Annotation of src/usr.bin/grep/grep.c, Revision 1.49
1.49 ! tedu 1: /* $OpenBSD: grep.c,v 1.48 2014/12/01 06:36:04 deraadt Exp $ */
1.6 deraadt 2:
1.1 deraadt 3: /*-
1.3 deraadt 4: * Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav
5: * All rights reserved.
1.1 deraadt 6: *
7: * Redistribution and use in source and binary forms, with or without
8: * modification, are permitted provided that the following conditions
9: * are met:
10: * 1. Redistributions of source code must retain the above copyright
11: * notice, this list of conditions and the following disclaimer.
12: * 2. Redistributions in binary form must reproduce the above copyright
13: * notice, this list of conditions and the following disclaimer in the
14: * documentation and/or other materials provided with the distribution.
15: *
16: * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19: * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26: * SUCH DAMAGE.
27: */
28:
1.3 deraadt 29: #include <sys/types.h>
1.9 tedu 30: #include <sys/limits.h>
1.3 deraadt 31: #include <sys/stat.h>
1.33 jaredy 32: #include <sys/queue.h>
1.1 deraadt 33:
1.18 millert 34: #include <ctype.h>
1.3 deraadt 35: #include <err.h>
36: #include <errno.h>
37: #include <getopt.h>
38: #include <regex.h>
1.1 deraadt 39: #include <stdio.h>
40: #include <stdlib.h>
1.3 deraadt 41: #include <string.h>
1.1 deraadt 42: #include <unistd.h>
43:
1.3 deraadt 44: #include "grep.h"
45:
46: /* Flags passed to regcomp() and regexec() */
47: int cflags;
48: int eflags = REG_STARTEND;
49:
50: int matchall; /* shortcut */
51: int patterns, pattern_sz;
52: char **pattern;
53: regex_t *r_pattern;
1.13 tedu 54: fastgrep_t *fg_pattern;
1.3 deraadt 55:
56: /* For regex errors */
57: char re_error[RE_ERROR_BUF + 1];
58:
59: /* Command-line flags */
60: int Aflag; /* -A x: print x lines trailing each match */
61: int Bflag; /* -B x: print x lines leading each match */
62: int Eflag; /* -E: interpret pattern as extended regexp */
63: int Fflag; /* -F: interpret pattern as list of fixed strings */
1.43 tedu 64: int Hflag; /* -H: always print filename header */
1.3 deraadt 65: int Lflag; /* -L: only show names of files with no matches */
66: int Rflag; /* -R: recursively search directory trees */
1.5 deraadt 67: #ifndef NOZ
1.3 deraadt 68: int Zflag; /* -Z: decompress input before processing */
1.5 deraadt 69: #endif
1.3 deraadt 70: int bflag; /* -b: show block numbers for each match */
71: int cflag; /* -c: only show a count of matching lines */
72: int hflag; /* -h: don't print filename headers */
73: int iflag; /* -i: ignore case */
74: int lflag; /* -l: only show names of files with matches */
75: int nflag; /* -n: show line numbers in front of matching lines */
1.44 tedu 76: int oflag; /* -o: print each match */
1.3 deraadt 77: int qflag; /* -q: quiet mode (don't output anything) */
78: int sflag; /* -s: silent mode (ignore errors) */
79: int vflag; /* -v: only show non-matching lines */
80: int wflag; /* -w: pattern must start and end on word boundaries */
81: int xflag; /* -x: pattern must match entire line */
1.27 otto 82: int lbflag; /* --line-buffered */
1.3 deraadt 83:
1.9 tedu 84: int binbehave = BIN_FILE_BIN;
85:
86: enum {
87: BIN_OPT = CHAR_MAX + 1,
88: HELP_OPT,
1.27 otto 89: MMAP_OPT,
90: LINEBUF_OPT
1.9 tedu 91: };
92:
1.3 deraadt 93: /* Housekeeping */
1.10 tedu 94: int first; /* flag whether or not this is our first match */
1.3 deraadt 95: int tail; /* lines left to print */
1.45 millert 96: int file_err; /* file reading error */
1.3 deraadt 97:
1.33 jaredy 98: struct patfile {
99: const char *pf_file;
100: SLIST_ENTRY(patfile) pf_next;
101: };
102: SLIST_HEAD(, patfile) patfilelh;
103:
1.8 tedu 104: extern char *__progname;
1.3 deraadt 105:
106: static void
107: usage(void)
108: {
1.5 deraadt 109: fprintf(stderr,
110: #ifdef NOZ
1.44 tedu 111: "usage: %s [-abcEFGHhIiLlnoqRsUVvwx] [-A num] [-B num] [-C[num]]\n"
1.5 deraadt 112: #else
1.44 tedu 113: "usage: %s [-abcEFGHhIiLlnoqRsUVvwxZ] [-A num] [-B num] [-C[num]]\n"
1.5 deraadt 114: #endif
1.30 jmc 115: "\t[-e pattern] [-f file] [--binary-files=value] [--context[=num]]\n"
116: "\t[--line-buffered] [pattern] [file ...]\n", __progname);
1.3 deraadt 117: exit(2);
118: }
119:
1.5 deraadt 120: #ifdef NOZ
1.46 millert 121: static const char optstr[] = "0123456789A:B:CEFGHILRUVabce:f:hilnoqrsuvwxy";
1.5 deraadt 122: #else
1.46 millert 123: static const char optstr[] = "0123456789A:B:CEFGHILRUVZabce:f:hilnoqrsuvwxy";
1.5 deraadt 124: #endif
1.3 deraadt 125:
1.46 millert 126: static const struct option long_options[] =
1.3 deraadt 127: {
1.19 deraadt 128: {"binary-files", required_argument, NULL, BIN_OPT},
129: {"help", no_argument, NULL, HELP_OPT},
130: {"mmap", no_argument, NULL, MMAP_OPT},
1.27 otto 131: {"line-buffered", no_argument, NULL, LINEBUF_OPT},
1.19 deraadt 132: {"after-context", required_argument, NULL, 'A'},
133: {"before-context", required_argument, NULL, 'B'},
134: {"context", optional_argument, NULL, 'C'},
135: {"devices", required_argument, NULL, 'D'},
136: {"extended-regexp", no_argument, NULL, 'E'},
137: {"fixed-strings", no_argument, NULL, 'F'},
138: {"basic-regexp", no_argument, NULL, 'G'},
1.43 tedu 139: {"with-filename", no_argument, NULL, 'H'},
1.19 deraadt 140: {"binary", no_argument, NULL, 'U'},
141: {"version", no_argument, NULL, 'V'},
142: {"text", no_argument, NULL, 'a'},
143: {"byte-offset", no_argument, NULL, 'b'},
144: {"count", no_argument, NULL, 'c'},
145: {"regexp", required_argument, NULL, 'e'},
146: {"file", required_argument, NULL, 'f'},
147: {"no-filename", no_argument, NULL, 'h'},
148: {"ignore-case", no_argument, NULL, 'i'},
149: {"files-without-match", no_argument, NULL, 'L'},
150: {"files-with-matches", no_argument, NULL, 'l'},
151: {"line-number", no_argument, NULL, 'n'},
152: {"quiet", no_argument, NULL, 'q'},
153: {"silent", no_argument, NULL, 'q'},
154: {"recursive", no_argument, NULL, 'r'},
155: {"no-messages", no_argument, NULL, 's'},
156: {"revert-match", no_argument, NULL, 'v'},
157: {"word-regexp", no_argument, NULL, 'w'},
158: {"line-regexp", no_argument, NULL, 'x'},
159: {"unix-byte-offsets", no_argument, NULL, 'u'},
1.5 deraadt 160: #ifndef NOZ
1.19 deraadt 161: {"decompress", no_argument, NULL, 'Z'},
1.5 deraadt 162: #endif
1.19 deraadt 163: {NULL, no_argument, NULL, 0}
1.3 deraadt 164: };
165:
1.1 deraadt 166:
1.3 deraadt 167: static void
168: add_pattern(char *pat, size_t len)
169: {
1.36 jaredy 170: if (!xflag && (len == 0 || matchall)) {
1.3 deraadt 171: matchall = 1;
172: return;
173: }
174: if (patterns == pattern_sz) {
175: pattern_sz *= 2;
1.48 deraadt 176: pattern = grep_reallocarray(pattern, ++pattern_sz, sizeof(*pattern));
1.3 deraadt 177: }
1.37 ray 178: if (len > 0 && pat[len - 1] == '\n')
1.3 deraadt 179: --len;
1.15 tedu 180: /* pat may not be NUL-terminated */
1.28 millert 181: if (wflag && !Fflag) {
1.32 otto 182: int bol = 0, eol = 0, extra;
1.26 millert 183: if (pat[0] == '^')
184: bol = 1;
1.36 jaredy 185: if (len > 0 && pat[len - 1] == '$')
1.26 millert 186: eol = 1;
1.32 otto 187: extra = Eflag ? 2 : 4;
188: pattern[patterns] = grep_malloc(len + 15 + extra);
189: snprintf(pattern[patterns], len + 15 + extra,
190: "%s[[:<:]]%s%.*s%s[[:>:]]%s",
191: bol ? "^" : "",
192: Eflag ? "(" : "\\(",
193: (int)len - bol - eol, pat + bol,
194: Eflag ? ")" : "\\)",
1.26 millert 195: eol ? "$" : "");
1.32 otto 196: len += 14 + extra;
1.25 millert 197: } else {
198: pattern[patterns] = grep_malloc(len + 1);
199: memcpy(pattern[patterns], pat, len);
200: pattern[patterns][len] = '\0';
201: }
1.3 deraadt 202: ++patterns;
203: }
1.1 deraadt 204:
1.3 deraadt 205: static void
1.35 otto 206: add_patterns(char *pats)
207: {
208: char *nl;
209:
210: while ((nl = strchr(pats, '\n')) != NULL) {
211: add_pattern(pats, nl - pats);
212: pats = nl + 1;
213: }
214: add_pattern(pats, strlen(pats));
215: }
216:
217: static void
1.33 jaredy 218: read_patterns(const char *fn)
1.3 deraadt 219: {
220: FILE *f;
221: char *line;
222: size_t len;
223:
224: if ((f = fopen(fn, "r")) == NULL)
1.20 millert 225: err(2, "%s", fn);
1.38 kili 226: while ((line = fgetln(f, &len)) != NULL)
227: add_pattern(line, *line == '\n' ? 0 : len);
1.3 deraadt 228: if (ferror(f))
1.20 millert 229: err(2, "%s", fn);
1.3 deraadt 230: fclose(f);
231: }
1.1 deraadt 232:
233: int
1.3 deraadt 234: main(int argc, char *argv[])
1.1 deraadt 235: {
1.41 jacekm 236: int c, lastc, prevoptind, newarg, i, needpattern, exprs, expr_sz;
1.33 jaredy 237: struct patfile *patfile, *pf_next;
1.21 millert 238: long l;
1.41 jacekm 239: char *ep, **expr;
1.3 deraadt 240:
1.33 jaredy 241: SLIST_INIT(&patfilelh);
1.16 tedu 242: switch (__progname[0]) {
243: case 'e':
1.49 ! tedu 244: Eflag = 1;
1.16 tedu 245: break;
246: case 'f':
1.49 ! tedu 247: Fflag = 1;
1.16 tedu 248: break;
249: #ifndef NOZ
250: case 'z':
1.49 ! tedu 251: Zflag = 1;
1.16 tedu 252: switch(__progname[1]) {
253: case 'e':
1.49 ! tedu 254: Eflag = 1;
1.16 tedu 255: break;
256: case 'f':
1.49 ! tedu 257: Fflag = 1;
1.16 tedu 258: break;
259: }
260: break;
261: #endif
262: }
263:
1.18 millert 264: lastc = '\0';
1.23 millert 265: newarg = 1;
266: prevoptind = 1;
1.38 kili 267: needpattern = 1;
1.41 jacekm 268: expr_sz = exprs = 0;
269: expr = NULL;
1.5 deraadt 270: while ((c = getopt_long(argc, argv, optstr,
1.17 millert 271: long_options, NULL)) != -1) {
1.3 deraadt 272: switch (c) {
273: case '0': case '1': case '2': case '3': case '4':
274: case '5': case '6': case '7': case '8': case '9':
1.23 millert 275: if (newarg || !isdigit(lastc))
276: Aflag = 0;
277: else if (Aflag > INT_MAX / 10)
278: errx(2, "context out of range");
279: Aflag = Bflag = (Aflag * 10) + (c - '0');
1.3 deraadt 280: break;
281: case 'A':
282: case 'B':
1.21 millert 283: l = strtol(optarg, &ep, 10);
284: if (ep == optarg || *ep != '\0' ||
285: l <= 0 || l >= INT_MAX)
286: errx(2, "context out of range");
287: if (c == 'A')
288: Aflag = (int)l;
289: else
290: Bflag = (int)l;
1.3 deraadt 291: break;
292: case 'C':
1.5 deraadt 293: if (optarg == NULL)
1.3 deraadt 294: Aflag = Bflag = 2;
1.21 millert 295: else {
296: l = strtol(optarg, &ep, 10);
297: if (ep == optarg || *ep != '\0' ||
298: l <= 0 || l >= INT_MAX)
299: errx(2, "context out of range");
300: Aflag = Bflag = (int)l;
301: }
1.1 deraadt 302: break;
303: case 'E':
1.49 ! tedu 304: Fflag = 0;
! 305: Eflag = 1;
1.1 deraadt 306: break;
307: case 'F':
1.49 ! tedu 308: Eflag = 0;
! 309: Fflag = 1;
1.3 deraadt 310: break;
311: case 'G':
1.16 tedu 312: Eflag = Fflag = 0;
1.1 deraadt 313: break;
1.43 tedu 314: case 'H':
1.49 ! tedu 315: Hflag = 1;
1.43 tedu 316: break;
1.9 tedu 317: case 'I':
318: binbehave = BIN_FILE_SKIP;
319: break;
1.1 deraadt 320: case 'L':
1.3 deraadt 321: lflag = 0;
322: Lflag = qflag = 1;
1.1 deraadt 323: break;
324: case 'R':
1.3 deraadt 325: case 'r':
1.49 ! tedu 326: Rflag = 1;
1.3 deraadt 327: break;
328: case 'U':
1.9 tedu 329: binbehave = BIN_FILE_BIN;
1.3 deraadt 330: break;
331: case 'V':
332: fprintf(stderr, "grep version %u.%u\n", VER_MAJ, VER_MIN);
1.9 tedu 333: exit(0);
1.1 deraadt 334: break;
1.5 deraadt 335: #ifndef NOZ
1.3 deraadt 336: case 'Z':
1.49 ! tedu 337: Zflag = 1;
1.1 deraadt 338: break;
1.5 deraadt 339: #endif
1.1 deraadt 340: case 'a':
1.9 tedu 341: binbehave = BIN_FILE_TEXT;
1.1 deraadt 342: break;
343: case 'b':
1.3 deraadt 344: bflag = 1;
1.1 deraadt 345: break;
346: case 'c':
1.3 deraadt 347: cflag = 1;
1.1 deraadt 348: break;
349: case 'e':
1.41 jacekm 350: /* defer adding of expressions until all arguments are parsed */
351: if (exprs == expr_sz) {
352: expr_sz *= 2;
1.48 deraadt 353: expr = grep_reallocarray(expr, ++expr_sz,
354: sizeof(*expr));
1.41 jacekm 355: }
1.38 kili 356: needpattern = 0;
1.41 jacekm 357: expr[exprs] = optarg;
358: ++exprs;
1.1 deraadt 359: break;
360: case 'f':
1.33 jaredy 361: patfile = grep_malloc(sizeof(*patfile));
362: patfile->pf_file = optarg;
363: SLIST_INSERT_HEAD(&patfilelh, patfile, pf_next);
1.38 kili 364: needpattern = 0;
1.1 deraadt 365: break;
366: case 'h':
1.3 deraadt 367: hflag = 1;
1.1 deraadt 368: break;
369: case 'i':
1.3 deraadt 370: case 'y':
1.13 tedu 371: iflag = 1;
1.1 deraadt 372: cflags |= REG_ICASE;
373: break;
374: case 'l':
1.3 deraadt 375: Lflag = 0;
376: lflag = qflag = 1;
1.1 deraadt 377: break;
378: case 'n':
1.3 deraadt 379: nflag = 1;
1.44 tedu 380: break;
381: case 'o':
382: oflag = 1;
1.3 deraadt 383: break;
1.1 deraadt 384: case 'q':
1.3 deraadt 385: qflag = 1;
1.1 deraadt 386: break;
387: case 's':
1.3 deraadt 388: sflag = 1;
1.1 deraadt 389: break;
390: case 'v':
1.3 deraadt 391: vflag = 1;
1.1 deraadt 392: break;
393: case 'w':
1.3 deraadt 394: wflag = 1;
1.1 deraadt 395: break;
396: case 'x':
1.3 deraadt 397: xflag = 1;
1.1 deraadt 398: break;
1.9 tedu 399: case BIN_OPT:
400: if (strcmp("binary", optarg) == 0)
401: binbehave = BIN_FILE_BIN;
402: else if (strcmp("without-match", optarg) == 0)
403: binbehave = BIN_FILE_SKIP;
404: else if (strcmp("text", optarg) == 0)
405: binbehave = BIN_FILE_TEXT;
406: else
407: errx(2, "Unknown binary-files option");
408: break;
409: case 'u':
410: case MMAP_OPT:
411: /* default, compatibility */
412: break;
1.27 otto 413: case LINEBUF_OPT:
414: lbflag = 1;
415: break;
1.9 tedu 416: case HELP_OPT:
1.1 deraadt 417: default:
418: usage();
419: }
1.18 millert 420: lastc = c;
1.23 millert 421: newarg = optind != prevoptind;
1.18 millert 422: prevoptind = optind;
1.1 deraadt 423: }
1.3 deraadt 424: argc -= optind;
425: argv += optind;
1.41 jacekm 426:
427: for (i = 0; i < exprs; i++)
428: add_patterns(expr[i]);
429: free(expr);
430: expr = NULL;
1.33 jaredy 431:
432: for (patfile = SLIST_FIRST(&patfilelh); patfile != NULL;
433: patfile = pf_next) {
434: pf_next = SLIST_NEXT(patfile, pf_next);
435: read_patterns(patfile->pf_file);
436: free(patfile);
437: }
1.3 deraadt 438:
1.38 kili 439: if (argc == 0 && needpattern)
1.1 deraadt 440: usage();
441:
1.38 kili 442: if (argc != 0 && needpattern) {
1.35 otto 443: add_patterns(*argv);
1.3 deraadt 444: --argc;
445: ++argv;
446: }
1.5 deraadt 447:
1.16 tedu 448: if (Eflag)
449: cflags |= REG_EXTENDED;
1.42 tedu 450: if (Fflag)
451: cflags |= REG_NOSPEC;
452: #ifdef SMALL
453: /* Sorry, this won't work */
454: if (Fflag && wflag)
455: errx(1, "Can't use small fgrep with -w");
456: #endif
1.39 deraadt 457: fg_pattern = grep_calloc(patterns, sizeof(*fg_pattern));
458: r_pattern = grep_calloc(patterns, sizeof(*r_pattern));
1.3 deraadt 459: for (i = 0; i < patterns; ++i) {
1.28 millert 460: /* Check if cheating is allowed (always is for fgrep). */
1.42 tedu 461: #ifndef SMALL
1.28 millert 462: if (Fflag) {
463: fgrepcomp(&fg_pattern[i], pattern[i]);
1.42 tedu 464: } else
465: #endif
466: {
1.28 millert 467: if (fastcomp(&fg_pattern[i], pattern[i])) {
468: /* Fall back to full regex library */
469: c = regcomp(&r_pattern[i], pattern[i], cflags);
470: if (c != 0) {
471: regerror(c, &r_pattern[i], re_error,
472: RE_ERROR_BUF);
473: errx(2, "%s", re_error);
474: }
1.13 tedu 475: }
1.1 deraadt 476: }
477: }
1.29 deraadt 478:
1.27 otto 479: if (lbflag)
1.47 millert 480: setvbuf(stdout, NULL, _IOLBF, 0);
1.1 deraadt 481:
1.43 tedu 482: if ((argc == 0 || argc == 1) && !Rflag && !Hflag)
1.3 deraadt 483: hflag = 1;
1.1 deraadt 484:
1.3 deraadt 485: if (argc == 0)
486: exit(!procfile(NULL));
1.5 deraadt 487:
1.3 deraadt 488: if (Rflag)
489: c = grep_tree(argv);
1.1 deraadt 490: else
1.3 deraadt 491: for (c = 0; argc--; ++argv)
492: c += procfile(*argv);
1.1 deraadt 493:
1.45 millert 494: exit(c ? (file_err ? (qflag ? 0 : 2) : 0) : (file_err ? 2 : 1));
1.1 deraadt 495: }