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