builtin-rev-list.c

   1 #include "cache.h"
   2 #include "refs.h"
   3 #include "tag.h"
   4 #include "commit.h"
   5 #include "tree.h"
   6 #include "blob.h"
   7 #include "tree-walk.h"
   8 #include "diff.h"
   9 #include "revision.h"
  10 #include "list-objects.h"
  11 #include "builtin.h"
  12 #include "log-tree.h"
  13 #include "graph.h"
  14
  15 /* bits #0-15 in revision.h */
  16
  17 #define COUNTED         (1u<<16)
  18
  19 static const char rev_list_usage[] =
  20 "git-rev-list [OPTION] <commit-id>... [ -- paths... ]\n"
  21 "  limiting output:\n"
  22 "    --max-count=nr\n"
  23 "    --max-age=epoch\n"
  24 "    --min-age=epoch\n"
  25 "    --sparse\n"
  26 "    --no-merges\n"
  27 "    --remove-empty\n"
  28 "    --all\n"
  29 "    --branches\n"
  30 "    --tags\n"
  31 "    --remotes\n"
  32 "    --stdin\n"
  33 "    --quiet\n"
  34 "  ordering output:\n"
  35 "    --topo-order\n"
  36 "    --date-order\n"
  37 "    --reverse\n"
  38 "  formatting output:\n"
  39 "    --parents\n"
  40 "    --children\n"
  41 "    --objects | --objects-edge\n"
  42 "    --unpacked\n"
  43 "    --header | --pretty\n"
  44 "    --abbrev=nr | --no-abbrev\n"
  45 "    --abbrev-commit\n"
  46 "    --left-right\n"
  47 "  special purpose:\n"
  48 "    --bisect\n"
  49 "    --bisect-vars\n"
  50 "    --bisect-all"
  51 ;
  52
  53 static struct rev_info revs;
  54
  55 static int bisect_list;
  56 static int show_timestamp;
  57 static int hdr_termination;
  58 static const char *header_prefix;
  59
  60 static void finish_commit(struct commit *commit);
  61 static void show_commit(struct commit *commit)
  62 {
  63         graph_show_commit(revs.graph);
  64
  65         if (show_timestamp)
  66                 printf("%lu ", commit->date);
  67         if (header_prefix)
  68                 fputs(header_prefix, stdout);
  69         if (commit->object.flags & BOUNDARY)
  70                 putchar('-');
  71         else if (commit->object.flags & UNINTERESTING)
  72                 putchar('^');
  73         else if (revs.left_right) {
  74                 if (commit->object.flags & SYMMETRIC_LEFT)
  75                         putchar('<');
  76                 else
  77                         putchar('>');
  78         }
  79         if (revs.abbrev_commit && revs.abbrev)
  80                 fputs(find_unique_abbrev(commit->object.sha1, revs.abbrev),
  81                       stdout);
  82         else
  83                 fputs(sha1_to_hex(commit->object.sha1), stdout);
  84         if (revs.print_parents) {
  85                 struct commit_list *parents = commit->parents;
  86                 while (parents) {
  87                         printf(" %s", sha1_to_hex(parents->item->object.sha1));
  88                         parents = parents->next;
  89                 }
  90         }
  91         if (revs.children.name) {
  92                 struct commit_list *children;
  93
  94                 children = lookup_decoration(&revs.children, &commit->object);
  95                 while (children) {
  96                         printf(" %s", sha1_to_hex(children->item->object.sha1));
  97                         children = children->next;
  98                 }
  99         }
 100         show_decorations(commit);
 101         if (revs.commit_format == CMIT_FMT_ONELINE)
 102                 putchar(' ');
 103         else
 104                 putchar('\n');
 105
 106         if (revs.verbose_header && commit->buffer) {
 107                 struct strbuf buf;
 108                 strbuf_init(&buf, 0);
 109                 pretty_print_commit(revs.commit_format, commit,
 110                                     &buf, revs.abbrev, NULL, NULL,
 111                                     revs.date_mode, 0);
 112                 if (revs.graph) {
 113                         if (buf.len) {
 114                                 if (revs.commit_format != CMIT_FMT_ONELINE)
 115                                         graph_show_oneline(revs.graph);
 116
 117                                 graph_show_commit_msg(revs.graph, &buf);
 118
 119                                 /*
 120                                  * Add a newline after the commit message.
 121                                  *
 122                                  * Usually, this newline produces a blank
 123                                  * padding line between entries, in which case
 124                                  * we need to add graph padding on this line.
 125                                  *
 126                                  * However, the commit message may not end in a
 127                                  * newline.  In this case the newline simply
 128                                  * ends the last line of the commit message,
 129                                  * and we don't need any graph output.  (This
 130                                  * always happens with CMIT_FMT_ONELINE, and it
 131                                  * happens with CMIT_FMT_USERFORMAT when the
 132                                  * format doesn't explicitly end in a newline.)
 133                                  */
 134                                 if (buf.len && buf.buf[buf.len - 1] == '\n')
 135                                         graph_show_padding(revs.graph);
 136                                 putchar('\n');
 137                         } else {
 138                                 /*
 139                                  * If the message buffer is empty, just show
 140                                  * the rest of the graph output for this
 141                                  * commit.
 142                                  */
 143                                 if (graph_show_remainder(revs.graph))
 144                                         putchar('\n');
 145                         }
 146                 } else {
 147                         if (buf.len)
 148                                 printf("%s%c", buf.buf, hdr_termination);
 149                 }
 150                 strbuf_release(&buf);
 151         } else {
 152                 if (graph_show_remainder(revs.graph))
 153                         putchar('\n');
 154         }
 155         maybe_flush_or_die(stdout, "stdout");
 156         finish_commit(commit);
 157 }
 158
 159 static void finish_commit(struct commit *commit)
 160 {
 161         if (commit->parents) {
 162                 free_commit_list(commit->parents);
 163                 commit->parents = NULL;
 164         }
 165         free(commit->buffer);
 166         commit->buffer = NULL;
 167 }
 168
 169 static void finish_object(struct object_array_entry *p)
 170 {
 171         if (p->item->type == OBJ_BLOB && !has_sha1_file(p->item->sha1))
 172                 die("missing blob object '%s'", sha1_to_hex(p->item->sha1));
 173 }
 174
 175 static void show_object(struct object_array_entry *p)
 176 {
 177         /* An object with name "foo\n0000000..." can be used to
 178          * confuse downstream git-pack-objects very badly.
 179          */
 180         const char *ep = strchr(p->name, '\n');
 181
 182         finish_object(p);
 183         if (ep) {
 184                 printf("%s %.*s\n", sha1_to_hex(p->item->sha1),
 185                        (int) (ep - p->name),
 186                        p->name);
 187         }
 188         else
 189                 printf("%s %s\n", sha1_to_hex(p->item->sha1), p->name);
 190 }
 191
 192 static void show_edge(struct commit *commit)
 193 {
 194         printf("-%s\n", sha1_to_hex(commit->object.sha1));
 195 }
 196
 197 /*
 198  * This is a truly stupid algorithm, but it's only
 199  * used for bisection, and we just don't care enough.
 200  *
 201  * We care just barely enough to avoid recursing for
 202  * non-merge entries.
 203  */
 204 static int count_distance(struct commit_list *entry)
 205 {
 206         int nr = 0;
 207
 208         while (entry) {
 209                 struct commit *commit = entry->item;
 210                 struct commit_list *p;
 211
 212                 if (commit->object.flags & (UNINTERESTING | COUNTED))
 213                         break;
 214                 if (!(commit->object.flags & TREESAME))
 215                         nr++;
 216                 commit->object.flags |= COUNTED;
 217                 p = commit->parents;
 218                 entry = p;
 219                 if (p) {
 220                         p = p->next;
 221                         while (p) {
 222                                 nr += count_distance(p);
 223                                 p = p->next;
 224                         }
 225                 }
 226         }
 227
 228         return nr;
 229 }
 230
 231 static void clear_distance(struct commit_list *list)
 232 {
 233         while (list) {
 234                 struct commit *commit = list->item;
 235                 commit->object.flags &= ~COUNTED;
 236                 list = list->next;
 237         }
 238 }
 239
 240 #define DEBUG_BISECT 0
 241
 242 static inline int weight(struct commit_list *elem)
 243 {
 244         return *((int*)(elem->item->util));
 245 }
 246
 247 static inline void weight_set(struct commit_list *elem, int weight)
 248 {
 249         *((int*)(elem->item->util)) = weight;
 250 }
 251
 252 static int count_interesting_parents(struct commit *commit)
 253 {
 254         struct commit_list *p;
 255         int count;
 256
 257         for (count = 0, p = commit->parents; p; p = p->next) {
 258                 if (p->item->object.flags & UNINTERESTING)
 259                         continue;
 260                 count++;
 261         }
 262         return count;
 263 }
 264
 265 static inline int halfway(struct commit_list *p, int nr)
 266 {
 267         /*
 268          * Don't short-cut something we are not going to return!
 269          */
 270         if (p->item->object.flags & TREESAME)
 271                 return 0;
 272         if (DEBUG_BISECT)
 273                 return 0;
 274         /*
 275          * 2 and 3 are halfway of 5.
 276          * 3 is halfway of 6 but 2 and 4 are not.
 277          */
 278         switch (2 * weight(p) - nr) {
 279         case -1: case 0: case 1:
 280                 return 1;
 281         default:
 282                 return 0;
 283         }
 284 }
 285
 286 #if !DEBUG_BISECT
 287 #define show_list(a,b,c,d) do { ; } while (0)
 288 #else
 289 static void show_list(const char *debug, int counted, int nr,
 290                       struct commit_list *list)
 291 {
 292         struct commit_list *p;
 293
 294         fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
 295
 296         for (p = list; p; p = p->next) {
 297                 struct commit_list *pp;
 298                 struct commit *commit = p->item;
 299                 unsigned flags = commit->object.flags;
 300                 enum object_type type;
 301                 unsigned long size;
 302                 char *buf = read_sha1_file(commit->object.sha1, &type, &size);
 303                 char *ep, *sp;
 304
 305                 fprintf(stderr, "%c%c%c ",
 306                         (flags & TREESAME) ? ' ' : 'T',
 307                         (flags & UNINTERESTING) ? 'U' : ' ',
 308                         (flags & COUNTED) ? 'C' : ' ');
 309                 if (commit->util)
 310                         fprintf(stderr, "%3d", weight(p));
 311                 else
 312                         fprintf(stderr, "---");
 313                 fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 314                 for (pp = commit->parents; pp; pp = pp->next)
 315                         fprintf(stderr, " %.*s", 8,
 316                                 sha1_to_hex(pp->item->object.sha1));
 317
 318                 sp = strstr(buf, "\n\n");
 319                 if (sp) {
 320                         sp += 2;
 321                         for (ep = sp; *ep && *ep != '\n'; ep++)
 322                                 ;
 323                         fprintf(stderr, " %.*s", (int)(ep - sp), sp);
 324                 }
 325                 fprintf(stderr, "\n");
 326         }
 327 }
 328 #endif /* DEBUG_BISECT */
 329
 330 static struct commit_list *best_bisection(struct commit_list *list, int nr)
 331 {
 332         struct commit_list *p, *best;
 333         int best_distance = -1;
 334
 335         best = list;
 336         for (p = list; p; p = p->next) {
 337                 int distance;
 338                 unsigned flags = p->item->object.flags;
 339
 340                 if (flags & TREESAME)
 341                         continue;
 342                 distance = weight(p);
 343                 if (nr - distance < distance)
 344                         distance = nr - distance;
 345                 if (distance > best_distance) {
 346                         best = p;
 347                         best_distance = distance;
 348                 }
 349         }
 350
 351         return best;
 352 }
 353
 354 struct commit_dist {
 355         struct commit *commit;
 356         int distance;
 357 };
 358
 359 static int compare_commit_dist(const void *a_, const void *b_)
 360 {
 361         struct commit_dist *a, *b;
 362
 363         a = (struct commit_dist *)a_;
 364         b = (struct commit_dist *)b_;
 365         if (a->distance != b->distance)
 366                 return b->distance - a->distance; /* desc sort */
 367         return hashcmp(a->commit->object.sha1, b->commit->object.sha1);
 368 }
 369
 370 static struct commit_list *best_bisection_sorted(struct commit_list *list, int nr)
 371 {
 372         struct commit_list *p;
 373         struct commit_dist *array = xcalloc(nr, sizeof(*array));
 374         int cnt, i;
 375
 376         for (p = list, cnt = 0; p; p = p->next) {
 377                 int distance;
 378                 unsigned flags = p->item->object.flags;
 379
 380                 if (flags & TREESAME)
 381                         continue;
 382                 distance = weight(p);
 383                 if (nr - distance < distance)
 384                         distance = nr - distance;
 385                 array[cnt].commit = p->item;
 386                 array[cnt].distance = distance;
 387                 cnt++;
 388         }
 389         qsort(array, cnt, sizeof(*array), compare_commit_dist);
 390         for (p = list, i = 0; i < cnt; i++) {
 391                 struct name_decoration *r = xmalloc(sizeof(*r) + 100);
 392                 struct object *obj = &(array[i].commit->object);
 393
 394                 sprintf(r->name, "dist=%d", array[i].distance);
 395                 r->next = add_decoration(&name_decoration, obj, r);
 396                 p->item = array[i].commit;
 397                 p = p->next;
 398         }
 399         if (p)
 400                 p->next = NULL;
 401         free(array);
 402         return list;
 403 }
 404
 405 /*
 406  * zero or positive weight is the number of interesting commits it can
 407  * reach, including itself.  Especially, weight = 0 means it does not
 408  * reach any tree-changing commits (e.g. just above uninteresting one
 409  * but traversal is with pathspec).
 410  *
 411  * weight = -1 means it has one parent and its distance is yet to
 412  * be computed.
 413  *
 414  * weight = -2 means it has more than one parent and its distance is
 415  * unknown.  After running count_distance() first, they will get zero
 416  * or positive distance.
 417  */
 418 static struct commit_list *do_find_bisection(struct commit_list *list,
 419                                              int nr, int *weights,
 420                                              int find_all)
 421 {
 422         int n, counted;
 423         struct commit_list *p;
 424
 425         counted = 0;
 426
 427         for (n = 0, p = list; p; p = p->next) {
 428                 struct commit *commit = p->item;
 429                 unsigned flags = commit->object.flags;
 430
 431                 p->item->util = &weights[n++];
 432                 switch (count_interesting_parents(commit)) {
 433                 case 0:
 434                         if (!(flags & TREESAME)) {
 435                                 weight_set(p, 1);
 436                                 counted++;
 437                                 show_list("bisection 2 count one",
 438                                           counted, nr, list);
 439                         }
 440                         /*
 441                          * otherwise, it is known not to reach any
 442                          * tree-changing commit and gets weight 0.
 443                          */
 444                         break;
 445                 case 1:
 446                         weight_set(p, -1);
 447                         break;
 448                 default:
 449                         weight_set(p, -2);
 450                         break;
 451                 }
 452         }
 453
 454         show_list("bisection 2 initialize", counted, nr, list);
 455
 456         /*
 457          * If you have only one parent in the resulting set
 458          * then you can reach one commit more than that parent
 459          * can reach.  So we do not have to run the expensive
 460          * count_distance() for single strand of pearls.
 461          *
 462          * However, if you have more than one parents, you cannot
 463          * just add their distance and one for yourself, since
 464          * they usually reach the same ancestor and you would
 465          * end up counting them twice that way.
 466          *
 467          * So we will first count distance of merges the usual
 468          * way, and then fill the blanks using cheaper algorithm.
 469          */
 470         for (p = list; p; p = p->next) {
 471                 if (p->item->object.flags & UNINTERESTING)
 472                         continue;
 473                 if (weight(p) != -2)
 474                         continue;
 475                 weight_set(p, count_distance(p));
 476                 clear_distance(list);
 477
 478                 /* Does it happen to be at exactly half-way? */
 479                 if (!find_all && halfway(p, nr))
 480                         return p;
 481                 counted++;
 482         }
 483
 484         show_list("bisection 2 count_distance", counted, nr, list);
 485
 486         while (counted < nr) {
 487                 for (p = list; p; p = p->next) {
 488                         struct commit_list *q;
 489                         unsigned flags = p->item->object.flags;
 490
 491                         if (0 <= weight(p))
 492                                 continue;
 493                         for (q = p->item->parents; q; q = q->next) {
 494                                 if (q->item->object.flags & UNINTERESTING)
 495                                         continue;
 496                                 if (0 <= weight(q))
 497                                         break;
 498                         }
 499                         if (!q)
 500                                 continue;
 501
 502                         /*
 503                          * weight for p is unknown but q is known.
 504                          * add one for p itself if p is to be counted,
 505                          * otherwise inherit it from q directly.
 506                          */
 507                         if (!(flags & TREESAME)) {
 508                                 weight_set(p, weight(q)+1);
 509                                 counted++;
 510                                 show_list("bisection 2 count one",
 511                                           counted, nr, list);
 512                         }
 513                         else
 514                                 weight_set(p, weight(q));
 515
 516                         /* Does it happen to be at exactly half-way? */
 517                         if (!find_all && halfway(p, nr))
 518                                 return p;
 519                 }
 520         }
 521
 522         show_list("bisection 2 counted all", counted, nr, list);
 523
 524         if (!find_all)
 525                 return best_bisection(list, nr);
 526         else
 527                 return best_bisection_sorted(list, nr);
 528 }
 529
 530 static struct commit_list *find_bisection(struct commit_list *list,
 531                                           int *reaches, int *all,
 532                                           int find_all)
 533 {
 534         int nr, on_list;
 535         struct commit_list *p, *best, *next, *last;
 536         int *weights;
 537
 538         show_list("bisection 2 entry", 0, 0, list);
 539
 540         /*
 541          * Count the number of total and tree-changing items on the
 542          * list, while reversing the list.
 543          */
 544         for (nr = on_list = 0, last = NULL, p = list;
 545              p;
 546              p = next) {
 547                 unsigned flags = p->item->object.flags;
 548
 549                 next = p->next;
 550                 if (flags & UNINTERESTING)
 551                         continue;
 552                 p->next = last;
 553                 last = p;
 554                 if (!(flags & TREESAME))
 555                         nr++;
 556                 on_list++;
 557         }
 558         list = last;
 559         show_list("bisection 2 sorted", 0, nr, list);
 560
 561         *all = nr;
 562         weights = xcalloc(on_list, sizeof(*weights));
 563
 564         /* Do the real work of finding bisection commit. */
 565         best = do_find_bisection(list, nr, weights, find_all);
 566         if (best) {
 567                 if (!find_all)
 568                         best->next = NULL;
 569                 *reaches = weight(best);
 570         }
 571         free(weights);
 572         return best;
 573 }
 574
 575 static void read_revisions_from_stdin(struct rev_info *revs)
 576 {
 577         char line[1000];
 578
 579         while (fgets(line, sizeof(line), stdin) != NULL) {
 580                 int len = strlen(line);
 581                 if (len && line[len - 1] == '\n')
 582                         line[--len] = 0;
 583                 if (!len)
 584                         break;
 585                 if (line[0] == '-')
 586                         die("options not supported in --stdin mode");
 587                 if (handle_revision_arg(line, revs, 0, 1))
 588                         die("bad revision '%s'", line);
 589         }
 590 }
 591
 592 int cmd_rev_list(int argc, const char **argv, const char *prefix)
 593 {
 594         struct commit_list *list;
 595         int i;
 596         int read_from_stdin = 0;
 597         int bisect_show_vars = 0;
 598         int bisect_find_all = 0;
 599         int quiet = 0;
 600
 601         git_config(git_default_config);
 602         init_revisions(&revs, prefix);
 603         revs.abbrev = 0;
 604         revs.commit_format = CMIT_FMT_UNSPECIFIED;
 605         argc = setup_revisions(argc, argv, &revs, NULL);
 606
 607         for (i = 1 ; i < argc; i++) {
 608                 const char *arg = argv[i];
 609
 610                 if (!strcmp(arg, "--header")) {
 611                         revs.verbose_header = 1;
 612                         continue;
 613                 }
 614                 if (!strcmp(arg, "--timestamp")) {
 615                         show_timestamp = 1;
 616                         continue;
 617                 }
 618                 if (!strcmp(arg, "--bisect")) {
 619                         bisect_list = 1;
 620                         continue;
 621                 }
 622                 if (!strcmp(arg, "--bisect-all")) {
 623                         bisect_list = 1;
 624                         bisect_find_all = 1;
 625                         continue;
 626                 }
 627                 if (!strcmp(arg, "--bisect-vars")) {
 628                         bisect_list = 1;
 629                         bisect_show_vars = 1;
 630                         continue;
 631                 }
 632                 if (!strcmp(arg, "--stdin")) {
 633                         if (read_from_stdin++)
 634                                 die("--stdin given twice?");
 635                         read_revisions_from_stdin(&revs);
 636                         continue;
 637                 }
 638                 if (!strcmp(arg, "--quiet")) {
 639                         quiet = 1;
 640                         continue;
 641                 }
 642                 usage(rev_list_usage);
 643
 644         }
 645         if (revs.commit_format != CMIT_FMT_UNSPECIFIED) {
 646                 /* The command line has a --pretty  */
 647                 hdr_termination = '\n';
 648                 if (revs.commit_format == CMIT_FMT_ONELINE)
 649                         header_prefix = "";
 650                 else
 651                         header_prefix = "commit ";
 652         }
 653         else if (revs.verbose_header)
 654                 /* Only --header was specified */
 655                 revs.commit_format = CMIT_FMT_RAW;
 656
 657         list = revs.commits;
 658
 659         if ((!list &&
 660              (!(revs.tag_objects||revs.tree_objects||revs.blob_objects) &&
 661               !revs.pending.nr)) ||
 662             revs.diff)
 663                 usage(rev_list_usage);
 664
 665         save_commit_buffer = revs.verbose_header || revs.grep_filter;
 666         if (bisect_list)
 667                 revs.limited = 1;
 668
 669         if (prepare_revision_walk(&revs))
 670                 die("revision walk setup failed");
 671         if (revs.tree_objects)
 672                 mark_edges_uninteresting(revs.commits, &revs, show_edge);
 673
 674         if (bisect_list) {
 675                 int reaches = reaches, all = all;
 676
 677                 revs.commits = find_bisection(revs.commits, &reaches, &all,
 678                                               bisect_find_all);
 679                 if (bisect_show_vars) {
 680                         int cnt;
 681                         char hex[41];
 682                         if (!revs.commits)
 683                                 return 1;
 684                         /*
 685                          * revs.commits can reach "reaches" commits among
 686                          * "all" commits.  If it is good, then there are
 687                          * (all-reaches) commits left to be bisected.
 688                          * On the other hand, if it is bad, then the set
 689                          * to bisect is "reaches".
 690                          * A bisect set of size N has (N-1) commits further
 691                          * to test, as we already know one bad one.
 692                          */
 693                         cnt = all - reaches;
 694                         if (cnt < reaches)
 695                                 cnt = reaches;
 696                         strcpy(hex, sha1_to_hex(revs.commits->item->object.sha1));
 697
 698                         if (bisect_find_all) {
 699                                 traverse_commit_list(&revs, show_commit, show_object);
 700                                 printf("------\n");
 701                         }
 702
 703                         printf("bisect_rev=%s\n"
 704                                "bisect_nr=%d\n"
 705                                "bisect_good=%d\n"
 706                                "bisect_bad=%d\n"
 707                                "bisect_all=%d\n",
 708                                hex,
 709                                cnt - 1,
 710                                all - reaches - 1,
 711                                reaches - 1,
 712                                all);
 713                         return 0;
 714                 }
 715         }
 716
 717         traverse_commit_list(&revs,
 718                 quiet ? finish_commit : show_commit,
 719                 quiet ? finish_object : show_object);
 720
 721         return 0;
 722 }