object-store.h

   1 #ifndef OBJECT_STORE_H
   2 #define OBJECT_STORE_H
   3
   4 #include "cache.h"
   5 #include "oidmap.h"
   6 #include "list.h"
   7 #include "sha1-array.h"
   8 #include "strbuf.h"
   9
  10 struct object_directory {
  11         struct object_directory *next;
  12
  13         /*
  14          * Used to store the results of readdir(3) calls when we are OK
  15          * sacrificing accuracy due to races for speed. That includes
  16          * object existence with OBJECT_INFO_QUICK, as well as
  17          * our search for unique abbreviated hashes. Don't use it for tasks
  18          * requiring greater accuracy!
  19          *
  20          * Be sure to call odb_load_loose_cache() before using.
  21          */
  22         char loose_objects_subdir_seen[256];
  23         struct oid_array loose_objects_cache[256];
  24
  25         /*
  26          * Path to the alternative object store. If this is a relative path,
  27          * it is relative to the current working directory.
  28          */
  29         char *path;
  30 };
  31
  32 void prepare_alt_odb(struct repository *r);
  33 char *compute_alternate_path(const char *path, struct strbuf *err);
  34 typedef int alt_odb_fn(struct object_directory *, void *);
  35 int foreach_alt_odb(alt_odb_fn, void*);
  36 typedef void alternate_ref_fn(const struct object_id *oid, void *);
  37 void for_each_alternate_ref(alternate_ref_fn, void *);
  38
  39 /*
  40  * Add the directory to the on-disk alternates file; the new entry will also
  41  * take effect in the current process.
  42  */
  43 void add_to_alternates_file(const char *dir);
  44
  45 /*
  46  * Add the directory to the in-memory list of alternates (along with any
  47  * recursive alternates it points to), but do not modify the on-disk alternates
  48  * file.
  49  */
  50 void add_to_alternates_memory(const char *dir);
  51
  52 /*
  53  * Populate and return the loose object cache array corresponding to the
  54  * given object ID.
  55  */
  56 struct oid_array *odb_loose_cache(struct object_directory *odb,
  57                                   const struct object_id *oid);
  58
  59 /* Empty the loose object cache for the specified object directory. */
  60 void odb_clear_loose_cache(struct object_directory *odb);
  61
  62 struct packed_git {
  63         struct packed_git *next;
  64         struct list_head mru;
  65         struct pack_window *windows;
  66         off_t pack_size;
  67         const void *index_data;
  68         size_t index_size;
  69         uint32_t num_objects;
  70         uint32_t num_bad_objects;
  71         unsigned char *bad_object_sha1;
  72         int index_version;
  73         time_t mtime;
  74         int pack_fd;
  75         int index;              /* for builtin/pack-objects.c */
  76         unsigned pack_local:1,
  77                  pack_keep:1,
  78                  pack_keep_in_core:1,
  79                  freshened:1,
  80                  do_not_close:1,
  81                  pack_promisor:1,
  82                  multi_pack_index:1;
  83         unsigned char hash[GIT_MAX_RAWSZ];
  84         struct revindex_entry *revindex;
  85         /* something like ".git/objects/pack/xxxxx.pack" */
  86         char pack_name[FLEX_ARRAY]; /* more */
  87 };
  88
  89 struct multi_pack_index;
  90
  91 struct raw_object_store {
  92         /*
  93          * Set of all object directories; the main directory is first (and
  94          * cannot be NULL after initialization). Subsequent directories are
  95          * alternates.
  96          */
  97         struct object_directory *odb;
  98         struct object_directory **odb_tail;
  99         int loaded_alternates;
 100
 101         /*
 102          * A list of alternate object directories loaded from the environment;
 103          * this should not generally need to be accessed directly, but will
 104          * populate the "odb" list when prepare_alt_odb() is run.
 105          */
 106         char *alternate_db;
 107
 108         /*
 109          * Objects that should be substituted by other objects
 110          * (see git-replace(1)).
 111          */
 112         struct oidmap *replace_map;
 113
 114         struct commit_graph *commit_graph;
 115         unsigned commit_graph_attempted : 1; /* if loading has been attempted */
 116
 117         /*
 118          * private data
 119          *
 120          * should only be accessed directly by packfile.c and midx.c
 121          */
 122         struct multi_pack_index *multi_pack_index;
 123
 124         /*
 125          * private data
 126          *
 127          * should only be accessed directly by packfile.c
 128          */
 129
 130         struct packed_git *packed_git;
 131         /* A most-recently-used ordered version of the packed_git list. */
 132         struct list_head packed_git_mru;
 133
 134         /*
 135          * A fast, rough count of the number of objects in the repository.
 136          * These two fields are not meant for direct access. Use
 137          * approximate_object_count() instead.
 138          */
 139         unsigned long approximate_object_count;
 140         unsigned approximate_object_count_valid : 1;
 141
 142         /*
 143          * Whether packed_git has already been populated with this repository's
 144          * packs.
 145          */
 146         unsigned packed_git_initialized : 1;
 147 };
 148
 149 struct raw_object_store *raw_object_store_new(void);
 150 void raw_object_store_clear(struct raw_object_store *o);
 151
 152 /*
 153  * Put in `buf` the name of the file in the local object database that
 154  * would be used to store a loose object with the specified oid.
 155  */
 156 const char *loose_object_path(struct repository *r, struct strbuf *buf,
 157                               const struct object_id *oid);
 158
 159 void *map_loose_object(struct repository *r, const struct object_id *oid,
 160                        unsigned long *size);
 161
 162 void *read_object_file_extended(struct repository *r,
 163                                 const struct object_id *oid,
 164                                 enum object_type *type,
 165                                 unsigned long *size, int lookup_replace);
 166 static inline void *repo_read_object_file(struct repository *r,
 167                                           const struct object_id *oid,
 168                                           enum object_type *type,
 169                                           unsigned long *size)
 170 {
 171         return read_object_file_extended(r, oid, type, size, 1);
 172 }
 173 #ifndef NO_THE_REPOSITORY_COMPATIBILITY_MACROS
 174 #define read_object_file(oid, type, size) repo_read_object_file(the_repository, oid, type, size)
 175 #endif
 176
 177 /* Read and unpack an object file into memory, write memory to an object file */
 178 int oid_object_info(struct repository *r, const struct object_id *, unsigned long *);
 179
 180 int hash_object_file(const void *buf, unsigned long len,
 181                      const char *type, struct object_id *oid);
 182
 183 int write_object_file(const void *buf, unsigned long len,
 184                       const char *type, struct object_id *oid);
 185
 186 int hash_object_file_literally(const void *buf, unsigned long len,
 187                                const char *type, struct object_id *oid,
 188                                unsigned flags);
 189
 190 int pretend_object_file(void *, unsigned long, enum object_type,
 191                         struct object_id *oid);
 192
 193 int force_object_loose(const struct object_id *oid, time_t mtime);
 194
 195 /*
 196  * Open the loose object at path, check its hash, and return the contents,
 197  * type, and size. If the object is a blob, then "contents" may return NULL,
 198  * to allow streaming of large blobs.
 199  *
 200  * Returns 0 on success, negative on error (details may be written to stderr).
 201  */
 202 int read_loose_object(const char *path,
 203                       const struct object_id *expected_oid,
 204                       enum object_type *type,
 205                       unsigned long *size,
 206                       void **contents);
 207
 208 #ifndef NO_THE_REPOSITORY_COMPATIBILITY_MACROS
 209 #define has_sha1_file_with_flags(sha1, flags) repo_has_sha1_file_with_flags(the_repository, sha1, flags)
 210 #define has_sha1_file(sha1) repo_has_sha1_file(the_repository, sha1)
 211 #endif
 212
 213 /* Same as the above, except for struct object_id. */
 214 int repo_has_object_file(struct repository *r, const struct object_id *oid);
 215 int repo_has_object_file_with_flags(struct repository *r,
 216                                     const struct object_id *oid, int flags);
 217 #ifndef NO_THE_REPOSITORY_COMPATIBILITY_MACROS
 218 #define has_object_file(oid) repo_has_object_file(the_repository, oid)
 219 #define has_object_file_with_flags(oid, flags) repo_has_object_file_with_flags(the_repository, oid, flags)
 220 #endif
 221
 222 /*
 223  * Return true iff an alternate object database has a loose object
 224  * with the specified name.  This function does not respect replace
 225  * references.
 226  */
 227 int has_loose_object_nonlocal(const struct object_id *);
 228
 229 void assert_oid_type(const struct object_id *oid, enum object_type expect);
 230
 231 struct object_info {
 232         /* Request */
 233         enum object_type *typep;
 234         unsigned long *sizep;
 235         off_t *disk_sizep;
 236         unsigned char *delta_base_sha1;
 237         struct strbuf *type_name;
 238         void **contentp;
 239
 240         /* Response */
 241         enum {
 242                 OI_CACHED,
 243                 OI_LOOSE,
 244                 OI_PACKED,
 245                 OI_DBCACHED
 246         } whence;
 247         union {
 248                 /*
 249                  * struct {
 250                  *      ... Nothing to expose in this case
 251                  * } cached;
 252                  * struct {
 253                  *      ... Nothing to expose in this case
 254                  * } loose;
 255                  */
 256                 struct {
 257                         struct packed_git *pack;
 258                         off_t offset;
 259                         unsigned int is_delta;
 260                 } packed;
 261         } u;
 262 };
 263
 264 /*
 265  * Initializer for a "struct object_info" that wants no items. You may
 266  * also memset() the memory to all-zeroes.
 267  */
 268 #define OBJECT_INFO_INIT {NULL}
 269
 270 /* Invoke lookup_replace_object() on the given hash */
 271 #define OBJECT_INFO_LOOKUP_REPLACE 1
 272 /* Allow reading from a loose object file of unknown/bogus type */
 273 #define OBJECT_INFO_ALLOW_UNKNOWN_TYPE 2
 274 /* Do not check cached storage */
 275 #define OBJECT_INFO_SKIP_CACHED 4
 276 /* Do not retry packed storage after checking packed and loose storage */
 277 #define OBJECT_INFO_QUICK 8
 278 /* Do not check loose object */
 279 #define OBJECT_INFO_IGNORE_LOOSE 16
 280 /*
 281  * Do not attempt to fetch the object if missing (even if fetch_is_missing is
 282  * nonzero).
 283  */
 284 #define OBJECT_INFO_SKIP_FETCH_OBJECT 32
 285 /*
 286  * This is meant for bulk prefetching of missing blobs in a partial
 287  * clone. Implies OBJECT_INFO_SKIP_FETCH_OBJECT and OBJECT_INFO_QUICK
 288  */
 289 #define OBJECT_INFO_FOR_PREFETCH (OBJECT_INFO_SKIP_FETCH_OBJECT | OBJECT_INFO_QUICK)
 290
 291 int oid_object_info_extended(struct repository *r,
 292                              const struct object_id *,
 293                              struct object_info *, unsigned flags);
 294
 295 /*
 296  * Iterate over the files in the loose-object parts of the object
 297  * directory "path", triggering the following callbacks:
 298  *
 299  *  - loose_object is called for each loose object we find.
 300  *
 301  *  - loose_cruft is called for any files that do not appear to be
 302  *    loose objects. Note that we only look in the loose object
 303  *    directories "objects/[0-9a-f]{2}/", so we will not report
 304  *    "objects/foobar" as cruft.
 305  *
 306  *  - loose_subdir is called for each top-level hashed subdirectory
 307  *    of the object directory (e.g., "$OBJDIR/f0"). It is called
 308  *    after the objects in the directory are processed.
 309  *
 310  * Any callback that is NULL will be ignored. Callbacks returning non-zero
 311  * will end the iteration.
 312  *
 313  * In the "buf" variant, "path" is a strbuf which will also be used as a
 314  * scratch buffer, but restored to its original contents before
 315  * the function returns.
 316  */
 317 typedef int each_loose_object_fn(const struct object_id *oid,
 318                                  const char *path,
 319                                  void *data);
 320 typedef int each_loose_cruft_fn(const char *basename,
 321                                 const char *path,
 322                                 void *data);
 323 typedef int each_loose_subdir_fn(unsigned int nr,
 324                                  const char *path,
 325                                  void *data);
 326 int for_each_file_in_obj_subdir(unsigned int subdir_nr,
 327                                 struct strbuf *path,
 328                                 each_loose_object_fn obj_cb,
 329                                 each_loose_cruft_fn cruft_cb,
 330                                 each_loose_subdir_fn subdir_cb,
 331                                 void *data);
 332 int for_each_loose_file_in_objdir(const char *path,
 333                                   each_loose_object_fn obj_cb,
 334                                   each_loose_cruft_fn cruft_cb,
 335                                   each_loose_subdir_fn subdir_cb,
 336                                   void *data);
 337 int for_each_loose_file_in_objdir_buf(struct strbuf *path,
 338                                       each_loose_object_fn obj_cb,
 339                                       each_loose_cruft_fn cruft_cb,
 340                                       each_loose_subdir_fn subdir_cb,
 341                                       void *data);
 342
 343 /* Flags for for_each_*_object() below. */
 344 enum for_each_object_flags {
 345         /* Iterate only over local objects, not alternates. */
 346         FOR_EACH_OBJECT_LOCAL_ONLY = (1<<0),
 347
 348         /* Only iterate over packs obtained from the promisor remote. */
 349         FOR_EACH_OBJECT_PROMISOR_ONLY = (1<<1),
 350
 351         /*
 352          * Visit objects within a pack in packfile order rather than .idx order
 353          */
 354         FOR_EACH_OBJECT_PACK_ORDER = (1<<2),
 355 };
 356
 357 /*
 358  * Iterate over all accessible loose objects without respect to
 359  * reachability. By default, this includes both local and alternate objects.
 360  * The order in which objects are visited is unspecified.
 361  *
 362  * Any flags specific to packs are ignored.
 363  */
 364 int for_each_loose_object(each_loose_object_fn, void *,
 365                           enum for_each_object_flags flags);
 366
 367 /*
 368  * Iterate over all accessible packed objects without respect to reachability.
 369  * By default, this includes both local and alternate packs.
 370  *
 371  * Note that some objects may appear twice if they are found in multiple packs.
 372  * Each pack is visited in an unspecified order. By default, objects within a
 373  * pack are visited in pack-idx order (i.e., sorted by oid).
 374  */
 375 typedef int each_packed_object_fn(const struct object_id *oid,
 376                                   struct packed_git *pack,
 377                                   uint32_t pos,
 378                                   void *data);
 379 int for_each_object_in_pack(struct packed_git *p,
 380                             each_packed_object_fn, void *data,
 381                             enum for_each_object_flags flags);
 382 int for_each_packed_object(each_packed_object_fn, void *,
 383                            enum for_each_object_flags flags);
 384
 385 #endif /* OBJECT_STORE_H */