read-cache.c 31 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
#include "dir.h"
11 12 13 14 15 16 17 18 19 20 21 22

/* 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" */
23

24
struct index_state the_index;
25

26 27 28 29 30 31 32
/*
 * 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)
{
33 34 35 36 37 38 39
	ce->ce_ctime = st->st_ctime;
	ce->ce_mtime = st->st_mtime;
	ce->ce_dev = st->st_dev;
	ce->ce_ino = st->st_ino;
	ce->ce_uid = st->st_uid;
	ce->ce_gid = st->st_gid;
	ce->ce_size = st->st_size;
Junio C Hamano's avatar
Junio C Hamano committed
40 41

	if (assume_unchanged)
42
		ce->ce_flags |= CE_VALID;
43 44
}

Junio C Hamano's avatar
Junio C Hamano committed
45 46 47 48 49 50 51
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];
52
		if (!index_fd(sha1, fd, st, 0, OBJ_BLOB, ce->name))
53
			match = hashcmp(sha1, ce->sha1);
54
		/* index_fd() closed the file descriptor already */
Junio C Hamano's avatar
Junio C Hamano committed
55 56 57 58
	}
	return match;
}

59
static int ce_compare_link(struct cache_entry *ce, size_t expected_size)
Junio C Hamano's avatar
Junio C Hamano committed
60 61 62 63 64
{
	int match = -1;
	char *target;
	void *buffer;
	unsigned long size;
65
	enum object_type type;
Junio C Hamano's avatar
Junio C Hamano committed
66 67 68 69 70 71 72 73
	int len;

	target = xmalloc(expected_size);
	len = readlink(ce->name, target, expected_size);
	if (len != expected_size) {
		free(target);
		return -1;
	}
74
	buffer = read_sha1_file(ce->sha1, &type, &size);
Junio C Hamano's avatar
Junio C Hamano committed
75 76 77 78 79 80 81 82 83 84 85
	if (!buffer) {
		free(target);
		return -1;
	}
	if (size == expected_size)
		match = memcmp(buffer, target, size);
	free(buffer);
	free(target);
	return match;
}

86 87 88 89 90 91
static int ce_compare_gitlink(struct cache_entry *ce)
{
	unsigned char sha1[20];

	/*
	 * We don't actually require that the .git directory
Martin Waitz's avatar
Martin Waitz committed
92
	 * under GITLINK directory be a valid git directory. It
93 94 95 96 97 98 99 100 101 102
	 * 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
103 104 105 106 107 108 109 110
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:
111
		if (ce_compare_link(ce, xsize_t(st->st_size)))
Junio C Hamano's avatar
Junio C Hamano committed
112 113
			return DATA_CHANGED;
		break;
114
	case S_IFDIR:
115
		if (S_ISGITLINK(ce->ce_mode))
116
			return 0;
Junio C Hamano's avatar
Junio C Hamano committed
117 118 119 120 121 122
	default:
		return TYPE_CHANGED;
	}
	return 0;
}

Junio C Hamano's avatar
Junio C Hamano committed
123
static int ce_match_stat_basic(struct cache_entry *ce, struct stat *st)
124 125 126
{
	unsigned int changed = 0;

127 128 129 130
	if (ce->ce_flags & CE_REMOVE)
		return MODE_CHANGED | DATA_CHANGED | TYPE_CHANGED;

	switch (ce->ce_mode & S_IFMT) {
131 132
	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
		/* We consider only the owner x bit to be relevant for
		 * "mode changes"
		 */
		if (trust_executable_bit &&
137
		    (0100 & (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;
Martin Waitz's avatar
Martin Waitz committed
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
	default:
152
		die("internal error: ce_mode is %o", ce->ce_mode);
153
	}
154
	if (ce->ce_mtime != (unsigned int) st->st_mtime)
155
		changed |= MTIME_CHANGED;
156
	if (ce->ce_ctime != (unsigned int) st->st_ctime)
157
		changed |= CTIME_CHANGED;
158

159 160
	if (ce->ce_uid != (unsigned int) st->st_uid ||
	    ce->ce_gid != (unsigned int) st->st_gid)
161
		changed |= OWNER_CHANGED;
162
	if (ce->ce_ino != (unsigned int) st->st_ino)
163
		changed |= INODE_CHANGED;
164 165 166 167 168 169 170

#ifdef USE_STDEV
	/*
	 * st_dev breaks on network filesystems where different
	 * clients will have different views of what "device"
	 * the filesystem is on
	 */
171
	if (ce->ce_dev != (unsigned int) st->st_dev)
172 173 174
		changed |= INODE_CHANGED;
#endif

175
	if (ce->ce_size != (unsigned int) st->st_size)
176
		changed |= DATA_CHANGED;
177

Junio C Hamano's avatar
Junio C Hamano committed
178 179 180
	return changed;
}

181
int ie_match_stat(struct index_state *istate,
182 183
		  struct cache_entry *ce, struct stat *st,
		  unsigned int options)
Junio C Hamano's avatar
Junio C Hamano committed
184
{
Junio C Hamano's avatar
Junio C Hamano committed
185
	unsigned int changed;
186 187
	int ignore_valid = options & CE_MATCH_IGNORE_VALID;
	int assume_racy_is_modified = options & CE_MATCH_RACY_IS_DIRTY;
Junio C Hamano's avatar
Junio C Hamano committed
188 189 190 191 192

	/*
	 * If it's marked as always valid in the index, it's
	 * valid whatever the checked-out copy says.
	 */
193
	if (!ignore_valid && (ce->ce_flags & CE_VALID))
Junio C Hamano's avatar
Junio C Hamano committed
194 195 196
		return 0;

	changed = ce_match_stat_basic(ce, st);
Junio C Hamano's avatar
Junio C Hamano committed
197

Junio C Hamano's avatar
Junio C Hamano committed
198 199 200 201 202 203 204 205 206 207 208 209 210 211
	/*
	 * 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
212
	 * carefully than others.
Junio C Hamano's avatar
Junio C Hamano committed
213 214
	 */
	if (!changed &&
215
	    istate->timestamp &&
216
	    istate->timestamp <= ce->ce_mtime) {
217 218 219 220 221
		if (assume_racy_is_modified)
			changed |= DATA_CHANGED;
		else
			changed |= ce_modified_check_fs(ce, st);
	}
222

Junio C Hamano's avatar
Junio C Hamano committed
223
	return changed;
224 225
}

226
int ie_modified(struct index_state *istate,
227
		struct cache_entry *ce, struct stat *st, unsigned int options)
228
{
Junio C Hamano's avatar
Junio C Hamano committed
229
	int changed, changed_fs;
230 231

	changed = ie_match_stat(istate, ce, st, options);
232 233 234 235 236 237 238 239 240 241 242 243 244
	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.
	 */
245
	if ((changed & DATA_CHANGED) && ce->ce_size != 0)
246 247
		return changed;

Junio C Hamano's avatar
Junio C Hamano committed
248 249 250
	changed_fs = ce_modified_check_fs(ce, st);
	if (changed_fs)
		return changed | changed_fs;
251 252 253
	return 0;
}

254 255 256 257 258 259 260 261 262 263 264 265
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];
266
	if (!c1 && S_ISDIR(mode1))
267
		c1 = '/';
268
	if (!c2 && S_ISDIR(mode2))
269 270 271 272
		c2 = '/';
	return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0;
}

273
int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
274
{
275 276
	int len1 = flags1 & CE_NAMEMASK;
	int len2 = flags2 & CE_NAMEMASK;
277 278 279 280 281 282 283 284 285 286
	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
287

288 289 290
	/* Compare stages  */
	flags1 &= CE_STAGEMASK;
	flags2 &= CE_STAGEMASK;
Junio C Hamano's avatar
Junio C Hamano committed
291

292 293 294 295
	if (flags1 < flags2)
		return -1;
	if (flags1 > flags2)
		return 1;
296 297 298
	return 0;
}

299
int index_name_pos(struct index_state *istate, const char *name, int namelen)
300 301 302 303
{
	int first, last;

	first = 0;
304
	last = istate->cache_nr;
305 306
	while (last > first) {
		int next = (last + first) >> 1;
307
		struct cache_entry *ce = istate->cache[next];
308
		int cmp = cache_name_compare(name, namelen, ce->name, ce->ce_flags);
309
		if (!cmp)
310
			return next;
311 312 313 314 315 316
		if (cmp < 0) {
			last = next;
			continue;
		}
		first = next+1;
	}
317
	return -first-1;
318 319
}

320
/* Remove entry, return true if there are more entries to go.. */
321
int remove_index_entry_at(struct index_state *istate, int pos)
322
{
323 324 325
	istate->cache_changed = 1;
	istate->cache_nr--;
	if (pos >= istate->cache_nr)
326
		return 0;
327 328 329
	memmove(istate->cache + pos,
		istate->cache + pos + 1,
		(istate->cache_nr - pos) * sizeof(struct cache_entry *));
330 331 332
	return 1;
}

333
int remove_file_from_index(struct index_state *istate, const char *path)
334
{
335
	int pos = index_name_pos(istate, path, strlen(path));
336 337
	if (pos < 0)
		pos = -pos-1;
Junio C Hamano's avatar
Junio C Hamano committed
338
	cache_tree_invalidate_path(istate->cache_tree, path);
339 340
	while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path))
		remove_index_entry_at(istate, pos);
341 342 343
	return 0;
}

344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371
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;
}

372
int add_file_to_index(struct index_state *istate, const char *path, int verbose)
Johannes Schindelin's avatar
Johannes Schindelin committed
373
{
374
	int size, namelen, pos;
Johannes Schindelin's avatar
Johannes Schindelin committed
375 376
	struct stat st;
	struct cache_entry *ce;
377
	unsigned ce_option = CE_MATCH_IGNORE_VALID|CE_MATCH_RACY_IS_DIRTY;
Johannes Schindelin's avatar
Johannes Schindelin committed
378 379 380 381

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

382 383
	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
384 385

	namelen = strlen(path);
386 387 388 389
	if (S_ISDIR(st.st_mode)) {
		while (namelen && path[namelen-1] == '/')
			namelen--;
	}
Johannes Schindelin's avatar
Johannes Schindelin committed
390 391 392
	size = cache_entry_size(namelen);
	ce = xcalloc(1, size);
	memcpy(ce->name, path, namelen);
393
	ce->ce_flags = namelen;
Johannes Schindelin's avatar
Johannes Schindelin committed
394 395
	fill_stat_cache_info(ce, &st);

396
	if (trust_executable_bit && has_symlinks)
397 398
		ce->ce_mode = create_ce_mode(st.st_mode);
	else {
399 400
		/* 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
401
		 */
402
		struct cache_entry *ent;
403
		int pos = index_name_pos_also_unmerged(istate, path, namelen);
404

405
		ent = (0 <= pos) ? istate->cache[pos] : NULL;
406
		ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
Johannes Schindelin's avatar
Johannes Schindelin committed
407 408
	}

409 410 411
	pos = index_name_pos(istate, ce->name, namelen);
	if (0 <= pos &&
	    !ce_stage(istate->cache[pos]) &&
412
	    !ie_match_stat(istate, istate->cache[pos], &st, ce_option)) {
413 414 415 416 417
		/* Nothing changed, really */
		free(ce);
		return 0;
	}

Johannes Schindelin's avatar
Johannes Schindelin committed
418 419
	if (index_path(ce->sha1, path, &st, 1))
		die("unable to index file %s", path);
420
	if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
Johannes Schindelin's avatar
Johannes Schindelin committed
421 422 423 424 425 426
		die("unable to add %s to index",path);
	if (verbose)
		printf("add '%s'\n", path);
	return 0;
}

427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451
struct cache_entry *make_cache_entry(unsigned int mode,
		const unsigned char *sha1, const char *path, int stage,
		int refresh)
{
	int size, len;
	struct cache_entry *ce;

	if (!verify_path(path))
		return NULL;

	len = strlen(path);
	size = cache_entry_size(len);
	ce = xcalloc(1, size);

	hashcpy(ce->sha1, sha1);
	memcpy(ce->name, path, len);
	ce->ce_flags = create_ce_flags(len, stage);
	ce->ce_mode = create_ce_mode(mode);

	if (refresh)
		return refresh_cache_entry(ce, 0);

	return ce;
}

452
int ce_same_name(struct cache_entry *a, struct cache_entry *b)
453 454 455 456 457
{
	int len = ce_namelen(a);
	return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
}

458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
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;
478 479
		if (!matchlen)
			return 1;
480 481 482 483
	}
	return 0;
}

484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547
/*
 * 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++;
	}
}

548 549 550
/*
 * Do we have another file that has the beginning components being a
 * proper superset of the name we're trying to add?
551
 */
552 553
static int has_file_name(struct index_state *istate,
			 const struct cache_entry *ce, int pos, int ok_to_replace)
554
{
555 556
	int retval = 0;
	int len = ce_namelen(ce);
557
	int stage = ce_stage(ce);
558
	const char *name = ce->name;
559

560 561
	while (pos < istate->cache_nr) {
		struct cache_entry *p = istate->cache[pos++];
562

563
		if (len >= ce_namelen(p))
564
			break;
565 566
		if (memcmp(name, p->name, len))
			break;
567 568
		if (ce_stage(p) != stage)
			continue;
569 570
		if (p->name[len] != '/')
			continue;
571
		if (p->ce_flags & CE_REMOVE)
572
			continue;
573 574 575
		retval = -1;
		if (!ok_to_replace)
			break;
576
		remove_index_entry_at(istate, --pos);
577
	}
578 579
	return retval;
}
580

581 582 583 584
/*
 * Do we have another file with a pathname that is a proper
 * subset of the name we're trying to add?
 */
585 586
static int has_dir_name(struct index_state *istate,
			const struct cache_entry *ce, int pos, int ok_to_replace)
587 588
{
	int retval = 0;
589
	int stage = ce_stage(ce);
590 591
	const char *name = ce->name;
	const char *slash = name + ce_namelen(ce);
592

593 594
	for (;;) {
		int len;
595

596 597 598 599 600 601 602
		for (;;) {
			if (*--slash == '/')
				break;
			if (slash <= ce->name)
				return retval;
		}
		len = slash - name;
603

604
		pos = index_name_pos(istate, name, create_ce_flags(len, stage));
605
		if (pos >= 0) {
606 607 608 609 610 611 612 613
			/*
			 * 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.
			 */
614
			if (stage || istate->cache[pos]->ce_mode) {
615 616 617
				retval = -1;
				if (!ok_to_replace)
					break;
618
				remove_index_entry_at(istate, pos);
619 620
				continue;
			}
621
		}
622 623
		else
			pos = -pos-1;
624 625 626 627

		/*
		 * Trivial optimization: if we find an entry that
		 * already matches the sub-directory, then we know
628
		 * we're ok, and we can exit.
629
		 */
630 631
		while (pos < istate->cache_nr) {
			struct cache_entry *p = istate->cache[pos];
632 633 634 635
			if ((ce_namelen(p) <= len) ||
			    (p->name[len] != '/') ||
			    memcmp(p->name, name, len))
				break; /* not our subdirectory */
636
			if (ce_stage(p) == stage && (stage || p->ce_mode))
637 638 639 640 641 642 643
				/* 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++;
644
		}
645
	}
646 647 648 649 650 651 652
	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.
Junio C Hamano's avatar
Junio C Hamano committed
653
 *
654 655 656 657
 * 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.
 */
658 659 660
static int check_file_directory_conflict(struct index_state *istate,
					 const struct cache_entry *ce,
					 int pos, int ok_to_replace)
661
{
662 663 664 665 666
	int retval;

	/*
	 * When ce is an "I am going away" entry, we allow it to be added
	 */
667
	if (ce->ce_flags & CE_REMOVE)
668 669
		return 0;

670 671 672
	/*
	 * We check if the path is a sub-path of a subsequent pathname
	 * first, since removing those will not change the position
673
	 * in the array.
674
	 */
675
	retval = has_file_name(istate, ce, pos, ok_to_replace);
676

677 678 679 680
	/*
	 * Then check if the path might have a clashing sub-directory
	 * before it.
	 */
681
	return retval + has_dir_name(istate, ce, pos, ok_to_replace);
682 683
}

684
static int add_index_entry_with_check(struct index_state *istate, struct cache_entry *ce, int option)
685 686
{
	int pos;
687 688
	int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
	int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
689
	int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
Junio C Hamano's avatar
Junio C Hamano committed
690

Junio C Hamano's avatar
Junio C Hamano committed
691
	cache_tree_invalidate_path(istate->cache_tree, ce->name);
692
	pos = index_name_pos(istate, ce->name, ce->ce_flags);
693

Junio C Hamano's avatar
Junio C Hamano committed
694
	/* existing match? Just replace it. */
695
	if (pos >= 0) {
696 697
		istate->cache_changed = 1;
		istate->cache[pos] = ce;
698 699
		return 0;
	}
700
	pos = -pos-1;
701

702 703 704 705
	/*
	 * Inserting a merged entry ("stage 0") into the index
	 * will always replace all non-merged entries..
	 */
706 707
	if (pos < istate->cache_nr && ce_stage(ce) == 0) {
		while (ce_same_name(istate->cache[pos], ce)) {
708
			ok_to_add = 1;
709
			if (!remove_index_entry_at(istate, pos))
710 711 712 713
				break;
		}
	}

714 715
	if (!ok_to_add)
		return -1;
716 717
	if (!verify_path(ce->name))
		return -1;
718

Junio C Hamano's avatar
Junio C Hamano committed
719
	if (!skip_df_check &&
720
	    check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
721
		if (!ok_to_replace)
722 723
			return error("'%s' appears as both a file and as a directory",
				     ce->name);
724
		pos = index_name_pos(istate, ce->name, ce->ce_flags);
725 726
		pos = -pos-1;
	}
727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742
	return pos + 1;
}

int add_index_entry(struct index_state *istate, struct cache_entry *ce, int option)
{
	int pos;

	if (option & ADD_CACHE_JUST_APPEND)
		pos = istate->cache_nr;
	else {
		int ret;
		ret = add_index_entry_with_check(istate, ce, option);
		if (ret <= 0)
			return ret;
		pos = ret - 1;
	}
743

744
	/* Make sure the array is big enough .. */
745 746 747 748
	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 *));
749 750 751
	}

	/* Add it in.. */
752
	istate->cache_nr++;
753
	if (istate->cache_nr > pos + 1)
754 755 756 757 758
		memmove(istate->cache + pos + 1,
			istate->cache + pos,
			(istate->cache_nr - pos - 1) * sizeof(ce));
	istate->cache[pos] = ce;
	istate->cache_changed = 1;
759 760 761
	return 0;
}

762 763 764 765 766 767 768 769 770 771 772
/*
 * "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.
 */
773
static struct cache_entry *refresh_cache_ent(struct index_state *istate,
774 775
					     struct cache_entry *ce,
					     unsigned int options, int *err)
776 777 778 779
{
	struct stat st;
	struct cache_entry *updated;
	int changed, size;
780
	int ignore_valid = options & CE_MATCH_IGNORE_VALID;
781

782
	if (lstat(ce->name, &st) < 0) {
783 784
		if (err)
			*err = errno;
785 786
		return NULL;
	}
787

788
	changed = ie_match_stat(istate, ce, &st, options);
789
	if (!changed) {
790 791 792 793 794 795 796 797
		/*
		 * The path is unchanged.  If we were told to ignore
		 * valid bit, then we did the actual stat check and
		 * found that the entry is unmodified.  If the entry
		 * is not marked VALID, this is the place to mark it
		 * valid again, under "assume unchanged" mode.
		 */
		if (ignore_valid && assume_unchanged &&
798
		    !(ce->ce_flags & CE_VALID))
799 800
			; /* mark this one VALID again */
		else
801
			return ce;
802 803
	}

804
	if (ie_modified(istate, ce, &st, options)) {
805 806
		if (err)
			*err = EINVAL;
807 808
		return NULL;
	}
809 810 811 812 813

	size = ce_size(ce);
	updated = xmalloc(size);
	memcpy(updated, ce, size);
	fill_stat_cache_info(updated, &st);
814 815 816 817 818
	/*
	 * If ignore_valid 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.
819
	 */
820
	if (!ignore_valid && assume_unchanged &&
821 822
	    !(ce->ce_flags & CE_VALID))
		updated->ce_flags &= ~CE_VALID;
823 824 825 826

	return updated;
}

827
int refresh_index(struct index_state *istate, unsigned int flags, const char **pathspec, char *seen)
828 829 830 831 832 833 834
{
	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;
835
	unsigned int options = really ? CE_MATCH_IGNORE_VALID : 0;
836

837
	for (i = 0; i < istate->cache_nr; i++) {
838
		struct cache_entry *ce, *new;
839 840
		int cache_errno = 0;

841
		ce = istate->cache[i];
842
		if (ce_stage(ce)) {
843 844
			while ((i < istate->cache_nr) &&
			       ! strcmp(istate->cache[i]->name, ce->name))
845 846 847 848 849 850 851 852 853
				i++;
			i--;
			if (allow_unmerged)
				continue;
			printf("%s: needs merge\n", ce->name);
			has_errors = 1;
			continue;
		}

854 855 856
		if (pathspec && !match_pathspec(pathspec, ce->name, strlen(ce->name), 0, seen))
			continue;

857
		new = refresh_cache_ent(istate, ce, options, &cache_errno);
858
		if (new == ce)
859
			continue;
860 861
		if (!new) {
			if (not_new && cache_errno == ENOENT)
862
				continue;
863
			if (really && cache_errno == EINVAL) {
864 865 866
				/* If we are doing --really-refresh that
				 * means the index is not valid anymore.
				 */
867
				ce->ce_flags &= ~CE_VALID;
868
				istate->cache_changed = 1;
869 870 871 872 873 874 875
			}
			if (quiet)
				continue;
			printf("%s: needs update\n", ce->name);
			has_errors = 1;
			continue;
		}
876 877
		istate->cache_changed = 1;
		/* You can NOT just free istate->cache[i] here, since it
878 879
		 * might not be necessarily malloc()ed but can also come
		 * from mmap(). */
880
		istate->cache[i] = new;
881 882 883 884
	}
	return has_errors;
}

885 886
struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
{
887
	return refresh_cache_ent(&the_index, ce, really, NULL);
888 889
}

890
static int verify_hdr(struct cache_header *hdr, unsigned long size)
891 892
{
	SHA_CTX c;
893
	unsigned char sha1[20];
894

895
	if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
896
		return error("bad signature");
897 898
	if (hdr->hdr_version != htonl(2))
		return error("bad index version");
899
	SHA1_Init(&c);
900
	SHA1_Update(&c, hdr, size - 20);
901
	SHA1_Final(sha1, &c);
902
	if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
903
		return error("bad index file sha1 signature");
904 905 906
	return 0;
}

907 908
static int read_index_extension(struct index_state *istate,
				const char *ext, void *data, unsigned long sz)
909 910 911
{
	switch (CACHE_EXT(ext)) {
	case CACHE_EXT_TREE:
912
		istate->cache_tree = cache_tree_read(data, sz);
913 914 915 916 917 918 919 920 921 922 923
		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;
}

924
int read_index(struct index_state *istate)
925
{
926
	return read_index_from(istate, get_index_file());
927 928
}

929 930
static void convert_from_disk(struct ondisk_cache_entry *ondisk, struct cache_entry *ce)
{
931 932
	size_t len;

933 934 935 936 937 938 939 940 941 942 943
	ce->ce_ctime = ntohl(ondisk->ctime.sec);
	ce->ce_mtime = ntohl(ondisk->mtime.sec);
	ce->ce_dev   = ntohl(ondisk->dev);
	ce->ce_ino   = ntohl(ondisk->ino);
	ce->ce_mode  = ntohl(ondisk->mode);
	ce->ce_uid   = ntohl(ondisk->uid);
	ce->ce_gid   = ntohl(ondisk->gid);
	ce->ce_size  = ntohl(ondisk->size);
	/* On-disk flags are just 16 bits */
	ce->ce_flags = ntohs(ondisk->flags);
	hashcpy(ce->sha1, ondisk->sha1);
944 945 946 947 948 949 950 951 952

	len = ce->ce_flags & CE_NAMEMASK;
	if (len == CE_NAMEMASK)
		len = strlen(ondisk->name);
	/*
	 * NEEDSWORK: If the original index is crafted, this copy could
	 * go unchecked.
	 */
	memcpy(ce->name, ondisk->name, len + 1);
953 954
}

955
/* remember to discard_cache() before reading a different cache! */
956
int read_index_from(struct index_state *istate, const char *path)
957 958 959
{
	int fd, i;
	struct stat st;
960
	unsigned long src_offset, dst_offset;
961
	struct cache_header *hdr;
962 963
	void *mmap;
	size_t mmap_size;
964 965

	errno = EBUSY;
966
	if (istate->alloc)
967
		return istate->cache_nr;
968

969
	errno = ENOENT;
970
	istate->timestamp = 0;
971
	fd = open(path, O_RDONLY);
972 973 974 975 976
	if (fd < 0) {
		if (errno == ENOENT)
			return 0;
		die("index file open failed (%s)", strerror(errno));
	}
977

978
	if (fstat(fd, &st))
979
		die("cannot stat the open index (%s)", strerror(errno));
980 981

	errno = EINVAL;
982 983
	mmap_size = xsize_t(st.st_size);
	if (mmap_size < sizeof(struct cache_header) + 20)
984 985
		die("index file smaller than expected");

986
	mmap = xmmap(NULL, mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
987
	close(fd);
988 989
	if (mmap == MAP_FAILED)
		die("unable to map index file");
990

991 992
	hdr = mmap;
	if (verify_hdr(hdr, mmap_size) < 0)
993 994
		goto unmap;

995 996 997
	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 *));
998

999 1000 1001 1002 1003 1004 1005 1006 1007 1008
	/*
	 * The disk format is actually larger than the in-memory format,
	 * due to space for nsec etc, so even though the in-memory one
	 * has room for a few  more flags, we can allocate using the same
	 * index size
	 */
	istate->alloc = xmalloc(mmap_size);

	src_offset = sizeof(*hdr);
	dst_offset = 0;
1009
	for (i = 0; i < istate->cache_nr; i++) {
1010
		struct ondisk_cache_entry *disk_ce;
1011 1012
		struct cache_entry *ce;

1013 1014 1015
		disk_ce = (struct ondisk_cache_entry *)((char *)mmap + src_offset);
		ce = (struct cache_entry *)((char *)istate->alloc + dst_offset);
		convert_from_disk(disk_ce, ce);
1016
		istate->cache[i] = ce;
1017 1018 1019

		src_offset += ondisk_ce_size(ce);
		dst_offset += ce_size(ce);
1020
	}
1021
	istate->timestamp = st.st_mtime;
1022
	while (src_offset <= mmap_size - 20 - 8) {
1023 1024 1025 1026 1027 1028 1029
		/* 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;
1030
		memcpy(&extsize, (char *)mmap + src_offset + 4, 4);
1031
		extsize = ntohl(extsize);
1032
		if (read_index_extension(istate,
1033 1034
					 (const char *) mmap + src_offset,
					 (char *) mmap + src_offset + 8,
1035
					 extsize) < 0)
1036
			goto unmap;
1037 1038
		src_offset += 8;
		src_offset += extsize;
1039
	}
1040
	munmap(mmap, mmap_size);
1041
	return istate->cache_nr;
1042 1043

unmap:
1044
	munmap(mmap, mmap_size);
1045
	errno = EINVAL;
1046
	die("index file corrupt");
1047 1048
}

1049
int discard_index(struct index_state *istate)
1050
{
1051 1052 1053 1054
	istate->cache_nr = 0;
	istate->cache_changed = 0;
	istate->timestamp = 0;
	cache_tree_free(&(istate->cache_tree));
1055 1056
	free(istate->alloc);
	istate->alloc = NULL;
1057 1058

	/* no need to throw away allocated active_cache */
1059
	return 0;
1060 1061
}

1062
#define WRITE_BUFFER_SIZE 8192
1063
static unsigned char write_buffer[WRITE_BUFFER_SIZE];
1064 1065
static unsigned long write_buffer_len;

1066 1067 1068 1069 1070
static int ce_write_flush(SHA_CTX *context, int fd)
{
	unsigned int buffered = write_buffer_len;
	if (buffered) {
		SHA1_Update(context, write_buffer, buffered);
1071
		if (write_in_full(fd, write_buffer, buffered) != buffered)
1072 1073 1074 1075 1076 1077
			return -1;
		write_buffer_len = 0;
	}
	return 0;
}

1078
static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
1079 1080 1081 1082 1083 1084 1085 1086 1087
{
	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) {
1088 1089
			write_buffer_len = buffered;
			if (ce_write_flush(context, fd))
1090 1091 1092 1093 1094
				return -1;
			buffered = 0;
		}
		write_buffer_len = buffered;
		len -= partial;
1095
		data = (char *) data + partial;
Junio C Hamano's avatar
Junio C Hamano committed
1096 1097
	}
	return 0;
1098 1099
}

1100
static int write_index_ext_header(SHA_CTX *context, int fd,
1101
				  unsigned int ext, unsigned int sz)
1102 1103 1104
{
	ext = htonl(ext);
	sz = htonl(sz);
David Rientjes's avatar
David Rientjes committed
1105 1106
	return ((ce_write(context, fd, &ext, 4) < 0) ||
		(ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
1107 1108 1109
}

static int ce_flush(SHA_CTX *context, int fd)
1110 1111
{
	unsigned int left = write_buffer_len;
1112

1113 1114
	if (left) {
		write_buffer_len = 0;
1115
		SHA1_Update(context, write_buffer, left);
1116
	}
1117

1118 1119
	/* Flush first if not enough space for SHA1 signature */
	if (left + 20 > WRITE_BUFFER_SIZE) {
1120
		if (write_in_full(fd, write_buffer, left) != left)
1121 1122 1123 1124
			return -1;
		left = 0;
	}

1125
	/* Append the SHA1 signature at the end */