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