bisect.c

   1 #include "cache.h"
   2 #include "commit.h"
   3 #include "diff.h"
   4 #include "revision.h"
   5 #include "refs.h"
   6 #include "list-objects.h"
   7 #include "quote.h"
   8 #include "sha1-lookup.h"
   9 #include "run-command.h"
  10 #include "bisect.h"
  11
  12 static unsigned char (*skipped_sha1)[20];
  13 static int skipped_sha1_nr;
  14 static int skipped_sha1_alloc;
  15
  16 static const char **rev_argv;
  17 static int rev_argv_nr;
  18 static int rev_argv_alloc;
  19
  20 static const unsigned char *current_bad_sha1;
  21
  22 static const char *argv_diff_tree[] = {"diff-tree", "--pretty", NULL, NULL};
  23 static const char *argv_checkout[] = {"checkout", "-q", NULL, "--", NULL};
  24 static const char *argv_show_branch[] = {"show-branch", NULL, NULL};
  25
  26 /* bits #0-15 in revision.h */
  27
  28 #define COUNTED         (1u<<16)
  29
  30 /*
  31  * This is a truly stupid algorithm, but it's only
  32  * used for bisection, and we just don't care enough.
  33  *
  34  * We care just barely enough to avoid recursing for
  35  * non-merge entries.
  36  */
  37 static int count_distance(struct commit_list *entry)
  38 {
  39         int nr = 0;
  40
  41         while (entry) {
  42                 struct commit *commit = entry->item;
  43                 struct commit_list *p;
  44
  45                 if (commit->object.flags & (UNINTERESTING | COUNTED))
  46                         break;
  47                 if (!(commit->object.flags & TREESAME))
  48                         nr++;
  49                 commit->object.flags |= COUNTED;
  50                 p = commit->parents;
  51                 entry = p;
  52                 if (p) {
  53                         p = p->next;
  54                         while (p) {
  55                                 nr += count_distance(p);
  56                                 p = p->next;
  57                         }
  58                 }
  59         }
  60
  61         return nr;
  62 }
  63
  64 static void clear_distance(struct commit_list *list)
  65 {
  66         while (list) {
  67                 struct commit *commit = list->item;
  68                 commit->object.flags &= ~COUNTED;
  69                 list = list->next;
  70         }
  71 }
  72
  73 #define DEBUG_BISECT 0
  74
  75 static inline int weight(struct commit_list *elem)
  76 {
  77         return *((int*)(elem->item->util));
  78 }
  79
  80 static inline void weight_set(struct commit_list *elem, int weight)
  81 {
  82         *((int*)(elem->item->util)) = weight;
  83 }
  84
  85 static int count_interesting_parents(struct commit *commit)
  86 {
  87         struct commit_list *p;
  88         int count;
  89
  90         for (count = 0, p = commit->parents; p; p = p->next) {
  91                 if (p->item->object.flags & UNINTERESTING)
  92                         continue;
  93                 count++;
  94         }
  95         return count;
  96 }
  97
  98 static inline int halfway(struct commit_list *p, int nr)
  99 {
 100         /*
 101          * Don't short-cut something we are not going to return!
 102          */
 103         if (p->item->object.flags & TREESAME)
 104                 return 0;
 105         if (DEBUG_BISECT)
 106                 return 0;
 107         /*
 108          * 2 and 3 are halfway of 5.
 109          * 3 is halfway of 6 but 2 and 4 are not.
 110          */
 111         switch (2 * weight(p) - nr) {
 112         case -1: case 0: case 1:
 113                 return 1;
 114         default:
 115                 return 0;
 116         }
 117 }
 118
 119 #if !DEBUG_BISECT
 120 #define show_list(a,b,c,d) do { ; } while (0)
 121 #else
 122 static void show_list(const char *debug, int counted, int nr,
 123                       struct commit_list *list)
 124 {
 125         struct commit_list *p;
 126
 127         fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
 128
 129         for (p = list; p; p = p->next) {
 130                 struct commit_list *pp;
 131                 struct commit *commit = p->item;
 132                 unsigned flags = commit->object.flags;
 133                 enum object_type type;
 134                 unsigned long size;
 135                 char *buf = read_sha1_file(commit->object.sha1, &type, &size);
 136                 char *ep, *sp;
 137
 138                 fprintf(stderr, "%c%c%c ",
 139                         (flags & TREESAME) ? ' ' : 'T',
 140                         (flags & UNINTERESTING) ? 'U' : ' ',
 141                         (flags & COUNTED) ? 'C' : ' ');
 142                 if (commit->util)
 143                         fprintf(stderr, "%3d", weight(p));
 144                 else
 145                         fprintf(stderr, "---");
 146                 fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 147                 for (pp = commit->parents; pp; pp = pp->next)
 148                         fprintf(stderr, " %.*s", 8,
 149                                 sha1_to_hex(pp->item->object.sha1));
 150
 151                 sp = strstr(buf, "\n\n");
 152                 if (sp) {
 153                         sp += 2;
 154                         for (ep = sp; *ep && *ep != '\n'; ep++)
 155                                 ;
 156                         fprintf(stderr, " %.*s", (int)(ep - sp), sp);
 157                 }
 158                 fprintf(stderr, "\n");
 159         }
 160 }
 161 #endif /* DEBUG_BISECT */
 162
 163 static struct commit_list *best_bisection(struct commit_list *list, int nr)
 164 {
 165         struct commit_list *p, *best;
 166         int best_distance = -1;
 167
 168         best = list;
 169         for (p = list; p; p = p->next) {
 170                 int distance;
 171                 unsigned flags = p->item->object.flags;
 172
 173                 if (flags & TREESAME)
 174                         continue;
 175                 distance = weight(p);
 176                 if (nr - distance < distance)
 177                         distance = nr - distance;
 178                 if (distance > best_distance) {
 179                         best = p;
 180                         best_distance = distance;
 181                 }
 182         }
 183
 184         return best;
 185 }
 186
 187 struct commit_dist {
 188         struct commit *commit;
 189         int distance;
 190 };
 191
 192 static int compare_commit_dist(const void *a_, const void *b_)
 193 {
 194         struct commit_dist *a, *b;
 195
 196         a = (struct commit_dist *)a_;
 197         b = (struct commit_dist *)b_;
 198         if (a->distance != b->distance)
 199                 return b->distance - a->distance; /* desc sort */
 200         return hashcmp(a->commit->object.sha1, b->commit->object.sha1);
 201 }
 202
 203 static struct commit_list *best_bisection_sorted(struct commit_list *list, int nr)
 204 {
 205         struct commit_list *p;
 206         struct commit_dist *array = xcalloc(nr, sizeof(*array));
 207         int cnt, i;
 208
 209         for (p = list, cnt = 0; p; p = p->next) {
 210                 int distance;
 211                 unsigned flags = p->item->object.flags;
 212
 213                 if (flags & TREESAME)
 214                         continue;
 215                 distance = weight(p);
 216                 if (nr - distance < distance)
 217                         distance = nr - distance;
 218                 array[cnt].commit = p->item;
 219                 array[cnt].distance = distance;
 220                 cnt++;
 221         }
 222         qsort(array, cnt, sizeof(*array), compare_commit_dist);
 223         for (p = list, i = 0; i < cnt; i++) {
 224                 struct name_decoration *r = xmalloc(sizeof(*r) + 100);
 225                 struct object *obj = &(array[i].commit->object);
 226
 227                 sprintf(r->name, "dist=%d", array[i].distance);
 228                 r->next = add_decoration(&name_decoration, obj, r);
 229                 p->item = array[i].commit;
 230                 p = p->next;
 231         }
 232         if (p)
 233                 p->next = NULL;
 234         free(array);
 235         return list;
 236 }
 237
 238 /*
 239  * zero or positive weight is the number of interesting commits it can
 240  * reach, including itself.  Especially, weight = 0 means it does not
 241  * reach any tree-changing commits (e.g. just above uninteresting one
 242  * but traversal is with pathspec).
 243  *
 244  * weight = -1 means it has one parent and its distance is yet to
 245  * be computed.
 246  *
 247  * weight = -2 means it has more than one parent and its distance is
 248  * unknown.  After running count_distance() first, they will get zero
 249  * or positive distance.
 250  */
 251 static struct commit_list *do_find_bisection(struct commit_list *list,
 252                                              int nr, int *weights,
 253                                              int find_all)
 254 {
 255         int n, counted;
 256         struct commit_list *p;
 257
 258         counted = 0;
 259
 260         for (n = 0, p = list; p; p = p->next) {
 261                 struct commit *commit = p->item;
 262                 unsigned flags = commit->object.flags;
 263
 264                 p->item->util = &weights[n++];
 265                 switch (count_interesting_parents(commit)) {
 266                 case 0:
 267                         if (!(flags & TREESAME)) {
 268                                 weight_set(p, 1);
 269                                 counted++;
 270                                 show_list("bisection 2 count one",
 271                                           counted, nr, list);
 272                         }
 273                         /*
 274                          * otherwise, it is known not to reach any
 275                          * tree-changing commit and gets weight 0.
 276                          */
 277                         break;
 278                 case 1:
 279                         weight_set(p, -1);
 280                         break;
 281                 default:
 282                         weight_set(p, -2);
 283                         break;
 284                 }
 285         }
 286
 287         show_list("bisection 2 initialize", counted, nr, list);
 288
 289         /*
 290          * If you have only one parent in the resulting set
 291          * then you can reach one commit more than that parent
 292          * can reach.  So we do not have to run the expensive
 293          * count_distance() for single strand of pearls.
 294          *
 295          * However, if you have more than one parents, you cannot
 296          * just add their distance and one for yourself, since
 297          * they usually reach the same ancestor and you would
 298          * end up counting them twice that way.
 299          *
 300          * So we will first count distance of merges the usual
 301          * way, and then fill the blanks using cheaper algorithm.
 302          */
 303         for (p = list; p; p = p->next) {
 304                 if (p->item->object.flags & UNINTERESTING)
 305                         continue;
 306                 if (weight(p) != -2)
 307                         continue;
 308                 weight_set(p, count_distance(p));
 309                 clear_distance(list);
 310
 311                 /* Does it happen to be at exactly half-way? */
 312                 if (!find_all && halfway(p, nr))
 313                         return p;
 314                 counted++;
 315         }
 316
 317         show_list("bisection 2 count_distance", counted, nr, list);
 318
 319         while (counted < nr) {
 320                 for (p = list; p; p = p->next) {
 321                         struct commit_list *q;
 322                         unsigned flags = p->item->object.flags;
 323
 324                         if (0 <= weight(p))
 325                                 continue;
 326                         for (q = p->item->parents; q; q = q->next) {
 327                                 if (q->item->object.flags & UNINTERESTING)
 328                                         continue;
 329                                 if (0 <= weight(q))
 330                                         break;
 331                         }
 332                         if (!q)
 333                                 continue;
 334
 335                         /*
 336                          * weight for p is unknown but q is known.
 337                          * add one for p itself if p is to be counted,
 338                          * otherwise inherit it from q directly.
 339                          */
 340                         if (!(flags & TREESAME)) {
 341                                 weight_set(p, weight(q)+1);
 342                                 counted++;
 343                                 show_list("bisection 2 count one",
 344                                           counted, nr, list);
 345                         }
 346                         else
 347                                 weight_set(p, weight(q));
 348
 349                         /* Does it happen to be at exactly half-way? */
 350                         if (!find_all && halfway(p, nr))
 351                                 return p;
 352                 }
 353         }
 354
 355         show_list("bisection 2 counted all", counted, nr, list);
 356
 357         if (!find_all)
 358                 return best_bisection(list, nr);
 359         else
 360                 return best_bisection_sorted(list, nr);
 361 }
 362
 363 struct commit_list *find_bisection(struct commit_list *list,
 364                                           int *reaches, int *all,
 365                                           int find_all)
 366 {
 367         int nr, on_list;
 368         struct commit_list *p, *best, *next, *last;
 369         int *weights;
 370
 371         show_list("bisection 2 entry", 0, 0, list);
 372
 373         /*
 374          * Count the number of total and tree-changing items on the
 375          * list, while reversing the list.
 376          */
 377         for (nr = on_list = 0, last = NULL, p = list;
 378              p;
 379              p = next) {
 380                 unsigned flags = p->item->object.flags;
 381
 382                 next = p->next;
 383                 if (flags & UNINTERESTING)
 384                         continue;
 385                 p->next = last;
 386                 last = p;
 387                 if (!(flags & TREESAME))
 388                         nr++;
 389                 on_list++;
 390         }
 391         list = last;
 392         show_list("bisection 2 sorted", 0, nr, list);
 393
 394         *all = nr;
 395         weights = xcalloc(on_list, sizeof(*weights));
 396
 397         /* Do the real work of finding bisection commit. */
 398         best = do_find_bisection(list, nr, weights, find_all);
 399         if (best) {
 400                 if (!find_all)
 401                         best->next = NULL;
 402                 *reaches = weight(best);
 403         }
 404         free(weights);
 405         return best;
 406 }
 407
 408 static int register_ref(const char *refname, const unsigned char *sha1,
 409                         int flags, void *cb_data)
 410 {
 411         if (!strcmp(refname, "bad")) {
 412                 ALLOC_GROW(rev_argv, rev_argv_nr + 1, rev_argv_alloc);
 413                 current_bad_sha1 = sha1;
 414                 rev_argv[rev_argv_nr++] = xstrdup(sha1_to_hex(sha1));
 415         } else if (!prefixcmp(refname, "good-")) {
 416                 const char *hex = sha1_to_hex(sha1);
 417                 char *good = xmalloc(strlen(hex) + 2);
 418                 *good = '^';
 419                 memcpy(good + 1, hex, strlen(hex) + 1);
 420                 ALLOC_GROW(rev_argv, rev_argv_nr + 1, rev_argv_alloc);
 421                 rev_argv[rev_argv_nr++] = good;
 422         } else if (!prefixcmp(refname, "skip-")) {
 423                 ALLOC_GROW(skipped_sha1, skipped_sha1_nr + 1,
 424                            skipped_sha1_alloc);
 425                 hashcpy(skipped_sha1[skipped_sha1_nr++], sha1);
 426         }
 427
 428         return 0;
 429 }
 430
 431 static int read_bisect_refs(void)
 432 {
 433         return for_each_ref_in("refs/bisect/", register_ref, NULL);
 434 }
 435
 436 void read_bisect_paths(void)
 437 {
 438         struct strbuf str = STRBUF_INIT;
 439         const char *filename = git_path("BISECT_NAMES");
 440         FILE *fp = fopen(filename, "r");
 441
 442         if (!fp)
 443                 die("Could not open file '%s': %s", filename, strerror(errno));
 444
 445         while (strbuf_getline(&str, fp, '\n') != EOF) {
 446                 char *quoted;
 447                 int res;
 448
 449                 strbuf_trim(&str);
 450                 quoted = strbuf_detach(&str, NULL);
 451                 res = sq_dequote_to_argv(quoted, &rev_argv,
 452                                          &rev_argv_nr, &rev_argv_alloc);
 453                 if (res)
 454                         die("Badly quoted content in file '%s': %s",
 455                             filename, quoted);
 456         }
 457
 458         strbuf_release(&str);
 459         fclose(fp);
 460 }
 461
 462 static int skipcmp(const void *a, const void *b)
 463 {
 464         return hashcmp(a, b);
 465 }
 466
 467 static void prepare_skipped(void)
 468 {
 469         qsort(skipped_sha1, skipped_sha1_nr, sizeof(*skipped_sha1), skipcmp);
 470 }
 471
 472 static const unsigned char *skipped_sha1_access(size_t index, void *table)
 473 {
 474         unsigned char (*skipped)[20] = table;
 475         return skipped[index];
 476 }
 477
 478 static int lookup_skipped(unsigned char *sha1)
 479 {
 480         return sha1_pos(sha1, skipped_sha1, skipped_sha1_nr,
 481                         skipped_sha1_access);
 482 }
 483
 484 struct commit_list *filter_skipped(struct commit_list *list,
 485                                    struct commit_list **tried,
 486                                    int show_all)
 487 {
 488         struct commit_list *filtered = NULL, **f = &filtered;
 489
 490         *tried = NULL;
 491
 492         if (!skipped_sha1_nr)
 493                 return list;
 494
 495         prepare_skipped();
 496
 497         while (list) {
 498                 struct commit_list *next = list->next;
 499                 list->next = NULL;
 500                 if (0 <= lookup_skipped(list->item->object.sha1)) {
 501                         /* Move current to tried list */
 502                         *tried = list;
 503                         tried = &list->next;
 504                 } else {
 505                         if (!show_all)
 506                                 return list;
 507                         /* Move current to filtered list */
 508                         *f = list;
 509                         f = &list->next;
 510                 }
 511                 list = next;
 512         }
 513
 514         return filtered;
 515 }
 516
 517 static void bisect_rev_setup(struct rev_info *revs, const char *prefix)
 518 {
 519         init_revisions(revs, prefix);
 520         revs->abbrev = 0;
 521         revs->commit_format = CMIT_FMT_UNSPECIFIED;
 522
 523         /* argv[0] will be ignored by setup_revisions */
 524         ALLOC_GROW(rev_argv, rev_argv_nr + 1, rev_argv_alloc);
 525         rev_argv[rev_argv_nr++] = xstrdup("bisect_rev_setup");
 526
 527         if (read_bisect_refs())
 528                 die("reading bisect refs failed");
 529
 530         ALLOC_GROW(rev_argv, rev_argv_nr + 1, rev_argv_alloc);
 531         rev_argv[rev_argv_nr++] = xstrdup("--");
 532
 533         read_bisect_paths();
 534
 535         ALLOC_GROW(rev_argv, rev_argv_nr + 1, rev_argv_alloc);
 536         rev_argv[rev_argv_nr++] = NULL;
 537
 538         setup_revisions(rev_argv_nr, rev_argv, revs, NULL);
 539
 540         revs->limited = 1;
 541 }
 542
 543 static void bisect_common(struct rev_info *revs, const char *prefix,
 544                           int *reaches, int *all)
 545 {
 546         bisect_rev_setup(revs, prefix);
 547
 548         if (prepare_revision_walk(revs))
 549                 die("revision walk setup failed");
 550         if (revs->tree_objects)
 551                 mark_edges_uninteresting(revs->commits, revs, NULL);
 552
 553         revs->commits = find_bisection(revs->commits, reaches, all,
 554                                        !!skipped_sha1_nr);
 555 }
 556
 557 static void exit_if_skipped_commits(struct commit_list *tried,
 558                                     const unsigned char *bad)
 559 {
 560         if (!tried)
 561                 return;
 562
 563         printf("There are only 'skip'ped commits left to test.\n"
 564                "The first bad commit could be any of:\n");
 565         print_commit_list(tried, "%s\n", "%s\n");
 566         if (bad)
 567                 printf("%s\n", sha1_to_hex(bad));
 568         printf("We cannot bisect more!\n");
 569         exit(2);
 570 }
 571
 572 static void mark_expected_rev(char *bisect_rev_hex)
 573 {
 574         int len = strlen(bisect_rev_hex);
 575         const char *filename = git_path("BISECT_EXPECTED_REV");
 576         int fd = open(filename, O_CREAT | O_TRUNC | O_WRONLY, 0600);
 577
 578         if (fd < 0)
 579                 die("could not create file '%s': %s",
 580                     filename, strerror(errno));
 581
 582         bisect_rev_hex[len] = '\n';
 583         write_or_die(fd, bisect_rev_hex, len + 1);
 584         bisect_rev_hex[len] = '\0';
 585
 586         if (close(fd) < 0)
 587                 die("closing file %s: %s", filename, strerror(errno));
 588 }
 589
 590 static int bisect_checkout(char *bisect_rev_hex)
 591 {
 592         int res;
 593
 594         mark_expected_rev(bisect_rev_hex);
 595
 596         argv_checkout[2] = bisect_rev_hex;
 597         res = run_command_v_opt(argv_checkout, RUN_GIT_CMD);
 598         if (res)
 599                 exit(res);
 600
 601         argv_show_branch[1] = bisect_rev_hex;
 602         return run_command_v_opt(argv_show_branch, RUN_GIT_CMD);
 603 }
 604
 605 /*
 606  * We use the convention that exiting with an exit code 10 means that
 607  * the bisection process finished successfully.
 608  * In this case the calling shell script should exit 0.
 609  */
 610 int bisect_next_exit(const char *prefix)
 611 {
 612         struct rev_info revs;
 613         struct commit_list *tried;
 614         int reaches = 0, all = 0, nr;
 615         const unsigned char *bisect_rev;
 616         char bisect_rev_hex[41];
 617
 618         bisect_common(&revs, prefix, &reaches, &all);
 619
 620         revs.commits = filter_skipped(revs.commits, &tried, 0);
 621
 622         if (!revs.commits) {
 623                 /*
 624                  * We should exit here only if the "bad"
 625                  * commit is also a "skip" commit.
 626                  */
 627                 exit_if_skipped_commits(tried, NULL);
 628
 629                 printf("%s was both good and bad\n",
 630                        sha1_to_hex(current_bad_sha1));
 631                 exit(1);
 632         }
 633
 634         bisect_rev = revs.commits->item->object.sha1;
 635         memcpy(bisect_rev_hex, sha1_to_hex(bisect_rev), 41);
 636
 637         if (!hashcmp(bisect_rev, current_bad_sha1)) {
 638                 exit_if_skipped_commits(tried, current_bad_sha1);
 639                 printf("%s is first bad commit\n", bisect_rev_hex);
 640                 argv_diff_tree[2] = bisect_rev_hex;
 641                 run_command_v_opt(argv_diff_tree, RUN_GIT_CMD);
 642                 /* This means the bisection process succeeded. */
 643                 exit(10);
 644         }
 645
 646         nr = all - reaches - 1;
 647         printf("Bisecting: %d revisions left to test after this "
 648                "(roughly %d steps)\n", nr, estimate_bisect_steps(all));
 649
 650         return bisect_checkout(bisect_rev_hex);
 651 }
 652