
284 	int c, i;
285 	struct stat sb;
286 
287 	stdin_label = "(standard input)";
288 
289 	progname = getprogname();
290 	switch (progname[0]) {
291 	case 'e':
292 		Eflag++;
293 		break;
294 	case 'f':
295 		Fflag++;
296 		break;
297 	case 'g':
298 		Gflag++;
299 		break;
300 	case 'z':
301 		zgrep++;
302 		switch (progname[1]) {
303 		case 'e':
304 			Eflag++;
305 			break;
306 		case 'f':
307 			Fflag++;
308 			break;
309 		case 'g':
310 			Gflag++;
311 			break;
312 		}
313 		break;
314 	}
315 
316 	while ((c = grep_getopt(argc, argv)) != -1) {
317 
318 		switch (c) {
319 
320 		case 'A':
321 			Aflag = check_context_arg(optarg);
322 			break;
323 		case 'B':
324 			Bflag = check_context_arg(optarg);
325 			break;
326 		case 'C':
327 			if (optarg == NULL)
328 				Aflag = Bflag = 2;
329 			else
330 				Aflag = Bflag = check_context_arg(optarg);
331 			break;
332 		case 'D':
333 			if (strcmp("read", optarg) == 0)
334 				devbehave = GREP_READ;
335 			else if (strcmp("skip", optarg) == 0)
336 				devbehave = GREP_SKIP;
337 			else {
338 				 errx(2, "Unknown device option");
339 			}
340 			break;
341 
342 		case 'E':
343 			Fflag = Gflag = 0;
344 			Eflag++;
345 			break;
346 		case 'F':
347 			Eflag = Gflag = 0;
348 			Fflag++;
349 			break;
350 		case 'G':
351 			Eflag = Fflag = 0;
352 			Gflag++;
353 			break;
354 		case 'H':
355 			Hflag = 1;
356 			break;
357 		case 'I':
358 			binbehave = BIN_FILE_SKIP;
359 			break;
360 		case 'L':
361 			lflag = 0;
362 			Lflag = qflag = 1;
363 			break;
364 /*		case 'P':
365 			linkbehave = LINK_SKIP;
366 			break;
367 		case 'S':
368 			linkbehave = LINK_FOLLOW;
369 			break;*/
370 		case 'R':
371 		case 'r':
372 			dirbehave = GREP_RECURSE;
373 			break;
374 		case 'U':
375 		case 'u':
376 			/* these are here for compatability */
377 			break;
378 		case 'V':
379 			fprintf(stdout, "grep version %s\n", VERSION);
380 			exit(0);
381 			break;
382 		case 'Z':
383 			fn_colonchar = fn_endchar = fn_dashchar = 0;
384 			break;
385 		case 'a':
386 			binbehave = BIN_FILE_TEXT;
387 			break;
388 		case 'b':
389 			bflag = 1;
390 			break;
391 		case 'c':
392 			cflag = 1;
393 			break;
394 		case 'd':
395 			if (strcmp("read", optarg) == 0)
396 				dirbehave = GREP_READ;
397 			else if (strcmp("skip", optarg) == 0)
398 				dirbehave = GREP_SKIP;
399 			else if (strcmp("recurse", optarg) == 0)
400 				dirbehave = GREP_RECURSE;
401 			else {
402 				errx(2, "Unknown directory option\n");
403 			}
404 			break;
405 
406 		case 'e':
407 			add_pattern(optarg, strlen(optarg));
408 			break;
409 		case 'f':
410 			read_patterns(optarg);
411 			break;
412 		case 'h':
413 			hflag = 1;
414 			break;
415 		case 'i':
416 		case 'y':
417 			cflags |= REG_ICASE;
418 			break;
419 		case 'l':
420 			Lflag = 0;
421 			lflag = qflag = 1;
422 			break;
423 		case 'm':
424 			mflag = 1;
425 			maxcount = strtol(optarg, (char **)NULL, 10);
426 			break;
427 		case 'n':
428 			nflag = 1;
429 			break;
430 		case 'o':
431 			oflag = 1;
432 			break;
433 		case 'q':
434 			qflag = 1;
435 			break;
436 		case 's':
437 			sflag = 1;
438 			break;
439 		case 'v':
440 			vflag = 1;
441 			break;
442 		case 'w':
443 			wflag = 1;
444 			break;
445 		case 'x':
446 			xflag = 1;
447 			break;
448 		case 'z':
449 			line_endchar = 0;
450 			break;
451 		case BIN_OPT:
452 			if (strcmp("binary", optarg) == 0)
453 				binbehave = BIN_FILE_BIN;
454 			else if (strcmp("without-match", optarg) == 0)
455 				binbehave = BIN_FILE_SKIP;
456 			else if (strcmp("text", optarg) == 0)
457 				binbehave = BIN_FILE_TEXT;
458 			else {
459                                 errx(2, "Unknown binary-files option\n");
460 			}
461                         break;
462 
463 		case COLOUR_OPT:
464 			if (optarg == NULL || strcmp("auto", optarg) == 0 ||
465 			      strcmp("tty", optarg) == 0 ||
466 			      strcmp("if-tty", optarg) == 0) {
467 
468 				/* Check that stdout is a terminal */
469 				if (isatty(STDOUT_FILENO) &&
470 					getenv("TERM") &&
471 					strcmp(getenv("TERM"), "dumb") != 0)
472 						colours = 1;
473 				else
474 					colours = 0;
475 
476 			} else if (strcmp("always", optarg) == 0 ||
477 				   strcmp("yes", optarg) == 0 ||
478 				   strcmp("force", optarg) == 0)
479 				colours = 1;
480 			else if (strcmp("never", optarg) == 0 ||
481 				 strcmp("no", optarg) == 0 ||
482 				 strcmp("none", optarg) == 0)
483 				colours = 0;
484 			else
485 				errx(2, "Unknown color option\n");
486 
487 			uc = getenv("GREP_COLOR");
488 			if (colours == 1 && uc != NULL && *uc != '\0')
489 				grep_colour = uc;
490 			break;
491 		case LABEL_OPT:
492 			stdin_label = optarg;
493 			break;
494 		case MMAP_OPT:
495 			break;
496 /*
497  * 		case LINK_OPT:
498  * 			if (strcmp("explicit", optarg) == 0)
499  * 				linkbehave = LINK_EXPLICIT;
500  * 			else if (strcmp("follow", optarg) == 0)
501  * 				linkbehave = LINK_FOLLOW;
502  * 			else if (strcmp("skip", optarg) == 0)
503  * 				linkbehave = LINK_SKIP;
504  *                         else {
505  *                                 errx(2, "Unknown links option\n");
506  * 			}
507  *                         break;
508  */
509 
510 		case HELP_OPT:
511 		default:
512 			usage();
513 		}
514 
515 	}
516 
517 	argc -= optind;
518 	argv += optind;
519 
520 	if (argc == 0 && patterns == 0)
521 		usage();
522 	if (patterns == 0) {
523 		add_pattern(*argv, strlen(*argv));
524 		--argc;
525 		++argv;
526 	}
527 
528 	if (Eflag)
529 		cflags |= REG_EXTENDED;
530 	else if (Fflag)
531 		cflags |= REG_NOSPEC;
532 	r_pattern = grep_malloc(patterns * sizeof(*r_pattern));
533 	for (i = 0; i < patterns; ++i) {
534 		if ((c = regcomp(&r_pattern[i], pattern[i], cflags))) {
535 			regerror(c, &r_pattern[i], re_error, RE_ERROR_BUF);
536 			errx(2, "%s", re_error);
537 		}
538 	}
539 
540 	if ((argc > 1 && !hflag) || Hflag)
541 		output_filenames = 1;
542 
543 	if (argc == 1 && !hflag && dirbehave == GREP_RECURSE)
544 		if (!stat(*argv, &sb) && (sb.st_mode & S_IFMT) == S_IFDIR)
545 			output_filenames = 1;
546 
547 	if (argc == 0)
548 		exit(!procfile(NULL));
549 
550 	if (dirbehave == GREP_RECURSE)
551 		c = grep_tree(argv);
552 	else
553 		for (c = 0; argc--; ++argv)
554 			c += procfile(*argv);
555 
556 	exit(!c);
557 }
558