diff-tree.c

   1 #include <ctype.h>
   2 #include "cache.h"
   3 #include "diff.h"
   4
   5 static int silent = 0;
   6 static int show_root_diff = 0;
   7 static int verbose_header = 0;
   8 static int ignore_merges = 1;
   9 static int recursive = 0;
  10 static int read_stdin = 0;
  11 static int diff_output_format = DIFF_FORMAT_HUMAN;
  12 static int detect_rename = 0;
  13 static int reverse_diff = 0;
  14 static int diff_score_opt = 0;
  15 static const char *pickaxe = NULL;
  16 static const char *header = NULL;
  17 static const char *header_prefix = "";
  18
  19 // What paths are we interested in?
  20 static int nr_paths = 0;
  21 static const char **paths = NULL;
  22 static int *pathlens = NULL;
  23
  24 static int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base);
  25
  26 static void update_tree_entry(void **bufp, unsigned long *sizep)
  27 {
  28         void *buf = *bufp;
  29         unsigned long size = *sizep;
  30         int len = strlen(buf) + 1 + 20;
  31
  32         if (size < len)
  33                 die("corrupt tree file");
  34         *bufp = buf + len;
  35         *sizep = size - len;
  36 }
  37
  38 static const unsigned char *extract(void *tree, unsigned long size, const char **pathp, unsigned int *modep)
  39 {
  40         int len = strlen(tree)+1;
  41         const unsigned char *sha1 = tree + len;
  42         const char *path = strchr(tree, ' ');
  43
  44         if (!path || size < len + 20 || sscanf(tree, "%o", modep) != 1)
  45                 die("corrupt tree file");
  46         *pathp = path+1;
  47         return sha1;
  48 }
  49
  50 static char *malloc_base(const char *base, const char *path, int pathlen)
  51 {
  52         int baselen = strlen(base);
  53         char *newbase = xmalloc(baselen + pathlen + 2);
  54         memcpy(newbase, base, baselen);
  55         memcpy(newbase + baselen, path, pathlen);
  56         memcpy(newbase + baselen + pathlen, "/", 2);
  57         return newbase;
  58 }
  59
  60 static void show_file(const char *prefix, void *tree, unsigned long size, const char *base);
  61 static void show_tree(const char *prefix, void *tree, unsigned long size, const char *base);
  62
  63 /* A file entry went away or appeared */
  64 static void show_file(const char *prefix, void *tree, unsigned long size, const char *base)
  65 {
  66         unsigned mode;
  67         const char *path;
  68         const unsigned char *sha1 = extract(tree, size, &path, &mode);
  69
  70         if (silent)
  71                 return;
  72
  73         if (recursive && S_ISDIR(mode)) {
  74                 char type[20];
  75                 unsigned long size;
  76                 char *newbase = malloc_base(base, path, strlen(path));
  77                 void *tree;
  78
  79                 tree = read_sha1_file(sha1, type, &size);
  80                 if (!tree || strcmp(type, "tree"))
  81                         die("corrupt tree sha %s", sha1_to_hex(sha1));
  82
  83                 show_tree(prefix, tree, size, newbase);
  84
  85                 free(tree);
  86                 free(newbase);
  87                 return;
  88         }
  89
  90         diff_addremove(prefix[0], mode, sha1, base, path);
  91 }
  92
  93 static int compare_tree_entry(void *tree1, unsigned long size1, void *tree2, unsigned long size2, const char *base)
  94 {
  95         unsigned mode1, mode2;
  96         const char *path1, *path2;
  97         const unsigned char *sha1, *sha2;
  98         int cmp, pathlen1, pathlen2;
  99
 100         sha1 = extract(tree1, size1, &path1, &mode1);
 101         sha2 = extract(tree2, size2, &path2, &mode2);
 102
 103         pathlen1 = strlen(path1);
 104         pathlen2 = strlen(path2);
 105         cmp = base_name_compare(path1, pathlen1, mode1, path2, pathlen2, mode2);
 106         if (cmp < 0) {
 107                 show_file("-", tree1, size1, base);
 108                 return -1;
 109         }
 110         if (cmp > 0) {
 111                 show_file("+", tree2, size2, base);
 112                 return 1;
 113         }
 114         if (!memcmp(sha1, sha2, 20) && mode1 == mode2)
 115                 return 0;
 116
 117         /*
 118          * If the filemode has changed to/from a directory from/to a regular
 119          * file, we need to consider it a remove and an add.
 120          */
 121         if (S_ISDIR(mode1) != S_ISDIR(mode2)) {
 122                 show_file("-", tree1, size1, base);
 123                 show_file("+", tree2, size2, base);
 124                 return 0;
 125         }
 126
 127         if (recursive && S_ISDIR(mode1)) {
 128                 int retval;
 129                 char *newbase = malloc_base(base, path1, pathlen1);
 130                 retval = diff_tree_sha1(sha1, sha2, newbase);
 131                 free(newbase);
 132                 return retval;
 133         }
 134
 135         if (silent)
 136                 return 0;
 137
 138         diff_change(mode1, mode2, sha1, sha2, base, path1);
 139         return 0;
 140 }
 141
 142 static int interesting(void *tree, unsigned long size, const char *base)
 143 {
 144         const char *path;
 145         unsigned mode;
 146         int i;
 147         int baselen, pathlen;
 148
 149         if (!nr_paths)
 150                 return 1;
 151
 152         (void)extract(tree, size, &path, &mode);
 153
 154         pathlen = strlen(path);
 155         baselen = strlen(base);
 156
 157         for (i=0; i < nr_paths; i++) {
 158                 const char *match = paths[i];
 159                 int matchlen = pathlens[i];
 160
 161                 if (baselen >= matchlen) {
 162                         /* If it doesn't match, move along... */
 163                         if (strncmp(base, match, matchlen))
 164                                 continue;
 165
 166                         /* The base is a subdirectory of a path which was specified. */
 167                         return 1;
 168                 }
 169
 170                 /* Does the base match? */
 171                 if (strncmp(base, match, baselen))
 172                         continue;
 173
 174                 match += baselen;
 175                 matchlen -= baselen;
 176
 177                 if (pathlen > matchlen)
 178                         continue;
 179
 180                 if (matchlen > pathlen) {
 181                         if (match[pathlen] != '/')
 182                                 continue;
 183                         if (!S_ISDIR(mode))
 184                                 continue;
 185                 }
 186
 187                 if (strncmp(path, match, pathlen))
 188                         continue;
 189
 190                 return 1;
 191         }
 192         return 0; /* No matches */
 193 }
 194
 195 /* A whole sub-tree went away or appeared */
 196 static void show_tree(const char *prefix, void *tree, unsigned long size, const char *base)
 197 {
 198         while (size) {
 199                 if (interesting(tree, size, base))
 200                         show_file(prefix, tree, size, base);
 201                 update_tree_entry(&tree, &size);
 202         }
 203 }
 204
 205 static int diff_tree(void *tree1, unsigned long size1, void *tree2, unsigned long size2, const char *base)
 206 {
 207         while (size1 | size2) {
 208                 if (nr_paths && size1 && !interesting(tree1, size1, base)) {
 209                         update_tree_entry(&tree1, &size1);
 210                         continue;
 211                 }
 212                 if (nr_paths && size2 && !interesting(tree2, size2, base)) {
 213                         update_tree_entry(&tree2, &size2);
 214                         continue;
 215                 }
 216                 if (!size1) {
 217                         show_file("+", tree2, size2, base);
 218                         update_tree_entry(&tree2, &size2);
 219                         continue;
 220                 }
 221                 if (!size2) {
 222                         show_file("-", tree1, size1, base);
 223                         update_tree_entry(&tree1, &size1);
 224                         continue;
 225                 }
 226                 switch (compare_tree_entry(tree1, size1, tree2, size2, base)) {
 227                 case -1:
 228                         update_tree_entry(&tree1, &size1);
 229                         continue;
 230                 case 0:
 231                         update_tree_entry(&tree1, &size1);
 232                         /* Fallthrough */
 233                 case 1:
 234                         update_tree_entry(&tree2, &size2);
 235                         continue;
 236                 }
 237                 die("git-diff-tree: internal error");
 238         }
 239         return 0;
 240 }
 241
 242 static int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base)
 243 {
 244         void *tree1, *tree2;
 245         unsigned long size1, size2;
 246         int retval;
 247
 248         tree1 = read_object_with_reference(old, "tree", &size1, NULL);
 249         if (!tree1)
 250                 die("unable to read source tree (%s)", sha1_to_hex(old));
 251         tree2 = read_object_with_reference(new, "tree", &size2, NULL);
 252         if (!tree2)
 253                 die("unable to read destination tree (%s)", sha1_to_hex(new));
 254         retval = diff_tree(tree1, size1, tree2, size2, base);
 255         free(tree1);
 256         free(tree2);
 257         return retval;
 258 }
 259
 260 static void call_diff_setup(void)
 261 {
 262         diff_setup(reverse_diff);
 263 }
 264
 265 static int call_diff_flush(void)
 266 {
 267         if (detect_rename)
 268                 diffcore_rename(detect_rename, diff_score_opt);
 269         diffcore_prune();
 270         if (pickaxe) {
 271                 diffcore_pickaxe(pickaxe);
 272                 if (diff_queue_is_empty()) {
 273                         diff_flush(DIFF_FORMAT_NO_OUTPUT, 0);
 274                         return 0;
 275                 }
 276         }
 277         if (nr_paths)
 278                 diffcore_pathspec(paths);
 279         if (header) {
 280                 if (diff_output_format == DIFF_FORMAT_MACHINE) {
 281                         const char *ep, *cp;
 282                         for (cp = header; *cp; cp = ep) {
 283                                 ep = strchr(cp, '\n');
 284                                 if (ep == 0) ep = cp + strlen(cp);
 285                                 printf("%.*s%c", ep-cp, cp, 0);
 286                                 if (*ep) ep++;
 287                         }
 288                 }
 289                 else {
 290                         printf("%s", header);
 291                 }
 292                 header = NULL;
 293         }
 294         diff_flush(diff_output_format, 1);
 295         return 1;
 296 }
 297
 298 static int diff_tree_sha1_top(const unsigned char *old,
 299                               const unsigned char *new, const char *base)
 300 {
 301         int ret;
 302
 303         call_diff_setup();
 304         ret = diff_tree_sha1(old, new, base);
 305         call_diff_flush();
 306         return ret;
 307 }
 308
 309 static int diff_root_tree(const unsigned char *new, const char *base)
 310 {
 311         int retval;
 312         void *tree;
 313         unsigned long size;
 314
 315         call_diff_setup();
 316         tree = read_object_with_reference(new, "tree", &size, NULL);
 317         if (!tree)
 318                 die("unable to read root tree (%s)", sha1_to_hex(new));
 319         retval = diff_tree("", 0, tree, size, base);
 320         free(tree);
 321         call_diff_flush();
 322         return retval;
 323 }
 324
 325 static int get_one_line(const char *msg, unsigned long len)
 326 {
 327         int ret = 0;
 328
 329         while (len--) {
 330                 ret++;
 331                 if (*msg++ == '\n')
 332                         break;
 333         }
 334         return ret;
 335 }
 336
 337 static int add_author_info(char *buf, const char *line, int len)
 338 {
 339         char *date;
 340         unsigned int namelen;
 341         unsigned long time;
 342         int tz;
 343
 344         line += strlen("author ");
 345         date = strchr(line, '>');
 346         if (!date)
 347                 return 0;
 348         namelen = ++date - line;
 349         time = strtoul(date, &date, 10);
 350         tz = strtol(date, NULL, 10);
 351
 352         return sprintf(buf, "Author: %.*s\nDate:   %s\n",
 353                 namelen, line,
 354                 show_date(time, tz));
 355 }
 356
 357 static char *generate_header(const char *commit, const char *parent, const char *msg, unsigned long len)
 358 {
 359         static char this_header[16384];
 360         int offset;
 361
 362         offset = sprintf(this_header, "%s%s (from %s)\n", header_prefix, commit, parent);
 363         if (verbose_header) {
 364                 int hdr = 1;
 365
 366                 for (;;) {
 367                         const char *line = msg;
 368                         int linelen = get_one_line(msg, len);
 369
 370                         if (!linelen)
 371                                 break;
 372
 373                         /*
 374                          * We want some slop for indentation and a possible
 375                          * final "...". Thus the "+ 20".
 376                          */
 377                         if (offset + linelen + 20 > sizeof(this_header)) {
 378                                 memcpy(this_header + offset, "    ...\n", 8);
 379                                 offset += 8;
 380                                 break;
 381                         }
 382
 383                         msg += linelen;
 384                         len -= linelen;
 385                         if (linelen == 1)
 386                                 hdr = 0;
 387                         if (hdr) {
 388                                 if (!memcmp(line, "author ", 7))
 389                                         offset += add_author_info(this_header + offset, line, linelen);
 390                                 continue;
 391                         }
 392                         memset(this_header + offset, ' ', 4);
 393                         memcpy(this_header + offset + 4, line, linelen);
 394                         offset += linelen + 4;
 395                 }
 396                 /* Make sure there is an EOLN */
 397                 if (this_header[offset-1] != '\n')
 398                         this_header[offset++] = '\n';
 399                 /* Add _another_ EOLN if we are doing diff output */
 400                 if (!silent)
 401                         this_header[offset++] = '\n';
 402                 this_header[offset] = 0;
 403         }
 404
 405         return this_header;
 406 }
 407
 408 static int diff_tree_commit(const unsigned char *commit, const char *name)
 409 {
 410         unsigned long size, offset;
 411         char *buf = read_object_with_reference(commit, "commit", &size, NULL);
 412
 413         if (!buf)
 414                 return -1;
 415
 416         if (!name) {
 417                 static char commit_name[60];
 418                 strcpy(commit_name, sha1_to_hex(commit));
 419                 name = commit_name;
 420         }
 421
 422         /* Root commit? */
 423         if (show_root_diff && memcmp(buf + 46, "parent ", 7)) {
 424                 header = generate_header(name, "root", buf, size);
 425                 diff_root_tree(commit, "");
 426         }
 427
 428         /* More than one parent? */
 429         if (ignore_merges) {
 430                 if (!memcmp(buf + 46 + 48, "parent ", 7))
 431                         return 0;
 432         }
 433
 434         offset = 46;
 435         while (offset + 48 < size && !memcmp(buf + offset, "parent ", 7)) {
 436                 unsigned char parent[20];
 437                 if (get_sha1_hex(buf + offset + 7, parent))
 438                         return -1;
 439                 header = generate_header(name, sha1_to_hex(parent), buf, size);
 440                 diff_tree_sha1_top(parent, commit, "");
 441                 if (!header && verbose_header) {
 442                         header_prefix = "\ndiff-tree ";
 443                         /*
 444                          * Don't print multiple merge entries if we
 445                          * don't print the diffs.
 446                          */
 447                         if (silent)
 448                                 break;
 449                 }
 450                 offset += 48;
 451         }
 452         return 0;
 453 }
 454
 455 static int diff_tree_stdin(char *line)
 456 {
 457         int len = strlen(line);
 458         unsigned char commit[20], parent[20];
 459         static char this_header[1000];
 460
 461         if (!len || line[len-1] != '\n')
 462                 return -1;
 463         line[len-1] = 0;
 464         if (get_sha1_hex(line, commit))
 465                 return -1;
 466         if (isspace(line[40]) && !get_sha1_hex(line+41, parent)) {
 467                 line[40] = 0;
 468                 line[81] = 0;
 469                 sprintf(this_header, "%s (from %s)\n", line, line+41);
 470                 header = this_header;
 471                 return diff_tree_sha1_top(parent, commit, "");
 472         }
 473         line[40] = 0;
 474         return diff_tree_commit(commit, line);
 475 }
 476
 477 static char *diff_tree_usage =
 478 "git-diff-tree [-p] [-r] [-z] [--stdin] [-M] [-C] [-R] [-S<string>] [-m] [-s] [-v] <tree-ish> <tree-ish>";
 479
 480 int main(int argc, const char **argv)
 481 {
 482         int nr_sha1;
 483         char line[1000];
 484         unsigned char sha1[2][20];
 485
 486         nr_sha1 = 0;
 487         for (;;) {
 488                 const char *arg;
 489
 490                 argv++;
 491                 argc--;
 492                 arg = *argv;
 493                 if (!arg)
 494                         break;
 495
 496                 if (*arg != '-') {
 497                         if (nr_sha1 < 2 && !get_sha1(arg, sha1[nr_sha1])) {
 498                                 nr_sha1++;
 499                                 continue;
 500                         }
 501                         break;
 502                 }
 503
 504                 if (!strcmp(arg, "--")) {
 505                         argv++;
 506                         argc--;
 507                         break;
 508                 }
 509                 if (!strcmp(arg, "-r")) {
 510                         recursive = 1;
 511                         continue;
 512                 }
 513                 if (!strcmp(arg, "-R")) {
 514                         reverse_diff = 1;
 515                         continue;
 516                 }
 517                 if (!strcmp(arg, "-p")) {
 518                         diff_output_format = DIFF_FORMAT_PATCH;
 519                         recursive = 1;
 520                         continue;
 521                 }
 522                 if (!strncmp(arg, "-S", 2)) {
 523                         pickaxe = arg + 2;
 524                         continue;
 525                 }
 526                 if (!strncmp(arg, "-M", 2)) {
 527                         detect_rename = DIFF_DETECT_RENAME;
 528                         diff_score_opt = diff_scoreopt_parse(arg);
 529                         continue;
 530                 }
 531                 if (!strncmp(arg, "-C", 2)) {
 532                         detect_rename = DIFF_DETECT_COPY;
 533                         diff_score_opt = diff_scoreopt_parse(arg);
 534                         continue;
 535                 }
 536                 if (!strcmp(arg, "-z")) {
 537                         diff_output_format = DIFF_FORMAT_MACHINE;
 538                         continue;
 539                 }
 540                 if (!strcmp(arg, "-m")) {
 541                         ignore_merges = 0;
 542                         continue;
 543                 }
 544                 if (!strcmp(arg, "-s")) {
 545                         silent = 1;
 546                         continue;
 547                 }
 548                 if (!strcmp(arg, "-v")) {
 549                         verbose_header = 1;
 550                         header_prefix = "diff-tree ";
 551                         continue;
 552                 }
 553                 if (!strcmp(arg, "--stdin")) {
 554                         read_stdin = 1;
 555                         continue;
 556                 }
 557                 if (!strcmp(arg, "--root")) {
 558                         show_root_diff = 1;
 559                         continue;
 560                 }
 561                 usage(diff_tree_usage);
 562         }
 563
 564         if (argc > 0) {
 565                 int i;
 566
 567                 paths = argv;
 568                 nr_paths = argc;
 569                 pathlens = xmalloc(nr_paths * sizeof(int));
 570                 for (i=0; i<nr_paths; i++)
 571                         pathlens[i] = strlen(paths[i]);
 572         }
 573
 574         switch (nr_sha1) {
 575         case 0:
 576                 if (!read_stdin)
 577                         usage(diff_tree_usage);
 578                 break;
 579         case 1:
 580                 diff_tree_commit(sha1[0], NULL);
 581                 break;
 582         case 2:
 583                 diff_tree_sha1_top(sha1[0], sha1[1], "");
 584                 break;
 585         }
 586
 587         if (!read_stdin)
 588                 return 0;
 589
 590         while (fgets(line, sizeof(line), stdin))
 591                 diff_tree_stdin(line);
 592
 593         return 0;
 594 }