1 /* $OpenBSD: diff.c,v 1.67 2019/06/28 13:35:00 deraadt Exp $ */ 2 3 /* 4 * Copyright (c) 2003 Todd C. Miller <Todd.Miller@courtesan.com> 5 * 6 * Permission to use, copy, modify, and distribute this software for any 7 * purpose with or without fee is hereby granted, provided that the above 8 * copyright notice and this permission notice appear in all copies. 9 * 10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES 11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF 12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR 13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES 14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN 15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF 16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. 17 * 18 * Sponsored in part by the Defense Advanced Research Projects 19 * Agency (DARPA) and Air Force Research Laboratory, Air Force 20 * Materiel Command, USAF, under agreement number F39502-99-1-0512. 21 */ 22 23 #include <sys/cdefs.h> 24 __FBSDID("$FreeBSD$"); 25 26 #include <sys/stat.h> 27 28 #include <ctype.h> 29 #include <err.h> 30 #include <errno.h> 31 #include <getopt.h> 32 #include <stdlib.h> 33 #include <stdio.h> 34 #include <string.h> 35 #include <unistd.h> 36 #include <limits.h> 37 38 #include "diff.h" 39 #include "xmalloc.h" 40 41 bool lflag, Nflag, Pflag, rflag, sflag, Tflag, cflag; 42 bool ignore_file_case, suppress_common; 43 int diff_format, diff_context, status; 44 int tabsize = 8, width = 130; 45 char *start, *ifdefname, *diffargs, *label[2], *ignore_pats; 46 char *group_format = NULL; 47 struct stat stb1, stb2; 48 struct excludes *excludes_list; 49 regex_t ignore_re; 50 51 #define OPTIONS "0123456789aBbC:cdD:efHhI:iL:lnNPpqrS:sTtU:uwW:X:x:y" 52 enum { 53 OPT_TSIZE = CHAR_MAX + 1, 54 OPT_STRIPCR, 55 OPT_IGN_FN_CASE, 56 OPT_NO_IGN_FN_CASE, 57 OPT_NORMAL, 58 OPT_HORIZON_LINES, 59 OPT_CHANGED_GROUP_FORMAT, 60 OPT_SUPPRESS_COMMON, 61 }; 62 63 static struct option longopts[] = { 64 { "text", no_argument, 0, 'a' }, 65 { "ignore-space-change", no_argument, 0, 'b' }, 66 { "context", optional_argument, 0, 'C' }, 67 { "ifdef", required_argument, 0, 'D' }, 68 { "minimal", no_argument, 0, 'd' }, 69 { "ed", no_argument, 0, 'e' }, 70 { "forward-ed", no_argument, 0, 'f' }, 71 { "speed-large-files", no_argument, NULL, 'H' }, 72 { "ignore-blank-lines", no_argument, 0, 'B' }, 73 { "ignore-matching-lines", required_argument, 0, 'I' }, 74 { "ignore-case", no_argument, 0, 'i' }, 75 { "paginate", no_argument, NULL, 'l' }, 76 { "label", required_argument, 0, 'L' }, 77 { "new-file", no_argument, 0, 'N' }, 78 { "rcs", no_argument, 0, 'n' }, 79 { "unidirectional-new-file", no_argument, 0, 'P' }, 80 { "show-c-function", no_argument, 0, 'p' }, 81 { "brief", no_argument, 0, 'q' }, 82 { "recursive", no_argument, 0, 'r' }, 83 { "report-identical-files", no_argument, 0, 's' }, 84 { "starting-file", required_argument, 0, 'S' }, 85 { "expand-tabs", no_argument, 0, 't' }, 86 { "initial-tab", no_argument, 0, 'T' }, 87 { "unified", optional_argument, 0, 'U' }, 88 { "ignore-all-space", no_argument, 0, 'w' }, 89 { "width", required_argument, 0, 'W' }, 90 { "exclude", required_argument, 0, 'x' }, 91 { "exclude-from", required_argument, 0, 'X' }, 92 { "side-by-side", no_argument, NULL, 'y' }, 93 { "ignore-file-name-case", no_argument, NULL, OPT_IGN_FN_CASE }, 94 { "horizon-lines", required_argument, NULL, OPT_HORIZON_LINES }, 95 { "no-ignore-file-name-case", no_argument, NULL, OPT_NO_IGN_FN_CASE }, 96 { "normal", no_argument, NULL, OPT_NORMAL }, 97 { "strip-trailing-cr", no_argument, NULL, OPT_STRIPCR }, 98 { "tabsize", required_argument, NULL, OPT_TSIZE }, 99 { "changed-group-format", required_argument, NULL, OPT_CHANGED_GROUP_FORMAT}, 100 { "suppress-common-lines", no_argument, NULL, OPT_SUPPRESS_COMMON }, 101 { NULL, 0, 0, '\0'} 102 }; 103 104 static void usage(void) __dead2; 105 static void conflicting_format(void) __dead2; 106 static void push_excludes(char *); 107 static void push_ignore_pats(char *); 108 static void read_excludes_file(char *file); 109 static void set_argstr(char **, char **); 110 static char *splice(char *, char *); 111 112 int 113 main(int argc, char **argv) 114 { 115 const char *errstr = NULL; 116 char *ep, **oargv; 117 long l; 118 int ch, dflags, lastch, gotstdin, prevoptind, newarg; 119 120 oargv = argv; 121 gotstdin = 0; 122 dflags = 0; 123 lastch = '\0'; 124 prevoptind = 1; 125 newarg = 1; 126 diff_context = 3; 127 diff_format = D_UNSET; 128 #define FORMAT_MISMATCHED(type) \ 129 (diff_format != D_UNSET && diff_format != (type)) 130 while ((ch = getopt_long(argc, argv, OPTIONS, longopts, NULL)) != -1) { 131 switch (ch) { 132 case '0': case '1': case '2': case '3': case '4': 133 case '5': case '6': case '7': case '8': case '9': 134 if (newarg) 135 usage(); /* disallow -[0-9]+ */ 136 else if (lastch == 'c' || lastch == 'u') 137 diff_context = 0; 138 else if (!isdigit(lastch) || diff_context > INT_MAX / 10) 139 usage(); 140 diff_context = (diff_context * 10) + (ch - '0'); 141 break; 142 case 'a': 143 dflags |= D_FORCEASCII; 144 break; 145 case 'b': 146 dflags |= D_FOLDBLANKS; 147 break; 148 case 'C': 149 case 'c': 150 if (FORMAT_MISMATCHED(D_CONTEXT)) 151 conflicting_format(); 152 diff_format = D_CONTEXT; 153 if (optarg != NULL) { 154 l = strtol(optarg, &ep, 10); 155 if (*ep != '\0' || l < 0 || l >= INT_MAX) 156 usage(); 157 diff_context = (int)l; 158 } 159 break; 160 case 'd': 161 dflags |= D_MINIMAL; 162 break; 163 case 'D': 164 if (FORMAT_MISMATCHED(D_IFDEF)) 165 conflicting_format(); 166 diff_format = D_IFDEF; 167 ifdefname = optarg; 168 break; 169 case 'e': 170 if (FORMAT_MISMATCHED(D_EDIT)) 171 conflicting_format(); 172 diff_format = D_EDIT; 173 break; 174 case 'f': 175 if (FORMAT_MISMATCHED(D_REVERSE)) 176 conflicting_format(); 177 diff_format = D_REVERSE; 178 break; 179 case 'H': 180 /* ignore but needed for compatibility with GNU diff */ 181 break; 182 case 'h': 183 /* silently ignore for backwards compatibility */ 184 break; 185 case 'B': 186 dflags |= D_SKIPBLANKLINES; 187 break; 188 case 'I': 189 push_ignore_pats(optarg); 190 break; 191 case 'i': 192 dflags |= D_IGNORECASE; 193 break; 194 case 'L': 195 if (label[0] == NULL) 196 label[0] = optarg; 197 else if (label[1] == NULL) 198 label[1] = optarg; 199 else 200 usage(); 201 break; 202 case 'l': 203 lflag = true; 204 break; 205 case 'N': 206 Nflag = true; 207 break; 208 case 'n': 209 if (FORMAT_MISMATCHED(D_NREVERSE)) 210 conflicting_format(); 211 diff_format = D_NREVERSE; 212 break; 213 case 'p': 214 dflags |= D_PROTOTYPE; 215 break; 216 case 'P': 217 Pflag = true; 218 break; 219 case 'r': 220 rflag = true; 221 break; 222 case 'q': 223 if (FORMAT_MISMATCHED(D_BRIEF)) 224 conflicting_format(); 225 diff_format = D_BRIEF; 226 break; 227 case 'S': 228 start = optarg; 229 break; 230 case 's': 231 sflag = true; 232 break; 233 case 'T': 234 Tflag = true; 235 break; 236 case 't': 237 dflags |= D_EXPANDTABS; 238 break; 239 case 'U': 240 case 'u': 241 if (FORMAT_MISMATCHED(D_UNIFIED)) 242 conflicting_format(); 243 diff_format = D_UNIFIED; 244 if (optarg != NULL) { 245 l = strtol(optarg, &ep, 10); 246 if (*ep != '\0' || l < 0 || l >= INT_MAX) 247 usage(); 248 diff_context = (int)l; 249 } 250 break; 251 case 'w': 252 dflags |= D_IGNOREBLANKS; 253 break; 254 case 'W': 255 width = (int) strtonum(optarg, 1, INT_MAX, &errstr); 256 if (errstr) { 257 warnx("Invalid argument for width"); 258 usage(); 259 } 260 break; 261 case 'X': 262 read_excludes_file(optarg); 263 break; 264 case 'x': 265 push_excludes(optarg); 266 break; 267 case 'y': 268 if (FORMAT_MISMATCHED(D_SIDEBYSIDE)) 269 conflicting_format(); 270 diff_format = D_SIDEBYSIDE; 271 break; 272 case OPT_CHANGED_GROUP_FORMAT: 273 if (FORMAT_MISMATCHED(D_GFORMAT)) 274 conflicting_format(); 275 diff_format = D_GFORMAT; 276 group_format = optarg; 277 break; 278 case OPT_HORIZON_LINES: 279 break; /* XXX TODO for compatibility with GNU diff3 */ 280 case OPT_IGN_FN_CASE: 281 ignore_file_case = true; 282 break; 283 case OPT_NO_IGN_FN_CASE: 284 ignore_file_case = false; 285 break; 286 case OPT_NORMAL: 287 if (FORMAT_MISMATCHED(D_NORMAL)) 288 conflicting_format(); 289 diff_format = D_NORMAL; 290 break; 291 case OPT_TSIZE: 292 tabsize = (int) strtonum(optarg, 1, INT_MAX, &errstr); 293 if (errstr) { 294 warnx("Invalid argument for tabsize"); 295 usage(); 296 } 297 break; 298 case OPT_STRIPCR: 299 dflags |= D_STRIPCR; 300 break; 301 case OPT_SUPPRESS_COMMON: 302 suppress_common = 1; 303 break; 304 default: 305 usage(); 306 break; 307 } 308 lastch = ch; 309 newarg = optind != prevoptind; 310 prevoptind = optind; 311 } 312 if (diff_format == D_UNSET && (dflags & D_PROTOTYPE) != 0) 313 diff_format = D_CONTEXT; 314 if (diff_format == D_UNSET) 315 diff_format = D_NORMAL; 316 argc -= optind; 317 argv += optind; 318 319 #ifdef __OpenBSD__ 320 if (pledge("stdio rpath tmppath", NULL) == -1) 321 err(2, "pledge"); 322 #endif 323 324 /* 325 * Do sanity checks, fill in stb1 and stb2 and call the appropriate 326 * driver routine. Both drivers use the contents of stb1 and stb2. 327 */ 328 if (argc != 2) 329 usage(); 330 if (ignore_pats != NULL) { 331 char buf[BUFSIZ]; 332 int error; 333 334 if ((error = regcomp(&ignore_re, ignore_pats, 335 REG_NEWLINE | REG_EXTENDED)) != 0) { 336 regerror(error, &ignore_re, buf, sizeof(buf)); 337 if (*ignore_pats != '\0') 338 errx(2, "%s: %s", ignore_pats, buf); 339 else 340 errx(2, "%s", buf); 341 } 342 } 343 if (strcmp(argv[0], "-") == 0) { 344 fstat(STDIN_FILENO, &stb1); 345 gotstdin = 1; 346 } else if (stat(argv[0], &stb1) != 0) { 347 if (!Nflag || errno != ENOENT) 348 err(2, "%s", argv[0]); 349 dflags |= D_EMPTY1; 350 memset(&stb1, 0, sizeof(struct stat)); 351 } 352 353 if (strcmp(argv[1], "-") == 0) { 354 fstat(STDIN_FILENO, &stb2); 355 gotstdin = 1; 356 } else if (stat(argv[1], &stb2) != 0) { 357 if (!Nflag || errno != ENOENT) 358 err(2, "%s", argv[1]); 359 dflags |= D_EMPTY2; 360 memset(&stb2, 0, sizeof(stb2)); 361 stb2.st_mode = stb1.st_mode; 362 } 363 364 if (dflags & D_EMPTY1 && dflags & D_EMPTY2){ 365 warn("%s", argv[0]); 366 warn("%s", argv[1]); 367 exit(2); 368 } 369 370 if (stb1.st_mode == 0) 371 stb1.st_mode = stb2.st_mode; 372 373 if (gotstdin && (S_ISDIR(stb1.st_mode) || S_ISDIR(stb2.st_mode))) 374 errx(2, "can't compare - to a directory"); 375 set_argstr(oargv, argv); 376 if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) { 377 if (diff_format == D_IFDEF) 378 errx(2, "-D option not supported with directories"); 379 diffdir(argv[0], argv[1], dflags); 380 } else { 381 if (S_ISDIR(stb1.st_mode)) { 382 argv[0] = splice(argv[0], argv[1]); 383 if (stat(argv[0], &stb1) == -1) 384 err(2, "%s", argv[0]); 385 } 386 if (S_ISDIR(stb2.st_mode)) { 387 argv[1] = splice(argv[1], argv[0]); 388 if (stat(argv[1], &stb2) == -1) 389 err(2, "%s", argv[1]); 390 } 391 print_status(diffreg(argv[0], argv[1], dflags, 1), argv[0], 392 argv[1], ""); 393 } 394 exit(status); 395 } 396 397 static void 398 set_argstr(char **av, char **ave) 399 { 400 size_t argsize; 401 char **ap; 402 403 argsize = 4 + *ave - *av + 1; 404 diffargs = xmalloc(argsize); 405 strlcpy(diffargs, "diff", argsize); 406 for (ap = av + 1; ap < ave; ap++) { 407 if (strcmp(*ap, "--") != 0) { 408 strlcat(diffargs, " ", argsize); 409 strlcat(diffargs, *ap, argsize); 410 } 411 } 412 } 413 414 /* 415 * Read in an excludes file and push each line. 416 */ 417 static void 418 read_excludes_file(char *file) 419 { 420 FILE *fp; 421 char *buf, *pattern; 422 size_t len; 423 424 if (strcmp(file, "-") == 0) 425 fp = stdin; 426 else if ((fp = fopen(file, "r")) == NULL) 427 err(2, "%s", file); 428 while ((buf = fgetln(fp, &len)) != NULL) { 429 if (buf[len - 1] == '\n') 430 len--; 431 if ((pattern = strndup(buf, len)) == NULL) 432 err(2, "xstrndup"); 433 push_excludes(pattern); 434 } 435 if (strcmp(file, "-") != 0) 436 fclose(fp); 437 } 438 439 /* 440 * Push a pattern onto the excludes list. 441 */ 442 static void 443 push_excludes(char *pattern) 444 { 445 struct excludes *entry; 446 447 entry = xmalloc(sizeof(*entry)); 448 entry->pattern = pattern; 449 entry->next = excludes_list; 450 excludes_list = entry; 451 } 452 453 static void 454 push_ignore_pats(char *pattern) 455 { 456 size_t len; 457 458 if (ignore_pats == NULL) 459 ignore_pats = xstrdup(pattern); 460 else { 461 /* old + "|" + new + NUL */ 462 len = strlen(ignore_pats) + strlen(pattern) + 2; 463 ignore_pats = xreallocarray(ignore_pats, 1, len); 464 strlcat(ignore_pats, "|", len); 465 strlcat(ignore_pats, pattern, len); 466 } 467 } 468 469 void 470 print_status(int val, char *path1, char *path2, const char *entry) 471 { 472 if (label[0] != NULL) 473 path1 = label[0]; 474 if (label[1] != NULL) 475 path2 = label[1]; 476 477 switch (val) { 478 case D_BINARY: 479 printf("Binary files %s%s and %s%s differ\n", 480 path1, entry, path2, entry); 481 break; 482 case D_DIFFER: 483 if (diff_format == D_BRIEF) 484 printf("Files %s%s and %s%s differ\n", 485 path1, entry, path2, entry); 486 break; 487 case D_SAME: 488 if (sflag) 489 printf("Files %s%s and %s%s are identical\n", 490 path1, entry, path2, entry); 491 break; 492 case D_MISMATCH1: 493 printf("File %s%s is a directory while file %s%s is a regular file\n", 494 path1, entry, path2, entry); 495 break; 496 case D_MISMATCH2: 497 printf("File %s%s is a regular file while file %s%s is a directory\n", 498 path1, entry, path2, entry); 499 break; 500 case D_SKIPPED1: 501 printf("File %s%s is not a regular file or directory and was skipped\n", 502 path1, entry); 503 break; 504 case D_SKIPPED2: 505 printf("File %s%s is not a regular file or directory and was skipped\n", 506 path2, entry); 507 break; 508 case D_ERROR: 509 break; 510 } 511 } 512 513 static void 514 usage(void) 515 { 516 (void)fprintf(stderr, 517 "usage: diff [-aBbdilpTtw] [-c | -e | -f | -n | -q | -u] [--ignore-case]\n" 518 " [--no-ignore-case] [--normal] [--strip-trailing-cr] [--tabsize]\n" 519 " [-I pattern] [-L label] file1 file2\n" 520 " diff [-aBbdilpTtw] [-I pattern] [-L label] [--ignore-case]\n" 521 " [--no-ignore-case] [--normal] [--strip-trailing-cr] [--tabsize]\n" 522 " -C number file1 file2\n" 523 " diff [-aBbdiltw] [-I pattern] [--ignore-case] [--no-ignore-case]\n" 524 " [--normal] [--strip-trailing-cr] [--tabsize] -D string file1 file2\n" 525 " diff [-aBbdilpTtw] [-I pattern] [-L label] [--ignore-case]\n" 526 " [--no-ignore-case] [--normal] [--tabsize] [--strip-trailing-cr]\n" 527 " -U number file1 file2\n" 528 " diff [-aBbdilNPprsTtw] [-c | -e | -f | -n | -q | -u] [--ignore-case]\n" 529 " [--no-ignore-case] [--normal] [--tabsize] [-I pattern] [-L label]\n" 530 " [-S name] [-X file] [-x pattern] dir1 dir2\n" 531 " diff [-aBbditwW] [--expand-tabs] [--ignore-all-blanks]\n" 532 " [--ignore-blank-lines] [--ignore-case] [--minimal]\n" 533 " [--no-ignore-file-name-case] [--strip-trailing-cr]\n" 534 " [--suppress-common-lines] [--tabsize] [--text] [--width]\n" 535 " -y | --side-by-side file1 file2\n"); 536 537 exit(2); 538 } 539 540 static void 541 conflicting_format(void) 542 { 543 544 fprintf(stderr, "error: conflicting output format options.\n"); 545 usage(); 546 } 547 548 static char * 549 splice(char *dir, char *path) 550 { 551 char *tail, *buf; 552 size_t dirlen; 553 554 dirlen = strlen(dir); 555 while (dirlen != 0 && dir[dirlen - 1] == '/') 556 dirlen--; 557 if ((tail = strrchr(path, '/')) == NULL) 558 tail = path; 559 else 560 tail++; 561 xasprintf(&buf, "%.*s/%s", (int)dirlen, dir, tail); 562 return (buf); 563 } 564