read-cache.c 31.4 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 45

	if (S_ISREG(st->st_mode))
		ce_mark_uptodate(ce);
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
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
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);
}

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
	case S_IFDIR:
118
		if (S_ISGITLINK(ce->ce_mode))
119
			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;

130 131 132 133
	if (ce->ce_flags & CE_REMOVE)
		return MODE_CHANGED | DATA_CHANGED | TYPE_CHANGED;

	switch (ce->ce_mode & S_IFMT) {
134 135
	case S_IFREG:
		changed |= !S_ISREG(st->st_mode) ? TYPE_CHANGED : 0;
Junio C Hamano's avatar
Junio C Hamano committed
136 137 138 139
		/* We consider only the owner x bit to be relevant for
		 * "mode changes"
		 */
		if (trust_executable_bit &&
140
		    (0100 & (ce->ce_mode ^ st->st_mode)))
141
			changed |= MODE_CHANGED;
142 143
		break;
	case S_IFLNK:
144 145 146
		if (!S_ISLNK(st->st_mode) &&
		    (has_symlinks || !S_ISREG(st->st_mode)))
			changed |= TYPE_CHANGED;
147
		break;
Martin Waitz's avatar
Martin Waitz committed
148
	case S_IFGITLINK:
149 150 151 152
		if (!S_ISDIR(st->st_mode))
			changed |= TYPE_CHANGED;
		else if (ce_compare_gitlink(ce))
			changed |= DATA_CHANGED;
153
		return changed;
154
	default:
155
		die("internal error: ce_mode is %o", ce->ce_mode);
156
	}
157
	if (ce->ce_mtime != (unsigned int) st->st_mtime)
158
		changed |= MTIME_CHANGED;
159
	if (ce->ce_ctime != (unsigned int) st->st_ctime)
160
		changed |= CTIME_CHANGED;
161

162 163
	if (ce->ce_uid != (unsigned int) st->st_uid ||
	    ce->ce_gid != (unsigned int) st->st_gid)
164
		changed |= OWNER_CHANGED;
165
	if (ce->ce_ino != (unsigned int) st->st_ino)
166
		changed |= INODE_CHANGED;
167 168 169 170 171 172 173

#ifdef USE_STDEV
	/*
	 * st_dev breaks on network filesystems where different
	 * clients will have different views of what "device"
	 * the filesystem is on
	 */
174
	if (ce->ce_dev != (unsigned int) st->st_dev)
175 176 177
		changed |= INODE_CHANGED;
#endif

178
	if (ce->ce_size != (unsigned int) st->st_size)
179
		changed |= DATA_CHANGED;
180

Junio C Hamano's avatar
Junio C Hamano committed
181 182 183
	return changed;
}

184
int ie_match_stat(struct index_state *istate,
185 186
		  struct cache_entry *ce, struct stat *st,
		  unsigned int options)
Junio C Hamano's avatar
Junio C Hamano committed
187
{
Junio C Hamano's avatar
Junio C Hamano committed
188
	unsigned int changed;
189 190
	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
191 192 193 194 195

	/*
	 * If it's marked as always valid in the index, it's
	 * valid whatever the checked-out copy says.
	 */
196
	if (!ignore_valid && (ce->ce_flags & CE_VALID))
Junio C Hamano's avatar
Junio C Hamano committed
197 198 199
		return 0;

	changed = ce_match_stat_basic(ce, st);
Junio C Hamano's avatar
Junio C Hamano committed
200

Junio C Hamano's avatar
Junio C Hamano committed
201 202 203 204 205 206 207 208 209 210 211 212 213 214
	/*
	 * 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
215
	 * carefully than others.
Junio C Hamano's avatar
Junio C Hamano committed
216 217
	 */
	if (!changed &&
218
	    istate->timestamp &&
219
	    istate->timestamp <= ce->ce_mtime) {
220 221 222 223 224
		if (assume_racy_is_modified)
			changed |= DATA_CHANGED;
		else
			changed |= ce_modified_check_fs(ce, st);
	}
225

Junio C Hamano's avatar
Junio C Hamano committed
226
	return changed;
227 228
}

229
int ie_modified(struct index_state *istate,
230
		struct cache_entry *ce, struct stat *st, unsigned int options)
231
{
Junio C Hamano's avatar
Junio C Hamano committed
232
	int changed, changed_fs;
233 234

	changed = ie_match_stat(istate, ce, st, options);
235 236 237 238 239 240 241 242 243 244 245 246 247
	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.
	 */
248
	if ((changed & DATA_CHANGED) && ce->ce_size != 0)
249 250
		return changed;

Junio C Hamano's avatar
Junio C Hamano committed
251 252 253
	changed_fs = ce_modified_check_fs(ce, st);
	if (changed_fs)
		return changed | changed_fs;
254 255 256
	return 0;
}

257 258 259 260 261 262 263 264 265 266 267 268
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];
269
	if (!c1 && S_ISDIR(mode1))
270
		c1 = '/';
271
	if (!c2 && S_ISDIR(mode2))
272 273 274 275
		c2 = '/';
	return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0;
}

276
int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
277
{
278 279
	int len1 = flags1 & CE_NAMEMASK;
	int len2 = flags2 & CE_NAMEMASK;
280 281 282 283 284 285 286 287 288 289
	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
290

291 292 293
	/* Compare stages  */
	flags1 &= CE_STAGEMASK;
	flags2 &= CE_STAGEMASK;
Junio C Hamano's avatar
Junio C Hamano committed
294

295 296 297 298
	if (flags1 < flags2)
		return -1;
	if (flags1 > flags2)
		return 1;
299 300 301
	return 0;
}

302
int index_name_pos(struct index_state *istate, const char *name, int namelen)
303 304 305 306
{
	int first, last;

	first = 0;
307
	last = istate->cache_nr;
308 309
	while (last > first) {
		int next = (last + first) >> 1;
310
		struct cache_entry *ce = istate->cache[next];
311
		int cmp = cache_name_compare(name, namelen, ce->name, ce->ce_flags);
312
		if (!cmp)
313
			return next;
314 315 316 317 318 319
		if (cmp < 0) {
			last = next;
			continue;
		}
		first = next+1;
	}
320
	return -first-1;
321 322
}

323
/* Remove entry, return true if there are more entries to go.. */
324
int remove_index_entry_at(struct index_state *istate, int pos)
325
{
326 327 328
	istate->cache_changed = 1;
	istate->cache_nr--;
	if (pos >= istate->cache_nr)
329
		return 0;
330 331 332
	memmove(istate->cache + pos,
		istate->cache + pos + 1,
		(istate->cache_nr - pos) * sizeof(struct cache_entry *));
333 334 335
	return 1;
}

336
int remove_file_from_index(struct index_state *istate, const char *path)
337
{
338
	int pos = index_name_pos(istate, path, strlen(path));
339 340
	if (pos < 0)
		pos = -pos-1;
Junio C Hamano's avatar
Junio C Hamano committed
341
	cache_tree_invalidate_path(istate->cache_tree, path);
342 343
	while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path))
		remove_index_entry_at(istate, pos);
344 345 346
	return 0;
}

347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374
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;
}

375
int add_file_to_index(struct index_state *istate, const char *path, int verbose)
Johannes Schindelin's avatar
Johannes Schindelin committed
376
{
377
	int size, namelen, pos;
Johannes Schindelin's avatar
Johannes Schindelin committed
378 379
	struct stat st;
	struct cache_entry *ce;
380
	unsigned ce_option = CE_MATCH_IGNORE_VALID|CE_MATCH_RACY_IS_DIRTY;
Johannes Schindelin's avatar
Johannes Schindelin committed
381 382 383 384

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

385 386
	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
387 388

	namelen = strlen(path);
389 390 391 392
	if (S_ISDIR(st.st_mode)) {
		while (namelen && path[namelen-1] == '/')
			namelen--;
	}
Johannes Schindelin's avatar
Johannes Schindelin committed
393 394 395
	size = cache_entry_size(namelen);
	ce = xcalloc(1, size);
	memcpy(ce->name, path, namelen);
396
	ce->ce_flags = namelen;
Johannes Schindelin's avatar
Johannes Schindelin committed
397 398
	fill_stat_cache_info(ce, &st);

399
	if (trust_executable_bit && has_symlinks)
400 401
		ce->ce_mode = create_ce_mode(st.st_mode);
	else {
402 403
		/* 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
404
		 */
405
		struct cache_entry *ent;
406
		int pos = index_name_pos_also_unmerged(istate, path, namelen);
407

408
		ent = (0 <= pos) ? istate->cache[pos] : NULL;
409
		ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
Johannes Schindelin's avatar
Johannes Schindelin committed
410 411
	}

412 413 414
	pos = index_name_pos(istate, ce->name, namelen);
	if (0 <= pos &&
	    !ce_stage(istate->cache[pos]) &&
415
	    !ie_match_stat(istate, istate->cache[pos], &st, ce_option)) {
416 417
		/* Nothing changed, really */
		free(ce);
418
		ce_mark_uptodate(istate->cache[pos]);
419 420 421
		return 0;
	}

Johannes Schindelin's avatar
Johannes Schindelin committed
422 423
	if (index_path(ce->sha1, path, &st, 1))
		die("unable to index file %s", path);
424
	if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
Johannes Schindelin's avatar
Johannes Schindelin committed
425 426 427 428 429 430
		die("unable to add %s to index",path);
	if (verbose)
		printf("add '%s'\n", path);
	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
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;
}

456
int ce_same_name(struct cache_entry *a, struct cache_entry *b)
457 458 459 460 461
{
	int len = ce_namelen(a);
	return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
}

462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481
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;
482 483
		if (!matchlen)
			return 1;
484 485 486 487
	}
	return 0;
}

488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551
/*
 * 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++;
	}
}

552 553 554
/*
 * Do we have another file that has the beginning components being a
 * proper superset of the name we're trying to add?
555
 */
556 557
static int has_file_name(struct index_state *istate,
			 const struct cache_entry *ce, int pos, int ok_to_replace)
558
{
559 560
	int retval = 0;
	int len = ce_namelen(ce);
561
	int stage = ce_stage(ce);
562
	const char *name = ce->name;
563

564 565
	while (pos < istate->cache_nr) {
		struct cache_entry *p = istate->cache[pos++];
566

567
		if (len >= ce_namelen(p))
568
			break;
569 570
		if (memcmp(name, p->name, len))
			break;
571 572
		if (ce_stage(p) != stage)
			continue;
573 574
		if (p->name[len] != '/')
			continue;
575
		if (p->ce_flags & CE_REMOVE)
576
			continue;
577 578 579
		retval = -1;
		if (!ok_to_replace)
			break;
580
		remove_index_entry_at(istate, --pos);
581
	}
582 583
	return retval;
}
584

585 586 587 588
/*
 * Do we have another file with a pathname that is a proper
 * subset of the name we're trying to add?
 */
589 590
static int has_dir_name(struct index_state *istate,
			const struct cache_entry *ce, int pos, int ok_to_replace)
591 592
{
	int retval = 0;
593
	int stage = ce_stage(ce);
594 595
	const char *name = ce->name;
	const char *slash = name + ce_namelen(ce);
596

597 598
	for (;;) {
		int len;
599

600 601 602 603 604 605 606
		for (;;) {
			if (*--slash == '/')
				break;
			if (slash <= ce->name)
				return retval;
		}
		len = slash - name;
607

608
		pos = index_name_pos(istate, name, create_ce_flags(len, stage));
609
		if (pos >= 0) {
610 611 612 613 614 615 616 617
			/*
			 * 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.
			 */
618
			if (!(istate->cache[pos]->ce_flags & CE_REMOVE)) {
619 620 621
				retval = -1;
				if (!ok_to_replace)
					break;
622
				remove_index_entry_at(istate, pos);
623 624
				continue;
			}
625
		}
626 627
		else
			pos = -pos-1;
628 629 630 631

		/*
		 * Trivial optimization: if we find an entry that
		 * already matches the sub-directory, then we know
632
		 * we're ok, and we can exit.
633
		 */
634 635
		while (pos < istate->cache_nr) {
			struct cache_entry *p = istate->cache[pos];
636 637 638 639
			if ((ce_namelen(p) <= len) ||
			    (p->name[len] != '/') ||
			    memcmp(p->name, name, len))
				break; /* not our subdirectory */
640 641 642
			if (ce_stage(p) == stage && !(p->ce_flags & CE_REMOVE))
				/*
				 * p is at the same stage as our entry, and
643 644 645 646 647 648
				 * is a subdirectory of what we are looking
				 * at, so we cannot have conflicts at our
				 * level or anything shorter.
				 */
				return retval;
			pos++;
649
		}
650
	}
651 652 653 654 655 656 657
	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
658
 *
659 660 661 662
 * 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.
 */
663 664 665
static int check_file_directory_conflict(struct index_state *istate,
					 const struct cache_entry *ce,
					 int pos, int ok_to_replace)
666
{
667 668 669 670 671
	int retval;

	/*
	 * When ce is an "I am going away" entry, we allow it to be added
	 */
672
	if (ce->ce_flags & CE_REMOVE)
673 674
		return 0;

675 676 677
	/*
	 * We check if the path is a sub-path of a subsequent pathname
	 * first, since removing those will not change the position
678
	 * in the array.
679
	 */
680
	retval = has_file_name(istate, ce, pos, ok_to_replace);
681

682 683 684 685
	/*
	 * Then check if the path might have a clashing sub-directory
	 * before it.
	 */
686
	return retval + has_dir_name(istate, ce, pos, ok_to_replace);
687 688
}

689
static int add_index_entry_with_check(struct index_state *istate, struct cache_entry *ce, int option)
690 691
{
	int pos;
692 693
	int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
	int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
694
	int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
Junio C Hamano's avatar
Junio C Hamano committed
695

Junio C Hamano's avatar
Junio C Hamano committed
696
	cache_tree_invalidate_path(istate->cache_tree, ce->name);
697
	pos = index_name_pos(istate, ce->name, ce->ce_flags);
698

Junio C Hamano's avatar
Junio C Hamano committed
699
	/* existing match? Just replace it. */
700
	if (pos >= 0) {
701 702
		istate->cache_changed = 1;
		istate->cache[pos] = ce;
703 704
		return 0;
	}
705
	pos = -pos-1;
706

707 708 709 710
	/*
	 * Inserting a merged entry ("stage 0") into the index
	 * will always replace all non-merged entries..
	 */
711 712
	if (pos < istate->cache_nr && ce_stage(ce) == 0) {
		while (ce_same_name(istate->cache[pos], ce)) {
713
			ok_to_add = 1;
714
			if (!remove_index_entry_at(istate, pos))
715 716 717 718
				break;
		}
	}

719 720
	if (!ok_to_add)
		return -1;
721 722
	if (!verify_path(ce->name))
		return -1;
723

Junio C Hamano's avatar
Junio C Hamano committed
724
	if (!skip_df_check &&
725
	    check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
726
		if (!ok_to_replace)
727 728
			return error("'%s' appears as both a file and as a directory",
				     ce->name);
729
		pos = index_name_pos(istate, ce->name, ce->ce_flags);
730 731
		pos = -pos-1;
	}
732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747
	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;
	}
748

749
	/* Make sure the array is big enough .. */
750 751 752 753
	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 *));
754 755 756
	}

	/* Add it in.. */
757
	istate->cache_nr++;
758
	if (istate->cache_nr > pos + 1)
759 760 761 762 763
		memmove(istate->cache + pos + 1,
			istate->cache + pos,
			(istate->cache_nr - pos - 1) * sizeof(ce));
	istate->cache[pos] = ce;
	istate->cache_changed = 1;
764 765 766
	return 0;
}

767 768 769 770 771 772 773 774 775 776 777
/*
 * "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.
 */
778
static struct cache_entry *refresh_cache_ent(struct index_state *istate,
779 780
					     struct cache_entry *ce,
					     unsigned int options, int *err)
781 782 783 784
{
	struct stat st;
	struct cache_entry *updated;
	int changed, size;
785
	int ignore_valid = options & CE_MATCH_IGNORE_VALID;
786

787 788 789
	if (ce_uptodate(ce))
		return ce;

790
	if (lstat(ce->name, &st) < 0) {
791 792
		if (err)
			*err = errno;
793 794
		return NULL;
	}
795

796
	changed = ie_match_stat(istate, ce, &st, options);
797
	if (!changed) {
798 799 800 801 802 803 804 805
		/*
		 * 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 &&
806
		    !(ce->ce_flags & CE_VALID))
807
			; /* mark this one VALID again */
808 809 810 811 812 813 814
		else {
			/*
			 * We do not mark the index itself "modified"
			 * because CE_UPTODATE flag is in-core only;
			 * we are not going to write this change out.
			 */
			ce_mark_uptodate(ce);
815
			return ce;
816
		}
817 818
	}

819
	if (ie_modified(istate, ce, &st, options)) {
820 821
		if (err)
			*err = EINVAL;
822 823
		return NULL;
	}
824 825 826 827 828

	size = ce_size(ce);
	updated = xmalloc(size);
	memcpy(updated, ce, size);
	fill_stat_cache_info(updated, &st);
829 830 831 832 833
	/*
	 * 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.
834
	 */
835
	if (!ignore_valid && assume_unchanged &&
836 837
	    !(ce->ce_flags & CE_VALID))
		updated->ce_flags &= ~CE_VALID;
838 839 840 841

	return updated;
}

842
int refresh_index(struct index_state *istate, unsigned int flags, const char **pathspec, char *seen)
843 844 845 846 847 848 849
{
	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;
850
	unsigned int options = really ? CE_MATCH_IGNORE_VALID : 0;
851

852
	for (i = 0; i < istate->cache_nr; i++) {
853
		struct cache_entry *ce, *new;
854 855
		int cache_errno = 0;

856
		ce = istate->cache[i];
857
		if (ce_stage(ce)) {
858 859
			while ((i < istate->cache_nr) &&
			       ! strcmp(istate->cache[i]->name, ce->name))
860 861 862 863 864 865 866 867 868
				i++;
			i--;
			if (allow_unmerged)
				continue;
			printf("%s: needs merge\n", ce->name);
			has_errors = 1;
			continue;
		}

869 870 871
		if (pathspec && !match_pathspec(pathspec, ce->name, strlen(ce->name), 0, seen))
			continue;

872
		new = refresh_cache_ent(istate, ce, options, &cache_errno);
873
		if (new == ce)
874
			continue;
875 876
		if (!new) {
			if (not_new && cache_errno == ENOENT)
877
				continue;
878
			if (really && cache_errno == EINVAL) {
879 880 881
				/* If we are doing --really-refresh that
				 * means the index is not valid anymore.
				 */
882
				ce->ce_flags &= ~CE_VALID;
883
				istate->cache_changed = 1;
884 885 886 887 888 889 890
			}
			if (quiet)
				continue;
			printf("%s: needs update\n", ce->name);
			has_errors = 1;
			continue;
		}
891 892
		istate->cache_changed = 1;
		/* You can NOT just free istate->cache[i] here, since it
893 894
		 * might not be necessarily malloc()ed but can also come
		 * from mmap(). */
895
		istate->cache[i] = new;
896 897 898 899
	}
	return has_errors;
}

900 901
struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
{
902
	return refresh_cache_ent(&the_index, ce, really, NULL);
903 904
}

905
static int verify_hdr(struct cache_header *hdr, unsigned long size)
906 907
{
	SHA_CTX c;
908
	unsigned char sha1[20];
909

910
	if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
911
		return error("bad signature");
912 913
	if (hdr->hdr_version != htonl(2))
		return error("bad index version");
914
	SHA1_Init(&c);
915
	SHA1_Update(&c, hdr, size - 20);
916
	SHA1_Final(sha1, &c);
917
	if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
918
		return error("bad index file sha1 signature");
919 920 921
	return 0;
}

922 923
static int read_index_extension(struct index_state *istate,
				const char *ext, void *data, unsigned long sz)
924 925 926
{
	switch (CACHE_EXT(ext)) {
	case CACHE_EXT_TREE:
927
		istate->cache_tree = cache_tree_read(data, sz);
928 929 930 931 932 933 934 935 936 937 938
		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;
}

939
int read_index(struct index_state *istate)
940
{
941
	return read_index_from(istate, get_index_file());
942 943
}

944 945
static void convert_from_disk(struct ondisk_cache_entry *ondisk, struct cache_entry *ce)
{
946 947
	size_t len;

948 949 950 951 952 953 954 955 956 957 958
	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);
959 960 961 962 963 964 965 966 967

	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);
968 969
}

970
/* remember to discard_cache() before reading a different cache! */
971
int read_index_from(struct index_state *istate, const char *path)
972 973 974
{
	int fd, i;
	struct stat st;
975
	unsigned long src_offset, dst_offset;
976
	struct cache_header *hdr;
977 978
	void *mmap;
	size_t mmap_size;
979 980

	errno = EBUSY;
981
	if (istate->alloc)
982
		return istate->cache_nr;
983

984
	errno = ENOENT;
985
	istate->timestamp = 0;
986
	fd = open(path, O_RDONLY);
987 988 989 990 991
	if (fd < 0) {
		if (errno == ENOENT)
			return 0;
		die("index file open failed (%s)", strerror(errno));
	}
992

993
	if (fstat(fd, &st))
994
		die("cannot stat the open index (%s)", strerror(errno));
995 996

	errno = EINVAL;
997 998
	mmap_size = xsize_t(st.st_size);
	if (mmap_size < sizeof(struct cache_header) + 20)
999 1000
		die("index file smaller than expected");

1001
	mmap = xmmap(NULL, mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
1002
	close(fd);
1003 1004
	if (mmap == MAP_FAILED)
		die("unable to map index file");
1005

1006 1007
	hdr = mmap;
	if (verify_hdr(hdr, mmap_size) < 0)
1008 1009
		goto unmap;

1010 1011 1012
	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 *));
1013

1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
	/*
	 * 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;
1024
	for (i = 0; i < istate->cache_nr; i++) {
1025
		struct ondisk_cache_entry *disk_ce;
1026 1027
		struct cache_entry *ce;

1028 1029 1030
		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);
1031
		istate->cache[i] = ce;
1032 1033 1034

		src_offset += ondisk_ce_size(ce);
		dst_offset += ce_size(ce);
1035
	}
1036
	istate->timestamp = st.st_mtime;
1037
	while (src_offset <= mmap_size - 20 - 8) {
1038 1039 1040 1041 1042 1043 1044
		/* 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;
1045
		memcpy(&extsize, (char *)mmap + src_offset + 4, 4);
1046
		extsize = ntohl(extsize);
1047
		if (read_index_extension(istate,
1048 1049
					 (const char *) mmap + src_offset,
					 (char *) mmap + src_offset + 8,
1050
					 extsize) < 0)
1051
			goto unmap;
1052 1053
		src_offset += 8;
		src_offset += extsize;
1054
	}
1055
	munmap(mmap, mmap_size);
1056
	return istate->cache_nr;
1057 1058

unmap:
1059
	munmap(mmap, mmap_size);
1060
	errno = EINVAL;
1061
	die("index file corrupt");
1062 1063
}

1064
int discard_index(struct index_state *istate)
1065
{
1066 1067 1068 1069
	istate->cache_nr = 0;
	istate->cache_changed = 0;
	istate->timestamp = 0;
	cache_tree_free(&(istate->cache_tree));
1070 1071
	free(istate->alloc);
	istate->alloc = NULL;
1072 1073

	/* no need to throw away allocated active_cache */
1074
	return 0;
1075 1076
}

1077
#define WRITE_BUFFER_SIZE 8192
1078
static unsigned char write_buffer[WRITE_BUFFER_SIZE];
1079 1080
static unsigned long write_buffer_len;

1081 1082 1083 1084 1085
static int ce_write_flush(SHA_CTX *context, int fd)
{
	unsigned int buffered = write_buffer_len;
	if (buffered) {
		SHA1_Update(context, write_buffer, buffered);
1086
		if (write_in_full(fd, write_buffer, buffered) != buffered)
1087 1088 1089 1090 1091 1092
			return -1;
		write_buffer_len = 0;
	}
	return 0;
}

1093
static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
1094 1095 1096 1097 1098 1099 1100 1101 1102
{
	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) {
1103 1104
			write_buffer_len = buffered;
			if (ce_write_flush(context, fd))
1105 1106 1107 1108 1109
				return -1;
			buffered = 0;
		}
		write_buffer_len = buffered;
		len -= partial;
1110
		data = (char *) data + partial;
Junio C Hamano's avatar
Junio C Hamano committed
1111 1112
	}
	return 0;
1113 1114
}

1115
static int write_index_ext_header(SHA_CTX *context, int fd,
1116
				  unsigned int ext, unsigned int sz)
1117 1118 1119
{
	ext = htonl(ext);
	sz = htonl(sz);
David Rientjes's avatar
David Rientjes committed
1120 1121
	return ((ce_write(context, fd, &ext, 4) < 0) ||
		(ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
1122 1123 1124
}

static int ce_flush(SHA_CTX *context, int fd)
1125 1126
{
	unsigned int left = write_buffer_len;
1127

1128 1129
	if (left) {
		write_buffer_len = 0;
1130
		SHA1_Update(context, write_buffer, left);<