[821] | 1 | /* vi: set sw=4 ts=4: */
|
---|
| 2 | /*
|
---|
| 3 | * Mini diff implementation for busybox, adapted from OpenBSD diff.
|
---|
| 4 | *
|
---|
| 5 | * Copyright (C) 2006 by Robert Sullivan <cogito.ergo.cogito@hotmail.com>
|
---|
| 6 | * Copyright (c) 2003 Todd C. Miller <Todd.Miller@courtesan.com>
|
---|
| 7 | *
|
---|
| 8 | * Sponsored in part by the Defense Advanced Research Projects
|
---|
| 9 | * Agency (DARPA) and Air Force Research Laboratory, Air Force
|
---|
| 10 | * Materiel Command, USAF, under agreement number F39502-99-1-0512.
|
---|
| 11 | *
|
---|
| 12 | * Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
|
---|
| 13 | */
|
---|
| 14 |
|
---|
| 15 | #include <time.h>
|
---|
| 16 | #include <sys/types.h>
|
---|
| 17 | #include <sys/param.h>
|
---|
| 18 | #include <sys/stat.h>
|
---|
| 19 | #include <ctype.h>
|
---|
| 20 | #include <errno.h>
|
---|
| 21 | #include <signal.h>
|
---|
| 22 | #include <stdlib.h>
|
---|
| 23 | #include <stdio.h>
|
---|
| 24 | #include <stdarg.h>
|
---|
| 25 | #include <string.h>
|
---|
| 26 | #include <unistd.h>
|
---|
| 27 | #include <sys/wait.h>
|
---|
| 28 | #include <fcntl.h>
|
---|
| 29 | #include <stddef.h>
|
---|
| 30 | #include <paths.h>
|
---|
| 31 | #include <dirent.h>
|
---|
| 32 | #include "busybox.h"
|
---|
| 33 |
|
---|
| 34 | #define FSIZE_MAX 32768
|
---|
| 35 |
|
---|
| 36 | /*
|
---|
| 37 | * Output flags
|
---|
| 38 | */
|
---|
| 39 | #define D_HEADER 1 /* Print a header/footer between files */
|
---|
| 40 | #define D_EMPTY1 2 /* Treat first file as empty (/dev/null) */
|
---|
| 41 | #define D_EMPTY2 4 /* Treat second file as empty (/dev/null) */
|
---|
| 42 |
|
---|
| 43 | /*
|
---|
| 44 | * Status values for print_status() and diffreg() return values
|
---|
| 45 | * Guide:
|
---|
| 46 | * D_SAME - files are the same
|
---|
| 47 | * D_DIFFER - files differ
|
---|
| 48 | * D_BINARY - binary files differ
|
---|
| 49 | * D_COMMON - subdirectory common to both dirs
|
---|
| 50 | * D_ONLY - file only exists in one dir
|
---|
| 51 | * D_MISMATCH1 - path1 a dir, path2 a file
|
---|
| 52 | * D_MISMATCH2 - path1 a file, path2 a dir
|
---|
| 53 | * D_ERROR - error occurred
|
---|
| 54 | * D_SKIPPED1 - skipped path1 as it is a special file
|
---|
| 55 | * D_SKIPPED2 - skipped path2 as it is a special file
|
---|
| 56 | */
|
---|
| 57 |
|
---|
| 58 | #define D_SAME 0
|
---|
| 59 | #define D_DIFFER (1<<0)
|
---|
| 60 | #define D_BINARY (1<<1)
|
---|
| 61 | #define D_COMMON (1<<2)
|
---|
| 62 | #define D_ONLY (1<<3)
|
---|
| 63 | #define D_MISMATCH1 (1<<4)
|
---|
| 64 | #define D_MISMATCH2 (1<<5)
|
---|
| 65 | #define D_ERROR (1<<6)
|
---|
| 66 | #define D_SKIPPED1 (1<<7)
|
---|
| 67 | #define D_SKIPPED2 (1<<8)
|
---|
| 68 |
|
---|
| 69 | /* Command line options */
|
---|
| 70 | static unsigned long cmd_flags;
|
---|
| 71 |
|
---|
| 72 | #define FLAG_a (1<<0)
|
---|
| 73 | #define FLAG_b (1<<1)
|
---|
| 74 | #define FLAG_d (1<<2)
|
---|
| 75 | #define FLAG_i (1<<3)
|
---|
| 76 | #define FLAG_L (1<<4)
|
---|
| 77 | #define FLAG_N (1<<5)
|
---|
| 78 | #define FLAG_q (1<<6)
|
---|
| 79 | #define FLAG_r (1<<7)
|
---|
| 80 | #define FLAG_s (1<<8)
|
---|
| 81 | #define FLAG_S (1<<9)
|
---|
| 82 | #define FLAG_t (1<<10)
|
---|
| 83 | #define FLAG_T (1<<11)
|
---|
| 84 | #define FLAG_U (1<<12)
|
---|
| 85 | #define FLAG_w (1<<13)
|
---|
| 86 |
|
---|
| 87 | int context, status;
|
---|
| 88 | char *start, *label[2];
|
---|
| 89 | struct stat stb1, stb2;
|
---|
| 90 | char **dl;
|
---|
| 91 | int dl_count = 0;
|
---|
| 92 |
|
---|
| 93 | struct cand {
|
---|
| 94 | int x;
|
---|
| 95 | int y;
|
---|
| 96 | int pred;
|
---|
| 97 | };
|
---|
| 98 |
|
---|
| 99 | struct line {
|
---|
| 100 | int serial;
|
---|
| 101 | int value;
|
---|
| 102 | } *file[2];
|
---|
| 103 |
|
---|
| 104 | /*
|
---|
| 105 | * The following struct is used to record change information
|
---|
| 106 | * doing a "context" or "unified" diff. (see routine "change" to
|
---|
| 107 | * understand the highly mnemonic field names)
|
---|
| 108 | */
|
---|
| 109 | struct context_vec {
|
---|
| 110 | int a; /* start line in old file */
|
---|
| 111 | int b; /* end line in old file */
|
---|
| 112 | int c; /* start line in new file */
|
---|
| 113 | int d; /* end line in new file */
|
---|
| 114 | };
|
---|
| 115 |
|
---|
| 116 | static int *J; /* will be overlaid on class */
|
---|
| 117 | static int *class; /* will be overlaid on file[0] */
|
---|
| 118 | static int *klist; /* will be overlaid on file[0] after class */
|
---|
| 119 | static int *member; /* will be overlaid on file[1] */
|
---|
| 120 | static int clen;
|
---|
| 121 | static int len[2];
|
---|
| 122 | static int pref, suff; /* length of prefix and suffix */
|
---|
| 123 | static int slen[2];
|
---|
| 124 | static int anychange;
|
---|
| 125 | static long *ixnew; /* will be overlaid on file[1] */
|
---|
| 126 | static long *ixold; /* will be overlaid on klist */
|
---|
| 127 | static struct cand *clist; /* merely a free storage pot for candidates */
|
---|
| 128 | static int clistlen; /* the length of clist */
|
---|
| 129 | static struct line *sfile[2]; /* shortened by pruning common prefix/suffix */
|
---|
| 130 | static struct context_vec *context_vec_start;
|
---|
| 131 | static struct context_vec *context_vec_end;
|
---|
| 132 | static struct context_vec *context_vec_ptr;
|
---|
| 133 |
|
---|
| 134 | static void print_only(const char *path, size_t dirlen, const char *entry)
|
---|
| 135 | {
|
---|
| 136 | if (dirlen > 1)
|
---|
| 137 | dirlen--;
|
---|
| 138 | printf("Only in %.*s: %s\n", (int) dirlen, path, entry);
|
---|
| 139 | }
|
---|
| 140 |
|
---|
| 141 | static void print_status(int val, char *path1, char *path2, char *entry)
|
---|
| 142 | {
|
---|
| 143 | const char *const _entry = entry ? entry : "";
|
---|
| 144 | char *_path1 = entry ? concat_path_file(path1, _entry) : path1;
|
---|
| 145 | char *_path2 = entry ? concat_path_file(path2, _entry) : path2;
|
---|
| 146 |
|
---|
| 147 | switch (val) {
|
---|
| 148 | case D_ONLY:
|
---|
| 149 | print_only(path1, strlen(path1), entry);
|
---|
| 150 | break;
|
---|
| 151 | case D_COMMON:
|
---|
| 152 | printf("Common subdirectories: %s and %s\n", _path1, _path2);
|
---|
| 153 | break;
|
---|
| 154 | case D_BINARY:
|
---|
| 155 | printf("Binary files %s and %s differ\n", _path1, _path2);
|
---|
| 156 | break;
|
---|
| 157 | case D_DIFFER:
|
---|
| 158 | if (cmd_flags & FLAG_q)
|
---|
| 159 | printf("Files %s and %s differ\n", _path1, _path2);
|
---|
| 160 | break;
|
---|
| 161 | case D_SAME:
|
---|
| 162 | if (cmd_flags & FLAG_s)
|
---|
| 163 | printf("Files %s and %s are identical\n", _path1, _path2);
|
---|
| 164 | break;
|
---|
| 165 | case D_MISMATCH1:
|
---|
| 166 | printf("File %s is a directory while file %s is a regular file\n",
|
---|
| 167 | _path1, _path2);
|
---|
| 168 | break;
|
---|
| 169 | case D_MISMATCH2:
|
---|
| 170 | printf("File %s is a regular file while file %s is a directory\n",
|
---|
| 171 | _path1, _path2);
|
---|
| 172 | break;
|
---|
| 173 | case D_SKIPPED1:
|
---|
| 174 | printf("File %s is not a regular file or directory and was skipped\n",
|
---|
| 175 | _path1);
|
---|
| 176 | break;
|
---|
| 177 | case D_SKIPPED2:
|
---|
| 178 | printf("File %s is not a regular file or directory and was skipped\n",
|
---|
| 179 | _path2);
|
---|
| 180 | break;
|
---|
| 181 | }
|
---|
| 182 | if (entry) {
|
---|
| 183 | free(_path1);
|
---|
| 184 | free(_path2);
|
---|
| 185 | }
|
---|
| 186 | }
|
---|
| 187 |
|
---|
| 188 | /*
|
---|
| 189 | * Hash function taken from Robert Sedgewick, Algorithms in C, 3d ed., p 578.
|
---|
| 190 | */
|
---|
| 191 | static int readhash(FILE * f)
|
---|
| 192 | {
|
---|
| 193 | int i, t, space;
|
---|
| 194 | int sum;
|
---|
| 195 |
|
---|
| 196 | sum = 1;
|
---|
| 197 | space = 0;
|
---|
| 198 | if (!(cmd_flags & FLAG_b) && !(cmd_flags & FLAG_w)) {
|
---|
| 199 | if (FLAG_i)
|
---|
| 200 | for (i = 0; (t = getc(f)) != '\n'; i++) {
|
---|
| 201 | if (t == EOF) {
|
---|
| 202 | if (i == 0)
|
---|
| 203 | return (0);
|
---|
| 204 | break;
|
---|
| 205 | }
|
---|
| 206 | sum = sum * 127 + t;
|
---|
| 207 | } else
|
---|
| 208 | for (i = 0; (t = getc(f)) != '\n'; i++) {
|
---|
| 209 | if (t == EOF) {
|
---|
| 210 | if (i == 0)
|
---|
| 211 | return (0);
|
---|
| 212 | break;
|
---|
| 213 | }
|
---|
| 214 | sum = sum * 127 + t;
|
---|
| 215 | }
|
---|
| 216 | } else {
|
---|
| 217 | for (i = 0;;) {
|
---|
| 218 | switch (t = getc(f)) {
|
---|
| 219 | case '\t':
|
---|
| 220 | case '\r':
|
---|
| 221 | case '\v':
|
---|
| 222 | case '\f':
|
---|
| 223 | case ' ':
|
---|
| 224 | space++;
|
---|
| 225 | continue;
|
---|
| 226 | default:
|
---|
| 227 | if (space && !(cmd_flags & FLAG_w)) {
|
---|
| 228 | i++;
|
---|
| 229 | space = 0;
|
---|
| 230 | }
|
---|
| 231 | sum = sum * 127 + t;
|
---|
| 232 | i++;
|
---|
| 233 | continue;
|
---|
| 234 | case EOF:
|
---|
| 235 | if (i == 0)
|
---|
| 236 | return (0);
|
---|
| 237 | /* FALLTHROUGH */
|
---|
| 238 | case '\n':
|
---|
| 239 | break;
|
---|
| 240 | }
|
---|
| 241 | break;
|
---|
| 242 | }
|
---|
| 243 | }
|
---|
| 244 | /*
|
---|
| 245 | * There is a remote possibility that we end up with a zero sum.
|
---|
| 246 | * Zero is used as an EOF marker, so return 1 instead.
|
---|
| 247 | */
|
---|
| 248 | return (sum == 0 ? 1 : sum);
|
---|
| 249 | }
|
---|
| 250 |
|
---|
| 251 |
|
---|
| 252 |
|
---|
| 253 | /*
|
---|
| 254 | * Check to see if the given files differ.
|
---|
| 255 | * Returns 0 if they are the same, 1 if different, and -1 on error.
|
---|
| 256 | */
|
---|
| 257 | static int files_differ(FILE * f1, FILE * f2, int flags)
|
---|
| 258 | {
|
---|
| 259 | char buf1[BUFSIZ], buf2[BUFSIZ];
|
---|
| 260 | size_t i, j;
|
---|
| 261 |
|
---|
| 262 | if ((flags & (D_EMPTY1 | D_EMPTY2)) || stb1.st_size != stb2.st_size ||
|
---|
| 263 | (stb1.st_mode & S_IFMT) != (stb2.st_mode & S_IFMT))
|
---|
| 264 | return (1);
|
---|
| 265 | while (1) {
|
---|
| 266 | i = fread(buf1, 1, sizeof(buf1), f1);
|
---|
| 267 | j = fread(buf2, 1, sizeof(buf2), f2);
|
---|
| 268 | if (i != j)
|
---|
| 269 | return (1);
|
---|
| 270 | if (i == 0 && j == 0) {
|
---|
| 271 | if (ferror(f1) || ferror(f2))
|
---|
| 272 | return (1);
|
---|
| 273 | return (0);
|
---|
| 274 | }
|
---|
| 275 | if (memcmp(buf1, buf2, i) != 0)
|
---|
| 276 | return (1);
|
---|
| 277 | }
|
---|
| 278 | }
|
---|
| 279 |
|
---|
| 280 | static void prepare(int i, FILE * fd, off_t filesize)
|
---|
| 281 | {
|
---|
| 282 | struct line *p;
|
---|
| 283 | int h;
|
---|
| 284 | size_t j, sz;
|
---|
| 285 |
|
---|
| 286 | rewind(fd);
|
---|
| 287 |
|
---|
| 288 | sz = (filesize <= FSIZE_MAX ? filesize : FSIZE_MAX) / 25;
|
---|
| 289 | if (sz < 100)
|
---|
| 290 | sz = 100;
|
---|
| 291 |
|
---|
| 292 | p = xmalloc((sz + 3) * sizeof(struct line));
|
---|
| 293 | for (j = 0; (h = readhash(fd));) {
|
---|
| 294 | if (j == sz) {
|
---|
| 295 | sz = sz * 3 / 2;
|
---|
| 296 | p = xrealloc(p, (sz + 3) * sizeof(struct line));
|
---|
| 297 | }
|
---|
| 298 | p[++j].value = h;
|
---|
| 299 | }
|
---|
| 300 | len[i] = j;
|
---|
| 301 | file[i] = p;
|
---|
| 302 | }
|
---|
| 303 |
|
---|
| 304 | static void prune(void)
|
---|
| 305 | {
|
---|
| 306 | int i, j;
|
---|
| 307 |
|
---|
| 308 | for (pref = 0; pref < len[0] && pref < len[1] &&
|
---|
| 309 | file[0][pref + 1].value == file[1][pref + 1].value; pref++);
|
---|
| 310 | for (suff = 0; suff < len[0] - pref && suff < len[1] - pref &&
|
---|
| 311 | file[0][len[0] - suff].value == file[1][len[1] - suff].value;
|
---|
| 312 | suff++);
|
---|
| 313 | for (j = 0; j < 2; j++) {
|
---|
| 314 | sfile[j] = file[j] + pref;
|
---|
| 315 | slen[j] = len[j] - pref - suff;
|
---|
| 316 | for (i = 0; i <= slen[j]; i++)
|
---|
| 317 | sfile[j][i].serial = i;
|
---|
| 318 | }
|
---|
| 319 | }
|
---|
| 320 |
|
---|
| 321 | static void equiv(struct line *a, int n, struct line *b, int m, int *c)
|
---|
| 322 | {
|
---|
| 323 | int i, j;
|
---|
| 324 |
|
---|
| 325 | i = j = 1;
|
---|
| 326 | while (i <= n && j <= m) {
|
---|
| 327 | if (a[i].value < b[j].value)
|
---|
| 328 | a[i++].value = 0;
|
---|
| 329 | else if (a[i].value == b[j].value)
|
---|
| 330 | a[i++].value = j;
|
---|
| 331 | else
|
---|
| 332 | j++;
|
---|
| 333 | }
|
---|
| 334 | while (i <= n)
|
---|
| 335 | a[i++].value = 0;
|
---|
| 336 | b[m + 1].value = 0;
|
---|
| 337 | j = 0;
|
---|
| 338 | while (++j <= m) {
|
---|
| 339 | c[j] = -b[j].serial;
|
---|
| 340 | while (b[j + 1].value == b[j].value) {
|
---|
| 341 | j++;
|
---|
| 342 | c[j] = b[j].serial;
|
---|
| 343 | }
|
---|
| 344 | }
|
---|
| 345 | c[j] = -1;
|
---|
| 346 | }
|
---|
| 347 |
|
---|
| 348 | static int isqrt(int n)
|
---|
| 349 | {
|
---|
| 350 | int y, x = 1;
|
---|
| 351 |
|
---|
| 352 | if (n == 0)
|
---|
| 353 | return (0);
|
---|
| 354 |
|
---|
| 355 | do {
|
---|
| 356 | y = x;
|
---|
| 357 | x = n / x;
|
---|
| 358 | x += y;
|
---|
| 359 | x /= 2;
|
---|
| 360 | } while ((x - y) > 1 || (x - y) < -1);
|
---|
| 361 |
|
---|
| 362 | return (x);
|
---|
| 363 | }
|
---|
| 364 |
|
---|
| 365 | static int newcand(int x, int y, int pred)
|
---|
| 366 | {
|
---|
| 367 | struct cand *q;
|
---|
| 368 |
|
---|
| 369 | if (clen == clistlen) {
|
---|
| 370 | clistlen = clistlen * 11 / 10;
|
---|
| 371 | clist = xrealloc(clist, clistlen * sizeof(struct cand));
|
---|
| 372 | }
|
---|
| 373 | q = clist + clen;
|
---|
| 374 | q->x = x;
|
---|
| 375 | q->y = y;
|
---|
| 376 | q->pred = pred;
|
---|
| 377 | return (clen++);
|
---|
| 378 | }
|
---|
| 379 |
|
---|
| 380 |
|
---|
| 381 | static int search(int *c, int k, int y)
|
---|
| 382 | {
|
---|
| 383 | int i, j, l, t;
|
---|
| 384 |
|
---|
| 385 | if (clist[c[k]].y < y) /* quick look for typical case */
|
---|
| 386 | return (k + 1);
|
---|
| 387 | i = 0;
|
---|
| 388 | j = k + 1;
|
---|
| 389 | while (1) {
|
---|
| 390 | l = i + j;
|
---|
| 391 | if ((l >>= 1) <= i)
|
---|
| 392 | break;
|
---|
| 393 | t = clist[c[l]].y;
|
---|
| 394 | if (t > y)
|
---|
| 395 | j = l;
|
---|
| 396 | else if (t < y)
|
---|
| 397 | i = l;
|
---|
| 398 | else
|
---|
| 399 | return (l);
|
---|
| 400 | }
|
---|
| 401 | return (l + 1);
|
---|
| 402 | }
|
---|
| 403 |
|
---|
| 404 |
|
---|
| 405 | static int stone(int *a, int n, int *b, int *c)
|
---|
| 406 | {
|
---|
| 407 | int i, k, y, j, l;
|
---|
| 408 | int oldc, tc, oldl;
|
---|
| 409 | unsigned int numtries;
|
---|
| 410 |
|
---|
| 411 | #if ENABLE_FEATURE_DIFF_MINIMAL
|
---|
| 412 | const unsigned int bound =
|
---|
| 413 | (cmd_flags & FLAG_d) ? UINT_MAX : MAX(256, isqrt(n));
|
---|
| 414 | #else
|
---|
| 415 | const unsigned int bound = MAX(256, isqrt(n));
|
---|
| 416 | #endif
|
---|
| 417 | k = 0;
|
---|
| 418 | c[0] = newcand(0, 0, 0);
|
---|
| 419 | for (i = 1; i <= n; i++) {
|
---|
| 420 | j = a[i];
|
---|
| 421 | if (j == 0)
|
---|
| 422 | continue;
|
---|
| 423 | y = -b[j];
|
---|
| 424 | oldl = 0;
|
---|
| 425 | oldc = c[0];
|
---|
| 426 | numtries = 0;
|
---|
| 427 | do {
|
---|
| 428 | if (y <= clist[oldc].y)
|
---|
| 429 | continue;
|
---|
| 430 | l = search(c, k, y);
|
---|
| 431 | if (l != oldl + 1)
|
---|
| 432 | oldc = c[l - 1];
|
---|
| 433 | if (l <= k) {
|
---|
| 434 | if (clist[c[l]].y <= y)
|
---|
| 435 | continue;
|
---|
| 436 | tc = c[l];
|
---|
| 437 | c[l] = newcand(i, y, oldc);
|
---|
| 438 | oldc = tc;
|
---|
| 439 | oldl = l;
|
---|
| 440 | numtries++;
|
---|
| 441 | } else {
|
---|
| 442 | c[l] = newcand(i, y, oldc);
|
---|
| 443 | k++;
|
---|
| 444 | break;
|
---|
| 445 | }
|
---|
| 446 | } while ((y = b[++j]) > 0 && numtries < bound);
|
---|
| 447 | }
|
---|
| 448 | return (k);
|
---|
| 449 | }
|
---|
| 450 |
|
---|
| 451 | static void unravel(int p)
|
---|
| 452 | {
|
---|
| 453 | struct cand *q;
|
---|
| 454 | int i;
|
---|
| 455 |
|
---|
| 456 | for (i = 0; i <= len[0]; i++)
|
---|
| 457 | J[i] = i <= pref ? i : i > len[0] - suff ? i + len[1] - len[0] : 0;
|
---|
| 458 | for (q = clist + p; q->y != 0; q = clist + q->pred)
|
---|
| 459 | J[q->x + pref] = q->y + pref;
|
---|
| 460 | }
|
---|
| 461 |
|
---|
| 462 |
|
---|
| 463 | static void unsort(struct line *f, int l, int *b)
|
---|
| 464 | {
|
---|
| 465 | int *a, i;
|
---|
| 466 |
|
---|
| 467 | a = xmalloc((l + 1) * sizeof(int));
|
---|
| 468 | for (i = 1; i <= l; i++)
|
---|
| 469 | a[f[i].serial] = f[i].value;
|
---|
| 470 | for (i = 1; i <= l; i++)
|
---|
| 471 | b[i] = a[i];
|
---|
| 472 | free(a);
|
---|
| 473 | }
|
---|
| 474 |
|
---|
| 475 | static int skipline(FILE * f)
|
---|
| 476 | {
|
---|
| 477 | int i, c;
|
---|
| 478 |
|
---|
| 479 | for (i = 1; (c = getc(f)) != '\n' && c != EOF; i++)
|
---|
| 480 | continue;
|
---|
| 481 | return (i);
|
---|
| 482 | }
|
---|
| 483 |
|
---|
| 484 |
|
---|
| 485 | /*
|
---|
| 486 | * Check does double duty:
|
---|
| 487 | * 1. ferret out any fortuitous correspondences due
|
---|
| 488 | * to confounding by hashing (which result in "jackpot")
|
---|
| 489 | * 2. collect random access indexes to the two files
|
---|
| 490 | */
|
---|
| 491 | static void check(FILE * f1, FILE * f2)
|
---|
| 492 | {
|
---|
| 493 | int i, j, jackpot, c, d;
|
---|
| 494 | long ctold, ctnew;
|
---|
| 495 |
|
---|
| 496 | rewind(f1);
|
---|
| 497 | rewind(f2);
|
---|
| 498 | j = 1;
|
---|
| 499 | ixold[0] = ixnew[0] = 0;
|
---|
| 500 | jackpot = 0;
|
---|
| 501 | ctold = ctnew = 0;
|
---|
| 502 | for (i = 1; i <= len[0]; i++) {
|
---|
| 503 | if (J[i] == 0) {
|
---|
| 504 | ixold[i] = ctold += skipline(f1);
|
---|
| 505 | continue;
|
---|
| 506 | }
|
---|
| 507 | while (j < J[i]) {
|
---|
| 508 | ixnew[j] = ctnew += skipline(f2);
|
---|
| 509 | j++;
|
---|
| 510 | }
|
---|
| 511 | if ((cmd_flags & FLAG_b) || (cmd_flags & FLAG_w)
|
---|
| 512 | || (cmd_flags & FLAG_i)) {
|
---|
| 513 | while (1) {
|
---|
| 514 | c = getc(f1);
|
---|
| 515 | d = getc(f2);
|
---|
| 516 | /*
|
---|
| 517 | * GNU diff ignores a missing newline
|
---|
| 518 | * in one file if bflag || wflag.
|
---|
| 519 | */
|
---|
| 520 | if (((cmd_flags & FLAG_b) || (cmd_flags & FLAG_w)) &&
|
---|
| 521 | ((c == EOF && d == '\n') || (c == '\n' && d == EOF))) {
|
---|
| 522 | break;
|
---|
| 523 | }
|
---|
| 524 | ctold++;
|
---|
| 525 | ctnew++;
|
---|
| 526 | if ((cmd_flags & FLAG_b) && isspace(c) && isspace(d)) {
|
---|
| 527 | do {
|
---|
| 528 | if (c == '\n')
|
---|
| 529 | break;
|
---|
| 530 | ctold++;
|
---|
| 531 | } while (isspace(c = getc(f1)));
|
---|
| 532 | do {
|
---|
| 533 | if (d == '\n')
|
---|
| 534 | break;
|
---|
| 535 | ctnew++;
|
---|
| 536 | } while (isspace(d = getc(f2)));
|
---|
| 537 | } else if (cmd_flags & FLAG_w) {
|
---|
| 538 | while (isspace(c) && c != '\n') {
|
---|
| 539 | c = getc(f1);
|
---|
| 540 | ctold++;
|
---|
| 541 | }
|
---|
| 542 | while (isspace(d) && d != '\n') {
|
---|
| 543 | d = getc(f2);
|
---|
| 544 | ctnew++;
|
---|
| 545 | }
|
---|
| 546 | }
|
---|
| 547 | if (c != d) {
|
---|
| 548 | jackpot++;
|
---|
| 549 | J[i] = 0;
|
---|
| 550 | if (c != '\n' && c != EOF)
|
---|
| 551 | ctold += skipline(f1);
|
---|
| 552 | if (d != '\n' && c != EOF)
|
---|
| 553 | ctnew += skipline(f2);
|
---|
| 554 | break;
|
---|
| 555 | }
|
---|
| 556 | if (c == '\n' || c == EOF)
|
---|
| 557 | break;
|
---|
| 558 | }
|
---|
| 559 | } else {
|
---|
| 560 | while (1) {
|
---|
| 561 | ctold++;
|
---|
| 562 | ctnew++;
|
---|
| 563 | if ((c = getc(f1)) != (d = getc(f2))) {
|
---|
| 564 | J[i] = 0;
|
---|
| 565 | if (c != '\n' && c != EOF)
|
---|
| 566 | ctold += skipline(f1);
|
---|
| 567 | if (d != '\n' && c != EOF)
|
---|
| 568 | ctnew += skipline(f2);
|
---|
| 569 | break;
|
---|
| 570 | }
|
---|
| 571 | if (c == '\n' || c == EOF)
|
---|
| 572 | break;
|
---|
| 573 | }
|
---|
| 574 | }
|
---|
| 575 | ixold[i] = ctold;
|
---|
| 576 | ixnew[j] = ctnew;
|
---|
| 577 | j++;
|
---|
| 578 | }
|
---|
| 579 | for (; j <= len[1]; j++)
|
---|
| 580 | ixnew[j] = ctnew += skipline(f2);
|
---|
| 581 | }
|
---|
| 582 |
|
---|
| 583 | /* shellsort CACM #201 */
|
---|
| 584 | static void sort(struct line *a, int n)
|
---|
| 585 | {
|
---|
| 586 | struct line *ai, *aim, w;
|
---|
| 587 | int j, m = 0, k;
|
---|
| 588 |
|
---|
| 589 | if (n == 0)
|
---|
| 590 | return;
|
---|
| 591 | for (j = 1; j <= n; j *= 2)
|
---|
| 592 | m = 2 * j - 1;
|
---|
| 593 | for (m /= 2; m != 0; m /= 2) {
|
---|
| 594 | k = n - m;
|
---|
| 595 | for (j = 1; j <= k; j++) {
|
---|
| 596 | for (ai = &a[j]; ai > a; ai -= m) {
|
---|
| 597 | aim = &ai[m];
|
---|
| 598 | if (aim < ai)
|
---|
| 599 | break; /* wraparound */
|
---|
| 600 | if (aim->value > ai[0].value ||
|
---|
| 601 | (aim->value == ai[0].value && aim->serial > ai[0].serial))
|
---|
| 602 | break;
|
---|
| 603 | w.value = ai[0].value;
|
---|
| 604 | ai[0].value = aim->value;
|
---|
| 605 | aim->value = w.value;
|
---|
| 606 | w.serial = ai[0].serial;
|
---|
| 607 | ai[0].serial = aim->serial;
|
---|
| 608 | aim->serial = w.serial;
|
---|
| 609 | }
|
---|
| 610 | }
|
---|
| 611 | }
|
---|
| 612 | }
|
---|
| 613 |
|
---|
| 614 |
|
---|
| 615 | static void uni_range(int a, int b)
|
---|
| 616 | {
|
---|
| 617 | if (a < b)
|
---|
| 618 | printf("%d,%d", a, b - a + 1);
|
---|
| 619 | else if (a == b)
|
---|
| 620 | printf("%d", b);
|
---|
| 621 | else
|
---|
| 622 | printf("%d,0", b);
|
---|
| 623 | }
|
---|
| 624 |
|
---|
| 625 | static int fetch(long *f, int a, int b, FILE * lb, int ch)
|
---|
| 626 | {
|
---|
| 627 | int i, j, c, lastc, col, nc;
|
---|
| 628 |
|
---|
| 629 | if (a > b)
|
---|
| 630 | return (0);
|
---|
| 631 | for (i = a; i <= b; i++) {
|
---|
| 632 | fseek(lb, f[i - 1], SEEK_SET);
|
---|
| 633 | nc = f[i] - f[i - 1];
|
---|
| 634 | if (ch != '\0') {
|
---|
| 635 | putchar(ch);
|
---|
| 636 | if (cmd_flags & FLAG_T)
|
---|
| 637 | putchar('\t');
|
---|
| 638 | }
|
---|
| 639 | col = 0;
|
---|
| 640 | for (j = 0, lastc = '\0'; j < nc; j++, lastc = c) {
|
---|
| 641 | if ((c = getc(lb)) == EOF) {
|
---|
| 642 | puts("\n\\ No newline at end of file");
|
---|
| 643 | return (0);
|
---|
| 644 | }
|
---|
| 645 | if (c == '\t' && (cmd_flags & FLAG_t)) {
|
---|
| 646 | do {
|
---|
| 647 | putchar(' ');
|
---|
| 648 | } while (++col & 7);
|
---|
| 649 | } else {
|
---|
| 650 | putchar(c);
|
---|
| 651 | col++;
|
---|
| 652 | }
|
---|
| 653 | }
|
---|
| 654 | }
|
---|
| 655 | return (0);
|
---|
| 656 | }
|
---|
| 657 |
|
---|
| 658 | static int asciifile(FILE * f)
|
---|
| 659 | {
|
---|
| 660 | #if ENABLE_FEATURE_DIFF_BINARY
|
---|
| 661 | unsigned char buf[BUFSIZ];
|
---|
| 662 | int i, cnt;
|
---|
| 663 | #endif
|
---|
| 664 |
|
---|
| 665 | if ((cmd_flags & FLAG_a) || f == NULL)
|
---|
| 666 | return (1);
|
---|
| 667 |
|
---|
| 668 | #if ENABLE_FEATURE_DIFF_BINARY
|
---|
| 669 | rewind(f);
|
---|
| 670 | cnt = fread(buf, 1, sizeof(buf), f);
|
---|
| 671 | for (i = 0; i < cnt; i++) {
|
---|
| 672 | if (!isprint(buf[i]) && !isspace(buf[i])) {
|
---|
| 673 | return (0);
|
---|
| 674 | }
|
---|
| 675 | }
|
---|
| 676 | #endif
|
---|
| 677 | return (1);
|
---|
| 678 | }
|
---|
| 679 |
|
---|
| 680 | /* dump accumulated "unified" diff changes */
|
---|
| 681 | static void dump_unified_vec(FILE * f1, FILE * f2)
|
---|
| 682 | {
|
---|
| 683 | struct context_vec *cvp = context_vec_start;
|
---|
| 684 | int lowa, upb, lowc, upd;
|
---|
| 685 | int a, b, c, d;
|
---|
| 686 | char ch;
|
---|
| 687 |
|
---|
| 688 | if (context_vec_start > context_vec_ptr)
|
---|
| 689 | return;
|
---|
| 690 |
|
---|
| 691 | b = d = 0; /* gcc */
|
---|
| 692 | lowa = MAX(1, cvp->a - context);
|
---|
| 693 | upb = MIN(len[0], context_vec_ptr->b + context);
|
---|
| 694 | lowc = MAX(1, cvp->c - context);
|
---|
| 695 | upd = MIN(len[1], context_vec_ptr->d + context);
|
---|
| 696 |
|
---|
| 697 | fputs("@@ -", stdout);
|
---|
| 698 | uni_range(lowa, upb);
|
---|
| 699 | fputs(" +", stdout);
|
---|
| 700 | uni_range(lowc, upd);
|
---|
| 701 | fputs(" @@", stdout);
|
---|
| 702 | putchar('\n');
|
---|
| 703 |
|
---|
| 704 | /*
|
---|
| 705 | * Output changes in "unified" diff format--the old and new lines
|
---|
| 706 | * are printed together.
|
---|
| 707 | */
|
---|
| 708 | for (; cvp <= context_vec_ptr; cvp++) {
|
---|
| 709 | a = cvp->a;
|
---|
| 710 | b = cvp->b;
|
---|
| 711 | c = cvp->c;
|
---|
| 712 | d = cvp->d;
|
---|
| 713 |
|
---|
| 714 | /*
|
---|
| 715 | * c: both new and old changes
|
---|
| 716 | * d: only changes in the old file
|
---|
| 717 | * a: only changes in the new file
|
---|
| 718 | */
|
---|
| 719 | if (a <= b && c <= d)
|
---|
| 720 | ch = 'c';
|
---|
| 721 | else
|
---|
| 722 | ch = (a <= b) ? 'd' : 'a';
|
---|
| 723 | #if 0
|
---|
| 724 | switch (ch) {
|
---|
| 725 | case 'c':
|
---|
| 726 | fetch(ixold, lowa, a - 1, f1, ' ');
|
---|
| 727 | fetch(ixold, a, b, f1, '-');
|
---|
| 728 | fetch(ixnew, c, d, f2, '+');
|
---|
| 729 | break;
|
---|
| 730 | case 'd':
|
---|
| 731 | fetch(ixold, lowa, a - 1, f1, ' ');
|
---|
| 732 | fetch(ixold, a, b, f1, '-');
|
---|
| 733 | break;
|
---|
| 734 | case 'a':
|
---|
| 735 | fetch(ixnew, lowc, c - 1, f2, ' ');
|
---|
| 736 | fetch(ixnew, c, d, f2, '+');
|
---|
| 737 | break;
|
---|
| 738 | }
|
---|
| 739 | #else
|
---|
| 740 | if (ch == 'c' || ch == 'd') {
|
---|
| 741 | fetch(ixold, lowa, a - 1, f1, ' ');
|
---|
| 742 | fetch(ixold, a, b, f1, '-');
|
---|
| 743 | }
|
---|
| 744 | if (ch == 'a')
|
---|
| 745 | fetch(ixnew, lowc, c - 1, f2, ' ');
|
---|
| 746 | if (ch == 'c' || ch == 'a')
|
---|
| 747 | fetch(ixnew, c, d, f2, '+');
|
---|
| 748 | #endif
|
---|
| 749 | lowa = b + 1;
|
---|
| 750 | lowc = d + 1;
|
---|
| 751 | }
|
---|
| 752 | fetch(ixnew, d + 1, upd, f2, ' ');
|
---|
| 753 |
|
---|
| 754 | context_vec_ptr = context_vec_start - 1;
|
---|
| 755 | }
|
---|
| 756 |
|
---|
| 757 |
|
---|
| 758 | static void print_header(const char *file1, const char *file2)
|
---|
| 759 | {
|
---|
| 760 | if (label[0] != NULL)
|
---|
| 761 | printf("%s %s\n", "---", label[0]);
|
---|
| 762 | else
|
---|
| 763 | printf("%s %s\t%s", "---", file1, ctime(&stb1.st_mtime));
|
---|
| 764 | if (label[1] != NULL)
|
---|
| 765 | printf("%s %s\n", "+++", label[1]);
|
---|
| 766 | else
|
---|
| 767 | printf("%s %s\t%s", "+++", file2, ctime(&stb2.st_mtime));
|
---|
| 768 | }
|
---|
| 769 |
|
---|
| 770 |
|
---|
| 771 |
|
---|
| 772 | /*
|
---|
| 773 | * Indicate that there is a difference between lines a and b of the from file
|
---|
| 774 | * to get to lines c to d of the to file. If a is greater then b then there
|
---|
| 775 | * are no lines in the from file involved and this means that there were
|
---|
| 776 | * lines appended (beginning at b). If c is greater than d then there are
|
---|
| 777 | * lines missing from the to file.
|
---|
| 778 | */
|
---|
| 779 | static void change(char *file1, FILE * f1, char *file2, FILE * f2, int a,
|
---|
| 780 | int b, int c, int d)
|
---|
| 781 | {
|
---|
| 782 | static size_t max_context = 64;
|
---|
| 783 |
|
---|
| 784 | if (a > b && c > d)
|
---|
| 785 | return;
|
---|
| 786 | if (cmd_flags & FLAG_q)
|
---|
| 787 | return;
|
---|
| 788 |
|
---|
| 789 | /*
|
---|
| 790 | * Allocate change records as needed.
|
---|
| 791 | */
|
---|
| 792 | if (context_vec_ptr == context_vec_end - 1) {
|
---|
| 793 | ptrdiff_t offset = context_vec_ptr - context_vec_start;
|
---|
| 794 |
|
---|
| 795 | max_context <<= 1;
|
---|
| 796 | context_vec_start = xrealloc(context_vec_start,
|
---|
| 797 | max_context *
|
---|
| 798 | sizeof(struct context_vec));
|
---|
| 799 | context_vec_end = context_vec_start + max_context;
|
---|
| 800 | context_vec_ptr = context_vec_start + offset;
|
---|
| 801 | }
|
---|
| 802 | if (anychange == 0) {
|
---|
| 803 | /*
|
---|
| 804 | * Print the context/unidiff header first time through.
|
---|
| 805 | */
|
---|
| 806 | print_header(file1, file2);
|
---|
| 807 | anychange = 1;
|
---|
| 808 | } else if (a > context_vec_ptr->b + (2 * context) + 1 &&
|
---|
| 809 | c > context_vec_ptr->d + (2 * context) + 1) {
|
---|
| 810 | /*
|
---|
| 811 | * If this change is more than 'context' lines from the
|
---|
| 812 | * previous change, dump the record and reset it.
|
---|
| 813 | */
|
---|
| 814 | dump_unified_vec(f1, f2);
|
---|
| 815 | }
|
---|
| 816 | context_vec_ptr++;
|
---|
| 817 | context_vec_ptr->a = a;
|
---|
| 818 | context_vec_ptr->b = b;
|
---|
| 819 | context_vec_ptr->c = c;
|
---|
| 820 | context_vec_ptr->d = d;
|
---|
| 821 | return;
|
---|
| 822 |
|
---|
| 823 | }
|
---|
| 824 |
|
---|
| 825 |
|
---|
| 826 | static void output(char *file1, FILE * f1, char *file2, FILE * f2)
|
---|
| 827 | {
|
---|
| 828 |
|
---|
| 829 | /* Note that j0 and j1 can't be used as they are defined in math.h.
|
---|
| 830 | * This also allows the rather amusing variable 'j00'... */
|
---|
| 831 | int m, i0, i1, j00, j01;
|
---|
| 832 |
|
---|
| 833 | rewind(f1);
|
---|
| 834 | rewind(f2);
|
---|
| 835 | m = len[0];
|
---|
| 836 | J[0] = 0;
|
---|
| 837 | J[m + 1] = len[1] + 1;
|
---|
| 838 | for (i0 = 1; i0 <= m; i0 = i1 + 1) {
|
---|
| 839 | while (i0 <= m && J[i0] == J[i0 - 1] + 1)
|
---|
| 840 | i0++;
|
---|
| 841 | j00 = J[i0 - 1] + 1;
|
---|
| 842 | i1 = i0 - 1;
|
---|
| 843 | while (i1 < m && J[i1 + 1] == 0)
|
---|
| 844 | i1++;
|
---|
| 845 | j01 = J[i1 + 1] - 1;
|
---|
| 846 | J[i1] = j01;
|
---|
| 847 | change(file1, f1, file2, f2, i0, i1, j00, j01);
|
---|
| 848 | }
|
---|
| 849 | if (m == 0) {
|
---|
| 850 | change(file1, f1, file2, f2, 1, 0, 1, len[1]);
|
---|
| 851 | }
|
---|
| 852 | if (anychange != 0) {
|
---|
| 853 | dump_unified_vec(f1, f2);
|
---|
| 854 | }
|
---|
| 855 | }
|
---|
| 856 |
|
---|
| 857 | /*
|
---|
| 858 | * The following code uses an algorithm due to Harold Stone,
|
---|
| 859 | * which finds a pair of longest identical subsequences in
|
---|
| 860 | * the two files.
|
---|
| 861 | *
|
---|
| 862 | * The major goal is to generate the match vector J.
|
---|
| 863 | * J[i] is the index of the line in file1 corresponding
|
---|
| 864 | * to line i file0. J[i] = 0 if there is no
|
---|
| 865 | * such line in file1.
|
---|
| 866 | *
|
---|
| 867 | * Lines are hashed so as to work in core. All potential
|
---|
| 868 | * matches are located by sorting the lines of each file
|
---|
| 869 | * on the hash (called ``value''). In particular, this
|
---|
| 870 | * collects the equivalence classes in file1 together.
|
---|
| 871 | * Subroutine equiv replaces the value of each line in
|
---|
| 872 | * file0 by the index of the first element of its
|
---|
| 873 | * matching equivalence in (the reordered) file1.
|
---|
| 874 | * To save space equiv squeezes file1 into a single
|
---|
| 875 | * array member in which the equivalence classes
|
---|
| 876 | * are simply concatenated, except that their first
|
---|
| 877 | * members are flagged by changing sign.
|
---|
| 878 | *
|
---|
| 879 | * Next the indices that point into member are unsorted into
|
---|
| 880 | * array class according to the original order of file0.
|
---|
| 881 | *
|
---|
| 882 | * The cleverness lies in routine stone. This marches
|
---|
| 883 | * through the lines of file0, developing a vector klist
|
---|
| 884 | * of "k-candidates". At step i a k-candidate is a matched
|
---|
| 885 | * pair of lines x,y (x in file0 y in file1) such that
|
---|
| 886 | * there is a common subsequence of length k
|
---|
| 887 | * between the first i lines of file0 and the first y
|
---|
| 888 | * lines of file1, but there is no such subsequence for
|
---|
| 889 | * any smaller y. x is the earliest possible mate to y
|
---|
| 890 | * that occurs in such a subsequence.
|
---|
| 891 | *
|
---|
| 892 | * Whenever any of the members of the equivalence class of
|
---|
| 893 | * lines in file1 matable to a line in file0 has serial number
|
---|
| 894 | * less than the y of some k-candidate, that k-candidate
|
---|
| 895 | * with the smallest such y is replaced. The new
|
---|
| 896 | * k-candidate is chained (via pred) to the current
|
---|
| 897 | * k-1 candidate so that the actual subsequence can
|
---|
| 898 | * be recovered. When a member has serial number greater
|
---|
| 899 | * that the y of all k-candidates, the klist is extended.
|
---|
| 900 | * At the end, the longest subsequence is pulled out
|
---|
| 901 | * and placed in the array J by unravel
|
---|
| 902 | *
|
---|
| 903 | * With J in hand, the matches there recorded are
|
---|
| 904 | * checked against reality to assure that no spurious
|
---|
| 905 | * matches have crept in due to hashing. If they have,
|
---|
| 906 | * they are broken, and "jackpot" is recorded--a harmless
|
---|
| 907 | * matter except that a true match for a spuriously
|
---|
| 908 | * mated line may now be unnecessarily reported as a change.
|
---|
| 909 | *
|
---|
| 910 | * Much of the complexity of the program comes simply
|
---|
| 911 | * from trying to minimize core utilization and
|
---|
| 912 | * maximize the range of doable problems by dynamically
|
---|
| 913 | * allocating what is needed and reusing what is not.
|
---|
| 914 | * The core requirements for problems larger than somewhat
|
---|
| 915 | * are (in words) 2*length(file0) + length(file1) +
|
---|
| 916 | * 3*(number of k-candidates installed), typically about
|
---|
| 917 | * 6n words for files of length n.
|
---|
| 918 | */
|
---|
| 919 |
|
---|
| 920 | static int diffreg(char *ofile1, char *ofile2, int flags)
|
---|
| 921 | {
|
---|
| 922 | char *file1 = ofile1;
|
---|
| 923 | char *file2 = ofile2;
|
---|
| 924 | FILE *f1 = NULL;
|
---|
| 925 | FILE *f2 = NULL;
|
---|
| 926 | int rval = D_SAME;
|
---|
| 927 | int i;
|
---|
| 928 |
|
---|
| 929 | anychange = 0;
|
---|
| 930 | context_vec_ptr = context_vec_start - 1;
|
---|
| 931 |
|
---|
| 932 | if (S_ISDIR(stb1.st_mode) != S_ISDIR(stb2.st_mode))
|
---|
| 933 | return (S_ISDIR(stb1.st_mode) ? D_MISMATCH1 : D_MISMATCH2);
|
---|
| 934 | if (strcmp(file1, "-") == 0 && strcmp(file2, "-") == 0)
|
---|
| 935 | goto closem;
|
---|
| 936 |
|
---|
| 937 | if (flags & D_EMPTY1)
|
---|
| 938 | f1 = bb_xfopen(bb_dev_null, "r");
|
---|
| 939 | else {
|
---|
| 940 | if (strcmp(file1, "-") == 0)
|
---|
| 941 | f1 = stdin;
|
---|
| 942 | else
|
---|
| 943 | f1 = bb_xfopen(file1, "r");
|
---|
| 944 | }
|
---|
| 945 |
|
---|
| 946 | if (flags & D_EMPTY2)
|
---|
| 947 | f2 = bb_xfopen(bb_dev_null, "r");
|
---|
| 948 | else {
|
---|
| 949 | if (strcmp(file2, "-") == 0)
|
---|
| 950 | f2 = stdin;
|
---|
| 951 | else
|
---|
| 952 | f2 = bb_xfopen(file2, "r");
|
---|
| 953 | }
|
---|
| 954 |
|
---|
| 955 | if ((i = files_differ(f1, f2, flags)) == 0)
|
---|
| 956 | goto closem;
|
---|
| 957 | else if (i != 1) { /* 1 == ok */
|
---|
| 958 | /* error */
|
---|
| 959 | status |= 2;
|
---|
| 960 | goto closem;
|
---|
| 961 | }
|
---|
| 962 |
|
---|
| 963 | if (!asciifile(f1) || !asciifile(f2)) {
|
---|
| 964 | rval = D_BINARY;
|
---|
| 965 | status |= 1;
|
---|
| 966 | goto closem;
|
---|
| 967 | }
|
---|
| 968 |
|
---|
| 969 | prepare(0, f1, stb1.st_size);
|
---|
| 970 | prepare(1, f2, stb2.st_size);
|
---|
| 971 | prune();
|
---|
| 972 | sort(sfile[0], slen[0]);
|
---|
| 973 | sort(sfile[1], slen[1]);
|
---|
| 974 |
|
---|
| 975 | member = (int *) file[1];
|
---|
| 976 | equiv(sfile[0], slen[0], sfile[1], slen[1], member);
|
---|
| 977 | member = xrealloc(member, (slen[1] + 2) * sizeof(int));
|
---|
| 978 |
|
---|
| 979 | class = (int *) file[0];
|
---|
| 980 | unsort(sfile[0], slen[0], class);
|
---|
| 981 | class = xrealloc(class, (slen[0] + 2) * sizeof(int));
|
---|
| 982 |
|
---|
| 983 | klist = xmalloc((slen[0] + 2) * sizeof(int));
|
---|
| 984 | clen = 0;
|
---|
| 985 | clistlen = 100;
|
---|
| 986 | clist = xmalloc(clistlen * sizeof(struct cand));
|
---|
| 987 | i = stone(class, slen[0], member, klist);
|
---|
| 988 | free(member);
|
---|
| 989 | free(class);
|
---|
| 990 |
|
---|
| 991 | J = xrealloc(J, (len[0] + 2) * sizeof(int));
|
---|
| 992 | unravel(klist[i]);
|
---|
| 993 | free(clist);
|
---|
| 994 | free(klist);
|
---|
| 995 |
|
---|
| 996 | ixold = xrealloc(ixold, (len[0] + 2) * sizeof(long));
|
---|
| 997 | ixnew = xrealloc(ixnew, (len[1] + 2) * sizeof(long));
|
---|
| 998 | check(f1, f2);
|
---|
| 999 | output(file1, f1, file2, f2);
|
---|
| 1000 |
|
---|
| 1001 | closem:
|
---|
| 1002 | if (anychange) {
|
---|
| 1003 | status |= 1;
|
---|
| 1004 | if (rval == D_SAME)
|
---|
| 1005 | rval = D_DIFFER;
|
---|
| 1006 | }
|
---|
| 1007 | if (f1 != NULL)
|
---|
| 1008 | fclose(f1);
|
---|
| 1009 | if (f2 != NULL)
|
---|
| 1010 | fclose(f2);
|
---|
| 1011 | if (file1 != ofile1)
|
---|
| 1012 | free(file1);
|
---|
| 1013 | if (file2 != ofile2)
|
---|
| 1014 | free(file2);
|
---|
| 1015 | return (rval);
|
---|
| 1016 | }
|
---|
| 1017 |
|
---|
| 1018 | #if ENABLE_FEATURE_DIFF_DIR
|
---|
| 1019 | static void do_diff(char *dir1, char *path1, char *dir2, char *path2)
|
---|
| 1020 | {
|
---|
| 1021 |
|
---|
| 1022 | int flags = D_HEADER;
|
---|
| 1023 | int val;
|
---|
| 1024 |
|
---|
| 1025 | char *fullpath1 = bb_xasprintf("%s/%s", dir1, path1);
|
---|
| 1026 | char *fullpath2 = bb_xasprintf("%s/%s", dir2, path2);
|
---|
| 1027 |
|
---|
| 1028 | if (stat(fullpath1, &stb1) != 0) {
|
---|
| 1029 | flags |= D_EMPTY1;
|
---|
| 1030 | memset(&stb1, 0, sizeof(stb1));
|
---|
| 1031 | fullpath1 = bb_xasprintf("%s/%s", dir1, path2);
|
---|
| 1032 | }
|
---|
| 1033 | if (stat(fullpath2, &stb2) != 0) {
|
---|
| 1034 | flags |= D_EMPTY2;
|
---|
| 1035 | memset(&stb2, 0, sizeof(stb2));
|
---|
| 1036 | stb2.st_mode = stb1.st_mode;
|
---|
| 1037 | fullpath2 = bb_xasprintf("%s/%s", dir2, path1);
|
---|
| 1038 | }
|
---|
| 1039 |
|
---|
| 1040 | if (stb1.st_mode == 0)
|
---|
| 1041 | stb1.st_mode = stb2.st_mode;
|
---|
| 1042 |
|
---|
| 1043 | if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) {
|
---|
| 1044 | printf("Common subdirectories: %s and %s\n", fullpath1, fullpath2);
|
---|
| 1045 | return;
|
---|
| 1046 | }
|
---|
| 1047 |
|
---|
| 1048 | if (!S_ISREG(stb1.st_mode) && !S_ISDIR(stb1.st_mode))
|
---|
| 1049 | val = D_SKIPPED1;
|
---|
| 1050 | else if (!S_ISREG(stb2.st_mode) && !S_ISDIR(stb2.st_mode))
|
---|
| 1051 | val = D_SKIPPED2;
|
---|
| 1052 | else
|
---|
| 1053 | val = diffreg(fullpath1, fullpath2, flags);
|
---|
| 1054 |
|
---|
| 1055 | print_status(val, fullpath1, fullpath2, NULL);
|
---|
| 1056 | }
|
---|
| 1057 | #endif
|
---|
| 1058 |
|
---|
| 1059 | #if ENABLE_FEATURE_DIFF_DIR
|
---|
| 1060 | static int dir_strcmp(const void *p1, const void *p2)
|
---|
| 1061 | {
|
---|
| 1062 | return strcmp(*(char *const *) p1, *(char *const *) p2);
|
---|
| 1063 | }
|
---|
| 1064 |
|
---|
| 1065 | /* This function adds a filename to dl, the directory listing. */
|
---|
| 1066 |
|
---|
| 1067 | static int add_to_dirlist(const char *filename,
|
---|
| 1068 | struct stat ATTRIBUTE_UNUSED * sb, void *userdata)
|
---|
| 1069 | {
|
---|
| 1070 | dl_count++;
|
---|
| 1071 | dl = xrealloc(dl, dl_count * sizeof(char *));
|
---|
| 1072 | dl[dl_count - 1] = bb_xstrdup(filename);
|
---|
| 1073 | if (cmd_flags & FLAG_r) {
|
---|
| 1074 | int *pp = (int *) userdata;
|
---|
| 1075 | int path_len = *pp + 1;
|
---|
| 1076 |
|
---|
| 1077 | dl[dl_count - 1] = &(dl[dl_count - 1])[path_len];
|
---|
| 1078 | }
|
---|
| 1079 | return TRUE;
|
---|
| 1080 | }
|
---|
| 1081 |
|
---|
| 1082 | /* This returns a sorted directory listing. */
|
---|
| 1083 | static char **get_dir(char *path)
|
---|
| 1084 | {
|
---|
| 1085 |
|
---|
| 1086 | int i;
|
---|
| 1087 | char **retval;
|
---|
| 1088 |
|
---|
| 1089 | /* If -r has been set, then the recursive_action function will be
|
---|
| 1090 | * used. Unfortunately, this outputs the root directory along with
|
---|
| 1091 | * the recursed paths, so use void *userdata to specify the string
|
---|
| 1092 | * length of the root directory. It can then be removed in
|
---|
| 1093 | * add_to_dirlist. */
|
---|
| 1094 |
|
---|
| 1095 | int path_len = strlen(path);
|
---|
| 1096 | void *userdata = &path_len;
|
---|
| 1097 |
|
---|
| 1098 | /* Reset dl_count - there's no need to free dl as bb_xrealloc does
|
---|
| 1099 | * the job nicely. */
|
---|
| 1100 | dl_count = 0;
|
---|
| 1101 |
|
---|
| 1102 | /* Now fill dl with a listing. */
|
---|
| 1103 | if (cmd_flags & FLAG_r)
|
---|
| 1104 | recursive_action(path, TRUE, TRUE, FALSE, add_to_dirlist, NULL,
|
---|
| 1105 | userdata);
|
---|
| 1106 | else {
|
---|
| 1107 | DIR *dp;
|
---|
| 1108 | struct dirent *ep;
|
---|
| 1109 |
|
---|
| 1110 | dp = bb_opendir(path);
|
---|
| 1111 | while ((ep = readdir(dp))) {
|
---|
| 1112 | if ((!strcmp(ep->d_name, "..")) || (!strcmp(ep->d_name, ".")))
|
---|
| 1113 | continue;
|
---|
| 1114 | add_to_dirlist(ep->d_name, NULL, NULL);
|
---|
| 1115 | }
|
---|
| 1116 | closedir(dp);
|
---|
| 1117 | }
|
---|
| 1118 |
|
---|
| 1119 | /* Sort dl alphabetically. */
|
---|
| 1120 | qsort(dl, dl_count, sizeof(char *), dir_strcmp);
|
---|
| 1121 |
|
---|
| 1122 | /* Copy dl so that we can return it. */
|
---|
| 1123 | retval = xmalloc(dl_count * sizeof(char *));
|
---|
| 1124 | for (i = 0; i < dl_count; i++)
|
---|
| 1125 | retval[i] = bb_xstrdup(dl[i]);
|
---|
| 1126 |
|
---|
| 1127 | return retval;
|
---|
| 1128 | }
|
---|
| 1129 |
|
---|
| 1130 | static void diffdir(char *p1, char *p2)
|
---|
| 1131 | {
|
---|
| 1132 |
|
---|
| 1133 | char **dirlist1, **dirlist2;
|
---|
| 1134 | char *dp1, *dp2;
|
---|
| 1135 | int dirlist1_count, dirlist2_count;
|
---|
| 1136 | int pos;
|
---|
| 1137 |
|
---|
| 1138 | /* Check for trailing slashes. */
|
---|
| 1139 |
|
---|
| 1140 | if (p1[strlen(p1) - 1] == '/')
|
---|
| 1141 | p1[strlen(p1) - 1] = '\0';
|
---|
| 1142 | if (p2[strlen(p2) - 1] == '/')
|
---|
| 1143 | p2[strlen(p2) - 1] = '\0';
|
---|
| 1144 |
|
---|
| 1145 | /* Get directory listings for p1 and p2. */
|
---|
| 1146 |
|
---|
| 1147 | dirlist1 = get_dir(p1);
|
---|
| 1148 | dirlist1_count = dl_count;
|
---|
| 1149 | dirlist1[dirlist1_count] = NULL;
|
---|
| 1150 | dirlist2 = get_dir(p2);
|
---|
| 1151 | dirlist2_count = dl_count;
|
---|
| 1152 | dirlist2[dirlist2_count] = NULL;
|
---|
| 1153 |
|
---|
| 1154 | /* If -S was set, find the starting point. */
|
---|
| 1155 | if (start) {
|
---|
| 1156 | while (*dirlist1 != NULL && strcmp(*dirlist1, start) < 0)
|
---|
| 1157 | dirlist1++;
|
---|
| 1158 | while (*dirlist2 != NULL && strcmp(*dirlist2, start) < 0)
|
---|
| 1159 | dirlist2++;
|
---|
| 1160 | if ((*dirlist1 == NULL) || (*dirlist2 == NULL))
|
---|
| 1161 | bb_error_msg(bb_msg_invalid_arg, "NULL", "-S");
|
---|
| 1162 | }
|
---|
| 1163 |
|
---|
| 1164 | /* Now that both dirlist1 and dirlist2 contain sorted directory
|
---|
| 1165 | * listings, we can start to go through dirlist1. If both listings
|
---|
| 1166 | * contain the same file, then do a normal diff. Otherwise, behaviour
|
---|
| 1167 | * is determined by whether the -N flag is set. */
|
---|
| 1168 | while (*dirlist1 != NULL || *dirlist2 != NULL) {
|
---|
| 1169 | dp1 = *dirlist1;
|
---|
| 1170 | dp2 = *dirlist2;
|
---|
| 1171 | pos = dp1 == NULL ? 1 : dp2 == NULL ? -1 : strcmp(dp1, dp2);
|
---|
| 1172 | if (pos == 0) {
|
---|
| 1173 | do_diff(p1, dp1, p2, dp2);
|
---|
| 1174 | dirlist1++;
|
---|
| 1175 | dirlist2++;
|
---|
| 1176 | } else if (pos < 0) {
|
---|
| 1177 | if (cmd_flags & FLAG_N)
|
---|
| 1178 | do_diff(p1, dp1, p2, NULL);
|
---|
| 1179 | else
|
---|
| 1180 | print_only(p1, strlen(p1) + 1, dp1);
|
---|
| 1181 | dirlist1++;
|
---|
| 1182 | } else {
|
---|
| 1183 | if (cmd_flags & FLAG_N)
|
---|
| 1184 | do_diff(p1, NULL, p2, dp2);
|
---|
| 1185 | else
|
---|
| 1186 | print_only(p2, strlen(p2) + 1, dp2);
|
---|
| 1187 | dirlist2++;
|
---|
| 1188 | }
|
---|
| 1189 | }
|
---|
| 1190 | }
|
---|
| 1191 | #endif
|
---|
| 1192 |
|
---|
| 1193 |
|
---|
| 1194 |
|
---|
| 1195 | int diff_main(int argc, char **argv)
|
---|
| 1196 | {
|
---|
| 1197 | int gotstdin = 0;
|
---|
| 1198 |
|
---|
| 1199 | char *U_opt;
|
---|
| 1200 | llist_t *L_arg = NULL;
|
---|
| 1201 |
|
---|
| 1202 | bb_opt_complementally = "L::";
|
---|
| 1203 | cmd_flags =
|
---|
| 1204 | bb_getopt_ulflags(argc, argv, "abdiL:NqrsS:tTU:wu", &L_arg, &start,
|
---|
| 1205 | &U_opt);
|
---|
| 1206 |
|
---|
| 1207 | if (cmd_flags & FLAG_L) {
|
---|
| 1208 | while (L_arg) {
|
---|
| 1209 | if (label[0] == NULL)
|
---|
| 1210 | label[0] = L_arg->data;
|
---|
| 1211 | else if (label[1] == NULL)
|
---|
| 1212 | label[1] = L_arg->data;
|
---|
| 1213 | else
|
---|
| 1214 | bb_show_usage();
|
---|
| 1215 |
|
---|
| 1216 | L_arg = L_arg->link;
|
---|
| 1217 | }
|
---|
| 1218 |
|
---|
| 1219 | /* If both label[0] and label[1] were set, they need to be swapped. */
|
---|
| 1220 | if (label[0] && label[1]) {
|
---|
| 1221 | char *tmp;
|
---|
| 1222 |
|
---|
| 1223 | tmp = label[1];
|
---|
| 1224 | label[1] = label[0];
|
---|
| 1225 | label[0] = tmp;
|
---|
| 1226 | }
|
---|
| 1227 | }
|
---|
| 1228 |
|
---|
| 1229 | context = 3; /* This is the default number of lines of context. */
|
---|
| 1230 | if (cmd_flags & FLAG_U) {
|
---|
| 1231 | context = bb_xgetlarg(U_opt, 10, 1, INT_MAX);
|
---|
| 1232 | }
|
---|
| 1233 | argc -= optind;
|
---|
| 1234 | argv += optind;
|
---|
| 1235 |
|
---|
| 1236 | /*
|
---|
| 1237 | * Do sanity checks, fill in stb1 and stb2 and call the appropriate
|
---|
| 1238 | * driver routine. Both drivers use the contents of stb1 and stb2.
|
---|
| 1239 | */
|
---|
| 1240 | if (argc < 2) {
|
---|
| 1241 | bb_error_msg("Missing filename");
|
---|
| 1242 | bb_show_usage();
|
---|
| 1243 | }
|
---|
| 1244 | if (strcmp(argv[0], "-") == 0) {
|
---|
| 1245 | fstat(STDIN_FILENO, &stb1);
|
---|
| 1246 | gotstdin = 1;
|
---|
| 1247 | } else
|
---|
| 1248 | xstat(argv[0], &stb1);
|
---|
| 1249 | if (strcmp(argv[1], "-") == 0) {
|
---|
| 1250 | fstat(STDIN_FILENO, &stb2);
|
---|
| 1251 | gotstdin = 1;
|
---|
| 1252 | } else
|
---|
| 1253 | xstat(argv[1], &stb2);
|
---|
| 1254 | if (gotstdin && (S_ISDIR(stb1.st_mode) || S_ISDIR(stb2.st_mode)))
|
---|
| 1255 | bb_error_msg_and_die("Can't compare - to a directory");
|
---|
| 1256 | if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) {
|
---|
| 1257 | #if ENABLE_FEATURE_DIFF_DIR
|
---|
| 1258 | diffdir(argv[0], argv[1]);
|
---|
| 1259 | #else
|
---|
| 1260 | bb_error_msg_and_die("Directory comparison not supported");
|
---|
| 1261 | #endif
|
---|
| 1262 | } else {
|
---|
| 1263 | if (S_ISDIR(stb1.st_mode)) {
|
---|
| 1264 | argv[0] = concat_path_file(argv[0], argv[1]);
|
---|
| 1265 | xstat(argv[0], &stb1);
|
---|
| 1266 | }
|
---|
| 1267 | if (S_ISDIR(stb2.st_mode)) {
|
---|
| 1268 | argv[1] = concat_path_file(argv[1], argv[0]);
|
---|
| 1269 | xstat(argv[1], &stb2);
|
---|
| 1270 | }
|
---|
| 1271 | print_status(diffreg(argv[0], argv[1], 0), argv[0], argv[1], NULL);
|
---|
| 1272 | }
|
---|
| 1273 | exit(status);
|
---|
| 1274 | }
|
---|