bisect.con commit bisect: fix memory leak in `find_bisection()` (fc5c40b)
   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                p = p->next;
 230        }
 231        if (p)
 232                p->next = NULL;
 233        strbuf_release(&buf);
 234        free(array);
 235        return list;
 236}
 237
 238/*
 239 * zero or positive weight is the number of interesting commits it can
 240 * reach, including itself.  Especially, weight = 0 means it does not
 241 * reach any tree-changing commits (e.g. just above uninteresting one
 242 * but traversal is with pathspec).
 243 *
 244 * weight = -1 means it has one parent and its distance is yet to
 245 * be computed.
 246 *
 247 * weight = -2 means it has more than one parent and its distance is
 248 * unknown.  After running count_distance() first, they will get zero
 249 * or positive distance.
 250 */
 251static struct commit_list *do_find_bisection(struct commit_list *list,
 252                                             int nr, int *weights,
 253                                             int find_all)
 254{
 255        int n, counted;
 256        struct commit_list *p;
 257
 258        counted = 0;
 259
 260        for (n = 0, p = list; p; p = p->next) {
 261                struct commit *commit = p->item;
 262                unsigned flags = commit->object.flags;
 263
 264                p->item->util = &weights[n++];
 265                switch (count_interesting_parents(commit)) {
 266                case 0:
 267                        if (!(flags & TREESAME)) {
 268                                weight_set(p, 1);
 269                                counted++;
 270                                show_list("bisection 2 count one",
 271                                          counted, nr, list);
 272                        }
 273                        /*
 274                         * otherwise, it is known not to reach any
 275                         * tree-changing commit and gets weight 0.
 276                         */
 277                        break;
 278                case 1:
 279                        weight_set(p, -1);
 280                        break;
 281                default:
 282                        weight_set(p, -2);
 283                        break;
 284                }
 285        }
 286
 287        show_list("bisection 2 initialize", counted, nr, list);
 288
 289        /*
 290         * If you have only one parent in the resulting set
 291         * then you can reach one commit more than that parent
 292         * can reach.  So we do not have to run the expensive
 293         * count_distance() for single strand of pearls.
 294         *
 295         * However, if you have more than one parents, you cannot
 296         * just add their distance and one for yourself, since
 297         * they usually reach the same ancestor and you would
 298         * end up counting them twice that way.
 299         *
 300         * So we will first count distance of merges the usual
 301         * way, and then fill the blanks using cheaper algorithm.
 302         */
 303        for (p = list; p; p = p->next) {
 304                if (p->item->object.flags & UNINTERESTING)
 305                        continue;
 306                if (weight(p) != -2)
 307                        continue;
 308                weight_set(p, count_distance(p));
 309                clear_distance(list);
 310
 311                /* Does it happen to be at exactly half-way? */
 312                if (!find_all && halfway(p, nr))
 313                        return p;
 314                counted++;
 315        }
 316
 317        show_list("bisection 2 count_distance", counted, nr, list);
 318
 319        while (counted < nr) {
 320                for (p = list; p; p = p->next) {
 321                        struct commit_list *q;
 322                        unsigned flags = p->item->object.flags;
 323
 324                        if (0 <= weight(p))
 325                                continue;
 326                        for (q = p->item->parents; q; q = q->next) {
 327                                if (q->item->object.flags & UNINTERESTING)
 328                                        continue;
 329                                if (0 <= weight(q))
 330                                        break;
 331                        }
 332                        if (!q)
 333                                continue;
 334
 335                        /*
 336                         * weight for p is unknown but q is known.
 337                         * add one for p itself if p is to be counted,
 338                         * otherwise inherit it from q directly.
 339                         */
 340                        if (!(flags & TREESAME)) {
 341                                weight_set(p, weight(q)+1);
 342                                counted++;
 343                                show_list("bisection 2 count one",
 344                                          counted, nr, list);
 345                        }
 346                        else
 347                                weight_set(p, weight(q));
 348
 349                        /* Does it happen to be at exactly half-way? */
 350                        if (!find_all && halfway(p, nr))
 351                                return p;
 352                }
 353        }
 354
 355        show_list("bisection 2 counted all", counted, nr, list);
 356
 357        if (!find_all)
 358                return best_bisection(list, nr);
 359        else
 360                return best_bisection_sorted(list, nr);
 361}
 362
 363void find_bisection(struct commit_list **commit_list, int *reaches,
 364                    int *all, int find_all)
 365{
 366        int nr, on_list;
 367        struct commit_list *list, *p, *best, *next, *last;
 368        int *weights;
 369
 370        show_list("bisection 2 entry", 0, 0, *commit_list);
 371
 372        /*
 373         * Count the number of total and tree-changing items on the
 374         * list, while reversing the list.
 375         */
 376        for (nr = on_list = 0, last = NULL, p = *commit_list;
 377             p;
 378             p = next) {
 379                unsigned flags = p->item->object.flags;
 380
 381                next = p->next;
 382                if (flags & UNINTERESTING) {
 383                        free(p);
 384                        continue;
 385                }
 386                p->next = last;
 387                last = p;
 388                if (!(flags & TREESAME))
 389                        nr++;
 390                on_list++;
 391        }
 392        list = last;
 393        show_list("bisection 2 sorted", 0, nr, list);
 394
 395        *all = nr;
 396        weights = xcalloc(on_list, sizeof(*weights));
 397
 398        /* Do the real work of finding bisection commit. */
 399        best = do_find_bisection(list, nr, weights, find_all);
 400        if (best) {
 401                if (!find_all)
 402                        best->next = NULL;
 403                *reaches = weight(best);
 404        }
 405        free(weights);
 406        *commit_list = best;
 407}
 408
 409static int register_ref(const char *refname, const struct object_id *oid,
 410                        int flags, void *cb_data)
 411{
 412        struct strbuf good_prefix = STRBUF_INIT;
 413        strbuf_addstr(&good_prefix, term_good);
 414        strbuf_addstr(&good_prefix, "-");
 415
 416        if (!strcmp(refname, term_bad)) {
 417                current_bad_oid = xmalloc(sizeof(*current_bad_oid));
 418                oidcpy(current_bad_oid, oid);
 419        } else if (starts_with(refname, good_prefix.buf)) {
 420                oid_array_append(&good_revs, oid);
 421        } else if (starts_with(refname, "skip-")) {
 422                oid_array_append(&skipped_revs, oid);
 423        }
 424
 425        strbuf_release(&good_prefix);
 426
 427        return 0;
 428}
 429
 430static int read_bisect_refs(void)
 431{
 432        return for_each_ref_in("refs/bisect/", register_ref, NULL);
 433}
 434
 435static GIT_PATH_FUNC(git_path_bisect_names, "BISECT_NAMES")
 436static GIT_PATH_FUNC(git_path_bisect_expected_rev, "BISECT_EXPECTED_REV")
 437static GIT_PATH_FUNC(git_path_bisect_terms, "BISECT_TERMS")
 438
 439static void read_bisect_paths(struct argv_array *array)
 440{
 441        struct strbuf str = STRBUF_INIT;
 442        const char *filename = git_path_bisect_names();
 443        FILE *fp = xfopen(filename, "r");
 444
 445        while (strbuf_getline_lf(&str, fp) != EOF) {
 446                strbuf_trim(&str);
 447                if (sq_dequote_to_argv_array(str.buf, array))
 448                        die(_("Badly quoted content in file '%s': %s"),
 449                            filename, str.buf);
 450        }
 451
 452        strbuf_release(&str);
 453        fclose(fp);
 454}
 455
 456static char *join_sha1_array_hex(struct oid_array *array, char delim)
 457{
 458        struct strbuf joined_hexs = STRBUF_INIT;
 459        int i;
 460
 461        for (i = 0; i < array->nr; i++) {
 462                strbuf_addstr(&joined_hexs, oid_to_hex(array->oid + i));
 463                if (i + 1 < array->nr)
 464                        strbuf_addch(&joined_hexs, delim);
 465        }
 466
 467        return strbuf_detach(&joined_hexs, NULL);
 468}
 469
 470/*
 471 * In this function, passing a not NULL skipped_first is very special.
 472 * It means that we want to know if the first commit in the list is
 473 * skipped because we will want to test a commit away from it if it is
 474 * indeed skipped.
 475 * So if the first commit is skipped, we cannot take the shortcut to
 476 * just "return list" when we find the first non skipped commit, we
 477 * have to return a fully filtered list.
 478 *
 479 * We use (*skipped_first == -1) to mean "it has been found that the
 480 * first commit is not skipped". In this case *skipped_first is set back
 481 * to 0 just before the function returns.
 482 */
 483struct commit_list *filter_skipped(struct commit_list *list,
 484                                   struct commit_list **tried,
 485                                   int show_all,
 486                                   int *count,
 487                                   int *skipped_first)
 488{
 489        struct commit_list *filtered = NULL, **f = &filtered;
 490
 491        *tried = NULL;
 492
 493        if (skipped_first)
 494                *skipped_first = 0;
 495        if (count)
 496                *count = 0;
 497
 498        if (!skipped_revs.nr)
 499                return list;
 500
 501        while (list) {
 502                struct commit_list *next = list->next;
 503                list->next = NULL;
 504                if (0 <= oid_array_lookup(&skipped_revs, &list->item->object.oid)) {
 505                        if (skipped_first && !*skipped_first)
 506                                *skipped_first = 1;
 507                        /* Move current to tried list */
 508                        *tried = list;
 509                        tried = &list->next;
 510                } else {
 511                        if (!show_all) {
 512                                if (!skipped_first || !*skipped_first)
 513                                        return list;
 514                        } else if (skipped_first && !*skipped_first) {
 515                                /* This means we know it's not skipped */
 516                                *skipped_first = -1;
 517                        }
 518                        /* Move current to filtered list */
 519                        *f = list;
 520                        f = &list->next;
 521                        if (count)
 522                                (*count)++;
 523                }
 524                list = next;
 525        }
 526
 527        if (skipped_first && *skipped_first == -1)
 528                *skipped_first = 0;
 529
 530        return filtered;
 531}
 532
 533#define PRN_MODULO 32768
 534
 535/*
 536 * This is a pseudo random number generator based on "man 3 rand".
 537 * It is not used properly because the seed is the argument and it
 538 * is increased by one between each call, but that should not matter
 539 * for this application.
 540 */
 541static unsigned get_prn(unsigned count) {
 542        count = count * 1103515245 + 12345;
 543        return (count/65536) % PRN_MODULO;
 544}
 545
 546/*
 547 * Custom integer square root from
 548 * https://en.wikipedia.org/wiki/Integer_square_root
 549 */
 550static int sqrti(int val)
 551{
 552        float d, x = val;
 553
 554        if (val == 0)
 555                return 0;
 556
 557        do {
 558                float y = (x + (float)val / x) / 2;
 559                d = (y > x) ? y - x : x - y;
 560                x = y;
 561        } while (d >= 0.5);
 562
 563        return (int)x;
 564}
 565
 566static struct commit_list *skip_away(struct commit_list *list, int count)
 567{
 568        struct commit_list *cur, *previous;
 569        int prn, index, i;
 570
 571        prn = get_prn(count);
 572        index = (count * prn / PRN_MODULO) * sqrti(prn) / sqrti(PRN_MODULO);
 573
 574        cur = list;
 575        previous = NULL;
 576
 577        for (i = 0; cur; cur = cur->next, i++) {
 578                if (i == index) {
 579                        if (oidcmp(&cur->item->object.oid, current_bad_oid))
 580                                return cur;
 581                        if (previous)
 582                                return previous;
 583                        return list;
 584                }
 585                previous = cur;
 586        }
 587
 588        return list;
 589}
 590
 591static struct commit_list *managed_skipped(struct commit_list *list,
 592                                           struct commit_list **tried)
 593{
 594        int count, skipped_first;
 595
 596        *tried = NULL;
 597
 598        if (!skipped_revs.nr)
 599                return list;
 600
 601        list = filter_skipped(list, tried, 0, &count, &skipped_first);
 602
 603        if (!skipped_first)
 604                return list;
 605
 606        return skip_away(list, count);
 607}
 608
 609static void bisect_rev_setup(struct rev_info *revs, const char *prefix,
 610                             const char *bad_format, const char *good_format,
 611                             int read_paths)
 612{
 613        struct argv_array rev_argv = ARGV_ARRAY_INIT;
 614        int i;
 615
 616        init_revisions(revs, prefix);
 617        revs->abbrev = 0;
 618        revs->commit_format = CMIT_FMT_UNSPECIFIED;
 619
 620        /* rev_argv.argv[0] will be ignored by setup_revisions */
 621        argv_array_push(&rev_argv, "bisect_rev_setup");
 622        argv_array_pushf(&rev_argv, bad_format, oid_to_hex(current_bad_oid));
 623        for (i = 0; i < good_revs.nr; i++)
 624                argv_array_pushf(&rev_argv, good_format,
 625                                 oid_to_hex(good_revs.oid + i));
 626        argv_array_push(&rev_argv, "--");
 627        if (read_paths)
 628                read_bisect_paths(&rev_argv);
 629
 630        setup_revisions(rev_argv.argc, rev_argv.argv, revs, NULL);
 631        /* XXX leak rev_argv, as "revs" may still be pointing to it */
 632}
 633
 634static void bisect_common(struct rev_info *revs)
 635{
 636        if (prepare_revision_walk(revs))
 637                die("revision walk setup failed");
 638        if (revs->tree_objects)
 639                mark_edges_uninteresting(revs, NULL);
 640}
 641
 642static void exit_if_skipped_commits(struct commit_list *tried,
 643                                    const struct object_id *bad)
 644{
 645        if (!tried)
 646                return;
 647
 648        printf("There are only 'skip'ped commits left to test.\n"
 649               "The first %s commit could be any of:\n", term_bad);
 650
 651        for ( ; tried; tried = tried->next)
 652                printf("%s\n", oid_to_hex(&tried->item->object.oid));
 653
 654        if (bad)
 655                printf("%s\n", oid_to_hex(bad));
 656        printf(_("We cannot bisect more!\n"));
 657        exit(2);
 658}
 659
 660static int is_expected_rev(const struct object_id *oid)
 661{
 662        const char *filename = git_path_bisect_expected_rev();
 663        struct stat st;
 664        struct strbuf str = STRBUF_INIT;
 665        FILE *fp;
 666        int res = 0;
 667
 668        if (stat(filename, &st) || !S_ISREG(st.st_mode))
 669                return 0;
 670
 671        fp = fopen_or_warn(filename, "r");
 672        if (!fp)
 673                return 0;
 674
 675        if (strbuf_getline_lf(&str, fp) != EOF)
 676                res = !strcmp(str.buf, oid_to_hex(oid));
 677
 678        strbuf_release(&str);
 679        fclose(fp);
 680
 681        return res;
 682}
 683
 684static int bisect_checkout(const struct object_id *bisect_rev, int no_checkout)
 685{
 686        char bisect_rev_hex[GIT_MAX_HEXSZ + 1];
 687
 688        memcpy(bisect_rev_hex, oid_to_hex(bisect_rev), GIT_SHA1_HEXSZ + 1);
 689        update_ref(NULL, "BISECT_EXPECTED_REV", bisect_rev->hash, NULL, 0, UPDATE_REFS_DIE_ON_ERR);
 690
 691        argv_checkout[2] = bisect_rev_hex;
 692        if (no_checkout) {
 693                update_ref(NULL, "BISECT_HEAD", bisect_rev->hash, NULL, 0, UPDATE_REFS_DIE_ON_ERR);
 694        } else {
 695                int res;
 696                res = run_command_v_opt(argv_checkout, RUN_GIT_CMD);
 697                if (res)
 698                        exit(res);
 699        }
 700
 701        argv_show_branch[1] = bisect_rev_hex;
 702        return run_command_v_opt(argv_show_branch, RUN_GIT_CMD);
 703}
 704
 705static struct commit *get_commit_reference(const struct object_id *oid)
 706{
 707        struct commit *r = lookup_commit_reference(oid);
 708        if (!r)
 709                die(_("Not a valid commit name %s"), oid_to_hex(oid));
 710        return r;
 711}
 712
 713static struct commit **get_bad_and_good_commits(int *rev_nr)
 714{
 715        struct commit **rev;
 716        int i, n = 0;
 717
 718        ALLOC_ARRAY(rev, 1 + good_revs.nr);
 719        rev[n++] = get_commit_reference(current_bad_oid);
 720        for (i = 0; i < good_revs.nr; i++)
 721                rev[n++] = get_commit_reference(good_revs.oid + i);
 722        *rev_nr = n;
 723
 724        return rev;
 725}
 726
 727static void handle_bad_merge_base(void)
 728{
 729        if (is_expected_rev(current_bad_oid)) {
 730                char *bad_hex = oid_to_hex(current_bad_oid);
 731                char *good_hex = join_sha1_array_hex(&good_revs, ' ');
 732                if (!strcmp(term_bad, "bad") && !strcmp(term_good, "good")) {
 733                        fprintf(stderr, _("The merge base %s is bad.\n"
 734                                "This means the bug has been fixed "
 735                                "between %s and [%s].\n"),
 736                                bad_hex, bad_hex, good_hex);
 737                } else if (!strcmp(term_bad, "new") && !strcmp(term_good, "old")) {
 738                        fprintf(stderr, _("The merge base %s is new.\n"
 739                                "The property has changed "
 740                                "between %s and [%s].\n"),
 741                                bad_hex, bad_hex, good_hex);
 742                } else {
 743                        fprintf(stderr, _("The merge base %s is %s.\n"
 744                                "This means the first '%s' commit is "
 745                                "between %s and [%s].\n"),
 746                                bad_hex, term_bad, term_good, bad_hex, good_hex);
 747                }
 748                exit(3);
 749        }
 750
 751        fprintf(stderr, _("Some %s revs are not ancestors of the %s rev.\n"
 752                "git bisect cannot work properly in this case.\n"
 753                "Maybe you mistook %s and %s revs?\n"),
 754                term_good, term_bad, term_good, term_bad);
 755        exit(1);
 756}
 757
 758static void handle_skipped_merge_base(const struct object_id *mb)
 759{
 760        char *mb_hex = oid_to_hex(mb);
 761        char *bad_hex = oid_to_hex(current_bad_oid);
 762        char *good_hex = join_sha1_array_hex(&good_revs, ' ');
 763
 764        warning(_("the merge base between %s and [%s] "
 765                "must be skipped.\n"
 766                "So we cannot be sure the first %s commit is "
 767                "between %s and %s.\n"
 768                "We continue anyway."),
 769                bad_hex, good_hex, term_bad, mb_hex, bad_hex);
 770        free(good_hex);
 771}
 772
 773/*
 774 * "check_merge_bases" checks that merge bases are not "bad" (or "new").
 775 *
 776 * - If one is "bad" (or "new"), it means the user assumed something wrong
 777 * and we must exit with a non 0 error code.
 778 * - If one is "good" (or "old"), that's good, we have nothing to do.
 779 * - If one is "skipped", we can't know but we should warn.
 780 * - If we don't know, we should check it out and ask the user to test.
 781 */
 782static void check_merge_bases(int no_checkout)
 783{
 784        struct commit_list *result;
 785        int rev_nr;
 786        struct commit **rev = get_bad_and_good_commits(&rev_nr);
 787
 788        result = get_merge_bases_many(rev[0], rev_nr - 1, rev + 1);
 789
 790        for (; result; result = result->next) {
 791                const struct object_id *mb = &result->item->object.oid;
 792                if (!oidcmp(mb, current_bad_oid)) {
 793                        handle_bad_merge_base();
 794                } else if (0 <= oid_array_lookup(&good_revs, mb)) {
 795                        continue;
 796                } else if (0 <= oid_array_lookup(&skipped_revs, mb)) {
 797                        handle_skipped_merge_base(mb);
 798                } else {
 799                        printf(_("Bisecting: a merge base must be tested\n"));
 800                        exit(bisect_checkout(mb, no_checkout));
 801                }
 802        }
 803
 804        free(rev);
 805        free_commit_list(result);
 806}
 807
 808static int check_ancestors(const char *prefix)
 809{
 810        struct rev_info revs;
 811        struct object_array pending_copy;
 812        int res;
 813
 814        bisect_rev_setup(&revs, prefix, "^%s", "%s", 0);
 815
 816        /* Save pending objects, so they can be cleaned up later. */
 817        pending_copy = revs.pending;
 818        revs.leak_pending = 1;
 819
 820        /*
 821         * bisect_common calls prepare_revision_walk right away, which
 822         * (together with .leak_pending = 1) makes us the sole owner of
 823         * the list of pending objects.
 824         */
 825        bisect_common(&revs);
 826        res = (revs.commits != NULL);
 827
 828        /* Clean up objects used, as they will be reused. */
 829        clear_commit_marks_for_object_array(&pending_copy, ALL_REV_FLAGS);
 830
 831        object_array_clear(&pending_copy);
 832
 833        return res;
 834}
 835
 836/*
 837 * "check_good_are_ancestors_of_bad" checks that all "good" revs are
 838 * ancestor of the "bad" rev.
 839 *
 840 * If that's not the case, we need to check the merge bases.
 841 * If a merge base must be tested by the user, its source code will be
 842 * checked out to be tested by the user and we will exit.
 843 */
 844static void check_good_are_ancestors_of_bad(const char *prefix, int no_checkout)
 845{
 846        char *filename = git_pathdup("BISECT_ANCESTORS_OK");
 847        struct stat st;
 848        int fd;
 849
 850        if (!current_bad_oid)
 851                die(_("a %s revision is needed"), term_bad);
 852
 853        /* Check if file BISECT_ANCESTORS_OK exists. */
 854        if (!stat(filename, &st) && S_ISREG(st.st_mode))
 855                goto done;
 856
 857        /* Bisecting with no good rev is ok. */
 858        if (good_revs.nr == 0)
 859                goto done;
 860
 861        /* Check if all good revs are ancestor of the bad rev. */
 862        if (check_ancestors(prefix))
 863                check_merge_bases(no_checkout);
 864
 865        /* Create file BISECT_ANCESTORS_OK. */
 866        fd = open(filename, O_CREAT | O_TRUNC | O_WRONLY, 0600);
 867        if (fd < 0)
 868                warning_errno(_("could not create file '%s'"),
 869                              filename);
 870        else
 871                close(fd);
 872 done:
 873        free(filename);
 874}
 875
 876/*
 877 * This does "git diff-tree --pretty COMMIT" without one fork+exec.
 878 */
 879static void show_diff_tree(const char *prefix, struct commit *commit)
 880{
 881        struct rev_info opt;
 882
 883        /* diff-tree init */
 884        init_revisions(&opt, prefix);
 885        git_config(git_diff_basic_config, NULL); /* no "diff" UI options */
 886        opt.abbrev = 0;
 887        opt.diff = 1;
 888
 889        /* This is what "--pretty" does */
 890        opt.verbose_header = 1;
 891        opt.use_terminator = 0;
 892        opt.commit_format = CMIT_FMT_DEFAULT;
 893
 894        /* diff-tree init */
 895        if (!opt.diffopt.output_format)
 896                opt.diffopt.output_format = DIFF_FORMAT_RAW;
 897
 898        setup_revisions(0, NULL, &opt, NULL);
 899        log_tree_commit(&opt, commit);
 900}
 901
 902/*
 903 * The terms used for this bisect session are stored in BISECT_TERMS.
 904 * We read them and store them to adapt the messages accordingly.
 905 * Default is bad/good.
 906 */
 907void read_bisect_terms(const char **read_bad, const char **read_good)
 908{
 909        struct strbuf str = STRBUF_INIT;
 910        const char *filename = git_path_bisect_terms();
 911        FILE *fp = fopen(filename, "r");
 912
 913        if (!fp) {
 914                if (errno == ENOENT) {
 915                        *read_bad = "bad";
 916                        *read_good = "good";
 917                        return;
 918                } else {
 919                        die_errno(_("could not read file '%s'"), filename);
 920                }
 921        } else {
 922                strbuf_getline_lf(&str, fp);
 923                *read_bad = strbuf_detach(&str, NULL);
 924                strbuf_getline_lf(&str, fp);
 925                *read_good = strbuf_detach(&str, NULL);
 926        }
 927        strbuf_release(&str);
 928        fclose(fp);
 929}
 930
 931/*
 932 * We use the convention that exiting with an exit code 10 means that
 933 * the bisection process finished successfully.
 934 * In this case the calling shell script should exit 0.
 935 *
 936 * If no_checkout is non-zero, the bisection process does not
 937 * checkout the trial commit but instead simply updates BISECT_HEAD.
 938 */
 939int bisect_next_all(const char *prefix, int no_checkout)
 940{
 941        struct rev_info revs;
 942        struct commit_list *tried;
 943        int reaches = 0, all = 0, nr, steps;
 944        struct object_id *bisect_rev;
 945        char *steps_msg;
 946
 947        read_bisect_terms(&term_bad, &term_good);
 948        if (read_bisect_refs())
 949                die(_("reading bisect refs failed"));
 950
 951        check_good_are_ancestors_of_bad(prefix, no_checkout);
 952
 953        bisect_rev_setup(&revs, prefix, "%s", "^%s", 1);
 954        revs.limited = 1;
 955
 956        bisect_common(&revs);
 957
 958        find_bisection(&revs.commits, &reaches, &all, !!skipped_revs.nr);
 959        revs.commits = managed_skipped(revs.commits, &tried);
 960
 961        if (!revs.commits) {
 962                /*
 963                 * We should exit here only if the "bad"
 964                 * commit is also a "skip" commit.
 965                 */
 966                exit_if_skipped_commits(tried, NULL);
 967
 968                printf(_("%s was both %s and %s\n"),
 969                       oid_to_hex(current_bad_oid),
 970                       term_good,
 971                       term_bad);
 972                exit(1);
 973        }
 974
 975        if (!all) {
 976                fprintf(stderr, _("No testable commit found.\n"
 977                        "Maybe you started with bad path parameters?\n"));
 978                exit(4);
 979        }
 980
 981        bisect_rev = &revs.commits->item->object.oid;
 982
 983        if (!oidcmp(bisect_rev, current_bad_oid)) {
 984                exit_if_skipped_commits(tried, current_bad_oid);
 985                printf("%s is the first %s commit\n", oid_to_hex(bisect_rev),
 986                        term_bad);
 987                show_diff_tree(prefix, revs.commits->item);
 988                /* This means the bisection process succeeded. */
 989                exit(10);
 990        }
 991
 992        nr = all - reaches - 1;
 993        steps = estimate_bisect_steps(all);
 994
 995        steps_msg = xstrfmt(Q_("(roughly %d step)", "(roughly %d steps)",
 996                  steps), steps);
 997        /*
 998         * TRANSLATORS: the last %s will be replaced with "(roughly %d
 999         * steps)" translation.
1000         */
1001        printf(Q_("Bisecting: %d revision left to test after this %s\n",
1002                  "Bisecting: %d revisions left to test after this %s\n",
1003                  nr), nr, steps_msg);
1004        free(steps_msg);
1005
1006        return bisect_checkout(bisect_rev, no_checkout);
1007}
1008
1009static inline int log2i(int n)
1010{
1011        int log2 = 0;
1012
1013        for (; n > 1; n >>= 1)
1014                log2++;
1015
1016        return log2;
1017}
1018
1019static inline int exp2i(int n)
1020{
1021        return 1 << n;
1022}
1023
1024/*
1025 * Estimate the number of bisect steps left (after the current step)
1026 *
1027 * For any x between 0 included and 2^n excluded, the probability for
1028 * n - 1 steps left looks like:
1029 *
1030 * P(2^n + x) == (2^n - x) / (2^n + x)
1031 *
1032 * and P(2^n + x) < 0.5 means 2^n < 3x
1033 */
1034int estimate_bisect_steps(int all)
1035{
1036        int n, x, e;
1037
1038        if (all < 3)
1039                return 0;
1040
1041        n = log2i(all);
1042        e = exp2i(n);
1043        x = all - e;
1044
1045        return (e < 3 * x) ? n : n - 1;
1046}