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 /* XXX TODO: This should be a chunk, not a line */
69 struct str ln; /* Current line */
70 size_t lnstart; /* Position in line */
71 size_t matchidx; /* Latest match index */
72 int printed; /* Metadata printed? */
73 bool binary; /* Binary file? */
77 * Match printing context
80 long long tail; /* Number of trailing lines to record */
81 int last_outed; /* Number of lines since last output */
82 bool doctx; /* Printing context? */
83 bool printmatch; /* Printing matches? */
84 bool same_file; /* Same file as previously printed? */
87 static void procmatch_match(struct mprintc *mc, struct parsec *pc);
88 static void procmatch_nomatch(struct mprintc *mc, struct parsec *pc);
89 static bool procmatches(struct mprintc *mc, struct parsec *pc, bool matched);
90 #ifdef WITH_INTERNAL_NOSPEC
91 static int litexec(const struct pat *pat, const char *string,
92 size_t nmatch, regmatch_t pmatch[]);
94 static int procline(struct parsec *pc);
95 static void printline(struct parsec *pc, int sep);
96 static void printline_metadata(struct str *line, int sep);
99 file_matching(const char *fname)
101 char *fname_base, *fname_buf;
104 ret = finclude ? false : true;
105 fname_buf = strdup(fname);
106 if (fname_buf == NULL)
108 fname_base = basename(fname_buf);
110 for (unsigned int i = 0; i < fpatterns; ++i) {
111 if (fnmatch(fpattern[i].pat, fname, 0) == 0 ||
112 fnmatch(fpattern[i].pat, fname_base, 0) == 0) {
113 if (fpattern[i].mode == EXCL_PAT) {
125 dir_matching(const char *dname)
129 ret = dinclude ? false : true;
131 for (unsigned int i = 0; i < dpatterns; ++i) {
133 fnmatch(dpattern[i].pat, dname, 0) == 0) {
134 if (dpattern[i].mode == EXCL_PAT)
144 * Processes a directory when a recursive search is performed with
145 * the -R option. Each appropriate file is passed to procfile().
148 grep_tree(char **argv)
154 const char *wd[] = { ".", NULL };
160 fts_flags = FTS_COMFOLLOW;
163 fts_flags = FTS_PHYSICAL;
166 fts_flags = FTS_LOGICAL;
170 fts_flags |= FTS_NOSTAT | FTS_NOCHDIR;
172 fts = fts_open((argv[0] == NULL) ?
173 __DECONST(char * const *, wd) : argv, fts_flags, NULL);
176 while ((p = fts_read(fts)) != NULL) {
177 switch (p->fts_info) {
183 warnx("%s: %s", p->fts_path, strerror(p->fts_errno));
188 if (dexclude || dinclude)
189 if (!dir_matching(p->fts_name) ||
190 !dir_matching(p->fts_path))
191 fts_set(fts, p, FTS_SKIP);
194 /* Print a warning for recursive directory loop */
195 warnx("warning: %s: recursive directory loop",
199 /* Check for file exclusion/inclusion */
201 if (fexclude || finclude)
202 ok &= file_matching(p->fts_path);
205 c += procfile(p->fts_path);
215 procmatch_match(struct mprintc *mc, struct parsec *pc)
219 if (!first_match && (!mc->same_file || mc->last_outed > 0))
226 /* Print the matching line, but only if not quiet/binary */
227 if (mc->printmatch) {
229 while (pc->matchidx >= MAX_MATCHES) {
230 /* Reset matchidx and try again */
232 if (procline(pc) == 0)
238 mc->same_file = true;
244 procmatch_nomatch(struct mprintc *mc, struct parsec *pc)
247 /* Deal with any -A context as needed */
249 grep_printline(&pc->ln, '-');
253 } else if (Bflag == 0 || (Bflag > 0 && enqueue(&pc->ln)))
255 * Enqueue non-matching lines for -B context. If we're not
256 * actually doing -B context or if the enqueue resulted in a
257 * line being rotated out, then go ahead and increment
258 * last_outed to signify a gap between context/match.
264 * Process any matches in the current parsing context, return a boolean
265 * indicating whether we should halt any further processing or not. 'true' to
266 * continue processing, 'false' to halt.
269 procmatches(struct mprintc *mc, struct parsec *pc, bool matched)
273 * XXX TODO: This should loop over pc->matches and handle things on a
274 * line-by-line basis, setting up a `struct str` as needed.
276 /* Deal with any -B context or context separators */
278 procmatch_match(mc, pc);
280 /* Count the matches if we have a match limit */
282 /* XXX TODO: Decrement by number of matched lines */
284 if (mflag && mcount <= 0)
287 } else if (mc->doctx)
288 procmatch_nomatch(mc, pc);
294 * Opens a file and processes it. Each file is processed line-by-line
295 * passing the lines to procline().
298 procfile(const char *fn)
307 if (strcmp(fn, "-") == 0) {
308 fn = label != NULL ? label : getstr(1);
311 if (!stat(fn, &sb)) {
312 /* Check if we need to process the file */
313 s = sb.st_mode & S_IFMT;
314 if (s == S_IFDIR && dirbehave == DIR_SKIP)
316 if ((s == S_IFIFO || s == S_IFCHR || s == S_IFBLK
317 || s == S_IFSOCK) && devbehave == DEV_SKIP)
329 pc.ln.file = grep_malloc(strlen(fn) + 1);
330 strcpy(pc.ln.file, fn);
335 pc.binary = f->binary;
336 memset(&mc, 0, sizeof(mc));
337 mc.printmatch = true;
338 if ((pc.binary && binbehave == BINFILE_BIN) || cflag || qflag ||
340 mc.printmatch = false;
341 if (mc.printmatch && (Aflag != 0 || Bflag != 0))
345 for (c = 0; c == 0 || !(lflag || qflag); ) {
347 * XXX TODO: We need to revisit this in a chunking world. We're
348 * not going to be doing per-line statistics because of the
349 * overhead involved. procmatches can figure that stuff out as
351 /* Reset per-line statistics */
356 pc.ln.off += pc.ln.len + 1;
357 /* XXX TODO: Grab a chunk */
358 if ((pc.ln.dat = grep_fgetln(f, &pc.ln.len)) == NULL ||
362 if (pc.ln.len > 0 && pc.ln.dat[pc.ln.len - 1] == fileeol)
366 /* Return if we need to skip a binary file */
367 if (pc.binary && binbehave == BINFILE_SKIP) {
374 if ((t = procline(&pc)) == 0)
377 /* Halt processing if we hit our match limit */
378 if (!procmatches(&mc, &pc, t == 0))
387 printf("%s:", pc.ln.file);
390 if (lflag && !qflag && c != 0)
391 printf("%s%c", fn, nullflag ? 0 : '\n');
392 if (Lflag && !qflag && c == 0)
393 printf("%s%c", fn, nullflag ? 0 : '\n');
394 if (c && !cflag && !lflag && !Lflag &&
395 binbehave == BINFILE_BIN && f->binary && !qflag)
396 printf(getstr(8), fn);
403 #ifdef WITH_INTERNAL_NOSPEC
405 * Internal implementation of literal string search within a string, modeled
406 * after regexec(3), for use when the regex(3) implementation doesn't offer
407 * either REG_NOSPEC or REG_LITERAL. This does not apply in the default FreeBSD
408 * config, but in other scenarios such as building against libgnuregex or on
409 * some non-FreeBSD OSes.
412 litexec(const struct pat *pat, const char *string, size_t nmatch,
415 char *(*strstr_fn)(const char *, const char *);
418 size_t idx, n, ofs, stringlen;
420 if (cflags & REG_ICASE)
421 strstr_fn = strcasestr;
425 ofs = pmatch[0].rm_so;
426 stringlen = pmatch[0].rm_eo;
427 if (ofs >= stringlen)
428 return (REG_NOMATCH);
429 subject = strndup(string, stringlen);
432 for (n = 0; ofs < stringlen;) {
433 search = (subject + ofs);
434 if ((unsigned long)pat->len > strlen(search))
436 sub = strstr_fn(search, pat->pat);
438 * Ignoring the empty string possibility due to context: grep optimizes
439 * for empty patterns and will never reach this point.
444 /* Fill in pmatch if necessary */
446 pmatch[idx].rm_so = ofs + (sub - search);
447 pmatch[idx].rm_eo = pmatch[idx].rm_so + pat->len;
450 ofs = pmatch[idx].rm_so + 1;
452 /* We only needed to know if we match or not */
456 if (n > 0 && nmatch > 0)
457 for (n = idx; n < nmatch; ++n)
458 pmatch[n].rm_so = pmatch[n].rm_eo = -1;
460 return (n > 0 ? 0 : REG_NOMATCH);
462 #endif /* WITH_INTERNAL_NOSPEC */
464 #define iswword(x) (iswalnum((x)) || (x) == L'_')
467 * Processes a line comparing it with the specified patterns. Each pattern
468 * is looped to be compared along with the full string, saving each and every
469 * match, which is necessary to colorize the output and to count the
470 * matches. The matching lines are passed to printline() to display the
471 * appropriate output.
474 procline(struct parsec *pc)
476 regmatch_t pmatch, lastmatch, chkmatch;
477 wchar_t wbegin, wend;
480 int c = 0, r = 0, lastmatches = 0, leflags = eflags;
481 size_t startm = 0, matchidx;
484 matchidx = pc->matchidx;
486 /* Special case: empty pattern with -w flag, check first character */
487 if (matchall && wflag) {
491 if (sscanf(&pc->ln.dat[0], "%lc", &wend) != 1 || iswword(wend))
500 /* Initialize to avoid a false positive warning from GCC. */
501 lastmatch.rm_so = lastmatch.rm_eo = 0;
503 /* Loop to process the whole line */
504 while (st <= pc->ln.len) {
508 if (st > 0 && pc->ln.dat[st - 1] != fileeol)
509 leflags |= REG_NOTBOL;
510 /* Loop to compare with all the patterns */
511 for (i = 0; i < patterns; i++) {
513 pmatch.rm_eo = pc->ln.len;
514 #ifdef WITH_INTERNAL_NOSPEC
515 if (grepbehave == GREP_FIXED)
516 r = litexec(&pattern[i], pc->ln.dat, 1, &pmatch);
519 #ifndef WITHOUT_FASTMATCH
520 if (fg_pattern[i].pattern)
521 r = fastexec(&fg_pattern[i],
522 pc->ln.dat, 1, &pmatch, leflags);
525 r = regexec(&r_pattern[i], pc->ln.dat, 1,
529 /* Check for full match */
530 if (xflag && (pmatch.rm_so != 0 ||
531 (size_t)pmatch.rm_eo != pc->ln.len))
533 /* Check for whole word match */
534 #ifndef WITHOUT_FASTMATCH
535 if (wflag || fg_pattern[i].word) {
539 wbegin = wend = L' ';
540 if (pmatch.rm_so != 0 &&
541 sscanf(&pc->ln.dat[pmatch.rm_so - 1],
542 "%lc", &wbegin) != 1)
544 else if ((size_t)pmatch.rm_eo !=
546 sscanf(&pc->ln.dat[pmatch.rm_eo],
549 else if (iswword(wbegin) ||
553 * If we're doing whole word matching and we
554 * matched once, then we should try the pattern
555 * again after advancing just past the start of
556 * the earliest match. This allows the pattern
557 * to match later on in the line and possibly
558 * still match a whole word.
560 if (r == REG_NOMATCH &&
561 (retry == pc->lnstart ||
562 (unsigned int)pmatch.rm_so + 1 < retry))
563 retry = pmatch.rm_so + 1;
564 if (r == REG_NOMATCH)
574 * Replace previous match if the new one is earlier
575 * and/or longer. This will lead to some amount of
576 * extra work if -o/--color are specified, but it's
577 * worth it from a correctness point of view.
579 if (matchidx > startm) {
580 chkmatch = pc->matches[matchidx - 1];
581 if (pmatch.rm_so < chkmatch.rm_so ||
582 (pmatch.rm_so == chkmatch.rm_so &&
583 (pmatch.rm_eo - pmatch.rm_so) >
584 (chkmatch.rm_eo - chkmatch.rm_so))) {
585 pc->matches[matchidx - 1] = pmatch;
589 /* Advance as normal if not */
590 pc->matches[matchidx++] = pmatch;
593 /* avoid excessive matching - skip further patterns */
594 if ((color == NULL && !oflag) || qflag || lflag ||
595 matchidx >= MAX_MATCHES) {
603 * Advance to just past the start of the earliest match, try
604 * again just in case we still have a chance to match later in
607 if (lastmatches == 0 && retry > pc->lnstart) {
612 /* XXX TODO: We will need to keep going, since we're chunky */
613 /* One pass if we are not recording matches */
614 if (!wflag && ((color == NULL && !oflag) || qflag || lflag || Lflag))
617 /* If we didn't have any matches or REG_NOSUB set */
618 if (lastmatches == 0 || (cflags & REG_NOSUB))
621 if (lastmatches == 0)
624 else if (st == nst && lastmatch.rm_so == lastmatch.rm_eo)
625 /* Zero-length match -- advance one more so we don't get stuck */
628 /* Advance st based on previous matches */
633 /* Reflect the new matchidx in the context */
634 pc->matchidx = matchidx;
641 * Safe malloc() for internal use.
644 grep_malloc(size_t size)
648 if ((ptr = malloc(size)) == NULL)
654 * Safe calloc() for internal use.
657 grep_calloc(size_t nmemb, size_t size)
661 if ((ptr = calloc(nmemb, size)) == NULL)
667 * Safe realloc() for internal use.
670 grep_realloc(void *ptr, size_t size)
673 if ((ptr = realloc(ptr, size)) == NULL)
679 * Safe strdup() for internal use.
682 grep_strdup(const char *str)
686 if ((ret = strdup(str)) == NULL)
692 * Print an entire line as-is, there are no inline matches to consider. This is
693 * used for printing context.
695 void grep_printline(struct str *line, int sep) {
696 printline_metadata(line, sep);
697 fwrite(line->dat, line->len, 1, stdout);
702 printline_metadata(struct str *line, int sep)
709 fputs(line->file, stdout);
712 printf("%s", line->file);
719 printf("%d", line->line_no);
725 printf("%lld", (long long)(line->off + line->boff));
733 * Prints a matching line according to the command line options.
736 printline(struct parsec *pc, int sep)
742 /* If matchall, everything matches but don't actually print for -o */
743 if (oflag && matchall)
746 matchidx = pc->matchidx;
749 if ((oflag || color) && matchidx > 0) {
750 /* Only print metadata once per line if --color */
751 if (!oflag && pc->printed == 0)
752 printline_metadata(&pc->ln, sep);
753 for (i = 0; i < matchidx; i++) {
754 match = pc->matches[i];
755 /* Don't output zero length matches */
756 if (match.rm_so == match.rm_eo)
759 * Metadata is printed on a per-line basis, so every
760 * match gets file metadata with the -o flag.
763 pc->ln.boff = match.rm_so;
764 printline_metadata(&pc->ln, sep);
766 fwrite(pc->ln.dat + a, match.rm_so - a, 1,
769 fprintf(stdout, "\33[%sm\33[K", color);
770 fwrite(pc->ln.dat + match.rm_so,
771 match.rm_eo - match.rm_so, 1, stdout);
773 fprintf(stdout, "\33[m\33[K");
779 if (pc->ln.len - a > 0)
780 fwrite(pc->ln.dat + a, pc->ln.len - a, 1,
785 grep_printline(&pc->ln, sep);