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 * Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav
7 * Copyright (C) 2008-2010 Gabor Kovesdan <gabor@FreeBSD.org>
8 * Copyright (C) 2017 Kyle Evans <kevans@FreeBSD.org>
11 * Redistribution and use in source and binary forms, with or without
12 * modification, are permitted provided that the following conditions
14 * 1. Redistributions of source code must retain the above copyright
15 * notice, this list of conditions and the following disclaimer.
16 * 2. Redistributions in binary form must reproduce the above copyright
17 * notice, this list of conditions and the following disclaimer in the
18 * documentation and/or other materials provided with the distribution.
20 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 #include <sys/cdefs.h>
34 __FBSDID("$FreeBSD$");
37 #include <sys/types.h>
53 #ifndef WITHOUT_FASTMATCH
54 #include "fastmatch.h"
58 static bool first_match = true;
61 * Parsing context; used to hold things like matches made and
65 regmatch_t matches[MAX_MATCHES]; /* Matches made */
66 struct str ln; /* Current line */
67 size_t lnstart; /* Position in line */
68 size_t matchidx; /* Latest match index */
69 int printed; /* Metadata printed? */
70 bool binary; /* Binary file? */
73 #ifdef WITH_INTERNAL_NOSPEC
74 static int litexec(const struct pat *pat, const char *string,
75 size_t nmatch, regmatch_t pmatch[]);
77 static int procline(struct parsec *pc);
78 static void printline(struct parsec *pc, int sep);
79 static void printline_metadata(struct str *line, int sep);
82 file_matching(const char *fname)
84 char *fname_base, *fname_buf;
87 ret = finclude ? false : true;
88 fname_buf = strdup(fname);
89 if (fname_buf == NULL)
91 fname_base = basename(fname_buf);
93 for (unsigned int i = 0; i < fpatterns; ++i) {
94 if (fnmatch(fpattern[i].pat, fname, 0) == 0 ||
95 fnmatch(fpattern[i].pat, fname_base, 0) == 0) {
96 if (fpattern[i].mode == EXCL_PAT) {
108 dir_matching(const char *dname)
112 ret = dinclude ? false : true;
114 for (unsigned int i = 0; i < dpatterns; ++i) {
116 fnmatch(dpattern[i].pat, dname, 0) == 0) {
117 if (dpattern[i].mode == EXCL_PAT)
127 * Processes a directory when a recursive search is performed with
128 * the -R option. Each appropriate file is passed to procfile().
131 grep_tree(char **argv)
137 const char *wd[] = { ".", NULL };
143 fts_flags = FTS_COMFOLLOW;
146 fts_flags = FTS_PHYSICAL;
149 fts_flags = FTS_LOGICAL;
153 fts_flags |= FTS_NOSTAT | FTS_NOCHDIR;
155 fts = fts_open((argv[0] == NULL) ?
156 __DECONST(char * const *, wd) : argv, fts_flags, NULL);
159 while ((p = fts_read(fts)) != NULL) {
160 switch (p->fts_info) {
166 warnx("%s: %s", p->fts_path, strerror(p->fts_errno));
171 if (dexclude || dinclude)
172 if (!dir_matching(p->fts_name) ||
173 !dir_matching(p->fts_path))
174 fts_set(fts, p, FTS_SKIP);
177 /* Print a warning for recursive directory loop */
178 warnx("warning: %s: recursive directory loop",
182 /* Check for file exclusion/inclusion */
184 if (fexclude || finclude)
185 ok &= file_matching(p->fts_path);
188 c += procfile(p->fts_path);
198 * Opens a file and processes it. Each file is processed line-by-line
199 * passing the lines to procline().
202 procfile(const char *fn)
210 int c, last_outed, t;
211 bool doctx, printmatch, same_file;
213 if (strcmp(fn, "-") == 0) {
214 fn = label != NULL ? label : getstr(1);
217 if (!stat(fn, &sb)) {
218 /* Check if we need to process the file */
219 s = sb.st_mode & S_IFMT;
220 if (s == S_IFDIR && dirbehave == DIR_SKIP)
222 if ((s == S_IFIFO || s == S_IFCHR || s == S_IFBLK
223 || s == S_IFSOCK) && devbehave == DEV_SKIP)
237 pc.ln.file = grep_malloc(strlen(fn) + 1);
238 strcpy(pc.ln.file, fn);
243 pc.binary = f->binary;
250 if ((pc.binary && binbehave == BINFILE_BIN) || cflag || qflag ||
253 if (printmatch && (Aflag != 0 || Bflag != 0))
257 for (c = 0; c == 0 || !(lflag || qflag); ) {
258 /* Reset per-line statistics */
263 pc.ln.off += pc.ln.len + 1;
264 if ((pc.ln.dat = grep_fgetln(f, &pc.ln.len)) == NULL ||
268 if (pc.ln.len > 0 && pc.ln.dat[pc.ln.len - 1] == fileeol)
272 /* Return if we need to skip a binary file */
273 if (pc.binary && binbehave == BINFILE_SKIP) {
280 if ((t = procline(&pc)) == 0)
283 /* Deal with any -B context or context separators */
284 if (t == 0 && doctx) {
285 if (!first_match && (!same_file || last_outed > 0))
291 /* Print the matching line, but only if not quiet/binary */
292 if (t == 0 && printmatch) {
294 while (pc.matchidx >= MAX_MATCHES) {
295 /* Reset matchidx and try again */
297 if (procline(&pc) == 0)
306 if (t != 0 && doctx) {
307 /* Deal with any -A context */
309 grep_printline(&pc.ln, '-');
315 * Enqueue non-matching lines for -B context.
316 * If we're not actually doing -B context or if
317 * the enqueue resulted in a line being rotated
318 * out, then go ahead and increment last_outed
319 * to signify a gap between context/match.
321 if (Bflag == 0 || (Bflag > 0 && enqueue(ln)))
326 /* Count the matches if we have a match limit */
327 if (t == 0 && mflag) {
329 if (mflag && mcount <= 0)
340 printf("%s:", pc.ln.file);
343 if (lflag && !qflag && c != 0)
344 printf("%s%c", fn, nullflag ? 0 : '\n');
345 if (Lflag && !qflag && c == 0)
346 printf("%s%c", fn, nullflag ? 0 : '\n');
347 if (c && !cflag && !lflag && !Lflag &&
348 binbehave == BINFILE_BIN && f->binary && !qflag)
349 printf(getstr(8), fn);
356 #ifdef WITH_INTERNAL_NOSPEC
358 * Internal implementation of literal string search within a string, modeled
359 * after regexec(3), for use when the regex(3) implementation doesn't offer
360 * either REG_NOSPEC or REG_LITERAL. This does not apply in the default FreeBSD
361 * config, but in other scenarios such as building against libgnuregex or on
362 * some non-FreeBSD OSes.
365 litexec(const struct pat *pat, const char *string, size_t nmatch,
368 char *(*strstr_fn)(const char *, const char *);
371 size_t idx, n, ofs, stringlen;
373 if (cflags & REG_ICASE)
374 strstr_fn = strcasestr;
378 ofs = pmatch[0].rm_so;
379 stringlen = pmatch[0].rm_eo;
380 if (ofs >= stringlen)
381 return (REG_NOMATCH);
382 subject = strndup(string, stringlen);
385 for (n = 0; ofs < stringlen;) {
386 search = (subject + ofs);
387 if ((unsigned long)pat->len > strlen(search))
389 sub = strstr_fn(search, pat->pat);
391 * Ignoring the empty string possibility due to context: grep optimizes
392 * for empty patterns and will never reach this point.
397 /* Fill in pmatch if necessary */
399 pmatch[idx].rm_so = ofs + (sub - search);
400 pmatch[idx].rm_eo = pmatch[idx].rm_so + pat->len;
403 ofs = pmatch[idx].rm_so + 1;
405 /* We only needed to know if we match or not */
409 if (n > 0 && nmatch > 0)
410 for (n = idx; n < nmatch; ++n)
411 pmatch[n].rm_so = pmatch[n].rm_eo = -1;
413 return (n > 0 ? 0 : REG_NOMATCH);
415 #endif /* WITH_INTERNAL_NOSPEC */
417 #define iswword(x) (iswalnum((x)) || (x) == L'_')
420 * Processes a line comparing it with the specified patterns. Each pattern
421 * is looped to be compared along with the full string, saving each and every
422 * match, which is necessary to colorize the output and to count the
423 * matches. The matching lines are passed to printline() to display the
424 * appropriate output.
427 procline(struct parsec *pc)
429 regmatch_t pmatch, lastmatch, chkmatch;
430 wchar_t wbegin, wend;
433 int c = 0, r = 0, lastmatches = 0, leflags = eflags;
434 size_t startm = 0, matchidx;
437 matchidx = pc->matchidx;
439 /* Special case: empty pattern with -w flag, check first character */
440 if (matchall && wflag) {
444 if (sscanf(&pc->ln.dat[0], "%lc", &wend) != 1 || iswword(wend))
453 /* Initialize to avoid a false positive warning from GCC. */
454 lastmatch.rm_so = lastmatch.rm_eo = 0;
456 /* Loop to process the whole line */
457 while (st <= pc->ln.len) {
461 if (st > 0 && pc->ln.dat[st - 1] != fileeol)
462 leflags |= REG_NOTBOL;
463 /* Loop to compare with all the patterns */
464 for (i = 0; i < patterns; i++) {
466 pmatch.rm_eo = pc->ln.len;
467 #ifdef WITH_INTERNAL_NOSPEC
468 if (grepbehave == GREP_FIXED)
469 r = litexec(&pattern[i], pc->ln.dat, 1, &pmatch);
472 #ifndef WITHOUT_FASTMATCH
473 if (fg_pattern[i].pattern)
474 r = fastexec(&fg_pattern[i],
475 pc->ln.dat, 1, &pmatch, leflags);
478 r = regexec(&r_pattern[i], pc->ln.dat, 1,
482 /* Check for full match */
483 if (xflag && (pmatch.rm_so != 0 ||
484 (size_t)pmatch.rm_eo != pc->ln.len))
486 /* Check for whole word match */
487 #ifndef WITHOUT_FASTMATCH
488 if (wflag || fg_pattern[i].word) {
492 wbegin = wend = L' ';
493 if (pmatch.rm_so != 0 &&
494 sscanf(&pc->ln.dat[pmatch.rm_so - 1],
495 "%lc", &wbegin) != 1)
497 else if ((size_t)pmatch.rm_eo !=
499 sscanf(&pc->ln.dat[pmatch.rm_eo],
502 else if (iswword(wbegin) ||
506 * If we're doing whole word matching and we
507 * matched once, then we should try the pattern
508 * again after advancing just past the start of
509 * the earliest match. This allows the pattern
510 * to match later on in the line and possibly
511 * still match a whole word.
513 if (r == REG_NOMATCH &&
514 (retry == pc->lnstart ||
515 (unsigned int)pmatch.rm_so + 1 < retry))
516 retry = pmatch.rm_so + 1;
517 if (r == REG_NOMATCH)
527 * Replace previous match if the new one is earlier
528 * and/or longer. This will lead to some amount of
529 * extra work if -o/--color are specified, but it's
530 * worth it from a correctness point of view.
532 if (matchidx > startm) {
533 chkmatch = pc->matches[matchidx - 1];
534 if (pmatch.rm_so < chkmatch.rm_so ||
535 (pmatch.rm_so == chkmatch.rm_so &&
536 (pmatch.rm_eo - pmatch.rm_so) >
537 (chkmatch.rm_eo - chkmatch.rm_so))) {
538 pc->matches[matchidx - 1] = pmatch;
542 /* Advance as normal if not */
543 pc->matches[matchidx++] = pmatch;
546 /* avoid excessive matching - skip further patterns */
547 if ((color == NULL && !oflag) || qflag || lflag ||
548 matchidx >= MAX_MATCHES) {
556 * Advance to just past the start of the earliest match, try
557 * again just in case we still have a chance to match later in
560 if (lastmatches == 0 && retry > pc->lnstart) {
565 /* One pass if we are not recording matches */
566 if (!wflag && ((color == NULL && !oflag) || qflag || lflag || Lflag))
569 /* If we didn't have any matches or REG_NOSUB set */
570 if (lastmatches == 0 || (cflags & REG_NOSUB))
573 if (lastmatches == 0)
576 else if (st == nst && lastmatch.rm_so == lastmatch.rm_eo)
577 /* Zero-length match -- advance one more so we don't get stuck */
580 /* Advance st based on previous matches */
585 /* Reflect the new matchidx in the context */
586 pc->matchidx = matchidx;
593 * Safe malloc() for internal use.
596 grep_malloc(size_t size)
600 if ((ptr = malloc(size)) == NULL)
606 * Safe calloc() for internal use.
609 grep_calloc(size_t nmemb, size_t size)
613 if ((ptr = calloc(nmemb, size)) == NULL)
619 * Safe realloc() for internal use.
622 grep_realloc(void *ptr, size_t size)
625 if ((ptr = realloc(ptr, size)) == NULL)
631 * Safe strdup() for internal use.
634 grep_strdup(const char *str)
638 if ((ret = strdup(str)) == NULL)
644 * Print an entire line as-is, there are no inline matches to consider. This is
645 * used for printing context.
647 void grep_printline(struct str *line, int sep) {
648 printline_metadata(line, sep);
649 fwrite(line->dat, line->len, 1, stdout);
654 printline_metadata(struct str *line, int sep)
661 fputs(line->file, stdout);
664 printf("%s", line->file);
671 printf("%d", line->line_no);
677 printf("%lld", (long long)(line->off + line->boff));
685 * Prints a matching line according to the command line options.
688 printline(struct parsec *pc, int sep)
694 /* If matchall, everything matches but don't actually print for -o */
695 if (oflag && matchall)
698 matchidx = pc->matchidx;
701 if ((oflag || color) && matchidx > 0) {
702 /* Only print metadata once per line if --color */
703 if (!oflag && pc->printed == 0)
704 printline_metadata(&pc->ln, sep);
705 for (i = 0; i < matchidx; i++) {
706 match = pc->matches[i];
707 /* Don't output zero length matches */
708 if (match.rm_so == match.rm_eo)
711 * Metadata is printed on a per-line basis, so every
712 * match gets file metadata with the -o flag.
715 pc->ln.boff = match.rm_so;
716 printline_metadata(&pc->ln, sep);
718 fwrite(pc->ln.dat + a, match.rm_so - a, 1,
721 fprintf(stdout, "\33[%sm\33[K", color);
722 fwrite(pc->ln.dat + match.rm_so,
723 match.rm_eo - match.rm_so, 1, stdout);
725 fprintf(stdout, "\33[m\33[K");
731 if (pc->ln.len - a > 0)
732 fwrite(pc->ln.dat + a, pc->ln.len - a, 1,
737 grep_printline(&pc->ln, sep);