util.c revision 1.10
1 /* $NetBSD: util.c,v 1.10 2011/04/18 03:27:40 joerg Exp $ */ 2 /* $FreeBSD: head/usr.bin/grep/util.c 211496 2010-08-19 09:28:59Z des $ */ 3 /* $OpenBSD: util.c,v 1.39 2010/07/02 22:18:03 tedu Exp $ */ 4 5 /*- 6 * Copyright (c) 1999 James Howard and Dag-Erling Codan Smrgrav 7 * Copyright (C) 2008-2010 Gabor Kovesdan <gabor (at) FreeBSD.org> 8 * All rights reserved. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 #if HAVE_NBTOOL_CONFIG_H 33 #include "nbtool_config.h" 34 #endif 35 36 #include <sys/cdefs.h> 37 __RCSID("$NetBSD: util.c,v 1.10 2011/04/18 03:27:40 joerg Exp $"); 38 39 #include <sys/stat.h> 40 #include <sys/types.h> 41 42 #include <ctype.h> 43 #include <err.h> 44 #include <errno.h> 45 #include <fnmatch.h> 46 #include <fts.h> 47 #include <libgen.h> 48 #include <stdbool.h> 49 #include <stdio.h> 50 #include <stdlib.h> 51 #include <string.h> 52 #include <unistd.h> 53 #include <wchar.h> 54 #include <wctype.h> 55 56 #include "grep.h" 57 58 static int linesqueued; 59 static int procline(struct str *l, int); 60 61 bool 62 file_matching(const char *fname) 63 { 64 char *fname_base, *fname_copy; 65 bool ret; 66 67 ret = finclude ? false : true; 68 fname_copy = grep_strdup(fname); 69 fname_base = basename(fname_copy); 70 71 for (unsigned int i = 0; i < fpatterns; ++i) { 72 if (fnmatch(fpattern[i].pat, fname, 0) == 0 || 73 fnmatch(fpattern[i].pat, fname_base, 0) == 0) { 74 if (fpattern[i].mode == EXCL_PAT) 75 return (false); 76 else 77 ret = true; 78 } 79 } 80 free(fname_copy); 81 return (ret); 82 } 83 84 static inline bool 85 dir_matching(const char *dname) 86 { 87 bool ret; 88 89 ret = dinclude ? false : true; 90 91 for (unsigned int i = 0; i < dpatterns; ++i) { 92 if (dname != NULL && 93 fnmatch(dname, dpattern[i].pat, 0) == 0) { 94 if (dpattern[i].mode == EXCL_PAT) 95 return (false); 96 else 97 ret = true; 98 } 99 } 100 return (ret); 101 } 102 103 /* 104 * Processes a directory when a recursive search is performed with 105 * the -R option. Each appropriate file is passed to procfile(). 106 */ 107 int 108 grep_tree(char **argv) 109 { 110 FTS *fts; 111 FTSENT *p; 112 char *d, *dir = NULL; 113 int c, fts_flags; 114 bool ok; 115 116 c = fts_flags = 0; 117 118 switch(linkbehave) { 119 case LINK_EXPLICIT: 120 fts_flags = FTS_COMFOLLOW; 121 break; 122 case LINK_SKIP: 123 fts_flags = FTS_PHYSICAL; 124 break; 125 default: 126 fts_flags = FTS_LOGICAL; 127 128 } 129 130 fts_flags |= FTS_NOSTAT | FTS_NOCHDIR; 131 132 if (!(fts = fts_open(argv, fts_flags, NULL))) 133 err(2, "fts_open"); 134 while ((p = fts_read(fts)) != NULL) { 135 switch (p->fts_info) { 136 case FTS_DNR: 137 /* FALLTHROUGH */ 138 case FTS_ERR: 139 errx(2, "%s: %s", p->fts_path, strerror(p->fts_errno)); 140 break; 141 case FTS_D: 142 /* FALLTHROUGH */ 143 case FTS_DP: 144 break; 145 case FTS_DC: 146 /* Print a warning for recursive directory loop */ 147 warnx("warning: %s: recursive directory loop", 148 p->fts_path); 149 break; 150 default: 151 /* Check for file exclusion/inclusion */ 152 ok = true; 153 if (dexclude || dinclude) { 154 if ((d = strrchr(p->fts_path, '/')) != NULL) { 155 dir = grep_malloc(sizeof(char) * 156 (d - p->fts_path + 1)); 157 memcpy(dir, p->fts_path, 158 d - p->fts_path); 159 dir[d - p->fts_path] = '\0'; 160 } 161 ok = dir_matching(dir); 162 free(dir); 163 dir = NULL; 164 } 165 if (fexclude || finclude) 166 ok &= file_matching(p->fts_path); 167 168 if (ok) 169 c += procfile(p->fts_path); 170 break; 171 } 172 } 173 174 fts_close(fts); 175 return (c); 176 } 177 178 /* 179 * Opens a file and processes it. Each file is processed line-by-line 180 * passing the lines to procline(). 181 */ 182 int 183 procfile(const char *fn) 184 { 185 struct file *f; 186 struct stat sb; 187 struct str ln; 188 mode_t s; 189 int c, t; 190 191 if (mflag && (mcount <= 0)) 192 return (0); 193 194 if (strcmp(fn, "-") == 0) { 195 fn = label != NULL ? label : getstr(1); 196 f = grep_open(NULL); 197 } else { 198 if (!stat(fn, &sb)) { 199 /* Check if we need to process the file */ 200 s = sb.st_mode & S_IFMT; 201 if (s == S_IFDIR && dirbehave == DIR_SKIP) 202 return (0); 203 if ((s == S_IFIFO || s == S_IFCHR || s == S_IFBLK 204 || s == S_IFSOCK) && devbehave == DEV_SKIP) 205 return (0); 206 } 207 f = grep_open(fn); 208 } 209 if (f == NULL) { 210 if (!sflag) 211 warn("%s", fn); 212 if (errno == ENOENT) 213 notfound = true; 214 return (0); 215 } 216 217 ln.file = grep_malloc(strlen(fn) + 1); 218 strcpy(ln.file, fn); 219 ln.line_no = 0; 220 ln.len = 0; 221 linesqueued = 0; 222 tail = 0; 223 ln.off = -1; 224 225 for (c = 0; c == 0 || !(lflag || qflag); ) { 226 ln.off += ln.len + 1; 227 if ((ln.dat = grep_fgetln(f, &ln.len)) == NULL || ln.len == 0) { 228 if (ln.line_no == 0 && matchall) 229 exit(0); 230 else 231 break; 232 } 233 if (ln.len > 0 && ln.dat[ln.len - 1] == '\n') 234 --ln.len; 235 ln.line_no++; 236 237 /* Return if we need to skip a binary file */ 238 if (f->binary && binbehave == BINFILE_SKIP) { 239 grep_close(f); 240 free(ln.file); 241 free(f); 242 return (0); 243 } 244 /* Process the file line-by-line */ 245 if ((t = procline(&ln, f->binary)) == 0 && Bflag > 0) { 246 enqueue(&ln); 247 linesqueued++; 248 } 249 c += t; 250 251 /* Count the matches if we have a match limit */ 252 if (mflag) { 253 mcount -= t; 254 if (mcount <= 0) 255 break; 256 } 257 } 258 if (Bflag > 0) 259 clearqueue(); 260 grep_close(f); 261 262 if (cflag) { 263 if (!hflag) 264 printf("%s:", ln.file); 265 printf("%u\n", c); 266 } 267 if (lflag && !qflag && c != 0) 268 printf("%s\n", fn); 269 if (Lflag && !qflag && c == 0) 270 printf("%s\n", fn); 271 if (c && !cflag && !lflag && !Lflag && 272 binbehave == BINFILE_BIN && f->binary && !qflag) 273 printf(getstr(8), fn); 274 275 free(ln.file); 276 free(f); 277 return (c); 278 } 279 280 #define iswword(x) (iswalnum((x)) || (x) == L'_') 281 282 /* 283 * Processes a line comparing it with the specified patterns. Each pattern 284 * is looped to be compared along with the full string, saving each and every 285 * match, which is necessary to colorize the output and to count the 286 * matches. The matching lines are passed to printline() to display the 287 * appropriate output. 288 */ 289 static int 290 procline(struct str *l, int nottext) 291 { 292 regmatch_t matches[MAX_LINE_MATCHES]; 293 regmatch_t pmatch; 294 size_t st = 0; 295 unsigned int i; 296 int c = 0, m = 0, r = 0; 297 298 if (!matchall) { 299 /* Loop to process the whole line */ 300 while (st <= l->len) { 301 pmatch.rm_so = st; 302 pmatch.rm_eo = l->len; 303 304 /* Loop to compare with all the patterns */ 305 for (i = 0; i < patterns; i++) { 306 /* 307 * XXX: grep_search() is a workaround for speed up and should be 308 * removed in the future. See fastgrep.c. 309 */ 310 if (fg_pattern[i].pattern) { 311 r = grep_search(&fg_pattern[i], 312 (unsigned char *)l->dat, 313 l->len, &pmatch); 314 r = (r == 0) ? 0 : REG_NOMATCH; 315 st = pmatch.rm_eo; 316 } else { 317 r = regexec(&r_pattern[i], l->dat, 1, 318 &pmatch, eflags); 319 r = (r == 0) ? 0 : REG_NOMATCH; 320 st = pmatch.rm_eo; 321 } 322 if (r == REG_NOMATCH) 323 continue; 324 /* Check for full match */ 325 if (r == 0 && xflag) 326 if (pmatch.rm_so != 0 || 327 (size_t)pmatch.rm_eo != l->len) 328 r = REG_NOMATCH; 329 /* Check for whole word match */ 330 if (r == 0 && fg_pattern[i].word && 331 pmatch.rm_so != 0) { 332 wint_t wbegin, wend; 333 334 wbegin = wend = L' '; 335 if (pmatch.rm_so != 0 && 336 sscanf(&l->dat[pmatch.rm_so - 1], 337 "%lc", &wbegin) != 1) 338 r = REG_NOMATCH; 339 else if ((size_t)pmatch.rm_eo != l->len && 340 sscanf(&l->dat[pmatch.rm_eo], 341 "%lc", &wend) != 1) 342 r = REG_NOMATCH; 343 else if (iswword(wbegin) || iswword(wend)) 344 r = REG_NOMATCH; 345 } 346 if (r == 0) { 347 if (m == 0) 348 c++; 349 if (m < MAX_LINE_MATCHES) 350 matches[m++] = pmatch; 351 /* matches - skip further patterns */ 352 if ((color != NULL && !oflag) || qflag || lflag) 353 break; 354 } 355 } 356 357 if (vflag) { 358 c = !c; 359 break; 360 } 361 /* One pass if we are not recording matches */ 362 if ((color != NULL && !oflag) || qflag || lflag) 363 break; 364 365 if (st == (size_t)pmatch.rm_so) 366 break; /* No matches */ 367 } 368 } else 369 c = !vflag; 370 371 if (c && binbehave == BINFILE_BIN && nottext) 372 return (c); /* Binary file */ 373 374 /* Dealing with the context */ 375 if ((tail || c) && !cflag && !qflag && !lflag && !Lflag) { 376 if (c) { 377 if (!first && !prev && !tail && Aflag) 378 printf("--\n"); 379 tail = Aflag; 380 if (Bflag > 0) { 381 if (!first && !prev) 382 printf("--\n"); 383 printqueue(); 384 } 385 linesqueued = 0; 386 printline(l, ':', matches, m); 387 } else { 388 printline(l, '-', matches, m); 389 tail--; 390 } 391 } 392 393 if (c) { 394 prev = true; 395 first = false; 396 } else 397 prev = false; 398 399 return (c); 400 } 401 402 /* 403 * Safe malloc() for internal use. 404 */ 405 void * 406 grep_malloc(size_t size) 407 { 408 void *ptr; 409 410 if ((ptr = malloc(size)) == NULL) 411 err(2, "malloc"); 412 return (ptr); 413 } 414 415 /* 416 * Safe calloc() for internal use. 417 */ 418 void * 419 grep_calloc(size_t nmemb, size_t size) 420 { 421 void *ptr; 422 423 if ((ptr = calloc(nmemb, size)) == NULL) 424 err(2, "calloc"); 425 return (ptr); 426 } 427 428 /* 429 * Safe realloc() for internal use. 430 */ 431 void * 432 grep_realloc(void *ptr, size_t size) 433 { 434 435 if ((ptr = realloc(ptr, size)) == NULL) 436 err(2, "realloc"); 437 return (ptr); 438 } 439 440 /* 441 * Safe strdup() for internal use. 442 */ 443 char * 444 grep_strdup(const char *str) 445 { 446 char *ret; 447 448 if ((ret = strdup(str)) == NULL) 449 err(2, "strdup"); 450 return (ret); 451 } 452 453 /* 454 * Prints a matching line according to the command line options. 455 */ 456 void 457 printline(struct str *line, int sep, regmatch_t *matches, int m) 458 { 459 size_t a = 0; 460 int i, n = 0; 461 462 if (!hflag) { 463 if (nullflag == 0) 464 fputs(line->file, stdout); 465 else { 466 printf("%s", line->file); 467 putchar(0); 468 } 469 ++n; 470 } 471 if (nflag) { 472 if (n > 0) 473 putchar(sep); 474 printf("%d", line->line_no); 475 ++n; 476 } 477 if (bflag) { 478 if (n > 0) 479 putchar(sep); 480 printf("%lld", (long long)line->off); 481 ++n; 482 } 483 if (n) 484 putchar(sep); 485 /* --color and -o */ 486 if ((oflag || color) && m > 0) { 487 for (i = 0; i < m; i++) { 488 if (!oflag) 489 fwrite(line->dat + a, matches[i].rm_so - a, 1, 490 stdout); 491 if (color) 492 fprintf(stdout, "\33[%sm\33[K", color); 493 494 fwrite(line->dat + matches[i].rm_so, 495 matches[i].rm_eo - matches[i].rm_so, 1, 496 stdout); 497 if (color) 498 fprintf(stdout, "\33[m\33[K"); 499 a = matches[i].rm_eo; 500 if (oflag) 501 putchar('\n'); 502 } 503 if (!oflag) { 504 if (line->len - a > 0) 505 fwrite(line->dat + a, line->len - a, 1, stdout); 506 putchar('\n'); 507 } 508 } else { 509 fwrite(line->dat, line->len, 1, stdout); 510 putchar('\n'); 511 } 512 } 513
Indexes created Tue Sep 23 21:09:50 GMT 2025