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 <kevans91@ksu.edu>
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? */
74 static int procline(struct parsec *pc);
75 static void printline(struct parsec *pc, int sep);
76 static void printline_metadata(struct str *line, int sep);
79 file_matching(const char *fname)
81 char *fname_base, *fname_buf;
84 ret = finclude ? false : true;
85 fname_buf = strdup(fname);
86 if (fname_buf == NULL)
88 fname_base = basename(fname_buf);
90 for (unsigned int i = 0; i < fpatterns; ++i) {
91 if (fnmatch(fpattern[i].pat, fname, 0) == 0 ||
92 fnmatch(fpattern[i].pat, fname_base, 0) == 0) {
93 if (fpattern[i].mode == EXCL_PAT) {
105 dir_matching(const char *dname)
109 ret = dinclude ? false : true;
111 for (unsigned int i = 0; i < dpatterns; ++i) {
113 fnmatch(dpattern[i].pat, dname, 0) == 0) {
114 if (dpattern[i].mode == EXCL_PAT)
124 * Processes a directory when a recursive search is performed with
125 * the -R option. Each appropriate file is passed to procfile().
128 grep_tree(char **argv)
134 const char *wd[] = { ".", NULL };
140 fts_flags = FTS_COMFOLLOW;
143 fts_flags = FTS_PHYSICAL;
146 fts_flags = FTS_LOGICAL;
150 fts_flags |= FTS_NOSTAT | FTS_NOCHDIR;
152 fts = fts_open((argv[0] == NULL) ?
153 __DECONST(char * const *, wd) : argv, fts_flags, NULL);
156 while ((p = fts_read(fts)) != NULL) {
157 switch (p->fts_info) {
163 warnx("%s: %s", p->fts_path, strerror(p->fts_errno));
168 if (dexclude || dinclude)
169 if (!dir_matching(p->fts_name) ||
170 !dir_matching(p->fts_path))
171 fts_set(fts, p, FTS_SKIP);
174 /* Print a warning for recursive directory loop */
175 warnx("warning: %s: recursive directory loop",
179 /* Check for file exclusion/inclusion */
181 if (fexclude || finclude)
182 ok &= file_matching(p->fts_path);
185 c += procfile(p->fts_path);
195 * Opens a file and processes it. Each file is processed line-by-line
196 * passing the lines to procline().
199 procfile(const char *fn)
207 int c, last_outed, t;
208 bool doctx, printmatch, same_file;
210 if (strcmp(fn, "-") == 0) {
211 fn = label != NULL ? label : getstr(1);
214 if (!stat(fn, &sb)) {
215 /* Check if we need to process the file */
216 s = sb.st_mode & S_IFMT;
217 if (s == S_IFDIR && dirbehave == DIR_SKIP)
219 if ((s == S_IFIFO || s == S_IFCHR || s == S_IFBLK
220 || s == S_IFSOCK) && devbehave == DEV_SKIP)
234 pc.ln.file = grep_malloc(strlen(fn) + 1);
235 strcpy(pc.ln.file, fn);
240 pc.binary = f->binary;
247 if ((pc.binary && binbehave == BINFILE_BIN) || cflag || qflag ||
250 if (printmatch && (Aflag != 0 || Bflag != 0))
254 for (c = 0; c == 0 || !(lflag || qflag); ) {
255 /* Reset per-line statistics */
260 pc.ln.off += pc.ln.len + 1;
261 if ((pc.ln.dat = grep_fgetln(f, &pc.ln.len)) == NULL ||
263 if (pc.ln.line_no == 0 && matchall)
265 * An empty file with an empty pattern and the
266 * -w flag does not match
268 exit(matchall && wflag ? 1 : 0);
273 if (pc.ln.len > 0 && pc.ln.dat[pc.ln.len - 1] == fileeol)
277 /* Return if we need to skip a binary file */
278 if (pc.binary && binbehave == BINFILE_SKIP) {
285 if ((t = procline(&pc)) == 0)
288 /* Deal with any -B context or context separators */
289 if (t == 0 && doctx) {
290 if (!first_match && (!same_file || last_outed > 0))
296 /* Print the matching line, but only if not quiet/binary */
297 if (t == 0 && printmatch) {
299 while (pc.matchidx >= MAX_MATCHES) {
300 /* Reset matchidx and try again */
302 if (procline(&pc) == 0)
311 if (t != 0 && doctx) {
312 /* Deal with any -A context */
314 grep_printline(&pc.ln, '-');
320 * Enqueue non-matching lines for -B context.
321 * If we're not actually doing -B context or if
322 * the enqueue resulted in a line being rotated
323 * out, then go ahead and increment last_outed
324 * to signify a gap between context/match.
326 if (Bflag == 0 || (Bflag > 0 && enqueue(ln)))
331 /* Count the matches if we have a match limit */
332 if (t == 0 && mflag) {
334 if (mflag && mcount <= 0)
345 printf("%s:", pc.ln.file);
348 if (lflag && !qflag && c != 0)
349 printf("%s%c", fn, nullflag ? 0 : '\n');
350 if (Lflag && !qflag && c == 0)
351 printf("%s%c", fn, nullflag ? 0 : '\n');
352 if (c && !cflag && !lflag && !Lflag &&
353 binbehave == BINFILE_BIN && f->binary && !qflag)
354 printf(getstr(8), fn);
361 #define iswword(x) (iswalnum((x)) || (x) == L'_')
364 * Processes a line comparing it with the specified patterns. Each pattern
365 * is looped to be compared along with the full string, saving each and every
366 * match, which is necessary to colorize the output and to count the
367 * matches. The matching lines are passed to printline() to display the
368 * appropriate output.
371 procline(struct parsec *pc)
373 regmatch_t pmatch, lastmatch, chkmatch;
374 wchar_t wbegin, wend;
377 int c = 0, r = 0, lastmatches = 0, leflags = eflags;
378 size_t startm = 0, matchidx;
381 matchidx = pc->matchidx;
383 /* Special case: empty pattern with -w flag, check first character */
384 if (matchall && wflag) {
388 if (sscanf(&pc->ln.dat[0], "%lc", &wend) != 1 || iswword(wend))
397 /* Initialize to avoid a false positive warning from GCC. */
398 lastmatch.rm_so = lastmatch.rm_eo = 0;
400 /* Loop to process the whole line */
401 while (st <= pc->ln.len) {
405 if (st > 0 && pc->ln.dat[st - 1] != fileeol)
406 leflags |= REG_NOTBOL;
407 /* Loop to compare with all the patterns */
408 for (i = 0; i < patterns; i++) {
410 pmatch.rm_eo = pc->ln.len;
411 #ifndef WITHOUT_FASTMATCH
412 if (fg_pattern[i].pattern)
413 r = fastexec(&fg_pattern[i],
414 pc->ln.dat, 1, &pmatch, leflags);
417 r = regexec(&r_pattern[i], pc->ln.dat, 1,
421 /* Check for full match */
422 if (xflag && (pmatch.rm_so != 0 ||
423 (size_t)pmatch.rm_eo != pc->ln.len))
425 /* Check for whole word match */
426 #ifndef WITHOUT_FASTMATCH
427 if (wflag || fg_pattern[i].word) {
431 wbegin = wend = L' ';
432 if (pmatch.rm_so != 0 &&
433 sscanf(&pc->ln.dat[pmatch.rm_so - 1],
434 "%lc", &wbegin) != 1)
436 else if ((size_t)pmatch.rm_eo !=
438 sscanf(&pc->ln.dat[pmatch.rm_eo],
441 else if (iswword(wbegin) ||
445 * If we're doing whole word matching and we
446 * matched once, then we should try the pattern
447 * again after advancing just past the start of
448 * the earliest match. This allows the pattern
449 * to match later on in the line and possibly
450 * still match a whole word.
452 if (r == REG_NOMATCH &&
453 (retry == pc->lnstart ||
454 pmatch.rm_so + 1 < retry))
455 retry = pmatch.rm_so + 1;
456 if (r == REG_NOMATCH)
466 * Replace previous match if the new one is earlier
467 * and/or longer. This will lead to some amount of
468 * extra work if -o/--color are specified, but it's
469 * worth it from a correctness point of view.
471 if (matchidx > startm) {
472 chkmatch = pc->matches[matchidx - 1];
473 if (pmatch.rm_so < chkmatch.rm_so ||
474 (pmatch.rm_so == chkmatch.rm_so &&
475 (pmatch.rm_eo - pmatch.rm_so) >
476 (chkmatch.rm_eo - chkmatch.rm_so))) {
477 pc->matches[matchidx - 1] = pmatch;
481 /* Advance as normal if not */
482 pc->matches[matchidx++] = pmatch;
485 /* avoid excessive matching - skip further patterns */
486 if ((color == NULL && !oflag) || qflag || lflag ||
487 matchidx >= MAX_MATCHES) {
495 * Advance to just past the start of the earliest match, try
496 * again just in case we still have a chance to match later in
499 if (lastmatches == 0 && retry > pc->lnstart) {
504 /* One pass if we are not recording matches */
505 if (!wflag && ((color == NULL && !oflag) || qflag || lflag || Lflag))
508 /* If we didn't have any matches or REG_NOSUB set */
509 if (lastmatches == 0 || (cflags & REG_NOSUB))
512 if (lastmatches == 0)
515 else if (st == nst && lastmatch.rm_so == lastmatch.rm_eo)
516 /* Zero-length match -- advance one more so we don't get stuck */
519 /* Advance st based on previous matches */
524 /* Reflect the new matchidx in the context */
525 pc->matchidx = matchidx;
532 * Safe malloc() for internal use.
535 grep_malloc(size_t size)
539 if ((ptr = malloc(size)) == NULL)
545 * Safe calloc() for internal use.
548 grep_calloc(size_t nmemb, size_t size)
552 if ((ptr = calloc(nmemb, size)) == NULL)
558 * Safe realloc() for internal use.
561 grep_realloc(void *ptr, size_t size)
564 if ((ptr = realloc(ptr, size)) == NULL)
570 * Safe strdup() for internal use.
573 grep_strdup(const char *str)
577 if ((ret = strdup(str)) == NULL)
583 * Print an entire line as-is, there are no inline matches to consider. This is
584 * used for printing context.
586 void grep_printline(struct str *line, int sep) {
587 printline_metadata(line, sep);
588 fwrite(line->dat, line->len, 1, stdout);
593 printline_metadata(struct str *line, int sep)
600 fputs(line->file, stdout);
603 printf("%s", line->file);
610 printf("%d", line->line_no);
616 printf("%lld", (long long)(line->off + line->boff));
624 * Prints a matching line according to the command line options.
627 printline(struct parsec *pc, int sep)
633 /* If matchall, everything matches but don't actually print for -o */
634 if (oflag && matchall)
637 matchidx = pc->matchidx;
640 if ((oflag || color) && matchidx > 0) {
641 /* Only print metadata once per line if --color */
642 if (!oflag && pc->printed == 0)
643 printline_metadata(&pc->ln, sep);
644 for (i = 0; i < matchidx; i++) {
645 match = pc->matches[i];
646 /* Don't output zero length matches */
647 if (match.rm_so == match.rm_eo)
650 * Metadata is printed on a per-line basis, so every
651 * match gets file metadata with the -o flag.
654 pc->ln.boff = match.rm_so;
655 printline_metadata(&pc->ln, sep);
657 fwrite(pc->ln.dat + a, match.rm_so - a, 1,
660 fprintf(stdout, "\33[%sm\33[K", color);
661 fwrite(pc->ln.dat + match.rm_so,
662 match.rm_eo - match.rm_so, 1, stdout);
664 fprintf(stdout, "\33[m\33[K");
670 if (pc->ln.len - a > 0)
671 fwrite(pc->ln.dat + a, pc->ln.len - a, 1,
676 grep_printline(&pc->ln, sep);