xref: /freebsd/usr.bin/grep/util.c (revision 3b3a8eb937bf8045231e8364bfd1b94cd4a95979)
1 /*	$NetBSD: util.c,v 1.9 2011/02/27 17:33:37 joerg Exp $	*/
2 /*	$FreeBSD$	*/
3 /*	$OpenBSD: util.c,v 1.39 2010/07/02 22:18:03 tedu Exp $	*/
4 
5 /*-
6  * Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav
7  * Copyright (C) 2008-2010 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("$FreeBSD$");
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 <fnmatch.h>
42 #include <fts.h>
43 #include <libgen.h>
44 #include <stdbool.h>
45 #include <stdio.h>
46 #include <stdlib.h>
47 #include <string.h>
48 #include <unistd.h>
49 #include <wchar.h>
50 #include <wctype.h>
51 
52 #include "fastmatch.h"
53 #include "grep.h"
54 
55 static int	 linesqueued;
56 static int	 procline(struct str *l, int);
57 
58 bool
59 file_matching(const char *fname)
60 {
61 	char *fname_base;
62 	bool ret;
63 
64 	ret = finclude ? false : true;
65 	fname_base = basename(fname);
66 
67 	for (unsigned int i = 0; i < fpatterns; ++i) {
68 		if (fnmatch(fpattern[i].pat, fname, 0) == 0 ||
69 		    fnmatch(fpattern[i].pat, fname_base, 0) == 0) {
70 			if (fpattern[i].mode == EXCL_PAT)
71 				return (false);
72 			else
73 				ret = true;
74 		}
75 	}
76 	return (ret);
77 }
78 
79 static inline bool
80 dir_matching(const char *dname)
81 {
82 	bool ret;
83 
84 	ret = dinclude ? false : true;
85 
86 	for (unsigned int i = 0; i < dpatterns; ++i) {
87 		if (dname != NULL &&
88 		    fnmatch(dpattern[i].pat, dname, 0) == 0) {
89 			if (dpattern[i].mode == EXCL_PAT)
90 				return (false);
91 			else
92 				ret = true;
93 		}
94 	}
95 	return (ret);
96 }
97 
98 /*
99  * Processes a directory when a recursive search is performed with
100  * the -R option.  Each appropriate file is passed to procfile().
101  */
102 int
103 grep_tree(char **argv)
104 {
105 	FTS *fts;
106 	FTSENT *p;
107 	int c, fts_flags;
108 	bool ok;
109 
110 	c = fts_flags = 0;
111 
112 	switch(linkbehave) {
113 	case LINK_EXPLICIT:
114 		fts_flags = FTS_COMFOLLOW;
115 		break;
116 	case LINK_SKIP:
117 		fts_flags = FTS_PHYSICAL;
118 		break;
119 	default:
120 		fts_flags = FTS_LOGICAL;
121 
122 	}
123 
124 	fts_flags |= FTS_NOSTAT | FTS_NOCHDIR;
125 
126 	if (!(fts = fts_open(argv, fts_flags, NULL)))
127 		err(2, "fts_open");
128 	while ((p = fts_read(fts)) != NULL) {
129 		switch (p->fts_info) {
130 		case FTS_DNR:
131 			/* FALLTHROUGH */
132 		case FTS_ERR:
133 			file_err = true;
134 			if(!sflag)
135 				warnx("%s: %s", p->fts_path, strerror(p->fts_errno));
136 			break;
137 		case FTS_D:
138 			/* FALLTHROUGH */
139 		case FTS_DP:
140 			if (dexclude || dinclude)
141 				if (!dir_matching(p->fts_name) ||
142 				    !dir_matching(p->fts_path))
143 					fts_set(fts, p, FTS_SKIP);
144 			break;
145 		case FTS_DC:
146 			/* Print a warning for recursive directory loop */
147 			warnx("warning: %s: recursive directory loop",
148 				p->fts_path);
149 			break;
150 		default:
151 			/* Check for file exclusion/inclusion */
152 			ok = true;
153 			if (fexclude || finclude)
154 				ok &= file_matching(p->fts_path);
155 
156 			if (ok)
157 				c += procfile(p->fts_path);
158 			break;
159 		}
160 	}
161 
162 	fts_close(fts);
163 	return (c);
164 }
165 
166 /*
167  * Opens a file and processes it.  Each file is processed line-by-line
168  * passing the lines to procline().
169  */
170 int
171 procfile(const char *fn)
172 {
173 	struct file *f;
174 	struct stat sb;
175 	struct str ln;
176 	mode_t s;
177 	int c, t;
178 
179 	if (mflag && (mcount <= 0))
180 		return (0);
181 
182 	if (strcmp(fn, "-") == 0) {
183 		fn = label != NULL ? label : getstr(1);
184 		f = grep_open(NULL);
185 	} else {
186 		if (!stat(fn, &sb)) {
187 			/* Check if we need to process the file */
188 			s = sb.st_mode & S_IFMT;
189 			if (s == S_IFDIR && dirbehave == DIR_SKIP)
190 				return (0);
191 			if ((s == S_IFIFO || s == S_IFCHR || s == S_IFBLK
192 				|| s == S_IFSOCK) && devbehave == DEV_SKIP)
193 					return (0);
194 		}
195 		f = grep_open(fn);
196 	}
197 	if (f == NULL) {
198 		file_err = true;
199 		if (!sflag)
200 			warn("%s", fn);
201 		return (0);
202 	}
203 
204 	ln.file = grep_malloc(strlen(fn) + 1);
205 	strcpy(ln.file, fn);
206 	ln.line_no = 0;
207 	ln.len = 0;
208 	linesqueued = 0;
209 	tail = 0;
210 	ln.off = -1;
211 
212 	for (c = 0;  c == 0 || !(lflag || qflag); ) {
213 		ln.off += ln.len + 1;
214 		if ((ln.dat = grep_fgetln(f, &ln.len)) == NULL || ln.len == 0) {
215 			if (ln.line_no == 0 && matchall)
216 				exit(0);
217 			else
218 				break;
219 		}
220 		if (ln.len > 0 && ln.dat[ln.len - 1] == '\n')
221 			--ln.len;
222 		ln.line_no++;
223 
224 		/* Return if we need to skip a binary file */
225 		if (f->binary && binbehave == BINFILE_SKIP) {
226 			grep_close(f);
227 			free(ln.file);
228 			free(f);
229 			return (0);
230 		}
231 		/* Process the file line-by-line */
232 		if ((t = procline(&ln, f->binary)) == 0 && Bflag > 0) {
233 			enqueue(&ln);
234 			linesqueued++;
235 		}
236 		c += t;
237 		if (mflag && mcount <= 0)
238 			break;
239 	}
240 	if (Bflag > 0)
241 		clearqueue();
242 	grep_close(f);
243 
244 	if (cflag) {
245 		if (!hflag)
246 			printf("%s:", ln.file);
247 		printf("%u\n", c);
248 	}
249 	if (lflag && !qflag && c != 0)
250 		printf("%s%c", fn, nullflag ? 0 : '\n');
251 	if (Lflag && !qflag && c == 0)
252 		printf("%s%c", fn, nullflag ? 0 : '\n');
253 	if (c && !cflag && !lflag && !Lflag &&
254 	    binbehave == BINFILE_BIN && f->binary && !qflag)
255 		printf(getstr(8), fn);
256 
257 	free(ln.file);
258 	free(f);
259 	return (c);
260 }
261 
262 #define iswword(x)	(iswalnum((x)) || (x) == L'_')
263 
264 /*
265  * Processes a line comparing it with the specified patterns.  Each pattern
266  * is looped to be compared along with the full string, saving each and every
267  * match, which is necessary to colorize the output and to count the
268  * matches.  The matching lines are passed to printline() to display the
269  * appropriate output.
270  */
271 static int
272 procline(struct str *l, int nottext)
273 {
274 	regmatch_t matches[MAX_LINE_MATCHES];
275 	regmatch_t pmatch;
276 	size_t st = 0;
277 	unsigned int i;
278 	int c = 0, m = 0, r = 0;
279 
280 	/* Loop to process the whole line */
281 	while (st <= l->len) {
282 		pmatch.rm_so = st;
283 		pmatch.rm_eo = l->len;
284 
285 		/* Loop to compare with all the patterns */
286 		for (i = 0; i < patterns; i++) {
287 			if (fg_pattern[i].pattern)
288 				r = fastexec(&fg_pattern[i],
289 				    l->dat, 1, &pmatch, eflags);
290 			else
291 				r = regexec(&r_pattern[i], l->dat, 1,
292 				    &pmatch, eflags);
293 			r = (r == 0) ? 0 : REG_NOMATCH;
294 			st = (cflags & REG_NOSUB)
295 				? (size_t)l->len
296 				: (size_t)pmatch.rm_eo;
297 			if (r == REG_NOMATCH)
298 				continue;
299 			/* Check for full match */
300 			if (r == 0 && xflag)
301 				if (pmatch.rm_so != 0 ||
302 				    (size_t)pmatch.rm_eo != l->len)
303 					r = REG_NOMATCH;
304 			/* Check for whole word match */
305 			if (r == 0 && (wflag || fg_pattern[i].word)) {
306 				wint_t wbegin, wend;
307 
308 				wbegin = wend = L' ';
309 				if (pmatch.rm_so != 0 &&
310 				    sscanf(&l->dat[pmatch.rm_so - 1],
311 				    "%lc", &wbegin) != 1)
312 					r = REG_NOMATCH;
313 				else if ((size_t)pmatch.rm_eo !=
314 				    l->len &&
315 				    sscanf(&l->dat[pmatch.rm_eo],
316 				    "%lc", &wend) != 1)
317 					r = REG_NOMATCH;
318 				else if (iswword(wbegin) ||
319 				    iswword(wend))
320 					r = REG_NOMATCH;
321 			}
322 			if (r == 0) {
323 				if (m == 0)
324 					c++;
325 				if (m < MAX_LINE_MATCHES)
326 					matches[m++] = pmatch;
327 				/* matches - skip further patterns */
328 				if ((color == NULL && !oflag) ||
329 				    qflag || lflag)
330 					break;
331 			}
332 		}
333 
334 		if (vflag) {
335 			c = !c;
336 			break;
337 		}
338 
339 		/* One pass if we are not recording matches */
340 		if ((color == NULL && !oflag) || qflag || lflag)
341 			break;
342 
343 		if (st == (size_t)pmatch.rm_so)
344 			break; 	/* No matches */
345 	}
346 
347 
348 	/* Count the matches if we have a match limit */
349 	if (mflag)
350 		mcount -= c;
351 
352 	if (c && binbehave == BINFILE_BIN && nottext)
353 		return (c); /* Binary file */
354 
355 	/* Dealing with the context */
356 	if ((tail || c) && !cflag && !qflag && !lflag && !Lflag) {
357 		if (c) {
358 			if (!first && !prev && !tail && Aflag)
359 				printf("--\n");
360 			tail = Aflag;
361 			if (Bflag > 0) {
362 				if (!first && !prev)
363 					printf("--\n");
364 				printqueue();
365 			}
366 			linesqueued = 0;
367 			printline(l, ':', matches, m);
368 		} else {
369 			printline(l, '-', matches, m);
370 			tail--;
371 		}
372 	}
373 
374 	if (c) {
375 		prev = true;
376 		first = false;
377 	} else
378 		prev = false;
379 
380 	return (c);
381 }
382 
383 /*
384  * Safe malloc() for internal use.
385  */
386 void *
387 grep_malloc(size_t size)
388 {
389 	void *ptr;
390 
391 	if ((ptr = malloc(size)) == NULL)
392 		err(2, "malloc");
393 	return (ptr);
394 }
395 
396 /*
397  * Safe calloc() for internal use.
398  */
399 void *
400 grep_calloc(size_t nmemb, size_t size)
401 {
402 	void *ptr;
403 
404 	if ((ptr = calloc(nmemb, size)) == NULL)
405 		err(2, "calloc");
406 	return (ptr);
407 }
408 
409 /*
410  * Safe realloc() for internal use.
411  */
412 void *
413 grep_realloc(void *ptr, size_t size)
414 {
415 
416 	if ((ptr = realloc(ptr, size)) == NULL)
417 		err(2, "realloc");
418 	return (ptr);
419 }
420 
421 /*
422  * Safe strdup() for internal use.
423  */
424 char *
425 grep_strdup(const char *str)
426 {
427 	char *ret;
428 
429 	if ((ret = strdup(str)) == NULL)
430 		err(2, "strdup");
431 	return (ret);
432 }
433 
434 /*
435  * Prints a matching line according to the command line options.
436  */
437 void
438 printline(struct str *line, int sep, regmatch_t *matches, int m)
439 {
440 	size_t a = 0;
441 	int i, n = 0;
442 
443 	if (!hflag) {
444 		if (!nullflag) {
445 			fputs(line->file, stdout);
446 			++n;
447 		} else {
448 			printf("%s", line->file);
449 			putchar(0);
450 		}
451 	}
452 	if (nflag) {
453 		if (n > 0)
454 			putchar(sep);
455 		printf("%d", line->line_no);
456 		++n;
457 	}
458 	if (bflag) {
459 		if (n > 0)
460 			putchar(sep);
461 		printf("%lld", (long long)line->off);
462 		++n;
463 	}
464 	if (n)
465 		putchar(sep);
466 	/* --color and -o */
467 	if ((oflag || color) && m > 0) {
468 		for (i = 0; i < m; i++) {
469 			if (!oflag)
470 				fwrite(line->dat + a, matches[i].rm_so - a, 1,
471 				    stdout);
472 			if (color)
473 				fprintf(stdout, "\33[%sm\33[K", color);
474 
475 				fwrite(line->dat + matches[i].rm_so,
476 				    matches[i].rm_eo - matches[i].rm_so, 1,
477 				    stdout);
478 			if (color)
479 				fprintf(stdout, "\33[m\33[K");
480 			a = matches[i].rm_eo;
481 			if (oflag)
482 				putchar('\n');
483 		}
484 		if (!oflag) {
485 			if (line->len - a > 0)
486 				fwrite(line->dat + a, line->len - a, 1, stdout);
487 			putchar('\n');
488 		}
489 	} else {
490 		fwrite(line->dat, line->len, 1, stdout);
491 		putchar('\n');
492 	}
493 }
494