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