object.c

   1 #include "git-compat-util.h"
   2 #include "gettext.h"
   3 #include "hex.h"
   4 #include "object.h"
   5 #include "replace-object.h"
   6 #include "object-file.h"
   7 #include "object-store.h"
   8 #include "blob.h"
   9 #include "statinfo.h"
  10 #include "tree.h"
  11 #include "commit.h"
  12 #include "tag.h"
  13 #include "alloc.h"
  14 #include "packfile.h"
  15 #include "commit-graph.h"
  16 #include "loose.h"
  17
  18 unsigned int get_max_object_index(void)
  19 {
  20         return the_repository->parsed_objects->obj_hash_size;
  21 }
  22
  23 struct object *get_indexed_object(unsigned int idx)
  24 {
  25         return the_repository->parsed_objects->obj_hash[idx];
  26 }
  27
  28 static const char *object_type_strings[] = {
  29         NULL,           /* OBJ_NONE = 0 */
  30         "commit",       /* OBJ_COMMIT = 1 */
  31         "tree",         /* OBJ_TREE = 2 */
  32         "blob",         /* OBJ_BLOB = 3 */
  33         "tag",          /* OBJ_TAG = 4 */
  34 };
  35
  36 const char *type_name(unsigned int type)
  37 {
  38         if (type >= ARRAY_SIZE(object_type_strings))
  39                 return NULL;
  40         return object_type_strings[type];
  41 }
  42
  43 int type_from_string_gently(const char *str, ssize_t len, int gentle)
  44 {
  45         int i;
  46
  47         if (len < 0)
  48                 len = strlen(str);
  49
  50         for (i = 1; i < ARRAY_SIZE(object_type_strings); i++)
  51                 if (!xstrncmpz(object_type_strings[i], str, len))
  52                         return i;
  53
  54         if (gentle)
  55                 return -1;
  56
  57         die(_("invalid object type \"%s\""), str);
  58 }
  59
  60 /*
  61  * Return a numerical hash value between 0 and n-1 for the object with
  62  * the specified sha1.  n must be a power of 2.  Please note that the
  63  * return value is *not* consistent across computer architectures.
  64  */
  65 static unsigned int hash_obj(const struct object_id *oid, unsigned int n)
  66 {
  67         return oidhash(oid) & (n - 1);
  68 }
  69
  70 /*
  71  * Insert obj into the hash table hash, which has length size (which
  72  * must be a power of 2).  On collisions, simply overflow to the next
  73  * empty bucket.
  74  */
  75 static void insert_obj_hash(struct object *obj, struct object **hash, unsigned int size)
  76 {
  77         unsigned int j = hash_obj(&obj->oid, size);
  78
  79         while (hash[j]) {
  80                 j++;
  81                 if (j >= size)
  82                         j = 0;
  83         }
  84         hash[j] = obj;
  85 }
  86
  87 /*
  88  * Look up the record for the given sha1 in the hash map stored in
  89  * obj_hash.  Return NULL if it was not found.
  90  */
  91 struct object *lookup_object(struct repository *r, const struct object_id *oid)
  92 {
  93         unsigned int i, first;
  94         struct object *obj;
  95
  96         if (!r->parsed_objects->obj_hash)
  97                 return NULL;
  98
  99         first = i = hash_obj(oid, r->parsed_objects->obj_hash_size);
 100         while ((obj = r->parsed_objects->obj_hash[i]) != NULL) {
 101                 if (oideq(oid, &obj->oid))
 102                         break;
 103                 i++;
 104                 if (i == r->parsed_objects->obj_hash_size)
 105                         i = 0;
 106         }
 107         if (obj && i != first) {
 108                 /*
 109                  * Move object to where we started to look for it so
 110                  * that we do not need to walk the hash table the next
 111                  * time we look for it.
 112                  */
 113                 SWAP(r->parsed_objects->obj_hash[i],
 114                      r->parsed_objects->obj_hash[first]);
 115         }
 116         return obj;
 117 }
 118
 119 /*
 120  * Increase the size of the hash map stored in obj_hash to the next
 121  * power of 2 (but at least 32).  Copy the existing values to the new
 122  * hash map.
 123  */
 124 static void grow_object_hash(struct repository *r)
 125 {
 126         int i;
 127         /*
 128          * Note that this size must always be power-of-2 to match hash_obj
 129          * above.
 130          */
 131         int new_hash_size = r->parsed_objects->obj_hash_size < 32 ? 32 : 2 * r->parsed_objects->obj_hash_size;
 132         struct object **new_hash;
 133
 134         CALLOC_ARRAY(new_hash, new_hash_size);
 135         for (i = 0; i < r->parsed_objects->obj_hash_size; i++) {
 136                 struct object *obj = r->parsed_objects->obj_hash[i];
 137
 138                 if (!obj)
 139                         continue;
 140                 insert_obj_hash(obj, new_hash, new_hash_size);
 141         }
 142         free(r->parsed_objects->obj_hash);
 143         r->parsed_objects->obj_hash = new_hash;
 144         r->parsed_objects->obj_hash_size = new_hash_size;
 145 }
 146
 147 void *create_object(struct repository *r, const struct object_id *oid, void *o)
 148 {
 149         struct object *obj = o;
 150
 151         obj->parsed = 0;
 152         obj->flags = 0;
 153         oidcpy(&obj->oid, oid);
 154
 155         if (r->parsed_objects->obj_hash_size - 1 <= r->parsed_objects->nr_objs * 2)
 156                 grow_object_hash(r);
 157
 158         insert_obj_hash(obj, r->parsed_objects->obj_hash,
 159                         r->parsed_objects->obj_hash_size);
 160         r->parsed_objects->nr_objs++;
 161         return obj;
 162 }
 163
 164 void *object_as_type(struct object *obj, enum object_type type, int quiet)
 165 {
 166         if (obj->type == type)
 167                 return obj;
 168         else if (obj->type == OBJ_NONE) {
 169                 if (type == OBJ_COMMIT)
 170                         init_commit_node((struct commit *) obj);
 171                 else
 172                         obj->type = type;
 173                 return obj;
 174         }
 175         else {
 176                 if (!quiet)
 177                         error(_("object %s is a %s, not a %s"),
 178                               oid_to_hex(&obj->oid),
 179                               type_name(obj->type), type_name(type));
 180                 return NULL;
 181         }
 182 }
 183
 184 struct object *lookup_unknown_object(struct repository *r, const struct object_id *oid)
 185 {
 186         struct object *obj = lookup_object(r, oid);
 187         if (!obj)
 188                 obj = create_object(r, oid, alloc_object_node(r));
 189         return obj;
 190 }
 191
 192 struct object *lookup_object_by_type(struct repository *r,
 193                             const struct object_id *oid,
 194                             enum object_type type)
 195 {
 196         switch (type) {
 197         case OBJ_COMMIT:
 198                 return (struct object *)lookup_commit(r, oid);
 199         case OBJ_TREE:
 200                 return (struct object *)lookup_tree(r, oid);
 201         case OBJ_TAG:
 202                 return (struct object *)lookup_tag(r, oid);
 203         case OBJ_BLOB:
 204                 return (struct object *)lookup_blob(r, oid);
 205         default:
 206                 BUG("unknown object type %d", type);
 207         }
 208 }
 209
 210 enum peel_status peel_object(struct repository *r,
 211                              const struct object_id *name,
 212                              struct object_id *oid)
 213 {
 214         struct object *o = lookup_unknown_object(r, name);
 215
 216         if (o->type == OBJ_NONE) {
 217                 int type = oid_object_info(r, name, NULL);
 218                 if (type < 0 || !object_as_type(o, type, 0))
 219                         return PEEL_INVALID;
 220         }
 221
 222         if (o->type != OBJ_TAG)
 223                 return PEEL_NON_TAG;
 224
 225         o = deref_tag_noverify(r, o);
 226         if (!o)
 227                 return PEEL_INVALID;
 228
 229         oidcpy(oid, &o->oid);
 230         return PEEL_PEELED;
 231 }
 232
 233 struct object *parse_object_buffer(struct repository *r, const struct object_id *oid, enum object_type type, unsigned long size, void *buffer, int *eaten_p)
 234 {
 235         struct object *obj;
 236         *eaten_p = 0;
 237
 238         obj = NULL;
 239         if (type == OBJ_BLOB) {
 240                 struct blob *blob = lookup_blob(r, oid);
 241                 if (blob) {
 242                         parse_blob_buffer(blob);
 243                         obj = &blob->object;
 244                 }
 245         } else if (type == OBJ_TREE) {
 246                 struct tree *tree = lookup_tree(r, oid);
 247                 if (tree) {
 248                         obj = &tree->object;
 249                         if (!tree->buffer)
 250                                 tree->object.parsed = 0;
 251                         if (!tree->object.parsed) {
 252                                 if (parse_tree_buffer(tree, buffer, size))
 253                                         return NULL;
 254                                 *eaten_p = 1;
 255                         }
 256                 }
 257         } else if (type == OBJ_COMMIT) {
 258                 struct commit *commit = lookup_commit(r, oid);
 259                 if (commit) {
 260                         if (parse_commit_buffer(r, commit, buffer, size, 1))
 261                                 return NULL;
 262                         if (save_commit_buffer &&
 263                             !get_cached_commit_buffer(r, commit, NULL)) {
 264                                 set_commit_buffer(r, commit, buffer, size);
 265                                 *eaten_p = 1;
 266                         }
 267                         obj = &commit->object;
 268                 }
 269         } else if (type == OBJ_TAG) {
 270                 struct tag *tag = lookup_tag(r, oid);
 271                 if (tag) {
 272                         if (parse_tag_buffer(r, tag, buffer, size))
 273                                return NULL;
 274                         obj = &tag->object;
 275                 }
 276         } else {
 277                 warning(_("object %s has unknown type id %d"), oid_to_hex(oid), type);
 278                 obj = NULL;
 279         }
 280         return obj;
 281 }
 282
 283 struct object *parse_object_or_die(const struct object_id *oid,
 284                                    const char *name)
 285 {
 286         struct object *o = parse_object(the_repository, oid);
 287         if (o)
 288                 return o;
 289
 290         die(_("unable to parse object: %s"), name ? name : oid_to_hex(oid));
 291 }
 292
 293 struct object *parse_object_with_flags(struct repository *r,
 294                                        const struct object_id *oid,
 295                                        enum parse_object_flags flags)
 296 {
 297         int skip_hash = !!(flags & PARSE_OBJECT_SKIP_HASH_CHECK);
 298         int discard_tree = !!(flags & PARSE_OBJECT_DISCARD_TREE);
 299         unsigned long size;
 300         enum object_type type;
 301         int eaten;
 302         const struct object_id *repl = lookup_replace_object(r, oid);
 303         void *buffer;
 304         struct object *obj;
 305
 306         obj = lookup_object(r, oid);
 307         if (obj && obj->parsed)
 308                 return obj;
 309
 310         if (skip_hash) {
 311                 struct commit *commit = lookup_commit_in_graph(r, repl);
 312                 if (commit)
 313                         return &commit->object;
 314         }
 315
 316         if ((!obj || obj->type == OBJ_BLOB) &&
 317             oid_object_info(r, oid, NULL) == OBJ_BLOB) {
 318                 if (!skip_hash && stream_object_signature(r, repl) < 0) {
 319                         error(_("hash mismatch %s"), oid_to_hex(oid));
 320                         return NULL;
 321                 }
 322                 parse_blob_buffer(lookup_blob(r, oid));
 323                 return lookup_object(r, oid);
 324         }
 325
 326         /*
 327          * If the caller does not care about the tree buffer and does not
 328          * care about checking the hash, we can simply verify that we
 329          * have the on-disk object with the correct type.
 330          */
 331         if (skip_hash && discard_tree &&
 332             (!obj || obj->type == OBJ_TREE) &&
 333             oid_object_info(r, oid, NULL) == OBJ_TREE) {
 334                 return &lookup_tree(r, oid)->object;
 335         }
 336
 337         buffer = repo_read_object_file(r, oid, &type, &size);
 338         if (buffer) {
 339                 if (!skip_hash &&
 340                     check_object_signature(r, repl, buffer, size, type) < 0) {
 341                         free(buffer);
 342                         error(_("hash mismatch %s"), oid_to_hex(repl));
 343                         return NULL;
 344                 }
 345
 346                 obj = parse_object_buffer(r, oid, type, size,
 347                                           buffer, &eaten);
 348                 if (!eaten)
 349                         free(buffer);
 350                 if (discard_tree && type == OBJ_TREE)
 351                         free_tree_buffer((struct tree *)obj);
 352                 return obj;
 353         }
 354         return NULL;
 355 }
 356
 357 struct object *parse_object(struct repository *r, const struct object_id *oid)
 358 {
 359         return parse_object_with_flags(r, oid, 0);
 360 }
 361
 362 struct object_list *object_list_insert(struct object *item,
 363                                        struct object_list **list_p)
 364 {
 365         struct object_list *new_list = xmalloc(sizeof(struct object_list));
 366         new_list->item = item;
 367         new_list->next = *list_p;
 368         *list_p = new_list;
 369         return new_list;
 370 }
 371
 372 int object_list_contains(struct object_list *list, struct object *obj)
 373 {
 374         while (list) {
 375                 if (list->item == obj)
 376                         return 1;
 377                 list = list->next;
 378         }
 379         return 0;
 380 }
 381
 382 void object_list_free(struct object_list **list)
 383 {
 384         while (*list) {
 385                 struct object_list *p = *list;
 386                 *list = p->next;
 387                 free(p);
 388         }
 389 }
 390
 391 /*
 392  * A zero-length string to which object_array_entry::name can be
 393  * initialized without requiring a malloc/free.
 394  */
 395 static char object_array_slopbuf[1];
 396
 397 void object_array_init(struct object_array *array)
 398 {
 399         struct object_array blank = OBJECT_ARRAY_INIT;
 400         memcpy(array, &blank, sizeof(*array));
 401 }
 402
 403 void add_object_array_with_path(struct object *obj, const char *name,
 404                                 struct object_array *array,
 405                                 unsigned mode, const char *path)
 406 {
 407         unsigned nr = array->nr;
 408         unsigned alloc = array->alloc;
 409         struct object_array_entry *objects = array->objects;
 410         struct object_array_entry *entry;
 411
 412         if (nr >= alloc) {
 413                 alloc = (alloc + 32) * 2;
 414                 REALLOC_ARRAY(objects, alloc);
 415                 array->alloc = alloc;
 416                 array->objects = objects;
 417         }
 418         entry = &objects[nr];
 419         entry->item = obj;
 420         if (!name)
 421                 entry->name = NULL;
 422         else if (!*name)
 423                 /* Use our own empty string instead of allocating one: */
 424                 entry->name = object_array_slopbuf;
 425         else
 426                 entry->name = xstrdup(name);
 427         entry->mode = mode;
 428         if (path)
 429                 entry->path = xstrdup(path);
 430         else
 431                 entry->path = NULL;
 432         array->nr = ++nr;
 433 }
 434
 435 void add_object_array(struct object *obj, const char *name, struct object_array *array)
 436 {
 437         add_object_array_with_path(obj, name, array, S_IFINVALID, NULL);
 438 }
 439
 440 /*
 441  * Free all memory associated with an entry; the result is
 442  * in an unspecified state and should not be examined.
 443  */
 444 static void object_array_release_entry(struct object_array_entry *ent)
 445 {
 446         if (ent->name != object_array_slopbuf)
 447                 free(ent->name);
 448         free(ent->path);
 449 }
 450
 451 struct object *object_array_pop(struct object_array *array)
 452 {
 453         struct object *ret;
 454
 455         if (!array->nr)
 456                 return NULL;
 457
 458         ret = array->objects[array->nr - 1].item;
 459         object_array_release_entry(&array->objects[array->nr - 1]);
 460         array->nr--;
 461         return ret;
 462 }
 463
 464 void object_array_filter(struct object_array *array,
 465                          object_array_each_func_t want, void *cb_data)
 466 {
 467         unsigned nr = array->nr, src, dst;
 468         struct object_array_entry *objects = array->objects;
 469
 470         for (src = dst = 0; src < nr; src++) {
 471                 if (want(&objects[src], cb_data)) {
 472                         if (src != dst)
 473                                 objects[dst] = objects[src];
 474                         dst++;
 475                 } else {
 476                         object_array_release_entry(&objects[src]);
 477                 }
 478         }
 479         array->nr = dst;
 480 }
 481
 482 void object_array_clear(struct object_array *array)
 483 {
 484         int i;
 485         for (i = 0; i < array->nr; i++)
 486                 object_array_release_entry(&array->objects[i]);
 487         FREE_AND_NULL(array->objects);
 488         array->nr = array->alloc = 0;
 489 }
 490
 491 /*
 492  * Return true if array already contains an entry.
 493  */
 494 static int contains_object(struct object_array *array,
 495                            const struct object *item, const char *name)
 496 {
 497         unsigned nr = array->nr, i;
 498         struct object_array_entry *object = array->objects;
 499
 500         for (i = 0; i < nr; i++, object++)
 501                 if (item == object->item && !strcmp(object->name, name))
 502                         return 1;
 503         return 0;
 504 }
 505
 506 void object_array_remove_duplicates(struct object_array *array)
 507 {
 508         unsigned nr = array->nr, src;
 509         struct object_array_entry *objects = array->objects;
 510
 511         array->nr = 0;
 512         for (src = 0; src < nr; src++) {
 513                 if (!contains_object(array, objects[src].item,
 514                                      objects[src].name)) {
 515                         if (src != array->nr)
 516                                 objects[array->nr] = objects[src];
 517                         array->nr++;
 518                 } else {
 519                         object_array_release_entry(&objects[src]);
 520                 }
 521         }
 522 }
 523
 524 void clear_object_flags(unsigned flags)
 525 {
 526         int i;
 527
 528         for (i=0; i < the_repository->parsed_objects->obj_hash_size; i++) {
 529                 struct object *obj = the_repository->parsed_objects->obj_hash[i];
 530                 if (obj)
 531                         obj->flags &= ~flags;
 532         }
 533 }
 534
 535 void repo_clear_commit_marks(struct repository *r, unsigned int flags)
 536 {
 537         int i;
 538
 539         for (i = 0; i < r->parsed_objects->obj_hash_size; i++) {
 540                 struct object *obj = r->parsed_objects->obj_hash[i];
 541                 if (obj && obj->type == OBJ_COMMIT)
 542                         obj->flags &= ~flags;
 543         }
 544 }
 545
 546 struct parsed_object_pool *parsed_object_pool_new(void)
 547 {
 548         struct parsed_object_pool *o = xmalloc(sizeof(*o));
 549         memset(o, 0, sizeof(*o));
 550
 551         o->blob_state = allocate_alloc_state();
 552         o->tree_state = allocate_alloc_state();
 553         o->commit_state = allocate_alloc_state();
 554         o->tag_state = allocate_alloc_state();
 555         o->object_state = allocate_alloc_state();
 556
 557         o->is_shallow = -1;
 558         CALLOC_ARRAY(o->shallow_stat, 1);
 559
 560         o->buffer_slab = allocate_commit_buffer_slab();
 561
 562         return o;
 563 }
 564
 565 struct raw_object_store *raw_object_store_new(void)
 566 {
 567         struct raw_object_store *o = xmalloc(sizeof(*o));
 568
 569         memset(o, 0, sizeof(*o));
 570         INIT_LIST_HEAD(&o->packed_git_mru);
 571         hashmap_init(&o->pack_map, pack_map_entry_cmp, NULL, 0);
 572         pthread_mutex_init(&o->replace_mutex, NULL);
 573         return o;
 574 }
 575
 576 void free_object_directory(struct object_directory *odb)
 577 {
 578         free(odb->path);
 579         odb_clear_loose_cache(odb);
 580         loose_object_map_clear(&odb->loose_map);
 581         free(odb);
 582 }
 583
 584 static void free_object_directories(struct raw_object_store *o)
 585 {
 586         while (o->odb) {
 587                 struct object_directory *next;
 588
 589                 next = o->odb->next;
 590                 free_object_directory(o->odb);
 591                 o->odb = next;
 592         }
 593         kh_destroy_odb_path_map(o->odb_by_path);
 594         o->odb_by_path = NULL;
 595 }
 596
 597 void raw_object_store_clear(struct raw_object_store *o)
 598 {
 599         FREE_AND_NULL(o->alternate_db);
 600
 601         oidmap_free(o->replace_map, 1);
 602         FREE_AND_NULL(o->replace_map);
 603         pthread_mutex_destroy(&o->replace_mutex);
 604
 605         free_commit_graph(o->commit_graph);
 606         o->commit_graph = NULL;
 607         o->commit_graph_attempted = 0;
 608
 609         free_object_directories(o);
 610         o->odb_tail = NULL;
 611         o->loaded_alternates = 0;
 612
 613         INIT_LIST_HEAD(&o->packed_git_mru);
 614         close_object_store(o);
 615         o->packed_git = NULL;
 616
 617         hashmap_clear(&o->pack_map);
 618 }
 619
 620 void parsed_object_pool_clear(struct parsed_object_pool *o)
 621 {
 622         /*
 623          * As objects are allocated in slabs (see alloc.c), we do
 624          * not need to free each object, but each slab instead.
 625          *
 626          * Before doing so, we need to free any additional memory
 627          * the objects may hold.
 628          */
 629         unsigned i;
 630
 631         for (i = 0; i < o->obj_hash_size; i++) {
 632                 struct object *obj = o->obj_hash[i];
 633
 634                 if (!obj)
 635                         continue;
 636
 637                 if (obj->type == OBJ_TREE)
 638                         free_tree_buffer((struct tree*)obj);
 639                 else if (obj->type == OBJ_COMMIT)
 640                         release_commit_memory(o, (struct commit*)obj);
 641                 else if (obj->type == OBJ_TAG)
 642                         release_tag_memory((struct tag*)obj);
 643         }
 644
 645         FREE_AND_NULL(o->obj_hash);
 646         o->obj_hash_size = 0;
 647
 648         free_commit_buffer_slab(o->buffer_slab);
 649         o->buffer_slab = NULL;
 650
 651         clear_alloc_state(o->blob_state);
 652         clear_alloc_state(o->tree_state);
 653         clear_alloc_state(o->commit_state);
 654         clear_alloc_state(o->tag_state);
 655         clear_alloc_state(o->object_state);
 656         stat_validity_clear(o->shallow_stat);
 657         FREE_AND_NULL(o->blob_state);
 658         FREE_AND_NULL(o->tree_state);
 659         FREE_AND_NULL(o->commit_state);
 660         FREE_AND_NULL(o->tag_state);
 661         FREE_AND_NULL(o->object_state);
 662         FREE_AND_NULL(o->shallow_stat);
 663 }