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