]> git.ipfire.org Git - thirdparty/git.git/blame - object-file.c
xmmap: inform Linux users of tuning knobs on ENOMEM
[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"
e05db0fd 35
1af64f73 36/* The maximum size for an object header. */
37#define MAX_HEADER_LEN 32
38
e1ccd7e2 39
40#define EMPTY_TREE_SHA1_BIN_LITERAL \
41 "\x4b\x82\x5d\xc6\x42\xcb\x6e\xb9\xa0\x60" \
42 "\xe5\x4b\xf8\xd6\x92\x88\xfb\xee\x49\x04"
13eeedb5 43#define EMPTY_TREE_SHA256_BIN_LITERAL \
44 "\x6e\xf1\x9b\x41\x22\x5c\x53\x69\xf1\xc1" \
45 "\x04\xd4\x5d\x8d\x85\xef\xa9\xb0\x57\xb5" \
46 "\x3b\x14\xb4\xb9\xb9\x39\xdd\x74\xde\xcc" \
47 "\x53\x21"
e1ccd7e2 48
49#define EMPTY_BLOB_SHA1_BIN_LITERAL \
50 "\xe6\x9d\xe2\x9b\xb2\xd1\xd6\x43\x4b\x8b" \
51 "\x29\xae\x77\x5a\xd8\xc2\xe4\x8c\x53\x91"
13eeedb5 52#define EMPTY_BLOB_SHA256_BIN_LITERAL \
53 "\x47\x3a\x0f\x4c\x3b\xe8\xa9\x36\x81\xa2" \
54 "\x67\xe3\xb1\xe9\xa7\xdc\xda\x11\x85\x43" \
55 "\x6f\xe1\x41\xf7\x74\x91\x20\xa3\x03\x72" \
56 "\x18\x13"
e1ccd7e2 57
e1ccd7e2 58static const struct object_id empty_tree_oid = {
5a6dce70 59 .hash = EMPTY_TREE_SHA1_BIN_LITERAL,
60 .algo = GIT_HASH_SHA1,
8576fde6 61};
e1ccd7e2 62static const struct object_id empty_blob_oid = {
5a6dce70 63 .hash = EMPTY_BLOB_SHA1_BIN_LITERAL,
64 .algo = GIT_HASH_SHA1,
8576fde6 65};
14228447 66static const struct object_id null_oid_sha1 = {
67 .hash = {0},
68 .algo = GIT_HASH_SHA1,
69};
13eeedb5 70static const struct object_id empty_tree_oid_sha256 = {
5a6dce70 71 .hash = EMPTY_TREE_SHA256_BIN_LITERAL,
72 .algo = GIT_HASH_SHA256,
13eeedb5 73};
74static const struct object_id empty_blob_oid_sha256 = {
5a6dce70 75 .hash = EMPTY_BLOB_SHA256_BIN_LITERAL,
76 .algo = GIT_HASH_SHA256,
13eeedb5 77};
14228447 78static const struct object_id null_oid_sha256 = {
79 .hash = {0},
80 .algo = GIT_HASH_SHA256,
81};
88cd621d 82
ac73cedf 83static void git_hash_sha1_init(git_hash_ctx *ctx)
f50e766b 84{
ac73cedf 85 git_SHA1_Init(&ctx->sha1);
f50e766b 86}
87
768e30ea 88static void git_hash_sha1_clone(git_hash_ctx *dst, const git_hash_ctx *src)
89{
90 git_SHA1_Clone(&dst->sha1, &src->sha1);
91}
92
ac73cedf 93static void git_hash_sha1_update(git_hash_ctx *ctx, const void *data, size_t len)
f50e766b 94{
ac73cedf 95 git_SHA1_Update(&ctx->sha1, data, len);
f50e766b 96}
97
ac73cedf 98static void git_hash_sha1_final(unsigned char *hash, git_hash_ctx *ctx)
f50e766b 99{
ac73cedf 100 git_SHA1_Final(hash, &ctx->sha1);
f50e766b 101}
102
ab795f0d 103static void git_hash_sha1_final_oid(struct object_id *oid, git_hash_ctx *ctx)
104{
105 git_SHA1_Final(oid->hash, &ctx->sha1);
106 memset(oid->hash + GIT_SHA1_RAWSZ, 0, GIT_MAX_RAWSZ - GIT_SHA1_RAWSZ);
5a6dce70 107 oid->algo = GIT_HASH_SHA1;
ab795f0d 108}
109
13eeedb5 110
111static void git_hash_sha256_init(git_hash_ctx *ctx)
112{
113 git_SHA256_Init(&ctx->sha256);
114}
115
768e30ea 116static void git_hash_sha256_clone(git_hash_ctx *dst, const git_hash_ctx *src)
117{
118 git_SHA256_Clone(&dst->sha256, &src->sha256);
119}
120
13eeedb5 121static void git_hash_sha256_update(git_hash_ctx *ctx, const void *data, size_t len)
122{
123 git_SHA256_Update(&ctx->sha256, data, len);
124}
125
126static void git_hash_sha256_final(unsigned char *hash, git_hash_ctx *ctx)
127{
128 git_SHA256_Final(hash, &ctx->sha256);
129}
130
ab795f0d 131static void git_hash_sha256_final_oid(struct object_id *oid, git_hash_ctx *ctx)
132{
133 git_SHA256_Final(oid->hash, &ctx->sha256);
134 /*
135 * This currently does nothing, so the compiler should optimize it out,
136 * but keep it in case we extend the hash size again.
137 */
138 memset(oid->hash + GIT_SHA256_RAWSZ, 0, GIT_MAX_RAWSZ - GIT_SHA256_RAWSZ);
5a6dce70 139 oid->algo = GIT_HASH_SHA256;
ab795f0d 140}
141
ac73cedf 142static void git_hash_unknown_init(git_hash_ctx *ctx)
f50e766b 143{
1a07e59c 144 BUG("trying to init unknown hash");
f50e766b 145}
146
768e30ea 147static void git_hash_unknown_clone(git_hash_ctx *dst, const git_hash_ctx *src)
148{
149 BUG("trying to clone unknown hash");
150}
151
ac73cedf 152static void git_hash_unknown_update(git_hash_ctx *ctx, const void *data, size_t len)
f50e766b 153{
1a07e59c 154 BUG("trying to update unknown hash");
f50e766b 155}
156
ac73cedf 157static void git_hash_unknown_final(unsigned char *hash, git_hash_ctx *ctx)
f50e766b 158{
1a07e59c 159 BUG("trying to finalize unknown hash");
f50e766b 160}
161
ab795f0d 162static void git_hash_unknown_final_oid(struct object_id *oid, git_hash_ctx *ctx)
163{
164 BUG("trying to finalize unknown hash");
165}
166
167
f50e766b 168const struct git_hash_algo hash_algos[GIT_HASH_NALGOS] = {
169 {
170 NULL,
171 0x00000000,
172 0,
173 0,
a2ce0a75 174 0,
f50e766b 175 git_hash_unknown_init,
768e30ea 176 git_hash_unknown_clone,
f50e766b 177 git_hash_unknown_update,
178 git_hash_unknown_final,
ab795f0d 179 git_hash_unknown_final_oid,
f50e766b 180 NULL,
181 NULL,
14228447 182 NULL,
f50e766b 183 },
184 {
1ccf07cb 185 "sha1",
f50e766b 186 /* "sha1", big-endian */
187 0x73686131,
f50e766b 188 GIT_SHA1_RAWSZ,
189 GIT_SHA1_HEXSZ,
a2ce0a75 190 GIT_SHA1_BLKSZ,
f50e766b 191 git_hash_sha1_init,
768e30ea 192 git_hash_sha1_clone,
f50e766b 193 git_hash_sha1_update,
194 git_hash_sha1_final,
ab795f0d 195 git_hash_sha1_final_oid,
f50e766b 196 &empty_tree_oid,
197 &empty_blob_oid,
14228447 198 &null_oid_sha1,
f50e766b 199 },
13eeedb5 200 {
201 "sha256",
202 /* "s256", big-endian */
203 0x73323536,
204 GIT_SHA256_RAWSZ,
205 GIT_SHA256_HEXSZ,
206 GIT_SHA256_BLKSZ,
207 git_hash_sha256_init,
768e30ea 208 git_hash_sha256_clone,
13eeedb5 209 git_hash_sha256_update,
210 git_hash_sha256_final,
ab795f0d 211 git_hash_sha256_final_oid,
13eeedb5 212 &empty_tree_oid_sha256,
213 &empty_blob_oid_sha256,
14228447 214 &null_oid_sha256,
13eeedb5 215 }
f50e766b 216};
217
14228447 218const struct object_id *null_oid(void)
219{
220 return the_hash_algo->null_oid;
221}
222
d8a92ced 223const char *empty_tree_oid_hex(void)
224{
225 static char buf[GIT_MAX_HEXSZ + 1];
226 return oid_to_hex_r(buf, the_hash_algo->empty_tree);
227}
228
229const char *empty_blob_oid_hex(void)
230{
231 static char buf[GIT_MAX_HEXSZ + 1];
232 return oid_to_hex_r(buf, the_hash_algo->empty_blob);
233}
234
2f90b9d9 235int hash_algo_by_name(const char *name)
236{
237 int i;
238 if (!name)
239 return GIT_HASH_UNKNOWN;
240 for (i = 1; i < GIT_HASH_NALGOS; i++)
241 if (!strcmp(name, hash_algos[i].name))
242 return i;
243 return GIT_HASH_UNKNOWN;
244}
245
246int hash_algo_by_id(uint32_t format_id)
247{
248 int i;
249 for (i = 1; i < GIT_HASH_NALGOS; i++)
250 if (format_id == hash_algos[i].format_id)
251 return i;
252 return GIT_HASH_UNKNOWN;
253}
254
95399788 255int hash_algo_by_length(int len)
256{
257 int i;
258 for (i = 1; i < GIT_HASH_NALGOS; i++)
259 if (len == hash_algos[i].rawsz)
260 return i;
261 return GIT_HASH_UNKNOWN;
262}
2f90b9d9 263
c597ba80
NTND
264/*
265 * This is meant to hold a *small* number of objects that you would
cb1c8d1d 266 * want read_object_file() to be able to return, but yet you do not want
c597ba80
NTND
267 * to write them into the object store (e.g. a browse-only
268 * application).
269 */
270static struct cached_object {
62ba93ea 271 struct object_id oid;
c597ba80
NTND
272 enum object_type type;
273 void *buf;
274 unsigned long size;
275} *cached_objects;
276static int cached_object_nr, cached_object_alloc;
277
278static struct cached_object empty_tree = {
62ba93ea 279 { EMPTY_TREE_SHA1_BIN_LITERAL },
c597ba80
NTND
280 OBJ_TREE,
281 "",
282 0
283};
284
62ba93ea 285static struct cached_object *find_cached_object(const struct object_id *oid)
c597ba80
NTND
286{
287 int i;
288 struct cached_object *co = cached_objects;
289
290 for (i = 0; i < cached_object_nr; i++, co++) {
4a7e27e9 291 if (oideq(&co->oid, oid))
c597ba80
NTND
292 return co;
293 }
4a7e27e9 294 if (oideq(oid, the_hash_algo->empty_tree))
c597ba80
NTND
295 return &empty_tree;
296 return NULL;
297}
298
9472935d 299
8462ff43 300static int get_conv_flags(unsigned flags)
9472935d
TB
301{
302 if (flags & HASH_RENORMALIZE)
8462ff43 303 return CONV_EOL_RENORMALIZE;
9472935d 304 else if (flags & HASH_WRITE_OBJECT)
107642fe 305 return global_conv_flags_eol | CONV_WRITE_OBJECT;
9472935d 306 else
8462ff43 307 return 0;
9472935d
TB
308}
309
310
90a6464b
JH
311int mkdir_in_gitdir(const char *path)
312{
313 if (mkdir(path, 0777)) {
314 int saved_errno = errno;
315 struct stat st;
316 struct strbuf sb = STRBUF_INIT;
317
318 if (errno != EEXIST)
319 return -1;
320 /*
321 * Are we looking at a path in a symlinked worktree
322 * whose original repository does not yet have it?
323 * e.g. .git/rr-cache pointing at its original
324 * repository in which the user hasn't performed any
325 * conflict resolution yet?
326 */
327 if (lstat(path, &st) || !S_ISLNK(st.st_mode) ||
328 strbuf_readlink(&sb, path, st.st_size) ||
329 !is_absolute_path(sb.buf) ||
330 mkdir(sb.buf, 0777)) {
331 strbuf_release(&sb);
332 errno = saved_errno;
333 return -1;
334 }
335 strbuf_release(&sb);
336 }
337 return adjust_shared_perm(path);
338}
339
eb3c027e 340static enum scld_error safe_create_leading_directories_1(char *path, int share)
b2cb9425 341{
26c8ae2a 342 char *next_component = path + offset_1st_component(path);
0be0521b 343 enum scld_error ret = SCLD_OK;
67d42212 344
0be0521b 345 while (ret == SCLD_OK && next_component) {
f0502332 346 struct stat st;
0f527403 347 char *slash = next_component, slash_character;
f0502332 348
0f527403
MH
349 while (*slash && !is_dir_sep(*slash))
350 slash++;
351
352 if (!*slash)
b2cb9425 353 break;
bf10cf70 354
26c8ae2a 355 next_component = slash + 1;
0f527403 356 while (is_dir_sep(*next_component))
bf10cf70 357 next_component++;
26c8ae2a 358 if (!*next_component)
5f0bdf50 359 break;
831651fd 360
0f527403 361 slash_character = *slash;
831651fd 362 *slash = '\0';
67d42212
JR
363 if (!stat(path, &st)) {
364 /* path exists */
204a047f
MH
365 if (!S_ISDIR(st.st_mode)) {
366 errno = ENOTDIR;
0be0521b 367 ret = SCLD_EXISTS;
204a047f 368 }
53a39721 369 } else if (mkdir(path, 0777)) {
928734d9 370 if (errno == EEXIST &&
9e6f885d 371 !stat(path, &st) && S_ISDIR(st.st_mode))
928734d9 372 ; /* somebody created it since we checked */
18d37e86
MH
373 else if (errno == ENOENT)
374 /*
375 * Either mkdir() failed because
376 * somebody just pruned the containing
377 * directory, or stat() failed because
378 * the file that was in our way was
379 * just removed. Either way, inform
380 * the caller that it might be worth
381 * trying again:
382 */
383 ret = SCLD_VANISHED;
9e6f885d 384 else
0be0521b 385 ret = SCLD_FAILED;
eb3c027e 386 } else if (share && adjust_shared_perm(path)) {
0be0521b 387 ret = SCLD_PERMS;
457f06d6 388 }
0f527403 389 *slash = slash_character;
b2cb9425 390 }
9e6f885d 391 return ret;
b2cb9425 392}
723c31fe 393
eb3c027e
MT
394enum scld_error safe_create_leading_directories(char *path)
395{
396 return safe_create_leading_directories_1(path, 1);
397}
398
399enum scld_error safe_create_leading_directories_no_share(char *path)
400{
401 return safe_create_leading_directories_1(path, 0);
402}
403
0be0521b 404enum scld_error safe_create_leading_directories_const(const char *path)
8e21d63b 405{
02944307 406 int save_errno;
8e21d63b
JK
407 /* path points to cache entries, so xstrdup before messing with it */
408 char *buf = xstrdup(path);
0be0521b 409 enum scld_error result = safe_create_leading_directories(buf);
02944307
MH
410
411 save_errno = errno;
8e21d63b 412 free(buf);
02944307 413 errno = save_errno;
8e21d63b
JK
414 return result;
415}
416
177978f5
MH
417int raceproof_create_file(const char *path, create_file_fn fn, void *cb)
418{
419 /*
420 * The number of times we will try to remove empty directories
421 * in the way of path. This is only 1 because if another
422 * process is racily creating directories that conflict with
423 * us, we don't want to fight against them.
424 */
425 int remove_directories_remaining = 1;
426
427 /*
428 * The number of times that we will try to create the
429 * directories containing path. We are willing to attempt this
430 * more than once, because another process could be trying to
431 * clean up empty directories at the same time as we are
432 * trying to create them.
433 */
434 int create_directories_remaining = 3;
435
436 /* A scratch copy of path, filled lazily if we need it: */
437 struct strbuf path_copy = STRBUF_INIT;
438
439 int ret, save_errno;
440
441 /* Sanity check: */
442 assert(*path);
443
444retry_fn:
445 ret = fn(path, cb);
446 save_errno = errno;
447 if (!ret)
448 goto out;
449
450 if (errno == EISDIR && remove_directories_remaining-- > 0) {
451 /*
452 * A directory is in the way. Maybe it is empty; try
453 * to remove it:
454 */
455 if (!path_copy.len)
456 strbuf_addstr(&path_copy, path);
457
458 if (!remove_dir_recursively(&path_copy, REMOVE_DIR_EMPTY_ONLY))
459 goto retry_fn;
460 } else if (errno == ENOENT && create_directories_remaining-- > 0) {
461 /*
462 * Maybe the containing directory didn't exist, or
463 * maybe it was just deleted by a process that is
464 * racing with us to clean up empty directories. Try
465 * to create it:
466 */
467 enum scld_error scld_result;
468
469 if (!path_copy.len)
470 strbuf_addstr(&path_copy, path);
471
472 do {
473 scld_result = safe_create_leading_directories(path_copy.buf);
474 if (scld_result == SCLD_OK)
475 goto retry_fn;
476 } while (scld_result == SCLD_VANISHED && create_directories_remaining-- > 0);
477 }
478
479out:
480 strbuf_release(&path_copy);
481 errno = save_errno;
482 return ret;
483}
484
514c5fdd 485static void fill_loose_path(struct strbuf *buf, const struct object_id *oid)
ace1534d
JH
486{
487 int i;
94b5e093 488 for (i = 0; i < the_hash_algo->rawsz; i++) {
ace1534d 489 static char hex[] = "0123456789abcdef";
514c5fdd 490 unsigned int val = oid->hash[i];
f7b7774f
JK
491 strbuf_addch(buf, hex[val >> 4]);
492 strbuf_addch(buf, hex[val & 0xf]);
afbba2f0 493 if (!i)
f7b7774f 494 strbuf_addch(buf, '/');
ace1534d
JH
495 }
496}
497
f0eaf638
JK
498static const char *odb_loose_path(struct object_directory *odb,
499 struct strbuf *buf,
514c5fdd 500 const struct object_id *oid)
0fcfd160 501{
b69fb867 502 strbuf_reset(buf);
f0eaf638 503 strbuf_addstr(buf, odb->path);
34498471 504 strbuf_addch(buf, '/');
514c5fdd 505 fill_loose_path(buf, oid);
f3f043a1 506 return buf->buf;
0fcfd160
LT
507}
508
f3f043a1 509const char *loose_object_path(struct repository *r, struct strbuf *buf,
514c5fdd 510 const struct object_id *oid)
29ec6af2 511{
514c5fdd 512 return odb_loose_path(r->objects->odb, buf, oid);
0fcfd160
LT
513}
514
4ea82473
JK
515/*
516 * Return non-zero iff the path is usable as an alternate object database.
517 */
13313fc3
SB
518static int alt_odb_usable(struct raw_object_store *o,
519 struct strbuf *path,
520 const char *normalized_objdir)
4ea82473 521{
263db403 522 struct object_directory *odb;
4ea82473
JK
523
524 /* Detect cases where alternate disappeared */
525 if (!is_directory(path->buf)) {
259328b7
NTND
526 error(_("object directory %s does not exist; "
527 "check .git/objects/info/alternates"),
4ea82473
JK
528 path->buf);
529 return 0;
530 }
531
532 /*
533 * Prevent the common mistake of listing the same
534 * thing twice, or object directory itself.
535 */
f0eaf638 536 for (odb = o->odb; odb; odb = odb->next) {
263db403 537 if (!fspathcmp(path->buf, odb->path))
4ea82473
JK
538 return 0;
539 }
540 if (!fspathcmp(path->buf, normalized_objdir))
541 return 0;
542
543 return 1;
544}
545
ddd5d056
JH
546/*
547 * Prepare alternate object database registry.
d5a63b99
JH
548 *
549 * The variable alt_odb_list points at the list of struct
263db403 550 * object_directory. The elements on this list come from
d5a63b99
JH
551 * non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT
552 * environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates,
1494e038
JH
553 * whose contents is similar to that environment variable but can be
554 * LF separated. Its base points at a statically allocated buffer that
d5a63b99
JH
555 * contains "/the/directory/corresponding/to/.git/objects/...", while
556 * its name points just after the slash at the end of ".git/objects/"
e5afd444
557 * in the example above, and has enough space to hold all hex characters
558 * of the object ID, an extra slash for the first level indirection, and
559 * the terminating NUL.
ddd5d056 560 */
77f012e8
SB
561static void read_info_alternates(struct repository *r,
562 const char *relative_base,
563 int depth);
564static int link_alt_odb_entry(struct repository *r, const char *entry,
cfc62fc9 565 const char *relative_base, int depth, const char *normalized_objdir)
ace1534d 566{
263db403 567 struct object_directory *ent;
5bdf0a84 568 struct strbuf pathbuf = STRBUF_INIT;
d5a63b99 569
85dadc38 570 if (!is_absolute_path(entry) && relative_base) {
4ac9006f 571 strbuf_realpath(&pathbuf, relative_base, 1);
5bdf0a84 572 strbuf_addch(&pathbuf, '/');
c2f493a4 573 }
6eac50d8 574 strbuf_addstr(&pathbuf, entry);
c2f493a4 575
37a95862 576 if (strbuf_normalize_path(&pathbuf) < 0 && relative_base) {
259328b7 577 error(_("unable to normalize alternate object path: %s"),
670c359d
JK
578 pathbuf.buf);
579 strbuf_release(&pathbuf);
580 return -1;
581 }
5bdf0a84
HW
582
583 /*
584 * The trailing slash after the directory name is given by
585 * this function at the end. Remove duplicates.
586 */
4ea82473
JK
587 while (pathbuf.len && pathbuf.buf[pathbuf.len - 1] == '/')
588 strbuf_setlen(&pathbuf, pathbuf.len - 1);
5bdf0a84 589
77f012e8 590 if (!alt_odb_usable(r->objects, &pathbuf, normalized_objdir)) {
4ea82473 591 strbuf_release(&pathbuf);
c2f493a4
MW
592 return -1;
593 }
594
ca56dadb 595 CALLOC_ARRAY(ent, 1);
f0eaf638 596 ent->path = xstrdup(pathbuf.buf);
c2f493a4
MW
597
598 /* add the alternate entry */
f0eaf638
JK
599 *r->objects->odb_tail = ent;
600 r->objects->odb_tail = &(ent->next);
c2f493a4
MW
601 ent->next = NULL;
602
603 /* recursively add alternates */
77f012e8 604 read_info_alternates(r, pathbuf.buf, depth + 1);
c2f493a4 605
4ea82473 606 strbuf_release(&pathbuf);
c2f493a4
MW
607 return 0;
608}
609
cf3c6352
JK
610static const char *parse_alt_odb_entry(const char *string,
611 int sep,
612 struct strbuf *out)
613{
614 const char *end;
615
616 strbuf_reset(out);
617
618 if (*string == '#') {
619 /* comment; consume up to next separator */
620 end = strchrnul(string, sep);
621 } else if (*string == '"' && !unquote_c_style(out, string, &end)) {
622 /*
623 * quoted path; unquote_c_style has copied the
624 * data for us and set "end". Broken quoting (e.g.,
625 * an entry that doesn't end with a quote) falls
626 * back to the unquoted case below.
627 */
628 } else {
629 /* normal, unquoted path */
630 end = strchrnul(string, sep);
631 strbuf_add(out, string, end - string);
632 }
633
634 if (*end)
635 end++;
636 return end;
637}
638
77f012e8
SB
639static void link_alt_odb_entries(struct repository *r, const char *alt,
640 int sep, const char *relative_base, int depth)
c2f493a4 641{
539e7506 642 struct strbuf objdirbuf = STRBUF_INIT;
cf3c6352 643 struct strbuf entry = STRBUF_INIT;
c2f493a4 644
f28e3668
JK
645 if (!alt || !*alt)
646 return;
647
c2f493a4 648 if (depth > 5) {
259328b7 649 error(_("%s: ignoring alternate object stores, nesting too deep"),
c2f493a4
MW
650 relative_base);
651 return;
652 }
653
f0eaf638 654 strbuf_add_absolute_path(&objdirbuf, r->objects->odb->path);
670c359d 655 if (strbuf_normalize_path(&objdirbuf) < 0)
259328b7 656 die(_("unable to normalize object directory: %s"),
670c359d 657 objdirbuf.buf);
539e7506 658
cf3c6352
JK
659 while (*alt) {
660 alt = parse_alt_odb_entry(alt, sep, &entry);
661 if (!entry.len)
9577e7e3 662 continue;
77f012e8 663 link_alt_odb_entry(r, entry.buf,
cfc62fc9 664 relative_base, depth, objdirbuf.buf);
9577e7e3 665 }
cf3c6352 666 strbuf_release(&entry);
539e7506 667 strbuf_release(&objdirbuf);
d5a63b99
JH
668}
669
77f012e8
SB
670static void read_info_alternates(struct repository *r,
671 const char *relative_base,
672 int depth)
d5a63b99 673{
5015f01c 674 char *path;
dc732bd5 675 struct strbuf buf = STRBUF_INIT;
d5a63b99 676
5015f01c 677 path = xstrfmt("%s/info/alternates", relative_base);
dc732bd5 678 if (strbuf_read_file(&buf, path, 1024) < 0) {
f0f7bebe 679 warn_on_fopen_errors(path);
dc732bd5 680 free(path);
9a217f2a 681 return;
ace1534d 682 }
d5a63b99 683
77f012e8 684 link_alt_odb_entries(r, buf.buf, '\n', relative_base, depth);
dc732bd5
JK
685 strbuf_release(&buf);
686 free(path);
ace1534d
JH
687}
688
bef70b22
DB
689void add_to_alternates_file(const char *reference)
690{
f132a127 691 struct lock_file lock = LOCK_INIT;
77b9b1d1
JK
692 char *alts = git_pathdup("objects/info/alternates");
693 FILE *in, *out;
f132a127 694 int found = 0;
77b9b1d1 695
f132a127
696 hold_lock_file_for_update(&lock, alts, LOCK_DIE_ON_ERROR);
697 out = fdopen_lock_file(&lock, "w");
77b9b1d1 698 if (!out)
259328b7 699 die_errno(_("unable to fdopen alternates lockfile"));
77b9b1d1
JK
700
701 in = fopen(alts, "r");
702 if (in) {
703 struct strbuf line = STRBUF_INIT;
77b9b1d1 704
3f163962 705 while (strbuf_getline(&line, in) != EOF) {
77b9b1d1
JK
706 if (!strcmp(reference, line.buf)) {
707 found = 1;
708 break;
709 }
710 fprintf_or_die(out, "%s\n", line.buf);
711 }
712
713 strbuf_release(&line);
714 fclose(in);
77b9b1d1
JK
715 }
716 else if (errno != ENOENT)
259328b7 717 die_errno(_("unable to read alternates file"));
77b9b1d1 718
f132a127
719 if (found) {
720 rollback_lock_file(&lock);
721 } else {
77b9b1d1 722 fprintf_or_die(out, "%s\n", reference);
f132a127 723 if (commit_lock_file(&lock))
259328b7 724 die_errno(_("unable to move new alternates file into place"));
f0eaf638 725 if (the_repository->objects->loaded_alternates)
93d8d1e2
SB
726 link_alt_odb_entries(the_repository, reference,
727 '\n', NULL, 0);
77b9b1d1
JK
728 }
729 free(alts);
bef70b22
DB
730}
731
a5b34d21
JK
732void add_to_alternates_memory(const char *reference)
733{
734 /*
735 * Make sure alternates are initialized, or else our entry may be
736 * overwritten when they are.
737 */
0b209034 738 prepare_alt_odb(the_repository);
a5b34d21 739
93d8d1e2
SB
740 link_alt_odb_entries(the_repository, reference,
741 '\n', NULL, 0);
a5b34d21
JK
742}
743
9eeea7d2
SB
744/*
745 * Compute the exact path an alternate is at and returns it. In case of
746 * error NULL is returned and the human readable error is added to `err`
efde7b72 747 * `path` may be relative and should point to $GIT_DIR.
9eeea7d2
SB
748 * `err` must not be null.
749 */
750char *compute_alternate_path(const char *path, struct strbuf *err)
751{
752 char *ref_git = NULL;
4530a85b 753 const char *repo;
9eeea7d2
SB
754 int seen_error = 0;
755
4530a85b
AM
756 ref_git = real_pathdup(path, 0);
757 if (!ref_git) {
9eeea7d2
SB
758 seen_error = 1;
759 strbuf_addf(err, _("path '%s' does not exist"), path);
760 goto out;
4530a85b 761 }
9eeea7d2
SB
762
763 repo = read_gitfile(ref_git);
764 if (!repo)
765 repo = read_gitfile(mkpath("%s/.git", ref_git));
766 if (repo) {
767 free(ref_git);
768 ref_git = xstrdup(repo);
769 }
770
771 if (!repo && is_directory(mkpath("%s/.git/objects", ref_git))) {
772 char *ref_git_git = mkpathdup("%s/.git", ref_git);
773 free(ref_git);
774 ref_git = ref_git_git;
775 } else if (!is_directory(mkpath("%s/objects", ref_git))) {
776 struct strbuf sb = STRBUF_INIT;
777 seen_error = 1;
778 if (get_common_dir(&sb, ref_git)) {
779 strbuf_addf(err,
780 _("reference repository '%s' as a linked "
781 "checkout is not supported yet."),
782 path);
783 goto out;
784 }
785
786 strbuf_addf(err, _("reference repository '%s' is not a "
787 "local repository."), path);
788 goto out;
789 }
790
791 if (!access(mkpath("%s/shallow", ref_git), F_OK)) {
792 strbuf_addf(err, _("reference repository '%s' is shallow"),
793 path);
794 seen_error = 1;
795 goto out;
796 }
797
798 if (!access(mkpath("%s/info/grafts", ref_git), F_OK)) {
799 strbuf_addf(err,
800 _("reference repository '%s' is grafted"),
801 path);
802 seen_error = 1;
803 goto out;
804 }
805
806out:
807 if (seen_error) {
6a83d902 808 FREE_AND_NULL(ref_git);
9eeea7d2
SB
809 }
810
811 return ref_git;
812}
813
709dfa69
JK
814static void fill_alternate_refs_command(struct child_process *cmd,
815 const char *repo_path)
816{
817 const char *value;
818
819 if (!git_config_get_value("core.alternateRefsCommand", &value)) {
820 cmd->use_shell = 1;
821
c972bf4c
JK
822 strvec_push(&cmd->args, value);
823 strvec_push(&cmd->args, repo_path);
709dfa69
JK
824 } else {
825 cmd->git_cmd = 1;
826
c972bf4c
JK
827 strvec_pushf(&cmd->args, "--git-dir=%s", repo_path);
828 strvec_push(&cmd->args, "for-each-ref");
829 strvec_push(&cmd->args, "--format=%(objectname)");
709dfa69
JK
830
831 if (!git_config_get_value("core.alternateRefsPrefixes", &value)) {
c972bf4c
JK
832 strvec_push(&cmd->args, "--");
833 strvec_split(&cmd->args, value);
709dfa69
JK
834 }
835 }
836
837 cmd->env = local_repo_env;
838 cmd->out = -1;
839}
840
841static void read_alternate_refs(const char *path,
842 alternate_ref_fn *cb,
843 void *data)
844{
845 struct child_process cmd = CHILD_PROCESS_INIT;
846 struct strbuf line = STRBUF_INIT;
847 FILE *fh;
848
849 fill_alternate_refs_command(&cmd, path);
850
851 if (start_command(&cmd))
852 return;
853
854 fh = xfdopen(cmd.out, "r");
855 while (strbuf_getline_lf(&line, fh) != EOF) {
856 struct object_id oid;
857 const char *p;
858
859 if (parse_oid_hex(line.buf, &oid, &p) || *p) {
860 warning(_("invalid line while parsing alternate refs: %s"),
861 line.buf);
862 break;
863 }
864
865 cb(&oid, data);
866 }
867
868 fclose(fh);
869 finish_command(&cmd);
86ad3ea5 870 strbuf_release(&line);
709dfa69
JK
871}
872
873struct alternate_refs_data {
874 alternate_ref_fn *fn;
875 void *data;
876};
877
878static int refs_from_alternate_cb(struct object_directory *e,
879 void *data)
880{
881 struct strbuf path = STRBUF_INIT;
882 size_t base_len;
883 struct alternate_refs_data *cb = data;
884
885 if (!strbuf_realpath(&path, e->path, 0))
886 goto out;
887 if (!strbuf_strip_suffix(&path, "/objects"))
888 goto out;
889 base_len = path.len;
890
891 /* Is this a git repository with refs? */
892 strbuf_addstr(&path, "/refs");
893 if (!is_directory(path.buf))
894 goto out;
895 strbuf_setlen(&path, base_len);
896
897 read_alternate_refs(path.buf, cb->fn, cb->data);
898
899out:
900 strbuf_release(&path);
901 return 0;
902}
903
904void for_each_alternate_ref(alternate_ref_fn fn, void *data)
905{
906 struct alternate_refs_data cb;
907 cb.fn = fn;
908 cb.data = data;
909 foreach_alt_odb(refs_from_alternate_cb, &cb);
910}
911
fe1b2268 912int foreach_alt_odb(alt_odb_fn fn, void *cb)
d79796bc 913{
263db403 914 struct object_directory *ent;
fe1b2268 915 int r = 0;
d79796bc 916
0b209034 917 prepare_alt_odb(the_repository);
f0eaf638 918 for (ent = the_repository->objects->odb->next; ent; ent = ent->next) {
fe1b2268
JK
919 r = fn(ent, cb);
920 if (r)
921 break;
922 }
923 return r;
d79796bc
JH
924}
925
13068bf0 926void prepare_alt_odb(struct repository *r)
c2f493a4 927{
f0eaf638 928 if (r->objects->loaded_alternates)
7dc24aa5
SP
929 return;
930
13068bf0 931 link_alt_odb_entries(r, r->objects->alternate_db, PATH_SEP, NULL, 0);
c2f493a4 932
f0eaf638
JK
933 read_info_alternates(r, r->objects->odb->path, 0);
934 r->objects->loaded_alternates = 1;
c2f493a4
MW
935}
936
3096b2ec 937/* Returns 1 if we have successfully freshened the file, 0 otherwise. */
33d4221c 938static int freshen_file(const char *fn)
ace1534d 939{
312cd761 940 return !utime(fn, NULL);
0f4dc14a 941}
ace1534d 942
3096b2ec
JK
943/*
944 * All of the check_and_freshen functions return 1 if the file exists and was
945 * freshened (if freshening was requested), 0 otherwise. If they return
946 * 0, you should not assume that it is safe to skip a write of the object (it
947 * either does not exist on disk, or has a stale mtime and may be subject to
948 * pruning).
949 */
6a5e6f5e 950int check_and_freshen_file(const char *fn, int freshen)
33d4221c
JK
951{
952 if (access(fn, F_OK))
953 return 0;
3096b2ec 954 if (freshen && !freshen_file(fn))
33d4221c
JK
955 return 0;
956 return 1;
957}
958
f0eaf638
JK
959static int check_and_freshen_odb(struct object_directory *odb,
960 const struct object_id *oid,
961 int freshen)
33d4221c 962{
f0eaf638 963 static struct strbuf path = STRBUF_INIT;
514c5fdd 964 odb_loose_path(odb, &path, oid);
f0eaf638
JK
965 return check_and_freshen_file(path.buf, freshen);
966}
ea657730 967
f0eaf638
JK
968static int check_and_freshen_local(const struct object_id *oid, int freshen)
969{
970 return check_and_freshen_odb(the_repository->objects->odb, oid, freshen);
33d4221c
JK
971}
972
6862ebbf 973static int check_and_freshen_nonlocal(const struct object_id *oid, int freshen)
0f4dc14a 974{
263db403 975 struct object_directory *odb;
f3f043a1 976
0b209034 977 prepare_alt_odb(the_repository);
f0eaf638
JK
978 for (odb = the_repository->objects->odb->next; odb; odb = odb->next) {
979 if (check_and_freshen_odb(odb, oid, freshen))
c529d75a 980 return 1;
ace1534d 981 }
c529d75a 982 return 0;
ace1534d
JH
983}
984
6862ebbf 985static int check_and_freshen(const struct object_id *oid, int freshen)
33d4221c 986{
6862ebbf 987 return check_and_freshen_local(oid, freshen) ||
988 check_and_freshen_nonlocal(oid, freshen);
33d4221c
JK
989}
990
6862ebbf 991int has_loose_object_nonlocal(const struct object_id *oid)
33d4221c 992{
6862ebbf 993 return check_and_freshen_nonlocal(oid, 0);
33d4221c
JK
994}
995
6862ebbf 996static int has_loose_object(const struct object_id *oid)
0f4dc14a 997{
6862ebbf 998 return check_and_freshen(oid, 0);
0f4dc14a
BC
999}
1000
02710228
SP
1001static void mmap_limit_check(size_t length)
1002{
1003 static size_t limit = 0;
1004 if (!limit) {
1005 limit = git_env_ulong("GIT_MMAP_LIMIT", 0);
1006 if (!limit)
1007 limit = SIZE_MAX;
1008 }
1009 if (length > limit)
259328b7 1010 die(_("attempting to mmap %"PRIuMAX" over limit %"PRIuMAX),
02710228
SP
1011 (uintmax_t)length, (uintmax_t)limit);
1012}
1013
1570856b
JK
1014void *xmmap_gently(void *start, size_t length,
1015 int prot, int flags, int fd, off_t offset)
58ecbd5e 1016{
02710228
SP
1017 void *ret;
1018
1019 mmap_limit_check(length);
1020 ret = mmap(start, length, prot, flags, fd, offset);
9827d4c1
JK
1021 if (ret == MAP_FAILED && !length)
1022 ret = NULL;
58ecbd5e
JN
1023 return ret;
1024}
1025
dc059294
EW
1026const char *mmap_os_err(void)
1027{
1028 static const char blank[] = "";
1029#if defined(__linux__)
1030 if (errno == ENOMEM) {
1031 /* this continues an existing error message: */
1032 static const char enomem[] =
1033", check sys.vm.max_map_count and/or RLIMIT_DATA";
1034 return enomem;
1035 }
1036#endif /* OS-specific bits */
1037 return blank;
1038}
1039
1570856b
JK
1040void *xmmap(void *start, size_t length,
1041 int prot, int flags, int fd, off_t offset)
1042{
1043 void *ret = xmmap_gently(start, length, prot, flags, fd, offset);
1044 if (ret == MAP_FAILED)
dc059294 1045 die_errno(_("mmap failed%s"), mmap_os_err());
1570856b
JK
1046 return ret;
1047}
1048
88d0db55 1049/*
090ea126 1050 * With an in-core object data in "map", rehash it to make sure the
cb1c8d1d
JK
1051 * object name actually matches "oid" to detect object corruption.
1052 * With "map" == NULL, try reading the object named with "oid" using
090ea126 1053 * the streaming interface and rehash it to do the same.
88d0db55 1054 */
b98d1885
MT
1055int check_object_signature(struct repository *r, const struct object_id *oid,
1056 void *map, unsigned long size, const char *type)
88d0db55 1057{
f070facc 1058 struct object_id real_oid;
090ea126
NTND
1059 enum object_type obj_type;
1060 struct git_istream *st;
18e2588e 1061 git_hash_ctx c;
1af64f73 1062 char hdr[MAX_HEADER_LEN];
090ea126 1063 int hdrlen;
88d0db55 1064
090ea126 1065 if (map) {
b98d1885 1066 hash_object_file(r->hash_algo, map, size, type, &real_oid);
9001dc2a 1067 return !oideq(oid, &real_oid) ? -1 : 0;
88d0db55
BC
1068 }
1069
b98d1885 1070 st = open_istream(r, oid, &obj_type, &size, NULL);
090ea126
NTND
1071 if (!st)
1072 return -1;
88d0db55 1073
090ea126 1074 /* Generate the header */
ca473cef 1075 hdrlen = xsnprintf(hdr, sizeof(hdr), "%s %"PRIuMAX , type_name(obj_type), (uintmax_t)size) + 1;
88d0db55 1076
090ea126 1077 /* Sha1.. */
b98d1885
MT
1078 r->hash_algo->init_fn(&c);
1079 r->hash_algo->update_fn(&c, hdr, hdrlen);
090ea126
NTND
1080 for (;;) {
1081 char buf[1024 * 16];
1082 ssize_t readlen = read_istream(st, buf, sizeof(buf));
1f688557 1083
f54fac53
JK
1084 if (readlen < 0) {
1085 close_istream(st);
1086 return -1;
1087 }
090ea126
NTND
1088 if (!readlen)
1089 break;
b98d1885 1090 r->hash_algo->update_fn(&c, buf, readlen);
fa5fc15d 1091 }
5951bf46 1092 r->hash_algo->final_oid_fn(&real_oid, &c);
090ea126 1093 close_istream(st);
9001dc2a 1094 return !oideq(oid, &real_oid) ? -1 : 0;
fa5fc15d
SP
1095}
1096
1b8ac5ea 1097int git_open_cloexec(const char *name, int flags)
a0788266 1098{
1e3001a8
JH
1099 int fd;
1100 static int o_cloexec = O_CLOEXEC;
a0788266 1101
1e3001a8
JH
1102 fd = open(name, flags | o_cloexec);
1103 if ((o_cloexec & O_CLOEXEC) && fd < 0 && errno == EINVAL) {
cd66ada0 1104 /* Try again w/o O_CLOEXEC: the kernel might not support it */
1e3001a8
JH
1105 o_cloexec &= ~O_CLOEXEC;
1106 fd = open(name, flags | o_cloexec);
491a8dec 1107 }
491a8dec 1108
9fb9495d 1109#if defined(F_GETFD) && defined(F_SETFD) && defined(FD_CLOEXEC)
491a8dec 1110 {
1e3001a8 1111 static int fd_cloexec = FD_CLOEXEC;
a0788266 1112
1e3001a8
JH
1113 if (!o_cloexec && 0 <= fd && fd_cloexec) {
1114 /* Opened w/o O_CLOEXEC? try with fcntl(2) to add it */
9fb9495d
EW
1115 int flags = fcntl(fd, F_GETFD);
1116 if (fcntl(fd, F_SETFD, flags | fd_cloexec))
1e3001a8 1117 fd_cloexec = 0;
cd66ada0 1118 }
44d1c19e 1119 }
a0788266 1120#endif
1b8ac5ea 1121 return fd;
a0788266
JS
1122}
1123
3cf8b462 1124/*
514c5fdd 1125 * Find "oid" as a loose object in the local repository or in an alternate.
771e7d57
JK
1126 * Returns 0 on success, negative on failure.
1127 *
1128 * The "path" out-parameter will give the path of the object we found (if any).
1129 * Note that it may point to static storage and is only valid until another
514c5fdd 1130 * call to stat_loose_object().
3cf8b462 1131 */
514c5fdd
JK
1132static int stat_loose_object(struct repository *r, const struct object_id *oid,
1133 struct stat *st, const char **path)
1f688557 1134{
263db403 1135 struct object_directory *odb;
ea657730
CC
1136 static struct strbuf buf = STRBUF_INIT;
1137
d2607fa0 1138 prepare_alt_odb(r);
f0eaf638 1139 for (odb = r->objects->odb; odb; odb = odb->next) {
514c5fdd 1140 *path = odb_loose_path(odb, &buf, oid);
771e7d57 1141 if (!lstat(*path, st))
052fe5ea 1142 return 0;
c7934306
SP
1143 }
1144
3cf8b462
SP
1145 return -1;
1146}
1147
771e7d57 1148/*
514c5fdd 1149 * Like stat_loose_object(), but actually open the object and return the
771e7d57
JK
1150 * descriptor. See the caveats on the "path" parameter above.
1151 */
514c5fdd
JK
1152static int open_loose_object(struct repository *r,
1153 const struct object_id *oid, const char **path)
60bb8b14 1154{
44d1c19e 1155 int fd;
263db403 1156 struct object_directory *odb;
f0eaf638 1157 int most_interesting_errno = ENOENT;
ea657730
CC
1158 static struct strbuf buf = STRBUF_INIT;
1159
ec7283e5 1160 prepare_alt_odb(r);
f0eaf638 1161 for (odb = r->objects->odb; odb; odb = odb->next) {
514c5fdd 1162 *path = odb_loose_path(odb, &buf, oid);
771e7d57 1163 fd = git_open(*path);
44d1c19e
LT
1164 if (fd >= 0)
1165 return fd;
f0eaf638 1166
d6c8a05b
JK
1167 if (most_interesting_errno == ENOENT)
1168 most_interesting_errno = errno;
03e79c88 1169 }
d6c8a05b 1170 errno = most_interesting_errno;
44d1c19e 1171 return -1;
1f688557
JH
1172}
1173
61c7711c 1174static int quick_has_loose(struct repository *r,
d7a24573 1175 const struct object_id *oid)
61c7711c 1176{
61c7711c
JK
1177 struct object_directory *odb;
1178
61c7711c
JK
1179 prepare_alt_odb(r);
1180 for (odb = r->objects->odb; odb; odb = odb->next) {
d7a24573 1181 if (oid_array_lookup(odb_loose_cache(odb, oid), oid) >= 0)
61c7711c
JK
1182 return 1;
1183 }
1184 return 0;
1185}
1186
f6371f92
JK
1187/*
1188 * Map the loose object at "path" if it is not NULL, or the path found by
514c5fdd 1189 * searching for a loose object named "oid".
f6371f92 1190 */
514c5fdd
JK
1191static void *map_loose_object_1(struct repository *r, const char *path,
1192 const struct object_id *oid, unsigned long *size)
27d69a46 1193{
0fcfd160 1194 void *map;
144bde78 1195 int fd;
84dd81c1 1196
f6371f92
JK
1197 if (path)
1198 fd = git_open(path);
8261e1f1 1199 else
514c5fdd 1200 fd = open_loose_object(r, oid, &path);
44d1c19e
LT
1201 map = NULL;
1202 if (fd >= 0) {
1203 struct stat st;
1f688557 1204
44d1c19e
LT
1205 if (!fstat(fd, &st)) {
1206 *size = xsize_t(st.st_size);
33e42de0
MM
1207 if (!*size) {
1208 /* mmap() is forbidden on empty files */
259328b7 1209 error(_("object file %s is empty"), path);
6881925e 1210 close(fd);
33e42de0
MM
1211 return NULL;
1212 }
44d1c19e 1213 map = xmmap(NULL, *size, PROT_READ, MAP_PRIVATE, fd, 0);
144bde78 1214 }
44d1c19e 1215 close(fd);
74e34e1f 1216 }
0fcfd160 1217 return map;
74e34e1f
NP
1218}
1219
514c5fdd
JK
1220void *map_loose_object(struct repository *r,
1221 const struct object_id *oid,
1222 unsigned long *size)
068f85e3 1223{
514c5fdd 1224 return map_loose_object_1(r, NULL, oid, size);
068f85e3 1225}
1226
00a7760e
JK
1227static int unpack_loose_short_header(git_zstream *stream,
1228 unsigned char *map, unsigned long mapsize,
1229 void *buffer, unsigned long bufsiz)
47fe3f6e 1230{
31877c9a
MT
1231 int ret;
1232
c4483576
LT
1233 /* Get the data stream */
1234 memset(stream, 0, sizeof(*stream));
1235 stream->next_in = map;
1236 stream->avail_in = mapsize;
1237 stream->next_out = buffer;
93821bd9 1238 stream->avail_out = bufsiz;
47fe3f6e 1239
39c68542 1240 git_inflate_init(stream);
31877c9a
MT
1241 obj_read_unlock();
1242 ret = git_inflate(stream, 0);
1243 obj_read_lock();
1244
1245 return ret;
9a217f2a
JH
1246}
1247
00a7760e
JK
1248int unpack_loose_header(git_zstream *stream,
1249 unsigned char *map, unsigned long mapsize,
1250 void *buffer, unsigned long bufsiz)
1f688557 1251{
00a7760e
JK
1252 int status = unpack_loose_short_header(stream, map, mapsize,
1253 buffer, bufsiz);
42873078 1254
d21f8426
JH
1255 if (status < Z_OK)
1256 return status;
628522ec 1257
d21f8426
JH
1258 /* Make sure we have the terminating NUL */
1259 if (!memchr(buffer, '\0', stream->next_out - (unsigned char *)buffer))
1260 return -1;
1f688557
JH
1261 return 0;
1262}
1263
00a7760e
JK
1264static int unpack_loose_header_to_strbuf(git_zstream *stream, unsigned char *map,
1265 unsigned long mapsize, void *buffer,
1266 unsigned long bufsiz, struct strbuf *header)
d131b7af 1267{
46f03448
KN
1268 int status;
1269
00a7760e 1270 status = unpack_loose_short_header(stream, map, mapsize, buffer, bufsiz);
d21f8426
JH
1271 if (status < Z_OK)
1272 return -1;
d131b7af 1273
46f03448
KN
1274 /*
1275 * Check if entire header is unpacked in the first iteration.
d131b7af 1276 */
46f03448
KN
1277 if (memchr(buffer, '\0', stream->next_out - (unsigned char *)buffer))
1278 return 0;
d131b7af 1279
46f03448
KN
1280 /*
1281 * buffer[0..bufsiz] was not large enough. Copy the partial
1282 * result out to header, and then append the result of further
1283 * reading the stream.
1284 */
1285 strbuf_add(header, buffer, stream->next_out - (unsigned char *)buffer);
1286 stream->next_out = buffer;
1287 stream->avail_out = bufsiz;
d131b7af 1288
46f03448 1289 do {
31877c9a 1290 obj_read_unlock();
46f03448 1291 status = git_inflate(stream, 0);
31877c9a 1292 obj_read_lock();
46f03448
KN
1293 strbuf_add(header, buffer, stream->next_out - (unsigned char *)buffer);
1294 if (memchr(buffer, '\0', stream->next_out - (unsigned char *)buffer))
1295 return 0;
1296 stream->next_out = buffer;
1297 stream->avail_out = bufsiz;
1298 } while (status != Z_STREAM_END);
1299 return -1;
d131b7af
SP
1300}
1301
00a7760e
JK
1302static void *unpack_loose_rest(git_zstream *stream,
1303 void *buffer, unsigned long size,
1304 const struct object_id *oid)
95099731 1305{
5180cacc 1306 int bytes = strlen(buffer) + 1;
3aee68aa 1307 unsigned char *buf = xmallocz(size);
93821bd9 1308 unsigned long n;
7efbff75 1309 int status = Z_OK;
95099731 1310
93821bd9
LT
1311 n = stream->total_out - bytes;
1312 if (n > size)
1313 n = size;
1314 memcpy(buf, (char *) buffer + bytes, n);
1315 bytes = n;
456cdf6e
LT
1316 if (bytes <= size) {
1317 /*
1318 * The above condition must be (bytes <= size), not
1319 * (bytes < size). In other words, even though we
ccf5ace0 1320 * expect no more output and set avail_out to zero,
456cdf6e
LT
1321 * the input zlib stream may have bytes that express
1322 * "this concludes the stream", and we *do* want to
1323 * eat that input.
1324 *
1325 * Otherwise we would not be able to test that we
1326 * consumed all the input to reach the expected size;
1327 * we also want to check that zlib tells us that all
1328 * went well with status == Z_STREAM_END at the end.
1329 */
5180cacc
LT
1330 stream->next_out = buf + bytes;
1331 stream->avail_out = size - bytes;
31877c9a
MT
1332 while (status == Z_OK) {
1333 obj_read_unlock();
39c68542 1334 status = git_inflate(stream, Z_FINISH);
31877c9a
MT
1335 obj_read_lock();
1336 }
5180cacc 1337 }
456cdf6e 1338 if (status == Z_STREAM_END && !stream->avail_in) {
39c68542 1339 git_inflate_end(stream);
7efbff75 1340 return buf;
95099731
NTND
1341 }
1342
7efbff75 1343 if (status < 0)
00a7760e 1344 error(_("corrupt loose object '%s'"), oid_to_hex(oid));
7efbff75 1345 else if (stream->avail_in)
259328b7 1346 error(_("garbage at end of loose object '%s'"),
00a7760e 1347 oid_to_hex(oid));
7efbff75
JH
1348 free(buf);
1349 return NULL;
95099731
NTND
1350}
1351
d40d535b 1352/*
5180cacc
LT
1353 * We used to just use "sscanf()", but that's actually way
1354 * too permissive for what we want to check. So do an anal
1355 * object header parse by hand.
d40d535b 1356 */
00a7760e
JK
1357static int parse_loose_header_extended(const char *hdr, struct object_info *oi,
1358 unsigned int flags)
1f688557 1359{
46f03448 1360 const char *type_buf = hdr;
5180cacc 1361 unsigned long size;
46f03448 1362 int type, type_len = 0;
43057304 1363
5180cacc 1364 /*
46f03448 1365 * The type can be of any size but is followed by
21666f1a 1366 * a space.
5180cacc 1367 */
5180cacc
LT
1368 for (;;) {
1369 char c = *hdr++;
d21f8426
JH
1370 if (!c)
1371 return -1;
5180cacc
LT
1372 if (c == ' ')
1373 break;
46f03448 1374 type_len++;
5180cacc 1375 }
1f688557 1376
46f03448 1377 type = type_from_string_gently(type_buf, type_len, 1);
6ca32f47
BW
1378 if (oi->type_name)
1379 strbuf_add(oi->type_name, type_buf, type_len);
46f03448
KN
1380 /*
1381 * Set type to 0 if its an unknown object and
2e3a16b2 1382 * we're obtaining the type using '--allow-unknown-type'
46f03448
KN
1383 * option.
1384 */
19fc5e84 1385 if ((flags & OBJECT_INFO_ALLOW_UNKNOWN_TYPE) && (type < 0))
46f03448
KN
1386 type = 0;
1387 else if (type < 0)
259328b7 1388 die(_("invalid object type"));
46f03448
KN
1389 if (oi->typep)
1390 *oi->typep = type;
5180cacc
LT
1391
1392 /*
1393 * The length must follow immediately, and be in canonical
1394 * decimal format (ie "010" is not valid).
1395 */
1396 size = *hdr++ - '0';
1397 if (size > 9)
1398 return -1;
1399 if (size) {
1400 for (;;) {
1401 unsigned long c = *hdr - '0';
1402 if (c > 9)
1403 break;
1404 hdr++;
1405 size = size * 10 + c;
1b1005d1 1406 }
c01f51cc 1407 }
46f03448
KN
1408
1409 if (oi->sizep)
1410 *oi->sizep = size;
5180cacc
LT
1411
1412 /*
1413 * The length must be followed by a zero byte
1414 */
46f03448 1415 return *hdr ? -1 : type;
1f688557
JH
1416}
1417
00a7760e 1418int parse_loose_header(const char *hdr, unsigned long *sizep)
bf592c50 1419{
27b5c1a0 1420 struct object_info oi = OBJECT_INFO_INIT;
21666f1a 1421
46f03448 1422 oi.sizep = sizep;
00a7760e 1423 return parse_loose_header_extended(hdr, &oi, 0);
bf592c50
DB
1424}
1425
514c5fdd
JK
1426static int loose_object_info(struct repository *r,
1427 const struct object_id *oid,
1428 struct object_info *oi, int flags)
65c2e0c3 1429{
46f03448
KN
1430 int status = 0;
1431 unsigned long mapsize;
65c2e0c3 1432 void *map;
ef49a7a0 1433 git_zstream stream;
1af64f73 1434 char hdr[MAX_HEADER_LEN];
46f03448 1435 struct strbuf hdrbuf = STRBUF_INIT;
c84a1f3e 1436 unsigned long size_scratch;
65c2e0c3 1437
b99b6bcc
JK
1438 if (oi->delta_base_oid)
1439 oidclr(oi->delta_base_oid);
5d642e75 1440
052fe5ea
JK
1441 /*
1442 * If we don't care about type or size, then we don't
4ef8d1dd
JH
1443 * need to look inside the object at all. Note that we
1444 * do not optimize out the stat call, even if the
1445 * caller doesn't care about the disk-size, since our
1446 * return value implicitly indicates whether the
1447 * object even exists.
052fe5ea 1448 */
6ca32f47 1449 if (!oi->typep && !oi->type_name && !oi->sizep && !oi->contentp) {
771e7d57 1450 const char *path;
4ef8d1dd 1451 struct stat st;
61c7711c 1452 if (!oi->disk_sizep && (flags & OBJECT_INFO_QUICK))
d7a24573 1453 return quick_has_loose(r, oid) ? 0 : -1;
514c5fdd 1454 if (stat_loose_object(r, oid, &st, &path) < 0)
4ef8d1dd
JH
1455 return -1;
1456 if (oi->disk_sizep)
23c339c0 1457 *oi->disk_sizep = st.st_size;
052fe5ea
JK
1458 return 0;
1459 }
1460
514c5fdd 1461 map = map_loose_object(r, oid, &mapsize);
f0df4ed5 1462 if (!map)
dbea72a8 1463 return -1;
c84a1f3e
JT
1464
1465 if (!oi->sizep)
1466 oi->sizep = &size_scratch;
1467
23c339c0
JK
1468 if (oi->disk_sizep)
1469 *oi->disk_sizep = mapsize;
19fc5e84 1470 if ((flags & OBJECT_INFO_ALLOW_UNKNOWN_TYPE)) {
00a7760e 1471 if (unpack_loose_header_to_strbuf(&stream, map, mapsize, hdr, sizeof(hdr), &hdrbuf) < 0)
259328b7 1472 status = error(_("unable to unpack %s header with --allow-unknown-type"),
514c5fdd 1473 oid_to_hex(oid));
00a7760e 1474 } else if (unpack_loose_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0)
259328b7 1475 status = error(_("unable to unpack %s header"),
514c5fdd 1476 oid_to_hex(oid));
46f03448
KN
1477 if (status < 0)
1478 ; /* Do nothing */
1479 else if (hdrbuf.len) {
00a7760e 1480 if ((status = parse_loose_header_extended(hdrbuf.buf, oi, flags)) < 0)
259328b7 1481 status = error(_("unable to parse %s header with --allow-unknown-type"),
514c5fdd 1482 oid_to_hex(oid));
00a7760e 1483 } else if ((status = parse_loose_header_extended(hdr, oi, flags)) < 0)
514c5fdd 1484 status = error(_("unable to parse %s header"), oid_to_hex(oid));
c84a1f3e 1485
b3ea7dd3 1486 if (status >= 0 && oi->contentp) {
00a7760e
JK
1487 *oi->contentp = unpack_loose_rest(&stream, hdr,
1488 *oi->sizep, oid);
b3ea7dd3
JK
1489 if (!*oi->contentp) {
1490 git_inflate_end(&stream);
1491 status = -1;
1492 }
1493 } else
c84a1f3e
JT
1494 git_inflate_end(&stream);
1495
65c2e0c3 1496 munmap(map, mapsize);
46f03448 1497 if (status && oi->typep)
23c339c0 1498 *oi->typep = status;
c84a1f3e
JT
1499 if (oi->sizep == &size_scratch)
1500 oi->sizep = NULL;
46f03448 1501 strbuf_release(&hdrbuf);
3ab0fb06 1502 oi->whence = OI_LOOSE;
93cff9a9 1503 return (status < 0) ? status : 0;
65c2e0c3
JH
1504}
1505
31877c9a
MT
1506int obj_read_use_lock = 0;
1507pthread_mutex_t obj_read_mutex;
1508
1509void enable_obj_read_lock(void)
1510{
1511 if (obj_read_use_lock)
1512 return;
1513
1514 obj_read_use_lock = 1;
1515 init_recursive_mutex(&obj_read_mutex);
1516}
1517
1518void disable_obj_read_lock(void)
1519{
1520 if (!obj_read_use_lock)
1521 return;
1522
1523 obj_read_use_lock = 0;
1524 pthread_mutex_destroy(&obj_read_mutex);
1525}
1526
8b4c0103
JT
1527int fetch_if_missing = 1;
1528
31877c9a
MT
1529static int do_oid_object_info_extended(struct repository *r,
1530 const struct object_id *oid,
1531 struct object_info *oi, unsigned flags)
f0df4ed5 1532{
cd585e2a 1533 static struct object_info blank_oi = OBJECT_INFO_INIT;
9c8a294a 1534 struct cached_object *co;
f0df4ed5 1535 struct pack_entry e;
5b086407 1536 int rtype;
b383a13c 1537 const struct object_id *real = oid;
8b4c0103 1538 int already_retried = 0;
f0df4ed5 1539
31877c9a 1540
b383a13c 1541 if (flags & OBJECT_INFO_LOOKUP_REPLACE)
9d98354f 1542 real = lookup_replace_object(r, oid);
f0df4ed5 1543
b383a13c 1544 if (is_null_oid(real))
87b5e236
JK
1545 return -1;
1546
cd585e2a
JT
1547 if (!oi)
1548 oi = &blank_oi;
1549
9c8a294a
JT
1550 co = find_cached_object(real);
1551 if (co) {
1552 if (oi->typep)
1553 *(oi->typep) = co->type;
1554 if (oi->sizep)
1555 *(oi->sizep) = co->size;
1556 if (oi->disk_sizep)
1557 *(oi->disk_sizep) = 0;
b99b6bcc
JK
1558 if (oi->delta_base_oid)
1559 oidclr(oi->delta_base_oid);
9c8a294a
JT
1560 if (oi->type_name)
1561 strbuf_addstr(oi->type_name, type_name(co->type));
1562 if (oi->contentp)
1563 *oi->contentp = xmemdupz(co->buf, co->size);
1564 oi->whence = OI_CACHED;
1565 return 0;
c4d9986f
NTND
1566 }
1567
8b4c0103 1568 while (1) {
42c8ce1c 1569 if (find_pack_entry(r, real, &e))
8b4c0103
JT
1570 break;
1571
024aa469
TI
1572 if (flags & OBJECT_INFO_IGNORE_LOOSE)
1573 return -1;
1574
ddd63e64 1575 /* Most likely it's a loose object. */
514c5fdd 1576 if (!loose_object_info(r, real, oi, flags))
5b086407 1577 return 0;
ddd63e64
SG
1578
1579 /* Not a loose object; someone else may have just packed it. */
2b7750c9 1580 if (!(flags & OBJECT_INFO_QUICK)) {
9d98354f 1581 reprepare_packed_git(r);
42c8ce1c 1582 if (find_pack_entry(r, real, &e))
2b7750c9
JT
1583 break;
1584 }
8b4c0103
JT
1585
1586 /* Check if it is a missing object */
b14ed5ad 1587 if (fetch_if_missing && has_promisor_remote() &&
0f4a4fb1 1588 !already_retried && r == the_repository &&
31f5256c 1589 !(flags & OBJECT_INFO_SKIP_FETCH_OBJECT)) {
8b4c0103 1590 /*
b14ed5ad
CC
1591 * TODO Investigate checking promisor_remote_get_direct()
1592 * TODO return value and stopping on error here.
1593 * TODO Pass a repository struct through
1594 * promisor_remote_get_direct(), such that arbitrary
1595 * repositories work.
8b4c0103 1596 */
b14ed5ad 1597 promisor_remote_get_direct(r, real, 1);
8b4c0103
JT
1598 already_retried = 1;
1599 continue;
dfdd4afc 1600 }
8b4c0103
JT
1601
1602 return -1;
f0df4ed5 1603 }
3d77d877 1604
cd585e2a
JT
1605 if (oi == &blank_oi)
1606 /*
1607 * We know that the caller doesn't actually need the
1608 * information below, so return early.
1609 */
1610 return 0;
9d98354f 1611 rtype = packed_object_info(r, e.p, e.offset, oi);
412916ee 1612 if (rtype < 0) {
b383a13c 1613 mark_bad_packed_object(e.p, real->hash);
31877c9a 1614 return do_oid_object_info_extended(r, real, oi, 0);
3ab0fb06 1615 } else if (oi->whence == OI_PACKED) {
9a490590
JH
1616 oi->u.packed.offset = e.offset;
1617 oi->u.packed.pack = e.p;
1618 oi->u.packed.is_delta = (rtype == OBJ_REF_DELTA ||
1619 rtype == OBJ_OFS_DELTA);
3d77d877
NP
1620 }
1621
5b086407 1622 return 0;
f0df4ed5
JS
1623}
1624
31877c9a
MT
1625int oid_object_info_extended(struct repository *r, const struct object_id *oid,
1626 struct object_info *oi, unsigned flags)
1627{
1628 int ret;
1629 obj_read_lock();
1630 ret = do_oid_object_info_extended(r, oid, oi, flags);
1631 obj_read_unlock();
1632 return ret;
1633}
1634
1635
3fc0dca9 1636/* returns enum object_type or negative */
9d98354f
SB
1637int oid_object_info(struct repository *r,
1638 const struct object_id *oid,
1639 unsigned long *sizep)
9a490590 1640{
5b086407 1641 enum object_type type;
27b5c1a0 1642 struct object_info oi = OBJECT_INFO_INIT;
9a490590 1643
5b086407 1644 oi.typep = &type;
9a490590 1645 oi.sizep = sizep;
9d98354f
SB
1646 if (oid_object_info_extended(r, oid, &oi,
1647 OBJECT_INFO_LOOKUP_REPLACE) < 0)
5b086407
JK
1648 return -1;
1649 return type;
9a490590
JH
1650}
1651
1b9b5c69 1652static void *read_object(struct repository *r,
cba595ab 1653 const struct object_id *oid, enum object_type *type,
f1d8130b
JT
1654 unsigned long *size)
1655{
1656 struct object_info oi = OBJECT_INFO_INIT;
1657 void *content;
1658 oi.typep = type;
1659 oi.sizep = size;
1660 oi.contentp = &content;
1661
cba595ab 1662 if (oid_object_info_extended(r, oid, &oi, 0) < 0)
f1d8130b
JT
1663 return NULL;
1664 return content;
1665}
1666
829e5c3b
PO
1667int pretend_object_file(void *buf, unsigned long len, enum object_type type,
1668 struct object_id *oid)
d66b37bb
JH
1669{
1670 struct cached_object *co;
1671
2dcde20e 1672 hash_object_file(the_hash_algo, buf, len, type_name(type), oid);
a64d2aae
JT
1673 if (has_object_file_with_flags(oid, OBJECT_INFO_QUICK | OBJECT_INFO_SKIP_FETCH_OBJECT) ||
1674 find_cached_object(oid))
d66b37bb 1675 return 0;
c7353967 1676 ALLOC_GROW(cached_objects, cached_object_nr + 1, cached_object_alloc);
d66b37bb
JH
1677 co = &cached_objects[cached_object_nr++];
1678 co->size = len;
21666f1a 1679 co->type = type;
efa13f7b
JH
1680 co->buf = xmalloc(len);
1681 memcpy(co->buf, buf, len);
62ba93ea 1682 oidcpy(&co->oid, oid);
d66b37bb
JH
1683 return 0;
1684}
1685
b6c4cecc
JH
1686/*
1687 * This function dies on corrupt objects; the callers who want to
1688 * deal with them should arrange to call read_object() and give error
1689 * messages themselves.
1690 */
a3b72c89
SB
1691void *read_object_file_extended(struct repository *r,
1692 const struct object_id *oid,
b4f5aca4 1693 enum object_type *type,
1694 unsigned long *size,
1695 int lookup_replace)
ac939109 1696{
3ba7a065 1697 void *data;
b6c4cecc 1698 const struct packed_git *p;
771e7d57
JK
1699 const char *path;
1700 struct stat st;
1f2e7cea 1701 const struct object_id *repl = lookup_replace ?
a3b72c89 1702 lookup_replace_object(r, oid) : oid;
b6c4cecc 1703
3ba7a065 1704 errno = 0;
cba595ab 1705 data = read_object(r, repl, type, size);
4bbf5a26 1706 if (data)
b6c4cecc 1707 return data;
68095570 1708
31877c9a 1709 obj_read_lock();
25f3af3f 1710 if (errno && errno != ENOENT)
259328b7 1711 die_errno(_("failed to read object %s"), oid_to_hex(oid));
3ba7a065 1712
68095570 1713 /* die if we replaced an object with one that does not exist */
b383a13c 1714 if (repl != oid)
259328b7 1715 die(_("replacement %s not found for %s"),
b383a13c 1716 oid_to_hex(repl), oid_to_hex(oid));
68095570 1717
cba595ab 1718 if (!stat_loose_object(r, repl, &st, &path))
259328b7 1719 die(_("loose object %s (stored in %s) is corrupt"),
b383a13c 1720 oid_to_hex(repl), path);
68095570 1721
a3b72c89 1722 if ((p = has_packed_and_bad(r, repl->hash)) != NULL)
259328b7 1723 die(_("packed object %s (stored in %s) is corrupt"),
b383a13c 1724 oid_to_hex(repl), p->pack_name);
31877c9a 1725 obj_read_unlock();
f5552aee 1726
b6c4cecc 1727 return NULL;
ac939109
NP
1728}
1729
d3b4705a
NTND
1730void *read_object_with_reference(struct repository *r,
1731 const struct object_id *oid,
21666f1a 1732 const char *required_type_name,
40469ee9 1733 unsigned long *size,
02f0547e 1734 struct object_id *actual_oid_return)
f4913f91 1735{
21666f1a 1736 enum object_type type, required_type;
f4913f91
JH
1737 void *buffer;
1738 unsigned long isize;
02f0547e 1739 struct object_id actual_oid;
f4913f91 1740
21666f1a 1741 required_type = type_from_string(required_type_name);
02f0547e 1742 oidcpy(&actual_oid, oid);
40469ee9
JH
1743 while (1) {
1744 int ref_length = -1;
1745 const char *ref_type = NULL;
f4913f91 1746
d3b4705a 1747 buffer = repo_read_object_file(r, &actual_oid, &type, &isize);
40469ee9
JH
1748 if (!buffer)
1749 return NULL;
21666f1a 1750 if (type == required_type) {
40469ee9 1751 *size = isize;
02f0547e 1752 if (actual_oid_return)
1753 oidcpy(actual_oid_return, &actual_oid);
40469ee9
JH
1754 return buffer;
1755 }
1756 /* Handle references */
21666f1a 1757 else if (type == OBJ_COMMIT)
40469ee9 1758 ref_type = "tree ";
21666f1a 1759 else if (type == OBJ_TAG)
40469ee9
JH
1760 ref_type = "object ";
1761 else {
1762 free(buffer);
1763 return NULL;
1764 }
1765 ref_length = strlen(ref_type);
f4913f91 1766
94b5e093 1767 if (ref_length + the_hash_algo->hexsz > isize ||
50974ec9 1768 memcmp(buffer, ref_type, ref_length) ||
02f0547e 1769 get_oid_hex((char *) buffer + ref_length, &actual_oid)) {
40469ee9
JH
1770 free(buffer);
1771 return NULL;
1772 }
1cf58e72 1773 free(buffer);
40469ee9 1774 /* Now we have the ID of the referred-to object in
02f0547e 1775 * actual_oid. Check again. */
f4913f91 1776 }
f4913f91
JH
1777}
1778
7ad5c44d
MT
1779static void write_object_file_prepare(const struct git_hash_algo *algo,
1780 const void *buf, unsigned long len,
a09c985e
PO
1781 const char *type, struct object_id *oid,
1782 char *hdr, int *hdrlen)
d410c0f5 1783{
18e2588e 1784 git_hash_ctx c;
d410c0f5
JH
1785
1786 /* Generate the header */
ca473cef 1787 *hdrlen = xsnprintf(hdr, *hdrlen, "%s %"PRIuMAX , type, (uintmax_t)len)+1;
d410c0f5
JH
1788
1789 /* Sha1.. */
7ad5c44d
MT
1790 algo->init_fn(&c);
1791 algo->update_fn(&c, hdr, *hdrlen);
1792 algo->update_fn(&c, buf, len);
5951bf46 1793 algo->final_oid_fn(oid, &c);
d410c0f5
JH
1794}
1795
230f1322 1796/*
5a688fe4 1797 * Move the just written object into its final resting place.
230f1322 1798 */
cb5add58 1799int finalize_object_file(const char *tmpfile, const char *filename)
230f1322 1800{
e32c0a9c 1801 int ret = 0;
5a688fe4 1802
348df166 1803 if (object_creation_mode == OBJECT_CREATION_USES_RENAMES)
be66a6c4
JS
1804 goto try_rename;
1805 else if (link(tmpfile, filename))
e32c0a9c 1806 ret = errno;
7ebb6fca
LT
1807
1808 /*
1809 * Coda hack - coda doesn't like cross-directory links,
1810 * so we fall back to a rename, which will mean that it
1811 * won't be able to check collisions, but that's not a
1812 * big deal.
1813 *
1814 * The same holds for FAT formatted media.
1815 *
3be1f18e 1816 * When this succeeds, we just return. We have nothing
7ebb6fca
LT
1817 * left to unlink.
1818 */
1819 if (ret && ret != EEXIST) {
be66a6c4 1820 try_rename:
7ebb6fca 1821 if (!rename(tmpfile, filename))
3be1f18e 1822 goto out;
9e48b389 1823 ret = errno;
230f1322 1824 }
691f1a28 1825 unlink_or_warn(tmpfile);
230f1322
LT
1826 if (ret) {
1827 if (ret != EEXIST) {
2c319886 1828 return error_errno(_("unable to write file %s"), filename);
230f1322
LT
1829 }
1830 /* FIXME!!! Collision check here ? */
1831 }
1832
3be1f18e 1833out:
5256b006 1834 if (adjust_shared_perm(filename))
259328b7 1835 return error(_("unable to set permission to '%s'"), filename);
230f1322
LT
1836 return 0;
1837}
1838
4d548150
LT
1839static int write_buffer(int fd, const void *buf, size_t len)
1840{
d34cf19b 1841 if (write_in_full(fd, buf, len) < 0)
259328b7 1842 return error_errno(_("file write error"));
4d548150
LT
1843 return 0;
1844}
1845
2dcde20e
MT
1846int hash_object_file(const struct git_hash_algo *algo, const void *buf,
1847 unsigned long len, const char *type,
f070facc 1848 struct object_id *oid)
abdc3fc8 1849{
1af64f73 1850 char hdr[MAX_HEADER_LEN];
ef1286d3 1851 int hdrlen = sizeof(hdr);
2dcde20e 1852 write_object_file_prepare(algo, buf, len, type, oid, hdr, &hdrlen);
abdc3fc8
RS
1853 return 0;
1854}
1855
e9039dd3 1856/* Finalize a file on disk, and close it. */
514c5fdd 1857static void close_loose_object(int fd)
e9039dd3 1858{
aafe9fba 1859 if (fsync_object_files)
76011357 1860 fsync_or_die(fd, "loose object file");
e9039dd3 1861 if (close(fd) != 0)
76011357 1862 die_errno(_("error when closing loose object file"));
e9039dd3
LT
1863}
1864
5723fe7e
LT
1865/* Size of directory component, including the ending '/' */
1866static inline int directory_size(const char *filename)
1867{
1868 const char *s = strrchr(filename, '/');
1869 if (!s)
1870 return 0;
1871 return s - filename + 1;
1872}
1873
1874/*
1875 * This creates a temporary file in the same directory as the final
1876 * 'filename'
1877 *
1878 * We want to avoid cross-directory filename renames, because those
1879 * can have problems on various filesystems (FAT, NFS, Coda).
1880 */
d4b3d11a 1881static int create_tmpfile(struct strbuf *tmp, const char *filename)
5723fe7e
LT
1882{
1883 int fd, dirlen = directory_size(filename);
1884
d4b3d11a
JK
1885 strbuf_reset(tmp);
1886 strbuf_add(tmp, filename, dirlen);
1887 strbuf_addstr(tmp, "tmp_obj_XXXXXX");
1888 fd = git_mkstemp_mode(tmp->buf, 0444);
cbacbf4e 1889 if (fd < 0 && dirlen && errno == ENOENT) {
d4b3d11a
JK
1890 /*
1891 * Make sure the directory exists; note that the contents
1892 * of the buffer are undefined after mkstemp returns an
1893 * error, so we have to rewrite the whole buffer from
1894 * scratch.
1895 */
1896 strbuf_reset(tmp);
1897 strbuf_add(tmp, filename, dirlen - 1);
1898 if (mkdir(tmp->buf, 0777) && errno != EEXIST)
b2476a60 1899 return -1;
d4b3d11a 1900 if (adjust_shared_perm(tmp->buf))
5723fe7e
LT
1901 return -1;
1902
1903 /* Try again */
d4b3d11a
JK
1904 strbuf_addstr(tmp, "/tmp_obj_XXXXXX");
1905 fd = git_mkstemp_mode(tmp->buf, 0444);
5723fe7e
LT
1906 }
1907 return fd;
1908}
1909
3fc7281f
PO
1910static int write_loose_object(const struct object_id *oid, char *hdr,
1911 int hdrlen, const void *buf, unsigned long len,
1912 time_t mtime)
0fcfd160 1913{
915308b1 1914 int fd, ret;
9892beba 1915 unsigned char compressed[4096];
ef49a7a0 1916 git_zstream stream;
18e2588e 1917 git_hash_ctx c;
3fc7281f 1918 struct object_id parano_oid;
d4b3d11a 1919 static struct strbuf tmp_file = STRBUF_INIT;
ea657730
CC
1920 static struct strbuf filename = STRBUF_INIT;
1921
514c5fdd 1922 loose_object_path(the_repository, &filename, oid);
a44c9a5e 1923
ea657730 1924 fd = create_tmpfile(&tmp_file, filename.buf);
aac17941 1925 if (fd < 0) {
35243577 1926 if (errno == EACCES)
259328b7 1927 return error(_("insufficient permission for adding an object to repository database %s"), get_object_directory());
916d081b 1928 else
259328b7 1929 return error_errno(_("unable to create temporary file"));
aac17941
LT
1930 }
1931
0fcfd160 1932 /* Set it up */
55bb5c91 1933 git_deflate_init(&stream, zlib_compression_level);
0fcfd160 1934 stream.next_out = compressed;
9892beba 1935 stream.avail_out = sizeof(compressed);
18e2588e 1936 the_hash_algo->init_fn(&c);
a44c9a5e
LT
1937
1938 /* First header.. */
d65a16f6 1939 stream.next_in = (unsigned char *)hdr;
a44c9a5e 1940 stream.avail_in = hdrlen;
55bb5c91
JH
1941 while (git_deflate(&stream, 0) == Z_OK)
1942 ; /* nothing */
18e2588e 1943 the_hash_algo->update_fn(&c, hdr, hdrlen);
a44c9a5e
LT
1944
1945 /* Then the data itself.. */
c00e657d 1946 stream.next_in = (void *)buf;
a44c9a5e 1947 stream.avail_in = len;
9892beba 1948 do {
748af44c 1949 unsigned char *in0 = stream.next_in;
55bb5c91 1950 ret = git_deflate(&stream, Z_FINISH);
18e2588e 1951 the_hash_algo->update_fn(&c, in0, stream.next_in - in0);
9892beba 1952 if (write_buffer(fd, compressed, stream.next_out - compressed) < 0)
76011357 1953 die(_("unable to write loose object file"));
9892beba
NP
1954 stream.next_out = compressed;
1955 stream.avail_out = sizeof(compressed);
1956 } while (ret == Z_OK);
1957
ac54c277 1958 if (ret != Z_STREAM_END)
259328b7 1959 die(_("unable to deflate new object %s (%d)"), oid_to_hex(oid),
3fc7281f 1960 ret);
55bb5c91 1961 ret = git_deflate_end_gently(&stream);
ac54c277 1962 if (ret != Z_OK)
259328b7 1963 die(_("deflateEnd on object %s failed (%d)"), oid_to_hex(oid),
3fc7281f 1964 ret);
5951bf46 1965 the_hash_algo->final_oid_fn(&parano_oid, &c);
9001dc2a 1966 if (!oideq(oid, &parano_oid))
259328b7 1967 die(_("confused by unstable object source data for %s"),
3fc7281f 1968 oid_to_hex(oid));
ac54c277 1969
514c5fdd 1970 close_loose_object(fd);
0fcfd160 1971
bbac7311
NP
1972 if (mtime) {
1973 struct utimbuf utb;
1974 utb.actime = mtime;
1975 utb.modtime = mtime;
d4b3d11a 1976 if (utime(tmp_file.buf, &utb) < 0)
259328b7 1977 warning_errno(_("failed utime() on %s"), tmp_file.buf);
bbac7311
NP
1978 }
1979
ea657730 1980 return finalize_object_file(tmp_file.buf, filename.buf);
0fcfd160 1981}
8237b185 1982
6862ebbf 1983static int freshen_loose_object(const struct object_id *oid)
33d4221c 1984{
6862ebbf 1985 return check_and_freshen(oid, 1);
33d4221c
JK
1986}
1987
6862ebbf 1988static int freshen_packed_object(const struct object_id *oid)
33d4221c
JK
1989{
1990 struct pack_entry e;
544443cb 1991 if (!find_pack_entry(the_repository, oid, &e))
ee1c6c34
JK
1992 return 0;
1993 if (e.p->freshened)
1994 return 1;
1995 if (!freshen_file(e.p->pack_name))
1996 return 0;
1997 e.p->freshened = 1;
1998 return 1;
33d4221c
JK
1999}
2000
a09c985e
PO
2001int write_object_file(const void *buf, unsigned long len, const char *type,
2002 struct object_id *oid)
bbac7311 2003{
1af64f73 2004 char hdr[MAX_HEADER_LEN];
ef1286d3 2005 int hdrlen = sizeof(hdr);
bbac7311
NP
2006
2007 /* Normally if we have it in the pack then we do not bother writing
2008 * it out into .git/objects/??/?{38} file.
2009 */
7ad5c44d
MT
2010 write_object_file_prepare(the_hash_algo, buf, len, type, oid, hdr,
2011 &hdrlen);
6862ebbf 2012 if (freshen_packed_object(oid) || freshen_loose_object(oid))
bbac7311 2013 return 0;
3fc7281f 2014 return write_loose_object(oid, hdr, hdrlen, buf, len, 0);
bbac7311
NP
2015}
2016
1752cbbc
PO
2017int hash_object_file_literally(const void *buf, unsigned long len,
2018 const char *type, struct object_id *oid,
2019 unsigned flags)
0c3db67c
ES
2020{
2021 char *header;
2022 int hdrlen, status = 0;
2023
2024 /* type string, SP, %lu of the length plus NUL must fit this */
1af64f73 2025 hdrlen = strlen(type) + MAX_HEADER_LEN;
ef1286d3 2026 header = xmalloc(hdrlen);
7ad5c44d
MT
2027 write_object_file_prepare(the_hash_algo, buf, len, type, oid, header,
2028 &hdrlen);
0c3db67c
ES
2029
2030 if (!(flags & HASH_WRITE_OBJECT))
2031 goto cleanup;
6862ebbf 2032 if (freshen_packed_object(oid) || freshen_loose_object(oid))
0c3db67c 2033 goto cleanup;
3fc7281f 2034 status = write_loose_object(oid, header, hdrlen, buf, len, 0);
0c3db67c
ES
2035
2036cleanup:
2037 free(header);
2038 return status;
2039}
2040
4bdb70a4 2041int force_object_loose(const struct object_id *oid, time_t mtime)
bbac7311 2042{
bbac7311
NP
2043 void *buf;
2044 unsigned long len;
2045 enum object_type type;
1af64f73 2046 char hdr[MAX_HEADER_LEN];
bbac7311 2047 int hdrlen;
1fb23e65 2048 int ret;
bbac7311 2049
6862ebbf 2050 if (has_loose_object(oid))
bbac7311 2051 return 0;
cba595ab 2052 buf = read_object(the_repository, oid, &type, &len);
bbac7311 2053 if (!buf)
2c319886 2054 return error(_("cannot read object for %s"), oid_to_hex(oid));
ca473cef 2055 hdrlen = xsnprintf(hdr, sizeof(hdr), "%s %"PRIuMAX , type_name(type), (uintmax_t)len) + 1;
3fc7281f 2056 ret = write_loose_object(oid, hdr, hdrlen, buf, len, mtime);
1fb23e65
BS
2057 free(buf);
2058
2059 return ret;
bbac7311
NP
2060}
2061
1d8d9cb6
JT
2062int has_object(struct repository *r, const struct object_id *oid,
2063 unsigned flags)
2064{
2065 int quick = !(flags & HAS_OBJECT_RECHECK_PACKED);
2066 unsigned object_info_flags = OBJECT_INFO_SKIP_FETCH_OBJECT |
2067 (quick ? OBJECT_INFO_QUICK : 0);
2068
2069 if (!startup_info->have_repository)
2070 return 0;
2071 return oid_object_info_extended(r, oid, NULL, object_info_flags) >= 0;
2072}
2073
cba595ab
JH
2074int repo_has_object_file_with_flags(struct repository *r,
2075 const struct object_id *oid, int flags)
8237b185 2076{
3e8b7d3c
JN
2077 if (!startup_info->have_repository)
2078 return 0;
9c8a294a 2079 return oid_object_info_extended(r, oid, NULL, flags) >= 0;
b419aa25 2080}
2081
9b45f499
SB
2082int repo_has_object_file(struct repository *r,
2083 const struct object_id *oid)
b419aa25 2084{
cba595ab 2085 return repo_has_object_file_with_flags(r, oid, 0);
5827a035
JK
2086}
2087
c879daa2
NTND
2088static void check_tree(const void *buf, size_t size)
2089{
2090 struct tree_desc desc;
2091 struct name_entry entry;
2092
2093 init_tree_desc(&desc, buf, size);
2094 while (tree_entry(&desc, &entry))
2095 /* do nothing
2096 * tree_entry() will die() on malformed entries */
2097 ;
2098}
2099
2100static void check_commit(const void *buf, size_t size)
2101{
2102 struct commit c;
2103 memset(&c, 0, sizeof(c));
08f4f445 2104 if (parse_commit_buffer(the_repository, &c, buf, size, 0))
259328b7 2105 die(_("corrupt commit"));
c879daa2
NTND
2106}
2107
2108static void check_tag(const void *buf, size_t size)
2109{
2110 struct tag t;
2111 memset(&t, 0, sizeof(t));
0e740fed 2112 if (parse_tag_buffer(the_repository, &t, buf, size))
259328b7 2113 die(_("corrupt tag"));
c879daa2
NTND
2114}
2115
58bf2a4c
NTND
2116static int index_mem(struct index_state *istate,
2117 struct object_id *oid, void *buf, size_t size,
c4ce46fc
JH
2118 enum object_type type,
2119 const char *path, unsigned flags)
e7332f96 2120{
6c510bee 2121 int ret, re_allocated = 0;
c4ce46fc 2122 int write_object = flags & HASH_WRITE_OBJECT;
74400e71 2123
7672db20 2124 if (!type)
edaec3fb 2125 type = OBJ_BLOB;
6c510bee
LT
2126
2127 /*
2128 * Convert blobs to git internal format
2129 */
43df4f86 2130 if ((type == OBJ_BLOB) && path) {
f285a2d7 2131 struct strbuf nbuf = STRBUF_INIT;
58bf2a4c 2132 if (convert_to_git(istate, path, buf, size, &nbuf,
8462ff43 2133 get_conv_flags(flags))) {
b315c5c0 2134 buf = strbuf_detach(&nbuf, &size);
6c510bee
LT
2135 re_allocated = 1;
2136 }
2137 }
c4ce46fc 2138 if (flags & HASH_FORMAT_CHECK) {
c879daa2
NTND
2139 if (type == OBJ_TREE)
2140 check_tree(buf, size);
2141 if (type == OBJ_COMMIT)
2142 check_commit(buf, size);
2143 if (type == OBJ_TAG)
2144 check_tag(buf, size);
2145 }
6c510bee 2146
7672db20 2147 if (write_object)
169c9c01 2148 ret = write_object_file(buf, size, type_name(type), oid);
abdc3fc8 2149 else
2dcde20e
MT
2150 ret = hash_object_file(the_hash_algo, buf, size,
2151 type_name(type), oid);
43df4f86 2152 if (re_allocated)
6c510bee 2153 free(buf);
43df4f86
DP
2154 return ret;
2155}
2156
58bf2a4c
NTND
2157static int index_stream_convert_blob(struct index_state *istate,
2158 struct object_id *oid,
2159 int fd,
2160 const char *path,
2161 unsigned flags)
9035d75a
SP
2162{
2163 int ret;
2164 const int write_object = flags & HASH_WRITE_OBJECT;
2165 struct strbuf sbuf = STRBUF_INIT;
2166
2167 assert(path);
58bf2a4c 2168 assert(would_convert_to_git_filter_fd(istate, path));
9035d75a 2169
58bf2a4c 2170 convert_to_git_filter_fd(istate, path, fd, &sbuf,
8462ff43 2171 get_conv_flags(flags));
9035d75a
SP
2172
2173 if (write_object)
169c9c01 2174 ret = write_object_file(sbuf.buf, sbuf.len, type_name(OBJ_BLOB),
a09c985e 2175 oid);
9035d75a 2176 else
2dcde20e
MT
2177 ret = hash_object_file(the_hash_algo, sbuf.buf, sbuf.len,
2178 type_name(OBJ_BLOB), oid);
9035d75a
SP
2179 strbuf_release(&sbuf);
2180 return ret;
2181}
2182
58bf2a4c
NTND
2183static int index_pipe(struct index_state *istate, struct object_id *oid,
2184 int fd, enum object_type type,
7b41e1e1
JH
2185 const char *path, unsigned flags)
2186{
2187 struct strbuf sbuf = STRBUF_INIT;
2188 int ret;
2189
2190 if (strbuf_read(&sbuf, fd, 4096) >= 0)
58bf2a4c 2191 ret = index_mem(istate, oid, sbuf.buf, sbuf.len, type, path, flags);
7b41e1e1
JH
2192 else
2193 ret = -1;
2194 strbuf_release(&sbuf);
2195 return ret;
2196}
2197
ea68b0ce
DP
2198#define SMALL_FILE_SIZE (32*1024)
2199
58bf2a4c
NTND
2200static int index_core(struct index_state *istate,
2201 struct object_id *oid, int fd, size_t size,
7b41e1e1
JH
2202 enum object_type type, const char *path,
2203 unsigned flags)
43df4f86
DP
2204{
2205 int ret;
43df4f86 2206
7b41e1e1 2207 if (!size) {
58bf2a4c 2208 ret = index_mem(istate, oid, "", size, type, path, flags);
ea68b0ce
DP
2209 } else if (size <= SMALL_FILE_SIZE) {
2210 char *buf = xmalloc(size);
90dca671
JK
2211 ssize_t read_result = read_in_full(fd, buf, size);
2212 if (read_result < 0)
259328b7 2213 ret = error_errno(_("read error while indexing %s"),
90dca671
JK
2214 path ? path : "<unknown>");
2215 else if (read_result != size)
259328b7 2216 ret = error(_("short read while indexing %s"),
90dca671 2217 path ? path : "<unknown>");
ea68b0ce 2218 else
58bf2a4c 2219 ret = index_mem(istate, oid, buf, size, type, path, flags);
ea68b0ce 2220 free(buf);
08bda208 2221 } else {
43df4f86 2222 void *buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
58bf2a4c 2223 ret = index_mem(istate, oid, buf, size, type, path, flags);
aac17941 2224 munmap(buf, size);
08bda208 2225 }
7b41e1e1
JH
2226 return ret;
2227}
2228
4dd1fbc7 2229/*
568508e7
JH
2230 * This creates one packfile per large blob unless bulk-checkin
2231 * machinery is "plugged".
4dd1fbc7
JH
2232 *
2233 * This also bypasses the usual "convert-to-git" dance, and that is on
2234 * purpose. We could write a streaming version of the converting
2235 * functions and insert that before feeding the data to fast-import
4f22b101
JK
2236 * (or equivalent in-core API described above). However, that is
2237 * somewhat complicated, as we do not know the size of the filter
2238 * result, which we need to know beforehand when writing a git object.
2239 * Since the primary motivation for trying to stream from the working
2240 * tree file and to avoid mmaping it in core is to deal with large
2241 * binary blobs, they generally do not want to get any conversion, and
2242 * callers should avoid this code path when filters are requested.
4dd1fbc7 2243 */
7d5e1dc3 2244static int index_stream(struct object_id *oid, int fd, size_t size,
4dd1fbc7
JH
2245 enum object_type type, const char *path,
2246 unsigned flags)
2247{
68ee6dfc 2248 return index_bulk_checkin(oid, fd, size, type, path, flags);
4dd1fbc7
JH
2249}
2250
58bf2a4c
NTND
2251int index_fd(struct index_state *istate, struct object_id *oid,
2252 int fd, struct stat *st,
7b41e1e1
JH
2253 enum object_type type, const char *path, unsigned flags)
2254{
2255 int ret;
7b41e1e1 2256
9079ab7c
SP
2257 /*
2258 * Call xsize_t() only when needed to avoid potentially unnecessary
2259 * die() for large files.
2260 */
58bf2a4c
NTND
2261 if (type == OBJ_BLOB && path && would_convert_to_git_filter_fd(istate, path))
2262 ret = index_stream_convert_blob(istate, oid, fd, path, flags);
9035d75a 2263 else if (!S_ISREG(st->st_mode))
58bf2a4c 2264 ret = index_pipe(istate, oid, fd, type, path, flags);
9079ab7c 2265 else if (st->st_size <= big_file_threshold || type != OBJ_BLOB ||
58bf2a4c
NTND
2266 (path && would_convert_to_git(istate, path)))
2267 ret = index_core(istate, oid, fd, xsize_t(st->st_size),
2268 type, path, flags);
4dd1fbc7 2269 else
7d5e1dc3 2270 ret = index_stream(oid, fd, xsize_t(st->st_size), type, path,
9079ab7c 2271 flags);
43df4f86 2272 close(fd);
aac17941 2273 return ret;
74400e71 2274}
ec1fcc16 2275
58bf2a4c
NTND
2276int index_path(struct index_state *istate, struct object_id *oid,
2277 const char *path, struct stat *st, unsigned flags)
ec1fcc16
JH
2278{
2279 int fd;
b760d3aa 2280 struct strbuf sb = STRBUF_INIT;
ea8e0297 2281 int rc = 0;
ec1fcc16
JH
2282
2283 switch (st->st_mode & S_IFMT) {
2284 case S_IFREG:
2285 fd = open(path, O_RDONLY);
2286 if (fd < 0)
7616c6ca 2287 return error_errno("open(\"%s\")", path);
58bf2a4c 2288 if (index_fd(istate, oid, fd, st, OBJ_BLOB, path, flags) < 0)
259328b7 2289 return error(_("%s: failed to insert into database"),
ec1fcc16
JH
2290 path);
2291 break;
2292 case S_IFLNK:
7616c6ca
NTND
2293 if (strbuf_readlink(&sb, path, st->st_size))
2294 return error_errno("readlink(\"%s\")", path);
c4ce46fc 2295 if (!(flags & HASH_WRITE_OBJECT))
2dcde20e
MT
2296 hash_object_file(the_hash_algo, sb.buf, sb.len,
2297 blob_type, oid);
a09c985e 2298 else if (write_object_file(sb.buf, sb.len, blob_type, oid))
259328b7 2299 rc = error(_("%s: failed to insert into database"), path);
b760d3aa 2300 strbuf_release(&sb);
ec1fcc16 2301 break;
f35a6d3b 2302 case S_IFDIR:
a98e6101 2303 return resolve_gitlink_ref(path, "HEAD", oid);
ec1fcc16 2304 default:
259328b7 2305 return error(_("%s: unsupported file type"), path);
ec1fcc16 2306 }
ea8e0297 2307 return rc;
ec1fcc16 2308}
a69e5429
JH
2309
2310int read_pack_header(int fd, struct pack_header *header)
2311{
f48ecd38 2312 if (read_in_full(fd, header, sizeof(*header)) != sizeof(*header))
c697ad14
HO
2313 /* "eof before pack header was fully read" */
2314 return PH_ERROR_EOF;
2315
a69e5429
JH
2316 if (header->hdr_signature != htonl(PACK_SIGNATURE))
2317 /* "protocol error (pack signature mismatch detected)" */
2318 return PH_ERROR_PACK_SIGNATURE;
2319 if (!pack_version_ok(header->hdr_version))
2320 /* "protocol error (pack version unsupported)" */
2321 return PH_ERROR_PROTOCOL;
2322 return 0;
2323}
40d52ff7 2324
e816caa0 2325void assert_oid_type(const struct object_id *oid, enum object_type expect)
40d52ff7 2326{
0df8e965 2327 enum object_type type = oid_object_info(the_repository, oid, NULL);
40d52ff7 2328 if (type < 0)
259328b7 2329 die(_("%s is not a valid object"), oid_to_hex(oid));
40d52ff7 2330 if (type != expect)
259328b7 2331 die(_("%s is not a valid '%s' object"), oid_to_hex(oid),
debca9d2 2332 type_name(expect));
40d52ff7 2333}
27e1e22d 2334
70c49050 2335int for_each_file_in_obj_subdir(unsigned int subdir_nr,
cc817ca3
RS
2336 struct strbuf *path,
2337 each_loose_object_fn obj_cb,
2338 each_loose_cruft_fn cruft_cb,
2339 each_loose_subdir_fn subdir_cb,
2340 void *data)
27e1e22d 2341{
0375f472
RS
2342 size_t origlen, baselen;
2343 DIR *dir;
27e1e22d
JK
2344 struct dirent *de;
2345 int r = 0;
62a24c89 2346 struct object_id oid;
27e1e22d 2347
70c49050
RS
2348 if (subdir_nr > 0xff)
2349 BUG("invalid loose object subdirectory: %x", subdir_nr);
2350
0375f472
RS
2351 origlen = path->len;
2352 strbuf_complete(path, '/');
2353 strbuf_addf(path, "%02x", subdir_nr);
0375f472
RS
2354
2355 dir = opendir(path->buf);
27e1e22d 2356 if (!dir) {
0375f472 2357 if (errno != ENOENT)
259328b7 2358 r = error_errno(_("unable to open %s"), path->buf);
0375f472
RS
2359 strbuf_setlen(path, origlen);
2360 return r;
27e1e22d
JK
2361 }
2362
62a24c89 2363 oid.hash[0] = subdir_nr;
163ee5e6
DS
2364 strbuf_addch(path, '/');
2365 baselen = path->len;
62a24c89 2366
b548f0f1 2367 while ((de = readdir_skip_dot_and_dotdot(dir))) {
163ee5e6 2368 size_t namelen;
27e1e22d 2369
163ee5e6 2370 namelen = strlen(de->d_name);
27e1e22d 2371 strbuf_setlen(path, baselen);
163ee5e6 2372 strbuf_add(path, de->d_name, namelen);
94b5e093 2373 if (namelen == the_hash_algo->hexsz - 2 &&
62a24c89 2374 !hex_to_bytes(oid.hash + 1, de->d_name,
94b5e093 2375 the_hash_algo->rawsz - 1)) {
5a6dce70 2376 oid_set_algo(&oid, the_hash_algo);
62a24c89
RS
2377 if (obj_cb) {
2378 r = obj_cb(&oid, path->buf, data);
2379 if (r)
2380 break;
27e1e22d 2381 }
62a24c89 2382 continue;
27e1e22d
JK
2383 }
2384
2385 if (cruft_cb) {
2386 r = cruft_cb(de->d_name, path->buf, data);
2387 if (r)
2388 break;
2389 }
2390 }
094c7e63 2391 closedir(dir);
27e1e22d 2392
163ee5e6 2393 strbuf_setlen(path, baselen - 1);
27e1e22d
JK
2394 if (!r && subdir_cb)
2395 r = subdir_cb(subdir_nr, path->buf, data);
2396
0375f472
RS
2397 strbuf_setlen(path, origlen);
2398
27e1e22d
JK
2399 return r;
2400}
2401
e6f875e0 2402int for_each_loose_file_in_objdir_buf(struct strbuf *path,
27e1e22d
JK
2403 each_loose_object_fn obj_cb,
2404 each_loose_cruft_fn cruft_cb,
2405 each_loose_subdir_fn subdir_cb,
2406 void *data)
2407{
27e1e22d
JK
2408 int r = 0;
2409 int i;
2410
27e1e22d 2411 for (i = 0; i < 256; i++) {
e6f875e0 2412 r = for_each_file_in_obj_subdir(i, path, obj_cb, cruft_cb,
27e1e22d 2413 subdir_cb, data);
27e1e22d
JK
2414 if (r)
2415 break;
2416 }
2417
e6f875e0
JK
2418 return r;
2419}
2420
2421int for_each_loose_file_in_objdir(const char *path,
2422 each_loose_object_fn obj_cb,
2423 each_loose_cruft_fn cruft_cb,
2424 each_loose_subdir_fn subdir_cb,
2425 void *data)
2426{
2427 struct strbuf buf = STRBUF_INIT;
2428 int r;
2429
2430 strbuf_addstr(&buf, path);
2431 r = for_each_loose_file_in_objdir_buf(&buf, obj_cb, cruft_cb,
2432 subdir_cb, data);
27e1e22d 2433 strbuf_release(&buf);
e6f875e0 2434
27e1e22d
JK
2435 return r;
2436}
660c889e 2437
a7ff6f5a
JK
2438int for_each_loose_object(each_loose_object_fn cb, void *data,
2439 enum for_each_object_flags flags)
660c889e 2440{
f0eaf638 2441 struct object_directory *odb;
b0a42642 2442
f0eaf638
JK
2443 prepare_alt_odb(the_repository);
2444 for (odb = the_repository->objects->odb; odb; odb = odb->next) {
2445 int r = for_each_loose_file_in_objdir(odb->path, cb, NULL,
2446 NULL, data);
2447 if (r)
2448 return r;
660c889e 2449
f0eaf638
JK
2450 if (flags & FOR_EACH_OBJECT_LOCAL_ONLY)
2451 break;
2452 }
1385bb7b 2453
f0eaf638 2454 return 0;
660c889e
JK
2455}
2456
3a2e0824
JK
2457static int append_loose_object(const struct object_id *oid, const char *path,
2458 void *data)
660c889e 2459{
3a2e0824
JK
2460 oid_array_append(data, oid);
2461 return 0;
2462}
660c889e 2463
0000d654
RS
2464struct oid_array *odb_loose_cache(struct object_directory *odb,
2465 const struct object_id *oid)
2466{
2467 int subdir_nr = oid->hash[0];
3a2e0824 2468 struct strbuf buf = STRBUF_INIT;
660c889e 2469
3a2e0824
JK
2470 if (subdir_nr < 0 ||
2471 subdir_nr >= ARRAY_SIZE(odb->loose_objects_subdir_seen))
2472 BUG("subdir_nr out of range");
1385bb7b 2473
3a2e0824 2474 if (odb->loose_objects_subdir_seen[subdir_nr])
8be88dbc 2475 return &odb->loose_objects_cache[subdir_nr];
3a2e0824
JK
2476
2477 strbuf_addstr(&buf, odb->path);
2478 for_each_file_in_obj_subdir(subdir_nr, &buf,
2479 append_loose_object,
2480 NULL, NULL,
4cea1ce0 2481 &odb->loose_objects_cache[subdir_nr]);
3a2e0824 2482 odb->loose_objects_subdir_seen[subdir_nr] = 1;
7317aa71 2483 strbuf_release(&buf);
8be88dbc 2484 return &odb->loose_objects_cache[subdir_nr];
660c889e
JK
2485}
2486
d4e19e51
RS
2487void odb_clear_loose_cache(struct object_directory *odb)
2488{
4cea1ce0
RS
2489 int i;
2490
2491 for (i = 0; i < ARRAY_SIZE(odb->loose_objects_cache); i++)
2492 oid_array_clear(&odb->loose_objects_cache[i]);
d4e19e51
RS
2493 memset(&odb->loose_objects_subdir_seen, 0,
2494 sizeof(odb->loose_objects_subdir_seen));
2495}
2496
00a7760e
JK
2497static int check_stream_oid(git_zstream *stream,
2498 const char *hdr,
2499 unsigned long size,
2500 const char *path,
2501 const struct object_id *expected_oid)
f6371f92 2502{
18e2588e 2503 git_hash_ctx c;
00a7760e 2504 struct object_id real_oid;
f6371f92
JK
2505 unsigned char buf[4096];
2506 unsigned long total_read;
2507 int status = Z_OK;
2508
18e2588e 2509 the_hash_algo->init_fn(&c);
2510 the_hash_algo->update_fn(&c, hdr, stream->total_out);
f6371f92
JK
2511
2512 /*
2513 * We already read some bytes into hdr, but the ones up to the NUL
2514 * do not count against the object's content size.
2515 */
2516 total_read = stream->total_out - strlen(hdr) - 1;
2517
2518 /*
2519 * This size comparison must be "<=" to read the final zlib packets;
00a7760e 2520 * see the comment in unpack_loose_rest for details.
f6371f92
JK
2521 */
2522 while (total_read <= size &&
18ad13e5
JH
2523 (status == Z_OK ||
2524 (status == Z_BUF_ERROR && !stream->avail_out))) {
f6371f92
JK
2525 stream->next_out = buf;
2526 stream->avail_out = sizeof(buf);
2527 if (size - total_read < stream->avail_out)
2528 stream->avail_out = size - total_read;
2529 status = git_inflate(stream, Z_FINISH);
18e2588e 2530 the_hash_algo->update_fn(&c, buf, stream->next_out - buf);
f6371f92
JK
2531 total_read += stream->next_out - buf;
2532 }
2533 git_inflate_end(stream);
2534
2535 if (status != Z_STREAM_END) {
00a7760e 2536 error(_("corrupt loose object '%s'"), oid_to_hex(expected_oid));
f6371f92
JK
2537 return -1;
2538 }
cce044df 2539 if (stream->avail_in) {
259328b7 2540 error(_("garbage at end of loose object '%s'"),
00a7760e 2541 oid_to_hex(expected_oid));
cce044df
JK
2542 return -1;
2543 }
f6371f92 2544
5951bf46 2545 the_hash_algo->final_oid_fn(&real_oid, &c);
00a7760e 2546 if (!oideq(expected_oid, &real_oid)) {
01f8d594 2547 error(_("hash mismatch for %s (expected %s)"), path,
00a7760e 2548 oid_to_hex(expected_oid));
f6371f92
JK
2549 return -1;
2550 }
2551
2552 return 0;
2553}
2554
2555int read_loose_object(const char *path,
d61d87bd 2556 const struct object_id *expected_oid,
f6371f92
JK
2557 enum object_type *type,
2558 unsigned long *size,
2559 void **contents)
2560{
2561 int ret = -1;
f6371f92
JK
2562 void *map = NULL;
2563 unsigned long mapsize;
2564 git_zstream stream;
1af64f73 2565 char hdr[MAX_HEADER_LEN];
f6371f92
JK
2566
2567 *contents = NULL;
2568
514c5fdd 2569 map = map_loose_object_1(the_repository, path, NULL, &mapsize);
f6371f92 2570 if (!map) {
259328b7 2571 error_errno(_("unable to mmap %s"), path);
f6371f92
JK
2572 goto out;
2573 }
2574
00a7760e 2575 if (unpack_loose_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0) {
259328b7 2576 error(_("unable to unpack header of %s"), path);
f6371f92
JK
2577 goto out;
2578 }
2579
00a7760e 2580 *type = parse_loose_header(hdr, size);
f6371f92 2581 if (*type < 0) {
259328b7 2582 error(_("unable to parse header of %s"), path);
f6371f92
JK
2583 git_inflate_end(&stream);
2584 goto out;
2585 }
2586
7ac4f3a0 2587 if (*type == OBJ_BLOB && *size > big_file_threshold) {
00a7760e 2588 if (check_stream_oid(&stream, hdr, *size, path, expected_oid) < 0)
f6371f92
JK
2589 goto out;
2590 } else {
00a7760e 2591 *contents = unpack_loose_rest(&stream, hdr, *size, expected_oid);
f6371f92 2592 if (!*contents) {
259328b7 2593 error(_("unable to unpack contents of %s"), path);
f6371f92
JK
2594 git_inflate_end(&stream);
2595 goto out;
2596 }
b98d1885
MT
2597 if (check_object_signature(the_repository, expected_oid,
2598 *contents, *size,
2599 type_name(*type))) {
01f8d594 2600 error(_("hash mismatch for %s (expected %s)"), path,
d61d87bd 2601 oid_to_hex(expected_oid));
f6371f92
JK
2602 free(*contents);
2603 goto out;
2604 }
2605 }
2606
2607 ret = 0; /* everything checks out */
2608
2609out:
2610 if (map)
2611 munmap(map, mapsize);
f6371f92
JK
2612 return ret;
2613}