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 "  special purpose:\n"
  39 "    --bisect\n"
  40 "    --bisect-vars"
  41 ;
  42
  43 static struct rev_info revs;
  44
  45 static int bisect_list;
  46 static int show_timestamp;
  47 static int hdr_termination;
  48 static const char *header_prefix;
  49
  50 static void show_commit(struct commit *commit)
  51 {
  52         if (show_timestamp)
  53                 printf("%lu ", commit->date);
  54         if (header_prefix)
  55                 fputs(header_prefix, stdout);
  56         if (commit->object.flags & BOUNDARY)
  57                 putchar('-');
  58         else if (revs.left_right) {
  59                 if (commit->object.flags & SYMMETRIC_LEFT)
  60                         putchar('<');
  61                 else
  62                         putchar('>');
  63         }
  64         if (revs.abbrev_commit && revs.abbrev)
  65                 fputs(find_unique_abbrev(commit->object.sha1, revs.abbrev),
  66                       stdout);
  67         else
  68                 fputs(sha1_to_hex(commit->object.sha1), stdout);
  69         if (revs.parents) {
  70                 struct commit_list *parents = commit->parents;
  71                 while (parents) {
  72                         struct object *o = &(parents->item->object);
  73                         parents = parents->next;
  74                         if (o->flags & TMP_MARK)
  75                                 continue;
  76                         printf(" %s", sha1_to_hex(o->sha1));
  77                         o->flags |= TMP_MARK;
  78                 }
  79                 /* TMP_MARK is a general purpose flag that can
  80                  * be used locally, but the user should clean
  81                  * things up after it is done with them.
  82                  */
  83                 for (parents = commit->parents;
  84                      parents;
  85                      parents = parents->next)
  86                         parents->item->object.flags &= ~TMP_MARK;
  87         }
  88         if (revs.commit_format == CMIT_FMT_ONELINE)
  89                 putchar(' ');
  90         else
  91                 putchar('\n');
  92
  93         if (revs.verbose_header) {
  94                 static char pretty_header[16384];
  95                 pretty_print_commit(revs.commit_format, commit, ~0,
  96                                     pretty_header, sizeof(pretty_header),
  97                                     revs.abbrev, NULL, NULL, revs.relative_date);
  98                 printf("%s%c", pretty_header, hdr_termination);
  99         }
 100         fflush(stdout);
 101         if (commit->parents) {
 102                 free_commit_list(commit->parents);
 103                 commit->parents = NULL;
 104         }
 105         free(commit->buffer);
 106         commit->buffer = NULL;
 107 }
 108
 109 static void show_object(struct object_array_entry *p)
 110 {
 111         /* An object with name "foo\n0000000..." can be used to
 112          * confuse downstream git-pack-objects very badly.
 113          */
 114         const char *ep = strchr(p->name, '\n');
 115         if (ep) {
 116                 printf("%s %.*s\n", sha1_to_hex(p->item->sha1),
 117                        (int) (ep - p->name),
 118                        p->name);
 119         }
 120         else
 121                 printf("%s %s\n", sha1_to_hex(p->item->sha1), p->name);
 122 }
 123
 124 static void show_edge(struct commit *commit)
 125 {
 126         printf("-%s\n", sha1_to_hex(commit->object.sha1));
 127 }
 128
 129 /*
 130  * This is a truly stupid algorithm, but it's only
 131  * used for bisection, and we just don't care enough.
 132  *
 133  * We care just barely enough to avoid recursing for
 134  * non-merge entries.
 135  */
 136 static int count_distance(struct commit_list *entry)
 137 {
 138         int nr = 0;
 139
 140         while (entry) {
 141                 struct commit *commit = entry->item;
 142                 struct commit_list *p;
 143
 144                 if (commit->object.flags & (UNINTERESTING | COUNTED))
 145                         break;
 146                 if (!revs.prune_fn || (commit->object.flags & TREECHANGE))
 147                         nr++;
 148                 commit->object.flags |= COUNTED;
 149                 p = commit->parents;
 150                 entry = p;
 151                 if (p) {
 152                         p = p->next;
 153                         while (p) {
 154                                 nr += count_distance(p);
 155                                 p = p->next;
 156                         }
 157                 }
 158         }
 159
 160         return nr;
 161 }
 162
 163 static void clear_distance(struct commit_list *list)
 164 {
 165         while (list) {
 166                 struct commit *commit = list->item;
 167                 commit->object.flags &= ~COUNTED;
 168                 list = list->next;
 169         }
 170 }
 171
 172 #define DEBUG_BISECT 0
 173
 174 static inline int weight(struct commit_list *elem)
 175 {
 176         return *((int*)(elem->item->util));
 177 }
 178
 179 static inline void weight_set(struct commit_list *elem, int weight)
 180 {
 181         *((int*)(elem->item->util)) = weight;
 182 }
 183
 184 static int count_interesting_parents(struct commit *commit)
 185 {
 186         struct commit_list *p;
 187         int count;
 188
 189         for (count = 0, p = commit->parents; p; p = p->next) {
 190                 if (p->item->object.flags & UNINTERESTING)
 191                         continue;
 192                 count++;
 193         }
 194         return count;
 195 }
 196
 197 static inline int halfway(struct commit_list *p, int distance, int nr)
 198 {
 199         /*
 200          * Don't short-cut something we are not going to return!
 201          */
 202         if (revs.prune_fn && !(p->item->object.flags & TREECHANGE))
 203                 return 0;
 204         if (DEBUG_BISECT)
 205                 return 0;
 206         /*
 207          * 2 and 3 are halfway of 5.
 208          * 3 is halfway of 6 but 2 and 4 are not.
 209          */
 210         distance *= 2;
 211         switch (distance - nr) {
 212         case -1: case 0: case 1:
 213                 return 1;
 214         default:
 215                 return 0;
 216         }
 217 }
 218
 219 #if !DEBUG_BISECT
 220 #define show_list(a,b,c,d) do { ; } while (0)
 221 #else
 222 static void show_list(const char *debug, int counted, int nr,
 223                       struct commit_list *list)
 224 {
 225         struct commit_list *p;
 226
 227         fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
 228
 229         for (p = list; p; p = p->next) {
 230                 struct commit_list *pp;
 231                 struct commit *commit = p->item;
 232                 unsigned flags = commit->object.flags;
 233                 enum object_type type;
 234                 unsigned long size;
 235                 char *buf = read_sha1_file(commit->object.sha1, &type, &size);
 236                 char *ep, *sp;
 237
 238                 fprintf(stderr, "%c%c%c ",
 239                         (flags & TREECHANGE) ? 'T' : ' ',
 240                         (flags & UNINTERESTING) ? 'U' : ' ',
 241                         (flags & COUNTED) ? 'C' : ' ');
 242                 if (commit->util)
 243                         fprintf(stderr, "%3d", weight(p));
 244                 else
 245                         fprintf(stderr, "---");
 246                 fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 247                 for (pp = commit->parents; pp; pp = pp->next)
 248                         fprintf(stderr, " %.*s", 8,
 249                                 sha1_to_hex(pp->item->object.sha1));
 250
 251                 sp = strstr(buf, "\n\n");
 252                 if (sp) {
 253                         sp += 2;
 254                         for (ep = sp; *ep && *ep != '\n'; ep++)
 255                                 ;
 256                         fprintf(stderr, " %.*s", (int)(ep - sp), sp);
 257                 }
 258                 fprintf(stderr, "\n");
 259         }
 260 }
 261 #endif /* DEBUG_BISECT */
 262
 263 /*
 264  * zero or positive weight is the number of interesting commits it can
 265  * reach, including itself.  Especially, weight = 0 means it does not
 266  * reach any tree-changing commits (e.g. just above uninteresting one
 267  * but traversal is with pathspec).
 268  *
 269  * weight = -1 means it has one parent and its distance is yet to
 270  * be computed.
 271  *
 272  * weight = -2 means it has more than one parent and its distance is
 273  * unknown.  After running count_distance() first, they will get zero
 274  * or positive distance.
 275  */
 276
 277 static struct commit_list *find_bisection(struct commit_list *list,
 278                                           int *reaches, int *all)
 279 {
 280         int n, nr, on_list, counted, distance;
 281         struct commit_list *p, *best, *next, *last;
 282         int *weights;
 283
 284         show_list("bisection 2 entry", 0, 0, list);
 285
 286         /*
 287          * Count the number of total and tree-changing items on the
 288          * list, while reversing the list.
 289          */
 290         for (nr = on_list = 0, last = NULL, p = list;
 291              p;
 292              p = next) {
 293                 unsigned flags = p->item->object.flags;
 294
 295                 next = p->next;
 296                 if (flags & UNINTERESTING)
 297                         continue;
 298                 p->next = last;
 299                 last = p;
 300                 if (!revs.prune_fn || (flags & TREECHANGE))
 301                         nr++;
 302                 on_list++;
 303         }
 304         list = last;
 305         show_list("bisection 2 sorted", 0, nr, list);
 306
 307         *all = nr;
 308         weights = xcalloc(on_list, sizeof(int*));
 309         counted = 0;
 310
 311         for (n = 0, p = list; p; p = p->next) {
 312                 struct commit *commit = p->item;
 313                 unsigned flags = commit->object.flags;
 314
 315                 p->item->util = &weights[n++];
 316                 switch (count_interesting_parents(commit)) {
 317                 case 0:
 318                         if (!revs.prune_fn || (flags & TREECHANGE)) {
 319                                 weight_set(p, 1);
 320                                 counted++;
 321                                 show_list("bisection 2 count one",
 322                                           counted, nr, list);
 323                         }
 324                         /*
 325                          * otherwise, it is known not to reach any
 326                          * tree-changing commit and gets weight 0.
 327                          */
 328                         break;
 329                 case 1:
 330                         weight_set(p, -1);
 331                         break;
 332                 default:
 333                         weight_set(p, -2);
 334                         break;
 335                 }
 336         }
 337
 338         show_list("bisection 2 initialize", counted, nr, list);
 339
 340         /*
 341          * If you have only one parent in the resulting set
 342          * then you can reach one commit more than that parent
 343          * can reach.  So we do not have to run the expensive
 344          * count_distance() for single strand of pearls.
 345          *
 346          * However, if you have more than one parents, you cannot
 347          * just add their distance and one for yourself, since
 348          * they usually reach the same ancestor and you would
 349          * end up counting them twice that way.
 350          *
 351          * So we will first count distance of merges the usual
 352          * way, and then fill the blanks using cheaper algorithm.
 353          */
 354         for (p = list; p; p = p->next) {
 355                 if (p->item->object.flags & UNINTERESTING)
 356                         continue;
 357                 n = weight(p);
 358                 if (n != -2)
 359                         continue;
 360                 distance = count_distance(p);
 361                 clear_distance(list);
 362                 weight_set(p, distance);
 363
 364                 /* Does it happen to be at exactly half-way? */
 365                 if (halfway(p, distance, nr)) {
 366                         p->next = NULL;
 367                         *reaches = distance;
 368                         free(weights);
 369                         return p;
 370                 }
 371                 counted++;
 372         }
 373
 374         show_list("bisection 2 count_distance", counted, nr, list);
 375
 376         while (counted < nr) {
 377                 for (p = list; p; p = p->next) {
 378                         struct commit_list *q;
 379                         unsigned flags = p->item->object.flags;
 380
 381                         if (0 <= weight(p))
 382                                 continue;
 383                         for (q = p->item->parents; q; q = q->next) {
 384                                 if (q->item->object.flags & UNINTERESTING)
 385                                         continue;
 386                                 if (0 <= weight(q))
 387                                         break;
 388                         }
 389                         if (!q)
 390                                 continue;
 391
 392                         /*
 393                          * weight for p is unknown but q is known.
 394                          * add one for p itself if p is to be counted,
 395                          * otherwise inherit it from q directly.
 396                          */
 397                         if (!revs.prune_fn || (flags & TREECHANGE)) {
 398                                 weight_set(p, weight(q)+1);
 399                                 counted++;
 400                                 show_list("bisection 2 count one",
 401                                           counted, nr, list);
 402                         }
 403                         else
 404                                 weight_set(p, weight(q));
 405
 406                         /* Does it happen to be at exactly half-way? */
 407                         distance = weight(p);
 408                         if (halfway(p, distance, nr)) {
 409                                 p->next = NULL;
 410                                 *reaches = distance;
 411                                 free(weights);
 412                                 return p;
 413                         }
 414                 }
 415         }
 416
 417         show_list("bisection 2 counted all", counted, nr, list);
 418
 419         /* Then find the best one */
 420         counted = -1;
 421         best = list;
 422         for (p = list; p; p = p->next) {
 423                 unsigned flags = p->item->object.flags;
 424
 425                 if (revs.prune_fn && !(flags & TREECHANGE))
 426                         continue;
 427                 distance = weight(p);
 428                 if (nr - distance < distance)
 429                         distance = nr - distance;
 430                 if (distance > counted) {
 431                         best = p;
 432                         counted = distance;
 433                         *reaches = weight(p);
 434                 }
 435         }
 436         if (best)
 437                 best->next = NULL;
 438         free(weights);
 439         return best;
 440 }
 441
 442 static void read_revisions_from_stdin(struct rev_info *revs)
 443 {
 444         char line[1000];
 445
 446         while (fgets(line, sizeof(line), stdin) != NULL) {
 447                 int len = strlen(line);
 448                 if (line[len - 1] == '\n')
 449                         line[--len] = 0;
 450                 if (!len)
 451                         break;
 452                 if (line[0] == '-')
 453                         die("options not supported in --stdin mode");
 454                 if (handle_revision_arg(line, revs, 0, 1))
 455                         die("bad revision '%s'", line);
 456         }
 457 }
 458
 459 int cmd_rev_list(int argc, const char **argv, const char *prefix)
 460 {
 461         struct commit_list *list;
 462         int i;
 463         int read_from_stdin = 0;
 464         int bisect_show_vars = 0;
 465
 466         git_config(git_default_config);
 467         init_revisions(&revs, prefix);
 468         revs.abbrev = 0;
 469         revs.commit_format = CMIT_FMT_UNSPECIFIED;
 470         argc = setup_revisions(argc, argv, &revs, NULL);
 471
 472         for (i = 1 ; i < argc; i++) {
 473                 const char *arg = argv[i];
 474
 475                 if (!strcmp(arg, "--header")) {
 476                         revs.verbose_header = 1;
 477                         continue;
 478                 }
 479                 if (!strcmp(arg, "--timestamp")) {
 480                         show_timestamp = 1;
 481                         continue;
 482                 }
 483                 if (!strcmp(arg, "--bisect")) {
 484                         bisect_list = 1;
 485                         continue;
 486                 }
 487                 if (!strcmp(arg, "--bisect-vars")) {
 488                         bisect_list = 1;
 489                         bisect_show_vars = 1;
 490                         continue;
 491                 }
 492                 if (!strcmp(arg, "--stdin")) {
 493                         if (read_from_stdin++)
 494                                 die("--stdin given twice?");
 495                         read_revisions_from_stdin(&revs);
 496                         continue;
 497                 }
 498                 usage(rev_list_usage);
 499
 500         }
 501         if (revs.commit_format != CMIT_FMT_UNSPECIFIED) {
 502                 /* The command line has a --pretty  */
 503                 hdr_termination = '\n';
 504                 if (revs.commit_format == CMIT_FMT_ONELINE)
 505                         header_prefix = "";
 506                 else
 507                         header_prefix = "commit ";
 508         }
 509         else if (revs.verbose_header)
 510                 /* Only --header was specified */
 511                 revs.commit_format = CMIT_FMT_RAW;
 512
 513         list = revs.commits;
 514
 515         if ((!list &&
 516              (!(revs.tag_objects||revs.tree_objects||revs.blob_objects) &&
 517               !revs.pending.nr)) ||
 518             revs.diff)
 519                 usage(rev_list_usage);
 520
 521         save_commit_buffer = revs.verbose_header || revs.grep_filter;
 522         track_object_refs = 0;
 523         if (bisect_list)
 524                 revs.limited = 1;
 525
 526         prepare_revision_walk(&revs);
 527         if (revs.tree_objects)
 528                 mark_edges_uninteresting(revs.commits, &revs, show_edge);
 529
 530         if (bisect_list) {
 531                 int reaches = reaches, all = all;
 532
 533                 revs.commits = find_bisection(revs.commits, &reaches, &all);
 534                 if (bisect_show_vars) {
 535                         int cnt;
 536                         if (!revs.commits)
 537                                 return 1;
 538                         /*
 539                          * revs.commits can reach "reaches" commits among
 540                          * "all" commits.  If it is good, then there are
 541                          * (all-reaches) commits left to be bisected.
 542                          * On the other hand, if it is bad, then the set
 543                          * to bisect is "reaches".
 544                          * A bisect set of size N has (N-1) commits further
 545                          * to test, as we already know one bad one.
 546                          */
 547                         cnt = all-reaches;
 548                         if (cnt < reaches)
 549                                 cnt = reaches;
 550                         printf("bisect_rev=%s\n"
 551                                "bisect_nr=%d\n"
 552                                "bisect_good=%d\n"
 553                                "bisect_bad=%d\n"
 554                                "bisect_all=%d\n",
 555                                sha1_to_hex(revs.commits->item->object.sha1),
 556                                cnt - 1,
 557                                all - reaches - 1,
 558                                reaches - 1,
 559                                all);
 560                         return 0;
 561                 }
 562         }
 563
 564         traverse_commit_list(&revs, show_commit, show_object);
 565
 566         return 0;
 567 }