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                         struct object *o = &(parents->item->object);
  74                         parents = parents->next;
  75                         if (o->flags & TMP_MARK)
  76                                 continue;
  77                         printf(" %s", sha1_to_hex(o->sha1));
  78                         o->flags |= TMP_MARK;
  79                 }
  80                 /* TMP_MARK is a general purpose flag that can
  81                  * be used locally, but the user should clean
  82                  * things up after it is done with them.
  83                  */
  84                 for (parents = commit->parents;
  85                      parents;
  86                      parents = parents->next)
  87                         parents->item->object.flags &= ~TMP_MARK;
  88         }
  89         if (revs.commit_format == CMIT_FMT_ONELINE)
  90                 putchar(' ');
  91         else
  92                 putchar('\n');
  93
  94         if (revs.verbose_header) {
  95                 char *buf = NULL;
  96                 unsigned long buflen = 0;
  97                 pretty_print_commit(revs.commit_format, commit, ~0,
  98                                     &buf, &buflen,
  99                                     revs.abbrev, NULL, NULL, revs.date_mode);
 100                 printf("%s%c", buf, hdr_termination);
 101                 free(buf);
 102         }
 103         fflush(stdout);
 104         if (commit->parents) {
 105                 free_commit_list(commit->parents);
 106                 commit->parents = NULL;
 107         }
 108         free(commit->buffer);
 109         commit->buffer = NULL;
 110 }
 111
 112 static void show_object(struct object_array_entry *p)
 113 {
 114         /* An object with name "foo\n0000000..." can be used to
 115          * confuse downstream git-pack-objects very badly.
 116          */
 117         const char *ep = strchr(p->name, '\n');
 118
 119         if (p->item->type == OBJ_BLOB && !has_sha1_file(p->item->sha1))
 120                 die("missing blob object '%s'", sha1_to_hex(p->item->sha1));
 121
 122         if (ep) {
 123                 printf("%s %.*s\n", sha1_to_hex(p->item->sha1),
 124                        (int) (ep - p->name),
 125                        p->name);
 126         }
 127         else
 128                 printf("%s %s\n", sha1_to_hex(p->item->sha1), p->name);
 129 }
 130
 131 static void show_edge(struct commit *commit)
 132 {
 133         printf("-%s\n", sha1_to_hex(commit->object.sha1));
 134 }
 135
 136 /*
 137  * This is a truly stupid algorithm, but it's only
 138  * used for bisection, and we just don't care enough.
 139  *
 140  * We care just barely enough to avoid recursing for
 141  * non-merge entries.
 142  */
 143 static int count_distance(struct commit_list *entry)
 144 {
 145         int nr = 0;
 146
 147         while (entry) {
 148                 struct commit *commit = entry->item;
 149                 struct commit_list *p;
 150
 151                 if (commit->object.flags & (UNINTERESTING | COUNTED))
 152                         break;
 153                 if (!revs.prune_fn || (commit->object.flags & TREECHANGE))
 154                         nr++;
 155                 commit->object.flags |= COUNTED;
 156                 p = commit->parents;
 157                 entry = p;
 158                 if (p) {
 159                         p = p->next;
 160                         while (p) {
 161                                 nr += count_distance(p);
 162                                 p = p->next;
 163                         }
 164                 }
 165         }
 166
 167         return nr;
 168 }
 169
 170 static void clear_distance(struct commit_list *list)
 171 {
 172         while (list) {
 173                 struct commit *commit = list->item;
 174                 commit->object.flags &= ~COUNTED;
 175                 list = list->next;
 176         }
 177 }
 178
 179 #define DEBUG_BISECT 0
 180
 181 static inline int weight(struct commit_list *elem)
 182 {
 183         return *((int*)(elem->item->util));
 184 }
 185
 186 static inline void weight_set(struct commit_list *elem, int weight)
 187 {
 188         *((int*)(elem->item->util)) = weight;
 189 }
 190
 191 static int count_interesting_parents(struct commit *commit)
 192 {
 193         struct commit_list *p;
 194         int count;
 195
 196         for (count = 0, p = commit->parents; p; p = p->next) {
 197                 if (p->item->object.flags & UNINTERESTING)
 198                         continue;
 199                 count++;
 200         }
 201         return count;
 202 }
 203
 204 static inline int halfway(struct commit_list *p, int distance, int nr)
 205 {
 206         /*
 207          * Don't short-cut something we are not going to return!
 208          */
 209         if (revs.prune_fn && !(p->item->object.flags & TREECHANGE))
 210                 return 0;
 211         if (DEBUG_BISECT)
 212                 return 0;
 213         /*
 214          * 2 and 3 are halfway of 5.
 215          * 3 is halfway of 6 but 2 and 4 are not.
 216          */
 217         distance *= 2;
 218         switch (distance - nr) {
 219         case -1: case 0: case 1:
 220                 return 1;
 221         default:
 222                 return 0;
 223         }
 224 }
 225
 226 #if !DEBUG_BISECT
 227 #define show_list(a,b,c,d) do { ; } while (0)
 228 #else
 229 static void show_list(const char *debug, int counted, int nr,
 230                       struct commit_list *list)
 231 {
 232         struct commit_list *p;
 233
 234         fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
 235
 236         for (p = list; p; p = p->next) {
 237                 struct commit_list *pp;
 238                 struct commit *commit = p->item;
 239                 unsigned flags = commit->object.flags;
 240                 enum object_type type;
 241                 unsigned long size;
 242                 char *buf = read_sha1_file(commit->object.sha1, &type, &size);
 243                 char *ep, *sp;
 244
 245                 fprintf(stderr, "%c%c%c ",
 246                         (flags & TREECHANGE) ? 'T' : ' ',
 247                         (flags & UNINTERESTING) ? 'U' : ' ',
 248                         (flags & COUNTED) ? 'C' : ' ');
 249                 if (commit->util)
 250                         fprintf(stderr, "%3d", weight(p));
 251                 else
 252                         fprintf(stderr, "---");
 253                 fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 254                 for (pp = commit->parents; pp; pp = pp->next)
 255                         fprintf(stderr, " %.*s", 8,
 256                                 sha1_to_hex(pp->item->object.sha1));
 257
 258                 sp = strstr(buf, "\n\n");
 259                 if (sp) {
 260                         sp += 2;
 261                         for (ep = sp; *ep && *ep != '\n'; ep++)
 262                                 ;
 263                         fprintf(stderr, " %.*s", (int)(ep - sp), sp);
 264                 }
 265                 fprintf(stderr, "\n");
 266         }
 267 }
 268 #endif /* DEBUG_BISECT */
 269
 270 /*
 271  * zero or positive weight is the number of interesting commits it can
 272  * reach, including itself.  Especially, weight = 0 means it does not
 273  * reach any tree-changing commits (e.g. just above uninteresting one
 274  * but traversal is with pathspec).
 275  *
 276  * weight = -1 means it has one parent and its distance is yet to
 277  * be computed.
 278  *
 279  * weight = -2 means it has more than one parent and its distance is
 280  * unknown.  After running count_distance() first, they will get zero
 281  * or positive distance.
 282  */
 283
 284 static struct commit_list *find_bisection(struct commit_list *list,
 285                                           int *reaches, int *all)
 286 {
 287         int n, nr, on_list, counted, distance;
 288         struct commit_list *p, *best, *next, *last;
 289         int *weights;
 290
 291         show_list("bisection 2 entry", 0, 0, list);
 292
 293         /*
 294          * Count the number of total and tree-changing items on the
 295          * list, while reversing the list.
 296          */
 297         for (nr = on_list = 0, last = NULL, p = list;
 298              p;
 299              p = next) {
 300                 unsigned flags = p->item->object.flags;
 301
 302                 next = p->next;
 303                 if (flags & UNINTERESTING)
 304                         continue;
 305                 p->next = last;
 306                 last = p;
 307                 if (!revs.prune_fn || (flags & TREECHANGE))
 308                         nr++;
 309                 on_list++;
 310         }
 311         list = last;
 312         show_list("bisection 2 sorted", 0, nr, list);
 313
 314         *all = nr;
 315         weights = xcalloc(on_list, sizeof(int*));
 316         counted = 0;
 317
 318         for (n = 0, p = list; p; p = p->next) {
 319                 struct commit *commit = p->item;
 320                 unsigned flags = commit->object.flags;
 321
 322                 p->item->util = &weights[n++];
 323                 switch (count_interesting_parents(commit)) {
 324                 case 0:
 325                         if (!revs.prune_fn || (flags & TREECHANGE)) {
 326                                 weight_set(p, 1);
 327                                 counted++;
 328                                 show_list("bisection 2 count one",
 329                                           counted, nr, list);
 330                         }
 331                         /*
 332                          * otherwise, it is known not to reach any
 333                          * tree-changing commit and gets weight 0.
 334                          */
 335                         break;
 336                 case 1:
 337                         weight_set(p, -1);
 338                         break;
 339                 default:
 340                         weight_set(p, -2);
 341                         break;
 342                 }
 343         }
 344
 345         show_list("bisection 2 initialize", counted, nr, list);
 346
 347         /*
 348          * If you have only one parent in the resulting set
 349          * then you can reach one commit more than that parent
 350          * can reach.  So we do not have to run the expensive
 351          * count_distance() for single strand of pearls.
 352          *
 353          * However, if you have more than one parents, you cannot
 354          * just add their distance and one for yourself, since
 355          * they usually reach the same ancestor and you would
 356          * end up counting them twice that way.
 357          *
 358          * So we will first count distance of merges the usual
 359          * way, and then fill the blanks using cheaper algorithm.
 360          */
 361         for (p = list; p; p = p->next) {
 362                 if (p->item->object.flags & UNINTERESTING)
 363                         continue;
 364                 n = weight(p);
 365                 if (n != -2)
 366                         continue;
 367                 distance = count_distance(p);
 368                 clear_distance(list);
 369                 weight_set(p, distance);
 370
 371                 /* Does it happen to be at exactly half-way? */
 372                 if (halfway(p, distance, nr)) {
 373                         p->next = NULL;
 374                         *reaches = distance;
 375                         free(weights);
 376                         return p;
 377                 }
 378                 counted++;
 379         }
 380
 381         show_list("bisection 2 count_distance", counted, nr, list);
 382
 383         while (counted < nr) {
 384                 for (p = list; p; p = p->next) {
 385                         struct commit_list *q;
 386                         unsigned flags = p->item->object.flags;
 387
 388                         if (0 <= weight(p))
 389                                 continue;
 390                         for (q = p->item->parents; q; q = q->next) {
 391                                 if (q->item->object.flags & UNINTERESTING)
 392                                         continue;
 393                                 if (0 <= weight(q))
 394                                         break;
 395                         }
 396                         if (!q)
 397                                 continue;
 398
 399                         /*
 400                          * weight for p is unknown but q is known.
 401                          * add one for p itself if p is to be counted,
 402                          * otherwise inherit it from q directly.
 403                          */
 404                         if (!revs.prune_fn || (flags & TREECHANGE)) {
 405                                 weight_set(p, weight(q)+1);
 406                                 counted++;
 407                                 show_list("bisection 2 count one",
 408                                           counted, nr, list);
 409                         }
 410                         else
 411                                 weight_set(p, weight(q));
 412
 413                         /* Does it happen to be at exactly half-way? */
 414                         distance = weight(p);
 415                         if (halfway(p, distance, nr)) {
 416                                 p->next = NULL;
 417                                 *reaches = distance;
 418                                 free(weights);
 419                                 return p;
 420                         }
 421                 }
 422         }
 423
 424         show_list("bisection 2 counted all", counted, nr, list);
 425
 426         /* Then find the best one */
 427         counted = -1;
 428         best = list;
 429         for (p = list; p; p = p->next) {
 430                 unsigned flags = p->item->object.flags;
 431
 432                 if (revs.prune_fn && !(flags & TREECHANGE))
 433                         continue;
 434                 distance = weight(p);
 435                 if (nr - distance < distance)
 436                         distance = nr - distance;
 437                 if (distance > counted) {
 438                         best = p;
 439                         counted = distance;
 440                         *reaches = weight(p);
 441                 }
 442         }
 443         if (best)
 444                 best->next = NULL;
 445         free(weights);
 446         return best;
 447 }
 448
 449 static void read_revisions_from_stdin(struct rev_info *revs)
 450 {
 451         char line[1000];
 452
 453         while (fgets(line, sizeof(line), stdin) != NULL) {
 454                 int len = strlen(line);
 455                 if (line[len - 1] == '\n')
 456                         line[--len] = 0;
 457                 if (!len)
 458                         break;
 459                 if (line[0] == '-')
 460                         die("options not supported in --stdin mode");
 461                 if (handle_revision_arg(line, revs, 0, 1))
 462                         die("bad revision '%s'", line);
 463         }
 464 }
 465
 466 int cmd_rev_list(int argc, const char **argv, const char *prefix)
 467 {
 468         struct commit_list *list;
 469         int i;
 470         int read_from_stdin = 0;
 471         int bisect_show_vars = 0;
 472
 473         git_config(git_default_config);
 474         init_revisions(&revs, prefix);
 475         revs.abbrev = 0;
 476         revs.commit_format = CMIT_FMT_UNSPECIFIED;
 477         argc = setup_revisions(argc, argv, &revs, NULL);
 478
 479         for (i = 1 ; i < argc; i++) {
 480                 const char *arg = argv[i];
 481
 482                 if (!strcmp(arg, "--header")) {
 483                         revs.verbose_header = 1;
 484                         continue;
 485                 }
 486                 if (!strcmp(arg, "--timestamp")) {
 487                         show_timestamp = 1;
 488                         continue;
 489                 }
 490                 if (!strcmp(arg, "--bisect")) {
 491                         bisect_list = 1;
 492                         continue;
 493                 }
 494                 if (!strcmp(arg, "--bisect-vars")) {
 495                         bisect_list = 1;
 496                         bisect_show_vars = 1;
 497                         continue;
 498                 }
 499                 if (!strcmp(arg, "--stdin")) {
 500                         if (read_from_stdin++)
 501                                 die("--stdin given twice?");
 502                         read_revisions_from_stdin(&revs);
 503                         continue;
 504                 }
 505                 usage(rev_list_usage);
 506
 507         }
 508         if (revs.commit_format != CMIT_FMT_UNSPECIFIED) {
 509                 /* The command line has a --pretty  */
 510                 hdr_termination = '\n';
 511                 if (revs.commit_format == CMIT_FMT_ONELINE)
 512                         header_prefix = "";
 513                 else
 514                         header_prefix = "commit ";
 515         }
 516         else if (revs.verbose_header)
 517                 /* Only --header was specified */
 518                 revs.commit_format = CMIT_FMT_RAW;
 519
 520         list = revs.commits;
 521
 522         if ((!list &&
 523              (!(revs.tag_objects||revs.tree_objects||revs.blob_objects) &&
 524               !revs.pending.nr)) ||
 525             revs.diff)
 526                 usage(rev_list_usage);
 527
 528         save_commit_buffer = revs.verbose_header || revs.grep_filter;
 529         track_object_refs = 0;
 530         if (bisect_list)
 531                 revs.limited = 1;
 532
 533         prepare_revision_walk(&revs);
 534         if (revs.tree_objects)
 535                 mark_edges_uninteresting(revs.commits, &revs, show_edge);
 536
 537         if (bisect_list) {
 538                 int reaches = reaches, all = all;
 539
 540                 revs.commits = find_bisection(revs.commits, &reaches, &all);
 541                 if (bisect_show_vars) {
 542                         int cnt;
 543                         if (!revs.commits)
 544                                 return 1;
 545                         /*
 546                          * revs.commits can reach "reaches" commits among
 547                          * "all" commits.  If it is good, then there are
 548                          * (all-reaches) commits left to be bisected.
 549                          * On the other hand, if it is bad, then the set
 550                          * to bisect is "reaches".
 551                          * A bisect set of size N has (N-1) commits further
 552                          * to test, as we already know one bad one.
 553                          */
 554                         cnt = all-reaches;
 555                         if (cnt < reaches)
 556                                 cnt = reaches;
 557                         printf("bisect_rev=%s\n"
 558                                "bisect_nr=%d\n"
 559                                "bisect_good=%d\n"
 560                                "bisect_bad=%d\n"
 561                                "bisect_all=%d\n",
 562                                sha1_to_hex(revs.commits->item->object.sha1),
 563                                cnt - 1,
 564                                all - reaches - 1,
 565                                reaches - 1,
 566                                all);
 567                         return 0;
 568                 }
 569         }
 570
 571         traverse_commit_list(&revs, show_commit, show_object);
 572
 573         return 0;
 574 }