]> git.ipfire.org Git - thirdparty/git.git/blame - object-file.c
tree-wide: apply equals-null.cocci
[thirdparty/git.git] / object-file.c
CommitLineData
0fcfd160
LT
1/*
2 * GIT - The information manager from hell
3 *
4 * Copyright (C) Linus Torvalds, 2005
5 *
e5afd444 6 * This handles basic git object files - packing, unpacking,
0fcfd160
LT
7 * creation etc.
8 */
0fcfd160 9#include "cache.h"
b2141fc1 10#include "config.h"
6eac50d8 11#include "string-list.h"
697cc8ef 12#include "lockfile.h"
1f688557 13#include "delta.h"
a733cb60 14#include "pack.h"
8e440259
PE
15#include "blob.h"
16#include "commit.h"
4dd1fbc7 17#include "run-command.h"
8e440259
PE
18#include "tag.h"
19#include "tree.h"
c879daa2 20#include "tree-walk.h"
f35a6d3b 21#include "refs.h"
70f5d5d3 22#include "pack-revindex.h"
bc626927 23#include "hash-lookup.h"
568508e7 24#include "bulk-checkin.h"
031dc927 25#include "repository.h"
47f351e9 26#include "replace-object.h"
090ea126 27#include "streaming.h"
543c5caa 28#include "dir.h"
12d95ef6 29#include "list.h"
c4c6effa 30#include "mergesort.h"
cf3c6352 31#include "quote.h"
4f39cd82 32#include "packfile.h"
90c62155 33#include "object-store.h"
b14ed5ad 34#include "promisor-remote.h"
a35e03de 35#include "submodule.h"
e05db0fd 36
1af64f73 37/* The maximum size for an object header. */
38#define MAX_HEADER_LEN 32
39
e1ccd7e2 40
41#define EMPTY_TREE_SHA1_BIN_LITERAL \
42 "\x4b\x82\x5d\xc6\x42\xcb\x6e\xb9\xa0\x60" \
43 "\xe5\x4b\xf8\xd6\x92\x88\xfb\xee\x49\x04"
13eeedb5 44#define EMPTY_TREE_SHA256_BIN_LITERAL \
45 "\x6e\xf1\x9b\x41\x22\x5c\x53\x69\xf1\xc1" \
46 "\x04\xd4\x5d\x8d\x85\xef\xa9\xb0\x57\xb5" \
47 "\x3b\x14\xb4\xb9\xb9\x39\xdd\x74\xde\xcc" \
48 "\x53\x21"
e1ccd7e2 49
50#define EMPTY_BLOB_SHA1_BIN_LITERAL \
51 "\xe6\x9d\xe2\x9b\xb2\xd1\xd6\x43\x4b\x8b" \
52 "\x29\xae\x77\x5a\xd8\xc2\xe4\x8c\x53\x91"
13eeedb5 53#define EMPTY_BLOB_SHA256_BIN_LITERAL \
54 "\x47\x3a\x0f\x4c\x3b\xe8\xa9\x36\x81\xa2" \
55 "\x67\xe3\xb1\xe9\xa7\xdc\xda\x11\x85\x43" \
56 "\x6f\xe1\x41\xf7\x74\x91\x20\xa3\x03\x72" \
57 "\x18\x13"
e1ccd7e2 58
e1ccd7e2 59static const struct object_id empty_tree_oid = {
5a6dce70 60 .hash = EMPTY_TREE_SHA1_BIN_LITERAL,
61 .algo = GIT_HASH_SHA1,
8576fde6 62};
e1ccd7e2 63static const struct object_id empty_blob_oid = {
5a6dce70 64 .hash = EMPTY_BLOB_SHA1_BIN_LITERAL,
65 .algo = GIT_HASH_SHA1,
8576fde6 66};
14228447 67static const struct object_id null_oid_sha1 = {
68 .hash = {0},
69 .algo = GIT_HASH_SHA1,
70};
13eeedb5 71static const struct object_id empty_tree_oid_sha256 = {
5a6dce70 72 .hash = EMPTY_TREE_SHA256_BIN_LITERAL,
73 .algo = GIT_HASH_SHA256,
13eeedb5 74};
75static const struct object_id empty_blob_oid_sha256 = {
5a6dce70 76 .hash = EMPTY_BLOB_SHA256_BIN_LITERAL,
77 .algo = GIT_HASH_SHA256,
13eeedb5 78};
14228447 79static const struct object_id null_oid_sha256 = {
80 .hash = {0},
81 .algo = GIT_HASH_SHA256,
82};
88cd621d 83
ac73cedf 84static void git_hash_sha1_init(git_hash_ctx *ctx)
f50e766b 85{
ac73cedf 86 git_SHA1_Init(&ctx->sha1);
f50e766b 87}
88
768e30ea 89static void git_hash_sha1_clone(git_hash_ctx *dst, const git_hash_ctx *src)
90{
91 git_SHA1_Clone(&dst->sha1, &src->sha1);
92}
93
ac73cedf 94static void git_hash_sha1_update(git_hash_ctx *ctx, const void *data, size_t len)
f50e766b 95{
ac73cedf 96 git_SHA1_Update(&ctx->sha1, data, len);
f50e766b 97}
98
ac73cedf 99static void git_hash_sha1_final(unsigned char *hash, git_hash_ctx *ctx)
f50e766b 100{
ac73cedf 101 git_SHA1_Final(hash, &ctx->sha1);
f50e766b 102}
103
ab795f0d 104static void git_hash_sha1_final_oid(struct object_id *oid, git_hash_ctx *ctx)
105{
106 git_SHA1_Final(oid->hash, &ctx->sha1);
107 memset(oid->hash + GIT_SHA1_RAWSZ, 0, GIT_MAX_RAWSZ - GIT_SHA1_RAWSZ);
5a6dce70 108 oid->algo = GIT_HASH_SHA1;
ab795f0d 109}
110
13eeedb5 111
112static void git_hash_sha256_init(git_hash_ctx *ctx)
113{
114 git_SHA256_Init(&ctx->sha256);
115}
116
768e30ea 117static void git_hash_sha256_clone(git_hash_ctx *dst, const git_hash_ctx *src)
118{
119 git_SHA256_Clone(&dst->sha256, &src->sha256);
120}
121
13eeedb5 122static void git_hash_sha256_update(git_hash_ctx *ctx, const void *data, size_t len)
123{
124 git_SHA256_Update(&ctx->sha256, data, len);
125}
126
127static void git_hash_sha256_final(unsigned char *hash, git_hash_ctx *ctx)
128{
129 git_SHA256_Final(hash, &ctx->sha256);
130}
131
ab795f0d 132static void git_hash_sha256_final_oid(struct object_id *oid, git_hash_ctx *ctx)
133{
134 git_SHA256_Final(oid->hash, &ctx->sha256);
135 /*
136 * This currently does nothing, so the compiler should optimize it out,
137 * but keep it in case we extend the hash size again.
138 */
139 memset(oid->hash + GIT_SHA256_RAWSZ, 0, GIT_MAX_RAWSZ - GIT_SHA256_RAWSZ);
5a6dce70 140 oid->algo = GIT_HASH_SHA256;
ab795f0d 141}
142
ac73cedf 143static void git_hash_unknown_init(git_hash_ctx *ctx)
f50e766b 144{
1a07e59c 145 BUG("trying to init unknown hash");
f50e766b 146}
147
768e30ea 148static void git_hash_unknown_clone(git_hash_ctx *dst, const git_hash_ctx *src)
149{
150 BUG("trying to clone unknown hash");
151}
152
ac73cedf 153static void git_hash_unknown_update(git_hash_ctx *ctx, const void *data, size_t len)
f50e766b 154{
1a07e59c 155 BUG("trying to update unknown hash");
f50e766b 156}
157
ac73cedf 158static void git_hash_unknown_final(unsigned char *hash, git_hash_ctx *ctx)
f50e766b 159{
1a07e59c 160 BUG("trying to finalize unknown hash");
f50e766b 161}
162
ab795f0d 163static void git_hash_unknown_final_oid(struct object_id *oid, git_hash_ctx *ctx)
164{
165 BUG("trying to finalize unknown hash");
166}
167
168
f50e766b 169const struct git_hash_algo hash_algos[GIT_HASH_NALGOS] = {
170 {
171 NULL,
172 0x00000000,
173 0,
174 0,
a2ce0a75 175 0,
f50e766b 176 git_hash_unknown_init,
768e30ea 177 git_hash_unknown_clone,
f50e766b 178 git_hash_unknown_update,
179 git_hash_unknown_final,
ab795f0d 180 git_hash_unknown_final_oid,
f50e766b 181 NULL,
182 NULL,
14228447 183 NULL,
f50e766b 184 },
185 {
1ccf07cb 186 "sha1",
f50e766b 187 /* "sha1", big-endian */
188 0x73686131,
f50e766b 189 GIT_SHA1_RAWSZ,
190 GIT_SHA1_HEXSZ,
a2ce0a75 191 GIT_SHA1_BLKSZ,
f50e766b 192 git_hash_sha1_init,
768e30ea 193 git_hash_sha1_clone,
f50e766b 194 git_hash_sha1_update,
195 git_hash_sha1_final,
ab795f0d 196 git_hash_sha1_final_oid,
f50e766b 197 &empty_tree_oid,
198 &empty_blob_oid,
14228447 199 &null_oid_sha1,
f50e766b 200 },
13eeedb5 201 {
202 "sha256",
203 /* "s256", big-endian */
204 0x73323536,
205 GIT_SHA256_RAWSZ,
206 GIT_SHA256_HEXSZ,
207 GIT_SHA256_BLKSZ,
208 git_hash_sha256_init,
768e30ea 209 git_hash_sha256_clone,
13eeedb5 210 git_hash_sha256_update,
211 git_hash_sha256_final,
ab795f0d 212 git_hash_sha256_final_oid,
13eeedb5 213 &empty_tree_oid_sha256,
214 &empty_blob_oid_sha256,
14228447 215 &null_oid_sha256,
13eeedb5 216 }
f50e766b 217};
218
14228447 219const struct object_id *null_oid(void)
220{
221 return the_hash_algo->null_oid;
222}
223
d8a92ced 224const char *empty_tree_oid_hex(void)
225{
226 static char buf[GIT_MAX_HEXSZ + 1];
227 return oid_to_hex_r(buf, the_hash_algo->empty_tree);
228}
229
230const char *empty_blob_oid_hex(void)
231{
232 static char buf[GIT_MAX_HEXSZ + 1];
233 return oid_to_hex_r(buf, the_hash_algo->empty_blob);
234}
235
2f90b9d9 236int hash_algo_by_name(const char *name)
237{
238 int i;
239 if (!name)
240 return GIT_HASH_UNKNOWN;
241 for (i = 1; i < GIT_HASH_NALGOS; i++)
242 if (!strcmp(name, hash_algos[i].name))
243 return i;
244 return GIT_HASH_UNKNOWN;
245}
246
247int hash_algo_by_id(uint32_t format_id)
248{
249 int i;
250 for (i = 1; i < GIT_HASH_NALGOS; i++)
251 if (format_id == hash_algos[i].format_id)
252 return i;
253 return GIT_HASH_UNKNOWN;
254}
255
95399788 256int hash_algo_by_length(int len)
257{
258 int i;
259 for (i = 1; i < GIT_HASH_NALGOS; i++)
260 if (len == hash_algos[i].rawsz)
261 return i;
262 return GIT_HASH_UNKNOWN;
263}
2f90b9d9 264
c597ba80
NTND
265/*
266 * This is meant to hold a *small* number of objects that you would
cb1c8d1d 267 * want read_object_file() to be able to return, but yet you do not want
c597ba80
NTND
268 * to write them into the object store (e.g. a browse-only
269 * application).
270 */
271static struct cached_object {
62ba93ea 272 struct object_id oid;
c597ba80
NTND
273 enum object_type type;
274 void *buf;
275 unsigned long size;
276} *cached_objects;
277static int cached_object_nr, cached_object_alloc;
278
279static struct cached_object empty_tree = {
62ba93ea 280 { EMPTY_TREE_SHA1_BIN_LITERAL },
c597ba80
NTND
281 OBJ_TREE,
282 "",
283 0
284};
285
62ba93ea 286static struct cached_object *find_cached_object(const struct object_id *oid)
c597ba80
NTND
287{
288 int i;
289 struct cached_object *co = cached_objects;
290
291 for (i = 0; i < cached_object_nr; i++, co++) {
4a7e27e9 292 if (oideq(&co->oid, oid))
c597ba80
NTND
293 return co;
294 }
4a7e27e9 295 if (oideq(oid, the_hash_algo->empty_tree))
c597ba80
NTND
296 return &empty_tree;
297 return NULL;
298}
299
9472935d 300
8462ff43 301static int get_conv_flags(unsigned flags)
9472935d
TB
302{
303 if (flags & HASH_RENORMALIZE)
8462ff43 304 return CONV_EOL_RENORMALIZE;
9472935d 305 else if (flags & HASH_WRITE_OBJECT)
107642fe 306 return global_conv_flags_eol | CONV_WRITE_OBJECT;
9472935d 307 else
8462ff43 308 return 0;
9472935d
TB
309}
310
311
90a6464b
JH
312int mkdir_in_gitdir(const char *path)
313{
314 if (mkdir(path, 0777)) {
315 int saved_errno = errno;
316 struct stat st;
317 struct strbuf sb = STRBUF_INIT;
318
319 if (errno != EEXIST)
320 return -1;
321 /*
322 * Are we looking at a path in a symlinked worktree
323 * whose original repository does not yet have it?
324 * e.g. .git/rr-cache pointing at its original
325 * repository in which the user hasn't performed any
326 * conflict resolution yet?
327 */
328 if (lstat(path, &st) || !S_ISLNK(st.st_mode) ||
329 strbuf_readlink(&sb, path, st.st_size) ||
330 !is_absolute_path(sb.buf) ||
331 mkdir(sb.buf, 0777)) {
332 strbuf_release(&sb);
333 errno = saved_errno;
334 return -1;
335 }
336 strbuf_release(&sb);
337 }
338 return adjust_shared_perm(path);
339}
340
eb3c027e 341static enum scld_error safe_create_leading_directories_1(char *path, int share)
b2cb9425 342{
26c8ae2a 343 char *next_component = path + offset_1st_component(path);
0be0521b 344 enum scld_error ret = SCLD_OK;
67d42212 345
0be0521b 346 while (ret == SCLD_OK && next_component) {
f0502332 347 struct stat st;
0f527403 348 char *slash = next_component, slash_character;
f0502332 349
0f527403
MH
350 while (*slash && !is_dir_sep(*slash))
351 slash++;
352
353 if (!*slash)
b2cb9425 354 break;
bf10cf70 355
26c8ae2a 356 next_component = slash + 1;
0f527403 357 while (is_dir_sep(*next_component))
bf10cf70 358 next_component++;
26c8ae2a 359 if (!*next_component)
5f0bdf50 360 break;
831651fd 361
0f527403 362 slash_character = *slash;
831651fd 363 *slash = '\0';
67d42212
JR
364 if (!stat(path, &st)) {
365 /* path exists */
204a047f
MH
366 if (!S_ISDIR(st.st_mode)) {
367 errno = ENOTDIR;
0be0521b 368 ret = SCLD_EXISTS;
204a047f 369 }
53a39721 370 } else if (mkdir(path, 0777)) {
928734d9 371 if (errno == EEXIST &&
9e6f885d 372 !stat(path, &st) && S_ISDIR(st.st_mode))
928734d9 373 ; /* somebody created it since we checked */
18d37e86
MH
374 else if (errno == ENOENT)
375 /*
376 * Either mkdir() failed because
377 * somebody just pruned the containing
378 * directory, or stat() failed because
379 * the file that was in our way was
380 * just removed. Either way, inform
381 * the caller that it might be worth
382 * trying again:
383 */
384 ret = SCLD_VANISHED;
9e6f885d 385 else
0be0521b 386 ret = SCLD_FAILED;
eb3c027e 387 } else if (share && adjust_shared_perm(path)) {
0be0521b 388 ret = SCLD_PERMS;
457f06d6 389 }
0f527403 390 *slash = slash_character;
b2cb9425 391 }
9e6f885d 392 return ret;
b2cb9425 393}
723c31fe 394
eb3c027e
MT
395enum scld_error safe_create_leading_directories(char *path)
396{
397 return safe_create_leading_directories_1(path, 1);
398}
399
400enum scld_error safe_create_leading_directories_no_share(char *path)
401{
402 return safe_create_leading_directories_1(path, 0);
403}
404
0be0521b 405enum scld_error safe_create_leading_directories_const(const char *path)
8e21d63b 406{
02944307 407 int save_errno;
8e21d63b
JK
408 /* path points to cache entries, so xstrdup before messing with it */
409 char *buf = xstrdup(path);
0be0521b 410 enum scld_error result = safe_create_leading_directories(buf);
02944307
MH
411
412 save_errno = errno;
8e21d63b 413 free(buf);
02944307 414 errno = save_errno;
8e21d63b
JK
415 return result;
416}
417
514c5fdd 418static void fill_loose_path(struct strbuf *buf, const struct object_id *oid)
ace1534d
JH
419{
420 int i;
94b5e093 421 for (i = 0; i < the_hash_algo->rawsz; i++) {
ace1534d 422 static char hex[] = "0123456789abcdef";
514c5fdd 423 unsigned int val = oid->hash[i];
f7b7774f
JK
424 strbuf_addch(buf, hex[val >> 4]);
425 strbuf_addch(buf, hex[val & 0xf]);
afbba2f0 426 if (!i)
f7b7774f 427 strbuf_addch(buf, '/');
ace1534d
JH
428 }
429}
430
f0eaf638
JK
431static const char *odb_loose_path(struct object_directory *odb,
432 struct strbuf *buf,
514c5fdd 433 const struct object_id *oid)
0fcfd160 434{
b69fb867 435 strbuf_reset(buf);
f0eaf638 436 strbuf_addstr(buf, odb->path);
34498471 437 strbuf_addch(buf, '/');
514c5fdd 438 fill_loose_path(buf, oid);
f3f043a1 439 return buf->buf;
0fcfd160
LT
440}
441
f3f043a1 442const char *loose_object_path(struct repository *r, struct strbuf *buf,
514c5fdd 443 const struct object_id *oid)
29ec6af2 444{
514c5fdd 445 return odb_loose_path(r->objects->odb, buf, oid);
0fcfd160
LT
446}
447
4ea82473
JK
448/*
449 * Return non-zero iff the path is usable as an alternate object database.
450 */
13313fc3
SB
451static int alt_odb_usable(struct raw_object_store *o,
452 struct strbuf *path,
cf2dc1c2 453 const char *normalized_objdir, khiter_t *pos)
4ea82473 454{
cf2dc1c2 455 int r;
4ea82473
JK
456
457 /* Detect cases where alternate disappeared */
458 if (!is_directory(path->buf)) {
259328b7
NTND
459 error(_("object directory %s does not exist; "
460 "check .git/objects/info/alternates"),
4ea82473
JK
461 path->buf);
462 return 0;
463 }
464
465 /*
466 * Prevent the common mistake of listing the same
467 * thing twice, or object directory itself.
468 */
cf2dc1c2
EW
469 if (!o->odb_by_path) {
470 khiter_t p;
471
472 o->odb_by_path = kh_init_odb_path_map();
473 assert(!o->odb->next);
474 p = kh_put_odb_path_map(o->odb_by_path, o->odb->path, &r);
475 assert(r == 1); /* never used */
476 kh_value(o->odb_by_path, p) = o->odb;
4ea82473 477 }
cf2dc1c2 478 if (fspatheq(path->buf, normalized_objdir))
4ea82473 479 return 0;
cf2dc1c2
EW
480 *pos = kh_put_odb_path_map(o->odb_by_path, path->buf, &r);
481 /* r: 0 = exists, 1 = never used, 2 = deleted */
482 return r == 0 ? 0 : 1;
4ea82473
JK
483}
484
ddd5d056
JH
485/*
486 * Prepare alternate object database registry.
d5a63b99
JH
487 *
488 * The variable alt_odb_list points at the list of struct
263db403 489 * object_directory. The elements on this list come from
d5a63b99
JH
490 * non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT
491 * environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates,
1494e038
JH
492 * whose contents is similar to that environment variable but can be
493 * LF separated. Its base points at a statically allocated buffer that
d5a63b99
JH
494 * contains "/the/directory/corresponding/to/.git/objects/...", while
495 * its name points just after the slash at the end of ".git/objects/"
e5afd444
496 * in the example above, and has enough space to hold all hex characters
497 * of the object ID, an extra slash for the first level indirection, and
498 * the terminating NUL.
ddd5d056 499 */
77f012e8
SB
500static void read_info_alternates(struct repository *r,
501 const char *relative_base,
502 int depth);
407532f8 503static int link_alt_odb_entry(struct repository *r, const struct strbuf *entry,
cfc62fc9 504 const char *relative_base, int depth, const char *normalized_objdir)
ace1534d 505{
263db403 506 struct object_directory *ent;
5bdf0a84 507 struct strbuf pathbuf = STRBUF_INIT;
cf2dc1c2 508 khiter_t pos;
d5a63b99 509
407532f8 510 if (!is_absolute_path(entry->buf) && relative_base) {
4ac9006f 511 strbuf_realpath(&pathbuf, relative_base, 1);
5bdf0a84 512 strbuf_addch(&pathbuf, '/');
c2f493a4 513 }
407532f8 514 strbuf_addbuf(&pathbuf, entry);
c2f493a4 515
37a95862 516 if (strbuf_normalize_path(&pathbuf) < 0 && relative_base) {
259328b7 517 error(_("unable to normalize alternate object path: %s"),
670c359d
JK
518 pathbuf.buf);
519 strbuf_release(&pathbuf);
520 return -1;
521 }
5bdf0a84
HW
522
523 /*
524 * The trailing slash after the directory name is given by
525 * this function at the end. Remove duplicates.
526 */
4ea82473
JK
527 while (pathbuf.len && pathbuf.buf[pathbuf.len - 1] == '/')
528 strbuf_setlen(&pathbuf, pathbuf.len - 1);
5bdf0a84 529
cf2dc1c2 530 if (!alt_odb_usable(r->objects, &pathbuf, normalized_objdir, &pos)) {
4ea82473 531 strbuf_release(&pathbuf);
c2f493a4
MW
532 return -1;
533 }
534
ca56dadb 535 CALLOC_ARRAY(ent, 1);
cf2dc1c2
EW
536 /* pathbuf.buf is already in r->objects->odb_by_path */
537 ent->path = strbuf_detach(&pathbuf, NULL);
c2f493a4
MW
538
539 /* add the alternate entry */
f0eaf638
JK
540 *r->objects->odb_tail = ent;
541 r->objects->odb_tail = &(ent->next);
c2f493a4 542 ent->next = NULL;
cf2dc1c2
EW
543 assert(r->objects->odb_by_path);
544 kh_value(r->objects->odb_by_path, pos) = ent;
c2f493a4
MW
545
546 /* recursively add alternates */
cf2dc1c2 547 read_info_alternates(r, ent->path, depth + 1);
c2f493a4
MW
548
549 return 0;
550}
551
cf3c6352
JK
552static const char *parse_alt_odb_entry(const char *string,
553 int sep,
554 struct strbuf *out)
555{
556 const char *end;
557
558 strbuf_reset(out);
559
560 if (*string == '#') {
561 /* comment; consume up to next separator */
562 end = strchrnul(string, sep);
563 } else if (*string == '"' && !unquote_c_style(out, string, &end)) {
564 /*
565 * quoted path; unquote_c_style has copied the
566 * data for us and set "end". Broken quoting (e.g.,
567 * an entry that doesn't end with a quote) falls
568 * back to the unquoted case below.
569 */
570 } else {
571 /* normal, unquoted path */
572 end = strchrnul(string, sep);
573 strbuf_add(out, string, end - string);
574 }
575
576 if (*end)
577 end++;
578 return end;
579}
580
77f012e8
SB
581static void link_alt_odb_entries(struct repository *r, const char *alt,
582 int sep, const char *relative_base, int depth)
c2f493a4 583{
539e7506 584 struct strbuf objdirbuf = STRBUF_INIT;
cf3c6352 585 struct strbuf entry = STRBUF_INIT;
c2f493a4 586
f28e3668
JK
587 if (!alt || !*alt)
588 return;
589
c2f493a4 590 if (depth > 5) {
259328b7 591 error(_("%s: ignoring alternate object stores, nesting too deep"),
c2f493a4
MW
592 relative_base);
593 return;
594 }
595
f0eaf638 596 strbuf_add_absolute_path(&objdirbuf, r->objects->odb->path);
670c359d 597 if (strbuf_normalize_path(&objdirbuf) < 0)
259328b7 598 die(_("unable to normalize object directory: %s"),
670c359d 599 objdirbuf.buf);
539e7506 600
cf3c6352
JK
601 while (*alt) {
602 alt = parse_alt_odb_entry(alt, sep, &entry);
603 if (!entry.len)
9577e7e3 604 continue;
407532f8 605 link_alt_odb_entry(r, &entry,
cfc62fc9 606 relative_base, depth, objdirbuf.buf);
9577e7e3 607 }
cf3c6352 608 strbuf_release(&entry);
539e7506 609 strbuf_release(&objdirbuf);
d5a63b99
JH
610}
611
77f012e8
SB
612static void read_info_alternates(struct repository *r,
613 const char *relative_base,
614 int depth)
d5a63b99 615{
5015f01c 616 char *path;
dc732bd5 617 struct strbuf buf = STRBUF_INIT;
d5a63b99 618
5015f01c 619 path = xstrfmt("%s/info/alternates", relative_base);
dc732bd5 620 if (strbuf_read_file(&buf, path, 1024) < 0) {
f0f7bebe 621 warn_on_fopen_errors(path);
dc732bd5 622 free(path);
9a217f2a 623 return;
ace1534d 624 }
d5a63b99 625
77f012e8 626 link_alt_odb_entries(r, buf.buf, '\n', relative_base, depth);
dc732bd5
JK
627 strbuf_release(&buf);
628 free(path);
ace1534d
JH
629}
630
bef70b22
DB
631void add_to_alternates_file(const char *reference)
632{
f132a127 633 struct lock_file lock = LOCK_INIT;
77b9b1d1
JK
634 char *alts = git_pathdup("objects/info/alternates");
635 FILE *in, *out;
f132a127 636 int found = 0;
77b9b1d1 637
f132a127
638 hold_lock_file_for_update(&lock, alts, LOCK_DIE_ON_ERROR);
639 out = fdopen_lock_file(&lock, "w");
77b9b1d1 640 if (!out)
259328b7 641 die_errno(_("unable to fdopen alternates lockfile"));
77b9b1d1
JK
642
643 in = fopen(alts, "r");
644 if (in) {
645 struct strbuf line = STRBUF_INIT;
77b9b1d1 646
3f163962 647 while (strbuf_getline(&line, in) != EOF) {
77b9b1d1
JK
648 if (!strcmp(reference, line.buf)) {
649 found = 1;
650 break;
651 }
652 fprintf_or_die(out, "%s\n", line.buf);
653 }
654
655 strbuf_release(&line);
656 fclose(in);
77b9b1d1
JK
657 }
658 else if (errno != ENOENT)
259328b7 659 die_errno(_("unable to read alternates file"));
77b9b1d1 660
f132a127
661 if (found) {
662 rollback_lock_file(&lock);
663 } else {
77b9b1d1 664 fprintf_or_die(out, "%s\n", reference);
f132a127 665 if (commit_lock_file(&lock))
259328b7 666 die_errno(_("unable to move new alternates file into place"));
f0eaf638 667 if (the_repository->objects->loaded_alternates)
93d8d1e2
SB
668 link_alt_odb_entries(the_repository, reference,
669 '\n', NULL, 0);
77b9b1d1
JK
670 }
671 free(alts);
bef70b22
DB
672}
673
a5b34d21
JK
674void add_to_alternates_memory(const char *reference)
675{
676 /*
677 * Make sure alternates are initialized, or else our entry may be
678 * overwritten when they are.
679 */
0b209034 680 prepare_alt_odb(the_repository);
a5b34d21 681
93d8d1e2
SB
682 link_alt_odb_entries(the_repository, reference,
683 '\n', NULL, 0);
a5b34d21
JK
684}
685
9eeea7d2
SB
686/*
687 * Compute the exact path an alternate is at and returns it. In case of
688 * error NULL is returned and the human readable error is added to `err`
efde7b72 689 * `path` may be relative and should point to $GIT_DIR.
9eeea7d2
SB
690 * `err` must not be null.
691 */
692char *compute_alternate_path(const char *path, struct strbuf *err)
693{
694 char *ref_git = NULL;
4530a85b 695 const char *repo;
9eeea7d2
SB
696 int seen_error = 0;
697
4530a85b
AM
698 ref_git = real_pathdup(path, 0);
699 if (!ref_git) {
9eeea7d2
SB
700 seen_error = 1;
701 strbuf_addf(err, _("path '%s' does not exist"), path);
702 goto out;
4530a85b 703 }
9eeea7d2
SB
704
705 repo = read_gitfile(ref_git);
706 if (!repo)
707 repo = read_gitfile(mkpath("%s/.git", ref_git));
708 if (repo) {
709 free(ref_git);
710 ref_git = xstrdup(repo);
711 }
712
713 if (!repo && is_directory(mkpath("%s/.git/objects", ref_git))) {
714 char *ref_git_git = mkpathdup("%s/.git", ref_git);
715 free(ref_git);
716 ref_git = ref_git_git;
717 } else if (!is_directory(mkpath("%s/objects", ref_git))) {
718 struct strbuf sb = STRBUF_INIT;
719 seen_error = 1;
720 if (get_common_dir(&sb, ref_git)) {
721 strbuf_addf(err,
722 _("reference repository '%s' as a linked "
723 "checkout is not supported yet."),
724 path);
725 goto out;
726 }
727
728 strbuf_addf(err, _("reference repository '%s' is not a "
729 "local repository."), path);
730 goto out;
731 }
732
733 if (!access(mkpath("%s/shallow", ref_git), F_OK)) {
734 strbuf_addf(err, _("reference repository '%s' is shallow"),
735 path);
736 seen_error = 1;
737 goto out;
738 }
739
740 if (!access(mkpath("%s/info/grafts", ref_git), F_OK)) {
741 strbuf_addf(err,
742 _("reference repository '%s' is grafted"),
743 path);
744 seen_error = 1;
745 goto out;
746 }
747
748out:
749 if (seen_error) {
6a83d902 750 FREE_AND_NULL(ref_git);
9eeea7d2
SB
751 }
752
753 return ref_git;
754}
755
f57a7396
TB
756struct object_directory *find_odb(struct repository *r, const char *obj_dir)
757{
758 struct object_directory *odb;
759 char *obj_dir_real = real_pathdup(obj_dir, 1);
760 struct strbuf odb_path_real = STRBUF_INIT;
761
762 prepare_alt_odb(r);
763 for (odb = r->objects->odb; odb; odb = odb->next) {
764 strbuf_realpath(&odb_path_real, odb->path, 1);
765 if (!strcmp(obj_dir_real, odb_path_real.buf))
766 break;
767 }
768
769 free(obj_dir_real);
770 strbuf_release(&odb_path_real);
771
772 if (!odb)
773 die(_("could not find object directory matching %s"), obj_dir);
774 return odb;
775}
776
709dfa69
JK
777static void fill_alternate_refs_command(struct child_process *cmd,
778 const char *repo_path)
779{
780 const char *value;
781
782 if (!git_config_get_value("core.alternateRefsCommand", &value)) {
783 cmd->use_shell = 1;
784
c972bf4c
JK
785 strvec_push(&cmd->args, value);
786 strvec_push(&cmd->args, repo_path);
709dfa69
JK
787 } else {
788 cmd->git_cmd = 1;
789
c972bf4c
JK
790 strvec_pushf(&cmd->args, "--git-dir=%s", repo_path);
791 strvec_push(&cmd->args, "for-each-ref");
792 strvec_push(&cmd->args, "--format=%(objectname)");
709dfa69
JK
793
794 if (!git_config_get_value("core.alternateRefsPrefixes", &value)) {
c972bf4c
JK
795 strvec_push(&cmd->args, "--");
796 strvec_split(&cmd->args, value);
709dfa69
JK
797 }
798 }
799
800 cmd->env = local_repo_env;
801 cmd->out = -1;
802}
803
804static void read_alternate_refs(const char *path,
805 alternate_ref_fn *cb,
806 void *data)
807{
808 struct child_process cmd = CHILD_PROCESS_INIT;
809 struct strbuf line = STRBUF_INIT;
810 FILE *fh;
811
812 fill_alternate_refs_command(&cmd, path);
813
814 if (start_command(&cmd))
815 return;
816
817 fh = xfdopen(cmd.out, "r");
818 while (strbuf_getline_lf(&line, fh) != EOF) {
819 struct object_id oid;
820 const char *p;
821
822 if (parse_oid_hex(line.buf, &oid, &p) || *p) {
823 warning(_("invalid line while parsing alternate refs: %s"),
824 line.buf);
825 break;
826 }
827
828 cb(&oid, data);
829 }
830
831 fclose(fh);
832 finish_command(&cmd);
86ad3ea5 833 strbuf_release(&line);
709dfa69
JK
834}
835
836struct alternate_refs_data {
837 alternate_ref_fn *fn;
838 void *data;
839};
840
841static int refs_from_alternate_cb(struct object_directory *e,
842 void *data)
843{
844 struct strbuf path = STRBUF_INIT;
845 size_t base_len;
846 struct alternate_refs_data *cb = data;
847
848 if (!strbuf_realpath(&path, e->path, 0))
849 goto out;
850 if (!strbuf_strip_suffix(&path, "/objects"))
851 goto out;
852 base_len = path.len;
853
854 /* Is this a git repository with refs? */
855 strbuf_addstr(&path, "/refs");
856 if (!is_directory(path.buf))
857 goto out;
858 strbuf_setlen(&path, base_len);
859
860 read_alternate_refs(path.buf, cb->fn, cb->data);
861
862out:
863 strbuf_release(&path);
864 return 0;
865}
866
867void for_each_alternate_ref(alternate_ref_fn fn, void *data)
868{
869 struct alternate_refs_data cb;
870 cb.fn = fn;
871 cb.data = data;
872 foreach_alt_odb(refs_from_alternate_cb, &cb);
873}
874
fe1b2268 875int foreach_alt_odb(alt_odb_fn fn, void *cb)
d79796bc 876{
263db403 877 struct object_directory *ent;
fe1b2268 878 int r = 0;
d79796bc 879
0b209034 880 prepare_alt_odb(the_repository);
f0eaf638 881 for (ent = the_repository->objects->odb->next; ent; ent = ent->next) {
fe1b2268
JK
882 r = fn(ent, cb);
883 if (r)
884 break;
885 }
886 return r;
d79796bc
JH
887}
888
13068bf0 889void prepare_alt_odb(struct repository *r)
c2f493a4 890{
f0eaf638 891 if (r->objects->loaded_alternates)
7dc24aa5
SP
892 return;
893
13068bf0 894 link_alt_odb_entries(r, r->objects->alternate_db, PATH_SEP, NULL, 0);
c2f493a4 895
f0eaf638
JK
896 read_info_alternates(r, r->objects->odb->path, 0);
897 r->objects->loaded_alternates = 1;
c2f493a4
MW
898}
899
3096b2ec 900/* Returns 1 if we have successfully freshened the file, 0 otherwise. */
33d4221c 901static int freshen_file(const char *fn)
ace1534d 902{
312cd761 903 return !utime(fn, NULL);
0f4dc14a 904}
ace1534d 905
3096b2ec
JK
906/*
907 * All of the check_and_freshen functions return 1 if the file exists and was
908 * freshened (if freshening was requested), 0 otherwise. If they return
909 * 0, you should not assume that it is safe to skip a write of the object (it
910 * either does not exist on disk, or has a stale mtime and may be subject to
911 * pruning).
912 */
6a5e6f5e 913int check_and_freshen_file(const char *fn, int freshen)
33d4221c
JK
914{
915 if (access(fn, F_OK))
916 return 0;
3096b2ec 917 if (freshen && !freshen_file(fn))
33d4221c
JK
918 return 0;
919 return 1;
920}
921
f0eaf638
JK
922static int check_and_freshen_odb(struct object_directory *odb,
923 const struct object_id *oid,
924 int freshen)
33d4221c 925{
f0eaf638 926 static struct strbuf path = STRBUF_INIT;
514c5fdd 927 odb_loose_path(odb, &path, oid);
f0eaf638
JK
928 return check_and_freshen_file(path.buf, freshen);
929}
ea657730 930
f0eaf638
JK
931static int check_and_freshen_local(const struct object_id *oid, int freshen)
932{
933 return check_and_freshen_odb(the_repository->objects->odb, oid, freshen);
33d4221c
JK
934}
935
6862ebbf 936static int check_and_freshen_nonlocal(const struct object_id *oid, int freshen)
0f4dc14a 937{
263db403 938 struct object_directory *odb;
f3f043a1 939
0b209034 940 prepare_alt_odb(the_repository);
f0eaf638
JK
941 for (odb = the_repository->objects->odb->next; odb; odb = odb->next) {
942 if (check_and_freshen_odb(odb, oid, freshen))
c529d75a 943 return 1;
ace1534d 944 }
c529d75a 945 return 0;
ace1534d
JH
946}
947
6862ebbf 948static int check_and_freshen(const struct object_id *oid, int freshen)
33d4221c 949{
6862ebbf 950 return check_and_freshen_local(oid, freshen) ||
951 check_and_freshen_nonlocal(oid, freshen);
33d4221c
JK
952}
953
6862ebbf 954int has_loose_object_nonlocal(const struct object_id *oid)
33d4221c 955{
6862ebbf 956 return check_and_freshen_nonlocal(oid, 0);
33d4221c
JK
957}
958
6862ebbf 959static int has_loose_object(const struct object_id *oid)
0f4dc14a 960{
6862ebbf 961 return check_and_freshen(oid, 0);
0f4dc14a
BC
962}
963
02710228
SP
964static void mmap_limit_check(size_t length)
965{
966 static size_t limit = 0;
967 if (!limit) {
968 limit = git_env_ulong("GIT_MMAP_LIMIT", 0);
969 if (!limit)
970 limit = SIZE_MAX;
971 }
972 if (length > limit)
259328b7 973 die(_("attempting to mmap %"PRIuMAX" over limit %"PRIuMAX),
02710228
SP
974 (uintmax_t)length, (uintmax_t)limit);
975}
976
1570856b
JK
977void *xmmap_gently(void *start, size_t length,
978 int prot, int flags, int fd, off_t offset)
58ecbd5e 979{
02710228
SP
980 void *ret;
981
982 mmap_limit_check(length);
983 ret = mmap(start, length, prot, flags, fd, offset);
9827d4c1
JK
984 if (ret == MAP_FAILED && !length)
985 ret = NULL;
58ecbd5e
JN
986 return ret;
987}
988
dc059294
EW
989const char *mmap_os_err(void)
990{
991 static const char blank[] = "";
992#if defined(__linux__)
993 if (errno == ENOMEM) {
994 /* this continues an existing error message: */
995 static const char enomem[] =
996", check sys.vm.max_map_count and/or RLIMIT_DATA";
997 return enomem;
998 }
999#endif /* OS-specific bits */
1000 return blank;
1001}
1002
1570856b
JK
1003void *xmmap(void *start, size_t length,
1004 int prot, int flags, int fd, off_t offset)
1005{
1006 void *ret = xmmap_gently(start, length, prot, flags, fd, offset);
1007 if (ret == MAP_FAILED)
dc059294 1008 die_errno(_("mmap failed%s"), mmap_os_err());
1570856b
JK
1009 return ret;
1010}
1011
88d0db55 1012/*
090ea126 1013 * With an in-core object data in "map", rehash it to make sure the
cb1c8d1d
JK
1014 * object name actually matches "oid" to detect object corruption.
1015 * With "map" == NULL, try reading the object named with "oid" using
090ea126 1016 * the streaming interface and rehash it to do the same.
88d0db55 1017 */
b98d1885 1018int check_object_signature(struct repository *r, const struct object_id *oid,
96e41f58
ÆAB
1019 void *map, unsigned long size, const char *type,
1020 struct object_id *real_oidp)
88d0db55 1021{
96e41f58
ÆAB
1022 struct object_id tmp;
1023 struct object_id *real_oid = real_oidp ? real_oidp : &tmp;
090ea126
NTND
1024 enum object_type obj_type;
1025 struct git_istream *st;
18e2588e 1026 git_hash_ctx c;
1af64f73 1027 char hdr[MAX_HEADER_LEN];
090ea126 1028 int hdrlen;
88d0db55 1029
090ea126 1030 if (map) {
96e41f58
ÆAB
1031 hash_object_file(r->hash_algo, map, size, type, real_oid);
1032 return !oideq(oid, real_oid) ? -1 : 0;
88d0db55
BC
1033 }
1034
b98d1885 1035 st = open_istream(r, oid, &obj_type, &size, NULL);
090ea126
NTND
1036 if (!st)
1037 return -1;
88d0db55 1038
090ea126 1039 /* Generate the header */
ca473cef 1040 hdrlen = xsnprintf(hdr, sizeof(hdr), "%s %"PRIuMAX , type_name(obj_type), (uintmax_t)size) + 1;
88d0db55 1041
090ea126 1042 /* Sha1.. */
b98d1885
MT
1043 r->hash_algo->init_fn(&c);
1044 r->hash_algo->update_fn(&c, hdr, hdrlen);
090ea126
NTND
1045 for (;;) {
1046 char buf[1024 * 16];
1047 ssize_t readlen = read_istream(st, buf, sizeof(buf));
1f688557 1048
f54fac53
JK
1049 if (readlen < 0) {
1050 close_istream(st);
1051 return -1;
1052 }
090ea126
NTND
1053 if (!readlen)
1054 break;
b98d1885 1055 r->hash_algo->update_fn(&c, buf, readlen);
fa5fc15d 1056 }
96e41f58 1057 r->hash_algo->final_oid_fn(real_oid, &c);
090ea126 1058 close_istream(st);
96e41f58 1059 return !oideq(oid, real_oid) ? -1 : 0;
fa5fc15d
SP
1060}
1061
1b8ac5ea 1062int git_open_cloexec(const char *name, int flags)
a0788266 1063{
1e3001a8
JH
1064 int fd;
1065 static int o_cloexec = O_CLOEXEC;
a0788266 1066
1e3001a8
JH
1067 fd = open(name, flags | o_cloexec);
1068 if ((o_cloexec & O_CLOEXEC) && fd < 0 && errno == EINVAL) {
cd66ada0 1069 /* Try again w/o O_CLOEXEC: the kernel might not support it */
1e3001a8
JH
1070 o_cloexec &= ~O_CLOEXEC;
1071 fd = open(name, flags | o_cloexec);
491a8dec 1072 }
491a8dec 1073
9fb9495d 1074#if defined(F_GETFD) && defined(F_SETFD) && defined(FD_CLOEXEC)
491a8dec 1075 {
1e3001a8 1076 static int fd_cloexec = FD_CLOEXEC;
a0788266 1077
1e3001a8
JH
1078 if (!o_cloexec && 0 <= fd && fd_cloexec) {
1079 /* Opened w/o O_CLOEXEC? try with fcntl(2) to add it */
9fb9495d
EW
1080 int flags = fcntl(fd, F_GETFD);
1081 if (fcntl(fd, F_SETFD, flags | fd_cloexec))
1e3001a8 1082 fd_cloexec = 0;
cd66ada0 1083 }
44d1c19e 1084 }
a0788266 1085#endif
1b8ac5ea 1086 return fd;
a0788266
JS
1087}
1088
3cf8b462 1089/*
514c5fdd 1090 * Find "oid" as a loose object in the local repository or in an alternate.
771e7d57
JK
1091 * Returns 0 on success, negative on failure.
1092 *
1093 * The "path" out-parameter will give the path of the object we found (if any).
1094 * Note that it may point to static storage and is only valid until another
514c5fdd 1095 * call to stat_loose_object().
3cf8b462 1096 */
514c5fdd
JK
1097static int stat_loose_object(struct repository *r, const struct object_id *oid,
1098 struct stat *st, const char **path)
1f688557 1099{
263db403 1100 struct object_directory *odb;
ea657730
CC
1101 static struct strbuf buf = STRBUF_INIT;
1102
d2607fa0 1103 prepare_alt_odb(r);
f0eaf638 1104 for (odb = r->objects->odb; odb; odb = odb->next) {
514c5fdd 1105 *path = odb_loose_path(odb, &buf, oid);
771e7d57 1106 if (!lstat(*path, st))
052fe5ea 1107 return 0;
c7934306
SP
1108 }
1109
3cf8b462
SP
1110 return -1;
1111}
1112
771e7d57 1113/*
514c5fdd 1114 * Like stat_loose_object(), but actually open the object and return the
771e7d57
JK
1115 * descriptor. See the caveats on the "path" parameter above.
1116 */
514c5fdd
JK
1117static int open_loose_object(struct repository *r,
1118 const struct object_id *oid, const char **path)
60bb8b14 1119{
44d1c19e 1120 int fd;
263db403 1121 struct object_directory *odb;
f0eaf638 1122 int most_interesting_errno = ENOENT;
ea657730
CC
1123 static struct strbuf buf = STRBUF_INIT;
1124
ec7283e5 1125 prepare_alt_odb(r);
f0eaf638 1126 for (odb = r->objects->odb; odb; odb = odb->next) {
514c5fdd 1127 *path = odb_loose_path(odb, &buf, oid);
771e7d57 1128 fd = git_open(*path);
44d1c19e
LT
1129 if (fd >= 0)
1130 return fd;
f0eaf638 1131
d6c8a05b
JK
1132 if (most_interesting_errno == ENOENT)
1133 most_interesting_errno = errno;
03e79c88 1134 }
d6c8a05b 1135 errno = most_interesting_errno;
44d1c19e 1136 return -1;
1f688557
JH
1137}
1138
61c7711c 1139static int quick_has_loose(struct repository *r,
d7a24573 1140 const struct object_id *oid)
61c7711c 1141{
61c7711c
JK
1142 struct object_directory *odb;
1143
61c7711c
JK
1144 prepare_alt_odb(r);
1145 for (odb = r->objects->odb; odb; odb = odb->next) {
92d8ed8a 1146 if (oidtree_contains(odb_loose_cache(odb, oid), oid))
61c7711c
JK
1147 return 1;
1148 }
1149 return 0;
1150}
1151
f6371f92
JK
1152/*
1153 * Map the loose object at "path" if it is not NULL, or the path found by
514c5fdd 1154 * searching for a loose object named "oid".
f6371f92 1155 */
514c5fdd
JK
1156static void *map_loose_object_1(struct repository *r, const char *path,
1157 const struct object_id *oid, unsigned long *size)
27d69a46 1158{
0fcfd160 1159 void *map;
144bde78 1160 int fd;
84dd81c1 1161
f6371f92
JK
1162 if (path)
1163 fd = git_open(path);
8261e1f1 1164 else
514c5fdd 1165 fd = open_loose_object(r, oid, &path);
44d1c19e
LT
1166 map = NULL;
1167 if (fd >= 0) {
1168 struct stat st;
1f688557 1169
44d1c19e
LT
1170 if (!fstat(fd, &st)) {
1171 *size = xsize_t(st.st_size);
33e42de0
MM
1172 if (!*size) {
1173 /* mmap() is forbidden on empty files */
259328b7 1174 error(_("object file %s is empty"), path);
6881925e 1175 close(fd);
33e42de0
MM
1176 return NULL;
1177 }
44d1c19e 1178 map = xmmap(NULL, *size, PROT_READ, MAP_PRIVATE, fd, 0);
144bde78 1179 }
44d1c19e 1180 close(fd);
74e34e1f 1181 }
0fcfd160 1182 return map;
74e34e1f
NP
1183}
1184
514c5fdd
JK
1185void *map_loose_object(struct repository *r,
1186 const struct object_id *oid,
1187 unsigned long *size)
068f85e3 1188{
514c5fdd 1189 return map_loose_object_1(r, NULL, oid, size);
068f85e3 1190}
1191
3b6a8db3
ÆAB
1192enum unpack_loose_header_result unpack_loose_header(git_zstream *stream,
1193 unsigned char *map,
1194 unsigned long mapsize,
1195 void *buffer,
1196 unsigned long bufsiz,
1197 struct strbuf *header)
47fe3f6e 1198{
01cab976 1199 int status;
31877c9a 1200
c4483576
LT
1201 /* Get the data stream */
1202 memset(stream, 0, sizeof(*stream));
1203 stream->next_in = map;
1204 stream->avail_in = mapsize;
1205 stream->next_out = buffer;
93821bd9 1206 stream->avail_out = bufsiz;
47fe3f6e 1207
39c68542 1208 git_inflate_init(stream);
31877c9a 1209 obj_read_unlock();
01cab976 1210 status = git_inflate(stream, 0);
31877c9a 1211 obj_read_lock();
d21f8426 1212 if (status < Z_OK)
3b6a8db3 1213 return ULHR_BAD;
d131b7af 1214
46f03448
KN
1215 /*
1216 * Check if entire header is unpacked in the first iteration.
d131b7af 1217 */
46f03448 1218 if (memchr(buffer, '\0', stream->next_out - (unsigned char *)buffer))
3b6a8db3 1219 return ULHR_OK;
d131b7af 1220
01cab976
ÆAB
1221 /*
1222 * We have a header longer than MAX_HEADER_LEN. The "header"
1223 * here is only non-NULL when we run "cat-file
1224 * --allow-unknown-type".
1225 */
1226 if (!header)
5848fb11 1227 return ULHR_TOO_LONG;
d131b7af 1228
46f03448
KN
1229 /*
1230 * buffer[0..bufsiz] was not large enough. Copy the partial
1231 * result out to header, and then append the result of further
1232 * reading the stream.
1233 */
1234 strbuf_add(header, buffer, stream->next_out - (unsigned char *)buffer);
1235 stream->next_out = buffer;
1236 stream->avail_out = bufsiz;
d131b7af 1237
46f03448 1238 do {
31877c9a 1239 obj_read_unlock();
46f03448 1240 status = git_inflate(stream, 0);
31877c9a 1241 obj_read_lock();
46f03448
KN
1242 strbuf_add(header, buffer, stream->next_out - (unsigned char *)buffer);
1243 if (memchr(buffer, '\0', stream->next_out - (unsigned char *)buffer))
1244 return 0;
1245 stream->next_out = buffer;
1246 stream->avail_out = bufsiz;
1247 } while (status != Z_STREAM_END);
5848fb11 1248 return ULHR_TOO_LONG;
d131b7af
SP
1249}
1250
00a7760e
JK
1251static void *unpack_loose_rest(git_zstream *stream,
1252 void *buffer, unsigned long size,
1253 const struct object_id *oid)
95099731 1254{
5180cacc 1255 int bytes = strlen(buffer) + 1;
3aee68aa 1256 unsigned char *buf = xmallocz(size);
93821bd9 1257 unsigned long n;
7efbff75 1258 int status = Z_OK;
95099731 1259
93821bd9
LT
1260 n = stream->total_out - bytes;
1261 if (n > size)
1262 n = size;
1263 memcpy(buf, (char *) buffer + bytes, n);
1264 bytes = n;
456cdf6e
LT
1265 if (bytes <= size) {
1266 /*
1267 * The above condition must be (bytes <= size), not
1268 * (bytes < size). In other words, even though we
ccf5ace0 1269 * expect no more output and set avail_out to zero,
456cdf6e
LT
1270 * the input zlib stream may have bytes that express
1271 * "this concludes the stream", and we *do* want to
1272 * eat that input.
1273 *
1274 * Otherwise we would not be able to test that we
1275 * consumed all the input to reach the expected size;
1276 * we also want to check that zlib tells us that all
1277 * went well with status == Z_STREAM_END at the end.
1278 */
5180cacc
LT
1279 stream->next_out = buf + bytes;
1280 stream->avail_out = size - bytes;
31877c9a
MT
1281 while (status == Z_OK) {
1282 obj_read_unlock();
39c68542 1283 status = git_inflate(stream, Z_FINISH);
31877c9a
MT
1284 obj_read_lock();
1285 }
5180cacc 1286 }
456cdf6e 1287 if (status == Z_STREAM_END && !stream->avail_in) {
39c68542 1288 git_inflate_end(stream);
7efbff75 1289 return buf;
95099731
NTND
1290 }
1291
7efbff75 1292 if (status < 0)
00a7760e 1293 error(_("corrupt loose object '%s'"), oid_to_hex(oid));
7efbff75 1294 else if (stream->avail_in)
259328b7 1295 error(_("garbage at end of loose object '%s'"),
00a7760e 1296 oid_to_hex(oid));
7efbff75
JH
1297 free(buf);
1298 return NULL;
95099731
NTND
1299}
1300
d40d535b 1301/*
5180cacc
LT
1302 * We used to just use "sscanf()", but that's actually way
1303 * too permissive for what we want to check. So do an anal
1304 * object header parse by hand.
d40d535b 1305 */
dccb32bf 1306int parse_loose_header(const char *hdr, struct object_info *oi)
1f688557 1307{
46f03448 1308 const char *type_buf = hdr;
5180cacc 1309 unsigned long size;
46f03448 1310 int type, type_len = 0;
43057304 1311
5180cacc 1312 /*
46f03448 1313 * The type can be of any size but is followed by
21666f1a 1314 * a space.
5180cacc 1315 */
5180cacc
LT
1316 for (;;) {
1317 char c = *hdr++;
d21f8426
JH
1318 if (!c)
1319 return -1;
5180cacc
LT
1320 if (c == ' ')
1321 break;
46f03448 1322 type_len++;
5180cacc 1323 }
1f688557 1324
46f03448 1325 type = type_from_string_gently(type_buf, type_len, 1);
6ca32f47
BW
1326 if (oi->type_name)
1327 strbuf_add(oi->type_name, type_buf, type_len);
46f03448
KN
1328 if (oi->typep)
1329 *oi->typep = type;
5180cacc
LT
1330
1331 /*
1332 * The length must follow immediately, and be in canonical
1333 * decimal format (ie "010" is not valid).
1334 */
1335 size = *hdr++ - '0';
1336 if (size > 9)
1337 return -1;
1338 if (size) {
1339 for (;;) {
1340 unsigned long c = *hdr - '0';
1341 if (c > 9)
1342 break;
1343 hdr++;
1344 size = size * 10 + c;
1b1005d1 1345 }
c01f51cc 1346 }
46f03448
KN
1347
1348 if (oi->sizep)
1349 *oi->sizep = size;
5180cacc
LT
1350
1351 /*
1352 * The length must be followed by a zero byte
1353 */
dccb32bf
ÆAB
1354 if (*hdr)
1355 return -1;
21666f1a 1356
dccb32bf
ÆAB
1357 /*
1358 * The format is valid, but the type may still be bogus. The
1359 * Caller needs to check its oi->typep.
1360 */
1361 return 0;
bf592c50
DB
1362}
1363
514c5fdd
JK
1364static int loose_object_info(struct repository *r,
1365 const struct object_id *oid,
1366 struct object_info *oi, int flags)
65c2e0c3 1367{
46f03448
KN
1368 int status = 0;
1369 unsigned long mapsize;
65c2e0c3 1370 void *map;
ef49a7a0 1371 git_zstream stream;
1af64f73 1372 char hdr[MAX_HEADER_LEN];
46f03448 1373 struct strbuf hdrbuf = STRBUF_INIT;
c84a1f3e 1374 unsigned long size_scratch;
dccb32bf 1375 enum object_type type_scratch;
01cab976 1376 int allow_unknown = flags & OBJECT_INFO_ALLOW_UNKNOWN_TYPE;
65c2e0c3 1377
b99b6bcc
JK
1378 if (oi->delta_base_oid)
1379 oidclr(oi->delta_base_oid);
5d642e75 1380
052fe5ea
JK
1381 /*
1382 * If we don't care about type or size, then we don't
4ef8d1dd
JH
1383 * need to look inside the object at all. Note that we
1384 * do not optimize out the stat call, even if the
1385 * caller doesn't care about the disk-size, since our
1386 * return value implicitly indicates whether the
1387 * object even exists.
052fe5ea 1388 */
6ca32f47 1389 if (!oi->typep && !oi->type_name && !oi->sizep && !oi->contentp) {
771e7d57 1390 const char *path;
4ef8d1dd 1391 struct stat st;
61c7711c 1392 if (!oi->disk_sizep && (flags & OBJECT_INFO_QUICK))
d7a24573 1393 return quick_has_loose(r, oid) ? 0 : -1;
514c5fdd 1394 if (stat_loose_object(r, oid, &st, &path) < 0)
4ef8d1dd
JH
1395 return -1;
1396 if (oi->disk_sizep)
23c339c0 1397 *oi->disk_sizep = st.st_size;
052fe5ea
JK
1398 return 0;
1399 }
1400
514c5fdd 1401 map = map_loose_object(r, oid, &mapsize);
f0df4ed5 1402 if (!map)
dbea72a8 1403 return -1;
c84a1f3e
JT
1404
1405 if (!oi->sizep)
1406 oi->sizep = &size_scratch;
dccb32bf
ÆAB
1407 if (!oi->typep)
1408 oi->typep = &type_scratch;
c84a1f3e 1409
23c339c0
JK
1410 if (oi->disk_sizep)
1411 *oi->disk_sizep = mapsize;
01cab976 1412
3b6a8db3
ÆAB
1413 switch (unpack_loose_header(&stream, map, mapsize, hdr, sizeof(hdr),
1414 allow_unknown ? &hdrbuf : NULL)) {
1415 case ULHR_OK:
dccb32bf
ÆAB
1416 if (parse_loose_header(hdrbuf.len ? hdrbuf.buf : hdr, oi) < 0)
1417 status = error(_("unable to parse %s header"), oid_to_hex(oid));
1418 else if (!allow_unknown && *oi->typep < 0)
1419 die(_("invalid object type"));
1420
1421 if (!oi->contentp)
1422 break;
1423 *oi->contentp = unpack_loose_rest(&stream, hdr, *oi->sizep, oid);
1424 if (*oi->contentp)
1425 goto cleanup;
1426
1427 status = -1;
3b6a8db3
ÆAB
1428 break;
1429 case ULHR_BAD:
259328b7 1430 status = error(_("unable to unpack %s header"),
514c5fdd 1431 oid_to_hex(oid));
3b6a8db3 1432 break;
5848fb11
ÆAB
1433 case ULHR_TOO_LONG:
1434 status = error(_("header for %s too long, exceeds %d bytes"),
1435 oid_to_hex(oid), MAX_HEADER_LEN);
1436 break;
3b6a8db3 1437 }
c84a1f3e 1438
dccb32bf
ÆAB
1439 git_inflate_end(&stream);
1440cleanup:
65c2e0c3 1441 munmap(map, mapsize);
c84a1f3e
JT
1442 if (oi->sizep == &size_scratch)
1443 oi->sizep = NULL;
46f03448 1444 strbuf_release(&hdrbuf);
dccb32bf
ÆAB
1445 if (oi->typep == &type_scratch)
1446 oi->typep = NULL;
3ab0fb06 1447 oi->whence = OI_LOOSE;
dccb32bf 1448 return status;
65c2e0c3
JH
1449}
1450
31877c9a
MT
1451int obj_read_use_lock = 0;
1452pthread_mutex_t obj_read_mutex;
1453
1454void enable_obj_read_lock(void)
1455{
1456 if (obj_read_use_lock)
1457 return;
1458
1459 obj_read_use_lock = 1;
1460 init_recursive_mutex(&obj_read_mutex);
1461}
1462
1463void disable_obj_read_lock(void)
1464{
1465 if (!obj_read_use_lock)
1466 return;
1467
1468 obj_read_use_lock = 0;
1469 pthread_mutex_destroy(&obj_read_mutex);
1470}
1471
8b4c0103
JT
1472int fetch_if_missing = 1;
1473
31877c9a
MT
1474static int do_oid_object_info_extended(struct repository *r,
1475 const struct object_id *oid,
1476 struct object_info *oi, unsigned flags)
f0df4ed5 1477{
cd585e2a 1478 static struct object_info blank_oi = OBJECT_INFO_INIT;
9c8a294a 1479 struct cached_object *co;
f0df4ed5 1480 struct pack_entry e;
5b086407 1481 int rtype;
b383a13c 1482 const struct object_id *real = oid;
8b4c0103 1483 int already_retried = 0;
f0df4ed5 1484
31877c9a 1485
b383a13c 1486 if (flags & OBJECT_INFO_LOOKUP_REPLACE)
9d98354f 1487 real = lookup_replace_object(r, oid);
f0df4ed5 1488
b383a13c 1489 if (is_null_oid(real))
87b5e236
JK
1490 return -1;
1491
cd585e2a
JT
1492 if (!oi)
1493 oi = &blank_oi;
1494
9c8a294a
JT
1495 co = find_cached_object(real);
1496 if (co) {
1497 if (oi->typep)
1498 *(oi->typep) = co->type;
1499 if (oi->sizep)
1500 *(oi->sizep) = co->size;
1501 if (oi->disk_sizep)
1502 *(oi->disk_sizep) = 0;
b99b6bcc
JK
1503 if (oi->delta_base_oid)
1504 oidclr(oi->delta_base_oid);
9c8a294a
JT
1505 if (oi->type_name)
1506 strbuf_addstr(oi->type_name, type_name(co->type));
1507 if (oi->contentp)
1508 *oi->contentp = xmemdupz(co->buf, co->size);
1509 oi->whence = OI_CACHED;
1510 return 0;
c4d9986f
NTND
1511 }
1512
8b4c0103 1513 while (1) {
42c8ce1c 1514 if (find_pack_entry(r, real, &e))
8b4c0103
JT
1515 break;
1516
024aa469
TI
1517 if (flags & OBJECT_INFO_IGNORE_LOOSE)
1518 return -1;
1519
ddd63e64 1520 /* Most likely it's a loose object. */
514c5fdd 1521 if (!loose_object_info(r, real, oi, flags))
5b086407 1522 return 0;
ddd63e64
SG
1523
1524 /* Not a loose object; someone else may have just packed it. */
2b7750c9 1525 if (!(flags & OBJECT_INFO_QUICK)) {
9d98354f 1526 reprepare_packed_git(r);
42c8ce1c 1527 if (find_pack_entry(r, real, &e))
2b7750c9
JT
1528 break;
1529 }
8b4c0103 1530
eef71904
JT
1531 /*
1532 * If r is the_repository, this might be an attempt at
1533 * accessing a submodule object as if it were in the_repository
1534 * (having called add_submodule_odb() on that submodule's ODB).
1535 * If any such ODBs exist, register them and try again.
1536 */
1537 if (r == the_repository &&
1538 register_all_submodule_odb_as_alternates())
a35e03de
JT
1539 /* We added some alternates; retry */
1540 continue;
1541
8b4c0103 1542 /* Check if it is a missing object */
ef830cc4
JT
1543 if (fetch_if_missing && repo_has_promisor_remote(r) &&
1544 !already_retried &&
31f5256c 1545 !(flags & OBJECT_INFO_SKIP_FETCH_OBJECT)) {
8b4c0103 1546 /*
b14ed5ad
CC
1547 * TODO Investigate checking promisor_remote_get_direct()
1548 * TODO return value and stopping on error here.
8b4c0103 1549 */
b14ed5ad 1550 promisor_remote_get_direct(r, real, 1);
8b4c0103
JT
1551 already_retried = 1;
1552 continue;
dfdd4afc 1553 }
8b4c0103
JT
1554
1555 return -1;
f0df4ed5 1556 }
3d77d877 1557
cd585e2a
JT
1558 if (oi == &blank_oi)
1559 /*
1560 * We know that the caller doesn't actually need the
1561 * information below, so return early.
1562 */
1563 return 0;
9d98354f 1564 rtype = packed_object_info(r, e.p, e.offset, oi);
412916ee 1565 if (rtype < 0) {
751530de 1566 mark_bad_packed_object(e.p, real);
31877c9a 1567 return do_oid_object_info_extended(r, real, oi, 0);
3ab0fb06 1568 } else if (oi->whence == OI_PACKED) {
9a490590
JH
1569 oi->u.packed.offset = e.offset;
1570 oi->u.packed.pack = e.p;
1571 oi->u.packed.is_delta = (rtype == OBJ_REF_DELTA ||
1572 rtype == OBJ_OFS_DELTA);
3d77d877
NP
1573 }
1574
5b086407 1575 return 0;
f0df4ed5
JS
1576}
1577
31877c9a
MT
1578int oid_object_info_extended(struct repository *r, const struct object_id *oid,
1579 struct object_info *oi, unsigned flags)
1580{
1581 int ret;
1582 obj_read_lock();
1583 ret = do_oid_object_info_extended(r, oid, oi, flags);
1584 obj_read_unlock();
1585 return ret;
1586}
1587
1588
3fc0dca9 1589/* returns enum object_type or negative */
9d98354f
SB
1590int oid_object_info(struct repository *r,
1591 const struct object_id *oid,
1592 unsigned long *sizep)
9a490590 1593{
5b086407 1594 enum object_type type;
27b5c1a0 1595 struct object_info oi = OBJECT_INFO_INIT;
9a490590 1596
5b086407 1597 oi.typep = &type;
9a490590 1598 oi.sizep = sizep;
9d98354f
SB
1599 if (oid_object_info_extended(r, oid, &oi,
1600 OBJECT_INFO_LOOKUP_REPLACE) < 0)
5b086407
JK
1601 return -1;
1602 return type;
9a490590
JH
1603}
1604
1b9b5c69 1605static void *read_object(struct repository *r,
cba595ab 1606 const struct object_id *oid, enum object_type *type,
f1d8130b
JT
1607 unsigned long *size)
1608{
1609 struct object_info oi = OBJECT_INFO_INIT;
1610 void *content;
1611 oi.typep = type;
1612 oi.sizep = size;
1613 oi.contentp = &content;
1614
cba595ab 1615 if (oid_object_info_extended(r, oid, &oi, 0) < 0)
f1d8130b
JT
1616 return NULL;
1617 return content;
1618}
1619
829e5c3b
PO
1620int pretend_object_file(void *buf, unsigned long len, enum object_type type,
1621 struct object_id *oid)
d66b37bb
JH
1622{
1623 struct cached_object *co;
1624
2dcde20e 1625 hash_object_file(the_hash_algo, buf, len, type_name(type), oid);
a64d2aae
JT
1626 if (has_object_file_with_flags(oid, OBJECT_INFO_QUICK | OBJECT_INFO_SKIP_FETCH_OBJECT) ||
1627 find_cached_object(oid))
d66b37bb 1628 return 0;
c7353967 1629 ALLOC_GROW(cached_objects, cached_object_nr + 1, cached_object_alloc);
d66b37bb
JH
1630 co = &cached_objects[cached_object_nr++];
1631 co->size = len;
21666f1a 1632 co->type = type;
efa13f7b
JH
1633 co->buf = xmalloc(len);
1634 memcpy(co->buf, buf, len);
62ba93ea 1635 oidcpy(&co->oid, oid);
d66b37bb
JH
1636 return 0;
1637}
1638
b6c4cecc
JH
1639/*
1640 * This function dies on corrupt objects; the callers who want to
1641 * deal with them should arrange to call read_object() and give error
1642 * messages themselves.
1643 */
a3b72c89
SB
1644void *read_object_file_extended(struct repository *r,
1645 const struct object_id *oid,
b4f5aca4 1646 enum object_type *type,
1647 unsigned long *size,
1648 int lookup_replace)
ac939109 1649{
3ba7a065 1650 void *data;
b6c4cecc 1651 const struct packed_git *p;
771e7d57
JK
1652 const char *path;
1653 struct stat st;
1f2e7cea 1654 const struct object_id *repl = lookup_replace ?
a3b72c89 1655 lookup_replace_object(r, oid) : oid;
b6c4cecc 1656
3ba7a065 1657 errno = 0;
cba595ab 1658 data = read_object(r, repl, type, size);
4bbf5a26 1659 if (data)
b6c4cecc 1660 return data;
68095570 1661
31877c9a 1662 obj_read_lock();
25f3af3f 1663 if (errno && errno != ENOENT)
259328b7 1664 die_errno(_("failed to read object %s"), oid_to_hex(oid));
3ba7a065 1665
68095570 1666 /* die if we replaced an object with one that does not exist */
b383a13c 1667 if (repl != oid)
259328b7 1668 die(_("replacement %s not found for %s"),
b383a13c 1669 oid_to_hex(repl), oid_to_hex(oid));
68095570 1670
cba595ab 1671 if (!stat_loose_object(r, repl, &st, &path))
259328b7 1672 die(_("loose object %s (stored in %s) is corrupt"),
b383a13c 1673 oid_to_hex(repl), path);
68095570 1674
afe8a907 1675 if ((p = has_packed_and_bad(r, repl)))
259328b7 1676 die(_("packed object %s (stored in %s) is corrupt"),
b383a13c 1677 oid_to_hex(repl), p->pack_name);
31877c9a 1678 obj_read_unlock();
f5552aee 1679
b6c4cecc 1680 return NULL;
ac939109
NP
1681}
1682
d3b4705a
NTND
1683void *read_object_with_reference(struct repository *r,
1684 const struct object_id *oid,
21666f1a 1685 const char *required_type_name,
40469ee9 1686 unsigned long *size,
02f0547e 1687 struct object_id *actual_oid_return)
f4913f91 1688{
21666f1a 1689 enum object_type type, required_type;
f4913f91
JH
1690 void *buffer;
1691 unsigned long isize;
02f0547e 1692 struct object_id actual_oid;
f4913f91 1693
21666f1a 1694 required_type = type_from_string(required_type_name);
02f0547e 1695 oidcpy(&actual_oid, oid);
40469ee9
JH
1696 while (1) {
1697 int ref_length = -1;
1698 const char *ref_type = NULL;
f4913f91 1699
d3b4705a 1700 buffer = repo_read_object_file(r, &actual_oid, &type, &isize);
40469ee9
JH
1701 if (!buffer)
1702 return NULL;
21666f1a 1703 if (type == required_type) {
40469ee9 1704 *size = isize;
02f0547e 1705 if (actual_oid_return)
1706 oidcpy(actual_oid_return, &actual_oid);
40469ee9
JH
1707 return buffer;
1708 }
1709 /* Handle references */
21666f1a 1710 else if (type == OBJ_COMMIT)
40469ee9 1711 ref_type = "tree ";
21666f1a 1712 else if (type == OBJ_TAG)
40469ee9
JH
1713 ref_type = "object ";
1714 else {
1715 free(buffer);
1716 return NULL;
1717 }
1718 ref_length = strlen(ref_type);
f4913f91 1719
94b5e093 1720 if (ref_length + the_hash_algo->hexsz > isize ||
50974ec9 1721 memcmp(buffer, ref_type, ref_length) ||
02f0547e 1722 get_oid_hex((char *) buffer + ref_length, &actual_oid)) {
40469ee9
JH
1723 free(buffer);
1724 return NULL;
1725 }
1cf58e72 1726 free(buffer);
40469ee9 1727 /* Now we have the ID of the referred-to object in
02f0547e 1728 * actual_oid. Check again. */
f4913f91 1729 }
f4913f91
JH
1730}
1731
7ad5c44d
MT
1732static void write_object_file_prepare(const struct git_hash_algo *algo,
1733 const void *buf, unsigned long len,
a09c985e
PO
1734 const char *type, struct object_id *oid,
1735 char *hdr, int *hdrlen)
d410c0f5 1736{
18e2588e 1737 git_hash_ctx c;
d410c0f5
JH
1738
1739 /* Generate the header */
ca473cef 1740 *hdrlen = xsnprintf(hdr, *hdrlen, "%s %"PRIuMAX , type, (uintmax_t)len)+1;
d410c0f5
JH
1741
1742 /* Sha1.. */
7ad5c44d
MT
1743 algo->init_fn(&c);
1744 algo->update_fn(&c, hdr, *hdrlen);
1745 algo->update_fn(&c, buf, len);
5951bf46 1746 algo->final_oid_fn(oid, &c);
d410c0f5
JH
1747}
1748
230f1322 1749/*
5a688fe4 1750 * Move the just written object into its final resting place.
230f1322 1751 */
cb5add58 1752int finalize_object_file(const char *tmpfile, const char *filename)
230f1322 1753{
e32c0a9c 1754 int ret = 0;
5a688fe4 1755
348df166 1756 if (object_creation_mode == OBJECT_CREATION_USES_RENAMES)
be66a6c4
JS
1757 goto try_rename;
1758 else if (link(tmpfile, filename))
e32c0a9c 1759 ret = errno;
7ebb6fca
LT
1760
1761 /*
1762 * Coda hack - coda doesn't like cross-directory links,
1763 * so we fall back to a rename, which will mean that it
1764 * won't be able to check collisions, but that's not a
1765 * big deal.
1766 *
1767 * The same holds for FAT formatted media.
1768 *
3be1f18e 1769 * When this succeeds, we just return. We have nothing
7ebb6fca
LT
1770 * left to unlink.
1771 */
1772 if (ret && ret != EEXIST) {
be66a6c4 1773 try_rename:
7ebb6fca 1774 if (!rename(tmpfile, filename))
3be1f18e 1775 goto out;
9e48b389 1776 ret = errno;
230f1322 1777 }
691f1a28 1778 unlink_or_warn(tmpfile);
230f1322
LT
1779 if (ret) {
1780 if (ret != EEXIST) {
2c319886 1781 return error_errno(_("unable to write file %s"), filename);
230f1322
LT
1782 }
1783 /* FIXME!!! Collision check here ? */
1784 }
1785
3be1f18e 1786out:
5256b006 1787 if (adjust_shared_perm(filename))
259328b7 1788 return error(_("unable to set permission to '%s'"), filename);
230f1322
LT
1789 return 0;
1790}
1791
4d548150
LT
1792static int write_buffer(int fd, const void *buf, size_t len)
1793{
d34cf19b 1794 if (write_in_full(fd, buf, len) < 0)
259328b7 1795 return error_errno(_("file write error"));
4d548150
LT
1796 return 0;
1797}
1798
2dcde20e
MT
1799int hash_object_file(const struct git_hash_algo *algo, const void *buf,
1800 unsigned long len, const char *type,
f070facc 1801 struct object_id *oid)
abdc3fc8 1802{
1af64f73 1803 char hdr[MAX_HEADER_LEN];
ef1286d3 1804 int hdrlen = sizeof(hdr);
2dcde20e 1805 write_object_file_prepare(algo, buf, len, type, oid, hdr, &hdrlen);
abdc3fc8
RS
1806 return 0;
1807}
1808
e9039dd3 1809/* Finalize a file on disk, and close it. */
514c5fdd 1810static void close_loose_object(int fd)
e9039dd3 1811{
aafe9fba 1812 if (fsync_object_files)
76011357 1813 fsync_or_die(fd, "loose object file");
e9039dd3 1814 if (close(fd) != 0)
76011357 1815 die_errno(_("error when closing loose object file"));
e9039dd3
LT
1816}
1817
5723fe7e
LT
1818/* Size of directory component, including the ending '/' */
1819static inline int directory_size(const char *filename)
1820{
1821 const char *s = strrchr(filename, '/');
1822 if (!s)
1823 return 0;
1824 return s - filename + 1;
1825}
1826
1827/*
1828 * This creates a temporary file in the same directory as the final
1829 * 'filename'
1830 *
1831 * We want to avoid cross-directory filename renames, because those
1832 * can have problems on various filesystems (FAT, NFS, Coda).
1833 */
d4b3d11a 1834static int create_tmpfile(struct strbuf *tmp, const char *filename)
5723fe7e
LT
1835{
1836 int fd, dirlen = directory_size(filename);
1837
d4b3d11a
JK
1838 strbuf_reset(tmp);
1839 strbuf_add(tmp, filename, dirlen);
1840 strbuf_addstr(tmp, "tmp_obj_XXXXXX");
1841 fd = git_mkstemp_mode(tmp->buf, 0444);
cbacbf4e 1842 if (fd < 0 && dirlen && errno == ENOENT) {
d4b3d11a
JK
1843 /*
1844 * Make sure the directory exists; note that the contents
1845 * of the buffer are undefined after mkstemp returns an
1846 * error, so we have to rewrite the whole buffer from
1847 * scratch.
1848 */
1849 strbuf_reset(tmp);
1850 strbuf_add(tmp, filename, dirlen - 1);
1851 if (mkdir(tmp->buf, 0777) && errno != EEXIST)
b2476a60 1852 return -1;
d4b3d11a 1853 if (adjust_shared_perm(tmp->buf))
5723fe7e
LT
1854 return -1;
1855
1856 /* Try again */
d4b3d11a
JK
1857 strbuf_addstr(tmp, "/tmp_obj_XXXXXX");
1858 fd = git_mkstemp_mode(tmp->buf, 0444);
5723fe7e
LT
1859 }
1860 return fd;
1861}
1862
3fc7281f
PO
1863static int write_loose_object(const struct object_id *oid, char *hdr,
1864 int hdrlen, const void *buf, unsigned long len,
4ef91a2d 1865 time_t mtime, unsigned flags)
0fcfd160 1866{
915308b1 1867 int fd, ret;
9892beba 1868 unsigned char compressed[4096];
ef49a7a0 1869 git_zstream stream;
18e2588e 1870 git_hash_ctx c;
3fc7281f 1871 struct object_id parano_oid;
d4b3d11a 1872 static struct strbuf tmp_file = STRBUF_INIT;
ea657730
CC
1873 static struct strbuf filename = STRBUF_INIT;
1874
514c5fdd 1875 loose_object_path(the_repository, &filename, oid);
a44c9a5e 1876
ea657730 1877 fd = create_tmpfile(&tmp_file, filename.buf);
aac17941 1878 if (fd < 0) {
4ef91a2d
ÆAB
1879 if (flags & HASH_SILENT)
1880 return -1;
1881 else if (errno == EACCES)
259328b7 1882 return error(_("insufficient permission for adding an object to repository database %s"), get_object_directory());
916d081b 1883 else
259328b7 1884 return error_errno(_("unable to create temporary file"));
aac17941
LT
1885 }
1886
0fcfd160 1887 /* Set it up */
55bb5c91 1888 git_deflate_init(&stream, zlib_compression_level);
0fcfd160 1889 stream.next_out = compressed;
9892beba 1890 stream.avail_out = sizeof(compressed);
18e2588e 1891 the_hash_algo->init_fn(&c);
a44c9a5e
LT
1892
1893 /* First header.. */
d65a16f6 1894 stream.next_in = (unsigned char *)hdr;
a44c9a5e 1895 stream.avail_in = hdrlen;
55bb5c91
JH
1896 while (git_deflate(&stream, 0) == Z_OK)
1897 ; /* nothing */
18e2588e 1898 the_hash_algo->update_fn(&c, hdr, hdrlen);
a44c9a5e
LT
1899
1900 /* Then the data itself.. */
c00e657d 1901 stream.next_in = (void *)buf;
a44c9a5e 1902 stream.avail_in = len;
9892beba 1903 do {
748af44c 1904 unsigned char *in0 = stream.next_in;
55bb5c91 1905 ret = git_deflate(&stream, Z_FINISH);
18e2588e 1906 the_hash_algo->update_fn(&c, in0, stream.next_in - in0);
9892beba 1907 if (write_buffer(fd, compressed, stream.next_out - compressed) < 0)
76011357 1908 die(_("unable to write loose object file"));
9892beba
NP
1909 stream.next_out = compressed;
1910 stream.avail_out = sizeof(compressed);
1911 } while (ret == Z_OK);
1912
ac54c277 1913 if (ret != Z_STREAM_END)
259328b7 1914 die(_("unable to deflate new object %s (%d)"), oid_to_hex(oid),
3fc7281f 1915 ret);
55bb5c91 1916 ret = git_deflate_end_gently(&stream);
ac54c277 1917 if (ret != Z_OK)
259328b7 1918 die(_("deflateEnd on object %s failed (%d)"), oid_to_hex(oid),
3fc7281f 1919 ret);
5951bf46 1920 the_hash_algo->final_oid_fn(&parano_oid, &c);
9001dc2a 1921 if (!oideq(oid, &parano_oid))
259328b7 1922 die(_("confused by unstable object source data for %s"),
3fc7281f 1923 oid_to_hex(oid));
ac54c277 1924
514c5fdd 1925 close_loose_object(fd);
0fcfd160 1926
bbac7311
NP
1927 if (mtime) {
1928 struct utimbuf utb;
1929 utb.actime = mtime;
1930 utb.modtime = mtime;
4ef91a2d
ÆAB
1931 if (utime(tmp_file.buf, &utb) < 0 &&
1932 !(flags & HASH_SILENT))
259328b7 1933 warning_errno(_("failed utime() on %s"), tmp_file.buf);
bbac7311
NP
1934 }
1935
ea657730 1936 return finalize_object_file(tmp_file.buf, filename.buf);
0fcfd160 1937}
8237b185 1938
6862ebbf 1939static int freshen_loose_object(const struct object_id *oid)
33d4221c 1940{
6862ebbf 1941 return check_and_freshen(oid, 1);
33d4221c
JK
1942}
1943
6862ebbf 1944static int freshen_packed_object(const struct object_id *oid)
33d4221c
JK
1945{
1946 struct pack_entry e;
544443cb 1947 if (!find_pack_entry(the_repository, oid, &e))
ee1c6c34
JK
1948 return 0;
1949 if (e.p->freshened)
1950 return 1;
1951 if (!freshen_file(e.p->pack_name))
1952 return 0;
1953 e.p->freshened = 1;
1954 return 1;
33d4221c
JK
1955}
1956
4ef91a2d
ÆAB
1957int write_object_file_flags(const void *buf, unsigned long len,
1958 const char *type, struct object_id *oid,
1959 unsigned flags)
bbac7311 1960{
1af64f73 1961 char hdr[MAX_HEADER_LEN];
ef1286d3 1962 int hdrlen = sizeof(hdr);
bbac7311
NP
1963
1964 /* Normally if we have it in the pack then we do not bother writing
1965 * it out into .git/objects/??/?{38} file.
1966 */
7ad5c44d
MT
1967 write_object_file_prepare(the_hash_algo, buf, len, type, oid, hdr,
1968 &hdrlen);
6862ebbf 1969 if (freshen_packed_object(oid) || freshen_loose_object(oid))
bbac7311 1970 return 0;
4ef91a2d 1971 return write_loose_object(oid, hdr, hdrlen, buf, len, 0, flags);
bbac7311
NP
1972}
1973
1752cbbc
PO
1974int hash_object_file_literally(const void *buf, unsigned long len,
1975 const char *type, struct object_id *oid,
1976 unsigned flags)
0c3db67c
ES
1977{
1978 char *header;
1979 int hdrlen, status = 0;
1980
1981 /* type string, SP, %lu of the length plus NUL must fit this */
1af64f73 1982 hdrlen = strlen(type) + MAX_HEADER_LEN;
ef1286d3 1983 header = xmalloc(hdrlen);
7ad5c44d
MT
1984 write_object_file_prepare(the_hash_algo, buf, len, type, oid, header,
1985 &hdrlen);
0c3db67c
ES
1986
1987 if (!(flags & HASH_WRITE_OBJECT))
1988 goto cleanup;
6862ebbf 1989 if (freshen_packed_object(oid) || freshen_loose_object(oid))
0c3db67c 1990 goto cleanup;
4ef91a2d 1991 status = write_loose_object(oid, header, hdrlen, buf, len, 0, 0);
0c3db67c
ES
1992
1993cleanup:
1994 free(header);
1995 return status;
1996}
1997
4bdb70a4 1998int force_object_loose(const struct object_id *oid, time_t mtime)
bbac7311 1999{
bbac7311
NP
2000 void *buf;
2001 unsigned long len;
2002 enum object_type type;
1af64f73 2003 char hdr[MAX_HEADER_LEN];
bbac7311 2004 int hdrlen;
1fb23e65 2005 int ret;
bbac7311 2006
6862ebbf 2007 if (has_loose_object(oid))
bbac7311 2008 return 0;
cba595ab 2009 buf = read_object(the_repository, oid, &type, &len);
bbac7311 2010 if (!buf)
2c319886 2011 return error(_("cannot read object for %s"), oid_to_hex(oid));
ca473cef 2012 hdrlen = xsnprintf(hdr, sizeof(hdr), "%s %"PRIuMAX , type_name(type), (uintmax_t)len) + 1;
4ef91a2d 2013 ret = write_loose_object(oid, hdr, hdrlen, buf, len, mtime, 0);
1fb23e65
BS
2014 free(buf);
2015
2016 return ret;
bbac7311
NP
2017}
2018
1d8d9cb6
JT
2019int has_object(struct repository *r, const struct object_id *oid,
2020 unsigned flags)
2021{
2022 int quick = !(flags & HAS_OBJECT_RECHECK_PACKED);
2023 unsigned object_info_flags = OBJECT_INFO_SKIP_FETCH_OBJECT |
2024 (quick ? OBJECT_INFO_QUICK : 0);
2025
2026 if (!startup_info->have_repository)
2027 return 0;
2028 return oid_object_info_extended(r, oid, NULL, object_info_flags) >= 0;
2029}
2030
cba595ab
JH
2031int repo_has_object_file_with_flags(struct repository *r,
2032 const struct object_id *oid, int flags)
8237b185 2033{
3e8b7d3c
JN
2034 if (!startup_info->have_repository)
2035 return 0;
9c8a294a 2036 return oid_object_info_extended(r, oid, NULL, flags) >= 0;
b419aa25 2037}
2038
9b45f499
SB
2039int repo_has_object_file(struct repository *r,
2040 const struct object_id *oid)
b419aa25 2041{
cba595ab 2042 return repo_has_object_file_with_flags(r, oid, 0);
5827a035
JK
2043}
2044
c879daa2
NTND
2045static void check_tree(const void *buf, size_t size)
2046{
2047 struct tree_desc desc;
2048 struct name_entry entry;
2049
2050 init_tree_desc(&desc, buf, size);
2051 while (tree_entry(&desc, &entry))
2052 /* do nothing
2053 * tree_entry() will die() on malformed entries */
2054 ;
2055}
2056
2057static void check_commit(const void *buf, size_t size)
2058{
2059 struct commit c;
2060 memset(&c, 0, sizeof(c));
08f4f445 2061 if (parse_commit_buffer(the_repository, &c, buf, size, 0))
259328b7 2062 die(_("corrupt commit"));
c879daa2
NTND
2063}
2064
2065static void check_tag(const void *buf, size_t size)
2066{
2067 struct tag t;
2068 memset(&t, 0, sizeof(t));
0e740fed 2069 if (parse_tag_buffer(the_repository, &t, buf, size))
259328b7 2070 die(_("corrupt tag"));
c879daa2
NTND
2071}
2072
58bf2a4c
NTND
2073static int index_mem(struct index_state *istate,
2074 struct object_id *oid, void *buf, size_t size,
c4ce46fc
JH
2075 enum object_type type,
2076 const char *path, unsigned flags)
e7332f96 2077{
6c510bee 2078 int ret, re_allocated = 0;
c4ce46fc 2079 int write_object = flags & HASH_WRITE_OBJECT;
74400e71 2080
7672db20 2081 if (!type)
edaec3fb 2082 type = OBJ_BLOB;
6c510bee
LT
2083
2084 /*
2085 * Convert blobs to git internal format
2086 */
43df4f86 2087 if ((type == OBJ_BLOB) && path) {
f285a2d7 2088 struct strbuf nbuf = STRBUF_INIT;
58bf2a4c 2089 if (convert_to_git(istate, path, buf, size, &nbuf,
8462ff43 2090 get_conv_flags(flags))) {
b315c5c0 2091 buf = strbuf_detach(&nbuf, &size);
6c510bee
LT
2092 re_allocated = 1;
2093 }
2094 }
c4ce46fc 2095 if (flags & HASH_FORMAT_CHECK) {
c879daa2
NTND
2096 if (type == OBJ_TREE)
2097 check_tree(buf, size);
2098 if (type == OBJ_COMMIT)
2099 check_commit(buf, size);
2100 if (type == OBJ_TAG)
2101 check_tag(buf, size);
2102 }
6c510bee 2103
7672db20 2104 if (write_object)
169c9c01 2105 ret = write_object_file(buf, size, type_name(type), oid);
abdc3fc8 2106 else
2dcde20e
MT
2107 ret = hash_object_file(the_hash_algo, buf, size,
2108 type_name(type), oid);
43df4f86 2109 if (re_allocated)
6c510bee 2110 free(buf);
43df4f86
DP
2111 return ret;
2112}
2113
58bf2a4c
NTND
2114static int index_stream_convert_blob(struct index_state *istate,
2115 struct object_id *oid,
2116 int fd,
2117 const char *path,
2118 unsigned flags)
9035d75a
SP
2119{
2120 int ret;
2121 const int write_object = flags & HASH_WRITE_OBJECT;
2122 struct strbuf sbuf = STRBUF_INIT;
2123
2124 assert(path);
58bf2a4c 2125 assert(would_convert_to_git_filter_fd(istate, path));
9035d75a 2126
58bf2a4c 2127 convert_to_git_filter_fd(istate, path, fd, &sbuf,
8462ff43 2128 get_conv_flags(flags));
9035d75a
SP
2129
2130 if (write_object)
169c9c01 2131 ret = write_object_file(sbuf.buf, sbuf.len, type_name(OBJ_BLOB),
a09c985e 2132 oid);
9035d75a 2133 else
2dcde20e
MT
2134 ret = hash_object_file(the_hash_algo, sbuf.buf, sbuf.len,
2135 type_name(OBJ_BLOB), oid);
9035d75a
SP
2136 strbuf_release(&sbuf);
2137 return ret;
2138}
2139
58bf2a4c
NTND
2140static int index_pipe(struct index_state *istate, struct object_id *oid,
2141 int fd, enum object_type type,
7b41e1e1
JH
2142 const char *path, unsigned flags)
2143{
2144 struct strbuf sbuf = STRBUF_INIT;
2145 int ret;
2146
2147 if (strbuf_read(&sbuf, fd, 4096) >= 0)
58bf2a4c 2148 ret = index_mem(istate, oid, sbuf.buf, sbuf.len, type, path, flags);
7b41e1e1
JH
2149 else
2150 ret = -1;
2151 strbuf_release(&sbuf);
2152 return ret;
2153}
2154
ea68b0ce
DP
2155#define SMALL_FILE_SIZE (32*1024)
2156
58bf2a4c
NTND
2157static int index_core(struct index_state *istate,
2158 struct object_id *oid, int fd, size_t size,
7b41e1e1
JH
2159 enum object_type type, const char *path,
2160 unsigned flags)
43df4f86
DP
2161{
2162 int ret;
43df4f86 2163
7b41e1e1 2164 if (!size) {
58bf2a4c 2165 ret = index_mem(istate, oid, "", size, type, path, flags);
ea68b0ce
DP
2166 } else if (size <= SMALL_FILE_SIZE) {
2167 char *buf = xmalloc(size);
90dca671
JK
2168 ssize_t read_result = read_in_full(fd, buf, size);
2169 if (read_result < 0)
259328b7 2170 ret = error_errno(_("read error while indexing %s"),
90dca671
JK
2171 path ? path : "<unknown>");
2172 else if (read_result != size)
259328b7 2173 ret = error(_("short read while indexing %s"),
90dca671 2174 path ? path : "<unknown>");
ea68b0ce 2175 else
58bf2a4c 2176 ret = index_mem(istate, oid, buf, size, type, path, flags);
ea68b0ce 2177 free(buf);
08bda208 2178 } else {
43df4f86 2179 void *buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
58bf2a4c 2180 ret = index_mem(istate, oid, buf, size, type, path, flags);
aac17941 2181 munmap(buf, size);
08bda208 2182 }
7b41e1e1
JH
2183 return ret;
2184}
2185
4dd1fbc7 2186/*
568508e7
JH
2187 * This creates one packfile per large blob unless bulk-checkin
2188 * machinery is "plugged".
4dd1fbc7
JH
2189 *
2190 * This also bypasses the usual "convert-to-git" dance, and that is on
2191 * purpose. We could write a streaming version of the converting
2192 * functions and insert that before feeding the data to fast-import
4f22b101
JK
2193 * (or equivalent in-core API described above). However, that is
2194 * somewhat complicated, as we do not know the size of the filter
2195 * result, which we need to know beforehand when writing a git object.
2196 * Since the primary motivation for trying to stream from the working
2197 * tree file and to avoid mmaping it in core is to deal with large
2198 * binary blobs, they generally do not want to get any conversion, and
2199 * callers should avoid this code path when filters are requested.
4dd1fbc7 2200 */
7d5e1dc3 2201static int index_stream(struct object_id *oid, int fd, size_t size,
4dd1fbc7
JH
2202 enum object_type type, const char *path,
2203 unsigned flags)
2204{
68ee6dfc 2205 return index_bulk_checkin(oid, fd, size, type, path, flags);
4dd1fbc7
JH
2206}
2207
58bf2a4c
NTND
2208int index_fd(struct index_state *istate, struct object_id *oid,
2209 int fd, struct stat *st,
7b41e1e1
JH
2210 enum object_type type, const char *path, unsigned flags)
2211{
2212 int ret;
7b41e1e1 2213
9079ab7c
SP
2214 /*
2215 * Call xsize_t() only when needed to avoid potentially unnecessary
2216 * die() for large files.
2217 */
58bf2a4c
NTND
2218 if (type == OBJ_BLOB && path && would_convert_to_git_filter_fd(istate, path))
2219 ret = index_stream_convert_blob(istate, oid, fd, path, flags);
9035d75a 2220 else if (!S_ISREG(st->st_mode))
58bf2a4c 2221 ret = index_pipe(istate, oid, fd, type, path, flags);
9079ab7c 2222 else if (st->st_size <= big_file_threshold || type != OBJ_BLOB ||
58bf2a4c
NTND
2223 (path && would_convert_to_git(istate, path)))
2224 ret = index_core(istate, oid, fd, xsize_t(st->st_size),
2225 type, path, flags);
4dd1fbc7 2226 else
7d5e1dc3 2227 ret = index_stream(oid, fd, xsize_t(st->st_size), type, path,
9079ab7c 2228 flags);
43df4f86 2229 close(fd);
aac17941 2230 return ret;
74400e71 2231}
ec1fcc16 2232
58bf2a4c
NTND
2233int index_path(struct index_state *istate, struct object_id *oid,
2234 const char *path, struct stat *st, unsigned flags)
ec1fcc16
JH
2235{
2236 int fd;
b760d3aa 2237 struct strbuf sb = STRBUF_INIT;
ea8e0297 2238 int rc = 0;
ec1fcc16
JH
2239
2240 switch (st->st_mode & S_IFMT) {
2241 case S_IFREG:
2242 fd = open(path, O_RDONLY);
2243 if (fd < 0)
7616c6ca 2244 return error_errno("open(\"%s\")", path);
58bf2a4c 2245 if (index_fd(istate, oid, fd, st, OBJ_BLOB, path, flags) < 0)
259328b7 2246 return error(_("%s: failed to insert into database"),
ec1fcc16
JH
2247 path);
2248 break;
2249 case S_IFLNK:
7616c6ca
NTND
2250 if (strbuf_readlink(&sb, path, st->st_size))
2251 return error_errno("readlink(\"%s\")", path);
c4ce46fc 2252 if (!(flags & HASH_WRITE_OBJECT))
2dcde20e
MT
2253 hash_object_file(the_hash_algo, sb.buf, sb.len,
2254 blob_type, oid);
a09c985e 2255 else if (write_object_file(sb.buf, sb.len, blob_type, oid))
259328b7 2256 rc = error(_("%s: failed to insert into database"), path);
b760d3aa 2257 strbuf_release(&sb);
ec1fcc16 2258 break;
f35a6d3b 2259 case S_IFDIR:
a98e6101 2260 return resolve_gitlink_ref(path, "HEAD", oid);
ec1fcc16 2261 default:
259328b7 2262 return error(_("%s: unsupported file type"), path);
ec1fcc16 2263 }
ea8e0297 2264 return rc;
ec1fcc16 2265}
a69e5429
JH
2266
2267int read_pack_header(int fd, struct pack_header *header)
2268{
f48ecd38 2269 if (read_in_full(fd, header, sizeof(*header)) != sizeof(*header))
c697ad14
HO
2270 /* "eof before pack header was fully read" */
2271 return PH_ERROR_EOF;
2272
a69e5429
JH
2273 if (header->hdr_signature != htonl(PACK_SIGNATURE))
2274 /* "protocol error (pack signature mismatch detected)" */
2275 return PH_ERROR_PACK_SIGNATURE;
2276 if (!pack_version_ok(header->hdr_version))
2277 /* "protocol error (pack version unsupported)" */
2278 return PH_ERROR_PROTOCOL;
2279 return 0;
2280}
40d52ff7 2281
e816caa0 2282void assert_oid_type(const struct object_id *oid, enum object_type expect)
40d52ff7 2283{
0df8e965 2284 enum object_type type = oid_object_info(the_repository, oid, NULL);
40d52ff7 2285 if (type < 0)
259328b7 2286 die(_("%s is not a valid object"), oid_to_hex(oid));
40d52ff7 2287 if (type != expect)
259328b7 2288 die(_("%s is not a valid '%s' object"), oid_to_hex(oid),
debca9d2 2289 type_name(expect));
40d52ff7 2290}
27e1e22d 2291
70c49050 2292int for_each_file_in_obj_subdir(unsigned int subdir_nr,
cc817ca3
RS
2293 struct strbuf *path,
2294 each_loose_object_fn obj_cb,
2295 each_loose_cruft_fn cruft_cb,
2296 each_loose_subdir_fn subdir_cb,
2297 void *data)
27e1e22d 2298{
0375f472
RS
2299 size_t origlen, baselen;
2300 DIR *dir;
27e1e22d
JK
2301 struct dirent *de;
2302 int r = 0;
62a24c89 2303 struct object_id oid;
27e1e22d 2304
70c49050
RS
2305 if (subdir_nr > 0xff)
2306 BUG("invalid loose object subdirectory: %x", subdir_nr);
2307
0375f472
RS
2308 origlen = path->len;
2309 strbuf_complete(path, '/');
2310 strbuf_addf(path, "%02x", subdir_nr);
0375f472
RS
2311
2312 dir = opendir(path->buf);
27e1e22d 2313 if (!dir) {
0375f472 2314 if (errno != ENOENT)
259328b7 2315 r = error_errno(_("unable to open %s"), path->buf);
0375f472
RS
2316 strbuf_setlen(path, origlen);
2317 return r;
27e1e22d
JK
2318 }
2319
62a24c89 2320 oid.hash[0] = subdir_nr;
163ee5e6
DS
2321 strbuf_addch(path, '/');
2322 baselen = path->len;
62a24c89 2323
b548f0f1 2324 while ((de = readdir_skip_dot_and_dotdot(dir))) {
163ee5e6 2325 size_t namelen;
27e1e22d 2326
163ee5e6 2327 namelen = strlen(de->d_name);
27e1e22d 2328 strbuf_setlen(path, baselen);
163ee5e6 2329 strbuf_add(path, de->d_name, namelen);
94b5e093 2330 if (namelen == the_hash_algo->hexsz - 2 &&
62a24c89 2331 !hex_to_bytes(oid.hash + 1, de->d_name,
94b5e093 2332 the_hash_algo->rawsz - 1)) {
5a6dce70 2333 oid_set_algo(&oid, the_hash_algo);
62a24c89
RS
2334 if (obj_cb) {
2335 r = obj_cb(&oid, path->buf, data);
2336 if (r)
2337 break;
27e1e22d 2338 }
62a24c89 2339 continue;
27e1e22d
JK
2340 }
2341
2342 if (cruft_cb) {
2343 r = cruft_cb(de->d_name, path->buf, data);
2344 if (r)
2345 break;
2346 }
2347 }
094c7e63 2348 closedir(dir);
27e1e22d 2349
163ee5e6 2350 strbuf_setlen(path, baselen - 1);
27e1e22d
JK
2351 if (!r && subdir_cb)
2352 r = subdir_cb(subdir_nr, path->buf, data);
2353
0375f472
RS
2354 strbuf_setlen(path, origlen);
2355
27e1e22d
JK
2356 return r;
2357}
2358
e6f875e0 2359int for_each_loose_file_in_objdir_buf(struct strbuf *path,
27e1e22d
JK
2360 each_loose_object_fn obj_cb,
2361 each_loose_cruft_fn cruft_cb,
2362 each_loose_subdir_fn subdir_cb,
2363 void *data)
2364{
27e1e22d
JK
2365 int r = 0;
2366 int i;
2367
27e1e22d 2368 for (i = 0; i < 256; i++) {
e6f875e0 2369 r = for_each_file_in_obj_subdir(i, path, obj_cb, cruft_cb,
27e1e22d 2370 subdir_cb, data);
27e1e22d
JK
2371 if (r)
2372 break;
2373 }
2374
e6f875e0
JK
2375 return r;
2376}
2377
2378int for_each_loose_file_in_objdir(const char *path,
2379 each_loose_object_fn obj_cb,
2380 each_loose_cruft_fn cruft_cb,
2381 each_loose_subdir_fn subdir_cb,
2382 void *data)
2383{
2384 struct strbuf buf = STRBUF_INIT;
2385 int r;
2386
2387 strbuf_addstr(&buf, path);
2388 r = for_each_loose_file_in_objdir_buf(&buf, obj_cb, cruft_cb,
2389 subdir_cb, data);
27e1e22d 2390 strbuf_release(&buf);
e6f875e0 2391
27e1e22d
JK
2392 return r;
2393}
660c889e 2394
a7ff6f5a
JK
2395int for_each_loose_object(each_loose_object_fn cb, void *data,
2396 enum for_each_object_flags flags)
660c889e 2397{
f0eaf638 2398 struct object_directory *odb;
b0a42642 2399
f0eaf638
JK
2400 prepare_alt_odb(the_repository);
2401 for (odb = the_repository->objects->odb; odb; odb = odb->next) {
2402 int r = for_each_loose_file_in_objdir(odb->path, cb, NULL,
2403 NULL, data);
2404 if (r)
2405 return r;
660c889e 2406
f0eaf638
JK
2407 if (flags & FOR_EACH_OBJECT_LOCAL_ONLY)
2408 break;
2409 }
1385bb7b 2410
f0eaf638 2411 return 0;
660c889e
JK
2412}
2413
3a2e0824
JK
2414static int append_loose_object(const struct object_id *oid, const char *path,
2415 void *data)
660c889e 2416{
92d8ed8a 2417 oidtree_insert(data, oid);
3a2e0824
JK
2418 return 0;
2419}
660c889e 2420
92d8ed8a 2421struct oidtree *odb_loose_cache(struct object_directory *odb,
0000d654
RS
2422 const struct object_id *oid)
2423{
2424 int subdir_nr = oid->hash[0];
3a2e0824 2425 struct strbuf buf = STRBUF_INIT;
33f379ee
EW
2426 size_t word_bits = bitsizeof(odb->loose_objects_subdir_seen[0]);
2427 size_t word_index = subdir_nr / word_bits;
581a3bb1 2428 size_t mask = 1u << (subdir_nr % word_bits);
33f379ee 2429 uint32_t *bitmap;
660c889e 2430
3a2e0824 2431 if (subdir_nr < 0 ||
33f379ee 2432 subdir_nr >= bitsizeof(odb->loose_objects_subdir_seen))
3a2e0824 2433 BUG("subdir_nr out of range");
1385bb7b 2434
33f379ee
EW
2435 bitmap = &odb->loose_objects_subdir_seen[word_index];
2436 if (*bitmap & mask)
92d8ed8a
EW
2437 return odb->loose_objects_cache;
2438 if (!odb->loose_objects_cache) {
2439 ALLOC_ARRAY(odb->loose_objects_cache, 1);
2440 oidtree_init(odb->loose_objects_cache);
2441 }
3a2e0824
JK
2442 strbuf_addstr(&buf, odb->path);
2443 for_each_file_in_obj_subdir(subdir_nr, &buf,
2444 append_loose_object,
2445 NULL, NULL,
92d8ed8a 2446 odb->loose_objects_cache);
33f379ee 2447 *bitmap |= mask;
7317aa71 2448 strbuf_release(&buf);
92d8ed8a 2449 return odb->loose_objects_cache;
660c889e
JK
2450}
2451
d4e19e51
RS
2452void odb_clear_loose_cache(struct object_directory *odb)
2453{
92d8ed8a
EW
2454 oidtree_clear(odb->loose_objects_cache);
2455 FREE_AND_NULL(odb->loose_objects_cache);
d4e19e51
RS
2456 memset(&odb->loose_objects_subdir_seen, 0,
2457 sizeof(odb->loose_objects_subdir_seen));
2458}
2459
00a7760e
JK
2460static int check_stream_oid(git_zstream *stream,
2461 const char *hdr,
2462 unsigned long size,
2463 const char *path,
2464 const struct object_id *expected_oid)
f6371f92 2465{
18e2588e 2466 git_hash_ctx c;
00a7760e 2467 struct object_id real_oid;
f6371f92
JK
2468 unsigned char buf[4096];
2469 unsigned long total_read;
2470 int status = Z_OK;
2471
18e2588e 2472 the_hash_algo->init_fn(&c);
2473 the_hash_algo->update_fn(&c, hdr, stream->total_out);
f6371f92
JK
2474
2475 /*
2476 * We already read some bytes into hdr, but the ones up to the NUL
2477 * do not count against the object's content size.
2478 */
2479 total_read = stream->total_out - strlen(hdr) - 1;
2480
2481 /*
2482 * This size comparison must be "<=" to read the final zlib packets;
00a7760e 2483 * see the comment in unpack_loose_rest for details.
f6371f92
JK
2484 */
2485 while (total_read <= size &&
18ad13e5
JH
2486 (status == Z_OK ||
2487 (status == Z_BUF_ERROR && !stream->avail_out))) {
f6371f92
JK
2488 stream->next_out = buf;
2489 stream->avail_out = sizeof(buf);
2490 if (size - total_read < stream->avail_out)
2491 stream->avail_out = size - total_read;
2492 status = git_inflate(stream, Z_FINISH);
18e2588e 2493 the_hash_algo->update_fn(&c, buf, stream->next_out - buf);
f6371f92
JK
2494 total_read += stream->next_out - buf;
2495 }
2496 git_inflate_end(stream);
2497
2498 if (status != Z_STREAM_END) {
00a7760e 2499 error(_("corrupt loose object '%s'"), oid_to_hex(expected_oid));
f6371f92
JK
2500 return -1;
2501 }
cce044df 2502 if (stream->avail_in) {
259328b7 2503 error(_("garbage at end of loose object '%s'"),
00a7760e 2504 oid_to_hex(expected_oid));
cce044df
JK
2505 return -1;
2506 }
f6371f92 2507
5951bf46 2508 the_hash_algo->final_oid_fn(&real_oid, &c);
00a7760e 2509 if (!oideq(expected_oid, &real_oid)) {
01f8d594 2510 error(_("hash mismatch for %s (expected %s)"), path,
00a7760e 2511 oid_to_hex(expected_oid));
f6371f92
JK
2512 return -1;
2513 }
2514
2515 return 0;
2516}
2517
2518int read_loose_object(const char *path,
d61d87bd 2519 const struct object_id *expected_oid,
96e41f58 2520 struct object_id *real_oid,
31deb28f
ÆAB
2521 void **contents,
2522 struct object_info *oi)
f6371f92
JK
2523{
2524 int ret = -1;
f6371f92
JK
2525 void *map = NULL;
2526 unsigned long mapsize;
2527 git_zstream stream;
1af64f73 2528 char hdr[MAX_HEADER_LEN];
31deb28f 2529 unsigned long *size = oi->sizep;
f6371f92 2530
514c5fdd 2531 map = map_loose_object_1(the_repository, path, NULL, &mapsize);
f6371f92 2532 if (!map) {
259328b7 2533 error_errno(_("unable to mmap %s"), path);
f6371f92
JK
2534 goto out;
2535 }
2536
01cab976
ÆAB
2537 if (unpack_loose_header(&stream, map, mapsize, hdr, sizeof(hdr),
2538 NULL) < 0) {
259328b7 2539 error(_("unable to unpack header of %s"), path);
f6371f92
JK
2540 goto out;
2541 }
2542
31deb28f 2543 if (parse_loose_header(hdr, oi) < 0) {
259328b7 2544 error(_("unable to parse header of %s"), path);
f6371f92
JK
2545 git_inflate_end(&stream);
2546 goto out;
2547 }
2548
31deb28f 2549 if (*oi->typep == OBJ_BLOB && *size > big_file_threshold) {
00a7760e 2550 if (check_stream_oid(&stream, hdr, *size, path, expected_oid) < 0)
f6371f92
JK
2551 goto out;
2552 } else {
00a7760e 2553 *contents = unpack_loose_rest(&stream, hdr, *size, expected_oid);
f6371f92 2554 if (!*contents) {
259328b7 2555 error(_("unable to unpack contents of %s"), path);
f6371f92
JK
2556 git_inflate_end(&stream);
2557 goto out;
2558 }
b98d1885 2559 if (check_object_signature(the_repository, expected_oid,
16235e3b
ÆAB
2560 *contents, *size,
2561 oi->type_name->buf, real_oid))
f6371f92 2562 goto out;
f6371f92
JK
2563 }
2564
2565 ret = 0; /* everything checks out */
2566
2567out:
2568 if (map)
2569 munmap(map, mapsize);
f6371f92
JK
2570 return ret;
2571}