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 distance, 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         distance *= 2;
 205         switch (distance - nr) {
 206         case -1: case 0: case 1:
 207                 return 1;
 208         default:
 209                 return 0;
 210         }
 211 }
 212
 213 #if !DEBUG_BISECT
 214 #define show_list(a,b,c,d) do { ; } while (0)
 215 #else
 216 static void show_list(const char *debug, int counted, int nr,
 217                       struct commit_list *list)
 218 {
 219         struct commit_list *p;
 220
 221         fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
 222
 223         for (p = list; p; p = p->next) {
 224                 struct commit_list *pp;
 225                 struct commit *commit = p->item;
 226                 unsigned flags = commit->object.flags;
 227                 enum object_type type;
 228                 unsigned long size;
 229                 char *buf = read_sha1_file(commit->object.sha1, &type, &size);
 230                 char *ep, *sp;
 231
 232                 fprintf(stderr, "%c%c%c ",
 233                         (flags & TREECHANGE) ? 'T' : ' ',
 234                         (flags & UNINTERESTING) ? 'U' : ' ',
 235                         (flags & COUNTED) ? 'C' : ' ');
 236                 if (commit->util)
 237                         fprintf(stderr, "%3d", weight(p));
 238                 else
 239                         fprintf(stderr, "---");
 240                 fprintf(stderr, " %.*s", 8, sha1_to_hex(commit->object.sha1));
 241                 for (pp = commit->parents; pp; pp = pp->next)
 242                         fprintf(stderr, " %.*s", 8,
 243                                 sha1_to_hex(pp->item->object.sha1));
 244
 245                 sp = strstr(buf, "\n\n");
 246                 if (sp) {
 247                         sp += 2;
 248                         for (ep = sp; *ep && *ep != '\n'; ep++)
 249                                 ;
 250                         fprintf(stderr, " %.*s", (int)(ep - sp), sp);
 251                 }
 252                 fprintf(stderr, "\n");
 253         }
 254 }
 255 #endif /* DEBUG_BISECT */
 256
 257 /*
 258  * zero or positive weight is the number of interesting commits it can
 259  * reach, including itself.  Especially, weight = 0 means it does not
 260  * reach any tree-changing commits (e.g. just above uninteresting one
 261  * but traversal is with pathspec).
 262  *
 263  * weight = -1 means it has one parent and its distance is yet to
 264  * be computed.
 265  *
 266  * weight = -2 means it has more than one parent and its distance is
 267  * unknown.  After running count_distance() first, they will get zero
 268  * or positive distance.
 269  */
 270
 271 static struct commit_list *find_bisection(struct commit_list *list,
 272                                           int *reaches, int *all)
 273 {
 274         int n, nr, on_list, counted, distance;
 275         struct commit_list *p, *best, *next, *last;
 276         int *weights;
 277
 278         show_list("bisection 2 entry", 0, 0, list);
 279
 280         /*
 281          * Count the number of total and tree-changing items on the
 282          * list, while reversing the list.
 283          */
 284         for (nr = on_list = 0, last = NULL, p = list;
 285              p;
 286              p = next) {
 287                 unsigned flags = p->item->object.flags;
 288
 289                 next = p->next;
 290                 if (flags & UNINTERESTING)
 291                         continue;
 292                 p->next = last;
 293                 last = p;
 294                 if (!revs.prune_fn || (flags & TREECHANGE))
 295                         nr++;
 296                 on_list++;
 297         }
 298         list = last;
 299         show_list("bisection 2 sorted", 0, nr, list);
 300
 301         *all = nr;
 302         weights = xcalloc(on_list, sizeof(*weights));
 303         counted = 0;
 304
 305         for (n = 0, p = list; p; p = p->next) {
 306                 struct commit *commit = p->item;
 307                 unsigned flags = commit->object.flags;
 308
 309                 p->item->util = &weights[n++];
 310                 switch (count_interesting_parents(commit)) {
 311                 case 0:
 312                         if (!revs.prune_fn || (flags & TREECHANGE)) {
 313                                 weight_set(p, 1);
 314                                 counted++;
 315                                 show_list("bisection 2 count one",
 316                                           counted, nr, list);
 317                         }
 318                         /*
 319                          * otherwise, it is known not to reach any
 320                          * tree-changing commit and gets weight 0.
 321                          */
 322                         break;
 323                 case 1:
 324                         weight_set(p, -1);
 325                         break;
 326                 default:
 327                         weight_set(p, -2);
 328                         break;
 329                 }
 330         }
 331
 332         show_list("bisection 2 initialize", counted, nr, list);
 333
 334         /*
 335          * If you have only one parent in the resulting set
 336          * then you can reach one commit more than that parent
 337          * can reach.  So we do not have to run the expensive
 338          * count_distance() for single strand of pearls.
 339          *
 340          * However, if you have more than one parents, you cannot
 341          * just add their distance and one for yourself, since
 342          * they usually reach the same ancestor and you would
 343          * end up counting them twice that way.
 344          *
 345          * So we will first count distance of merges the usual
 346          * way, and then fill the blanks using cheaper algorithm.
 347          */
 348         for (p = list; p; p = p->next) {
 349                 if (p->item->object.flags & UNINTERESTING)
 350                         continue;
 351                 n = weight(p);
 352                 if (n != -2)
 353                         continue;
 354                 distance = count_distance(p);
 355                 clear_distance(list);
 356                 weight_set(p, distance);
 357
 358                 /* Does it happen to be at exactly half-way? */
 359                 if (halfway(p, distance, nr)) {
 360                         p->next = NULL;
 361                         *reaches = distance;
 362                         free(weights);
 363                         return p;
 364                 }
 365                 counted++;
 366         }
 367
 368         show_list("bisection 2 count_distance", counted, nr, list);
 369
 370         while (counted < nr) {
 371                 for (p = list; p; p = p->next) {
 372                         struct commit_list *q;
 373                         unsigned flags = p->item->object.flags;
 374
 375                         if (0 <= weight(p))
 376                                 continue;
 377                         for (q = p->item->parents; q; q = q->next) {
 378                                 if (q->item->object.flags & UNINTERESTING)
 379                                         continue;
 380                                 if (0 <= weight(q))
 381                                         break;
 382                         }
 383                         if (!q)
 384                                 continue;
 385
 386                         /*
 387                          * weight for p is unknown but q is known.
 388                          * add one for p itself if p is to be counted,
 389                          * otherwise inherit it from q directly.
 390                          */
 391                         if (!revs.prune_fn || (flags & TREECHANGE)) {
 392                                 weight_set(p, weight(q)+1);
 393                                 counted++;
 394                                 show_list("bisection 2 count one",
 395                                           counted, nr, list);
 396                         }
 397                         else
 398                                 weight_set(p, weight(q));
 399
 400                         /* Does it happen to be at exactly half-way? */
 401                         distance = weight(p);
 402                         if (halfway(p, distance, nr)) {
 403                                 p->next = NULL;
 404                                 *reaches = distance;
 405                                 free(weights);
 406                                 return p;
 407                         }
 408                 }
 409         }
 410
 411         show_list("bisection 2 counted all", counted, nr, list);
 412
 413         /* Then find the best one */
 414         counted = -1;
 415         best = list;
 416         for (p = list; p; p = p->next) {
 417                 unsigned flags = p->item->object.flags;
 418
 419                 if (revs.prune_fn && !(flags & TREECHANGE))
 420                         continue;
 421                 distance = weight(p);
 422                 if (nr - distance < distance)
 423                         distance = nr - distance;
 424                 if (distance > counted) {
 425                         best = p;
 426                         counted = distance;
 427                         *reaches = weight(p);
 428                 }
 429         }
 430         if (best)
 431                 best->next = NULL;
 432         free(weights);
 433         return best;
 434 }
 435
 436 static void read_revisions_from_stdin(struct rev_info *revs)
 437 {
 438         char line[1000];
 439
 440         while (fgets(line, sizeof(line), stdin) != NULL) {
 441                 int len = strlen(line);
 442                 if (line[len - 1] == '\n')
 443                         line[--len] = 0;
 444                 if (!len)
 445                         break;
 446                 if (line[0] == '-')
 447                         die("options not supported in --stdin mode");
 448                 if (handle_revision_arg(line, revs, 0, 1))
 449                         die("bad revision '%s'", line);
 450         }
 451 }
 452
 453 int cmd_rev_list(int argc, const char **argv, const char *prefix)
 454 {
 455         struct commit_list *list;
 456         int i;
 457         int read_from_stdin = 0;
 458         int bisect_show_vars = 0;
 459
 460         git_config(git_default_config);
 461         init_revisions(&revs, prefix);
 462         revs.abbrev = 0;
 463         revs.commit_format = CMIT_FMT_UNSPECIFIED;
 464         argc = setup_revisions(argc, argv, &revs, NULL);
 465
 466         for (i = 1 ; i < argc; i++) {
 467                 const char *arg = argv[i];
 468
 469                 if (!strcmp(arg, "--header")) {
 470                         revs.verbose_header = 1;
 471                         continue;
 472                 }
 473                 if (!strcmp(arg, "--timestamp")) {
 474                         show_timestamp = 1;
 475                         continue;
 476                 }
 477                 if (!strcmp(arg, "--bisect")) {
 478                         bisect_list = 1;
 479                         continue;
 480                 }
 481                 if (!strcmp(arg, "--bisect-vars")) {
 482                         bisect_list = 1;
 483                         bisect_show_vars = 1;
 484                         continue;
 485                 }
 486                 if (!strcmp(arg, "--stdin")) {
 487                         if (read_from_stdin++)
 488                                 die("--stdin given twice?");
 489                         read_revisions_from_stdin(&revs);
 490                         continue;
 491                 }
 492                 usage(rev_list_usage);
 493
 494         }
 495         if (revs.commit_format != CMIT_FMT_UNSPECIFIED) {
 496                 /* The command line has a --pretty  */
 497                 hdr_termination = '\n';
 498                 if (revs.commit_format == CMIT_FMT_ONELINE)
 499                         header_prefix = "";
 500                 else
 501                         header_prefix = "commit ";
 502         }
 503         else if (revs.verbose_header)
 504                 /* Only --header was specified */
 505                 revs.commit_format = CMIT_FMT_RAW;
 506
 507         list = revs.commits;
 508
 509         if ((!list &&
 510              (!(revs.tag_objects||revs.tree_objects||revs.blob_objects) &&
 511               !revs.pending.nr)) ||
 512             revs.diff)
 513                 usage(rev_list_usage);
 514
 515         save_commit_buffer = revs.verbose_header || revs.grep_filter;
 516         track_object_refs = 0;
 517         if (bisect_list)
 518                 revs.limited = 1;
 519
 520         prepare_revision_walk(&revs);
 521         if (revs.tree_objects)
 522                 mark_edges_uninteresting(revs.commits, &revs, show_edge);
 523
 524         if (bisect_list) {
 525                 int reaches = reaches, all = all;
 526
 527                 revs.commits = find_bisection(revs.commits, &reaches, &all);
 528                 if (bisect_show_vars) {
 529                         int cnt;
 530                         if (!revs.commits)
 531                                 return 1;
 532                         /*
 533                          * revs.commits can reach "reaches" commits among
 534                          * "all" commits.  If it is good, then there are
 535                          * (all-reaches) commits left to be bisected.
 536                          * On the other hand, if it is bad, then the set
 537                          * to bisect is "reaches".
 538                          * A bisect set of size N has (N-1) commits further
 539                          * to test, as we already know one bad one.
 540                          */
 541                         cnt = all-reaches;
 542                         if (cnt < reaches)
 543                                 cnt = reaches;
 544                         printf("bisect_rev=%s\n"
 545                                "bisect_nr=%d\n"
 546                                "bisect_good=%d\n"
 547                                "bisect_bad=%d\n"
 548                                "bisect_all=%d\n",
 549                                sha1_to_hex(revs.commits->item->object.sha1),
 550                                cnt - 1,
 551                                all - reaches - 1,
 552                                reaches - 1,
 553                                all);
 554                         return 0;
 555                 }
 556         }
 557
 558         traverse_commit_list(&revs, show_commit, show_object);
 559
 560         return 0;
 561 }