1 /* $NetBSD: util.c,v 1.9 2011/02/27 17:33:37 joerg Exp $ */
3 /* $OpenBSD: util.c,v 1.39 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) 2008-2010 Gabor Kovesdan <gabor@FreeBSD.org>
10 * Copyright (C) 2017 Kyle Evans <kevans@FreeBSD.org>
11 * All rights reserved.
13 * Redistribution and use in source and binary forms, with or without
14 * modification, are permitted provided that the following conditions
16 * 1. Redistributions of source code must retain the above copyright
17 * notice, this list of conditions and the following disclaimer.
18 * 2. Redistributions in binary form must reproduce the above copyright
19 * notice, this list of conditions and the following disclaimer in the
20 * documentation and/or other materials provided with the distribution.
22 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 #include <sys/cdefs.h>
36 __FBSDID("$FreeBSD$");
39 #include <sys/types.h>
55 #ifndef WITHOUT_FASTMATCH
56 #include "fastmatch.h"
60 static bool first_match = true;
63 * Parsing context; used to hold things like matches made and
67 regmatch_t matches[MAX_MATCHES]; /* Matches made */
68 struct str ln; /* Current line */
69 size_t lnstart; /* Position in line */
70 size_t matchidx; /* Latest match index */
71 int printed; /* Metadata printed? */
72 bool binary; /* Binary file? */
75 #ifdef WITH_INTERNAL_NOSPEC
76 static int litexec(const struct pat *pat, const char *string,
77 size_t nmatch, regmatch_t pmatch[]);
79 static int procline(struct parsec *pc);
80 static void printline(struct parsec *pc, int sep);
81 static void printline_metadata(struct str *line, int sep);
84 file_matching(const char *fname)
86 char *fname_base, *fname_buf;
89 ret = finclude ? false : true;
90 fname_buf = strdup(fname);
91 if (fname_buf == NULL)
93 fname_base = basename(fname_buf);
95 for (unsigned int i = 0; i < fpatterns; ++i) {
96 if (fnmatch(fpattern[i].pat, fname, 0) == 0 ||
97 fnmatch(fpattern[i].pat, fname_base, 0) == 0) {
98 if (fpattern[i].mode == EXCL_PAT) {
110 dir_matching(const char *dname)
114 ret = dinclude ? false : true;
116 for (unsigned int i = 0; i < dpatterns; ++i) {
118 fnmatch(dpattern[i].pat, dname, 0) == 0) {
119 if (dpattern[i].mode == EXCL_PAT)
129 * Processes a directory when a recursive search is performed with
130 * the -R option. Each appropriate file is passed to procfile().
133 grep_tree(char **argv)
139 const char *wd[] = { ".", NULL };
145 fts_flags = FTS_COMFOLLOW;
148 fts_flags = FTS_PHYSICAL;
151 fts_flags = FTS_LOGICAL;
155 fts_flags |= FTS_NOSTAT | FTS_NOCHDIR;
157 fts = fts_open((argv[0] == NULL) ?
158 __DECONST(char * const *, wd) : argv, fts_flags, NULL);
161 while ((p = fts_read(fts)) != NULL) {
162 switch (p->fts_info) {
168 warnx("%s: %s", p->fts_path, strerror(p->fts_errno));
173 if (dexclude || dinclude)
174 if (!dir_matching(p->fts_name) ||
175 !dir_matching(p->fts_path))
176 fts_set(fts, p, FTS_SKIP);
179 /* Print a warning for recursive directory loop */
180 warnx("warning: %s: recursive directory loop",
184 /* Check for file exclusion/inclusion */
186 if (fexclude || finclude)
187 ok &= file_matching(p->fts_path);
190 c += procfile(p->fts_path);
200 * Opens a file and processes it. Each file is processed line-by-line
201 * passing the lines to procline().
204 procfile(const char *fn)
212 int c, last_outed, t;
213 bool doctx, printmatch, same_file;
215 if (strcmp(fn, "-") == 0) {
216 fn = label != NULL ? label : getstr(1);
219 if (!stat(fn, &sb)) {
220 /* Check if we need to process the file */
221 s = sb.st_mode & S_IFMT;
222 if (s == S_IFDIR && dirbehave == DIR_SKIP)
224 if ((s == S_IFIFO || s == S_IFCHR || s == S_IFBLK
225 || s == S_IFSOCK) && devbehave == DEV_SKIP)
239 pc.ln.file = grep_malloc(strlen(fn) + 1);
240 strcpy(pc.ln.file, fn);
245 pc.binary = f->binary;
252 if ((pc.binary && binbehave == BINFILE_BIN) || cflag || qflag ||
255 if (printmatch && (Aflag != 0 || Bflag != 0))
259 for (c = 0; c == 0 || !(lflag || qflag); ) {
260 /* Reset per-line statistics */
265 pc.ln.off += pc.ln.len + 1;
266 if ((pc.ln.dat = grep_fgetln(f, &pc.ln.len)) == NULL ||
270 if (pc.ln.len > 0 && pc.ln.dat[pc.ln.len - 1] == fileeol)
274 /* Return if we need to skip a binary file */
275 if (pc.binary && binbehave == BINFILE_SKIP) {
282 if ((t = procline(&pc)) == 0)
285 /* Deal with any -B context or context separators */
286 if (t == 0 && doctx) {
287 if (!first_match && (!same_file || last_outed > 0))
293 /* Print the matching line, but only if not quiet/binary */
294 if (t == 0 && printmatch) {
296 while (pc.matchidx >= MAX_MATCHES) {
297 /* Reset matchidx and try again */
299 if (procline(&pc) == 0)
308 if (t != 0 && doctx) {
309 /* Deal with any -A context */
311 grep_printline(&pc.ln, '-');
317 * Enqueue non-matching lines for -B context.
318 * If we're not actually doing -B context or if
319 * the enqueue resulted in a line being rotated
320 * out, then go ahead and increment last_outed
321 * to signify a gap between context/match.
323 if (Bflag == 0 || (Bflag > 0 && enqueue(ln)))
328 /* Count the matches if we have a match limit */
329 if (t == 0 && mflag) {
331 if (mflag && mcount <= 0)
342 printf("%s:", pc.ln.file);
345 if (lflag && !qflag && c != 0)
346 printf("%s%c", fn, nullflag ? 0 : '\n');
347 if (Lflag && !qflag && c == 0)
348 printf("%s%c", fn, nullflag ? 0 : '\n');
349 if (c && !cflag && !lflag && !Lflag &&
350 binbehave == BINFILE_BIN && f->binary && !qflag)
351 printf(getstr(8), fn);
358 #ifdef WITH_INTERNAL_NOSPEC
360 * Internal implementation of literal string search within a string, modeled
361 * after regexec(3), for use when the regex(3) implementation doesn't offer
362 * either REG_NOSPEC or REG_LITERAL. This does not apply in the default FreeBSD
363 * config, but in other scenarios such as building against libgnuregex or on
364 * some non-FreeBSD OSes.
367 litexec(const struct pat *pat, const char *string, size_t nmatch,
370 char *(*strstr_fn)(const char *, const char *);
373 size_t idx, n, ofs, stringlen;
375 if (cflags & REG_ICASE)
376 strstr_fn = strcasestr;
380 ofs = pmatch[0].rm_so;
381 stringlen = pmatch[0].rm_eo;
382 if (ofs >= stringlen)
383 return (REG_NOMATCH);
384 subject = strndup(string, stringlen);
387 for (n = 0; ofs < stringlen;) {
388 search = (subject + ofs);
389 if ((unsigned long)pat->len > strlen(search))
391 sub = strstr_fn(search, pat->pat);
393 * Ignoring the empty string possibility due to context: grep optimizes
394 * for empty patterns and will never reach this point.
399 /* Fill in pmatch if necessary */
401 pmatch[idx].rm_so = ofs + (sub - search);
402 pmatch[idx].rm_eo = pmatch[idx].rm_so + pat->len;
405 ofs = pmatch[idx].rm_so + 1;
407 /* We only needed to know if we match or not */
411 if (n > 0 && nmatch > 0)
412 for (n = idx; n < nmatch; ++n)
413 pmatch[n].rm_so = pmatch[n].rm_eo = -1;
415 return (n > 0 ? 0 : REG_NOMATCH);
417 #endif /* WITH_INTERNAL_NOSPEC */
419 #define iswword(x) (iswalnum((x)) || (x) == L'_')
422 * Processes a line comparing it with the specified patterns. Each pattern
423 * is looped to be compared along with the full string, saving each and every
424 * match, which is necessary to colorize the output and to count the
425 * matches. The matching lines are passed to printline() to display the
426 * appropriate output.
429 procline(struct parsec *pc)
431 regmatch_t pmatch, lastmatch, chkmatch;
432 wchar_t wbegin, wend;
435 int c = 0, r = 0, lastmatches = 0, leflags = eflags;
436 size_t startm = 0, matchidx;
439 matchidx = pc->matchidx;
441 /* Special case: empty pattern with -w flag, check first character */
442 if (matchall && wflag) {
446 if (sscanf(&pc->ln.dat[0], "%lc", &wend) != 1 || iswword(wend))
455 /* Initialize to avoid a false positive warning from GCC. */
456 lastmatch.rm_so = lastmatch.rm_eo = 0;
458 /* Loop to process the whole line */
459 while (st <= pc->ln.len) {
463 if (st > 0 && pc->ln.dat[st - 1] != fileeol)
464 leflags |= REG_NOTBOL;
465 /* Loop to compare with all the patterns */
466 for (i = 0; i < patterns; i++) {
468 pmatch.rm_eo = pc->ln.len;
469 #ifdef WITH_INTERNAL_NOSPEC
470 if (grepbehave == GREP_FIXED)
471 r = litexec(&pattern[i], pc->ln.dat, 1, &pmatch);
474 #ifndef WITHOUT_FASTMATCH
475 if (fg_pattern[i].pattern)
476 r = fastexec(&fg_pattern[i],
477 pc->ln.dat, 1, &pmatch, leflags);
480 r = regexec(&r_pattern[i], pc->ln.dat, 1,
484 /* Check for full match */
485 if (xflag && (pmatch.rm_so != 0 ||
486 (size_t)pmatch.rm_eo != pc->ln.len))
488 /* Check for whole word match */
489 #ifndef WITHOUT_FASTMATCH
490 if (wflag || fg_pattern[i].word) {
494 wbegin = wend = L' ';
495 if (pmatch.rm_so != 0 &&
496 sscanf(&pc->ln.dat[pmatch.rm_so - 1],
497 "%lc", &wbegin) != 1)
499 else if ((size_t)pmatch.rm_eo !=
501 sscanf(&pc->ln.dat[pmatch.rm_eo],
504 else if (iswword(wbegin) ||
508 * If we're doing whole word matching and we
509 * matched once, then we should try the pattern
510 * again after advancing just past the start of
511 * the earliest match. This allows the pattern
512 * to match later on in the line and possibly
513 * still match a whole word.
515 if (r == REG_NOMATCH &&
516 (retry == pc->lnstart ||
517 (unsigned int)pmatch.rm_so + 1 < retry))
518 retry = pmatch.rm_so + 1;
519 if (r == REG_NOMATCH)
529 * Replace previous match if the new one is earlier
530 * and/or longer. This will lead to some amount of
531 * extra work if -o/--color are specified, but it's
532 * worth it from a correctness point of view.
534 if (matchidx > startm) {
535 chkmatch = pc->matches[matchidx - 1];
536 if (pmatch.rm_so < chkmatch.rm_so ||
537 (pmatch.rm_so == chkmatch.rm_so &&
538 (pmatch.rm_eo - pmatch.rm_so) >
539 (chkmatch.rm_eo - chkmatch.rm_so))) {
540 pc->matches[matchidx - 1] = pmatch;
544 /* Advance as normal if not */
545 pc->matches[matchidx++] = pmatch;
548 /* avoid excessive matching - skip further patterns */
549 if ((color == NULL && !oflag) || qflag || lflag ||
550 matchidx >= MAX_MATCHES) {
558 * Advance to just past the start of the earliest match, try
559 * again just in case we still have a chance to match later in
562 if (lastmatches == 0 && retry > pc->lnstart) {
567 /* One pass if we are not recording matches */
568 if (!wflag && ((color == NULL && !oflag) || qflag || lflag || Lflag))
571 /* If we didn't have any matches or REG_NOSUB set */
572 if (lastmatches == 0 || (cflags & REG_NOSUB))
575 if (lastmatches == 0)
578 else if (st == nst && lastmatch.rm_so == lastmatch.rm_eo)
579 /* Zero-length match -- advance one more so we don't get stuck */
582 /* Advance st based on previous matches */
587 /* Reflect the new matchidx in the context */
588 pc->matchidx = matchidx;
595 * Safe malloc() for internal use.
598 grep_malloc(size_t size)
602 if ((ptr = malloc(size)) == NULL)
608 * Safe calloc() for internal use.
611 grep_calloc(size_t nmemb, size_t size)
615 if ((ptr = calloc(nmemb, size)) == NULL)
621 * Safe realloc() for internal use.
624 grep_realloc(void *ptr, size_t size)
627 if ((ptr = realloc(ptr, size)) == NULL)
633 * Safe strdup() for internal use.
636 grep_strdup(const char *str)
640 if ((ret = strdup(str)) == NULL)
646 * Print an entire line as-is, there are no inline matches to consider. This is
647 * used for printing context.
649 void grep_printline(struct str *line, int sep) {
650 printline_metadata(line, sep);
651 fwrite(line->dat, line->len, 1, stdout);
656 printline_metadata(struct str *line, int sep)
663 fputs(line->file, stdout);
666 printf("%s", line->file);
673 printf("%d", line->line_no);
679 printf("%lld", (long long)(line->off + line->boff));
687 * Prints a matching line according to the command line options.
690 printline(struct parsec *pc, int sep)
696 /* If matchall, everything matches but don't actually print for -o */
697 if (oflag && matchall)
700 matchidx = pc->matchidx;
703 if ((oflag || color) && matchidx > 0) {
704 /* Only print metadata once per line if --color */
705 if (!oflag && pc->printed == 0)
706 printline_metadata(&pc->ln, sep);
707 for (i = 0; i < matchidx; i++) {
708 match = pc->matches[i];
709 /* Don't output zero length matches */
710 if (match.rm_so == match.rm_eo)
713 * Metadata is printed on a per-line basis, so every
714 * match gets file metadata with the -o flag.
717 pc->ln.boff = match.rm_so;
718 printline_metadata(&pc->ln, sep);
720 fwrite(pc->ln.dat + a, match.rm_so - a, 1,
723 fprintf(stdout, "\33[%sm\33[K", color);
724 fwrite(pc->ln.dat + match.rm_so,
725 match.rm_eo - match.rm_so, 1, stdout);
727 fprintf(stdout, "\33[m\33[K");
733 if (pc->ln.len - a > 0)
734 fwrite(pc->ln.dat + a, pc->ln.len - a, 1,
739 grep_printline(&pc->ln, sep);