util.c revision 223009 
1/*	$NetBSD: util.c,v 1.9 2011/02/27 17:33:37 joerg Exp $	*/
2/*	$FreeBSD: head/usr.bin/grep/util.c 223009 2011-06-12 12:51:58Z gabor $	*/
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 Co��dan Sm��rgrav
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: head/usr.bin/grep/util.c 223009 2011-06-12 12:51:58Z 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 <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 "grep.h"
53
54static int	 linesqueued;
55static int	 procline(struct str *l, int);
56
57bool
58file_matching(const char *fname)
59{
60	char *fname_base;
61	bool ret;
62
63	ret = finclude ? false : true;
64	fname_base = basename(fname);
65
66	for (unsigned int i = 0; i < fpatterns; ++i) {
67		if (fnmatch(fpattern[i].pat, fname, 0) == 0 ||
68		    fnmatch(fpattern[i].pat, fname_base, 0) == 0) {
69			if (fpattern[i].mode == EXCL_PAT)
70				return (false);
71			else
72				ret = true;
73		}
74	}
75	return (ret);
76}
77
78static inline bool
79dir_matching(const char *dname)
80{
81	bool ret;
82
83	ret = dinclude ? false : true;
84
85	for (unsigned int i = 0; i < dpatterns; ++i) {
86		if (dname != NULL &&
87		    fnmatch(dname, dpattern[i].pat, 0) == 0) {
88			if (dpattern[i].mode == EXCL_PAT)
89				return (false);
90			else
91				ret = true;
92		}
93	}
94	return (ret);
95}
96
97/*
98 * Processes a directory when a recursive search is performed with
99 * the -R option.  Each appropriate file is passed to procfile().
100 */
101int
102grep_tree(char **argv)
103{
104	FTS *fts;
105	FTSENT *p;
106	char *d, *dir = NULL;
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			errx(2, "%s: %s", p->fts_path, strerror(p->fts_errno));
134			break;
135		case FTS_D:
136			/* FALLTHROUGH */
137		case FTS_DP:
138			break;
139		case FTS_DC:
140			/* Print a warning for recursive directory loop */
141			warnx("warning: %s: recursive directory loop",
142				p->fts_path);
143			break;
144		default:
145			/* Check for file exclusion/inclusion */
146			ok = true;
147			if (dexclude || dinclude) {
148				if ((d = strrchr(p->fts_path, '/')) != NULL) {
149					dir = grep_malloc(sizeof(char) *
150					    (d - p->fts_path + 1));
151					memcpy(dir, p->fts_path,
152					    d - p->fts_path);
153					dir[d - p->fts_path] = '\0';
154				}
155				ok = dir_matching(dir);
156				free(dir);
157				dir = NULL;
158			}
159			if (fexclude || finclude)
160				ok &= file_matching(p->fts_path);
161
162			if (ok)
163				c += procfile(p->fts_path);
164			break;
165		}
166	}
167
168	fts_close(fts);
169	return (c);
170}
171
172/*
173 * Opens a file and processes it.  Each file is processed line-by-line
174 * passing the lines to procline().
175 */
176int
177procfile(const char *fn)
178{
179	struct file *f;
180	struct stat sb;
181	struct str ln;
182	mode_t s;
183	int c, t;
184
185	if (mflag && (mcount <= 0))
186		return (0);
187
188	if (strcmp(fn, "-") == 0) {
189		fn = label != NULL ? label : getstr(1);
190		f = grep_open(NULL);
191	} else {
192		if (!stat(fn, &sb)) {
193			/* Check if we need to process the file */
194			s = sb.st_mode & S_IFMT;
195			if (s == S_IFDIR && dirbehave == DIR_SKIP)
196				return (0);
197			if ((s == S_IFIFO || s == S_IFCHR || s == S_IFBLK
198				|| s == S_IFSOCK) && devbehave == DEV_SKIP)
199					return (0);
200		}
201		f = grep_open(fn);
202	}
203	if (f == NULL) {
204		if (!sflag)
205			warn("%s", fn);
206		if (errno == ENOENT)
207			notfound = true;
208		return (0);
209	}
210
211	ln.file = grep_malloc(strlen(fn) + 1);
212	strcpy(ln.file, fn);
213	ln.line_no = 0;
214	ln.len = 0;
215	linesqueued = 0;
216	tail = 0;
217	ln.off = -1;
218
219	for (c = 0;  c == 0 || !(lflag || qflag); ) {
220		ln.off += ln.len + 1;
221		if ((ln.dat = grep_fgetln(f, &ln.len)) == NULL || ln.len == 0) {
222			if (ln.line_no == 0 && matchall)
223				exit(0);
224			else
225				break;
226		}
227		if (ln.len > 0 && ln.dat[ln.len - 1] == '\n')
228			--ln.len;
229		ln.line_no++;
230
231		/* Return if we need to skip a binary file */
232		if (f->binary && binbehave == BINFILE_SKIP) {
233			grep_close(f);
234			free(ln.file);
235			free(f);
236			return (0);
237		}
238		/* Process the file line-by-line */
239		if ((t = procline(&ln, f->binary)) == 0 && Bflag > 0) {
240			enqueue(&ln);
241			linesqueued++;
242		}
243		c += t;
244
245		/* Count the matches if we have a match limit */
246		if (mflag) {
247			mcount -= t;
248			if (mcount <= 0)
249				break;
250		}
251	}
252	if (Bflag > 0)
253		clearqueue();
254	grep_close(f);
255
256	if (cflag) {
257		if (!hflag)
258			printf("%s:", ln.file);
259		printf("%u\n", c);
260	}
261	if (lflag && !qflag && c != 0)
262		printf("%s\n", fn);
263	if (Lflag && !qflag && c == 0)
264		printf("%s\n", fn);
265	if (c && !cflag && !lflag && !Lflag &&
266	    binbehave == BINFILE_BIN && f->binary && !qflag)
267		printf(getstr(8), fn);
268
269	free(ln.file);
270	free(f);
271	return (c);
272}
273
274#define iswword(x)	(iswalnum((x)) || (x) == L'_')
275
276/*
277 * Processes a line comparing it with the specified patterns.  Each pattern
278 * is looped to be compared along with the full string, saving each and every
279 * match, which is necessary to colorize the output and to count the
280 * matches.  The matching lines are passed to printline() to display the
281 * appropriate output.
282 */
283static int
284procline(struct str *l, int nottext)
285{
286	regmatch_t matches[MAX_LINE_MATCHES];
287	regmatch_t pmatch;
288	size_t st = 0;
289	unsigned int i;
290	int c = 0, m = 0, r = 0;
291
292	if (!matchall) {
293		/* Loop to process the whole line */
294		while (st <= l->len) {
295			pmatch.rm_so = st;
296			pmatch.rm_eo = l->len;
297
298			/* Loop to compare with all the patterns */
299			for (i = 0; i < patterns; i++) {
300/*
301 * XXX: grep_search() is a workaround for speed up and should be
302 * removed in the future.  See fastgrep.c.
303 */
304				if (fg_pattern[i].pattern)
305					r = grep_search(&fg_pattern[i],
306					    (unsigned char *)l->dat,
307					    l->len, &pmatch);
308				else
309					r = regexec(&r_pattern[i], l->dat, 1,
310					    &pmatch, eflags);
311				r = (r == 0) ? 0 : REG_NOMATCH;
312				st = (cflags & REG_NOSUB)
313					? (size_t)l->len
314					: (size_t)pmatch.rm_eo;
315				if (r == REG_NOMATCH)
316					continue;
317				/* Check for full match */
318				if (r == 0 && xflag)
319					if (pmatch.rm_so != 0 ||
320					    (size_t)pmatch.rm_eo != l->len)
321						r = REG_NOMATCH;
322				/* Check for whole word match */
323				if (r == 0 && (wflag || fg_pattern[i].word)) {
324					wint_t wbegin, wend;
325
326					wbegin = wend = L' ';
327					if (pmatch.rm_so != 0 &&
328					    sscanf(&l->dat[pmatch.rm_so - 1],
329					    "%lc", &wbegin) != 1)
330						r = REG_NOMATCH;
331					else if ((size_t)pmatch.rm_eo !=
332					    l->len &&
333					    sscanf(&l->dat[pmatch.rm_eo],
334					    "%lc", &wend) != 1)
335						r = REG_NOMATCH;
336					else if (iswword(wbegin) ||
337					    iswword(wend))
338						r = REG_NOMATCH;
339				}
340				if (r == 0) {
341					if (m == 0)
342						c++;
343					if (m < MAX_LINE_MATCHES)
344						matches[m++] = pmatch;
345					/* matches - skip further patterns */
346					if ((color == NULL && !oflag) ||
347					    qflag || lflag)
348						break;
349				}
350			}
351
352			if (vflag) {
353				c = !c;
354				break;
355			}
356			/* One pass if we are not recording matches */
357			if ((color == NULL && !oflag) || qflag || lflag)
358				break;
359
360			if (st == (size_t)pmatch.rm_so)
361				break; 	/* No matches */
362		}
363	} else
364		c = !vflag;
365
366	if (c && binbehave == BINFILE_BIN && nottext)
367		return (c); /* Binary file */
368
369	/* Dealing with the context */
370	if ((tail || c) && !cflag && !qflag && !lflag && !Lflag) {
371		if (c) {
372			if (!first && !prev && !tail && Aflag)
373				printf("--\n");
374			tail = Aflag;
375			if (Bflag > 0) {
376				if (!first && !prev)
377					printf("--\n");
378				printqueue();
379			}
380			linesqueued = 0;
381			printline(l, ':', matches, m);
382		} else {
383			printline(l, '-', matches, m);
384			tail--;
385		}
386	}
387
388	if (c) {
389		prev = true;
390		first = false;
391	} else
392		prev = false;
393
394	return (c);
395}
396
397/*
398 * Safe malloc() for internal use.
399 */
400void *
401grep_malloc(size_t size)
402{
403	void *ptr;
404
405	if ((ptr = malloc(size)) == NULL)
406		err(2, "malloc");
407	return (ptr);
408}
409
410/*
411 * Safe calloc() for internal use.
412 */
413void *
414grep_calloc(size_t nmemb, size_t size)
415{
416	void *ptr;
417
418	if ((ptr = calloc(nmemb, size)) == NULL)
419		err(2, "calloc");
420	return (ptr);
421}
422
423/*
424 * Safe realloc() for internal use.
425 */
426void *
427grep_realloc(void *ptr, size_t size)
428{
429
430	if ((ptr = realloc(ptr, size)) == NULL)
431		err(2, "realloc");
432	return (ptr);
433}
434
435/*
436 * Safe strdup() for internal use.
437 */
438char *
439grep_strdup(const char *str)
440{
441	char *ret;
442
443	if ((ret = strdup(str)) == NULL)
444		err(2, "strdup");
445	return (ret);
446}
447
448/*
449 * Prints a matching line according to the command line options.
450 */
451void
452printline(struct str *line, int sep, regmatch_t *matches, int m)
453{
454	size_t a = 0;
455	int i, n = 0;
456
457	if (!hflag) {
458		if (nullflag == 0)
459			fputs(line->file, stdout);
460		else {
461			printf("%s", line->file);
462			putchar(0);
463		}
464		++n;
465	}
466	if (nflag) {
467		if (n > 0)
468			putchar(sep);
469		printf("%d", line->line_no);
470		++n;
471	}
472	if (bflag) {
473		if (n > 0)
474			putchar(sep);
475		printf("%lld", (long long)line->off);
476		++n;
477	}
478	if (n)
479		putchar(sep);
480	/* --color and -o */
481	if ((oflag || color) && m > 0) {
482		for (i = 0; i < m; i++) {
483			if (!oflag)
484				fwrite(line->dat + a, matches[i].rm_so - a, 1,
485				    stdout);
486			if (color)
487				fprintf(stdout, "\33[%sm\33[K", color);
488
489				fwrite(line->dat + matches[i].rm_so,
490				    matches[i].rm_eo - matches[i].rm_so, 1,
491				    stdout);
492			if (color)
493				fprintf(stdout, "\33[m\33[K");
494			a = matches[i].rm_eo;
495			if (oflag)
496				putchar('\n');
497		}
498		if (!oflag) {
499			if (line->len - a > 0)
500				fwrite(line->dat + a, line->len - a, 1, stdout);
501			putchar('\n');
502		}
503	} else {
504		fwrite(line->dat, line->len, 1, stdout);
505		putchar('\n');
506	}
507}
508