1 /* $OpenBSD: diff.c,v 1.67 2019/06/28 13:35:00 deraadt Exp $ */
2
3 /*
4 * Copyright (c) 2003 Todd C. Miller <[email protected]>
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 #include <sys/stat.h>
25
26 #include <ctype.h>
27 #include <err.h>
28 #include <errno.h>
29 #include <getopt.h>
30 #include <stdlib.h>
31 #include <stdio.h>
32 #include <string.h>
33 #include <unistd.h>
34 #include <limits.h>
35
36 #include "diff.h"
37 #include "xmalloc.h"
38
39 static const char diff_version[] = "FreeBSD diff 20220309";
40 bool lflag, Nflag, Pflag, rflag, sflag, Tflag, cflag;
41 bool ignore_file_case, suppress_common, color, noderef;
42 static bool help = false;
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_HELP,
62 OPT_HORIZON_LINES,
63 OPT_CHANGED_GROUP_FORMAT,
64 OPT_SUPPRESS_COMMON,
65 OPT_COLOR,
66 OPT_NO_DEREFERENCE,
67 OPT_VERSION,
68 };
69
70 static struct option longopts[] = {
71 { "text", no_argument, 0, 'a' },
72 { "ignore-space-change", no_argument, 0, 'b' },
73 { "context", optional_argument, 0, 'C' },
74 { "ifdef", required_argument, 0, 'D' },
75 { "minimal", no_argument, 0, 'd' },
76 { "ed", no_argument, 0, 'e' },
77 { "forward-ed", no_argument, 0, 'f' },
78 { "show-function-line", required_argument, 0, 'F' },
79 { "speed-large-files", no_argument, NULL, 'H' },
80 { "ignore-blank-lines", no_argument, 0, 'B' },
81 { "ignore-matching-lines", required_argument, 0, 'I' },
82 { "ignore-case", no_argument, 0, 'i' },
83 { "paginate", no_argument, NULL, 'l' },
84 { "label", required_argument, 0, 'L' },
85 { "new-file", no_argument, 0, 'N' },
86 { "rcs", no_argument, 0, 'n' },
87 { "unidirectional-new-file", no_argument, 0, 'P' },
88 { "show-c-function", no_argument, 0, 'p' },
89 { "brief", no_argument, 0, 'q' },
90 { "recursive", no_argument, 0, 'r' },
91 { "report-identical-files", no_argument, 0, 's' },
92 { "starting-file", required_argument, 0, 'S' },
93 { "expand-tabs", no_argument, 0, 't' },
94 { "initial-tab", no_argument, 0, 'T' },
95 { "unified", optional_argument, 0, 'U' },
96 { "ignore-all-space", no_argument, 0, 'w' },
97 { "width", required_argument, 0, 'W' },
98 { "exclude", required_argument, 0, 'x' },
99 { "exclude-from", required_argument, 0, 'X' },
100 { "side-by-side", no_argument, NULL, 'y' },
101 { "ignore-file-name-case", no_argument, NULL, OPT_IGN_FN_CASE },
102 { "help", no_argument, NULL, OPT_HELP},
103 { "horizon-lines", required_argument, NULL, OPT_HORIZON_LINES },
104 { "no-dereference", no_argument, NULL, OPT_NO_DEREFERENCE},
105 { "no-ignore-file-name-case", no_argument, NULL, OPT_NO_IGN_FN_CASE },
106 { "normal", no_argument, NULL, OPT_NORMAL },
107 { "strip-trailing-cr", no_argument, NULL, OPT_STRIPCR },
108 { "tabsize", required_argument, NULL, OPT_TSIZE },
109 { "changed-group-format", required_argument, NULL, OPT_CHANGED_GROUP_FORMAT},
110 { "suppress-common-lines", no_argument, NULL, OPT_SUPPRESS_COMMON },
111 { "color", optional_argument, NULL, OPT_COLOR },
112 { "version", no_argument, NULL, OPT_VERSION},
113 { NULL, 0, 0, '\0'}
114 };
115
116 static void checked_regcomp(char const *, regex_t *);
117 static void usage(void) __dead2;
118 static void conflicting_format(void) __dead2;
119 static void push_excludes(char *);
120 static void push_ignore_pats(char *);
121 static void read_excludes_file(char *file);
122 static void set_argstr(char **, char **);
123 static char *splice(char *, char *);
124 static bool do_color(void);
125
126 int
main(int argc,char ** argv)127 main(int argc, char **argv)
128 {
129 const char *errstr = NULL;
130 char *ep, **oargv;
131 long l;
132 int ch, dflags, lastch, gotstdin, prevoptind, newarg;
133
134 oargv = argv;
135 gotstdin = 0;
136 dflags = 0;
137 lastch = '\0';
138 prevoptind = 1;
139 newarg = 1;
140 diff_context = 3;
141 diff_format = D_UNSET;
142 #define FORMAT_MISMATCHED(type) \
143 (diff_format != D_UNSET && diff_format != (type))
144 while ((ch = getopt_long(argc, argv, OPTIONS, longopts, NULL)) != -1) {
145 switch (ch) {
146 case '0': case '1': case '2': case '3': case '4':
147 case '5': case '6': case '7': case '8': case '9':
148 if (newarg)
149 usage(); /* disallow -[0-9]+ */
150 else if (lastch == 'c' || lastch == 'u')
151 diff_context = 0;
152 else if (!isdigit(lastch) || diff_context > INT_MAX / 10)
153 usage();
154 diff_context = (diff_context * 10) + (ch - '0');
155 break;
156 case 'a':
157 dflags |= D_FORCEASCII;
158 break;
159 case 'b':
160 dflags |= D_FOLDBLANKS;
161 break;
162 case 'C':
163 case 'c':
164 if (FORMAT_MISMATCHED(D_CONTEXT))
165 conflicting_format();
166 cflag = true;
167 diff_format = D_CONTEXT;
168 if (optarg != NULL) {
169 l = strtol(optarg, &ep, 10);
170 if (*ep != '\0' || l < 0 || l >= INT_MAX)
171 usage();
172 diff_context = (int)l;
173 }
174 break;
175 case 'd':
176 dflags |= D_MINIMAL;
177 break;
178 case 'D':
179 if (FORMAT_MISMATCHED(D_IFDEF))
180 conflicting_format();
181 diff_format = D_IFDEF;
182 ifdefname = optarg;
183 break;
184 case 'e':
185 if (FORMAT_MISMATCHED(D_EDIT))
186 conflicting_format();
187 diff_format = D_EDIT;
188 break;
189 case 'f':
190 if (FORMAT_MISMATCHED(D_REVERSE))
191 conflicting_format();
192 diff_format = D_REVERSE;
193 break;
194 case 'H':
195 /* ignore but needed for compatibility with GNU diff */
196 break;
197 case 'h':
198 /* silently ignore for backwards compatibility */
199 break;
200 case 'B':
201 dflags |= D_SKIPBLANKLINES;
202 break;
203 case 'F':
204 if (dflags & D_PROTOTYPE)
205 conflicting_format();
206 dflags |= D_MATCHLAST;
207 most_recent_pat = xstrdup(optarg);
208 break;
209 case 'I':
210 push_ignore_pats(optarg);
211 break;
212 case 'i':
213 dflags |= D_IGNORECASE;
214 break;
215 case 'L':
216 if (label[0] == NULL)
217 label[0] = optarg;
218 else if (label[1] == NULL)
219 label[1] = optarg;
220 else
221 usage();
222 break;
223 case 'l':
224 lflag = true;
225 break;
226 case 'N':
227 Nflag = true;
228 break;
229 case 'n':
230 if (FORMAT_MISMATCHED(D_NREVERSE))
231 conflicting_format();
232 diff_format = D_NREVERSE;
233 break;
234 case 'p':
235 if (dflags & D_MATCHLAST)
236 conflicting_format();
237 dflags |= D_PROTOTYPE;
238 break;
239 case 'P':
240 Pflag = true;
241 break;
242 case 'r':
243 rflag = true;
244 break;
245 case 'q':
246 if (FORMAT_MISMATCHED(D_BRIEF))
247 conflicting_format();
248 diff_format = D_BRIEF;
249 break;
250 case 'S':
251 start = optarg;
252 break;
253 case 's':
254 sflag = true;
255 break;
256 case 'T':
257 Tflag = true;
258 break;
259 case 't':
260 dflags |= D_EXPANDTABS;
261 break;
262 case 'U':
263 case 'u':
264 if (FORMAT_MISMATCHED(D_UNIFIED))
265 conflicting_format();
266 diff_format = D_UNIFIED;
267 if (optarg != NULL) {
268 l = strtol(optarg, &ep, 10);
269 if (*ep != '\0' || l < 0 || l >= INT_MAX)
270 usage();
271 diff_context = (int)l;
272 }
273 break;
274 case 'w':
275 dflags |= D_IGNOREBLANKS;
276 break;
277 case 'W':
278 width = (int) strtonum(optarg, 1, INT_MAX, &errstr);
279 if (errstr)
280 errx(1, "width is %s: %s", errstr, optarg);
281 break;
282 case 'X':
283 read_excludes_file(optarg);
284 break;
285 case 'x':
286 push_excludes(optarg);
287 break;
288 case 'y':
289 if (FORMAT_MISMATCHED(D_SIDEBYSIDE))
290 conflicting_format();
291 diff_format = D_SIDEBYSIDE;
292 break;
293 case OPT_CHANGED_GROUP_FORMAT:
294 if (FORMAT_MISMATCHED(D_GFORMAT))
295 conflicting_format();
296 diff_format = D_GFORMAT;
297 group_format = optarg;
298 break;
299 case OPT_HELP:
300 help = true;
301 usage();
302 break;
303 case OPT_HORIZON_LINES:
304 break; /* XXX TODO for compatibility with GNU diff3 */
305 case OPT_IGN_FN_CASE:
306 ignore_file_case = true;
307 break;
308 case OPT_NO_IGN_FN_CASE:
309 ignore_file_case = false;
310 break;
311 case OPT_NORMAL:
312 if (FORMAT_MISMATCHED(D_NORMAL))
313 conflicting_format();
314 diff_format = D_NORMAL;
315 break;
316 case OPT_TSIZE:
317 tabsize = (int) strtonum(optarg, 1, INT_MAX, &errstr);
318 if (errstr)
319 errx(1, "tabsize is %s: %s", errstr, optarg);
320 break;
321 case OPT_STRIPCR:
322 dflags |= D_STRIPCR;
323 break;
324 case OPT_SUPPRESS_COMMON:
325 suppress_common = 1;
326 break;
327 case OPT_COLOR:
328 if (optarg == NULL || strncmp(optarg, "auto", 4) == 0)
329 colorflag = COLORFLAG_AUTO;
330 else if (strncmp(optarg, "always", 6) == 0)
331 colorflag = COLORFLAG_ALWAYS;
332 else if (strncmp(optarg, "never", 5) == 0)
333 colorflag = COLORFLAG_NEVER;
334 else
335 errx(2, "unsupported --color value '%s' (must be always, auto, or never)",
336 optarg);
337 break;
338 case OPT_NO_DEREFERENCE:
339 noderef = true;
340 break;
341 case OPT_VERSION:
342 printf("%s\n", diff_version);
343 exit(0);
344 default:
345 usage();
346 break;
347 }
348 lastch = ch;
349 newarg = optind != prevoptind;
350 prevoptind = optind;
351 }
352 if (diff_format == D_UNSET && (dflags & D_PROTOTYPE) != 0)
353 diff_format = D_CONTEXT;
354 if (diff_format == D_UNSET)
355 diff_format = D_NORMAL;
356 argc -= optind;
357 argv += optind;
358
359 if (do_color()) {
360 char *p;
361 const char *env;
362
363 color = true;
364 add_code = "32";
365 del_code = "31";
366 env = getenv("DIFFCOLORS");
367 if (env != NULL && *env != '\0' && (p = strdup(env))) {
368 add_code = p;
369 strsep(&p, ":");
370 if (p != NULL)
371 del_code = p;
372 }
373 }
374
375 #ifdef __OpenBSD__
376 if (pledge("stdio rpath tmppath", NULL) == -1)
377 err(2, "pledge");
378 #endif
379
380 /*
381 * Do sanity checks, fill in stb1 and stb2 and call the appropriate
382 * driver routine. Both drivers use the contents of stb1 and stb2.
383 */
384 if (argc != 2)
385 usage();
386 checked_regcomp(ignore_pats, &ignore_re);
387 checked_regcomp(most_recent_pat, &most_recent_re);
388 if (strcmp(argv[0], "-") == 0) {
389 fstat(STDIN_FILENO, &stb1);
390 gotstdin = 1;
391 } else if (stat(argv[0], &stb1) != 0) {
392 if (!Nflag || errno != ENOENT)
393 err(2, "%s", argv[0]);
394 dflags |= D_EMPTY1;
395 memset(&stb1, 0, sizeof(struct stat));
396 }
397
398 if (strcmp(argv[1], "-") == 0) {
399 fstat(STDIN_FILENO, &stb2);
400 gotstdin = 1;
401 } else if (stat(argv[1], &stb2) != 0) {
402 if (!Nflag || errno != ENOENT)
403 err(2, "%s", argv[1]);
404 dflags |= D_EMPTY2;
405 memset(&stb2, 0, sizeof(stb2));
406 stb2.st_mode = stb1.st_mode;
407 }
408
409 if (dflags & D_EMPTY1 && dflags & D_EMPTY2){
410 warn("%s", argv[0]);
411 warn("%s", argv[1]);
412 exit(2);
413 }
414
415 if (stb1.st_mode == 0)
416 stb1.st_mode = stb2.st_mode;
417
418 if (gotstdin && (S_ISDIR(stb1.st_mode) || S_ISDIR(stb2.st_mode)))
419 errx(2, "can't compare - to a directory");
420 set_argstr(oargv, argv);
421 if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) {
422 if (diff_format == D_IFDEF)
423 errx(2, "-D option not supported with directories");
424 diffdir(argv[0], argv[1], dflags);
425 } else {
426 if (S_ISDIR(stb1.st_mode)) {
427 argv[0] = splice(argv[0], argv[1]);
428 if (stat(argv[0], &stb1) == -1)
429 err(2, "%s", argv[0]);
430 }
431 if (S_ISDIR(stb2.st_mode)) {
432 argv[1] = splice(argv[1], argv[0]);
433 if (stat(argv[1], &stb2) == -1)
434 err(2, "%s", argv[1]);
435 }
436 print_status(diffreg(argv[0], argv[1], dflags, 1), argv[0],
437 argv[1], "");
438 }
439 exit(status);
440 }
441
442 static void
checked_regcomp(char const * pattern,regex_t * comp)443 checked_regcomp(char const *pattern, regex_t *comp)
444 {
445 char buf[BUFSIZ];
446 int error;
447
448 if (pattern == NULL)
449 return;
450
451 error = regcomp(comp, pattern, REG_NEWLINE | REG_EXTENDED);
452 if (error != 0) {
453 regerror(error, comp, buf, sizeof(buf));
454 if (*pattern != '\0')
455 errx(2, "%s: %s", pattern, buf);
456 else
457 errx(2, "%s", buf);
458 }
459 }
460
461 static void
set_argstr(char ** av,char ** ave)462 set_argstr(char **av, char **ave)
463 {
464 size_t argsize;
465 char **ap;
466
467 argsize = 4 + *ave - *av + 1;
468 diffargs = xmalloc(argsize);
469 strlcpy(diffargs, "diff", argsize);
470 for (ap = av + 1; ap < ave; ap++) {
471 if (strcmp(*ap, "--") != 0) {
472 strlcat(diffargs, " ", argsize);
473 strlcat(diffargs, *ap, argsize);
474 }
475 }
476 }
477
478 /*
479 * Read in an excludes file and push each line.
480 */
481 static void
read_excludes_file(char * file)482 read_excludes_file(char *file)
483 {
484 FILE *fp;
485 char *buf, *pattern;
486 size_t len;
487
488 if (strcmp(file, "-") == 0)
489 fp = stdin;
490 else if ((fp = fopen(file, "r")) == NULL)
491 err(2, "%s", file);
492 while ((buf = fgetln(fp, &len)) != NULL) {
493 if (buf[len - 1] == '\n')
494 len--;
495 if ((pattern = strndup(buf, len)) == NULL)
496 err(2, "xstrndup");
497 push_excludes(pattern);
498 }
499 if (strcmp(file, "-") != 0)
500 fclose(fp);
501 }
502
503 /*
504 * Push a pattern onto the excludes list.
505 */
506 static void
push_excludes(char * pattern)507 push_excludes(char *pattern)
508 {
509 struct excludes *entry;
510
511 entry = xmalloc(sizeof(*entry));
512 entry->pattern = pattern;
513 entry->next = excludes_list;
514 excludes_list = entry;
515 }
516
517 static void
push_ignore_pats(char * pattern)518 push_ignore_pats(char *pattern)
519 {
520 size_t len;
521
522 if (ignore_pats == NULL)
523 ignore_pats = xstrdup(pattern);
524 else {
525 /* old + "|" + new + NUL */
526 len = strlen(ignore_pats) + strlen(pattern) + 2;
527 ignore_pats = xreallocarray(ignore_pats, 1, len);
528 strlcat(ignore_pats, "|", len);
529 strlcat(ignore_pats, pattern, len);
530 }
531 }
532
533 void
print_status(int val,char * path1,char * path2,const char * entry)534 print_status(int val, char *path1, char *path2, const char *entry)
535 {
536 if (label[0] != NULL)
537 path1 = label[0];
538 if (label[1] != NULL)
539 path2 = label[1];
540
541 switch (val) {
542 case D_BINARY:
543 printf("Binary files %s%s and %s%s differ\n",
544 path1, entry, path2, entry);
545 break;
546 case D_DIFFER:
547 if (diff_format == D_BRIEF)
548 printf("Files %s%s and %s%s differ\n",
549 path1, entry, path2, entry);
550 break;
551 case D_SAME:
552 if (sflag)
553 printf("Files %s%s and %s%s are identical\n",
554 path1, entry, path2, entry);
555 break;
556 case D_MISMATCH1:
557 printf("File %s%s is a directory while file %s%s is a regular file\n",
558 path1, entry, path2, entry);
559 break;
560 case D_MISMATCH2:
561 printf("File %s%s is a regular file while file %s%s is a directory\n",
562 path1, entry, path2, entry);
563 break;
564 case D_SKIPPED1:
565 printf("File %s%s is not a regular file or directory and was skipped\n",
566 path1, entry);
567 break;
568 case D_SKIPPED2:
569 printf("File %s%s is not a regular file or directory and was skipped\n",
570 path2, entry);
571 break;
572 case D_ERROR:
573 break;
574 }
575 }
576
577 static void
usage(void)578 usage(void)
579 {
580 (void)fprintf(help ? stdout : stderr,
581 "usage: diff [-aBbdilpTtw] [-c | -e | -f | -n | -q | -u] [--ignore-case]\n"
582 " [--no-ignore-case] [--normal] [--strip-trailing-cr] [--tabsize]\n"
583 " [-I pattern] [-F pattern] [-L label] file1 file2\n"
584 " diff [-aBbdilpTtw] [-I pattern] [-L label] [--ignore-case]\n"
585 " [--no-ignore-case] [--normal] [--strip-trailing-cr] [--tabsize]\n"
586 " [-F pattern] -C number file1 file2\n"
587 " diff [-aBbdiltw] [-I pattern] [--ignore-case] [--no-ignore-case]\n"
588 " [--normal] [--strip-trailing-cr] [--tabsize] -D string file1 file2\n"
589 " diff [-aBbdilpTtw] [-I pattern] [-L label] [--ignore-case]\n"
590 " [--no-ignore-case] [--normal] [--tabsize] [--strip-trailing-cr]\n"
591 " [-F pattern] -U number file1 file2\n"
592 " diff [-aBbdilNPprsTtw] [-c | -e | -f | -n | -q | -u] [--ignore-case]\n"
593 " [--no-ignore-case] [--normal] [--tabsize] [-I pattern] [-L label]\n"
594 " [-F pattern] [-S name] [-X file] [-x pattern] dir1 dir2\n"
595 " diff [-aBbditwW] [--expand-tabs] [--ignore-all-space]\n"
596 " [--ignore-blank-lines] [--ignore-case] [--minimal]\n"
597 " [--no-ignore-file-name-case] [--strip-trailing-cr]\n"
598 " [--suppress-common-lines] [--tabsize] [--text] [--width]\n"
599 " -y | --side-by-side file1 file2\n"
600 " diff [--help] [--version]\n");
601
602 if (help)
603 exit(0);
604 else
605 exit(2);
606 }
607
608 static void
conflicting_format(void)609 conflicting_format(void)
610 {
611
612 fprintf(stderr, "error: conflicting output format options.\n");
613 usage();
614 }
615
616 static bool
do_color(void)617 do_color(void)
618 {
619 const char *p, *p2;
620
621 switch (colorflag) {
622 case COLORFLAG_AUTO:
623 p = getenv("CLICOLOR");
624 p2 = getenv("COLORTERM");
625 if ((p != NULL && *p != '\0') || (p2 != NULL && *p2 != '\0'))
626 return isatty(STDOUT_FILENO);
627 break;
628 case COLORFLAG_ALWAYS:
629 return (true);
630 case COLORFLAG_NEVER:
631 return (false);
632 }
633
634 return (false);
635 }
636
637 static char *
splice(char * dir,char * path)638 splice(char *dir, char *path)
639 {
640 char *tail, *buf;
641 size_t dirlen;
642
643 dirlen = strlen(dir);
644 while (dirlen != 0 && dir[dirlen - 1] == '/')
645 dirlen--;
646 if ((tail = strrchr(path, '/')) == NULL)
647 tail = path;
648 else
649 tail++;
650 xasprintf(&buf, "%.*s/%s", (int)dirlen, dir, tail);
651 return (buf);
652 }
653