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