builtin/fsck.c

   1 #define USE_THE_INDEX_COMPATIBILITY_MACROS
   2 #include "builtin.h"
   3 #include "cache.h"
   4 #include "repository.h"
   5 #include "config.h"
   6 #include "commit.h"
   7 #include "tree.h"
   8 #include "blob.h"
   9 #include "tag.h"
  10 #include "refs.h"
  11 #include "pack.h"
  12 #include "cache-tree.h"
  13 #include "tree-walk.h"
  14 #include "fsck.h"
  15 #include "parse-options.h"
  16 #include "dir.h"
  17 #include "progress.h"
  18 #include "streaming.h"
  19 #include "decorate.h"
  20 #include "packfile.h"
  21 #include "object-store.h"
  22 #include "run-command.h"
  23 #include "worktree.h"
  24
  25 #define REACHABLE 0x0001
  26 #define SEEN      0x0002
  27 #define HAS_OBJ   0x0004
  28 /* This flag is set if something points to this object. */
  29 #define USED      0x0008
  30
  31 static int show_root;
  32 static int show_tags;
  33 static int show_unreachable;
  34 static int include_reflogs = 1;
  35 static int check_full = 1;
  36 static int connectivity_only;
  37 static int check_strict;
  38 static int keep_cache_objects;
  39 static struct fsck_options fsck_walk_options = FSCK_OPTIONS_DEFAULT;
  40 static struct fsck_options fsck_obj_options = FSCK_OPTIONS_DEFAULT;
  41 static int errors_found;
  42 static int write_lost_and_found;
  43 static int verbose;
  44 static int show_progress = -1;
  45 static int show_dangling = 1;
  46 static int name_objects;
  47 #define ERROR_OBJECT 01
  48 #define ERROR_REACHABLE 02
  49 #define ERROR_PACK 04
  50 #define ERROR_REFS 010
  51 #define ERROR_COMMIT_GRAPH 020
  52 #define ERROR_MULTI_PACK_INDEX 040
  53
  54 static const char *describe_object(const struct object_id *oid)
  55 {
  56         return fsck_describe_object(&fsck_walk_options, oid);
  57 }
  58
  59 static const char *printable_type(const struct object_id *oid,
  60                                   enum object_type type)
  61 {
  62         const char *ret;
  63
  64         if (type == OBJ_NONE)
  65                 type = oid_object_info(the_repository, oid, NULL);
  66
  67         ret = type_name(type);
  68         if (!ret)
  69                 ret = _("unknown");
  70
  71         return ret;
  72 }
  73
  74 static int objerror(struct object *obj, const char *err)
  75 {
  76         errors_found |= ERROR_OBJECT;
  77         /* TRANSLATORS: e.g. error in tree 01bfda: <more explanation> */
  78         fprintf_ln(stderr, _("error in %s %s: %s"),
  79                    printable_type(&obj->oid, obj->type),
  80                    describe_object(&obj->oid), err);
  81         return -1;
  82 }
  83
  84 static int fsck_error_func(struct fsck_options *o,
  85                            const struct object_id *oid,
  86                            enum object_type object_type,
  87                            enum fsck_msg_type msg_type,
  88                            enum fsck_msg_id msg_id,
  89                            const char *message)
  90 {
  91         switch (msg_type) {
  92         case FSCK_WARN:
  93                 /* TRANSLATORS: e.g. warning in tree 01bfda: <more explanation> */
  94                 fprintf_ln(stderr, _("warning in %s %s: %s"),
  95                            printable_type(oid, object_type),
  96                            describe_object(oid), message);
  97                 return 0;
  98         case FSCK_ERROR:
  99                 /* TRANSLATORS: e.g. error in tree 01bfda: <more explanation> */
 100                 fprintf_ln(stderr, _("error in %s %s: %s"),
 101                            printable_type(oid, object_type),
 102                            describe_object(oid), message);
 103                 return 1;
 104         default:
 105                 BUG("%d (FSCK_IGNORE?) should never trigger this callback",
 106                     msg_type);
 107         }
 108 }
 109
 110 static struct object_array pending;
 111
 112 static int mark_object(struct object *obj, int type, void *data, struct fsck_options *options)
 113 {
 114         struct object *parent = data;
 115
 116         /*
 117          * The only case data is NULL or type is OBJ_ANY is when
 118          * mark_object_reachable() calls us.  All the callers of
 119          * that function has non-NULL obj hence ...
 120          */
 121         if (!obj) {
 122                 /* ... these references to parent->fld are safe here */
 123                 printf_ln(_("broken link from %7s %s"),
 124                           printable_type(&parent->oid, parent->type),
 125                           describe_object(&parent->oid));
 126                 printf_ln(_("broken link from %7s %s"),
 127                           (type == OBJ_ANY ? _("unknown") : type_name(type)),
 128                           _("unknown"));
 129                 errors_found |= ERROR_REACHABLE;
 130                 return 1;
 131         }
 132
 133         if (type != OBJ_ANY && obj->type != type)
 134                 /* ... and the reference to parent is safe here */
 135                 objerror(parent, _("wrong object type in link"));
 136
 137         if (obj->flags & REACHABLE)
 138                 return 0;
 139         obj->flags |= REACHABLE;
 140
 141         if (is_promisor_object(&obj->oid))
 142                 /*
 143                  * Further recursion does not need to be performed on this
 144                  * object since it is a promisor object (so it does not need to
 145                  * be added to "pending").
 146                  */
 147                 return 0;
 148
 149         if (!(obj->flags & HAS_OBJ)) {
 150                 if (parent && !has_object(the_repository, &obj->oid, 1)) {
 151                         printf_ln(_("broken link from %7s %s\n"
 152                                     "              to %7s %s"),
 153                                   printable_type(&parent->oid, parent->type),
 154                                   describe_object(&parent->oid),
 155                                   printable_type(&obj->oid, obj->type),
 156                                   describe_object(&obj->oid));
 157                         errors_found |= ERROR_REACHABLE;
 158                 }
 159                 return 1;
 160         }
 161
 162         add_object_array(obj, NULL, &pending);
 163         return 0;
 164 }
 165
 166 static void mark_object_reachable(struct object *obj)
 167 {
 168         mark_object(obj, OBJ_ANY, NULL, NULL);
 169 }
 170
 171 static int traverse_one_object(struct object *obj)
 172 {
 173         int result = fsck_walk(obj, obj, &fsck_walk_options);
 174
 175         if (obj->type == OBJ_TREE) {
 176                 struct tree *tree = (struct tree *)obj;
 177                 free_tree_buffer(tree);
 178         }
 179         return result;
 180 }
 181
 182 static int traverse_reachable(void)
 183 {
 184         struct progress *progress = NULL;
 185         unsigned int nr = 0;
 186         int result = 0;
 187         if (show_progress)
 188                 progress = start_delayed_progress(_("Checking connectivity"), 0);
 189         while (pending.nr) {
 190                 result |= traverse_one_object(object_array_pop(&pending));
 191                 display_progress(progress, ++nr);
 192         }
 193         stop_progress(&progress);
 194         return !!result;
 195 }
 196
 197 static int mark_used(struct object *obj, enum object_type object_type,
 198                      void *data, struct fsck_options *options)
 199 {
 200         if (!obj)
 201                 return 1;
 202         obj->flags |= USED;
 203         return 0;
 204 }
 205
 206 static void mark_unreachable_referents(const struct object_id *oid)
 207 {
 208         struct fsck_options options = FSCK_OPTIONS_DEFAULT;
 209         struct object *obj = lookup_object(the_repository, oid);
 210
 211         if (!obj || !(obj->flags & HAS_OBJ))
 212                 return; /* not part of our original set */
 213         if (obj->flags & REACHABLE)
 214                 return; /* reachable objects already traversed */
 215
 216         /*
 217          * Avoid passing OBJ_NONE to fsck_walk, which will parse the object
 218          * (and we want to avoid parsing blobs).
 219          */
 220         if (obj->type == OBJ_NONE) {
 221                 enum object_type type = oid_object_info(the_repository,
 222                                                         &obj->oid, NULL);
 223                 if (type > 0)
 224                         object_as_type(obj, type, 0);
 225         }
 226
 227         options.walk = mark_used;
 228         fsck_walk(obj, NULL, &options);
 229 }
 230
 231 static int mark_loose_unreachable_referents(const struct object_id *oid,
 232                                             const char *path,
 233                                             void *data)
 234 {
 235         mark_unreachable_referents(oid);
 236         return 0;
 237 }
 238
 239 static int mark_packed_unreachable_referents(const struct object_id *oid,
 240                                              struct packed_git *pack,
 241                                              uint32_t pos,
 242                                              void *data)
 243 {
 244         mark_unreachable_referents(oid);
 245         return 0;
 246 }
 247
 248 /*
 249  * Check a single reachable object
 250  */
 251 static void check_reachable_object(struct object *obj)
 252 {
 253         /*
 254          * We obviously want the object to be parsed,
 255          * except if it was in a pack-file and we didn't
 256          * do a full fsck
 257          */
 258         if (!(obj->flags & HAS_OBJ)) {
 259                 if (is_promisor_object(&obj->oid))
 260                         return;
 261                 if (has_object_pack(&obj->oid))
 262                         return; /* it is in pack - forget about it */
 263                 printf_ln(_("missing %s %s"),
 264                           printable_type(&obj->oid, obj->type),
 265                           describe_object(&obj->oid));
 266                 errors_found |= ERROR_REACHABLE;
 267                 return;
 268         }
 269 }
 270
 271 /*
 272  * Check a single unreachable object
 273  */
 274 static void check_unreachable_object(struct object *obj)
 275 {
 276         /*
 277          * Missing unreachable object? Ignore it. It's not like
 278          * we miss it (since it can't be reached), nor do we want
 279          * to complain about it being unreachable (since it does
 280          * not exist).
 281          */
 282         if (!(obj->flags & HAS_OBJ))
 283                 return;
 284
 285         /*
 286          * Unreachable object that exists? Show it if asked to,
 287          * since this is something that is prunable.
 288          */
 289         if (show_unreachable) {
 290                 printf_ln(_("unreachable %s %s"),
 291                           printable_type(&obj->oid, obj->type),
 292                           describe_object(&obj->oid));
 293                 return;
 294         }
 295
 296         /*
 297          * "!USED" means that nothing at all points to it, including
 298          * other unreachable objects. In other words, it's the "tip"
 299          * of some set of unreachable objects, usually a commit that
 300          * got dropped.
 301          *
 302          * Such starting points are more interesting than some random
 303          * set of unreachable objects, so we show them even if the user
 304          * hasn't asked for _all_ unreachable objects. If you have
 305          * deleted a branch by mistake, this is a prime candidate to
 306          * start looking at, for example.
 307          */
 308         if (!(obj->flags & USED)) {
 309                 if (show_dangling)
 310                         printf_ln(_("dangling %s %s"),
 311                                   printable_type(&obj->oid, obj->type),
 312                                   describe_object(&obj->oid));
 313                 if (write_lost_and_found) {
 314                         char *filename = git_pathdup("lost-found/%s/%s",
 315                                 obj->type == OBJ_COMMIT ? "commit" : "other",
 316                                 describe_object(&obj->oid));
 317                         FILE *f;
 318
 319                         if (safe_create_leading_directories_const(filename)) {
 320                                 error(_("could not create lost-found"));
 321                                 free(filename);
 322                                 return;
 323                         }
 324                         f = xfopen(filename, "w");
 325                         if (obj->type == OBJ_BLOB) {
 326                                 if (stream_blob_to_fd(fileno(f), &obj->oid, NULL, 1))
 327                                         die_errno(_("could not write '%s'"), filename);
 328                         } else
 329                                 fprintf(f, "%s\n", describe_object(&obj->oid));
 330                         if (fclose(f))
 331                                 die_errno(_("could not finish '%s'"),
 332                                           filename);
 333                         free(filename);
 334                 }
 335                 return;
 336         }
 337
 338         /*
 339          * Otherwise? It's there, it's unreachable, and some other unreachable
 340          * object points to it. Ignore it - it's not interesting, and we showed
 341          * all the interesting cases above.
 342          */
 343 }
 344
 345 static void check_object(struct object *obj)
 346 {
 347         if (verbose)
 348                 fprintf_ln(stderr, _("Checking %s"), describe_object(&obj->oid));
 349
 350         if (obj->flags & REACHABLE)
 351                 check_reachable_object(obj);
 352         else
 353                 check_unreachable_object(obj);
 354 }
 355
 356 static void check_connectivity(void)
 357 {
 358         int i, max;
 359
 360         /* Traverse the pending reachable objects */
 361         traverse_reachable();
 362
 363         /*
 364          * With --connectivity-only, we won't have actually opened and marked
 365          * unreachable objects with USED. Do that now to make --dangling, etc
 366          * accurate.
 367          */
 368         if (connectivity_only && (show_dangling || write_lost_and_found)) {
 369                 /*
 370                  * Even though we already have a "struct object" for each of
 371                  * these in memory, we must not iterate over the internal
 372                  * object hash as we do below. Our loop would potentially
 373                  * resize the hash, making our iteration invalid.
 374                  *
 375                  * Instead, we'll just go back to the source list of objects,
 376                  * and ignore any that weren't present in our earlier
 377                  * traversal.
 378                  */
 379                 for_each_loose_object(mark_loose_unreachable_referents, NULL, 0);
 380                 for_each_packed_object(mark_packed_unreachable_referents, NULL, 0);
 381         }
 382
 383         /* Look up all the requirements, warn about missing objects.. */
 384         max = get_max_object_index();
 385         if (verbose)
 386                 fprintf_ln(stderr, _("Checking connectivity (%d objects)"), max);
 387
 388         for (i = 0; i < max; i++) {
 389                 struct object *obj = get_indexed_object(i);
 390
 391                 if (obj)
 392                         check_object(obj);
 393         }
 394 }
 395
 396 static int fsck_obj(struct object *obj, void *buffer, unsigned long size)
 397 {
 398         int err;
 399
 400         if (obj->flags & SEEN)
 401                 return 0;
 402         obj->flags |= SEEN;
 403
 404         if (verbose)
 405                 fprintf_ln(stderr, _("Checking %s %s"),
 406                            printable_type(&obj->oid, obj->type),
 407                            describe_object(&obj->oid));
 408
 409         if (fsck_walk(obj, NULL, &fsck_obj_options))
 410                 objerror(obj, _("broken links"));
 411         err = fsck_object(obj, buffer, size, &fsck_obj_options);
 412         if (err)
 413                 goto out;
 414
 415         if (obj->type == OBJ_COMMIT) {
 416                 struct commit *commit = (struct commit *) obj;
 417
 418                 if (!commit->parents && show_root)
 419                         printf_ln(_("root %s"),
 420                                   describe_object(&commit->object.oid));
 421         }
 422
 423         if (obj->type == OBJ_TAG) {
 424                 struct tag *tag = (struct tag *) obj;
 425
 426                 if (show_tags && tag->tagged) {
 427                         printf_ln(_("tagged %s %s (%s) in %s"),
 428                                   printable_type(&tag->tagged->oid, tag->tagged->type),
 429                                   describe_object(&tag->tagged->oid),
 430                                   tag->tag,
 431                                   describe_object(&tag->object.oid));
 432                 }
 433         }
 434
 435 out:
 436         if (obj->type == OBJ_TREE)
 437                 free_tree_buffer((struct tree *)obj);
 438         if (obj->type == OBJ_COMMIT)
 439                 free_commit_buffer(the_repository->parsed_objects,
 440                                    (struct commit *)obj);
 441         return err;
 442 }
 443
 444 static int fsck_obj_buffer(const struct object_id *oid, enum object_type type,
 445                            unsigned long size, void *buffer, int *eaten)
 446 {
 447         /*
 448          * Note, buffer may be NULL if type is OBJ_BLOB. See
 449          * verify_packfile(), data_valid variable for details.
 450          */
 451         struct object *obj;
 452         obj = parse_object_buffer(the_repository, oid, type, size, buffer,
 453                                   eaten);
 454         if (!obj) {
 455                 errors_found |= ERROR_OBJECT;
 456                 return error(_("%s: object corrupt or missing"),
 457                              oid_to_hex(oid));
 458         }
 459         obj->flags &= ~(REACHABLE | SEEN);
 460         obj->flags |= HAS_OBJ;
 461         return fsck_obj(obj, buffer, size);
 462 }
 463
 464 static int default_refs;
 465
 466 static void fsck_handle_reflog_oid(const char *refname, struct object_id *oid,
 467         timestamp_t timestamp)
 468 {
 469         struct object *obj;
 470
 471         if (!is_null_oid(oid)) {
 472                 obj = lookup_object(the_repository, oid);
 473                 if (obj && (obj->flags & HAS_OBJ)) {
 474                         if (timestamp)
 475                                 fsck_put_object_name(&fsck_walk_options, oid,
 476                                                      "%s@{%"PRItime"}",
 477                                                      refname, timestamp);
 478                         obj->flags |= USED;
 479                         mark_object_reachable(obj);
 480                 } else if (!is_promisor_object(oid)) {
 481                         error(_("%s: invalid reflog entry %s"),
 482                               refname, oid_to_hex(oid));
 483                         errors_found |= ERROR_REACHABLE;
 484                 }
 485         }
 486 }
 487
 488 static int fsck_handle_reflog_ent(struct object_id *ooid, struct object_id *noid,
 489                 const char *email, timestamp_t timestamp, int tz,
 490                 const char *message, void *cb_data)
 491 {
 492         const char *refname = cb_data;
 493
 494         if (verbose)
 495                 fprintf_ln(stderr, _("Checking reflog %s->%s"),
 496                            oid_to_hex(ooid), oid_to_hex(noid));
 497
 498         fsck_handle_reflog_oid(refname, ooid, 0);
 499         fsck_handle_reflog_oid(refname, noid, timestamp);
 500         return 0;
 501 }
 502
 503 static int fsck_handle_reflog(const char *logname, const struct object_id *oid,
 504                               int flag, void *cb_data)
 505 {
 506         struct strbuf refname = STRBUF_INIT;
 507
 508         strbuf_worktree_ref(cb_data, &refname, logname);
 509         for_each_reflog_ent(refname.buf, fsck_handle_reflog_ent, refname.buf);
 510         strbuf_release(&refname);
 511         return 0;
 512 }
 513
 514 static int fsck_handle_ref(const char *refname, const struct object_id *oid,
 515                            int flag, void *cb_data)
 516 {
 517         struct object *obj;
 518
 519         obj = parse_object(the_repository, oid);
 520         if (!obj) {
 521                 if (is_promisor_object(oid)) {
 522                         /*
 523                          * Increment default_refs anyway, because this is a
 524                          * valid ref.
 525                          */
 526                          default_refs++;
 527                          return 0;
 528                 }
 529                 error(_("%s: invalid sha1 pointer %s"),
 530                       refname, oid_to_hex(oid));
 531                 errors_found |= ERROR_REACHABLE;
 532                 /* We'll continue with the rest despite the error.. */
 533                 return 0;
 534         }
 535         if (obj->type != OBJ_COMMIT && is_branch(refname)) {
 536                 error(_("%s: not a commit"), refname);
 537                 errors_found |= ERROR_REFS;
 538         }
 539         default_refs++;
 540         obj->flags |= USED;
 541         fsck_put_object_name(&fsck_walk_options,
 542                              oid, "%s", refname);
 543         mark_object_reachable(obj);
 544
 545         return 0;
 546 }
 547
 548 static int fsck_head_link(const char *head_ref_name,
 549                           const char **head_points_at,
 550                           struct object_id *head_oid);
 551
 552 static void get_default_heads(void)
 553 {
 554         struct worktree **worktrees, **p;
 555         const char *head_points_at;
 556         struct object_id head_oid;
 557
 558         for_each_rawref(fsck_handle_ref, NULL);
 559
 560         worktrees = get_worktrees();
 561         for (p = worktrees; *p; p++) {
 562                 struct worktree *wt = *p;
 563                 struct strbuf ref = STRBUF_INIT;
 564
 565                 strbuf_worktree_ref(wt, &ref, "HEAD");
 566                 fsck_head_link(ref.buf, &head_points_at, &head_oid);
 567                 if (head_points_at && !is_null_oid(&head_oid))
 568                         fsck_handle_ref(ref.buf, &head_oid, 0, NULL);
 569                 strbuf_release(&ref);
 570
 571                 if (include_reflogs)
 572                         refs_for_each_reflog(get_worktree_ref_store(wt),
 573                                              fsck_handle_reflog, wt);
 574         }
 575         free_worktrees(worktrees);
 576
 577         /*
 578          * Not having any default heads isn't really fatal, but
 579          * it does mean that "--unreachable" no longer makes any
 580          * sense (since in this case everything will obviously
 581          * be unreachable by definition.
 582          *
 583          * Showing dangling objects is valid, though (as those
 584          * dangling objects are likely lost heads).
 585          *
 586          * So we just print a warning about it, and clear the
 587          * "show_unreachable" flag.
 588          */
 589         if (!default_refs) {
 590                 fprintf_ln(stderr, _("notice: No default references"));
 591                 show_unreachable = 0;
 592         }
 593 }
 594
 595 static int fsck_loose(const struct object_id *oid, const char *path, void *data)
 596 {
 597         struct object *obj;
 598         enum object_type type;
 599         unsigned long size;
 600         void *contents;
 601         int eaten;
 602
 603         if (read_loose_object(path, oid, &type, &size, &contents) < 0) {
 604                 errors_found |= ERROR_OBJECT;
 605                 error(_("%s: object corrupt or missing: %s"),
 606                       oid_to_hex(oid), path);
 607                 return 0; /* keep checking other objects */
 608         }
 609
 610         if (!contents && type != OBJ_BLOB)
 611                 BUG("read_loose_object streamed a non-blob");
 612
 613         obj = parse_object_buffer(the_repository, oid, type, size,
 614                                   contents, &eaten);
 615
 616         if (!obj) {
 617                 errors_found |= ERROR_OBJECT;
 618                 error(_("%s: object could not be parsed: %s"),
 619                       oid_to_hex(oid), path);
 620                 if (!eaten)
 621                         free(contents);
 622                 return 0; /* keep checking other objects */
 623         }
 624
 625         obj->flags &= ~(REACHABLE | SEEN);
 626         obj->flags |= HAS_OBJ;
 627         if (fsck_obj(obj, contents, size))
 628                 errors_found |= ERROR_OBJECT;
 629
 630         if (!eaten)
 631                 free(contents);
 632         return 0; /* keep checking other objects, even if we saw an error */
 633 }
 634
 635 static int fsck_cruft(const char *basename, const char *path, void *data)
 636 {
 637         if (!starts_with(basename, "tmp_obj_"))
 638                 fprintf_ln(stderr, _("bad sha1 file: %s"), path);
 639         return 0;
 640 }
 641
 642 static int fsck_subdir(unsigned int nr, const char *path, void *progress)
 643 {
 644         display_progress(progress, nr + 1);
 645         return 0;
 646 }
 647
 648 static void fsck_object_dir(const char *path)
 649 {
 650         struct progress *progress = NULL;
 651
 652         if (verbose)
 653                 fprintf_ln(stderr, _("Checking object directory"));
 654
 655         if (show_progress)
 656                 progress = start_progress(_("Checking object directories"), 256);
 657
 658         for_each_loose_file_in_objdir(path, fsck_loose, fsck_cruft, fsck_subdir,
 659                                       progress);
 660         display_progress(progress, 256);
 661         stop_progress(&progress);
 662 }
 663
 664 static int fsck_head_link(const char *head_ref_name,
 665                           const char **head_points_at,
 666                           struct object_id *head_oid)
 667 {
 668         int null_is_error = 0;
 669
 670         if (verbose)
 671                 fprintf_ln(stderr, _("Checking %s link"), head_ref_name);
 672
 673         *head_points_at = resolve_ref_unsafe(head_ref_name, 0, head_oid, NULL);
 674         if (!*head_points_at) {
 675                 errors_found |= ERROR_REFS;
 676                 return error(_("invalid %s"), head_ref_name);
 677         }
 678         if (!strcmp(*head_points_at, head_ref_name))
 679                 /* detached HEAD */
 680                 null_is_error = 1;
 681         else if (!starts_with(*head_points_at, "refs/heads/")) {
 682                 errors_found |= ERROR_REFS;
 683                 return error(_("%s points to something strange (%s)"),
 684                              head_ref_name, *head_points_at);
 685         }
 686         if (is_null_oid(head_oid)) {
 687                 if (null_is_error) {
 688                         errors_found |= ERROR_REFS;
 689                         return error(_("%s: detached HEAD points at nothing"),
 690                                      head_ref_name);
 691                 }
 692                 fprintf_ln(stderr,
 693                            _("notice: %s points to an unborn branch (%s)"),
 694                            head_ref_name, *head_points_at + 11);
 695         }
 696         return 0;
 697 }
 698
 699 static int fsck_cache_tree(struct cache_tree *it)
 700 {
 701         int i;
 702         int err = 0;
 703
 704         if (verbose)
 705                 fprintf_ln(stderr, _("Checking cache tree"));
 706
 707         if (0 <= it->entry_count) {
 708                 struct object *obj = parse_object(the_repository, &it->oid);
 709                 if (!obj) {
 710                         error(_("%s: invalid sha1 pointer in cache-tree"),
 711                               oid_to_hex(&it->oid));
 712                         errors_found |= ERROR_REFS;
 713                         return 1;
 714                 }
 715                 obj->flags |= USED;
 716                 fsck_put_object_name(&fsck_walk_options, &it->oid, ":");
 717                 mark_object_reachable(obj);
 718                 if (obj->type != OBJ_TREE)
 719                         err |= objerror(obj, _("non-tree in cache-tree"));
 720         }
 721         for (i = 0; i < it->subtree_nr; i++)
 722                 err |= fsck_cache_tree(it->down[i]->cache_tree);
 723         return err;
 724 }
 725
 726 static void mark_object_for_connectivity(const struct object_id *oid)
 727 {
 728         struct object *obj = lookup_unknown_object(oid);
 729         obj->flags |= HAS_OBJ;
 730 }
 731
 732 static int mark_loose_for_connectivity(const struct object_id *oid,
 733                                        const char *path,
 734                                        void *data)
 735 {
 736         mark_object_for_connectivity(oid);
 737         return 0;
 738 }
 739
 740 static int mark_packed_for_connectivity(const struct object_id *oid,
 741                                         struct packed_git *pack,
 742                                         uint32_t pos,
 743                                         void *data)
 744 {
 745         mark_object_for_connectivity(oid);
 746         return 0;
 747 }
 748
 749 static char const * const fsck_usage[] = {
 750         N_("git fsck [<options>] [<object>...]"),
 751         NULL
 752 };
 753
 754 static struct option fsck_opts[] = {
 755         OPT__VERBOSE(&verbose, N_("be verbose")),
 756         OPT_BOOL(0, "unreachable", &show_unreachable, N_("show unreachable objects")),
 757         OPT_BOOL(0, "dangling", &show_dangling, N_("show dangling objects")),
 758         OPT_BOOL(0, "tags", &show_tags, N_("report tags")),
 759         OPT_BOOL(0, "root", &show_root, N_("report root nodes")),
 760         OPT_BOOL(0, "cache", &keep_cache_objects, N_("make index objects head nodes")),
 761         OPT_BOOL(0, "reflogs", &include_reflogs, N_("make reflogs head nodes (default)")),
 762         OPT_BOOL(0, "full", &check_full, N_("also consider packs and alternate objects")),
 763         OPT_BOOL(0, "connectivity-only", &connectivity_only, N_("check only connectivity")),
 764         OPT_BOOL(0, "strict", &check_strict, N_("enable more strict checking")),
 765         OPT_BOOL(0, "lost-found", &write_lost_and_found,
 766                                 N_("write dangling objects in .git/lost-found")),
 767         OPT_BOOL(0, "progress", &show_progress, N_("show progress")),
 768         OPT_BOOL(0, "name-objects", &name_objects, N_("show verbose names for reachable objects")),
 769         OPT_END(),
 770 };
 771
 772 int cmd_fsck(int argc, const char **argv, const char *prefix)
 773 {
 774         int i;
 775         struct object_directory *odb;
 776
 777         /* fsck knows how to handle missing promisor objects */
 778         fetch_if_missing = 0;
 779
 780         errors_found = 0;
 781         read_replace_refs = 0;
 782
 783         argc = parse_options(argc, argv, prefix, fsck_opts, fsck_usage, 0);
 784
 785         fsck_walk_options.walk = mark_object;
 786         fsck_obj_options.walk = mark_used;
 787         fsck_obj_options.error_func = fsck_error_func;
 788         if (check_strict)
 789                 fsck_obj_options.strict = 1;
 790
 791         if (show_progress == -1)
 792                 show_progress = isatty(2);
 793         if (verbose)
 794                 show_progress = 0;
 795
 796         if (write_lost_and_found) {
 797                 check_full = 1;
 798                 include_reflogs = 0;
 799         }
 800
 801         if (name_objects)
 802                 fsck_enable_object_names(&fsck_walk_options);
 803
 804         git_config(git_fsck_config, &fsck_obj_options);
 805
 806         if (connectivity_only) {
 807                 for_each_loose_object(mark_loose_for_connectivity, NULL, 0);
 808                 for_each_packed_object(mark_packed_for_connectivity, NULL, 0);
 809         } else {
 810                 prepare_alt_odb(the_repository);
 811                 for (odb = the_repository->objects->odb; odb; odb = odb->next)
 812                         fsck_object_dir(odb->path);
 813
 814                 if (check_full) {
 815                         struct packed_git *p;
 816                         uint32_t total = 0, count = 0;
 817                         struct progress *progress = NULL;
 818
 819                         if (show_progress) {
 820                                 for (p = get_all_packs(the_repository); p;
 821                                      p = p->next) {
 822                                         if (open_pack_index(p))
 823                                                 continue;
 824                                         total += p->num_objects;
 825                                 }
 826
 827                                 progress = start_progress(_("Checking objects"), total);
 828                         }
 829                         for (p = get_all_packs(the_repository); p;
 830                              p = p->next) {
 831                                 /* verify gives error messages itself */
 832                                 if (verify_pack(the_repository,
 833                                                 p, fsck_obj_buffer,
 834                                                 progress, count))
 835                                         errors_found |= ERROR_PACK;
 836                                 count += p->num_objects;
 837                         }
 838                         stop_progress(&progress);
 839                 }
 840
 841                 if (fsck_finish(&fsck_obj_options))
 842                         errors_found |= ERROR_OBJECT;
 843         }
 844
 845         for (i = 0; i < argc; i++) {
 846                 const char *arg = argv[i];
 847                 struct object_id oid;
 848                 if (!get_oid(arg, &oid)) {
 849                         struct object *obj = lookup_object(the_repository,
 850                                                            &oid);
 851
 852                         if (!obj || !(obj->flags & HAS_OBJ)) {
 853                                 if (is_promisor_object(&oid))
 854                                         continue;
 855                                 error(_("%s: object missing"), oid_to_hex(&oid));
 856                                 errors_found |= ERROR_OBJECT;
 857                                 continue;
 858                         }
 859
 860                         obj->flags |= USED;
 861                         fsck_put_object_name(&fsck_walk_options, &oid,
 862                                              "%s", arg);
 863                         mark_object_reachable(obj);
 864                         continue;
 865                 }
 866                 error(_("invalid parameter: expected sha1, got '%s'"), arg);
 867                 errors_found |= ERROR_OBJECT;
 868         }
 869
 870         /*
 871          * If we've not been given any explicit head information, do the
 872          * default ones from .git/refs. We also consider the index file
 873          * in this case (ie this implies --cache).
 874          */
 875         if (!argc) {
 876                 get_default_heads();
 877                 keep_cache_objects = 1;
 878         }
 879
 880         if (keep_cache_objects) {
 881                 verify_index_checksum = 1;
 882                 verify_ce_order = 1;
 883                 read_cache();
 884                 for (i = 0; i < active_nr; i++) {
 885                         unsigned int mode;
 886                         struct blob *blob;
 887                         struct object *obj;
 888
 889                         mode = active_cache[i]->ce_mode;
 890                         if (S_ISGITLINK(mode))
 891                                 continue;
 892                         blob = lookup_blob(the_repository,
 893                                            &active_cache[i]->oid);
 894                         if (!blob)
 895                                 continue;
 896                         obj = &blob->object;
 897                         obj->flags |= USED;
 898                         fsck_put_object_name(&fsck_walk_options, &obj->oid,
 899                                              ":%s", active_cache[i]->name);
 900                         mark_object_reachable(obj);
 901                 }
 902                 if (active_cache_tree)
 903                         fsck_cache_tree(active_cache_tree);
 904         }
 905
 906         check_connectivity();
 907
 908         if (!git_config_get_bool("core.commitgraph", &i) && i) {
 909                 struct child_process commit_graph_verify = CHILD_PROCESS_INIT;
 910                 const char *verify_argv[] = { "commit-graph", "verify", NULL, NULL, NULL };
 911
 912                 prepare_alt_odb(the_repository);
 913                 for (odb = the_repository->objects->odb; odb; odb = odb->next) {
 914                         child_process_init(&commit_graph_verify);
 915                         commit_graph_verify.argv = verify_argv;
 916                         commit_graph_verify.git_cmd = 1;
 917                         verify_argv[2] = "--object-dir";
 918                         verify_argv[3] = odb->path;
 919                         if (run_command(&commit_graph_verify))
 920                                 errors_found |= ERROR_COMMIT_GRAPH;
 921                 }
 922         }
 923
 924         if (!git_config_get_bool("core.multipackindex", &i) && i) {
 925                 struct child_process midx_verify = CHILD_PROCESS_INIT;
 926                 const char *midx_argv[] = { "multi-pack-index", "verify", NULL, NULL, NULL };
 927
 928                 prepare_alt_odb(the_repository);
 929                 for (odb = the_repository->objects->odb; odb; odb = odb->next) {
 930                         child_process_init(&midx_verify);
 931                         midx_verify.argv = midx_argv;
 932                         midx_verify.git_cmd = 1;
 933                         midx_argv[2] = "--object-dir";
 934                         midx_argv[3] = odb->path;
 935                         if (run_command(&midx_verify))
 936                                 errors_found |= ERROR_MULTI_PACK_INDEX;
 937                 }
 938         }
 939
 940         return errors_found;
 941 }