read-cache.c 27.3 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5
/*
 * GIT - The information manager from hell
 *
 * Copyright (C) Linus Torvalds, 2005
 */
6
#define NO_THE_INDEX_COMPATIBILITY_MACROS
7
#include "cache.h"
8
#include "cache-tree.h"
9
#include "refs.h"
10 11 12 13 14 15 16 17 18 19 20 21

/* Index extensions.
 *
 * The first letter should be 'A'..'Z' for extensions that are not
 * necessary for a correct operation (i.e. optimization data).
 * When new extensions are added that _needs_ to be understood in
 * order to correctly interpret the index file, pick character that
 * is outside the range, to cause the reader to abort.
 */

#define CACHE_EXT(s) ( (s[0]<<24)|(s[1]<<16)|(s[2]<<8)|(s[3]) )
#define CACHE_EXT_TREE 0x54524545	/* "TREE" */
22

23
struct index_state the_index;
24

25 26 27 28 29 30 31 32 33
/*
 * This only updates the "non-critical" parts of the directory
 * cache, ie the parts that aren't tracked by GIT, and only used
 * to validate the cache.
 */
void fill_stat_cache_info(struct cache_entry *ce, struct stat *st)
{
	ce->ce_ctime.sec = htonl(st->st_ctime);
	ce->ce_mtime.sec = htonl(st->st_mtime);
34
#ifdef USE_NSEC
35 36 37 38 39 40 41 42
	ce->ce_ctime.nsec = htonl(st->st_ctim.tv_nsec);
	ce->ce_mtime.nsec = htonl(st->st_mtim.tv_nsec);
#endif
	ce->ce_dev = htonl(st->st_dev);
	ce->ce_ino = htonl(st->st_ino);
	ce->ce_uid = htonl(st->st_uid);
	ce->ce_gid = htonl(st->st_gid);
	ce->ce_size = htonl(st->st_size);
Junio C Hamano's avatar
Junio C Hamano committed
43 44 45

	if (assume_unchanged)
		ce->ce_flags |= htons(CE_VALID);
46 47
}

Junio C Hamano's avatar
Junio C Hamano committed
48 49 50 51 52 53 54
static int ce_compare_data(struct cache_entry *ce, struct stat *st)
{
	int match = -1;
	int fd = open(ce->name, O_RDONLY);

	if (fd >= 0) {
		unsigned char sha1[20];
55
		if (!index_fd(sha1, fd, st, 0, OBJ_BLOB, ce->name))
56
			match = hashcmp(sha1, ce->sha1);
57
		/* index_fd() closed the file descriptor already */
Junio C Hamano's avatar
Junio C Hamano committed
58 59 60 61
	}
	return match;
}

62
static int ce_compare_link(struct cache_entry *ce, size_t expected_size)
Junio C Hamano's avatar
Junio C Hamano committed
63 64 65 66 67
{
	int match = -1;
	char *target;
	void *buffer;
	unsigned long size;
68
	enum object_type type;
Junio C Hamano's avatar
Junio C Hamano committed
69 70 71 72 73 74 75 76
	int len;

	target = xmalloc(expected_size);
	len = readlink(ce->name, target, expected_size);
	if (len != expected_size) {
		free(target);
		return -1;
	}
77
	buffer = read_sha1_file(ce->sha1, &type, &size);
Junio C Hamano's avatar
Junio C Hamano committed
78 79 80 81 82 83 84 85 86 87 88
	if (!buffer) {
		free(target);
		return -1;
	}
	if (size == expected_size)
		match = memcmp(buffer, target, size);
	free(buffer);
	free(target);
	return match;
}

89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
static int ce_compare_gitlink(struct cache_entry *ce)
{
	unsigned char sha1[20];

	/*
	 * We don't actually require that the .git directory
	 * under DIRLNK directory be a valid git directory. It
	 * might even be missing (in case nobody populated that
	 * sub-project).
	 *
	 * If so, we consider it always to match.
	 */
	if (resolve_gitlink_ref(ce->name, "HEAD", sha1) < 0)
		return 0;
	return hashcmp(sha1, ce->sha1);
}

Junio C Hamano's avatar
Junio C Hamano committed
106 107 108 109 110 111 112 113
static int ce_modified_check_fs(struct cache_entry *ce, struct stat *st)
{
	switch (st->st_mode & S_IFMT) {
	case S_IFREG:
		if (ce_compare_data(ce, st))
			return DATA_CHANGED;
		break;
	case S_IFLNK:
114
		if (ce_compare_link(ce, xsize_t(st->st_size)))
Junio C Hamano's avatar
Junio C Hamano committed
115 116
			return DATA_CHANGED;
		break;
117 118 119
	case S_IFDIR:
		if (S_ISDIRLNK(ntohl(ce->ce_mode)))
			return 0;
Junio C Hamano's avatar
Junio C Hamano committed
120 121 122 123 124 125
	default:
		return TYPE_CHANGED;
	}
	return 0;
}

Junio C Hamano's avatar
Junio C Hamano committed
126
static int ce_match_stat_basic(struct cache_entry *ce, struct stat *st)
127 128 129
{
	unsigned int changed = 0;

130 131 132
	switch (ntohl(ce->ce_mode) & S_IFMT) {
	case S_IFREG:
		changed |= !S_ISREG(st->st_mode) ? TYPE_CHANGED : 0;
Junio C Hamano's avatar
Junio C Hamano committed
133 134 135 136 137
		/* We consider only the owner x bit to be relevant for
		 * "mode changes"
		 */
		if (trust_executable_bit &&
		    (0100 & (ntohl(ce->ce_mode) ^ st->st_mode)))
138
			changed |= MODE_CHANGED;
139 140
		break;
	case S_IFLNK:
141 142 143
		if (!S_ISLNK(st->st_mode) &&
		    (has_symlinks || !S_ISREG(st->st_mode)))
			changed |= TYPE_CHANGED;
144
		break;
145 146 147 148 149
	case S_IFDIRLNK:
		if (!S_ISDIR(st->st_mode))
			changed |= TYPE_CHANGED;
		else if (ce_compare_gitlink(ce))
			changed |= DATA_CHANGED;
150
		return changed;
151 152 153
	default:
		die("internal error: ce_mode is %o", ntohl(ce->ce_mode));
	}
154
	if (ce->ce_mtime.sec != htonl(st->st_mtime))
155
		changed |= MTIME_CHANGED;
156 157 158
	if (ce->ce_ctime.sec != htonl(st->st_ctime))
		changed |= CTIME_CHANGED;

159
#ifdef USE_NSEC
160 161 162 163 164
	/*
	 * nsec seems unreliable - not all filesystems support it, so
	 * as long as it is in the inode cache you get right nsec
	 * but after it gets flushed, you get zero nsec.
	 */
165
	if (ce->ce_mtime.nsec != htonl(st->st_mtim.tv_nsec))
166
		changed |= MTIME_CHANGED;
167
	if (ce->ce_ctime.nsec != htonl(st->st_ctim.tv_nsec))
168
		changed |= CTIME_CHANGED;
169 170 171 172
#endif	

	if (ce->ce_uid != htonl(st->st_uid) ||
	    ce->ce_gid != htonl(st->st_gid))
173
		changed |= OWNER_CHANGED;
174
	if (ce->ce_ino != htonl(st->st_ino))
175
		changed |= INODE_CHANGED;
176 177 178 179 180 181 182 183 184 185 186

#ifdef USE_STDEV
	/*
	 * st_dev breaks on network filesystems where different
	 * clients will have different views of what "device"
	 * the filesystem is on
	 */
	if (ce->ce_dev != htonl(st->st_dev))
		changed |= INODE_CHANGED;
#endif

187
	if (ce->ce_size != htonl(st->st_size))
188
		changed |= DATA_CHANGED;
189

Junio C Hamano's avatar
Junio C Hamano committed
190 191 192
	return changed;
}

193 194
int ie_match_stat(struct index_state *istate,
		  struct cache_entry *ce, struct stat *st, int options)
Junio C Hamano's avatar
Junio C Hamano committed
195
{
Junio C Hamano's avatar
Junio C Hamano committed
196
	unsigned int changed;
197 198
	int ignore_valid = options & 01;
	int assume_racy_is_modified = options & 02;
Junio C Hamano's avatar
Junio C Hamano committed
199 200 201 202 203 204 205 206 207

	/*
	 * If it's marked as always valid in the index, it's
	 * valid whatever the checked-out copy says.
	 */
	if (!ignore_valid && (ce->ce_flags & htons(CE_VALID)))
		return 0;

	changed = ce_match_stat_basic(ce, st);
Junio C Hamano's avatar
Junio C Hamano committed
208

Junio C Hamano's avatar
Junio C Hamano committed
209 210 211 212 213 214 215 216 217 218 219 220 221 222
	/*
	 * Within 1 second of this sequence:
	 * 	echo xyzzy >file && git-update-index --add file
	 * running this command:
	 * 	echo frotz >file
	 * would give a falsely clean cache entry.  The mtime and
	 * length match the cache, and other stat fields do not change.
	 *
	 * We could detect this at update-index time (the cache entry
	 * being registered/updated records the same time as "now")
	 * and delay the return from git-update-index, but that would
	 * effectively mean we can make at most one commit per second,
	 * which is not acceptable.  Instead, we check cache entries
	 * whose mtime are the same as the index file timestamp more
Junio C Hamano's avatar
Junio C Hamano committed
223
	 * carefully than others.
Junio C Hamano's avatar
Junio C Hamano committed
224 225
	 */
	if (!changed &&
226 227
	    istate->timestamp &&
	    istate->timestamp <= ntohl(ce->ce_mtime.sec)) {
228 229 230 231 232
		if (assume_racy_is_modified)
			changed |= DATA_CHANGED;
		else
			changed |= ce_modified_check_fs(ce, st);
	}
233

Junio C Hamano's avatar
Junio C Hamano committed
234
	return changed;
235 236
}

237 238
int ie_modified(struct index_state *istate,
		struct cache_entry *ce, struct stat *st, int really)
239
{
Junio C Hamano's avatar
Junio C Hamano committed
240
	int changed, changed_fs;
241
	changed = ie_match_stat(istate, ce, st, really);
242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257
	if (!changed)
		return 0;
	/*
	 * If the mode or type has changed, there's no point in trying
	 * to refresh the entry - it's not going to match
	 */
	if (changed & (MODE_CHANGED | TYPE_CHANGED))
		return changed;

	/* Immediately after read-tree or update-index --cacheinfo,
	 * the length field is zero.  For other cases the ce_size
	 * should match the SHA1 recorded in the index entry.
	 */
	if ((changed & DATA_CHANGED) && ce->ce_size != htonl(0))
		return changed;

Junio C Hamano's avatar
Junio C Hamano committed
258 259 260
	changed_fs = ce_modified_check_fs(ce, st);
	if (changed_fs)
		return changed | changed_fs;
261 262 263
	return 0;
}

264 265 266 267 268 269 270 271 272 273 274 275
int base_name_compare(const char *name1, int len1, int mode1,
		      const char *name2, int len2, int mode2)
{
	unsigned char c1, c2;
	int len = len1 < len2 ? len1 : len2;
	int cmp;

	cmp = memcmp(name1, name2, len);
	if (cmp)
		return cmp;
	c1 = name1[len];
	c2 = name2[len];
276
	if (!c1 && S_ISDIR(mode1))
277
		c1 = '/';
278
	if (!c2 && S_ISDIR(mode2))
279 280 281 282
		c2 = '/';
	return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0;
}

283
int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
284
{
285 286
	int len1 = flags1 & CE_NAMEMASK;
	int len2 = flags2 & CE_NAMEMASK;
287 288 289 290 291 292 293 294 295 296
	int len = len1 < len2 ? len1 : len2;
	int cmp;

	cmp = memcmp(name1, name2, len);
	if (cmp)
		return cmp;
	if (len1 < len2)
		return -1;
	if (len1 > len2)
		return 1;
Junio C Hamano's avatar
Junio C Hamano committed
297

298 299 300
	/* Compare stages  */
	flags1 &= CE_STAGEMASK;
	flags2 &= CE_STAGEMASK;
Junio C Hamano's avatar
Junio C Hamano committed
301

302 303 304 305
	if (flags1 < flags2)
		return -1;
	if (flags1 > flags2)
		return 1;
306 307 308
	return 0;
}

309
int index_name_pos(struct index_state *istate, const char *name, int namelen)
310 311 312 313
{
	int first, last;

	first = 0;
314
	last = istate->cache_nr;
315 316
	while (last > first) {
		int next = (last + first) >> 1;
317
		struct cache_entry *ce = istate->cache[next];
318
		int cmp = cache_name_compare(name, namelen, ce->name, ntohs(ce->ce_flags));
319
		if (!cmp)
320
			return next;
321 322 323 324 325 326
		if (cmp < 0) {
			last = next;
			continue;
		}
		first = next+1;
	}
327
	return -first-1;
328 329
}

330
/* Remove entry, return true if there are more entries to go.. */
331
int remove_index_entry_at(struct index_state *istate, int pos)
332
{
333 334 335
	istate->cache_changed = 1;
	istate->cache_nr--;
	if (pos >= istate->cache_nr)
336
		return 0;
337 338 339
	memmove(istate->cache + pos,
		istate->cache + pos + 1,
		(istate->cache_nr - pos) * sizeof(struct cache_entry *));
340 341 342
	return 1;
}

343
int remove_file_from_index(struct index_state *istate, const char *path)
344
{
345
	int pos = index_name_pos(istate, path, strlen(path));
346 347
	if (pos < 0)
		pos = -pos-1;
348 349
	while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path))
		remove_index_entry_at(istate, pos);
350 351 352
	return 0;
}

353
int add_file_to_index(struct index_state *istate, const char *path, int verbose)
Johannes Schindelin's avatar
Johannes Schindelin committed
354 355 356 357 358 359 360 361
{
	int size, namelen;
	struct stat st;
	struct cache_entry *ce;

	if (lstat(path, &st))
		die("%s: unable to stat (%s)", path, strerror(errno));

362 363
	if (!S_ISREG(st.st_mode) && !S_ISLNK(st.st_mode) && !S_ISDIR(st.st_mode))
		die("%s: can only add regular files, symbolic links or git-directories", path);
Johannes Schindelin's avatar
Johannes Schindelin committed
364 365

	namelen = strlen(path);
366 367 368 369
	if (S_ISDIR(st.st_mode)) {
		while (namelen && path[namelen-1] == '/')
			namelen--;
	}
Johannes Schindelin's avatar
Johannes Schindelin committed
370 371 372 373 374 375
	size = cache_entry_size(namelen);
	ce = xcalloc(1, size);
	memcpy(ce->name, path, namelen);
	ce->ce_flags = htons(namelen);
	fill_stat_cache_info(ce, &st);

376
	if (trust_executable_bit && has_symlinks)
377 378
		ce->ce_mode = create_ce_mode(st.st_mode);
	else {
379 380
		/* If there is an existing entry, pick the mode bits and type
		 * from it, otherwise assume unexecutable regular file.
Johannes Schindelin's avatar
Johannes Schindelin committed
381
		 */
382
		struct cache_entry *ent;
383
		int pos = index_name_pos(istate, path, namelen);
384

385
		ent = (0 <= pos) ? istate->cache[pos] : NULL;
386
		ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
Johannes Schindelin's avatar
Johannes Schindelin committed
387 388 389 390
	}

	if (index_path(ce->sha1, path, &st, 1))
		die("unable to index file %s", path);
391
	if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
Johannes Schindelin's avatar
Johannes Schindelin committed
392 393 394
		die("unable to add %s to index",path);
	if (verbose)
		printf("add '%s'\n", path);
395
	cache_tree_invalidate_path(istate->cache_tree, path);
Johannes Schindelin's avatar
Johannes Schindelin committed
396 397 398
	return 0;
}

399
int ce_same_name(struct cache_entry *a, struct cache_entry *b)
400 401 402 403 404
{
	int len = ce_namelen(a);
	return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
}

405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424
int ce_path_match(const struct cache_entry *ce, const char **pathspec)
{
	const char *match, *name;
	int len;

	if (!pathspec)
		return 1;

	len = ce_namelen(ce);
	name = ce->name;
	while ((match = *pathspec++) != NULL) {
		int matchlen = strlen(match);
		if (matchlen > len)
			continue;
		if (memcmp(name, match, matchlen))
			continue;
		if (matchlen && name[matchlen-1] == '/')
			return 1;
		if (name[matchlen] == '/' || !name[matchlen])
			return 1;
425 426
		if (!matchlen)
			return 1;
427 428 429 430
	}
	return 0;
}

431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494
/*
 * We fundamentally don't like some paths: we don't want
 * dot or dot-dot anywhere, and for obvious reasons don't
 * want to recurse into ".git" either.
 *
 * Also, we don't want double slashes or slashes at the
 * end that can make pathnames ambiguous.
 */
static int verify_dotfile(const char *rest)
{
	/*
	 * The first character was '.', but that
	 * has already been discarded, we now test
	 * the rest.
	 */
	switch (*rest) {
	/* "." is not allowed */
	case '\0': case '/':
		return 0;

	/*
	 * ".git" followed by  NUL or slash is bad. This
	 * shares the path end test with the ".." case.
	 */
	case 'g':
		if (rest[1] != 'i')
			break;
		if (rest[2] != 't')
			break;
		rest += 2;
	/* fallthrough */
	case '.':
		if (rest[1] == '\0' || rest[1] == '/')
			return 0;
	}
	return 1;
}

int verify_path(const char *path)
{
	char c;

	goto inside;
	for (;;) {
		if (!c)
			return 1;
		if (c == '/') {
inside:
			c = *path++;
			switch (c) {
			default:
				continue;
			case '/': case '\0':
				break;
			case '.':
				if (verify_dotfile(path))
					continue;
			}
			return 0;
		}
		c = *path++;
	}
}

495 496 497
/*
 * Do we have another file that has the beginning components being a
 * proper superset of the name we're trying to add?
498
 */
499 500
static int has_file_name(struct index_state *istate,
			 const struct cache_entry *ce, int pos, int ok_to_replace)
501
{
502 503
	int retval = 0;
	int len = ce_namelen(ce);
504
	int stage = ce_stage(ce);
505
	const char *name = ce->name;
506

507 508
	while (pos < istate->cache_nr) {
		struct cache_entry *p = istate->cache[pos++];
509

510
		if (len >= ce_namelen(p))
511
			break;
512 513
		if (memcmp(name, p->name, len))
			break;
514 515
		if (ce_stage(p) != stage)
			continue;
516 517
		if (p->name[len] != '/')
			continue;
518 519
		if (!ce_stage(p) && !p->ce_mode)
			continue;
520 521 522
		retval = -1;
		if (!ok_to_replace)
			break;
523
		remove_index_entry_at(istate, --pos);
524
	}
525 526
	return retval;
}
527

528 529 530 531
/*
 * Do we have another file with a pathname that is a proper
 * subset of the name we're trying to add?
 */
532 533
static int has_dir_name(struct index_state *istate,
			const struct cache_entry *ce, int pos, int ok_to_replace)
534 535
{
	int retval = 0;
536
	int stage = ce_stage(ce);
537 538
	const char *name = ce->name;
	const char *slash = name + ce_namelen(ce);
539

540 541
	for (;;) {
		int len;
542

543 544 545 546 547 548 549
		for (;;) {
			if (*--slash == '/')
				break;
			if (slash <= ce->name)
				return retval;
		}
		len = slash - name;
550

551
		pos = index_name_pos(istate, name, ntohs(create_ce_flags(len, stage)));
552
		if (pos >= 0) {
553 554 555 556 557 558 559 560
			/*
			 * Found one, but not so fast.  This could
			 * be a marker that says "I was here, but
			 * I am being removed".  Such an entry is
			 * not a part of the resulting tree, and
			 * it is Ok to have a directory at the same
			 * path.
			 */
561
			if (stage || istate->cache[pos]->ce_mode) {
562 563 564
				retval = -1;
				if (!ok_to_replace)
					break;
565
				remove_index_entry_at(istate, pos);
566 567
				continue;
			}
568
		}
569 570
		else
			pos = -pos-1;
571 572 573 574

		/*
		 * Trivial optimization: if we find an entry that
		 * already matches the sub-directory, then we know
575
		 * we're ok, and we can exit.
576
		 */
577 578
		while (pos < istate->cache_nr) {
			struct cache_entry *p = istate->cache[pos];
579 580 581 582
			if ((ce_namelen(p) <= len) ||
			    (p->name[len] != '/') ||
			    memcmp(p->name, name, len))
				break; /* not our subdirectory */
583
			if (ce_stage(p) == stage && (stage || p->ce_mode))
584 585 586 587 588 589 590
				/* p is at the same stage as our entry, and
				 * is a subdirectory of what we are looking
				 * at, so we cannot have conflicts at our
				 * level or anything shorter.
				 */
				return retval;
			pos++;
591
		}
592
	}
593 594 595 596 597 598 599 600 601 602 603 604
	return retval;
}

/* We may be in a situation where we already have path/file and path
 * is being added, or we already have path and path/file is being
 * added.  Either one would result in a nonsense tree that has path
 * twice when git-write-tree tries to write it out.  Prevent it.
 * 
 * If ok-to-replace is specified, we remove the conflicting entries
 * from the cache so the caller should recompute the insert position.
 * When this happens, we return non-zero.
 */
605 606 607
static int check_file_directory_conflict(struct index_state *istate,
					 const struct cache_entry *ce,
					 int pos, int ok_to_replace)
608
{
609 610 611 612 613 614 615 616
	int retval;

	/*
	 * When ce is an "I am going away" entry, we allow it to be added
	 */
	if (!ce_stage(ce) && !ce->ce_mode)
		return 0;

617 618 619
	/*
	 * We check if the path is a sub-path of a subsequent pathname
	 * first, since removing those will not change the position
620
	 * in the array.
621
	 */
622
	retval = has_file_name(istate, ce, pos, ok_to_replace);
623

624 625 626 627
	/*
	 * Then check if the path might have a clashing sub-directory
	 * before it.
	 */
628
	return retval + has_dir_name(istate, ce, pos, ok_to_replace);
629 630
}

631
int add_index_entry(struct index_state *istate, struct cache_entry *ce, int option)
632 633
{
	int pos;
634 635
	int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
	int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
636
	int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
Junio C Hamano's avatar
Junio C Hamano committed
637

638
	pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
639

Junio C Hamano's avatar
Junio C Hamano committed
640
	/* existing match? Just replace it. */
641
	if (pos >= 0) {
642 643
		istate->cache_changed = 1;
		istate->cache[pos] = ce;
644 645
		return 0;
	}
646
	pos = -pos-1;
647

648 649 650 651
	/*
	 * Inserting a merged entry ("stage 0") into the index
	 * will always replace all non-merged entries..
	 */
652 653
	if (pos < istate->cache_nr && ce_stage(ce) == 0) {
		while (ce_same_name(istate->cache[pos], ce)) {
654
			ok_to_add = 1;
655
			if (!remove_index_entry_at(istate, pos))
656 657 658 659
				break;
		}
	}

660 661
	if (!ok_to_add)
		return -1;
662 663
	if (!verify_path(ce->name))
		return -1;
664

Junio C Hamano's avatar
Junio C Hamano committed
665
	if (!skip_df_check &&
666
	    check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
667
		if (!ok_to_replace)
668 669 670
			return error("'%s' appears as both a file and as a directory",
				     ce->name);
		pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
671 672
		pos = -pos-1;
	}
673

674
	/* Make sure the array is big enough .. */
675 676 677 678
	if (istate->cache_nr == istate->cache_alloc) {
		istate->cache_alloc = alloc_nr(istate->cache_alloc);
		istate->cache = xrealloc(istate->cache,
					istate->cache_alloc * sizeof(struct cache_entry *));
679 680 681
	}

	/* Add it in.. */
682 683 684 685 686 687 688
	istate->cache_nr++;
	if (istate->cache_nr > pos)
		memmove(istate->cache + pos + 1,
			istate->cache + pos,
			(istate->cache_nr - pos - 1) * sizeof(ce));
	istate->cache[pos] = ce;
	istate->cache_changed = 1;
689 690 691
	return 0;
}

692 693 694 695 696 697 698 699 700 701 702
/*
 * "refresh" does not calculate a new sha1 file or bring the
 * cache up-to-date for mode/content changes. But what it
 * _does_ do is to "re-match" the stat information of a file
 * with the cache, so that you can refresh the cache for a
 * file that hasn't been changed but where the stat entry is
 * out of date.
 *
 * For example, you'd want to do this after doing a "git-read-tree",
 * to link up the stat cache details with the proper files.
 */
703 704
static struct cache_entry *refresh_cache_ent(struct index_state *istate,
					     struct cache_entry *ce, int really, int *err)
705 706 707 708 709
{
	struct stat st;
	struct cache_entry *updated;
	int changed, size;

710
	if (lstat(ce->name, &st) < 0) {
711 712
		if (err)
			*err = errno;
713 714
		return NULL;
	}
715

716
	changed = ie_match_stat(istate, ce, &st, really);
717 718 719 720 721
	if (!changed) {
		if (really && assume_unchanged &&
		    !(ce->ce_flags & htons(CE_VALID)))
			; /* mark this one VALID again */
		else
722
			return ce;
723 724
	}

725
	if (ie_modified(istate, ce, &st, really)) {
726 727
		if (err)
			*err = EINVAL;
728 729
		return NULL;
	}
730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747

	size = ce_size(ce);
	updated = xmalloc(size);
	memcpy(updated, ce, size);
	fill_stat_cache_info(updated, &st);

	/* In this case, if really is not set, we should leave
	 * CE_VALID bit alone.  Otherwise, paths marked with
	 * --no-assume-unchanged (i.e. things to be edited) will
	 * reacquire CE_VALID bit automatically, which is not
	 * really what we want.
	 */
	if (!really && assume_unchanged && !(ce->ce_flags & htons(CE_VALID)))
		updated->ce_flags &= ~htons(CE_VALID);

	return updated;
}

748
int refresh_index(struct index_state *istate, unsigned int flags)
749 750 751 752 753 754 755 756
{
	int i;
	int has_errors = 0;
	int really = (flags & REFRESH_REALLY) != 0;
	int allow_unmerged = (flags & REFRESH_UNMERGED) != 0;
	int quiet = (flags & REFRESH_QUIET) != 0;
	int not_new = (flags & REFRESH_IGNORE_MISSING) != 0;

757
	for (i = 0; i < istate->cache_nr; i++) {
758
		struct cache_entry *ce, *new;
759 760
		int cache_errno = 0;

761
		ce = istate->cache[i];
762
		if (ce_stage(ce)) {
763 764
			while ((i < istate->cache_nr) &&
			       ! strcmp(istate->cache[i]->name, ce->name))
765 766 767 768 769 770 771 772 773
				i++;
			i--;
			if (allow_unmerged)
				continue;
			printf("%s: needs merge\n", ce->name);
			has_errors = 1;
			continue;
		}

774
		new = refresh_cache_ent(istate, ce, really, &cache_errno);
775
		if (new == ce)
776
			continue;
777 778
		if (!new) {
			if (not_new && cache_errno == ENOENT)
779
				continue;
780
			if (really && cache_errno == EINVAL) {
781 782 783 784
				/* If we are doing --really-refresh that
				 * means the index is not valid anymore.
				 */
				ce->ce_flags &= ~htons(CE_VALID);
785
				istate->cache_changed = 1;
786 787 788 789 790 791 792
			}
			if (quiet)
				continue;
			printf("%s: needs update\n", ce->name);
			has_errors = 1;
			continue;
		}
793 794
		istate->cache_changed = 1;
		/* You can NOT just free istate->cache[i] here, since it
795 796
		 * might not be necessarily malloc()ed but can also come
		 * from mmap(). */
797
		istate->cache[i] = new;
798 799 800 801
	}
	return has_errors;
}

802 803
struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
{
804
	return refresh_cache_ent(&the_index, ce, really, NULL);
805 806
}

807
static int verify_hdr(struct cache_header *hdr, unsigned long size)
808 809
{
	SHA_CTX c;
810
	unsigned char sha1[20];
811

812
	if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
813
		return error("bad signature");
814 815
	if (hdr->hdr_version != htonl(2))
		return error("bad index version");
816
	SHA1_Init(&c);
817
	SHA1_Update(&c, hdr, size - 20);
818
	SHA1_Final(sha1, &c);
819
	if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
820
		return error("bad index file sha1 signature");
821 822 823
	return 0;
}

824 825
static int read_index_extension(struct index_state *istate,
				const char *ext, void *data, unsigned long sz)
826 827 828
{
	switch (CACHE_EXT(ext)) {
	case CACHE_EXT_TREE:
829
		istate->cache_tree = cache_tree_read(data, sz);
830 831 832 833 834 835 836 837 838 839 840
		break;
	default:
		if (*ext < 'A' || 'Z' < *ext)
			return error("index uses %.4s extension, which we do not understand",
				     ext);
		fprintf(stderr, "ignoring %.4s extension\n", ext);
		break;
	}
	return 0;
}

841
int read_index(struct index_state *istate)
842
{
843
	return read_index_from(istate, get_index_file());
844 845 846
}

/* remember to discard_cache() before reading a different cache! */
847
int read_index_from(struct index_state *istate, const char *path)
848 849 850
{
	int fd, i;
	struct stat st;
851
	unsigned long offset;
852 853 854
	struct cache_header *hdr;

	errno = EBUSY;
855 856
	if (istate->mmap)
		return istate->cache_nr;
857

858
	errno = ENOENT;
859
	istate->timestamp = 0;
860
	fd = open(path, O_RDONLY);
861 862 863 864 865
	if (fd < 0) {
		if (errno == ENOENT)
			return 0;
		die("index file open failed (%s)", strerror(errno));
	}
866 867

	if (!fstat(fd, &st)) {
868
		istate->mmap_size = xsize_t(st.st_size);
869
		errno = EINVAL;
870 871 872
		if (istate->mmap_size >= sizeof(struct cache_header) + 20)
			istate->mmap = xmmap(NULL, istate->mmap_size,
					    PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
873 874 875 876
		else
			die("index file smaller than expected");
	} else
		die("cannot stat the open index (%s)", strerror(errno));
877 878
	close(fd);

879 880
	hdr = istate->mmap;
	if (verify_hdr(hdr, istate->mmap_size) < 0)
881 882
		goto unmap;

883 884 885
	istate->cache_nr = ntohl(hdr->hdr_entries);
	istate->cache_alloc = alloc_nr(istate->cache_nr);
	istate->cache = xcalloc(istate->cache_alloc, sizeof(struct cache_entry *));
886 887

	offset = sizeof(*hdr);
888 889 890 891
	for (i = 0; i < istate->cache_nr; i++) {
		struct cache_entry *ce;

		ce = (struct cache_entry *)((char *)(istate->mmap) + offset);
892
		offset = offset + ce_size(ce);
893
		istate->cache[i] = ce;
894
	}
895 896
	istate->timestamp = st.st_mtime;
	while (offset <= istate->mmap_size - 20 - 8) {
897 898 899 900 901 902 903
		/* After an array of active_nr index entries,
		 * there can be arbitrary number of extended
		 * sections, each of which is prefixed with
		 * extension name (4-byte) and section length
		 * in 4-byte network byte order.
		 */
		unsigned long extsize;
904
		memcpy(&extsize, (char *)(istate->mmap) + offset + 4, 4);
905
		extsize = ntohl(extsize);
906 907 908
		if (read_index_extension(istate,
					 ((const char *) (istate->mmap)) + offset,
					 (char *) (istate->mmap) + offset + 8,
909
					 extsize) < 0)
910 911 912 913
			goto unmap;
		offset += 8;
		offset += extsize;
	}
914
	return istate->cache_nr;
915 916

unmap:
917
	munmap(istate->mmap, istate->mmap_size);
918
	errno = EINVAL;
919
	die("index file corrupt");
920 921
}

922
int discard_index(struct index_state *istate)
923 924 925
{
	int ret;

926 927 928 929 930
	istate->cache_nr = 0;
	istate->cache_changed = 0;
	istate->timestamp = 0;
	cache_tree_free(&(istate->cache_tree));
	if (istate->mmap == NULL)
931
		return 0;
932 933 934
	ret = munmap(istate->mmap, istate->mmap_size);
	istate->mmap = NULL;
	istate->mmap_size = 0;
935 936 937 938 939

	/* no need to throw away allocated active_cache */
	return ret;
}

940
#define WRITE_BUFFER_SIZE 8192
941
static unsigned char write_buffer[WRITE_BUFFER_SIZE];
942 943
static unsigned long write_buffer_len;

944 945 946 947 948
static int ce_write_flush(SHA_CTX *context, int fd)
{
	unsigned int buffered = write_buffer_len;
	if (buffered) {
		SHA1_Update(context, write_buffer, buffered);
949
		if (write_in_full(fd, write_buffer, buffered) != buffered)
950 951 952 953 954 955
			return -1;
		write_buffer_len = 0;
	}
	return 0;
}

956
static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
957 958 959 960 961 962 963 964 965
{
	while (len) {
		unsigned int buffered = write_buffer_len;
		unsigned int partial = WRITE_BUFFER_SIZE - buffered;
		if (partial > len)
			partial = len;
		memcpy(write_buffer + buffered, data, partial);
		buffered += partial;
		if (buffered == WRITE_BUFFER_SIZE) {
966 967
			write_buffer_len = buffered;
			if (ce_write_flush(context, fd))
968 969 970 971 972
				return -1;
			buffered = 0;
		}
		write_buffer_len = buffered;
		len -= partial;
973
		data = (char *) data + partial;
974 975 976 977
 	}
 	return 0;
}

978
static int write_index_ext_header(SHA_CTX *context, int fd,
979
				  unsigned int ext, unsigned int sz)
980 981 982
{
	ext = htonl(ext);
	sz = htonl(sz);
David Rientjes's avatar
David Rientjes committed
983 984
	return ((ce_write(context, fd, &ext, 4) < 0) ||
		(ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
985 986 987
}

static int ce_flush(SHA_CTX *context, int fd)
988 989
{
	unsigned int left = write_buffer_len;
990

991 992
	if (left) {
		write_buffer_len = 0;
993
		SHA1_Update(context, write_buffer, left);
994
	}
995

996 997
	/* Flush first if not enough space for SHA1 signature */
	if (left + 20 > WRITE_BUFFER_SIZE) {
998
		if (write_in_full(fd, write_buffer, left) != left)
999 1000 1001 1002
			return -1;
		left = 0;
	}

1003
	/* Append the SHA1 signature at the end */
1004
	SHA1_Final(write_buffer + left, context);
1005
	left += 20;
1006
	return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
1007 1008
}

Junio C Hamano's avatar
Junio C Hamano committed
1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
static void ce_smudge_racily_clean_entry(struct cache_entry *ce)
{
	/*
	 * The only thing we care about in this function is to smudge the
	 * falsely clean entry due to touch-update-touch race, so we leave
	 * everything else as they are.  We are called for entries whose
	 * ce_mtime match the index file mtime.
	 */
	struct stat st;

	if (lstat(ce->name, &st) < 0)
		return;
	if (ce_match_stat_basic(ce, &st))
		return;
	if (ce_modified_check_fs(ce, &st)) {
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034
		/* This is "racily clean"; smudge it.  Note that this
		 * is a tricky code.  At first glance, it may appear
		 * that it can break with this sequence:
		 *
		 * $ echo xyzzy >frotz
		 * $ git-update-index --add frotz
		 * $ : >frotz
		 * $ sleep 3
		 * $ echo filfre >nitfol
		 * $ git-update-index --add nitfol
		 *
1035
		 * but it does not.  When the second update-index runs,
1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048
		 * it notices that the entry "frotz" has the same timestamp
		 * as index, and if we were to smudge it by resetting its
		 * size to zero here, then the object name recorded
		 * in index is the 6-byte file but the cached stat information
		 * becomes zero --- which would then match what we would
		 * obtain from the filesystem next time we stat("frotz"). 
		 *
		 * However, the second update-index, before calling
		 * this function, notices that the cached size is 6
		 * bytes and what is on the filesystem is an empty
		 * file, and never calls us, so the cached size information
		 * for "frotz" stays 6 which does not match the filesystem.
		 */
Junio C Hamano's avatar
Junio C Hamano committed
1049 1050 1051 1052
		ce->ce_size = htonl(0);
	}
}

1053
int write_index(struct index_state *istate, int newfd)
1054 1055 1056
{
	SHA_CTX c;
	struct cache_header hdr;
1057
	int i, removed;
1058 1059
	struct cache_entry **cache = istate->cache;
	int entries = istate->cache_nr;
1060 1061 1062 1063

	for (i = removed = 0; i < entries; i++)
		if (!cache[i]->ce_mode)
			removed++;
1064

1065
	hdr.hdr_signature = htonl(CACHE_SIGNATURE);
1066
	hdr.hdr_version = htonl(2);
1067
	hdr.hdr_entries = htonl(entries - removed);
1068 1069

	SHA1_Init(&c);
1070
	if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
1071 1072 1073 1074
		return -1;

	for (i = 0; i < entries; i++) {
		struct cache_entry *ce = cache[i];
1075 1076
		if (!ce->ce_mode)
			continue;
1077 1078
		if (istate->timestamp &&
		    istate->timestamp <= ntohl(ce->ce_mtime.sec))
Junio C Hamano's avatar
Junio C Hamano committed
1079
			ce_smudge_racily_clean_entry(ce);
1080
		if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
1081 1082
			return -1;
	}
1083

1084
	/* Write extension data here */
1085
	if (istate->cache_tree) {
1086
		unsigned long sz;
1087
		void *data = cache_tree_write(istate->cache_tree, &sz);
1088 1089 1090
		if (data &&
		    !write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sz) &&
		    !ce_write(&c, newfd, data, sz))
1091
			free(data);
1092 1093 1094 1095 1096 1097
		else {
			free(data);
			return -1;
		}
	}
	return ce_flush(&c, newfd);
1098
}