1 /* $OpenBSD: diff.c,v 1.48 2004/12/09 18:56:10 millert 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 #ifndef lint 24 static const char rcsid[] = "$OpenBSD: diff.c,v 1.48 2004/12/09 18:56:10 millert Exp $"; 25 #endif /* not lint */ 26 27 #include <sys/param.h> 28 #include <sys/stat.h> 29 30 #include <ctype.h> 31 #include <err.h> 32 #include <errno.h> 33 #include <getopt.h> 34 #include <signal.h> 35 #include <stdlib.h> 36 #include <stdio.h> 37 #include <stdarg.h> 38 #include <string.h> 39 #include <unistd.h> 40 41 #include "diff.h" 42 43 int aflag, bflag, dflag, iflag, lflag, Nflag, Pflag, pflag, rflag; 44 int sflag, tflag, Tflag, wflag; 45 int format, context, status; 46 char *start, *ifdefname, *diffargs, *label[2], *ignore_pats; 47 struct stat stb1, stb2; 48 struct excludes *excludes_list; 49 regex_t ignore_re; 50 51 #define OPTIONS "0123456789abC:cdD:efhI:iL:lnNPpqrS:sTtU:uwX:x:" 52 static struct option longopts[] = { 53 { "text", no_argument, 0, 'a' }, 54 { "ignore-space-change", no_argument, 0, 'b' }, 55 { "context", optional_argument, 0, 'C' }, 56 { "ifdef", required_argument, 0, 'D' }, 57 { "minimal", no_argument, 0, 'd' }, 58 { "ed", no_argument, 0, 'e' }, 59 { "forward-ed", no_argument, 0, 'f' }, 60 { "ignore-matching-lines", required_argument, 0, 'I' }, 61 { "ignore-case", no_argument, 0, 'i' }, 62 { "paginate", no_argument, 0, 'l' }, 63 { "label", required_argument, 0, 'L' }, 64 { "new-file", no_argument, 0, 'N' }, 65 { "rcs", no_argument, 0, 'n' }, 66 { "unidirectional-new-file", no_argument, 0, 'P' }, 67 { "show-c-function", no_argument, 0, 'p' }, 68 { "brief", no_argument, 0, 'q' }, 69 { "recursive", no_argument, 0, 'r' }, 70 { "report-identical-files", no_argument, 0, 's' }, 71 { "starting-file", required_argument, 0, 'S' }, 72 { "expand-tabs", no_argument, 0, 't' }, 73 { "initial-tab", no_argument, 0, 'T' }, 74 { "unified", optional_argument, 0, 'U' }, 75 { "ignore-all-space", no_argument, 0, 'w' }, 76 { "exclude", required_argument, 0, 'x' }, 77 { "exclude-from", required_argument, 0, 'X' }, 78 { NULL, 0, 0, '\0'} 79 }; 80 81 __dead void usage(void); 82 void push_excludes(char *); 83 void push_ignore_pats(char *); 84 void read_excludes_file(char *file); 85 void set_argstr(char **, char **); 86 87 int 88 main(int argc, char **argv) 89 { 90 char *ep, **oargv; 91 long l; 92 int ch, lastch, gotstdin, prevoptind, newarg; 93 94 oargv = argv; 95 gotstdin = 0; 96 97 lastch = '\0'; 98 prevoptind = 1; 99 newarg = 1; 100 while ((ch = getopt_long(argc, argv, OPTIONS, longopts, NULL)) != -1) { 101 switch (ch) { 102 case '0': case '1': case '2': case '3': case '4': 103 case '5': case '6': case '7': case '8': case '9': 104 if (newarg) 105 usage(); /* disallow -[0-9]+ */ 106 else if (lastch == 'c' || lastch == 'u') 107 context = 0; 108 else if (!isdigit(lastch) || context > INT_MAX / 10) 109 usage(); 110 context = (context * 10) + (ch - '0'); 111 break; 112 case 'a': 113 aflag = 1; 114 break; 115 case 'b': 116 bflag = 1; 117 break; 118 case 'C': 119 case 'c': 120 format = D_CONTEXT; 121 if (optarg != NULL) { 122 l = strtol(optarg, &ep, 10); 123 if (*ep != '\0' || l < 0 || l >= INT_MAX) 124 usage(); 125 context = (int)l; 126 } else 127 context = 3; 128 break; 129 case 'd': 130 dflag = 1; 131 break; 132 case 'D': 133 format = D_IFDEF; 134 ifdefname = optarg; 135 break; 136 case 'e': 137 format = D_EDIT; 138 break; 139 case 'f': 140 format = D_REVERSE; 141 break; 142 case 'h': 143 /* silently ignore for backwards compatibility */ 144 break; 145 case 'I': 146 push_ignore_pats(optarg); 147 break; 148 case 'i': 149 iflag = 1; 150 break; 151 case 'L': 152 if (label[0] == NULL) 153 label[0] = optarg; 154 else if (label[1] == NULL) 155 label[1] = optarg; 156 else 157 usage(); 158 break; 159 case 'l': 160 lflag = 1; 161 signal(SIGPIPE, SIG_IGN); 162 break; 163 case 'N': 164 Nflag = 1; 165 break; 166 case 'n': 167 format = D_NREVERSE; 168 break; 169 case 'p': 170 pflag = 1; 171 break; 172 case 'P': 173 Pflag = 1; 174 break; 175 case 'r': 176 rflag = 1; 177 break; 178 case 'q': 179 format = D_BRIEF; 180 break; 181 case 'S': 182 start = optarg; 183 break; 184 case 's': 185 sflag = 1; 186 break; 187 case 'T': 188 Tflag = 1; 189 break; 190 case 't': 191 tflag = 1; 192 break; 193 case 'U': 194 case 'u': 195 format = D_UNIFIED; 196 if (optarg != NULL) { 197 l = strtol(optarg, &ep, 10); 198 if (*ep != '\0' || l < 0 || l >= INT_MAX) 199 usage(); 200 context = (int)l; 201 } else 202 context = 3; 203 break; 204 case 'w': 205 wflag = 1; 206 break; 207 case 'X': 208 read_excludes_file(optarg); 209 break; 210 case 'x': 211 push_excludes(optarg); 212 break; 213 default: 214 usage(); 215 break; 216 } 217 lastch = ch; 218 newarg = optind != prevoptind; 219 prevoptind = optind; 220 } 221 argc -= optind; 222 argv += optind; 223 224 /* 225 * Do sanity checks, fill in stb1 and stb2 and call the appropriate 226 * driver routine. Both drivers use the contents of stb1 and stb2. 227 */ 228 if (argc != 2) 229 usage(); 230 if (ignore_pats != NULL) { 231 char buf[BUFSIZ]; 232 int error; 233 234 if ((error = regcomp(&ignore_re, ignore_pats, 235 REG_NEWLINE | REG_EXTENDED)) != 0) { 236 regerror(error, &ignore_re, buf, sizeof(buf)); 237 if (*ignore_pats != '\0') 238 errx(2, "%s: %s", ignore_pats, buf); 239 else 240 errx(2, "%s", buf); 241 } 242 } 243 if (strcmp(argv[0], "-") == 0) { 244 fstat(STDIN_FILENO, &stb1); 245 gotstdin = 1; 246 } else if (stat(argv[0], &stb1) != 0) 247 err(2, "%s", argv[0]); 248 if (strcmp(argv[1], "-") == 0) { 249 fstat(STDIN_FILENO, &stb2); 250 gotstdin = 1; 251 } else if (stat(argv[1], &stb2) != 0) 252 err(2, "%s", argv[1]); 253 if (gotstdin && (S_ISDIR(stb1.st_mode) || S_ISDIR(stb2.st_mode))) 254 errx(2, "can't compare - to a directory"); 255 set_argstr(oargv, argv); 256 if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) { 257 if (format == D_IFDEF) 258 errx(2, "-D option not supported with directories"); 259 diffdir(argv[0], argv[1]); 260 } else { 261 if (S_ISDIR(stb1.st_mode)) { 262 argv[0] = splice(argv[0], argv[1]); 263 if (stat(argv[0], &stb1) < 0) 264 err(2, "%s", argv[0]); 265 } 266 if (S_ISDIR(stb2.st_mode)) { 267 argv[1] = splice(argv[1], argv[0]); 268 if (stat(argv[1], &stb2) < 0) 269 err(2, "%s", argv[1]); 270 } 271 print_status(diffreg(argv[0], argv[1], 0), argv[0], argv[1], 272 NULL); 273 } 274 exit(status); 275 } 276 277 void * 278 emalloc(size_t n) 279 { 280 void *p; 281 282 if ((p = malloc(n)) == NULL) 283 err(2, NULL); 284 return (p); 285 } 286 287 void * 288 erealloc(void *p, size_t n) 289 { 290 void *q; 291 292 if ((q = realloc(p, n)) == NULL) 293 err(2, NULL); 294 return (q); 295 } 296 297 int 298 easprintf(char **ret, const char *fmt, ...) 299 { 300 int len; 301 va_list ap; 302 303 va_start(ap, fmt); 304 len = vasprintf(ret, fmt, ap); 305 va_end(ap); 306 307 if (len == -1) 308 err(2, NULL); 309 return (len); 310 } 311 312 void 313 set_argstr(char **av, char **ave) 314 { 315 size_t argsize; 316 char **ap; 317 318 argsize = 4 + *ave - *av + 1; 319 diffargs = emalloc(argsize); 320 strlcpy(diffargs, "diff", argsize); 321 for (ap = av + 1; ap < ave; ap++) { 322 if (strcmp(*ap, "--") != 0) { 323 strlcat(diffargs, " ", argsize); 324 strlcat(diffargs, *ap, argsize); 325 } 326 } 327 } 328 329 /* 330 * Read in an excludes file and push each line. 331 */ 332 void 333 read_excludes_file(char *file) 334 { 335 FILE *fp; 336 char *buf, *pattern; 337 size_t len; 338 339 if (strcmp(file, "-") == 0) 340 fp = stdin; 341 else if ((fp = fopen(file, "r")) == NULL) 342 err(2, "%s", file); 343 while ((buf = fgetln(fp, &len)) != NULL) { 344 if (buf[len - 1] == '\n') 345 len--; 346 pattern = emalloc(len + 1); 347 memcpy(pattern, buf, len); 348 pattern[len] = '\0'; 349 push_excludes(pattern); 350 } 351 if (strcmp(file, "-") != 0) 352 fclose(fp); 353 } 354 355 /* 356 * Push a pattern onto the excludes list. 357 */ 358 void 359 push_excludes(char *pattern) 360 { 361 struct excludes *entry; 362 363 entry = emalloc(sizeof(*entry)); 364 entry->pattern = pattern; 365 entry->next = excludes_list; 366 excludes_list = entry; 367 } 368 369 void 370 push_ignore_pats(char *pattern) 371 { 372 size_t len; 373 374 if (ignore_pats == NULL) { 375 /* XXX: estrdup */ 376 len = strlen(pattern) + 1; 377 ignore_pats = emalloc(len); 378 strlcpy(ignore_pats, pattern, len); 379 } else { 380 /* old + "|" + new + NUL */ 381 len = strlen(ignore_pats) + strlen(pattern) + 2; 382 ignore_pats = erealloc(ignore_pats, len); 383 strlcat(ignore_pats, "|", len); 384 strlcat(ignore_pats, pattern, len); 385 } 386 } 387 388 void 389 print_only(const char *path, size_t dirlen, const char *entry) 390 { 391 if (dirlen > 1) 392 dirlen--; 393 printf("Only in %.*s: %s\n", (int)dirlen, path, entry); 394 } 395 396 void 397 print_status(int val, char *path1, char *path2, char *entry) 398 { 399 switch (val) { 400 case D_ONLY: 401 print_only(path1, strlen(path1), entry); 402 break; 403 case D_COMMON: 404 printf("Common subdirectories: %s%s and %s%s\n", 405 path1, entry ? entry : "", path2, entry ? entry : ""); 406 break; 407 case D_BINARY: 408 printf("Binary files %s%s and %s%s differ\n", 409 path1, entry ? entry : "", path2, entry ? entry : ""); 410 break; 411 case D_DIFFER: 412 if (format == D_BRIEF) 413 printf("Files %s%s and %s%s differ\n", 414 path1, entry ? entry : "", 415 path2, entry ? entry : ""); 416 break; 417 case D_SAME: 418 if (sflag) 419 printf("Files %s%s and %s%s are identical\n", 420 path1, entry ? entry : "", 421 path2, entry ? entry : ""); 422 break; 423 case D_MISMATCH1: 424 printf("File %s%s is a directory while file %s%s is a regular file\n", 425 path1, entry ? entry : "", path2, entry ? entry : ""); 426 break; 427 case D_MISMATCH2: 428 printf("File %s%s is a regular file while file %s%s is a directory\n", 429 path1, entry ? entry : "", path2, entry ? entry : ""); 430 break; 431 case D_SKIPPED1: 432 printf("File %s%s is not a regular file or directory and was skipped\n", 433 path1, entry ? entry : ""); 434 break; 435 case D_SKIPPED2: 436 printf("File %s%s is not a regular file or directory and was skipped\n", 437 path2, entry ? entry : ""); 438 break; 439 } 440 } 441 442 __dead void 443 usage(void) 444 { 445 (void)fprintf(stderr, 446 "usage: diff [-abdilpqtTw] [-I pattern] [-c | -e | -f | -n | -u]\n" 447 " [-L label] file1 file2\n" 448 " diff [-abdilpqtTw] [-I pattern] [-L label] -C number file1 file2\n" 449 " diff [-abdilqtw] [-I pattern] -D string file1 file2\n" 450 " diff [-abdilpqtTw] [-I pattern] [-L label] -U number file1 file2\n" 451 " diff [-abdilNPpqtTw] [-I pattern] [-c | -e | -f | -n | -u]\n" 452 " [-L label] [-r] [-s] [-S name] [-X file] [-x pattern] dir1\n" 453 " dir2\n"); 454 455 exit(2); 456 } 457