read-cache.c 28.2 KB
Newer Older
L
Linus Torvalds 已提交
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);
J
Junio C Hamano 已提交
43 44 45

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

J
Junio C Hamano 已提交
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 */
J
Junio C Hamano 已提交
58 59 60 61
	}
	return match;
}

62
static int ce_compare_link(struct cache_entry *ce, size_t expected_size)
J
Junio C Hamano 已提交
63 64 65 66 67
{
	int match = -1;
	char *target;
	void *buffer;
	unsigned long size;
68
	enum object_type type;
J
Junio C Hamano 已提交
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);
J
Junio C Hamano 已提交
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
static int ce_compare_gitlink(struct cache_entry *ce)
{
	unsigned char sha1[20];

	/*
	 * We don't actually require that the .git directory
M
Martin Waitz 已提交
95
	 * under GITLINK directory be a valid git directory. It
96 97 98 99 100 101 102 103 104 105
	 * 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);
}

J
Junio C Hamano 已提交
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)))
J
Junio C Hamano 已提交
115 116
			return DATA_CHANGED;
		break;
117
	case S_IFDIR:
M
Martin Waitz 已提交
118
		if (S_ISGITLINK(ntohl(ce->ce_mode)))
119
			return 0;
J
Junio C Hamano 已提交
120 121 122 123 124 125
	default:
		return TYPE_CHANGED;
	}
	return 0;
}

J
Junio C Hamano 已提交
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;
J
Junio C Hamano 已提交
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;
M
Martin Waitz 已提交
145
	case S_IFGITLINK:
146 147 148 149
		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;
J
Junio C Hamano 已提交
169
#endif
170 171 172

	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

J
Junio C Hamano 已提交
190 191 192
	return changed;
}

193 194
int ie_match_stat(struct index_state *istate,
		  struct cache_entry *ce, struct stat *st, int options)
J
Junio C Hamano 已提交
195
{
J
Junio C Hamano 已提交
196
	unsigned int changed;
197 198
	int ignore_valid = options & 01;
	int assume_racy_is_modified = options & 02;
J
Junio C Hamano 已提交
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);
J
Junio C Hamano 已提交
208

J
Junio C Hamano 已提交
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
J
Junio C Hamano 已提交
223
	 * carefully than others.
J
Junio C Hamano 已提交
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

J
Junio C Hamano 已提交
234
	return changed;
235 236
}

237 238
int ie_modified(struct index_state *istate,
		struct cache_entry *ce, struct stat *st, int really)
239
{
J
Junio C Hamano 已提交
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;

J
Junio C Hamano 已提交
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;
J
Junio C Hamano 已提交
297

298 299 300
	/* Compare stages  */
	flags1 &= CE_STAGEMASK;
	flags2 &= CE_STAGEMASK;
J
Junio C Hamano 已提交
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;
}

353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380
static int compare_name(struct cache_entry *ce, const char *path, int namelen)
{
	return namelen != ce_namelen(ce) || memcmp(path, ce->name, namelen);
}

static int index_name_pos_also_unmerged(struct index_state *istate,
	const char *path, int namelen)
{
	int pos = index_name_pos(istate, path, namelen);
	struct cache_entry *ce;

	if (pos >= 0)
		return pos;

	/* maybe unmerged? */
	pos = -1 - pos;
	if (pos >= istate->cache_nr ||
			compare_name((ce = istate->cache[pos]), path, namelen))
		return -1;

	/* order of preference: stage 2, 1, 3 */
	if (ce_stage(ce) == 1 && pos + 1 < istate->cache_nr &&
			ce_stage((ce = istate->cache[pos + 1])) == 2 &&
			!compare_name(ce, path, namelen))
		pos++;
	return pos;
}

381
int add_file_to_index(struct index_state *istate, const char *path, int verbose)
J
Johannes Schindelin 已提交
382
{
383
	int size, namelen, pos;
J
Johannes Schindelin 已提交
384 385 386 387 388 389
	struct stat st;
	struct cache_entry *ce;

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

390 391
	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);
J
Johannes Schindelin 已提交
392 393

	namelen = strlen(path);
394 395 396 397
	if (S_ISDIR(st.st_mode)) {
		while (namelen && path[namelen-1] == '/')
			namelen--;
	}
J
Johannes Schindelin 已提交
398 399 400 401 402 403
	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);

404
	if (trust_executable_bit && has_symlinks)
405 406
		ce->ce_mode = create_ce_mode(st.st_mode);
	else {
407 408
		/* If there is an existing entry, pick the mode bits and type
		 * from it, otherwise assume unexecutable regular file.
J
Johannes Schindelin 已提交
409
		 */
410
		struct cache_entry *ent;
411
		int pos = index_name_pos_also_unmerged(istate, path, namelen);
412

413
		ent = (0 <= pos) ? istate->cache[pos] : NULL;
414
		ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
J
Johannes Schindelin 已提交
415 416
	}

417 418 419 420 421 422 423 424 425
	pos = index_name_pos(istate, ce->name, namelen);
	if (0 <= pos &&
	    !ce_stage(istate->cache[pos]) &&
	    !ie_modified(istate, istate->cache[pos], &st, 1)) {
		/* Nothing changed, really */
		free(ce);
		return 0;
	}

J
Johannes Schindelin 已提交
426 427
	if (index_path(ce->sha1, path, &st, 1))
		die("unable to index file %s", path);
428
	if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
J
Johannes Schindelin 已提交
429 430 431
		die("unable to add %s to index",path);
	if (verbose)
		printf("add '%s'\n", path);
432
	cache_tree_invalidate_path(istate->cache_tree, path);
J
Johannes Schindelin 已提交
433 434 435
	return 0;
}

436
int ce_same_name(struct cache_entry *a, struct cache_entry *b)
437 438 439 440 441
{
	int len = ce_namelen(a);
	return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
}

442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461
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;
462 463
		if (!matchlen)
			return 1;
464 465 466 467
	}
	return 0;
}

468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
/*
 * 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++;
	}
}

532 533 534
/*
 * Do we have another file that has the beginning components being a
 * proper superset of the name we're trying to add?
535
 */
536 537
static int has_file_name(struct index_state *istate,
			 const struct cache_entry *ce, int pos, int ok_to_replace)
538
{
539 540
	int retval = 0;
	int len = ce_namelen(ce);
541
	int stage = ce_stage(ce);
542
	const char *name = ce->name;
543

544 545
	while (pos < istate->cache_nr) {
		struct cache_entry *p = istate->cache[pos++];
546

547
		if (len >= ce_namelen(p))
548
			break;
549 550
		if (memcmp(name, p->name, len))
			break;
551 552
		if (ce_stage(p) != stage)
			continue;
553 554
		if (p->name[len] != '/')
			continue;
555 556
		if (!ce_stage(p) && !p->ce_mode)
			continue;
557 558 559
		retval = -1;
		if (!ok_to_replace)
			break;
560
		remove_index_entry_at(istate, --pos);
561
	}
562 563
	return retval;
}
564

565 566 567 568
/*
 * Do we have another file with a pathname that is a proper
 * subset of the name we're trying to add?
 */
569 570
static int has_dir_name(struct index_state *istate,
			const struct cache_entry *ce, int pos, int ok_to_replace)
571 572
{
	int retval = 0;
573
	int stage = ce_stage(ce);
574 575
	const char *name = ce->name;
	const char *slash = name + ce_namelen(ce);
576

577 578
	for (;;) {
		int len;
579

580 581 582 583 584 585 586
		for (;;) {
			if (*--slash == '/')
				break;
			if (slash <= ce->name)
				return retval;
		}
		len = slash - name;
587

588
		pos = index_name_pos(istate, name, ntohs(create_ce_flags(len, stage)));
589
		if (pos >= 0) {
590 591 592 593 594 595 596 597
			/*
			 * 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.
			 */
598
			if (stage || istate->cache[pos]->ce_mode) {
599 600 601
				retval = -1;
				if (!ok_to_replace)
					break;
602
				remove_index_entry_at(istate, pos);
603 604
				continue;
			}
605
		}
606 607
		else
			pos = -pos-1;
608 609 610 611

		/*
		 * Trivial optimization: if we find an entry that
		 * already matches the sub-directory, then we know
612
		 * we're ok, and we can exit.
613
		 */
614 615
		while (pos < istate->cache_nr) {
			struct cache_entry *p = istate->cache[pos];
616 617 618 619
			if ((ce_namelen(p) <= len) ||
			    (p->name[len] != '/') ||
			    memcmp(p->name, name, len))
				break; /* not our subdirectory */
620
			if (ce_stage(p) == stage && (stage || p->ce_mode))
621 622 623 624 625 626 627
				/* 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++;
628
		}
629
	}
630 631 632 633 634 635 636
	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.
J
Junio C Hamano 已提交
637
 *
638 639 640 641
 * 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.
 */
642 643 644
static int check_file_directory_conflict(struct index_state *istate,
					 const struct cache_entry *ce,
					 int pos, int ok_to_replace)
645
{
646 647 648 649 650 651 652 653
	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;

654 655 656
	/*
	 * We check if the path is a sub-path of a subsequent pathname
	 * first, since removing those will not change the position
657
	 * in the array.
658
	 */
659
	retval = has_file_name(istate, ce, pos, ok_to_replace);
660

661 662 663 664
	/*
	 * Then check if the path might have a clashing sub-directory
	 * before it.
	 */
665
	return retval + has_dir_name(istate, ce, pos, ok_to_replace);
666 667
}

668
int add_index_entry(struct index_state *istate, struct cache_entry *ce, int option)
669 670
{
	int pos;
671 672
	int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
	int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
673
	int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
J
Junio C Hamano 已提交
674

675
	pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
676

J
Junio C Hamano 已提交
677
	/* existing match? Just replace it. */
678
	if (pos >= 0) {
679 680
		istate->cache_changed = 1;
		istate->cache[pos] = ce;
681 682
		return 0;
	}
683
	pos = -pos-1;
684

685 686 687 688
	/*
	 * Inserting a merged entry ("stage 0") into the index
	 * will always replace all non-merged entries..
	 */
689 690
	if (pos < istate->cache_nr && ce_stage(ce) == 0) {
		while (ce_same_name(istate->cache[pos], ce)) {
691
			ok_to_add = 1;
692
			if (!remove_index_entry_at(istate, pos))
693 694 695 696
				break;
		}
	}

697 698
	if (!ok_to_add)
		return -1;
699 700
	if (!verify_path(ce->name))
		return -1;
701

J
Junio C Hamano 已提交
702
	if (!skip_df_check &&
703
	    check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
704
		if (!ok_to_replace)
705 706 707
			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));
708 709
		pos = -pos-1;
	}
710

711
	/* Make sure the array is big enough .. */
712 713 714 715
	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 *));
716 717 718
	}

	/* Add it in.. */
719 720 721 722 723 724 725
	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;
726 727 728
	return 0;
}

L
Linus Torvalds 已提交
729 730 731 732 733 734 735 736 737 738 739
/*
 * "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.
 */
740 741
static struct cache_entry *refresh_cache_ent(struct index_state *istate,
					     struct cache_entry *ce, int really, int *err)
L
Linus Torvalds 已提交
742 743 744 745 746
{
	struct stat st;
	struct cache_entry *updated;
	int changed, size;

747
	if (lstat(ce->name, &st) < 0) {
748 749
		if (err)
			*err = errno;
750 751
		return NULL;
	}
L
Linus Torvalds 已提交
752

753
	changed = ie_match_stat(istate, ce, &st, really);
L
Linus Torvalds 已提交
754 755 756 757 758
	if (!changed) {
		if (really && assume_unchanged &&
		    !(ce->ce_flags & htons(CE_VALID)))
			; /* mark this one VALID again */
		else
759
			return ce;
L
Linus Torvalds 已提交
760 761
	}

762
	if (ie_modified(istate, ce, &st, really)) {
763 764
		if (err)
			*err = EINVAL;
765 766
		return NULL;
	}
L
Linus Torvalds 已提交
767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784

	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;
}

785
int refresh_index(struct index_state *istate, unsigned int flags)
L
Linus Torvalds 已提交
786 787 788 789 790 791 792 793
{
	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;

794
	for (i = 0; i < istate->cache_nr; i++) {
L
Linus Torvalds 已提交
795
		struct cache_entry *ce, *new;
796 797
		int cache_errno = 0;

798
		ce = istate->cache[i];
L
Linus Torvalds 已提交
799
		if (ce_stage(ce)) {
800 801
			while ((i < istate->cache_nr) &&
			       ! strcmp(istate->cache[i]->name, ce->name))
L
Linus Torvalds 已提交
802 803 804 805 806 807 808 809 810
				i++;
			i--;
			if (allow_unmerged)
				continue;
			printf("%s: needs merge\n", ce->name);
			has_errors = 1;
			continue;
		}

811
		new = refresh_cache_ent(istate, ce, really, &cache_errno);
812
		if (new == ce)
L
Linus Torvalds 已提交
813
			continue;
814 815
		if (!new) {
			if (not_new && cache_errno == ENOENT)
L
Linus Torvalds 已提交
816
				continue;
817
			if (really && cache_errno == EINVAL) {
L
Linus Torvalds 已提交
818 819 820 821
				/* If we are doing --really-refresh that
				 * means the index is not valid anymore.
				 */
				ce->ce_flags &= ~htons(CE_VALID);
822
				istate->cache_changed = 1;
L
Linus Torvalds 已提交
823 824 825 826 827 828 829
			}
			if (quiet)
				continue;
			printf("%s: needs update\n", ce->name);
			has_errors = 1;
			continue;
		}
830 831
		istate->cache_changed = 1;
		/* You can NOT just free istate->cache[i] here, since it
L
Linus Torvalds 已提交
832 833
		 * might not be necessarily malloc()ed but can also come
		 * from mmap(). */
834
		istate->cache[i] = new;
L
Linus Torvalds 已提交
835 836 837 838
	}
	return has_errors;
}

839 840
struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
{
841
	return refresh_cache_ent(&the_index, ce, really, NULL);
842 843
}

844
static int verify_hdr(struct cache_header *hdr, unsigned long size)
845 846
{
	SHA_CTX c;
847
	unsigned char sha1[20];
848

849
	if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
850
		return error("bad signature");
851 852
	if (hdr->hdr_version != htonl(2))
		return error("bad index version");
853
	SHA1_Init(&c);
854
	SHA1_Update(&c, hdr, size - 20);
855
	SHA1_Final(sha1, &c);
856
	if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
857
		return error("bad index file sha1 signature");
858 859 860
	return 0;
}

861 862
static int read_index_extension(struct index_state *istate,
				const char *ext, void *data, unsigned long sz)
863 864 865
{
	switch (CACHE_EXT(ext)) {
	case CACHE_EXT_TREE:
866
		istate->cache_tree = cache_tree_read(data, sz);
867 868 869 870 871 872 873 874 875 876 877
		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;
}

878
int read_index(struct index_state *istate)
879
{
880
	return read_index_from(istate, get_index_file());
881 882 883
}

/* remember to discard_cache() before reading a different cache! */
884
int read_index_from(struct index_state *istate, const char *path)
885 886 887
{
	int fd, i;
	struct stat st;
888
	unsigned long offset;
889 890 891
	struct cache_header *hdr;

	errno = EBUSY;
892 893
	if (istate->mmap)
		return istate->cache_nr;
894

895
	errno = ENOENT;
896
	istate->timestamp = 0;
897
	fd = open(path, O_RDONLY);
898 899 900 901 902
	if (fd < 0) {
		if (errno == ENOENT)
			return 0;
		die("index file open failed (%s)", strerror(errno));
	}
903

904
	if (fstat(fd, &st))
905
		die("cannot stat the open index (%s)", strerror(errno));
906 907 908 909 910 911 912

	errno = EINVAL;
	istate->mmap_size = xsize_t(st.st_size);
	if (istate->mmap_size < sizeof(struct cache_header) + 20)
		die("index file smaller than expected");

	istate->mmap = xmmap(NULL, istate->mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
913 914
	close(fd);

915 916
	hdr = istate->mmap;
	if (verify_hdr(hdr, istate->mmap_size) < 0)
917 918
		goto unmap;

919 920 921
	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 *));
922 923

	offset = sizeof(*hdr);
924 925 926 927
	for (i = 0; i < istate->cache_nr; i++) {
		struct cache_entry *ce;

		ce = (struct cache_entry *)((char *)(istate->mmap) + offset);
928
		offset = offset + ce_size(ce);
929
		istate->cache[i] = ce;
930
	}
931 932
	istate->timestamp = st.st_mtime;
	while (offset <= istate->mmap_size - 20 - 8) {
933 934 935 936 937 938 939
		/* 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;
940
		memcpy(&extsize, (char *)(istate->mmap) + offset + 4, 4);
941
		extsize = ntohl(extsize);
942 943 944
		if (read_index_extension(istate,
					 ((const char *) (istate->mmap)) + offset,
					 (char *) (istate->mmap) + offset + 8,
945
					 extsize) < 0)
946 947 948 949
			goto unmap;
		offset += 8;
		offset += extsize;
	}
950
	return istate->cache_nr;
951 952

unmap:
953
	munmap(istate->mmap, istate->mmap_size);
954
	errno = EINVAL;
955
	die("index file corrupt");
956 957
}

958
int discard_index(struct index_state *istate)
959 960 961
{
	int ret;

962 963 964 965 966
	istate->cache_nr = 0;
	istate->cache_changed = 0;
	istate->timestamp = 0;
	cache_tree_free(&(istate->cache_tree));
	if (istate->mmap == NULL)
967
		return 0;
968 969 970
	ret = munmap(istate->mmap, istate->mmap_size);
	istate->mmap = NULL;
	istate->mmap_size = 0;
971 972 973 974 975

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

976
#define WRITE_BUFFER_SIZE 8192
977
static unsigned char write_buffer[WRITE_BUFFER_SIZE];
978 979
static unsigned long write_buffer_len;

980 981 982 983 984
static int ce_write_flush(SHA_CTX *context, int fd)
{
	unsigned int buffered = write_buffer_len;
	if (buffered) {
		SHA1_Update(context, write_buffer, buffered);
985
		if (write_in_full(fd, write_buffer, buffered) != buffered)
986 987 988 989 990 991
			return -1;
		write_buffer_len = 0;
	}
	return 0;
}

992
static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
993 994 995 996 997 998 999 1000 1001
{
	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) {
1002 1003
			write_buffer_len = buffered;
			if (ce_write_flush(context, fd))
1004 1005 1006 1007 1008
				return -1;
			buffered = 0;
		}
		write_buffer_len = buffered;
		len -= partial;
1009
		data = (char *) data + partial;
J
Junio C Hamano 已提交
1010 1011
	}
	return 0;
1012 1013
}

1014
static int write_index_ext_header(SHA_CTX *context, int fd,
1015
				  unsigned int ext, unsigned int sz)
1016 1017 1018
{
	ext = htonl(ext);
	sz = htonl(sz);
D
David Rientjes 已提交
1019 1020
	return ((ce_write(context, fd, &ext, 4) < 0) ||
		(ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
1021 1022 1023
}

static int ce_flush(SHA_CTX *context, int fd)
1024 1025
{
	unsigned int left = write_buffer_len;
1026

1027 1028
	if (left) {
		write_buffer_len = 0;
1029
		SHA1_Update(context, write_buffer, left);
1030
	}
1031

1032 1033
	/* Flush first if not enough space for SHA1 signature */
	if (left + 20 > WRITE_BUFFER_SIZE) {
1034
		if (write_in_full(fd, write_buffer, left) != left)
1035 1036 1037 1038
			return -1;
		left = 0;
	}

1039
	/* Append the SHA1 signature at the end */
1040
	SHA1_Final(write_buffer + left, context);
1041
	left += 20;
1042
	return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
1043 1044
}

J
Junio C Hamano 已提交
1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059
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)) {
1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070
		/* 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
		 *
1071
		 * but it does not.  When the second update-index runs,
1072 1073 1074 1075 1076
		 * 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
J
Junio C Hamano 已提交
1077
		 * obtain from the filesystem next time we stat("frotz").
1078 1079 1080 1081 1082 1083 1084
		 *
		 * 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.
		 */
J
Junio C Hamano 已提交
1085 1086 1087 1088
		ce->ce_size = htonl(0);
	}
}

1089
int write_index(struct index_state *istate, int newfd)
1090 1091 1092
{
	SHA_CTX c;
	struct cache_header hdr;
1093
	int i, removed;
1094 1095
	struct cache_entry **cache = istate->cache;
	int entries = istate->cache_nr;
1096 1097 1098 1099

	for (i = removed = 0; i < entries; i++)
		if (!cache[i]->ce_mode)
			removed++;
1100

1101
	hdr.hdr_signature = htonl(CACHE_SIGNATURE);
1102
	hdr.hdr_version = htonl(2);
1103
	hdr.hdr_entries = htonl(entries - removed);
1104 1105

	SHA1_Init(&c);
1106
	if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
1107 1108 1109 1110
		return -1;

	for (i = 0; i < entries; i++) {
		struct cache_entry *ce = cache[i];
1111 1112
		if (!ce->ce_mode)
			continue;
1113 1114
		if (istate->timestamp &&
		    istate->timestamp <= ntohl(ce->ce_mtime.sec))
J
Junio C Hamano 已提交
1115
			ce_smudge_racily_clean_entry(ce);
1116
		if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
1117 1118
			return -1;
	}
1119

1120
	/* Write extension data here */
1121
	if (istate->cache_tree) {
1122
		unsigned long sz;
1123
		void *data = cache_tree_write(istate->cache_tree, &sz);
1124 1125 1126
		if (data &&
		    !write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sz) &&
		    !ce_write(&c, newfd, data, sz))
1127
			free(data);
1128 1129 1130 1131 1132 1133
		else {
			free(data);
			return -1;
		}
	}
	return ce_flush(&c, newfd);
1134
}