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 "log-tree.h"
  11 #include "bisect.h"
  12 #include "sha1-array.h"
  13 #include "argv-array.h"
  14
  15 static struct sha1_array good_revs;
  16 static struct sha1_array skipped_revs;
  17
  18 static unsigned char *current_bad_sha1;
  19
  20 static const char *argv_checkout[] = {"checkout", "-q", NULL, "--", NULL};
  21 static const char *argv_show_branch[] = {"show-branch", NULL, NULL};
  22 static const char *argv_update_ref[] = {"update-ref", "--no-deref", "BISECT_HEAD", NULL, NULL};
  23
  24 /* Remember to update object flag allocation in object.h */
  25 #define COUNTED         (1u<<16)
  26
  27 /*
  28  * This is a truly stupid algorithm, but it's only
  29  * used for bisection, and we just don't care enough.
  30  *
  31  * We care just barely enough to avoid recursing for
  32  * non-merge entries.
  33  */
  34 static int count_distance(struct commit_list *entry)
  35 {
  36         int nr = 0;
  37
  38         while (entry) {
  39                 struct commit *commit = entry->item;
  40                 struct commit_list *p;
  41
  42                 if (commit->object.flags & (UNINTERESTING | COUNTED))
  43                         break;
  44                 if (!(commit->object.flags & TREESAME))
  45                         nr++;
  46                 commit->object.flags |= COUNTED;
  47                 p = commit->parents;
  48                 entry = p;
  49                 if (p) {
  50                         p = p->next;
  51                         while (p) {
  52                                 nr += count_distance(p);
  53                                 p = p->next;
  54                         }
  55                 }
  56         }
  57
  58         return nr;
  59 }
  60
  61 static void clear_distance(struct commit_list *list)
  62 {
  63         while (list) {
  64                 struct commit *commit = list->item;
  65                 commit->object.flags &= ~COUNTED;
  66                 list = list->next;
  67         }
  68 }
  69
  70 #define DEBUG_BISECT 0
  71
  72 static inline int weight(struct commit_list *elem)
  73 {
  74         return *((int*)(elem->item->util));
  75 }
  76
  77 static inline void weight_set(struct commit_list *elem, int weight)
  78 {
  79         *((int*)(elem->item->util)) = weight;
  80 }
  81
  82 static int count_interesting_parents(struct commit *commit)
  83 {
  84         struct commit_list *p;
  85         int count;
  86
  87         for (count = 0, p = commit->parents; p; p = p->next) {
  88                 if (p->item->object.flags & UNINTERESTING)
  89                         continue;
  90                 count++;
  91         }
  92         return count;
  93 }
  94
  95 static inline int halfway(struct commit_list *p, int nr)
  96 {
  97         /*
  98          * Don't short-cut something we are not going to return!
  99          */
 100         if (p->item->object.flags & TREESAME)
 101                 return 0;
 102         if (DEBUG_BISECT)
 103                 return 0;
 104         /*
 105          * 2 and 3 are halfway of 5.
 106          * 3 is halfway of 6 but 2 and 4 are not.
 107          */
 108         switch (2 * weight(p) - nr) {
 109         case -1: case 0: case 1:
 110                 return 1;
 111         default:
 112                 return 0;
 113         }
 114 }
 115
 116 #if !DEBUG_BISECT
 117 #define show_list(a,b,c,d) do { ; } while (0)
 118 #else
 119 static void show_list(const char *debug, int counted, int nr,
 120                       struct commit_list *list)
 121 {
 122         struct commit_list *p;
 123
 124         fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
 125
 126         for (p = list; p; p = p->next) {
 127                 struct commit_list *pp;
 128                 struct commit *commit = p->item;
 129                 unsigned flags = commit->object.flags;
 130                 enum object_type type;
 131                 unsigned long size;
 132                 char *buf = read_sha1_file(commit->object.sha1, &type, &size);
 133                 const char *subject_start;
 134                 int subject_len;
 135
 136                 fprintf(stderr, "%c%c%c ",
 137                         (flags & TREESAME) ? ' ' : 'T',
 138                         (flags & UNINTERESTING) ? 'U' : ' ',
 139                         (flags & COUNTED) ? 'C' : ' ');
 140                 if (commit->util)
 141                         fprintf(stderr, "%3d", weight(p));
 142                 else
 143                         fprintf(stderr, "---");
 144                 fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 145                 for (pp = commit->parents; pp; pp = pp->next)
 146                         fprintf(stderr, " %.*s", 8,
 147                                 sha1_to_hex(pp->item->object.sha1));
 148
 149                 subject_len = find_commit_subject(buf, &subject_start);
 150                 if (subject_len)
 151                         fprintf(stderr, " %.*s", subject_len, subject_start);
 152                 fprintf(stderr, "\n");
 153         }
 154 }
 155 #endif /* DEBUG_BISECT */
 156
 157 static struct commit_list *best_bisection(struct commit_list *list, int nr)
 158 {
 159         struct commit_list *p, *best;
 160         int best_distance = -1;
 161
 162         best = list;
 163         for (p = list; p; p = p->next) {
 164                 int distance;
 165                 unsigned flags = p->item->object.flags;
 166
 167                 if (flags & TREESAME)
 168                         continue;
 169                 distance = weight(p);
 170                 if (nr - distance < distance)
 171                         distance = nr - distance;
 172                 if (distance > best_distance) {
 173                         best = p;
 174                         best_distance = distance;
 175                 }
 176         }
 177
 178         return best;
 179 }
 180
 181 struct commit_dist {
 182         struct commit *commit;
 183         int distance;
 184 };
 185
 186 static int compare_commit_dist(const void *a_, const void *b_)
 187 {
 188         struct commit_dist *a, *b;
 189
 190         a = (struct commit_dist *)a_;
 191         b = (struct commit_dist *)b_;
 192         if (a->distance != b->distance)
 193                 return b->distance - a->distance; /* desc sort */
 194         return hashcmp(a->commit->object.sha1, b->commit->object.sha1);
 195 }
 196
 197 static struct commit_list *best_bisection_sorted(struct commit_list *list, int nr)
 198 {
 199         struct commit_list *p;
 200         struct commit_dist *array = xcalloc(nr, sizeof(*array));
 201         int cnt, i;
 202
 203         for (p = list, cnt = 0; p; p = p->next) {
 204                 int distance;
 205                 unsigned flags = p->item->object.flags;
 206
 207                 if (flags & TREESAME)
 208                         continue;
 209                 distance = weight(p);
 210                 if (nr - distance < distance)
 211                         distance = nr - distance;
 212                 array[cnt].commit = p->item;
 213                 array[cnt].distance = distance;
 214                 cnt++;
 215         }
 216         qsort(array, cnt, sizeof(*array), compare_commit_dist);
 217         for (p = list, i = 0; i < cnt; i++) {
 218                 struct name_decoration *r = xmalloc(sizeof(*r) + 100);
 219                 struct object *obj = &(array[i].commit->object);
 220
 221                 sprintf(r->name, "dist=%d", array[i].distance);
 222                 r->next = add_decoration(&name_decoration, obj, r);
 223                 p->item = array[i].commit;
 224                 p = p->next;
 225         }
 226         if (p)
 227                 p->next = NULL;
 228         free(array);
 229         return list;
 230 }
 231
 232 /*
 233  * zero or positive weight is the number of interesting commits it can
 234  * reach, including itself.  Especially, weight = 0 means it does not
 235  * reach any tree-changing commits (e.g. just above uninteresting one
 236  * but traversal is with pathspec).
 237  *
 238  * weight = -1 means it has one parent and its distance is yet to
 239  * be computed.
 240  *
 241  * weight = -2 means it has more than one parent and its distance is
 242  * unknown.  After running count_distance() first, they will get zero
 243  * or positive distance.
 244  */
 245 static struct commit_list *do_find_bisection(struct commit_list *list,
 246                                              int nr, int *weights,
 247                                              int find_all)
 248 {
 249         int n, counted;
 250         struct commit_list *p;
 251
 252         counted = 0;
 253
 254         for (n = 0, p = list; p; p = p->next) {
 255                 struct commit *commit = p->item;
 256                 unsigned flags = commit->object.flags;
 257
 258                 p->item->util = &weights[n++];
 259                 switch (count_interesting_parents(commit)) {
 260                 case 0:
 261                         if (!(flags & TREESAME)) {
 262                                 weight_set(p, 1);
 263                                 counted++;
 264                                 show_list("bisection 2 count one",
 265                                           counted, nr, list);
 266                         }
 267                         /*
 268                          * otherwise, it is known not to reach any
 269                          * tree-changing commit and gets weight 0.
 270                          */
 271                         break;
 272                 case 1:
 273                         weight_set(p, -1);
 274                         break;
 275                 default:
 276                         weight_set(p, -2);
 277                         break;
 278                 }
 279         }
 280
 281         show_list("bisection 2 initialize", counted, nr, list);
 282
 283         /*
 284          * If you have only one parent in the resulting set
 285          * then you can reach one commit more than that parent
 286          * can reach.  So we do not have to run the expensive
 287          * count_distance() for single strand of pearls.
 288          *
 289          * However, if you have more than one parents, you cannot
 290          * just add their distance and one for yourself, since
 291          * they usually reach the same ancestor and you would
 292          * end up counting them twice that way.
 293          *
 294          * So we will first count distance of merges the usual
 295          * way, and then fill the blanks using cheaper algorithm.
 296          */
 297         for (p = list; p; p = p->next) {
 298                 if (p->item->object.flags & UNINTERESTING)
 299                         continue;
 300                 if (weight(p) != -2)
 301                         continue;
 302                 weight_set(p, count_distance(p));
 303                 clear_distance(list);
 304
 305                 /* Does it happen to be at exactly half-way? */
 306                 if (!find_all && halfway(p, nr))
 307                         return p;
 308                 counted++;
 309         }
 310
 311         show_list("bisection 2 count_distance", counted, nr, list);
 312
 313         while (counted < nr) {
 314                 for (p = list; p; p = p->next) {
 315                         struct commit_list *q;
 316                         unsigned flags = p->item->object.flags;
 317
 318                         if (0 <= weight(p))
 319                                 continue;
 320                         for (q = p->item->parents; q; q = q->next) {
 321                                 if (q->item->object.flags & UNINTERESTING)
 322                                         continue;
 323                                 if (0 <= weight(q))
 324                                         break;
 325                         }
 326                         if (!q)
 327                                 continue;
 328
 329                         /*
 330                          * weight for p is unknown but q is known.
 331                          * add one for p itself if p is to be counted,
 332                          * otherwise inherit it from q directly.
 333                          */
 334                         if (!(flags & TREESAME)) {
 335                                 weight_set(p, weight(q)+1);
 336                                 counted++;
 337                                 show_list("bisection 2 count one",
 338                                           counted, nr, list);
 339                         }
 340                         else
 341                                 weight_set(p, weight(q));
 342
 343                         /* Does it happen to be at exactly half-way? */
 344                         if (!find_all && halfway(p, nr))
 345                                 return p;
 346                 }
 347         }
 348
 349         show_list("bisection 2 counted all", counted, nr, list);
 350
 351         if (!find_all)
 352                 return best_bisection(list, nr);
 353         else
 354                 return best_bisection_sorted(list, nr);
 355 }
 356
 357 struct commit_list *find_bisection(struct commit_list *list,
 358                                           int *reaches, int *all,
 359                                           int find_all)
 360 {
 361         int nr, on_list;
 362         struct commit_list *p, *best, *next, *last;
 363         int *weights;
 364
 365         show_list("bisection 2 entry", 0, 0, list);
 366
 367         /*
 368          * Count the number of total and tree-changing items on the
 369          * list, while reversing the list.
 370          */
 371         for (nr = on_list = 0, last = NULL, p = list;
 372              p;
 373              p = next) {
 374                 unsigned flags = p->item->object.flags;
 375
 376                 next = p->next;
 377                 if (flags & UNINTERESTING)
 378                         continue;
 379                 p->next = last;
 380                 last = p;
 381                 if (!(flags & TREESAME))
 382                         nr++;
 383                 on_list++;
 384         }
 385         list = last;
 386         show_list("bisection 2 sorted", 0, nr, list);
 387
 388         *all = nr;
 389         weights = xcalloc(on_list, sizeof(*weights));
 390
 391         /* Do the real work of finding bisection commit. */
 392         best = do_find_bisection(list, nr, weights, find_all);
 393         if (best) {
 394                 if (!find_all)
 395                         best->next = NULL;
 396                 *reaches = weight(best);
 397         }
 398         free(weights);
 399         return best;
 400 }
 401
 402 static int register_ref(const char *refname, const unsigned char *sha1,
 403                         int flags, void *cb_data)
 404 {
 405         if (!strcmp(refname, "bad")) {
 406                 current_bad_sha1 = xmalloc(20);
 407                 hashcpy(current_bad_sha1, sha1);
 408         } else if (starts_with(refname, "good-")) {
 409                 sha1_array_append(&good_revs, sha1);
 410         } else if (starts_with(refname, "skip-")) {
 411                 sha1_array_append(&skipped_revs, sha1);
 412         }
 413
 414         return 0;
 415 }
 416
 417 static int read_bisect_refs(void)
 418 {
 419         return for_each_ref_in("refs/bisect/", register_ref, NULL);
 420 }
 421
 422 static void read_bisect_paths(struct argv_array *array)
 423 {
 424         struct strbuf str = STRBUF_INIT;
 425         const char *filename = git_path("BISECT_NAMES");
 426         FILE *fp = fopen(filename, "r");
 427
 428         if (!fp)
 429                 die_errno("Could not open file '%s'", filename);
 430
 431         while (strbuf_getline(&str, fp, '\n') != EOF) {
 432                 strbuf_trim(&str);
 433                 if (sq_dequote_to_argv_array(str.buf, array))
 434                         die("Badly quoted content in file '%s': %s",
 435                             filename, str.buf);
 436         }
 437
 438         strbuf_release(&str);
 439         fclose(fp);
 440 }
 441
 442 static char *join_sha1_array_hex(struct sha1_array *array, char delim)
 443 {
 444         struct strbuf joined_hexs = STRBUF_INIT;
 445         int i;
 446
 447         for (i = 0; i < array->nr; i++) {
 448                 strbuf_addstr(&joined_hexs, sha1_to_hex(array->sha1[i]));
 449                 if (i + 1 < array->nr)
 450                         strbuf_addch(&joined_hexs, delim);
 451         }
 452
 453         return strbuf_detach(&joined_hexs, NULL);
 454 }
 455
 456 /*
 457  * In this function, passing a not NULL skipped_first is very special.
 458  * It means that we want to know if the first commit in the list is
 459  * skipped because we will want to test a commit away from it if it is
 460  * indeed skipped.
 461  * So if the first commit is skipped, we cannot take the shortcut to
 462  * just "return list" when we find the first non skipped commit, we
 463  * have to return a fully filtered list.
 464  *
 465  * We use (*skipped_first == -1) to mean "it has been found that the
 466  * first commit is not skipped". In this case *skipped_first is set back
 467  * to 0 just before the function returns.
 468  */
 469 struct commit_list *filter_skipped(struct commit_list *list,
 470                                    struct commit_list **tried,
 471                                    int show_all,
 472                                    int *count,
 473                                    int *skipped_first)
 474 {
 475         struct commit_list *filtered = NULL, **f = &filtered;
 476
 477         *tried = NULL;
 478
 479         if (skipped_first)
 480                 *skipped_first = 0;
 481         if (count)
 482                 *count = 0;
 483
 484         if (!skipped_revs.nr)
 485                 return list;
 486
 487         while (list) {
 488                 struct commit_list *next = list->next;
 489                 list->next = NULL;
 490                 if (0 <= sha1_array_lookup(&skipped_revs,
 491                                            list->item->object.sha1)) {
 492                         if (skipped_first && !*skipped_first)
 493                                 *skipped_first = 1;
 494                         /* Move current to tried list */
 495                         *tried = list;
 496                         tried = &list->next;
 497                 } else {
 498                         if (!show_all) {
 499                                 if (!skipped_first || !*skipped_first)
 500                                         return list;
 501                         } else if (skipped_first && !*skipped_first) {
 502                                 /* This means we know it's not skipped */
 503                                 *skipped_first = -1;
 504                         }
 505                         /* Move current to filtered list */
 506                         *f = list;
 507                         f = &list->next;
 508                         if (count)
 509                                 (*count)++;
 510                 }
 511                 list = next;
 512         }
 513
 514         if (skipped_first && *skipped_first == -1)
 515                 *skipped_first = 0;
 516
 517         return filtered;
 518 }
 519
 520 #define PRN_MODULO 32768
 521
 522 /*
 523  * This is a pseudo random number generator based on "man 3 rand".
 524  * It is not used properly because the seed is the argument and it
 525  * is increased by one between each call, but that should not matter
 526  * for this application.
 527  */
 528 static unsigned get_prn(unsigned count) {
 529         count = count * 1103515245 + 12345;
 530         return (count/65536) % PRN_MODULO;
 531 }
 532
 533 /*
 534  * Custom integer square root from
 535  * http://en.wikipedia.org/wiki/Integer_square_root
 536  */
 537 static int sqrti(int val)
 538 {
 539         float d, x = val;
 540
 541         if (val == 0)
 542                 return 0;
 543
 544         do {
 545                 float y = (x + (float)val / x) / 2;
 546                 d = (y > x) ? y - x : x - y;
 547                 x = y;
 548         } while (d >= 0.5);
 549
 550         return (int)x;
 551 }
 552
 553 static struct commit_list *skip_away(struct commit_list *list, int count)
 554 {
 555         struct commit_list *cur, *previous;
 556         int prn, index, i;
 557
 558         prn = get_prn(count);
 559         index = (count * prn / PRN_MODULO) * sqrti(prn) / sqrti(PRN_MODULO);
 560
 561         cur = list;
 562         previous = NULL;
 563
 564         for (i = 0; cur; cur = cur->next, i++) {
 565                 if (i == index) {
 566                         if (hashcmp(cur->item->object.sha1, current_bad_sha1))
 567                                 return cur;
 568                         if (previous)
 569                                 return previous;
 570                         return list;
 571                 }
 572                 previous = cur;
 573         }
 574
 575         return list;
 576 }
 577
 578 static struct commit_list *managed_skipped(struct commit_list *list,
 579                                            struct commit_list **tried)
 580 {
 581         int count, skipped_first;
 582
 583         *tried = NULL;
 584
 585         if (!skipped_revs.nr)
 586                 return list;
 587
 588         list = filter_skipped(list, tried, 0, &count, &skipped_first);
 589
 590         if (!skipped_first)
 591                 return list;
 592
 593         return skip_away(list, count);
 594 }
 595
 596 static void bisect_rev_setup(struct rev_info *revs, const char *prefix,
 597                              const char *bad_format, const char *good_format,
 598                              int read_paths)
 599 {
 600         struct argv_array rev_argv = ARGV_ARRAY_INIT;
 601         int i;
 602
 603         init_revisions(revs, prefix);
 604         revs->abbrev = 0;
 605         revs->commit_format = CMIT_FMT_UNSPECIFIED;
 606
 607         /* rev_argv.argv[0] will be ignored by setup_revisions */
 608         argv_array_push(&rev_argv, "bisect_rev_setup");
 609         argv_array_pushf(&rev_argv, bad_format, sha1_to_hex(current_bad_sha1));
 610         for (i = 0; i < good_revs.nr; i++)
 611                 argv_array_pushf(&rev_argv, good_format,
 612                                  sha1_to_hex(good_revs.sha1[i]));
 613         argv_array_push(&rev_argv, "--");
 614         if (read_paths)
 615                 read_bisect_paths(&rev_argv);
 616
 617         setup_revisions(rev_argv.argc, rev_argv.argv, revs, NULL);
 618         /* XXX leak rev_argv, as "revs" may still be pointing to it */
 619 }
 620
 621 static void bisect_common(struct rev_info *revs)
 622 {
 623         if (prepare_revision_walk(revs))
 624                 die("revision walk setup failed");
 625         if (revs->tree_objects)
 626                 mark_edges_uninteresting(revs, NULL);
 627 }
 628
 629 static void exit_if_skipped_commits(struct commit_list *tried,
 630                                     const unsigned char *bad)
 631 {
 632         if (!tried)
 633                 return;
 634
 635         printf("There are only 'skip'ped commits left to test.\n"
 636                "The first bad commit could be any of:\n");
 637         print_commit_list(tried, "%s\n", "%s\n");
 638         if (bad)
 639                 printf("%s\n", sha1_to_hex(bad));
 640         printf("We cannot bisect more!\n");
 641         exit(2);
 642 }
 643
 644 static int is_expected_rev(const unsigned char *sha1)
 645 {
 646         const char *filename = git_path("BISECT_EXPECTED_REV");
 647         struct stat st;
 648         struct strbuf str = STRBUF_INIT;
 649         FILE *fp;
 650         int res = 0;
 651
 652         if (stat(filename, &st) || !S_ISREG(st.st_mode))
 653                 return 0;
 654
 655         fp = fopen(filename, "r");
 656         if (!fp)
 657                 return 0;
 658
 659         if (strbuf_getline(&str, fp, '\n') != EOF)
 660                 res = !strcmp(str.buf, sha1_to_hex(sha1));
 661
 662         strbuf_release(&str);
 663         fclose(fp);
 664
 665         return res;
 666 }
 667
 668 static void mark_expected_rev(char *bisect_rev_hex)
 669 {
 670         int len = strlen(bisect_rev_hex);
 671         const char *filename = git_path("BISECT_EXPECTED_REV");
 672         int fd = open(filename, O_CREAT | O_TRUNC | O_WRONLY, 0600);
 673
 674         if (fd < 0)
 675                 die_errno("could not create file '%s'", filename);
 676
 677         bisect_rev_hex[len] = '\n';
 678         write_or_die(fd, bisect_rev_hex, len + 1);
 679         bisect_rev_hex[len] = '\0';
 680
 681         if (close(fd) < 0)
 682                 die("closing file %s: %s", filename, strerror(errno));
 683 }
 684
 685 static int bisect_checkout(char *bisect_rev_hex, int no_checkout)
 686 {
 687
 688         mark_expected_rev(bisect_rev_hex);
 689
 690         argv_checkout[2] = bisect_rev_hex;
 691         if (no_checkout) {
 692                 argv_update_ref[3] = bisect_rev_hex;
 693                 if (run_command_v_opt(argv_update_ref, RUN_GIT_CMD))
 694                         die("update-ref --no-deref HEAD failed on %s",
 695                             bisect_rev_hex);
 696         } else {
 697                 int res;
 698                 res = run_command_v_opt(argv_checkout, RUN_GIT_CMD);
 699                 if (res)
 700                         exit(res);
 701         }
 702
 703         argv_show_branch[1] = bisect_rev_hex;
 704         return run_command_v_opt(argv_show_branch, RUN_GIT_CMD);
 705 }
 706
 707 static struct commit *get_commit_reference(const unsigned char *sha1)
 708 {
 709         struct commit *r = lookup_commit_reference(sha1);
 710         if (!r)
 711                 die("Not a valid commit name %s", sha1_to_hex(sha1));
 712         return r;
 713 }
 714
 715 static struct commit **get_bad_and_good_commits(int *rev_nr)
 716 {
 717         int len = 1 + good_revs.nr;
 718         struct commit **rev = xmalloc(len * sizeof(*rev));
 719         int i, n = 0;
 720
 721         rev[n++] = get_commit_reference(current_bad_sha1);
 722         for (i = 0; i < good_revs.nr; i++)
 723                 rev[n++] = get_commit_reference(good_revs.sha1[i]);
 724         *rev_nr = n;
 725
 726         return rev;
 727 }
 728
 729 static void handle_bad_merge_base(void)
 730 {
 731         if (is_expected_rev(current_bad_sha1)) {
 732                 char *bad_hex = sha1_to_hex(current_bad_sha1);
 733                 char *good_hex = join_sha1_array_hex(&good_revs, ' ');
 734
 735                 fprintf(stderr, "The merge base %s is bad.\n"
 736                         "This means the bug has been fixed "
 737                         "between %s and [%s].\n",
 738                         bad_hex, bad_hex, good_hex);
 739
 740                 exit(3);
 741         }
 742
 743         fprintf(stderr, "Some good revs are not ancestor of the bad rev.\n"
 744                 "git bisect cannot work properly in this case.\n"
 745                 "Maybe you mistake good and bad revs?\n");
 746         exit(1);
 747 }
 748
 749 static void handle_skipped_merge_base(const unsigned char *mb)
 750 {
 751         char *mb_hex = sha1_to_hex(mb);
 752         char *bad_hex = sha1_to_hex(current_bad_sha1);
 753         char *good_hex = join_sha1_array_hex(&good_revs, ' ');
 754
 755         warning("the merge base between %s and [%s] "
 756                 "must be skipped.\n"
 757                 "So we cannot be sure the first bad commit is "
 758                 "between %s and %s.\n"
 759                 "We continue anyway.",
 760                 bad_hex, good_hex, mb_hex, bad_hex);
 761         free(good_hex);
 762 }
 763
 764 /*
 765  * "check_merge_bases" checks that merge bases are not "bad".
 766  *
 767  * - If one is "bad", it means the user assumed something wrong
 768  * and we must exit with a non 0 error code.
 769  * - If one is "good", that's good, we have nothing to do.
 770  * - If one is "skipped", we can't know but we should warn.
 771  * - If we don't know, we should check it out and ask the user to test.
 772  */
 773 static void check_merge_bases(int no_checkout)
 774 {
 775         struct commit_list *result;
 776         int rev_nr;
 777         struct commit **rev = get_bad_and_good_commits(&rev_nr);
 778
 779         result = get_merge_bases_many(rev[0], rev_nr - 1, rev + 1, 0);
 780
 781         for (; result; result = result->next) {
 782                 const unsigned char *mb = result->item->object.sha1;
 783                 if (!hashcmp(mb, current_bad_sha1)) {
 784                         handle_bad_merge_base();
 785                 } else if (0 <= sha1_array_lookup(&good_revs, mb)) {
 786                         continue;
 787                 } else if (0 <= sha1_array_lookup(&skipped_revs, mb)) {
 788                         handle_skipped_merge_base(mb);
 789                 } else {
 790                         printf("Bisecting: a merge base must be tested\n");
 791                         exit(bisect_checkout(sha1_to_hex(mb), no_checkout));
 792                 }
 793         }
 794
 795         free(rev);
 796         free_commit_list(result);
 797 }
 798
 799 static int check_ancestors(const char *prefix)
 800 {
 801         struct rev_info revs;
 802         struct object_array pending_copy;
 803         int res;
 804
 805         bisect_rev_setup(&revs, prefix, "^%s", "%s", 0);
 806
 807         /* Save pending objects, so they can be cleaned up later. */
 808         pending_copy = revs.pending;
 809         revs.leak_pending = 1;
 810
 811         /*
 812          * bisect_common calls prepare_revision_walk right away, which
 813          * (together with .leak_pending = 1) makes us the sole owner of
 814          * the list of pending objects.
 815          */
 816         bisect_common(&revs);
 817         res = (revs.commits != NULL);
 818
 819         /* Clean up objects used, as they will be reused. */
 820         clear_commit_marks_for_object_array(&pending_copy, ALL_REV_FLAGS);
 821         free(pending_copy.objects);
 822
 823         return res;
 824 }
 825
 826 /*
 827  * "check_good_are_ancestors_of_bad" checks that all "good" revs are
 828  * ancestor of the "bad" rev.
 829  *
 830  * If that's not the case, we need to check the merge bases.
 831  * If a merge base must be tested by the user, its source code will be
 832  * checked out to be tested by the user and we will exit.
 833  */
 834 static void check_good_are_ancestors_of_bad(const char *prefix, int no_checkout)
 835 {
 836         char *filename = git_pathdup("BISECT_ANCESTORS_OK");
 837         struct stat st;
 838         int fd;
 839
 840         if (!current_bad_sha1)
 841                 die("a bad revision is needed");
 842
 843         /* Check if file BISECT_ANCESTORS_OK exists. */
 844         if (!stat(filename, &st) && S_ISREG(st.st_mode))
 845                 goto done;
 846
 847         /* Bisecting with no good rev is ok. */
 848         if (good_revs.nr == 0)
 849                 goto done;
 850
 851         /* Check if all good revs are ancestor of the bad rev. */
 852         if (check_ancestors(prefix))
 853                 check_merge_bases(no_checkout);
 854
 855         /* Create file BISECT_ANCESTORS_OK. */
 856         fd = open(filename, O_CREAT | O_TRUNC | O_WRONLY, 0600);
 857         if (fd < 0)
 858                 warning("could not create file '%s': %s",
 859                         filename, strerror(errno));
 860         else
 861                 close(fd);
 862  done:
 863         free(filename);
 864 }
 865
 866 /*
 867  * This does "git diff-tree --pretty COMMIT" without one fork+exec.
 868  */
 869 static void show_diff_tree(const char *prefix, struct commit *commit)
 870 {
 871         struct rev_info opt;
 872
 873         /* diff-tree init */
 874         init_revisions(&opt, prefix);
 875         git_config(git_diff_basic_config, NULL); /* no "diff" UI options */
 876         opt.abbrev = 0;
 877         opt.diff = 1;
 878
 879         /* This is what "--pretty" does */
 880         opt.verbose_header = 1;
 881         opt.use_terminator = 0;
 882         opt.commit_format = CMIT_FMT_DEFAULT;
 883
 884         /* diff-tree init */
 885         if (!opt.diffopt.output_format)
 886                 opt.diffopt.output_format = DIFF_FORMAT_RAW;
 887
 888         log_tree_commit(&opt, commit);
 889 }
 890
 891 /*
 892  * We use the convention that exiting with an exit code 10 means that
 893  * the bisection process finished successfully.
 894  * In this case the calling shell script should exit 0.
 895  *
 896  * If no_checkout is non-zero, the bisection process does not
 897  * checkout the trial commit but instead simply updates BISECT_HEAD.
 898  */
 899 int bisect_next_all(const char *prefix, int no_checkout)
 900 {
 901         struct rev_info revs;
 902         struct commit_list *tried;
 903         int reaches = 0, all = 0, nr, steps;
 904         const unsigned char *bisect_rev;
 905         char bisect_rev_hex[41];
 906
 907         if (read_bisect_refs())
 908                 die("reading bisect refs failed");
 909
 910         check_good_are_ancestors_of_bad(prefix, no_checkout);
 911
 912         bisect_rev_setup(&revs, prefix, "%s", "^%s", 1);
 913         revs.limited = 1;
 914
 915         bisect_common(&revs);
 916
 917         revs.commits = find_bisection(revs.commits, &reaches, &all,
 918                                        !!skipped_revs.nr);
 919         revs.commits = managed_skipped(revs.commits, &tried);
 920
 921         if (!revs.commits) {
 922                 /*
 923                  * We should exit here only if the "bad"
 924                  * commit is also a "skip" commit.
 925                  */
 926                 exit_if_skipped_commits(tried, NULL);
 927
 928                 printf("%s was both good and bad\n",
 929                        sha1_to_hex(current_bad_sha1));
 930                 exit(1);
 931         }
 932
 933         if (!all) {
 934                 fprintf(stderr, "No testable commit found.\n"
 935                         "Maybe you started with bad path parameters?\n");
 936                 exit(4);
 937         }
 938
 939         bisect_rev = revs.commits->item->object.sha1;
 940         memcpy(bisect_rev_hex, sha1_to_hex(bisect_rev), 41);
 941
 942         if (!hashcmp(bisect_rev, current_bad_sha1)) {
 943                 exit_if_skipped_commits(tried, current_bad_sha1);
 944                 printf("%s is the first bad commit\n", bisect_rev_hex);
 945                 show_diff_tree(prefix, revs.commits->item);
 946                 /* This means the bisection process succeeded. */
 947                 exit(10);
 948         }
 949
 950         nr = all - reaches - 1;
 951         steps = estimate_bisect_steps(all);
 952         printf("Bisecting: %d revision%s left to test after this "
 953                "(roughly %d step%s)\n", nr, (nr == 1 ? "" : "s"),
 954                steps, (steps == 1 ? "" : "s"));
 955
 956         return bisect_checkout(bisect_rev_hex, no_checkout);
 957 }
 958
 959 static inline int log2i(int n)
 960 {
 961         int log2 = 0;
 962
 963         for (; n > 1; n >>= 1)
 964                 log2++;
 965
 966         return log2;
 967 }
 968
 969 static inline int exp2i(int n)
 970 {
 971         return 1 << n;
 972 }
 973
 974 /*
 975  * Estimate the number of bisect steps left (after the current step)
 976  *
 977  * For any x between 0 included and 2^n excluded, the probability for
 978  * n - 1 steps left looks like:
 979  *
 980  * P(2^n + x) == (2^n - x) / (2^n + x)
 981  *
 982  * and P(2^n + x) < 0.5 means 2^n < 3x
 983  */
 984 int estimate_bisect_steps(int all)
 985 {
 986         int n, x, e;
 987
 988         if (all < 3)
 989                 return 0;
 990
 991         n = log2i(all);
 992         e = exp2i(n);
 993         x = all - e;
 994
 995         return (e < 3 * x) ? n : n - 1;
 996 }