bisect.c

   1 #include "cache.h"
   2 #include "commit.h"
   3 #include "diff.h"
   4 #include "revision.h"
   5 #include "refs.h"
   6 #include "list-objects.h"
   7 #include "quote.h"
   8 #include "sha1-lookup.h"
   9 #include "run-command.h"
  10 #include "bisect.h"
  11
  12 struct sha1_array {
  13         unsigned char (*sha1)[20];
  14         int sha1_nr;
  15         int sha1_alloc;
  16         int sorted;
  17 };
  18
  19 static struct sha1_array good_revs;
  20 static struct sha1_array skipped_revs;
  21
  22 static const unsigned char *current_bad_sha1;
  23
  24 struct argv_array {
  25         const char **argv;
  26         int argv_nr;
  27         int argv_alloc;
  28 };
  29
  30 static const char *argv_diff_tree[] = {"diff-tree", "--pretty", NULL, NULL};
  31 static const char *argv_checkout[] = {"checkout", "-q", NULL, "--", NULL};
  32 static const char *argv_show_branch[] = {"show-branch", NULL, NULL};
  33
  34 /* bits #0-15 in revision.h */
  35
  36 #define COUNTED         (1u<<16)
  37
  38 /*
  39  * This is a truly stupid algorithm, but it's only
  40  * used for bisection, and we just don't care enough.
  41  *
  42  * We care just barely enough to avoid recursing for
  43  * non-merge entries.
  44  */
  45 static int count_distance(struct commit_list *entry)
  46 {
  47         int nr = 0;
  48
  49         while (entry) {
  50                 struct commit *commit = entry->item;
  51                 struct commit_list *p;
  52
  53                 if (commit->object.flags & (UNINTERESTING | COUNTED))
  54                         break;
  55                 if (!(commit->object.flags & TREESAME))
  56                         nr++;
  57                 commit->object.flags |= COUNTED;
  58                 p = commit->parents;
  59                 entry = p;
  60                 if (p) {
  61                         p = p->next;
  62                         while (p) {
  63                                 nr += count_distance(p);
  64                                 p = p->next;
  65                         }
  66                 }
  67         }
  68
  69         return nr;
  70 }
  71
  72 static void clear_distance(struct commit_list *list)
  73 {
  74         while (list) {
  75                 struct commit *commit = list->item;
  76                 commit->object.flags &= ~COUNTED;
  77                 list = list->next;
  78         }
  79 }
  80
  81 #define DEBUG_BISECT 0
  82
  83 static inline int weight(struct commit_list *elem)
  84 {
  85         return *((int*)(elem->item->util));
  86 }
  87
  88 static inline void weight_set(struct commit_list *elem, int weight)
  89 {
  90         *((int*)(elem->item->util)) = weight;
  91 }
  92
  93 static int count_interesting_parents(struct commit *commit)
  94 {
  95         struct commit_list *p;
  96         int count;
  97
  98         for (count = 0, p = commit->parents; p; p = p->next) {
  99                 if (p->item->object.flags & UNINTERESTING)
 100                         continue;
 101                 count++;
 102         }
 103         return count;
 104 }
 105
 106 static inline int halfway(struct commit_list *p, int nr)
 107 {
 108         /*
 109          * Don't short-cut something we are not going to return!
 110          */
 111         if (p->item->object.flags & TREESAME)
 112                 return 0;
 113         if (DEBUG_BISECT)
 114                 return 0;
 115         /*
 116          * 2 and 3 are halfway of 5.
 117          * 3 is halfway of 6 but 2 and 4 are not.
 118          */
 119         switch (2 * weight(p) - nr) {
 120         case -1: case 0: case 1:
 121                 return 1;
 122         default:
 123                 return 0;
 124         }
 125 }
 126
 127 #if !DEBUG_BISECT
 128 #define show_list(a,b,c,d) do { ; } while (0)
 129 #else
 130 static void show_list(const char *debug, int counted, int nr,
 131                       struct commit_list *list)
 132 {
 133         struct commit_list *p;
 134
 135         fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
 136
 137         for (p = list; p; p = p->next) {
 138                 struct commit_list *pp;
 139                 struct commit *commit = p->item;
 140                 unsigned flags = commit->object.flags;
 141                 enum object_type type;
 142                 unsigned long size;
 143                 char *buf = read_sha1_file(commit->object.sha1, &type, &size);
 144                 char *ep, *sp;
 145
 146                 fprintf(stderr, "%c%c%c ",
 147                         (flags & TREESAME) ? ' ' : 'T',
 148                         (flags & UNINTERESTING) ? 'U' : ' ',
 149                         (flags & COUNTED) ? 'C' : ' ');
 150                 if (commit->util)
 151                         fprintf(stderr, "%3d", weight(p));
 152                 else
 153                         fprintf(stderr, "---");
 154                 fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 155                 for (pp = commit->parents; pp; pp = pp->next)
 156                         fprintf(stderr, " %.*s", 8,
 157                                 sha1_to_hex(pp->item->object.sha1));
 158
 159                 sp = strstr(buf, "\n\n");
 160                 if (sp) {
 161                         sp += 2;
 162                         for (ep = sp; *ep && *ep != '\n'; ep++)
 163                                 ;
 164                         fprintf(stderr, " %.*s", (int)(ep - sp), sp);
 165                 }
 166                 fprintf(stderr, "\n");
 167         }
 168 }
 169 #endif /* DEBUG_BISECT */
 170
 171 static struct commit_list *best_bisection(struct commit_list *list, int nr)
 172 {
 173         struct commit_list *p, *best;
 174         int best_distance = -1;
 175
 176         best = list;
 177         for (p = list; p; p = p->next) {
 178                 int distance;
 179                 unsigned flags = p->item->object.flags;
 180
 181                 if (flags & TREESAME)
 182                         continue;
 183                 distance = weight(p);
 184                 if (nr - distance < distance)
 185                         distance = nr - distance;
 186                 if (distance > best_distance) {
 187                         best = p;
 188                         best_distance = distance;
 189                 }
 190         }
 191
 192         return best;
 193 }
 194
 195 struct commit_dist {
 196         struct commit *commit;
 197         int distance;
 198 };
 199
 200 static int compare_commit_dist(const void *a_, const void *b_)
 201 {
 202         struct commit_dist *a, *b;
 203
 204         a = (struct commit_dist *)a_;
 205         b = (struct commit_dist *)b_;
 206         if (a->distance != b->distance)
 207                 return b->distance - a->distance; /* desc sort */
 208         return hashcmp(a->commit->object.sha1, b->commit->object.sha1);
 209 }
 210
 211 static struct commit_list *best_bisection_sorted(struct commit_list *list, int nr)
 212 {
 213         struct commit_list *p;
 214         struct commit_dist *array = xcalloc(nr, sizeof(*array));
 215         int cnt, i;
 216
 217         for (p = list, cnt = 0; p; p = p->next) {
 218                 int distance;
 219                 unsigned flags = p->item->object.flags;
 220
 221                 if (flags & TREESAME)
 222                         continue;
 223                 distance = weight(p);
 224                 if (nr - distance < distance)
 225                         distance = nr - distance;
 226                 array[cnt].commit = p->item;
 227                 array[cnt].distance = distance;
 228                 cnt++;
 229         }
 230         qsort(array, cnt, sizeof(*array), compare_commit_dist);
 231         for (p = list, i = 0; i < cnt; i++) {
 232                 struct name_decoration *r = xmalloc(sizeof(*r) + 100);
 233                 struct object *obj = &(array[i].commit->object);
 234
 235                 sprintf(r->name, "dist=%d", array[i].distance);
 236                 r->next = add_decoration(&name_decoration, obj, r);
 237                 p->item = array[i].commit;
 238                 p = p->next;
 239         }
 240         if (p)
 241                 p->next = NULL;
 242         free(array);
 243         return list;
 244 }
 245
 246 /*
 247  * zero or positive weight is the number of interesting commits it can
 248  * reach, including itself.  Especially, weight = 0 means it does not
 249  * reach any tree-changing commits (e.g. just above uninteresting one
 250  * but traversal is with pathspec).
 251  *
 252  * weight = -1 means it has one parent and its distance is yet to
 253  * be computed.
 254  *
 255  * weight = -2 means it has more than one parent and its distance is
 256  * unknown.  After running count_distance() first, they will get zero
 257  * or positive distance.
 258  */
 259 static struct commit_list *do_find_bisection(struct commit_list *list,
 260                                              int nr, int *weights,
 261                                              int find_all)
 262 {
 263         int n, counted;
 264         struct commit_list *p;
 265
 266         counted = 0;
 267
 268         for (n = 0, p = list; p; p = p->next) {
 269                 struct commit *commit = p->item;
 270                 unsigned flags = commit->object.flags;
 271
 272                 p->item->util = &weights[n++];
 273                 switch (count_interesting_parents(commit)) {
 274                 case 0:
 275                         if (!(flags & TREESAME)) {
 276                                 weight_set(p, 1);
 277                                 counted++;
 278                                 show_list("bisection 2 count one",
 279                                           counted, nr, list);
 280                         }
 281                         /*
 282                          * otherwise, it is known not to reach any
 283                          * tree-changing commit and gets weight 0.
 284                          */
 285                         break;
 286                 case 1:
 287                         weight_set(p, -1);
 288                         break;
 289                 default:
 290                         weight_set(p, -2);
 291                         break;
 292                 }
 293         }
 294
 295         show_list("bisection 2 initialize", counted, nr, list);
 296
 297         /*
 298          * If you have only one parent in the resulting set
 299          * then you can reach one commit more than that parent
 300          * can reach.  So we do not have to run the expensive
 301          * count_distance() for single strand of pearls.
 302          *
 303          * However, if you have more than one parents, you cannot
 304          * just add their distance and one for yourself, since
 305          * they usually reach the same ancestor and you would
 306          * end up counting them twice that way.
 307          *
 308          * So we will first count distance of merges the usual
 309          * way, and then fill the blanks using cheaper algorithm.
 310          */
 311         for (p = list; p; p = p->next) {
 312                 if (p->item->object.flags & UNINTERESTING)
 313                         continue;
 314                 if (weight(p) != -2)
 315                         continue;
 316                 weight_set(p, count_distance(p));
 317                 clear_distance(list);
 318
 319                 /* Does it happen to be at exactly half-way? */
 320                 if (!find_all && halfway(p, nr))
 321                         return p;
 322                 counted++;
 323         }
 324
 325         show_list("bisection 2 count_distance", counted, nr, list);
 326
 327         while (counted < nr) {
 328                 for (p = list; p; p = p->next) {
 329                         struct commit_list *q;
 330                         unsigned flags = p->item->object.flags;
 331
 332                         if (0 <= weight(p))
 333                                 continue;
 334                         for (q = p->item->parents; q; q = q->next) {
 335                                 if (q->item->object.flags & UNINTERESTING)
 336                                         continue;
 337                                 if (0 <= weight(q))
 338                                         break;
 339                         }
 340                         if (!q)
 341                                 continue;
 342
 343                         /*
 344                          * weight for p is unknown but q is known.
 345                          * add one for p itself if p is to be counted,
 346                          * otherwise inherit it from q directly.
 347                          */
 348                         if (!(flags & TREESAME)) {
 349                                 weight_set(p, weight(q)+1);
 350                                 counted++;
 351                                 show_list("bisection 2 count one",
 352                                           counted, nr, list);
 353                         }
 354                         else
 355                                 weight_set(p, weight(q));
 356
 357                         /* Does it happen to be at exactly half-way? */
 358                         if (!find_all && halfway(p, nr))
 359                                 return p;
 360                 }
 361         }
 362
 363         show_list("bisection 2 counted all", counted, nr, list);
 364
 365         if (!find_all)
 366                 return best_bisection(list, nr);
 367         else
 368                 return best_bisection_sorted(list, nr);
 369 }
 370
 371 struct commit_list *find_bisection(struct commit_list *list,
 372                                           int *reaches, int *all,
 373                                           int find_all)
 374 {
 375         int nr, on_list;
 376         struct commit_list *p, *best, *next, *last;
 377         int *weights;
 378
 379         show_list("bisection 2 entry", 0, 0, list);
 380
 381         /*
 382          * Count the number of total and tree-changing items on the
 383          * list, while reversing the list.
 384          */
 385         for (nr = on_list = 0, last = NULL, p = list;
 386              p;
 387              p = next) {
 388                 unsigned flags = p->item->object.flags;
 389
 390                 next = p->next;
 391                 if (flags & UNINTERESTING)
 392                         continue;
 393                 p->next = last;
 394                 last = p;
 395                 if (!(flags & TREESAME))
 396                         nr++;
 397                 on_list++;
 398         }
 399         list = last;
 400         show_list("bisection 2 sorted", 0, nr, list);
 401
 402         *all = nr;
 403         weights = xcalloc(on_list, sizeof(*weights));
 404
 405         /* Do the real work of finding bisection commit. */
 406         best = do_find_bisection(list, nr, weights, find_all);
 407         if (best) {
 408                 if (!find_all)
 409                         best->next = NULL;
 410                 *reaches = weight(best);
 411         }
 412         free(weights);
 413         return best;
 414 }
 415
 416 static void argv_array_push(struct argv_array *array, const char *string)
 417 {
 418         ALLOC_GROW(array->argv, array->argv_nr + 1, array->argv_alloc);
 419         array->argv[array->argv_nr++] = string;
 420 }
 421
 422 static void argv_array_push_sha1(struct argv_array *array,
 423                                  const unsigned char *sha1,
 424                                  const char *format)
 425 {
 426         struct strbuf buf = STRBUF_INIT;
 427         strbuf_addf(&buf, format, sha1_to_hex(sha1));
 428         argv_array_push(array, strbuf_detach(&buf, NULL));
 429 }
 430
 431 static void sha1_array_push(struct sha1_array *array,
 432                             const unsigned char *sha1)
 433 {
 434         ALLOC_GROW(array->sha1, array->sha1_nr + 1, array->sha1_alloc);
 435         hashcpy(array->sha1[array->sha1_nr++], sha1);
 436 }
 437
 438 static int register_ref(const char *refname, const unsigned char *sha1,
 439                         int flags, void *cb_data)
 440 {
 441         if (!strcmp(refname, "bad")) {
 442                 current_bad_sha1 = sha1;
 443         } else if (!prefixcmp(refname, "good-")) {
 444                 sha1_array_push(&good_revs, sha1);
 445         } else if (!prefixcmp(refname, "skip-")) {
 446                 sha1_array_push(&skipped_revs, sha1);
 447         }
 448
 449         return 0;
 450 }
 451
 452 static int read_bisect_refs(void)
 453 {
 454         return for_each_ref_in("refs/bisect/", register_ref, NULL);
 455 }
 456
 457 void read_bisect_paths(struct argv_array *array)
 458 {
 459         struct strbuf str = STRBUF_INIT;
 460         const char *filename = git_path("BISECT_NAMES");
 461         FILE *fp = fopen(filename, "r");
 462
 463         if (!fp)
 464                 die("Could not open file '%s': %s", filename, strerror(errno));
 465
 466         while (strbuf_getline(&str, fp, '\n') != EOF) {
 467                 char *quoted;
 468                 int res;
 469
 470                 strbuf_trim(&str);
 471                 quoted = strbuf_detach(&str, NULL);
 472                 res = sq_dequote_to_argv(quoted, &array->argv,
 473                                          &array->argv_nr, &array->argv_alloc);
 474                 if (res)
 475                         die("Badly quoted content in file '%s': %s",
 476                             filename, quoted);
 477         }
 478
 479         strbuf_release(&str);
 480         fclose(fp);
 481 }
 482
 483 static int array_cmp(const void *a, const void *b)
 484 {
 485         return hashcmp(a, b);
 486 }
 487
 488 static void sort_sha1_array(struct sha1_array *array)
 489 {
 490         qsort(array->sha1, array->sha1_nr, sizeof(*array->sha1), array_cmp);
 491
 492         array->sorted = 1;
 493 }
 494
 495 static const unsigned char *sha1_access(size_t index, void *table)
 496 {
 497         unsigned char (*array)[20] = table;
 498         return array[index];
 499 }
 500
 501 static int lookup_sha1_array(struct sha1_array *array,
 502                              const unsigned char *sha1)
 503 {
 504         if (!array->sorted)
 505                 sort_sha1_array(array);
 506
 507         return sha1_pos(sha1, array->sha1, array->sha1_nr, sha1_access);
 508 }
 509
 510 struct commit_list *filter_skipped(struct commit_list *list,
 511                                    struct commit_list **tried,
 512                                    int show_all)
 513 {
 514         struct commit_list *filtered = NULL, **f = &filtered;
 515
 516         *tried = NULL;
 517
 518         if (!skipped_revs.sha1_nr)
 519                 return list;
 520
 521         while (list) {
 522                 struct commit_list *next = list->next;
 523                 list->next = NULL;
 524                 if (0 <= lookup_sha1_array(&skipped_revs,
 525                                            list->item->object.sha1)) {
 526                         /* Move current to tried list */
 527                         *tried = list;
 528                         tried = &list->next;
 529                 } else {
 530                         if (!show_all)
 531                                 return list;
 532                         /* Move current to filtered list */
 533                         *f = list;
 534                         f = &list->next;
 535                 }
 536                 list = next;
 537         }
 538
 539         return filtered;
 540 }
 541
 542 static void bisect_rev_setup(struct rev_info *revs, const char *prefix)
 543 {
 544         struct argv_array rev_argv = { NULL, 0, 0 };
 545         int i;
 546
 547         init_revisions(revs, prefix);
 548         revs->abbrev = 0;
 549         revs->commit_format = CMIT_FMT_UNSPECIFIED;
 550
 551         /* rev_argv.argv[0] will be ignored by setup_revisions */
 552         argv_array_push(&rev_argv, xstrdup("bisect_rev_setup"));
 553         argv_array_push_sha1(&rev_argv, current_bad_sha1, "%s");
 554         for (i = 0; i < good_revs.sha1_nr; i++)
 555                 argv_array_push_sha1(&rev_argv, good_revs.sha1[i], "^%s");
 556         argv_array_push(&rev_argv, xstrdup("--"));
 557         read_bisect_paths(&rev_argv);
 558         argv_array_push(&rev_argv, NULL);
 559
 560         setup_revisions(rev_argv.argv_nr, rev_argv.argv, revs, NULL);
 561         revs->limited = 1;
 562 }
 563
 564 static void bisect_common(struct rev_info *revs, int *reaches, int *all)
 565 {
 566         if (prepare_revision_walk(revs))
 567                 die("revision walk setup failed");
 568         if (revs->tree_objects)
 569                 mark_edges_uninteresting(revs->commits, revs, NULL);
 570
 571         revs->commits = find_bisection(revs->commits, reaches, all,
 572                                        !!skipped_revs.sha1_nr);
 573 }
 574
 575 static void exit_if_skipped_commits(struct commit_list *tried,
 576                                     const unsigned char *bad)
 577 {
 578         if (!tried)
 579                 return;
 580
 581         printf("There are only 'skip'ped commits left to test.\n"
 582                "The first bad commit could be any of:\n");
 583         print_commit_list(tried, "%s\n", "%s\n");
 584         if (bad)
 585                 printf("%s\n", sha1_to_hex(bad));
 586         printf("We cannot bisect more!\n");
 587         exit(2);
 588 }
 589
 590 static void mark_expected_rev(char *bisect_rev_hex)
 591 {
 592         int len = strlen(bisect_rev_hex);
 593         const char *filename = git_path("BISECT_EXPECTED_REV");
 594         int fd = open(filename, O_CREAT | O_TRUNC | O_WRONLY, 0600);
 595
 596         if (fd < 0)
 597                 die("could not create file '%s': %s",
 598                     filename, strerror(errno));
 599
 600         bisect_rev_hex[len] = '\n';
 601         write_or_die(fd, bisect_rev_hex, len + 1);
 602         bisect_rev_hex[len] = '\0';
 603
 604         if (close(fd) < 0)
 605                 die("closing file %s: %s", filename, strerror(errno));
 606 }
 607
 608 static int bisect_checkout(char *bisect_rev_hex)
 609 {
 610         int res;
 611
 612         mark_expected_rev(bisect_rev_hex);
 613
 614         argv_checkout[2] = bisect_rev_hex;
 615         res = run_command_v_opt(argv_checkout, RUN_GIT_CMD);
 616         if (res)
 617                 exit(res);
 618
 619         argv_show_branch[1] = bisect_rev_hex;
 620         return run_command_v_opt(argv_show_branch, RUN_GIT_CMD);
 621 }
 622
 623 /*
 624  * We use the convention that exiting with an exit code 10 means that
 625  * the bisection process finished successfully.
 626  * In this case the calling shell script should exit 0.
 627  */
 628 int bisect_next_exit(const char *prefix)
 629 {
 630         struct rev_info revs;
 631         struct commit_list *tried;
 632         int reaches = 0, all = 0, nr;
 633         const unsigned char *bisect_rev;
 634         char bisect_rev_hex[41];
 635
 636         if (read_bisect_refs())
 637                 die("reading bisect refs failed");
 638
 639         bisect_rev_setup(&revs, prefix);
 640
 641         bisect_common(&revs, &reaches, &all);
 642
 643         revs.commits = filter_skipped(revs.commits, &tried, 0);
 644
 645         if (!revs.commits) {
 646                 /*
 647                  * We should exit here only if the "bad"
 648                  * commit is also a "skip" commit.
 649                  */
 650                 exit_if_skipped_commits(tried, NULL);
 651
 652                 printf("%s was both good and bad\n",
 653                        sha1_to_hex(current_bad_sha1));
 654                 exit(1);
 655         }
 656
 657         bisect_rev = revs.commits->item->object.sha1;
 658         memcpy(bisect_rev_hex, sha1_to_hex(bisect_rev), 41);
 659
 660         if (!hashcmp(bisect_rev, current_bad_sha1)) {
 661                 exit_if_skipped_commits(tried, current_bad_sha1);
 662                 printf("%s is first bad commit\n", bisect_rev_hex);
 663                 argv_diff_tree[2] = bisect_rev_hex;
 664                 run_command_v_opt(argv_diff_tree, RUN_GIT_CMD);
 665                 /* This means the bisection process succeeded. */
 666                 exit(10);
 667         }
 668
 669         nr = all - reaches - 1;
 670         printf("Bisecting: %d revisions left to test after this "
 671                "(roughly %d steps)\n", nr, estimate_bisect_steps(all));
 672
 673         return bisect_checkout(bisect_rev_hex);
 674 }
 675