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