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 static const char diff_version[] = "FreeBSD diff 20220309"; 42 bool lflag, Nflag, Pflag, rflag, sflag, Tflag, cflag; 43 bool ignore_file_case, suppress_common, color, noderef; 44 static bool help = false; 45 int diff_format, diff_context, status; 46 int tabsize = 8, width = 130; 47 static int colorflag = COLORFLAG_NEVER; 48 char *start, *ifdefname, *diffargs, *label[2]; 49 char *ignore_pats, *most_recent_pat; 50 char *group_format = NULL; 51 const char *add_code, *del_code; 52 struct stat stb1, stb2; 53 struct excludes *excludes_list; 54 regex_t ignore_re, most_recent_re; 55 56 #define OPTIONS "0123456789aBbC:cdD:efF:HhI:iL:lnNPpqrS:sTtU:uwW:X:x:y" 57 enum { 58 OPT_TSIZE = CHAR_MAX + 1, 59 OPT_STRIPCR, 60 OPT_IGN_FN_CASE, 61 OPT_NO_IGN_FN_CASE, 62 OPT_NORMAL, 63 OPT_HELP, 64 OPT_HORIZON_LINES, 65 OPT_CHANGED_GROUP_FORMAT, 66 OPT_SUPPRESS_COMMON, 67 OPT_COLOR, 68 OPT_NO_DEREFERENCE, 69 OPT_VERSION, 70 }; 71 72 static struct option longopts[] = { 73 { "text", no_argument, 0, 'a' }, 74 { "ignore-space-change", no_argument, 0, 'b' }, 75 { "context", optional_argument, 0, 'C' }, 76 { "ifdef", required_argument, 0, 'D' }, 77 { "minimal", no_argument, 0, 'd' }, 78 { "ed", no_argument, 0, 'e' }, 79 { "forward-ed", no_argument, 0, 'f' }, 80 { "show-function-line", required_argument, 0, 'F' }, 81 { "speed-large-files", no_argument, NULL, 'H' }, 82 { "ignore-blank-lines", no_argument, 0, 'B' }, 83 { "ignore-matching-lines", required_argument, 0, 'I' }, 84 { "ignore-case", no_argument, 0, 'i' }, 85 { "paginate", no_argument, NULL, 'l' }, 86 { "label", required_argument, 0, 'L' }, 87 { "new-file", no_argument, 0, 'N' }, 88 { "rcs", no_argument, 0, 'n' }, 89 { "unidirectional-new-file", no_argument, 0, 'P' }, 90 { "show-c-function", no_argument, 0, 'p' }, 91 { "brief", no_argument, 0, 'q' }, 92 { "recursive", no_argument, 0, 'r' }, 93 { "report-identical-files", no_argument, 0, 's' }, 94 { "starting-file", required_argument, 0, 'S' }, 95 { "expand-tabs", no_argument, 0, 't' }, 96 { "initial-tab", no_argument, 0, 'T' }, 97 { "unified", optional_argument, 0, 'U' }, 98 { "ignore-all-space", no_argument, 0, 'w' }, 99 { "width", required_argument, 0, 'W' }, 100 { "exclude", required_argument, 0, 'x' }, 101 { "exclude-from", required_argument, 0, 'X' }, 102 { "side-by-side", no_argument, NULL, 'y' }, 103 { "ignore-file-name-case", no_argument, NULL, OPT_IGN_FN_CASE }, 104 { "help", no_argument, NULL, OPT_HELP}, 105 { "horizon-lines", required_argument, NULL, OPT_HORIZON_LINES }, 106 { "no-dereference", no_argument, NULL, OPT_NO_DEREFERENCE}, 107 { "no-ignore-file-name-case", no_argument, NULL, OPT_NO_IGN_FN_CASE }, 108 { "normal", no_argument, NULL, OPT_NORMAL }, 109 { "strip-trailing-cr", no_argument, NULL, OPT_STRIPCR }, 110 { "tabsize", required_argument, NULL, OPT_TSIZE }, 111 { "changed-group-format", required_argument, NULL, OPT_CHANGED_GROUP_FORMAT}, 112 { "suppress-common-lines", no_argument, NULL, OPT_SUPPRESS_COMMON }, 113 { "color", optional_argument, NULL, OPT_COLOR }, 114 { "version", no_argument, NULL, OPT_VERSION}, 115 { NULL, 0, 0, '\0'} 116 }; 117 118 static void checked_regcomp(char const *, regex_t *); 119 static void usage(void) __dead2; 120 static void conflicting_format(void) __dead2; 121 static void push_excludes(char *); 122 static void push_ignore_pats(char *); 123 static void read_excludes_file(char *file); 124 static void set_argstr(char **, char **); 125 static char *splice(char *, char *); 126 static bool do_color(void); 127 128 int 129 main(int argc, char **argv) 130 { 131 const char *errstr = NULL; 132 char *ep, **oargv; 133 long l; 134 int ch, dflags, lastch, gotstdin, prevoptind, newarg; 135 136 oargv = argv; 137 gotstdin = 0; 138 dflags = 0; 139 lastch = '\0'; 140 prevoptind = 1; 141 newarg = 1; 142 diff_context = 3; 143 diff_format = D_UNSET; 144 #define FORMAT_MISMATCHED(type) \ 145 (diff_format != D_UNSET && diff_format != (type)) 146 while ((ch = getopt_long(argc, argv, OPTIONS, longopts, NULL)) != -1) { 147 switch (ch) { 148 case '0': case '1': case '2': case '3': case '4': 149 case '5': case '6': case '7': case '8': case '9': 150 if (newarg) 151 usage(); /* disallow -[0-9]+ */ 152 else if (lastch == 'c' || lastch == 'u') 153 diff_context = 0; 154 else if (!isdigit(lastch) || diff_context > INT_MAX / 10) 155 usage(); 156 diff_context = (diff_context * 10) + (ch - '0'); 157 break; 158 case 'a': 159 dflags |= D_FORCEASCII; 160 break; 161 case 'b': 162 dflags |= D_FOLDBLANKS; 163 break; 164 case 'C': 165 case 'c': 166 if (FORMAT_MISMATCHED(D_CONTEXT)) 167 conflicting_format(); 168 cflag = true; 169 diff_format = D_CONTEXT; 170 if (optarg != NULL) { 171 l = strtol(optarg, &ep, 10); 172 if (*ep != '\0' || l < 0 || l >= INT_MAX) 173 usage(); 174 diff_context = (int)l; 175 } 176 break; 177 case 'd': 178 dflags |= D_MINIMAL; 179 break; 180 case 'D': 181 if (FORMAT_MISMATCHED(D_IFDEF)) 182 conflicting_format(); 183 diff_format = D_IFDEF; 184 ifdefname = optarg; 185 break; 186 case 'e': 187 if (FORMAT_MISMATCHED(D_EDIT)) 188 conflicting_format(); 189 diff_format = D_EDIT; 190 break; 191 case 'f': 192 if (FORMAT_MISMATCHED(D_REVERSE)) 193 conflicting_format(); 194 diff_format = D_REVERSE; 195 break; 196 case 'H': 197 /* ignore but needed for compatibility with GNU diff */ 198 break; 199 case 'h': 200 /* silently ignore for backwards compatibility */ 201 break; 202 case 'B': 203 dflags |= D_SKIPBLANKLINES; 204 break; 205 case 'F': 206 if (dflags & D_PROTOTYPE) 207 conflicting_format(); 208 dflags |= D_MATCHLAST; 209 most_recent_pat = xstrdup(optarg); 210 break; 211 case 'I': 212 push_ignore_pats(optarg); 213 break; 214 case 'i': 215 dflags |= D_IGNORECASE; 216 break; 217 case 'L': 218 if (label[0] == NULL) 219 label[0] = optarg; 220 else if (label[1] == NULL) 221 label[1] = optarg; 222 else 223 usage(); 224 break; 225 case 'l': 226 lflag = true; 227 break; 228 case 'N': 229 Nflag = true; 230 break; 231 case 'n': 232 if (FORMAT_MISMATCHED(D_NREVERSE)) 233 conflicting_format(); 234 diff_format = D_NREVERSE; 235 break; 236 case 'p': 237 if (dflags & D_MATCHLAST) 238 conflicting_format(); 239 dflags |= D_PROTOTYPE; 240 break; 241 case 'P': 242 Pflag = true; 243 break; 244 case 'r': 245 rflag = true; 246 break; 247 case 'q': 248 if (FORMAT_MISMATCHED(D_BRIEF)) 249 conflicting_format(); 250 diff_format = D_BRIEF; 251 break; 252 case 'S': 253 start = optarg; 254 break; 255 case 's': 256 sflag = true; 257 break; 258 case 'T': 259 Tflag = true; 260 break; 261 case 't': 262 dflags |= D_EXPANDTABS; 263 break; 264 case 'U': 265 case 'u': 266 if (FORMAT_MISMATCHED(D_UNIFIED)) 267 conflicting_format(); 268 diff_format = D_UNIFIED; 269 if (optarg != NULL) { 270 l = strtol(optarg, &ep, 10); 271 if (*ep != '\0' || l < 0 || l >= INT_MAX) 272 usage(); 273 diff_context = (int)l; 274 } 275 break; 276 case 'w': 277 dflags |= D_IGNOREBLANKS; 278 break; 279 case 'W': 280 width = (int) strtonum(optarg, 1, INT_MAX, &errstr); 281 if (errstr) { 282 warnx("Invalid argument for width"); 283 usage(); 284 } 285 break; 286 case 'X': 287 read_excludes_file(optarg); 288 break; 289 case 'x': 290 push_excludes(optarg); 291 break; 292 case 'y': 293 if (FORMAT_MISMATCHED(D_SIDEBYSIDE)) 294 conflicting_format(); 295 diff_format = D_SIDEBYSIDE; 296 break; 297 case OPT_CHANGED_GROUP_FORMAT: 298 if (FORMAT_MISMATCHED(D_GFORMAT)) 299 conflicting_format(); 300 diff_format = D_GFORMAT; 301 group_format = optarg; 302 break; 303 case OPT_HELP: 304 help = true; 305 usage(); 306 break; 307 case OPT_HORIZON_LINES: 308 break; /* XXX TODO for compatibility with GNU diff3 */ 309 case OPT_IGN_FN_CASE: 310 ignore_file_case = true; 311 break; 312 case OPT_NO_IGN_FN_CASE: 313 ignore_file_case = false; 314 break; 315 case OPT_NORMAL: 316 if (FORMAT_MISMATCHED(D_NORMAL)) 317 conflicting_format(); 318 diff_format = D_NORMAL; 319 break; 320 case OPT_TSIZE: 321 tabsize = (int) strtonum(optarg, 1, INT_MAX, &errstr); 322 if (errstr) { 323 warnx("Invalid argument for tabsize"); 324 usage(); 325 } 326 break; 327 case OPT_STRIPCR: 328 dflags |= D_STRIPCR; 329 break; 330 case OPT_SUPPRESS_COMMON: 331 suppress_common = 1; 332 break; 333 case OPT_COLOR: 334 if (optarg == NULL || strncmp(optarg, "auto", 4) == 0) 335 colorflag = COLORFLAG_AUTO; 336 else if (strncmp(optarg, "always", 6) == 0) 337 colorflag = COLORFLAG_ALWAYS; 338 else if (strncmp(optarg, "never", 5) == 0) 339 colorflag = COLORFLAG_NEVER; 340 else 341 errx(2, "unsupported --color value '%s' (must be always, auto, or never)", 342 optarg); 343 break; 344 case OPT_NO_DEREFERENCE: 345 rflag = true; 346 noderef = true; 347 break; 348 case OPT_VERSION: 349 printf("%s\n", diff_version); 350 exit(0); 351 default: 352 usage(); 353 break; 354 } 355 lastch = ch; 356 newarg = optind != prevoptind; 357 prevoptind = optind; 358 } 359 if (diff_format == D_UNSET && (dflags & D_PROTOTYPE) != 0) 360 diff_format = D_CONTEXT; 361 if (diff_format == D_UNSET) 362 diff_format = D_NORMAL; 363 argc -= optind; 364 argv += optind; 365 366 if (do_color()) { 367 char *p; 368 const char *env; 369 370 color = true; 371 add_code = "32"; 372 del_code = "31"; 373 env = getenv("DIFFCOLORS"); 374 if (env != NULL && *env != '\0' && (p = strdup(env))) { 375 add_code = p; 376 strsep(&p, ":"); 377 if (p != NULL) 378 del_code = p; 379 } 380 } 381 382 #ifdef __OpenBSD__ 383 if (pledge("stdio rpath tmppath", NULL) == -1) 384 err(2, "pledge"); 385 #endif 386 387 /* 388 * Do sanity checks, fill in stb1 and stb2 and call the appropriate 389 * driver routine. Both drivers use the contents of stb1 and stb2. 390 */ 391 if (argc != 2) 392 usage(); 393 checked_regcomp(ignore_pats, &ignore_re); 394 checked_regcomp(most_recent_pat, &most_recent_re); 395 if (strcmp(argv[0], "-") == 0) { 396 fstat(STDIN_FILENO, &stb1); 397 gotstdin = 1; 398 } else if (stat(argv[0], &stb1) != 0) { 399 if (!Nflag || errno != ENOENT) 400 err(2, "%s", argv[0]); 401 dflags |= D_EMPTY1; 402 memset(&stb1, 0, sizeof(struct stat)); 403 } 404 405 if (strcmp(argv[1], "-") == 0) { 406 fstat(STDIN_FILENO, &stb2); 407 gotstdin = 1; 408 } else if (stat(argv[1], &stb2) != 0) { 409 if (!Nflag || errno != ENOENT) 410 err(2, "%s", argv[1]); 411 dflags |= D_EMPTY2; 412 memset(&stb2, 0, sizeof(stb2)); 413 stb2.st_mode = stb1.st_mode; 414 } 415 416 if (dflags & D_EMPTY1 && dflags & D_EMPTY2){ 417 warn("%s", argv[0]); 418 warn("%s", argv[1]); 419 exit(2); 420 } 421 422 if (stb1.st_mode == 0) 423 stb1.st_mode = stb2.st_mode; 424 425 if (gotstdin && (S_ISDIR(stb1.st_mode) || S_ISDIR(stb2.st_mode))) 426 errx(2, "can't compare - to a directory"); 427 set_argstr(oargv, argv); 428 if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) { 429 if (diff_format == D_IFDEF) 430 errx(2, "-D option not supported with directories"); 431 diffdir(argv[0], argv[1], dflags); 432 } else { 433 if (S_ISDIR(stb1.st_mode)) { 434 argv[0] = splice(argv[0], argv[1]); 435 if (stat(argv[0], &stb1) == -1) 436 err(2, "%s", argv[0]); 437 } 438 if (S_ISDIR(stb2.st_mode)) { 439 argv[1] = splice(argv[1], argv[0]); 440 if (stat(argv[1], &stb2) == -1) 441 err(2, "%s", argv[1]); 442 } 443 print_status(diffreg(argv[0], argv[1], dflags, 1), argv[0], 444 argv[1], ""); 445 } 446 exit(status); 447 } 448 449 static void 450 checked_regcomp(char const *pattern, regex_t *comp) 451 { 452 char buf[BUFSIZ]; 453 int error; 454 455 if (pattern == NULL) 456 return; 457 458 error = regcomp(comp, pattern, REG_NEWLINE | REG_EXTENDED); 459 if (error != 0) { 460 regerror(error, comp, buf, sizeof(buf)); 461 if (*pattern != '\0') 462 errx(2, "%s: %s", pattern, buf); 463 else 464 errx(2, "%s", buf); 465 } 466 } 467 468 static void 469 set_argstr(char **av, char **ave) 470 { 471 size_t argsize; 472 char **ap; 473 474 argsize = 4 + *ave - *av + 1; 475 diffargs = xmalloc(argsize); 476 strlcpy(diffargs, "diff", argsize); 477 for (ap = av + 1; ap < ave; ap++) { 478 if (strcmp(*ap, "--") != 0) { 479 strlcat(diffargs, " ", argsize); 480 strlcat(diffargs, *ap, argsize); 481 } 482 } 483 } 484 485 /* 486 * Read in an excludes file and push each line. 487 */ 488 static void 489 read_excludes_file(char *file) 490 { 491 FILE *fp; 492 char *buf, *pattern; 493 size_t len; 494 495 if (strcmp(file, "-") == 0) 496 fp = stdin; 497 else if ((fp = fopen(file, "r")) == NULL) 498 err(2, "%s", file); 499 while ((buf = fgetln(fp, &len)) != NULL) { 500 if (buf[len - 1] == '\n') 501 len--; 502 if ((pattern = strndup(buf, len)) == NULL) 503 err(2, "xstrndup"); 504 push_excludes(pattern); 505 } 506 if (strcmp(file, "-") != 0) 507 fclose(fp); 508 } 509 510 /* 511 * Push a pattern onto the excludes list. 512 */ 513 static void 514 push_excludes(char *pattern) 515 { 516 struct excludes *entry; 517 518 entry = xmalloc(sizeof(*entry)); 519 entry->pattern = pattern; 520 entry->next = excludes_list; 521 excludes_list = entry; 522 } 523 524 static void 525 push_ignore_pats(char *pattern) 526 { 527 size_t len; 528 529 if (ignore_pats == NULL) 530 ignore_pats = xstrdup(pattern); 531 else { 532 /* old + "|" + new + NUL */ 533 len = strlen(ignore_pats) + strlen(pattern) + 2; 534 ignore_pats = xreallocarray(ignore_pats, 1, len); 535 strlcat(ignore_pats, "|", len); 536 strlcat(ignore_pats, pattern, len); 537 } 538 } 539 540 void 541 print_status(int val, char *path1, char *path2, const char *entry) 542 { 543 if (label[0] != NULL) 544 path1 = label[0]; 545 if (label[1] != NULL) 546 path2 = label[1]; 547 548 switch (val) { 549 case D_BINARY: 550 printf("Binary files %s%s and %s%s differ\n", 551 path1, entry, path2, entry); 552 break; 553 case D_DIFFER: 554 if (diff_format == D_BRIEF) 555 printf("Files %s%s and %s%s differ\n", 556 path1, entry, path2, entry); 557 break; 558 case D_SAME: 559 if (sflag) 560 printf("Files %s%s and %s%s are identical\n", 561 path1, entry, path2, entry); 562 break; 563 case D_MISMATCH1: 564 printf("File %s%s is a directory while file %s%s is a regular file\n", 565 path1, entry, path2, entry); 566 break; 567 case D_MISMATCH2: 568 printf("File %s%s is a regular file while file %s%s is a directory\n", 569 path1, entry, path2, entry); 570 break; 571 case D_SKIPPED1: 572 printf("File %s%s is not a regular file or directory and was skipped\n", 573 path1, entry); 574 break; 575 case D_SKIPPED2: 576 printf("File %s%s is not a regular file or directory and was skipped\n", 577 path2, entry); 578 break; 579 case D_ERROR: 580 break; 581 } 582 } 583 584 static void 585 usage(void) 586 { 587 (void)fprintf(help ? stdout : stderr, 588 "usage: diff [-aBbdilpTtw] [-c | -e | -f | -n | -q | -u] [--ignore-case]\n" 589 " [--no-ignore-case] [--normal] [--strip-trailing-cr] [--tabsize]\n" 590 " [-I pattern] [-F pattern] [-L label] file1 file2\n" 591 " diff [-aBbdilpTtw] [-I pattern] [-L label] [--ignore-case]\n" 592 " [--no-ignore-case] [--normal] [--strip-trailing-cr] [--tabsize]\n" 593 " [-F pattern] -C number file1 file2\n" 594 " diff [-aBbdiltw] [-I pattern] [--ignore-case] [--no-ignore-case]\n" 595 " [--normal] [--strip-trailing-cr] [--tabsize] -D string file1 file2\n" 596 " diff [-aBbdilpTtw] [-I pattern] [-L label] [--ignore-case]\n" 597 " [--no-ignore-case] [--normal] [--tabsize] [--strip-trailing-cr]\n" 598 " [-F pattern] -U number file1 file2\n" 599 " diff [-aBbdilNPprsTtw] [-c | -e | -f | -n | -q | -u] [--ignore-case]\n" 600 " [--no-ignore-case] [--normal] [--tabsize] [-I pattern] [-L label]\n" 601 " [-F pattern] [-S name] [-X file] [-x pattern] dir1 dir2\n" 602 " diff [-aBbditwW] [--expand-tabs] [--ignore-all-blanks]\n" 603 " [--ignore-blank-lines] [--ignore-case] [--minimal]\n" 604 " [--no-ignore-file-name-case] [--strip-trailing-cr]\n" 605 " [--suppress-common-lines] [--tabsize] [--text] [--width]\n" 606 " -y | --side-by-side file1 file2\n" 607 " diff [--help] [--version]\n"); 608 609 if (help) 610 exit(0); 611 else 612 exit(2); 613 } 614 615 static void 616 conflicting_format(void) 617 { 618 619 fprintf(stderr, "error: conflicting output format options.\n"); 620 usage(); 621 } 622 623 static bool 624 do_color(void) 625 { 626 const char *p, *p2; 627 628 switch (colorflag) { 629 case COLORFLAG_AUTO: 630 p = getenv("CLICOLOR"); 631 p2 = getenv("COLORTERM"); 632 if ((p != NULL && *p != '\0') || (p2 != NULL && *p2 != '\0')) 633 return isatty(STDOUT_FILENO); 634 break; 635 case COLORFLAG_ALWAYS: 636 return (true); 637 case COLORFLAG_NEVER: 638 return (false); 639 } 640 641 return (false); 642 } 643 644 static char * 645 splice(char *dir, char *path) 646 { 647 char *tail, *buf; 648 size_t dirlen; 649 650 dirlen = strlen(dir); 651 while (dirlen != 0 && dir[dirlen - 1] == '/') 652 dirlen--; 653 if ((tail = strrchr(path, '/')) == NULL) 654 tail = path; 655 else 656 tail++; 657 xasprintf(&buf, "%.*s/%s", (int)dirlen, dir, tail); 658 return (buf); 659 } 660