bisect.con commit bisect: fix off-by-one error in `best_bisection_sorted()` (7c11718)
   1#include "cache.h"
   2#include "config.h"
   3#include "commit.h"
   4#include "diff.h"
   5#include "revision.h"
   6#include "refs.h"
   7#include "list-objects.h"
   8#include "quote.h"
   9#include "sha1-lookup.h"
  10#include "run-command.h"
  11#include "log-tree.h"
  12#include "bisect.h"
  13#include "sha1-array.h"
  14#include "argv-array.h"
  15
  16static struct oid_array good_revs;
  17static struct oid_array skipped_revs;
  18
  19static struct object_id *current_bad_oid;
  20
  21static const char *argv_checkout[] = {"checkout", "-q", NULL, "--", NULL};
  22static const char *argv_show_branch[] = {"show-branch", NULL, NULL};
  23
  24static const char *term_bad;
  25static const char *term_good;
  26
  27/* Remember to update object flag allocation in object.h */
  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 */
  37static 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
  64static 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
  75static inline int weight(struct commit_list *elem)
  76{
  77        return *((int*)(elem->item->util));
  78}
  79
  80static inline void weight_set(struct commit_list *elem, int weight)
  81{
  82        *((int*)(elem->item->util)) = weight;
  83}
  84
  85static 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
  98static 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
 122static 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                const char *subject_start;
 137                int subject_len;
 138
 139                fprintf(stderr, "%c%c%c ",
 140                        (flags & TREESAME) ? ' ' : 'T',
 141                        (flags & UNINTERESTING) ? 'U' : ' ',
 142                        (flags & COUNTED) ? 'C' : ' ');
 143                if (commit->util)
 144                        fprintf(stderr, "%3d", weight(p));
 145                else
 146                        fprintf(stderr, "---");
 147                fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 148                for (pp = commit->parents; pp; pp = pp->next)
 149                        fprintf(stderr, " %.*s", 8,
 150                                sha1_to_hex(pp->item->object.sha1));
 151
 152                subject_len = find_commit_subject(buf, &subject_start);
 153                if (subject_len)
 154                        fprintf(stderr, " %.*s", subject_len, subject_start);
 155                fprintf(stderr, "\n");
 156        }
 157}
 158#endif /* DEBUG_BISECT */
 159
 160static struct commit_list *best_bisection(struct commit_list *list, int nr)
 161{
 162        struct commit_list *p, *best;
 163        int best_distance = -1;
 164
 165        best = list;
 166        for (p = list; p; p = p->next) {
 167                int distance;
 168                unsigned flags = p->item->object.flags;
 169
 170                if (flags & TREESAME)
 171                        continue;
 172                distance = weight(p);
 173                if (nr - distance < distance)
 174                        distance = nr - distance;
 175                if (distance > best_distance) {
 176                        best = p;
 177                        best_distance = distance;
 178                }
 179        }
 180
 181        return best;
 182}
 183
 184struct commit_dist {
 185        struct commit *commit;
 186        int distance;
 187};
 188
 189static int compare_commit_dist(const void *a_, const void *b_)
 190{
 191        struct commit_dist *a, *b;
 192
 193        a = (struct commit_dist *)a_;
 194        b = (struct commit_dist *)b_;
 195        if (a->distance != b->distance)
 196                return b->distance - a->distance; /* desc sort */
 197        return oidcmp(&a->commit->object.oid, &b->commit->object.oid);
 198}
 199
 200static struct commit_list *best_bisection_sorted(struct commit_list *list, int nr)
 201{
 202        struct commit_list *p;
 203        struct commit_dist *array = xcalloc(nr, sizeof(*array));
 204        struct strbuf buf = STRBUF_INIT;
 205        int cnt, i;
 206
 207        for (p = list, cnt = 0; p; p = p->next) {
 208                int distance;
 209                unsigned flags = p->item->object.flags;
 210
 211                if (flags & TREESAME)
 212                        continue;
 213                distance = weight(p);
 214                if (nr - distance < distance)
 215                        distance = nr - distance;
 216                array[cnt].commit = p->item;
 217                array[cnt].distance = distance;
 218                cnt++;
 219        }
 220        QSORT(array, cnt, compare_commit_dist);
 221        for (p = list, i = 0; i < cnt; i++) {
 222                struct object *obj = &(array[i].commit->object);
 223
 224                strbuf_reset(&buf);
 225                strbuf_addf(&buf, "dist=%d", array[i].distance);
 226                add_name_decoration(DECORATION_NONE, buf.buf, obj);
 227
 228                p->item = array[i].commit;
 229                if (i < cnt - 1)
 230                        p = p->next;
 231        }
 232        free_commit_list(p->next);
 233        p->next = NULL;
 234        strbuf_release(&buf);
 235        free(array);
 236        return list;
 237}
 238
 239/*
 240 * zero or positive weight is the number of interesting commits it can
 241 * reach, including itself.  Especially, weight = 0 means it does not
 242 * reach any tree-changing commits (e.g. just above uninteresting one
 243 * but traversal is with pathspec).
 244 *
 245 * weight = -1 means it has one parent and its distance is yet to
 246 * be computed.
 247 *
 248 * weight = -2 means it has more than one parent and its distance is
 249 * unknown.  After running count_distance() first, they will get zero
 250 * or positive distance.
 251 */
 252static struct commit_list *do_find_bisection(struct commit_list *list,
 253                                             int nr, int *weights,
 254                                             int find_all)
 255{
 256        int n, counted;
 257        struct commit_list *p;
 258
 259        counted = 0;
 260
 261        for (n = 0, p = list; p; p = p->next) {
 262                struct commit *commit = p->item;
 263                unsigned flags = commit->object.flags;
 264
 265                p->item->util = &weights[n++];
 266                switch (count_interesting_parents(commit)) {
 267                case 0:
 268                        if (!(flags & TREESAME)) {
 269                                weight_set(p, 1);
 270                                counted++;
 271                                show_list("bisection 2 count one",
 272                                          counted, nr, list);
 273                        }
 274                        /*
 275                         * otherwise, it is known not to reach any
 276                         * tree-changing commit and gets weight 0.
 277                         */
 278                        break;
 279                case 1:
 280                        weight_set(p, -1);
 281                        break;
 282                default:
 283                        weight_set(p, -2);
 284                        break;
 285                }
 286        }
 287
 288        show_list("bisection 2 initialize", counted, nr, list);
 289
 290        /*
 291         * If you have only one parent in the resulting set
 292         * then you can reach one commit more than that parent
 293         * can reach.  So we do not have to run the expensive
 294         * count_distance() for single strand of pearls.
 295         *
 296         * However, if you have more than one parents, you cannot
 297         * just add their distance and one for yourself, since
 298         * they usually reach the same ancestor and you would
 299         * end up counting them twice that way.
 300         *
 301         * So we will first count distance of merges the usual
 302         * way, and then fill the blanks using cheaper algorithm.
 303         */
 304        for (p = list; p; p = p->next) {
 305                if (p->item->object.flags & UNINTERESTING)
 306                        continue;
 307                if (weight(p) != -2)
 308                        continue;
 309                weight_set(p, count_distance(p));
 310                clear_distance(list);
 311
 312                /* Does it happen to be at exactly half-way? */
 313                if (!find_all && halfway(p, nr))
 314                        return p;
 315                counted++;
 316        }
 317
 318        show_list("bisection 2 count_distance", counted, nr, list);
 319
 320        while (counted < nr) {
 321                for (p = list; p; p = p->next) {
 322                        struct commit_list *q;
 323                        unsigned flags = p->item->object.flags;
 324
 325                        if (0 <= weight(p))
 326                                continue;
 327                        for (q = p->item->parents; q; q = q->next) {
 328                                if (q->item->object.flags & UNINTERESTING)
 329                                        continue;
 330                                if (0 <= weight(q))
 331                                        break;
 332                        }
 333                        if (!q)
 334                                continue;
 335
 336                        /*
 337                         * weight for p is unknown but q is known.
 338                         * add one for p itself if p is to be counted,
 339                         * otherwise inherit it from q directly.
 340                         */
 341                        if (!(flags & TREESAME)) {
 342                                weight_set(p, weight(q)+1);
 343                                counted++;
 344                                show_list("bisection 2 count one",
 345                                          counted, nr, list);
 346                        }
 347                        else
 348                                weight_set(p, weight(q));
 349
 350                        /* Does it happen to be at exactly half-way? */
 351                        if (!find_all && halfway(p, nr))
 352                                return p;
 353                }
 354        }
 355
 356        show_list("bisection 2 counted all", counted, nr, list);
 357
 358        if (!find_all)
 359                return best_bisection(list, nr);
 360        else
 361                return best_bisection_sorted(list, nr);
 362}
 363
 364void find_bisection(struct commit_list **commit_list, int *reaches,
 365                    int *all, int find_all)
 366{
 367        int nr, on_list;
 368        struct commit_list *list, *p, *best, *next, *last;
 369        int *weights;
 370
 371        show_list("bisection 2 entry", 0, 0, *commit_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 = *commit_list;
 378             p;
 379             p = next) {
 380                unsigned flags = p->item->object.flags;
 381
 382                next = p->next;
 383                if (flags & UNINTERESTING) {
 384                        free(p);
 385                        continue;
 386                }
 387                p->next = last;
 388                last = p;
 389                if (!(flags & TREESAME))
 390                        nr++;
 391                on_list++;
 392        }
 393        list = last;
 394        show_list("bisection 2 sorted", 0, nr, list);
 395
 396        *all = nr;
 397        weights = xcalloc(on_list, sizeof(*weights));
 398
 399        /* Do the real work of finding bisection commit. */
 400        best = do_find_bisection(list, nr, weights, find_all);
 401        if (best) {
 402                if (!find_all)
 403                        best->next = NULL;
 404                *reaches = weight(best);
 405        }
 406        free(weights);
 407        *commit_list = best;
 408}
 409
 410static int register_ref(const char *refname, const struct object_id *oid,
 411                        int flags, void *cb_data)
 412{
 413        struct strbuf good_prefix = STRBUF_INIT;
 414        strbuf_addstr(&good_prefix, term_good);
 415        strbuf_addstr(&good_prefix, "-");
 416
 417        if (!strcmp(refname, term_bad)) {
 418                current_bad_oid = xmalloc(sizeof(*current_bad_oid));
 419                oidcpy(current_bad_oid, oid);
 420        } else if (starts_with(refname, good_prefix.buf)) {
 421                oid_array_append(&good_revs, oid);
 422        } else if (starts_with(refname, "skip-")) {
 423                oid_array_append(&skipped_revs, oid);
 424        }
 425
 426        strbuf_release(&good_prefix);
 427
 428        return 0;
 429}
 430
 431static int read_bisect_refs(void)
 432{
 433        return for_each_ref_in("refs/bisect/", register_ref, NULL);
 434}
 435
 436static GIT_PATH_FUNC(git_path_bisect_names, "BISECT_NAMES")
 437static GIT_PATH_FUNC(git_path_bisect_expected_rev, "BISECT_EXPECTED_REV")
 438static GIT_PATH_FUNC(git_path_bisect_terms, "BISECT_TERMS")
 439
 440static void read_bisect_paths(struct argv_array *array)
 441{
 442        struct strbuf str = STRBUF_INIT;
 443        const char *filename = git_path_bisect_names();
 444        FILE *fp = xfopen(filename, "r");
 445
 446        while (strbuf_getline_lf(&str, fp) != EOF) {
 447                strbuf_trim(&str);
 448                if (sq_dequote_to_argv_array(str.buf, array))
 449                        die(_("Badly quoted content in file '%s': %s"),
 450                            filename, str.buf);
 451        }
 452
 453        strbuf_release(&str);
 454        fclose(fp);
 455}
 456
 457static char *join_sha1_array_hex(struct oid_array *array, char delim)
 458{
 459        struct strbuf joined_hexs = STRBUF_INIT;
 460        int i;
 461
 462        for (i = 0; i < array->nr; i++) {
 463                strbuf_addstr(&joined_hexs, oid_to_hex(array->oid + i));
 464                if (i + 1 < array->nr)
 465                        strbuf_addch(&joined_hexs, delim);
 466        }
 467
 468        return strbuf_detach(&joined_hexs, NULL);
 469}
 470
 471/*
 472 * In this function, passing a not NULL skipped_first is very special.
 473 * It means that we want to know if the first commit in the list is
 474 * skipped because we will want to test a commit away from it if it is
 475 * indeed skipped.
 476 * So if the first commit is skipped, we cannot take the shortcut to
 477 * just "return list" when we find the first non skipped commit, we
 478 * have to return a fully filtered list.
 479 *
 480 * We use (*skipped_first == -1) to mean "it has been found that the
 481 * first commit is not skipped". In this case *skipped_first is set back
 482 * to 0 just before the function returns.
 483 */
 484struct commit_list *filter_skipped(struct commit_list *list,
 485                                   struct commit_list **tried,
 486                                   int show_all,
 487                                   int *count,
 488                                   int *skipped_first)
 489{
 490        struct commit_list *filtered = NULL, **f = &filtered;
 491
 492        *tried = NULL;
 493
 494        if (skipped_first)
 495                *skipped_first = 0;
 496        if (count)
 497                *count = 0;
 498
 499        if (!skipped_revs.nr)
 500                return list;
 501
 502        while (list) {
 503                struct commit_list *next = list->next;
 504                list->next = NULL;
 505                if (0 <= oid_array_lookup(&skipped_revs, &list->item->object.oid)) {
 506                        if (skipped_first && !*skipped_first)
 507                                *skipped_first = 1;
 508                        /* Move current to tried list */
 509                        *tried = list;
 510                        tried = &list->next;
 511                } else {
 512                        if (!show_all) {
 513                                if (!skipped_first || !*skipped_first)
 514                                        return list;
 515                        } else if (skipped_first && !*skipped_first) {
 516                                /* This means we know it's not skipped */
 517                                *skipped_first = -1;
 518                        }
 519                        /* Move current to filtered list */
 520                        *f = list;
 521                        f = &list->next;
 522                        if (count)
 523                                (*count)++;
 524                }
 525                list = next;
 526        }
 527
 528        if (skipped_first && *skipped_first == -1)
 529                *skipped_first = 0;
 530
 531        return filtered;
 532}
 533
 534#define PRN_MODULO 32768
 535
 536/*
 537 * This is a pseudo random number generator based on "man 3 rand".
 538 * It is not used properly because the seed is the argument and it
 539 * is increased by one between each call, but that should not matter
 540 * for this application.
 541 */
 542static unsigned get_prn(unsigned count) {
 543        count = count * 1103515245 + 12345;
 544        return (count/65536) % PRN_MODULO;
 545}
 546
 547/*
 548 * Custom integer square root from
 549 * https://en.wikipedia.org/wiki/Integer_square_root
 550 */
 551static int sqrti(int val)
 552{
 553        float d, x = val;
 554
 555        if (val == 0)
 556                return 0;
 557
 558        do {
 559                float y = (x + (float)val / x) / 2;
 560                d = (y > x) ? y - x : x - y;
 561                x = y;
 562        } while (d >= 0.5);
 563
 564        return (int)x;
 565}
 566
 567static struct commit_list *skip_away(struct commit_list *list, int count)
 568{
 569        struct commit_list *cur, *previous;
 570        int prn, index, i;
 571
 572        prn = get_prn(count);
 573        index = (count * prn / PRN_MODULO) * sqrti(prn) / sqrti(PRN_MODULO);
 574
 575        cur = list;
 576        previous = NULL;
 577
 578        for (i = 0; cur; cur = cur->next, i++) {
 579                if (i == index) {
 580                        if (oidcmp(&cur->item->object.oid, current_bad_oid))
 581                                return cur;
 582                        if (previous)
 583                                return previous;
 584                        return list;
 585                }
 586                previous = cur;
 587        }
 588
 589        return list;
 590}
 591
 592static struct commit_list *managed_skipped(struct commit_list *list,
 593                                           struct commit_list **tried)
 594{
 595        int count, skipped_first;
 596
 597        *tried = NULL;
 598
 599        if (!skipped_revs.nr)
 600                return list;
 601
 602        list = filter_skipped(list, tried, 0, &count, &skipped_first);
 603
 604        if (!skipped_first)
 605                return list;
 606
 607        return skip_away(list, count);
 608}
 609
 610static void bisect_rev_setup(struct rev_info *revs, const char *prefix,
 611                             const char *bad_format, const char *good_format,
 612                             int read_paths)
 613{
 614        struct argv_array rev_argv = ARGV_ARRAY_INIT;
 615        int i;
 616
 617        init_revisions(revs, prefix);
 618        revs->abbrev = 0;
 619        revs->commit_format = CMIT_FMT_UNSPECIFIED;
 620
 621        /* rev_argv.argv[0] will be ignored by setup_revisions */
 622        argv_array_push(&rev_argv, "bisect_rev_setup");
 623        argv_array_pushf(&rev_argv, bad_format, oid_to_hex(current_bad_oid));
 624        for (i = 0; i < good_revs.nr; i++)
 625                argv_array_pushf(&rev_argv, good_format,
 626                                 oid_to_hex(good_revs.oid + i));
 627        argv_array_push(&rev_argv, "--");
 628        if (read_paths)
 629                read_bisect_paths(&rev_argv);
 630
 631        setup_revisions(rev_argv.argc, rev_argv.argv, revs, NULL);
 632        /* XXX leak rev_argv, as "revs" may still be pointing to it */
 633}
 634
 635static void bisect_common(struct rev_info *revs)
 636{
 637        if (prepare_revision_walk(revs))
 638                die("revision walk setup failed");
 639        if (revs->tree_objects)
 640                mark_edges_uninteresting(revs, NULL);
 641}
 642
 643static void exit_if_skipped_commits(struct commit_list *tried,
 644                                    const struct object_id *bad)
 645{
 646        if (!tried)
 647                return;
 648
 649        printf("There are only 'skip'ped commits left to test.\n"
 650               "The first %s commit could be any of:\n", term_bad);
 651
 652        for ( ; tried; tried = tried->next)
 653                printf("%s\n", oid_to_hex(&tried->item->object.oid));
 654
 655        if (bad)
 656                printf("%s\n", oid_to_hex(bad));
 657        printf(_("We cannot bisect more!\n"));
 658        exit(2);
 659}
 660
 661static int is_expected_rev(const struct object_id *oid)
 662{
 663        const char *filename = git_path_bisect_expected_rev();
 664        struct stat st;
 665        struct strbuf str = STRBUF_INIT;
 666        FILE *fp;
 667        int res = 0;
 668
 669        if (stat(filename, &st) || !S_ISREG(st.st_mode))
 670                return 0;
 671
 672        fp = fopen_or_warn(filename, "r");
 673        if (!fp)
 674                return 0;
 675
 676        if (strbuf_getline_lf(&str, fp) != EOF)
 677                res = !strcmp(str.buf, oid_to_hex(oid));
 678
 679        strbuf_release(&str);
 680        fclose(fp);
 681
 682        return res;
 683}
 684
 685static int bisect_checkout(const struct object_id *bisect_rev, int no_checkout)
 686{
 687        char bisect_rev_hex[GIT_MAX_HEXSZ + 1];
 688
 689        memcpy(bisect_rev_hex, oid_to_hex(bisect_rev), GIT_SHA1_HEXSZ + 1);
 690        update_ref(NULL, "BISECT_EXPECTED_REV", bisect_rev->hash, NULL, 0, UPDATE_REFS_DIE_ON_ERR);
 691
 692        argv_checkout[2] = bisect_rev_hex;
 693        if (no_checkout) {
 694                update_ref(NULL, "BISECT_HEAD", bisect_rev->hash, NULL, 0, UPDATE_REFS_DIE_ON_ERR);
 695        } else {
 696                int res;
 697                res = run_command_v_opt(argv_checkout, RUN_GIT_CMD);
 698                if (res)
 699                        exit(res);
 700        }
 701
 702        argv_show_branch[1] = bisect_rev_hex;
 703        return run_command_v_opt(argv_show_branch, RUN_GIT_CMD);
 704}
 705
 706static struct commit *get_commit_reference(const struct object_id *oid)
 707{
 708        struct commit *r = lookup_commit_reference(oid);
 709        if (!r)
 710                die(_("Not a valid commit name %s"), oid_to_hex(oid));
 711        return r;
 712}
 713
 714static struct commit **get_bad_and_good_commits(int *rev_nr)
 715{
 716        struct commit **rev;
 717        int i, n = 0;
 718
 719        ALLOC_ARRAY(rev, 1 + good_revs.nr);
 720        rev[n++] = get_commit_reference(current_bad_oid);
 721        for (i = 0; i < good_revs.nr; i++)
 722                rev[n++] = get_commit_reference(good_revs.oid + i);
 723        *rev_nr = n;
 724
 725        return rev;
 726}
 727
 728static void handle_bad_merge_base(void)
 729{
 730        if (is_expected_rev(current_bad_oid)) {
 731                char *bad_hex = oid_to_hex(current_bad_oid);
 732                char *good_hex = join_sha1_array_hex(&good_revs, ' ');
 733                if (!strcmp(term_bad, "bad") && !strcmp(term_good, "good")) {
 734                        fprintf(stderr, _("The merge base %s is bad.\n"
 735                                "This means the bug has been fixed "
 736                                "between %s and [%s].\n"),
 737                                bad_hex, bad_hex, good_hex);
 738                } else if (!strcmp(term_bad, "new") && !strcmp(term_good, "old")) {
 739                        fprintf(stderr, _("The merge base %s is new.\n"
 740                                "The property has changed "
 741                                "between %s and [%s].\n"),
 742                                bad_hex, bad_hex, good_hex);
 743                } else {
 744                        fprintf(stderr, _("The merge base %s is %s.\n"
 745                                "This means the first '%s' commit is "
 746                                "between %s and [%s].\n"),
 747                                bad_hex, term_bad, term_good, bad_hex, good_hex);
 748                }
 749                exit(3);
 750        }
 751
 752        fprintf(stderr, _("Some %s revs are not ancestors of the %s rev.\n"
 753                "git bisect cannot work properly in this case.\n"
 754                "Maybe you mistook %s and %s revs?\n"),
 755                term_good, term_bad, term_good, term_bad);
 756        exit(1);
 757}
 758
 759static void handle_skipped_merge_base(const struct object_id *mb)
 760{
 761        char *mb_hex = oid_to_hex(mb);
 762        char *bad_hex = oid_to_hex(current_bad_oid);
 763        char *good_hex = join_sha1_array_hex(&good_revs, ' ');
 764
 765        warning(_("the merge base between %s and [%s] "
 766                "must be skipped.\n"
 767                "So we cannot be sure the first %s commit is "
 768                "between %s and %s.\n"
 769                "We continue anyway."),
 770                bad_hex, good_hex, term_bad, mb_hex, bad_hex);
 771        free(good_hex);
 772}
 773
 774/*
 775 * "check_merge_bases" checks that merge bases are not "bad" (or "new").
 776 *
 777 * - If one is "bad" (or "new"), it means the user assumed something wrong
 778 * and we must exit with a non 0 error code.
 779 * - If one is "good" (or "old"), that's good, we have nothing to do.
 780 * - If one is "skipped", we can't know but we should warn.
 781 * - If we don't know, we should check it out and ask the user to test.
 782 */
 783static void check_merge_bases(int no_checkout)
 784{
 785        struct commit_list *result;
 786        int rev_nr;
 787        struct commit **rev = get_bad_and_good_commits(&rev_nr);
 788
 789        result = get_merge_bases_many(rev[0], rev_nr - 1, rev + 1);
 790
 791        for (; result; result = result->next) {
 792                const struct object_id *mb = &result->item->object.oid;
 793                if (!oidcmp(mb, current_bad_oid)) {
 794                        handle_bad_merge_base();
 795                } else if (0 <= oid_array_lookup(&good_revs, mb)) {
 796                        continue;
 797                } else if (0 <= oid_array_lookup(&skipped_revs, mb)) {
 798                        handle_skipped_merge_base(mb);
 799                } else {
 800                        printf(_("Bisecting: a merge base must be tested\n"));
 801                        exit(bisect_checkout(mb, no_checkout));
 802                }
 803        }
 804
 805        free(rev);
 806        free_commit_list(result);
 807}
 808
 809static int check_ancestors(const char *prefix)
 810{
 811        struct rev_info revs;
 812        struct object_array pending_copy;
 813        int res;
 814
 815        bisect_rev_setup(&revs, prefix, "^%s", "%s", 0);
 816
 817        /* Save pending objects, so they can be cleaned up later. */
 818        pending_copy = revs.pending;
 819        revs.leak_pending = 1;
 820
 821        /*
 822         * bisect_common calls prepare_revision_walk right away, which
 823         * (together with .leak_pending = 1) makes us the sole owner of
 824         * the list of pending objects.
 825         */
 826        bisect_common(&revs);
 827        res = (revs.commits != NULL);
 828
 829        /* Clean up objects used, as they will be reused. */
 830        clear_commit_marks_for_object_array(&pending_copy, ALL_REV_FLAGS);
 831
 832        object_array_clear(&pending_copy);
 833
 834        return res;
 835}
 836
 837/*
 838 * "check_good_are_ancestors_of_bad" checks that all "good" revs are
 839 * ancestor of the "bad" rev.
 840 *
 841 * If that's not the case, we need to check the merge bases.
 842 * If a merge base must be tested by the user, its source code will be
 843 * checked out to be tested by the user and we will exit.
 844 */
 845static void check_good_are_ancestors_of_bad(const char *prefix, int no_checkout)
 846{
 847        char *filename = git_pathdup("BISECT_ANCESTORS_OK");
 848        struct stat st;
 849        int fd;
 850
 851        if (!current_bad_oid)
 852                die(_("a %s revision is needed"), term_bad);
 853
 854        /* Check if file BISECT_ANCESTORS_OK exists. */
 855        if (!stat(filename, &st) && S_ISREG(st.st_mode))
 856                goto done;
 857
 858        /* Bisecting with no good rev is ok. */
 859        if (good_revs.nr == 0)
 860                goto done;
 861
 862        /* Check if all good revs are ancestor of the bad rev. */
 863        if (check_ancestors(prefix))
 864                check_merge_bases(no_checkout);
 865
 866        /* Create file BISECT_ANCESTORS_OK. */
 867        fd = open(filename, O_CREAT | O_TRUNC | O_WRONLY, 0600);
 868        if (fd < 0)
 869                warning_errno(_("could not create file '%s'"),
 870                              filename);
 871        else
 872                close(fd);
 873 done:
 874        free(filename);
 875}
 876
 877/*
 878 * This does "git diff-tree --pretty COMMIT" without one fork+exec.
 879 */
 880static void show_diff_tree(const char *prefix, struct commit *commit)
 881{
 882        struct rev_info opt;
 883
 884        /* diff-tree init */
 885        init_revisions(&opt, prefix);
 886        git_config(git_diff_basic_config, NULL); /* no "diff" UI options */
 887        opt.abbrev = 0;
 888        opt.diff = 1;
 889
 890        /* This is what "--pretty" does */
 891        opt.verbose_header = 1;
 892        opt.use_terminator = 0;
 893        opt.commit_format = CMIT_FMT_DEFAULT;
 894
 895        /* diff-tree init */
 896        if (!opt.diffopt.output_format)
 897                opt.diffopt.output_format = DIFF_FORMAT_RAW;
 898
 899        setup_revisions(0, NULL, &opt, NULL);
 900        log_tree_commit(&opt, commit);
 901}
 902
 903/*
 904 * The terms used for this bisect session are stored in BISECT_TERMS.
 905 * We read them and store them to adapt the messages accordingly.
 906 * Default is bad/good.
 907 */
 908void read_bisect_terms(const char **read_bad, const char **read_good)
 909{
 910        struct strbuf str = STRBUF_INIT;
 911        const char *filename = git_path_bisect_terms();
 912        FILE *fp = fopen(filename, "r");
 913
 914        if (!fp) {
 915                if (errno == ENOENT) {
 916                        *read_bad = "bad";
 917                        *read_good = "good";
 918                        return;
 919                } else {
 920                        die_errno(_("could not read file '%s'"), filename);
 921                }
 922        } else {
 923                strbuf_getline_lf(&str, fp);
 924                *read_bad = strbuf_detach(&str, NULL);
 925                strbuf_getline_lf(&str, fp);
 926                *read_good = strbuf_detach(&str, NULL);
 927        }
 928        strbuf_release(&str);
 929        fclose(fp);
 930}
 931
 932/*
 933 * We use the convention that exiting with an exit code 10 means that
 934 * the bisection process finished successfully.
 935 * In this case the calling shell script should exit 0.
 936 *
 937 * If no_checkout is non-zero, the bisection process does not
 938 * checkout the trial commit but instead simply updates BISECT_HEAD.
 939 */
 940int bisect_next_all(const char *prefix, int no_checkout)
 941{
 942        struct rev_info revs;
 943        struct commit_list *tried;
 944        int reaches = 0, all = 0, nr, steps;
 945        struct object_id *bisect_rev;
 946        char *steps_msg;
 947
 948        read_bisect_terms(&term_bad, &term_good);
 949        if (read_bisect_refs())
 950                die(_("reading bisect refs failed"));
 951
 952        check_good_are_ancestors_of_bad(prefix, no_checkout);
 953
 954        bisect_rev_setup(&revs, prefix, "%s", "^%s", 1);
 955        revs.limited = 1;
 956
 957        bisect_common(&revs);
 958
 959        find_bisection(&revs.commits, &reaches, &all, !!skipped_revs.nr);
 960        revs.commits = managed_skipped(revs.commits, &tried);
 961
 962        if (!revs.commits) {
 963                /*
 964                 * We should exit here only if the "bad"
 965                 * commit is also a "skip" commit.
 966                 */
 967                exit_if_skipped_commits(tried, NULL);
 968
 969                printf(_("%s was both %s and %s\n"),
 970                       oid_to_hex(current_bad_oid),
 971                       term_good,
 972                       term_bad);
 973                exit(1);
 974        }
 975
 976        if (!all) {
 977                fprintf(stderr, _("No testable commit found.\n"
 978                        "Maybe you started with bad path parameters?\n"));
 979                exit(4);
 980        }
 981
 982        bisect_rev = &revs.commits->item->object.oid;
 983
 984        if (!oidcmp(bisect_rev, current_bad_oid)) {
 985                exit_if_skipped_commits(tried, current_bad_oid);
 986                printf("%s is the first %s commit\n", oid_to_hex(bisect_rev),
 987                        term_bad);
 988                show_diff_tree(prefix, revs.commits->item);
 989                /* This means the bisection process succeeded. */
 990                exit(10);
 991        }
 992
 993        nr = all - reaches - 1;
 994        steps = estimate_bisect_steps(all);
 995
 996        steps_msg = xstrfmt(Q_("(roughly %d step)", "(roughly %d steps)",
 997                  steps), steps);
 998        /*
 999         * TRANSLATORS: the last %s will be replaced with "(roughly %d
1000         * steps)" translation.
1001         */
1002        printf(Q_("Bisecting: %d revision left to test after this %s\n",
1003                  "Bisecting: %d revisions left to test after this %s\n",
1004                  nr), nr, steps_msg);
1005        free(steps_msg);
1006
1007        return bisect_checkout(bisect_rev, no_checkout);
1008}
1009
1010static inline int log2i(int n)
1011{
1012        int log2 = 0;
1013
1014        for (; n > 1; n >>= 1)
1015                log2++;
1016
1017        return log2;
1018}
1019
1020static inline int exp2i(int n)
1021{
1022        return 1 << n;
1023}
1024
1025/*
1026 * Estimate the number of bisect steps left (after the current step)
1027 *
1028 * For any x between 0 included and 2^n excluded, the probability for
1029 * n - 1 steps left looks like:
1030 *
1031 * P(2^n + x) == (2^n - x) / (2^n + x)
1032 *
1033 * and P(2^n + x) < 0.5 means 2^n < 3x
1034 */
1035int estimate_bisect_steps(int all)
1036{
1037        int n, x, e;
1038
1039        if (all < 3)
1040                return 0;
1041
1042        n = log2i(all);
1043        e = exp2i(n);
1044        x = all - e;
1045
1046        return (e < 3 * x) ? n : n - 1;
1047}