diff-tree.c

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