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 const 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 /* bits #0-15 in revision.h */
  25
  26 #define COUNTED         (1u<<16)
  27
  28 /*
  29  * This is a truly stupid algorithm, but it's only
  30  * used for bisection, and we just don't care enough.
  31  *
  32  * We care just barely enough to avoid recursing for
  33  * non-merge entries.
  34  */
  35 static int count_distance(struct commit_list *entry)
  36 {
  37         int nr = 0;
  38
  39         while (entry) {
  40                 struct commit *commit = entry->item;
  41                 struct commit_list *p;
  42
  43                 if (commit->object.flags & (UNINTERESTING | COUNTED))
  44                         break;
  45                 if (!(commit->object.flags & TREESAME))
  46                         nr++;
  47                 commit->object.flags |= COUNTED;
  48                 p = commit->parents;
  49                 entry = p;
  50                 if (p) {
  51                         p = p->next;
  52                         while (p) {
  53                                 nr += count_distance(p);
  54                                 p = p->next;
  55                         }
  56                 }
  57         }
  58
  59         return nr;
  60 }
  61
  62 static void clear_distance(struct commit_list *list)
  63 {
  64         while (list) {
  65                 struct commit *commit = list->item;
  66                 commit->object.flags &= ~COUNTED;
  67                 list = list->next;
  68         }
  69 }
  70
  71 #define DEBUG_BISECT 0
  72
  73 static inline int weight(struct commit_list *elem)
  74 {
  75         return *((int*)(elem->item->util));
  76 }
  77
  78 static inline void weight_set(struct commit_list *elem, int weight)
  79 {
  80         *((int*)(elem->item->util)) = weight;
  81 }
  82
  83 static int count_interesting_parents(struct commit *commit)
  84 {
  85         struct commit_list *p;
  86         int count;
  87
  88         for (count = 0, p = commit->parents; p; p = p->next) {
  89                 if (p->item->object.flags & UNINTERESTING)
  90                         continue;
  91                 count++;
  92         }
  93         return count;
  94 }
  95
  96 static inline int halfway(struct commit_list *p, int nr)
  97 {
  98         /*
  99          * Don't short-cut something we are not going to return!
 100          */
 101         if (p->item->object.flags & TREESAME)
 102                 return 0;
 103         if (DEBUG_BISECT)
 104                 return 0;
 105         /*
 106          * 2 and 3 are halfway of 5.
 107          * 3 is halfway of 6 but 2 and 4 are not.
 108          */
 109         switch (2 * weight(p) - nr) {
 110         case -1: case 0: case 1:
 111                 return 1;
 112         default:
 113                 return 0;
 114         }
 115 }
 116
 117 #if !DEBUG_BISECT
 118 #define show_list(a,b,c,d) do { ; } while (0)
 119 #else
 120 static void show_list(const char *debug, int counted, int nr,
 121                       struct commit_list *list)
 122 {
 123         struct commit_list *p;
 124
 125         fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
 126
 127         for (p = list; p; p = p->next) {
 128                 struct commit_list *pp;
 129                 struct commit *commit = p->item;
 130                 unsigned flags = commit->object.flags;
 131                 enum object_type type;
 132                 unsigned long size;
 133                 char *buf = read_sha1_file(commit->object.sha1, &type, &size);
 134                 const char *subject_start;
 135                 int subject_len;
 136
 137                 fprintf(stderr, "%c%c%c ",
 138                         (flags & TREESAME) ? ' ' : 'T',
 139                         (flags & UNINTERESTING) ? 'U' : ' ',
 140                         (flags & COUNTED) ? 'C' : ' ');
 141                 if (commit->util)
 142                         fprintf(stderr, "%3d", weight(p));
 143                 else
 144                         fprintf(stderr, "---");
 145                 fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 146                 for (pp = commit->parents; pp; pp = pp->next)
 147                         fprintf(stderr, " %.*s", 8,
 148                                 sha1_to_hex(pp->item->object.sha1));
 149
 150                 subject_len = find_commit_subject(buf, &subject_start);
 151                 if (subject_len)
 152                         fprintf(stderr, " %.*s", subject_len, subject_start);
 153                 fprintf(stderr, "\n");
 154         }
 155 }
 156 #endif /* DEBUG_BISECT */
 157
 158 static struct commit_list *best_bisection(struct commit_list *list, int nr)
 159 {
 160         struct commit_list *p, *best;
 161         int best_distance = -1;
 162
 163         best = list;
 164         for (p = list; p; p = p->next) {
 165                 int distance;
 166                 unsigned flags = p->item->object.flags;
 167
 168                 if (flags & TREESAME)
 169                         continue;
 170                 distance = weight(p);
 171                 if (nr - distance < distance)
 172                         distance = nr - distance;
 173                 if (distance > best_distance) {
 174                         best = p;
 175                         best_distance = distance;
 176                 }
 177         }
 178
 179         return best;
 180 }
 181
 182 struct commit_dist {
 183         struct commit *commit;
 184         int distance;
 185 };
 186
 187 static int compare_commit_dist(const void *a_, const void *b_)
 188 {
 189         struct commit_dist *a, *b;
 190
 191         a = (struct commit_dist *)a_;
 192         b = (struct commit_dist *)b_;
 193         if (a->distance != b->distance)
 194                 return b->distance - a->distance; /* desc sort */
 195         return hashcmp(a->commit->object.sha1, b->commit->object.sha1);
 196 }
 197
 198 static struct commit_list *best_bisection_sorted(struct commit_list *list, int nr)
 199 {
 200         struct commit_list *p;
 201         struct commit_dist *array = xcalloc(nr, sizeof(*array));
 202         int cnt, i;
 203
 204         for (p = list, cnt = 0; p; p = p->next) {
 205                 int distance;
 206                 unsigned flags = p->item->object.flags;
 207
 208                 if (flags & TREESAME)
 209                         continue;
 210                 distance = weight(p);
 211                 if (nr - distance < distance)
 212                         distance = nr - distance;
 213                 array[cnt].commit = p->item;
 214                 array[cnt].distance = distance;
 215                 cnt++;
 216         }
 217         qsort(array, cnt, sizeof(*array), compare_commit_dist);
 218         for (p = list, i = 0; i < cnt; i++) {
 219                 struct name_decoration *r = xmalloc(sizeof(*r) + 100);
 220                 struct object *obj = &(array[i].commit->object);
 221
 222                 sprintf(r->name, "dist=%d", array[i].distance);
 223                 r->next = add_decoration(&name_decoration, obj, r);
 224                 p->item = array[i].commit;
 225                 p = p->next;
 226         }
 227         if (p)
 228                 p->next = NULL;
 229         free(array);
 230         return list;
 231 }
 232
 233 /*
 234  * zero or positive weight is the number of interesting commits it can
 235  * reach, including itself.  Especially, weight = 0 means it does not
 236  * reach any tree-changing commits (e.g. just above uninteresting one
 237  * but traversal is with pathspec).
 238  *
 239  * weight = -1 means it has one parent and its distance is yet to
 240  * be computed.
 241  *
 242  * weight = -2 means it has more than one parent and its distance is
 243  * unknown.  After running count_distance() first, they will get zero
 244  * or positive distance.
 245  */
 246 static struct commit_list *do_find_bisection(struct commit_list *list,
 247                                              int nr, int *weights,
 248                                              int find_all)
 249 {
 250         int n, counted;
 251         struct commit_list *p;
 252
 253         counted = 0;
 254
 255         for (n = 0, p = list; p; p = p->next) {
 256                 struct commit *commit = p->item;
 257                 unsigned flags = commit->object.flags;
 258
 259                 p->item->util = &weights[n++];
 260                 switch (count_interesting_parents(commit)) {
 261                 case 0:
 262                         if (!(flags & TREESAME)) {
 263                                 weight_set(p, 1);
 264                                 counted++;
 265                                 show_list("bisection 2 count one",
 266                                           counted, nr, list);
 267                         }
 268                         /*
 269                          * otherwise, it is known not to reach any
 270                          * tree-changing commit and gets weight 0.
 271                          */
 272                         break;
 273                 case 1:
 274                         weight_set(p, -1);
 275                         break;
 276                 default:
 277                         weight_set(p, -2);
 278                         break;
 279                 }
 280         }
 281
 282         show_list("bisection 2 initialize", counted, nr, list);
 283
 284         /*
 285          * If you have only one parent in the resulting set
 286          * then you can reach one commit more than that parent
 287          * can reach.  So we do not have to run the expensive
 288          * count_distance() for single strand of pearls.
 289          *
 290          * However, if you have more than one parents, you cannot
 291          * just add their distance and one for yourself, since
 292          * they usually reach the same ancestor and you would
 293          * end up counting them twice that way.
 294          *
 295          * So we will first count distance of merges the usual
 296          * way, and then fill the blanks using cheaper algorithm.
 297          */
 298         for (p = list; p; p = p->next) {
 299                 if (p->item->object.flags & UNINTERESTING)
 300                         continue;
 301                 if (weight(p) != -2)
 302                         continue;
 303                 weight_set(p, count_distance(p));
 304                 clear_distance(list);
 305
 306                 /* Does it happen to be at exactly half-way? */
 307                 if (!find_all && halfway(p, nr))
 308                         return p;
 309                 counted++;
 310         }
 311
 312         show_list("bisection 2 count_distance", counted, nr, list);
 313
 314         while (counted < nr) {
 315                 for (p = list; p; p = p->next) {
 316                         struct commit_list *q;
 317                         unsigned flags = p->item->object.flags;
 318
 319                         if (0 <= weight(p))
 320                                 continue;
 321                         for (q = p->item->parents; q; q = q->next) {
 322                                 if (q->item->object.flags & UNINTERESTING)
 323                                         continue;
 324                                 if (0 <= weight(q))
 325                                         break;
 326                         }
 327                         if (!q)
 328                                 continue;
 329
 330                         /*
 331                          * weight for p is unknown but q is known.
 332                          * add one for p itself if p is to be counted,
 333                          * otherwise inherit it from q directly.
 334                          */
 335                         if (!(flags & TREESAME)) {
 336                                 weight_set(p, weight(q)+1);
 337                                 counted++;
 338                                 show_list("bisection 2 count one",
 339                                           counted, nr, list);
 340                         }
 341                         else
 342                                 weight_set(p, weight(q));
 343
 344                         /* Does it happen to be at exactly half-way? */
 345                         if (!find_all && halfway(p, nr))
 346                                 return p;
 347                 }
 348         }
 349
 350         show_list("bisection 2 counted all", counted, nr, list);
 351
 352         if (!find_all)
 353                 return best_bisection(list, nr);
 354         else
 355                 return best_bisection_sorted(list, nr);
 356 }
 357
 358 struct commit_list *find_bisection(struct commit_list *list,
 359                                           int *reaches, int *all,
 360                                           int find_all)
 361 {
 362         int nr, on_list;
 363         struct commit_list *p, *best, *next, *last;
 364         int *weights;
 365
 366         show_list("bisection 2 entry", 0, 0, list);
 367
 368         /*
 369          * Count the number of total and tree-changing items on the
 370          * list, while reversing the list.
 371          */
 372         for (nr = on_list = 0, last = NULL, p = list;
 373              p;
 374              p = next) {
 375                 unsigned flags = p->item->object.flags;
 376
 377                 next = p->next;
 378                 if (flags & UNINTERESTING)
 379                         continue;
 380                 p->next = last;
 381                 last = p;
 382                 if (!(flags & TREESAME))
 383                         nr++;
 384                 on_list++;
 385         }
 386         list = last;
 387         show_list("bisection 2 sorted", 0, nr, list);
 388
 389         *all = nr;
 390         weights = xcalloc(on_list, sizeof(*weights));
 391
 392         /* Do the real work of finding bisection commit. */
 393         best = do_find_bisection(list, nr, weights, find_all);
 394         if (best) {
 395                 if (!find_all)
 396                         best->next = NULL;
 397                 *reaches = weight(best);
 398         }
 399         free(weights);
 400         return best;
 401 }
 402
 403 static int register_ref(const char *refname, const unsigned char *sha1,
 404                         int flags, void *cb_data)
 405 {
 406         if (!strcmp(refname, "bad")) {
 407                 current_bad_sha1 = sha1;
 408         } else if (!prefixcmp(refname, "good-")) {
 409                 sha1_array_append(&good_revs, sha1);
 410         } else if (!prefixcmp(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->commits, 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         int res;
 688
 689         mark_expected_rev(bisect_rev_hex);
 690
 691         argv_checkout[2] = bisect_rev_hex;
 692         if (no_checkout) {
 693                 argv_update_ref[3] = bisect_rev_hex;
 694                 if (run_command_v_opt(argv_update_ref, RUN_GIT_CMD))
 695                         die("update-ref --no-deref HEAD failed on %s",
 696                             bisect_rev_hex);
 697         } else {
 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 }