read-cache.c 45.2 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
#include "dir.h"
11 12
#include "tree.h"
#include "commit.h"
J
Junio C Hamano 已提交
13
#include "blob.h"
14
#include "resolve-undo.h"
15

16 17
static struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really);

18 19 20 21 22 23 24 25 26 27 28
/* 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" */
29
#define CACHE_EXT_RESOLVE_UNDO 0x52455543 /* "REUC" */
30

31
struct index_state the_index;
32

J
Junio C Hamano 已提交
33 34 35
static void set_index_entry(struct index_state *istate, int nr, struct cache_entry *ce)
{
	istate->cache[nr] = ce;
36
	add_name_hash(istate, ce);
J
Junio C Hamano 已提交
37 38
}

39 40 41 42
static void replace_index_entry(struct index_state *istate, int nr, struct cache_entry *ce)
{
	struct cache_entry *old = istate->cache[nr];

43
	remove_name_hash(old);
L
Linus Torvalds 已提交
44
	set_index_entry(istate, nr, ce);
45 46 47
	istate->cache_changed = 1;
}

48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63
void rename_index_entry_at(struct index_state *istate, int nr, const char *new_name)
{
	struct cache_entry *old = istate->cache[nr], *new;
	int namelen = strlen(new_name);

	new = xmalloc(cache_entry_size(namelen));
	copy_cache_entry(new, old);
	new->ce_flags &= ~(CE_STATE_MASK | CE_NAMEMASK);
	new->ce_flags |= (namelen >= CE_NAMEMASK ? CE_NAMEMASK : namelen);
	memcpy(new->name, new_name, namelen + 1);

	cache_tree_invalidate_path(istate->cache_tree, old->name);
	remove_index_entry_at(istate, nr);
	add_index_entry(istate, new, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE);
}

64 65 66 67 68 69 70
/*
 * 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)
{
K
Kjetil Barvik 已提交
71 72
	ce->ce_ctime.sec = (unsigned int)st->st_ctime;
	ce->ce_mtime.sec = (unsigned int)st->st_mtime;
73 74
	ce->ce_ctime.nsec = ST_CTIME_NSEC(*st);
	ce->ce_mtime.nsec = ST_MTIME_NSEC(*st);
75 76 77 78 79
	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;
J
Junio C Hamano 已提交
80 81

	if (assume_unchanged)
82
		ce->ce_flags |= CE_VALID;
83 84 85

	if (S_ISREG(st->st_mode))
		ce_mark_uptodate(ce);
86 87
}

J
Junio C Hamano 已提交
88 89 90 91 92 93 94
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];
95
		if (!index_fd(sha1, fd, st, 0, OBJ_BLOB, ce->name, 0))
96
			match = hashcmp(sha1, ce->sha1);
97
		/* index_fd() closed the file descriptor already */
J
Junio C Hamano 已提交
98 99 100 101
	}
	return match;
}

102
static int ce_compare_link(struct cache_entry *ce, size_t expected_size)
J
Junio C Hamano 已提交
103 104 105 106
{
	int match = -1;
	void *buffer;
	unsigned long size;
107
	enum object_type type;
108
	struct strbuf sb = STRBUF_INIT;
J
Junio C Hamano 已提交
109

110
	if (strbuf_readlink(&sb, ce->name, expected_size))
J
Junio C Hamano 已提交
111
		return -1;
112

113
	buffer = read_sha1_file(ce->sha1, &type, &size);
114 115 116 117
	if (buffer) {
		if (size == sb.len)
			match = memcmp(buffer, sb.buf, size);
		free(buffer);
J
Junio C Hamano 已提交
118
	}
119
	strbuf_release(&sb);
J
Junio C Hamano 已提交
120 121 122
	return match;
}

123 124 125 126 127 128
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
129
	 * under GITLINK directory be a valid git directory. It
130 131 132 133 134 135 136 137 138 139
	 * 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
140 141 142 143 144 145 146 147
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:
148
		if (ce_compare_link(ce, xsize_t(st->st_size)))
J
Junio C Hamano 已提交
149 150
			return DATA_CHANGED;
		break;
151
	case S_IFDIR:
152
		if (S_ISGITLINK(ce->ce_mode))
153
			return ce_compare_gitlink(ce) ? DATA_CHANGED : 0;
J
Junio C Hamano 已提交
154 155 156 157 158 159
	default:
		return TYPE_CHANGED;
	}
	return 0;
}

160
static int is_empty_blob_sha1(const unsigned char *sha1)
161 162 163 164 165 166 167 168 169
{
	static const unsigned char empty_blob_sha1[20] = {
		0xe6,0x9d,0xe2,0x9b,0xb2,0xd1,0xd6,0x43,0x4b,0x8b,
		0x29,0xae,0x77,0x5a,0xd8,0xc2,0xe4,0x8c,0x53,0x91
	};

	return !hashcmp(sha1, empty_blob_sha1);
}

J
Junio C Hamano 已提交
170
static int ce_match_stat_basic(struct cache_entry *ce, struct stat *st)
171 172 173
{
	unsigned int changed = 0;

174 175 176 177
	if (ce->ce_flags & CE_REMOVE)
		return MODE_CHANGED | DATA_CHANGED | TYPE_CHANGED;

	switch (ce->ce_mode & S_IFMT) {
178 179
	case S_IFREG:
		changed |= !S_ISREG(st->st_mode) ? TYPE_CHANGED : 0;
J
Junio C Hamano 已提交
180 181 182 183
		/* We consider only the owner x bit to be relevant for
		 * "mode changes"
		 */
		if (trust_executable_bit &&
184
		    (0100 & (ce->ce_mode ^ st->st_mode)))
185
			changed |= MODE_CHANGED;
186 187
		break;
	case S_IFLNK:
188 189 190
		if (!S_ISLNK(st->st_mode) &&
		    (has_symlinks || !S_ISREG(st->st_mode)))
			changed |= TYPE_CHANGED;
191
		break;
M
Martin Waitz 已提交
192
	case S_IFGITLINK:
193
		/* We ignore most of the st_xxx fields for gitlinks */
194 195 196 197
		if (!S_ISDIR(st->st_mode))
			changed |= TYPE_CHANGED;
		else if (ce_compare_gitlink(ce))
			changed |= DATA_CHANGED;
198
		return changed;
199
	default:
200
		die("internal error: ce_mode is %o", ce->ce_mode);
201
	}
K
Kjetil Barvik 已提交
202
	if (ce->ce_mtime.sec != (unsigned int)st->st_mtime)
203
		changed |= MTIME_CHANGED;
K
Kjetil Barvik 已提交
204
	if (trust_ctime && ce->ce_ctime.sec != (unsigned int)st->st_ctime)
205
		changed |= CTIME_CHANGED;
206

K
Kjetil Barvik 已提交
207
#ifdef USE_NSEC
208
	if (ce->ce_mtime.nsec != ST_MTIME_NSEC(*st))
K
Kjetil Barvik 已提交
209
		changed |= MTIME_CHANGED;
210
	if (trust_ctime && ce->ce_ctime.nsec != ST_CTIME_NSEC(*st))
K
Kjetil Barvik 已提交
211 212 213
		changed |= CTIME_CHANGED;
#endif

214 215
	if (ce->ce_uid != (unsigned int) st->st_uid ||
	    ce->ce_gid != (unsigned int) st->st_gid)
216
		changed |= OWNER_CHANGED;
217
	if (ce->ce_ino != (unsigned int) st->st_ino)
218
		changed |= INODE_CHANGED;
219 220 221 222 223 224 225

#ifdef USE_STDEV
	/*
	 * st_dev breaks on network filesystems where different
	 * clients will have different views of what "device"
	 * the filesystem is on
	 */
226
	if (ce->ce_dev != (unsigned int) st->st_dev)
227 228 229
		changed |= INODE_CHANGED;
#endif

230
	if (ce->ce_size != (unsigned int) st->st_size)
231
		changed |= DATA_CHANGED;
232

233 234 235 236 237 238
	/* Racily smudged entry? */
	if (!ce->ce_size) {
		if (!is_empty_blob_sha1(ce->sha1))
			changed |= DATA_CHANGED;
	}

J
Junio C Hamano 已提交
239 240 241
	return changed;
}

242
static int is_racy_timestamp(const struct index_state *istate, struct cache_entry *ce)
243
{
244
	return (!S_ISGITLINK(ce->ce_mode) &&
K
Kjetil Barvik 已提交
245 246 247 248 249 250 251 252 253 254
		istate->timestamp.sec &&
#ifdef USE_NSEC
		 /* nanosecond timestamped files can also be racy! */
		(istate->timestamp.sec < ce->ce_mtime.sec ||
		 (istate->timestamp.sec == ce->ce_mtime.sec &&
		  istate->timestamp.nsec <= ce->ce_mtime.nsec))
#else
		istate->timestamp.sec <= ce->ce_mtime.sec
#endif
		 );
255 256
}

257
int ie_match_stat(const struct index_state *istate,
258 259
		  struct cache_entry *ce, struct stat *st,
		  unsigned int options)
J
Junio C Hamano 已提交
260
{
J
Junio C Hamano 已提交
261
	unsigned int changed;
262
	int ignore_valid = options & CE_MATCH_IGNORE_VALID;
263
	int ignore_skip_worktree = options & CE_MATCH_IGNORE_SKIP_WORKTREE;
264
	int assume_racy_is_modified = options & CE_MATCH_RACY_IS_DIRTY;
J
Junio C Hamano 已提交
265 266 267 268

	/*
	 * If it's marked as always valid in the index, it's
	 * valid whatever the checked-out copy says.
269 270
	 *
	 * skip-worktree has the same effect with higher precedence
J
Junio C Hamano 已提交
271
	 */
272 273
	if (!ignore_skip_worktree && ce_skip_worktree(ce))
		return 0;
274
	if (!ignore_valid && (ce->ce_flags & CE_VALID))
J
Junio C Hamano 已提交
275 276
		return 0;

277 278 279 280 281 282 283 284
	/*
	 * Intent-to-add entries have not been added, so the index entry
	 * by definition never matches what is in the work tree until it
	 * actually gets added.
	 */
	if (ce->ce_flags & CE_INTENT_TO_ADD)
		return DATA_CHANGED | TYPE_CHANGED | MODE_CHANGED;

J
Junio C Hamano 已提交
285
	changed = ce_match_stat_basic(ce, st);
J
Junio C Hamano 已提交
286

J
Junio C Hamano 已提交
287 288 289 290 291 292 293 294 295 296 297 298 299 300
	/*
	 * 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
301
	 * carefully than others.
J
Junio C Hamano 已提交
302
	 */
303
	if (!changed && is_racy_timestamp(istate, ce)) {
304 305 306 307 308
		if (assume_racy_is_modified)
			changed |= DATA_CHANGED;
		else
			changed |= ce_modified_check_fs(ce, st);
	}
309

J
Junio C Hamano 已提交
310
	return changed;
311 312
}

313
int ie_modified(const struct index_state *istate,
314
		struct cache_entry *ce, struct stat *st, unsigned int options)
315
{
J
Junio C Hamano 已提交
316
	int changed, changed_fs;
317 318

	changed = ie_match_stat(istate, ce, st, options);
319 320 321 322 323 324 325 326 327
	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;

328 329 330 331 332 333 334 335 336 337 338 339 340
	/*
	 * Immediately after read-tree or update-index --cacheinfo,
	 * the length field is zero, as we have never even read the
	 * lstat(2) information once, and we cannot trust DATA_CHANGED
	 * returned by ie_match_stat() which in turn was returned by
	 * ce_match_stat_basic() to signal that the filesize of the
	 * blob changed.  We have to actually go to the filesystem to
	 * see if the contents match, and if so, should answer "unchanged".
	 *
	 * The logic does not apply to gitlinks, as ce_match_stat_basic()
	 * already has checked the actual HEAD from the filesystem in the
	 * subproject.  If ie_match_stat() already said it is different,
	 * then we know it is.
341
	 */
342 343
	if ((changed & DATA_CHANGED) &&
	    (S_ISGITLINK(ce->ce_mode) || ce->ce_size != 0))
344 345
		return changed;

J
Junio C Hamano 已提交
346 347 348
	changed_fs = ce_modified_check_fs(ce, st);
	if (changed_fs)
		return changed | changed_fs;
349 350 351
	return 0;
}

352 353 354 355 356 357 358 359 360 361 362 363
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];
364
	if (!c1 && S_ISDIR(mode1))
365
		c1 = '/';
366
	if (!c2 && S_ISDIR(mode2))
367 368 369 370
		c2 = '/';
	return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0;
}

371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
/*
 * df_name_compare() is identical to base_name_compare(), except it
 * compares conflicting directory/file entries as equal. Note that
 * while a directory name compares as equal to a regular file, they
 * then individually compare _differently_ to a filename that has
 * a dot after the basename (because '\0' < '.' < '/').
 *
 * This is used by routines that want to traverse the git namespace
 * but then handle conflicting entries together when possible.
 */
int df_name_compare(const char *name1, int len1, int mode1,
		    const char *name2, int len2, int mode2)
{
	int len = len1 < len2 ? len1 : len2, cmp;
	unsigned char c1, c2;

	cmp = memcmp(name1, name2, len);
	if (cmp)
		return cmp;
	/* Directories and files compare equal (same length, same name) */
	if (len1 == len2)
		return 0;
	c1 = name1[len];
	if (!c1 && S_ISDIR(mode1))
		c1 = '/';
	c2 = name2[len];
	if (!c2 && S_ISDIR(mode2))
		c2 = '/';
	if (c1 == '/' && !c2)
		return 0;
	if (c2 == '/' && !c1)
		return 0;
	return c1 - c2;
}

406
int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
407
{
408 409
	int len1 = flags1 & CE_NAMEMASK;
	int len2 = flags2 & CE_NAMEMASK;
410 411 412 413 414 415 416 417 418 419
	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
420

421 422 423
	/* Compare stages  */
	flags1 &= CE_STAGEMASK;
	flags2 &= CE_STAGEMASK;
J
Junio C Hamano 已提交
424

425 426 427 428
	if (flags1 < flags2)
		return -1;
	if (flags1 > flags2)
		return 1;
429 430 431
	return 0;
}

432
int index_name_pos(const struct index_state *istate, const char *name, int namelen)
433 434 435 436
{
	int first, last;

	first = 0;
437
	last = istate->cache_nr;
438 439
	while (last > first) {
		int next = (last + first) >> 1;
440
		struct cache_entry *ce = istate->cache[next];
441
		int cmp = cache_name_compare(name, namelen, ce->name, ce->ce_flags);
442
		if (!cmp)
443
			return next;
444 445 446 447 448 449
		if (cmp < 0) {
			last = next;
			continue;
		}
		first = next+1;
	}
450
	return -first-1;
451 452
}

453
/* Remove entry, return true if there are more entries to go.. */
454
int remove_index_entry_at(struct index_state *istate, int pos)
455
{
456 457
	struct cache_entry *ce = istate->cache[pos];

458
	record_resolve_undo(istate, ce);
459
	remove_name_hash(ce);
460 461 462
	istate->cache_changed = 1;
	istate->cache_nr--;
	if (pos >= istate->cache_nr)
463
		return 0;
464 465 466
	memmove(istate->cache + pos,
		istate->cache + pos + 1,
		(istate->cache_nr - pos) * sizeof(struct cache_entry *));
467 468 469
	return 1;
}

470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489
/*
 * Remove all cache ententries marked for removal, that is where
 * CE_REMOVE is set in ce_flags.  This is much more effective than
 * calling remove_index_entry_at() for each entry to be removed.
 */
void remove_marked_cache_entries(struct index_state *istate)
{
	struct cache_entry **ce_array = istate->cache;
	unsigned int i, j;

	for (i = j = 0; i < istate->cache_nr; i++) {
		if (ce_array[i]->ce_flags & CE_REMOVE)
			remove_name_hash(ce_array[i]);
		else
			ce_array[j++] = ce_array[i];
	}
	istate->cache_changed = 1;
	istate->cache_nr = j;
}

490
int remove_file_from_index(struct index_state *istate, const char *path)
491
{
492
	int pos = index_name_pos(istate, path, strlen(path));
493 494
	if (pos < 0)
		pos = -pos-1;
J
Junio C Hamano 已提交
495
	cache_tree_invalidate_path(istate->cache_tree, path);
496 497
	while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path))
		remove_index_entry_at(istate, pos);
498 499 500
	return 0;
}

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
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;
}

529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560
static int different_name(struct cache_entry *ce, struct cache_entry *alias)
{
	int len = ce_namelen(ce);
	return ce_namelen(alias) != len || memcmp(ce->name, alias->name, len);
}

/*
 * If we add a filename that aliases in the cache, we will use the
 * name that we already have - but we don't want to update the same
 * alias twice, because that implies that there were actually two
 * different files with aliasing names!
 *
 * So we use the CE_ADDED flag to verify that the alias was an old
 * one before we accept it as
 */
static struct cache_entry *create_alias_ce(struct cache_entry *ce, struct cache_entry *alias)
{
	int len;
	struct cache_entry *new;

	if (alias->ce_flags & CE_ADDED)
		die("Will not add file alias '%s' ('%s' already exists in index)", ce->name, alias->name);

	/* Ok, create the new entry using the name of the existing alias */
	len = ce_namelen(alias);
	new = xcalloc(1, cache_entry_size(len));
	memcpy(new->name, alias->name, len);
	copy_cache_entry(new, ce);
	free(ce);
	return new;
}

J
Junio C Hamano 已提交
561 562 563 564 565 566 567 568
static void record_intent_to_add(struct cache_entry *ce)
{
	unsigned char sha1[20];
	if (write_sha1_file("", 0, blob_type, sha1))
		die("cannot create an empty blob in the object database");
	hashcpy(ce->sha1, sha1);
}

569
int add_to_index(struct index_state *istate, const char *path, struct stat *st, int flags)
J
Johannes Schindelin 已提交
570
{
571
	int size, namelen, was_same;
572
	mode_t st_mode = st->st_mode;
573
	struct cache_entry *ce, *alias;
574
	unsigned ce_option = CE_MATCH_IGNORE_VALID|CE_MATCH_IGNORE_SKIP_WORKTREE|CE_MATCH_RACY_IS_DIRTY;
575 576
	int verbose = flags & (ADD_CACHE_VERBOSE | ADD_CACHE_PRETEND);
	int pretend = flags & ADD_CACHE_PRETEND;
J
Junio C Hamano 已提交
577 578 579
	int intent_only = flags & ADD_CACHE_INTENT;
	int add_option = (ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE|
			  (intent_only ? ADD_CACHE_NEW_ONLY : 0));
J
Johannes Schindelin 已提交
580

581
	if (!S_ISREG(st_mode) && !S_ISLNK(st_mode) && !S_ISDIR(st_mode))
582
		return error("%s: can only add regular files, symbolic links or git-directories", path);
J
Johannes Schindelin 已提交
583 584

	namelen = strlen(path);
585
	if (S_ISDIR(st_mode)) {
586 587 588
		while (namelen && path[namelen-1] == '/')
			namelen--;
	}
J
Johannes Schindelin 已提交
589 590 591
	size = cache_entry_size(namelen);
	ce = xcalloc(1, size);
	memcpy(ce->name, path, namelen);
592
	ce->ce_flags = namelen;
J
Junio C Hamano 已提交
593 594
	if (!intent_only)
		fill_stat_cache_info(ce, st);
595 596
	else
		ce->ce_flags |= CE_INTENT_TO_ADD;
J
Johannes Schindelin 已提交
597

598
	if (trust_executable_bit && has_symlinks)
599
		ce->ce_mode = create_ce_mode(st_mode);
600
	else {
601 602
		/* 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
603
		 */
604
		struct cache_entry *ent;
605
		int pos = index_name_pos_also_unmerged(istate, path, namelen);
606

607
		ent = (0 <= pos) ? istate->cache[pos] : NULL;
608
		ce->ce_mode = ce_mode_from_stat(ent, st_mode);
J
Johannes Schindelin 已提交
609 610
	}

611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633
	/* When core.ignorecase=true, determine if a directory of the same name but differing
	 * case already exists within the Git repository.  If it does, ensure the directory
	 * case of the file being added to the repository matches (is folded into) the existing
	 * entry's directory case.
	 */
	if (ignore_case) {
		const char *startPtr = ce->name;
		const char *ptr = startPtr;
		while (*ptr) {
			while (*ptr && *ptr != '/')
				++ptr;
			if (*ptr == '/') {
				struct cache_entry *foundce;
				++ptr;
				foundce = index_name_exists(&the_index, ce->name, ptr - ce->name, ignore_case);
				if (foundce) {
					memcpy((void *)startPtr, foundce->name + (startPtr - ce->name), ptr - startPtr);
					startPtr = ptr;
				}
			}
		}
	}

634
	alias = index_name_exists(istate, ce->name, ce_namelen(ce), ignore_case);
635
	if (alias && !ce_stage(alias) && !ie_match_stat(istate, alias, st, ce_option)) {
636 637
		/* Nothing changed, really */
		free(ce);
638 639
		if (!S_ISGITLINK(alias->ce_mode))
			ce_mark_uptodate(alias);
640
		alias->ce_flags |= CE_ADDED;
641 642
		return 0;
	}
J
Junio C Hamano 已提交
643 644 645 646 647 648
	if (!intent_only) {
		if (index_path(ce->sha1, path, st, 1))
			return error("unable to index file %s", path);
	} else
		record_intent_to_add(ce);

649 650 651
	if (ignore_case && alias && different_name(ce, alias))
		ce = create_alias_ce(ce, alias);
	ce->ce_flags |= CE_ADDED;
652

J
Junio C Hamano 已提交
653
	/* It was suspected to be racily clean, but it turns out to be Ok */
654 655 656 657 658 659 660
	was_same = (alias &&
		    !ce_stage(alias) &&
		    !hashcmp(alias->sha1, ce->sha1) &&
		    ce->ce_mode == alias->ce_mode);

	if (pretend)
		;
J
Junio C Hamano 已提交
661
	else if (add_index_entry(istate, ce, add_option))
662
		return error("unable to add %s to index",path);
663
	if (verbose && !was_same)
J
Johannes Schindelin 已提交
664 665 666 667
		printf("add '%s'\n", path);
	return 0;
}

668
int add_file_to_index(struct index_state *istate, const char *path, int flags)
669 670 671
{
	struct stat st;
	if (lstat(path, &st))
672
		die_errno("unable to stat '%s'", path);
673
	return add_to_index(istate, path, &st, flags);
674 675
}

676 677 678 679 680 681 682
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;

683 684
	if (!verify_path(path)) {
		error("Invalid path '%s'", path);
685
		return NULL;
686
	}
687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702

	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;
}

703
int ce_same_name(struct cache_entry *a, struct cache_entry *b)
704 705 706 707 708
{
	int len = ce_namelen(a);
	return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
}

709
int ce_path_match(const struct cache_entry *ce, const struct pathspec *pathspec)
710
{
711
	return match_pathspec_depth(pathspec, ce->name, ce_namelen(ce), 0, NULL);
712 713
}

714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777
/*
 * 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++;
	}
}

778 779 780
/*
 * Do we have another file that has the beginning components being a
 * proper superset of the name we're trying to add?
781
 */
782 783
static int has_file_name(struct index_state *istate,
			 const struct cache_entry *ce, int pos, int ok_to_replace)
784
{
785 786
	int retval = 0;
	int len = ce_namelen(ce);
787
	int stage = ce_stage(ce);
788
	const char *name = ce->name;
789

790 791
	while (pos < istate->cache_nr) {
		struct cache_entry *p = istate->cache[pos++];
792

793
		if (len >= ce_namelen(p))
794
			break;
795 796
		if (memcmp(name, p->name, len))
			break;
797 798
		if (ce_stage(p) != stage)
			continue;
799 800
		if (p->name[len] != '/')
			continue;
801
		if (p->ce_flags & CE_REMOVE)
802
			continue;
803 804 805
		retval = -1;
		if (!ok_to_replace)
			break;
806
		remove_index_entry_at(istate, --pos);
807
	}
808 809
	return retval;
}
810

811 812 813 814
/*
 * Do we have another file with a pathname that is a proper
 * subset of the name we're trying to add?
 */
815 816
static int has_dir_name(struct index_state *istate,
			const struct cache_entry *ce, int pos, int ok_to_replace)
817 818
{
	int retval = 0;
819
	int stage = ce_stage(ce);
820 821
	const char *name = ce->name;
	const char *slash = name + ce_namelen(ce);
822

823 824
	for (;;) {
		int len;
825

826 827 828 829 830 831 832
		for (;;) {
			if (*--slash == '/')
				break;
			if (slash <= ce->name)
				return retval;
		}
		len = slash - name;
833

834
		pos = index_name_pos(istate, name, create_ce_flags(len, stage));
835
		if (pos >= 0) {
836 837 838 839 840 841 842 843
			/*
			 * 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.
			 */
844
			if (!(istate->cache[pos]->ce_flags & CE_REMOVE)) {
845 846 847
				retval = -1;
				if (!ok_to_replace)
					break;
848
				remove_index_entry_at(istate, pos);
849 850
				continue;
			}
851
		}
852 853
		else
			pos = -pos-1;
854 855 856 857

		/*
		 * Trivial optimization: if we find an entry that
		 * already matches the sub-directory, then we know
858
		 * we're ok, and we can exit.
859
		 */
860 861
		while (pos < istate->cache_nr) {
			struct cache_entry *p = istate->cache[pos];
862 863 864 865
			if ((ce_namelen(p) <= len) ||
			    (p->name[len] != '/') ||
			    memcmp(p->name, name, len))
				break; /* not our subdirectory */
866 867 868
			if (ce_stage(p) == stage && !(p->ce_flags & CE_REMOVE))
				/*
				 * p is at the same stage as our entry, and
869 870 871 872 873 874
				 * is a subdirectory of what we are looking
				 * at, so we cannot have conflicts at our
				 * level or anything shorter.
				 */
				return retval;
			pos++;
875
		}
876
	}
877 878 879 880 881 882 883
	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
884
 *
885 886 887 888
 * 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.
 */
889 890 891
static int check_file_directory_conflict(struct index_state *istate,
					 const struct cache_entry *ce,
					 int pos, int ok_to_replace)
892
{
893 894 895 896 897
	int retval;

	/*
	 * When ce is an "I am going away" entry, we allow it to be added
	 */
898
	if (ce->ce_flags & CE_REMOVE)
899 900
		return 0;

901 902 903
	/*
	 * We check if the path is a sub-path of a subsequent pathname
	 * first, since removing those will not change the position
904
	 * in the array.
905
	 */
906
	retval = has_file_name(istate, ce, pos, ok_to_replace);
907

908 909 910 911
	/*
	 * Then check if the path might have a clashing sub-directory
	 * before it.
	 */
912
	return retval + has_dir_name(istate, ce, pos, ok_to_replace);
913 914
}

915
static int add_index_entry_with_check(struct index_state *istate, struct cache_entry *ce, int option)
916 917
{
	int pos;
918 919
	int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
	int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
920
	int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
J
Junio C Hamano 已提交
921
	int new_only = option & ADD_CACHE_NEW_ONLY;
J
Junio C Hamano 已提交
922

J
Junio C Hamano 已提交
923
	cache_tree_invalidate_path(istate->cache_tree, ce->name);
924
	pos = index_name_pos(istate, ce->name, ce->ce_flags);
925

J
Junio C Hamano 已提交
926
	/* existing match? Just replace it. */
927
	if (pos >= 0) {
J
Junio C Hamano 已提交
928 929
		if (!new_only)
			replace_index_entry(istate, pos, ce);
930 931
		return 0;
	}
932
	pos = -pos-1;
933

934 935 936 937
	/*
	 * Inserting a merged entry ("stage 0") into the index
	 * will always replace all non-merged entries..
	 */
938 939
	if (pos < istate->cache_nr && ce_stage(ce) == 0) {
		while (ce_same_name(istate->cache[pos], ce)) {
940
			ok_to_add = 1;
941
			if (!remove_index_entry_at(istate, pos))
942 943 944 945
				break;
		}
	}

946 947
	if (!ok_to_add)
		return -1;
948
	if (!verify_path(ce->name))
949
		return error("Invalid path '%s'", ce->name);
950

J
Junio C Hamano 已提交
951
	if (!skip_df_check &&
952
	    check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
953
		if (!ok_to_replace)
954 955
			return error("'%s' appears as both a file and as a directory",
				     ce->name);
956
		pos = index_name_pos(istate, ce->name, ce->ce_flags);
957 958
		pos = -pos-1;
	}
959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974
	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;
	}
975

976
	/* Make sure the array is big enough .. */
977 978 979 980
	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 *));
981 982 983
	}

	/* Add it in.. */
984
	istate->cache_nr++;
985
	if (istate->cache_nr > pos + 1)
986 987 988
		memmove(istate->cache + pos + 1,
			istate->cache + pos,
			(istate->cache_nr - pos - 1) * sizeof(ce));
989
	set_index_entry(istate, pos, ce);
990
	istate->cache_changed = 1;
991 992 993
	return 0;
}

L
Linus Torvalds 已提交
994 995 996 997 998 999 1000 1001 1002 1003 1004
/*
 * "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.
 */
1005
static struct cache_entry *refresh_cache_ent(struct index_state *istate,
1006 1007
					     struct cache_entry *ce,
					     unsigned int options, int *err)
L
Linus Torvalds 已提交
1008 1009 1010 1011
{
	struct stat st;
	struct cache_entry *updated;
	int changed, size;
1012
	int ignore_valid = options & CE_MATCH_IGNORE_VALID;
1013
	int ignore_skip_worktree = options & CE_MATCH_IGNORE_SKIP_WORKTREE;
L
Linus Torvalds 已提交
1014

1015 1016 1017
	if (ce_uptodate(ce))
		return ce;

1018
	/*
1019 1020 1021
	 * CE_VALID or CE_SKIP_WORKTREE means the user promised us
	 * that the change to the work tree does not matter and told
	 * us not to worry.
1022
	 */
1023 1024 1025 1026
	if (!ignore_skip_worktree && ce_skip_worktree(ce)) {
		ce_mark_uptodate(ce);
		return ce;
	}
1027 1028 1029 1030 1031
	if (!ignore_valid && (ce->ce_flags & CE_VALID)) {
		ce_mark_uptodate(ce);
		return ce;
	}

1032
	if (lstat(ce->name, &st) < 0) {
1033 1034
		if (err)
			*err = errno;
1035 1036
		return NULL;
	}
L
Linus Torvalds 已提交
1037

1038
	changed = ie_match_stat(istate, ce, &st, options);
L
Linus Torvalds 已提交
1039
	if (!changed) {
1040 1041 1042 1043 1044 1045 1046 1047
		/*
		 * 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 &&
1048
		    !(ce->ce_flags & CE_VALID))
L
Linus Torvalds 已提交
1049
			; /* mark this one VALID again */
1050 1051 1052 1053 1054 1055
		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.
			 */
1056 1057
			if (!S_ISGITLINK(ce->ce_mode))
				ce_mark_uptodate(ce);
1058
			return ce;
1059
		}
L
Linus Torvalds 已提交
1060 1061
	}

1062
	if (ie_modified(istate, ce, &st, options)) {
1063 1064
		if (err)
			*err = EINVAL;
1065 1066
		return NULL;
	}
L
Linus Torvalds 已提交
1067 1068 1069 1070 1071

	size = ce_size(ce);
	updated = xmalloc(size);
	memcpy(updated, ce, size);
	fill_stat_cache_info(updated, &st);
1072 1073 1074 1075 1076
	/*
	 * 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.
L
Linus Torvalds 已提交
1077
	 */
1078
	if (!ignore_valid && assume_unchanged &&
1079 1080
	    !(ce->ce_flags & CE_VALID))
		updated->ce_flags &= ~CE_VALID;
L
Linus Torvalds 已提交
1081 1082 1083 1084

	return updated;
}

1085
static void show_file(const char * fmt, const char * name, int in_porcelain,
1086
		      int * first, const char *header_msg)
1087 1088 1089 1090 1091 1092 1093 1094 1095
{
	if (in_porcelain && *first && header_msg) {
		printf("%s\n", header_msg);
		*first=0;
	}
	printf(fmt, name);
}

int refresh_index(struct index_state *istate, unsigned int flags, const char **pathspec,
1096
		  char *seen, const char *header_msg)
L
Linus Torvalds 已提交
1097 1098 1099 1100 1101 1102 1103
{
	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;
1104
	int ignore_submodules = (flags & REFRESH_IGNORE_SUBMODULES) != 0;
1105 1106
	int first = 1;
	int in_porcelain = (flags & REFRESH_IN_PORCELAIN);
1107
	unsigned int options = really ? CE_MATCH_IGNORE_VALID : 0;
1108 1109
	const char *needs_update_fmt;
	const char *needs_merge_fmt;
L
Linus Torvalds 已提交
1110

1111 1112
	needs_update_fmt = (in_porcelain ? "M\t%s\n" : "%s: needs update\n");
	needs_merge_fmt = (in_porcelain ? "U\t%s\n" : "%s: needs merge\n");
1113
	for (i = 0; i < istate->cache_nr; i++) {
L
Linus Torvalds 已提交
1114
		struct cache_entry *ce, *new;
1115 1116
		int cache_errno = 0;

1117
		ce = istate->cache[i];
1118 1119 1120
		if (ignore_submodules && S_ISGITLINK(ce->ce_mode))
			continue;

L
Linus Torvalds 已提交
1121
		if (ce_stage(ce)) {
1122 1123
			while ((i < istate->cache_nr) &&
			       ! strcmp(istate->cache[i]->name, ce->name))
L
Linus Torvalds 已提交
1124 1125 1126 1127
				i++;
			i--;
			if (allow_unmerged)
				continue;
1128
			show_file(needs_merge_fmt, ce->name, in_porcelain, &first, header_msg);
L
Linus Torvalds 已提交
1129 1130 1131 1132
			has_errors = 1;
			continue;
		}

1133 1134 1135
		if (pathspec && !match_pathspec(pathspec, ce->name, strlen(ce->name), 0, seen))
			continue;

1136
		new = refresh_cache_ent(istate, ce, options, &cache_errno);
1137
		if (new == ce)
L
Linus Torvalds 已提交
1138
			continue;
1139 1140
		if (!new) {
			if (not_new && cache_errno == ENOENT)
L
Linus Torvalds 已提交
1141
				continue;
1142
			if (really && cache_errno == EINVAL) {
L
Linus Torvalds 已提交
1143 1144 1145
				/* If we are doing --really-refresh that
				 * means the index is not valid anymore.
				 */
1146
				ce->ce_flags &= ~CE_VALID;
1147
				istate->cache_changed = 1;
L
Linus Torvalds 已提交
1148 1149 1150
			}
			if (quiet)
				continue;
1151
			show_file(needs_update_fmt, ce->name, in_porcelain, &first, header_msg);
L
Linus Torvalds 已提交
1152 1153 1154
			has_errors = 1;
			continue;
		}
1155 1156

		replace_index_entry(istate, i, new);
L
Linus Torvalds 已提交
1157 1158 1159 1160
	}
	return has_errors;
}

1161
static struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
1162
{
1163
	return refresh_cache_ent(&the_index, ce, really, NULL);
1164 1165
}

1166
static int verify_hdr(struct cache_header *hdr, unsigned long size)
1167
{
1168
	git_SHA_CTX c;
1169
	unsigned char sha1[20];
1170

1171
	if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
1172
		return error("bad signature");
1173
	if (hdr->hdr_version != htonl(2) && hdr->hdr_version != htonl(3))
1174
		return error("bad index version");
1175 1176 1177
	git_SHA1_Init(&c);
	git_SHA1_Update(&c, hdr, size - 20);
	git_SHA1_Final(sha1, &c);
1178
	if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
1179
		return error("bad index file sha1 signature");
1180 1181 1182
	return 0;
}

1183 1184
static int read_index_extension(struct index_state *istate,
				const char *ext, void *data, unsigned long sz)
1185 1186 1187
{
	switch (CACHE_EXT(ext)) {
	case CACHE_EXT_TREE:
1188
		istate->cache_tree = cache_tree_read(data, sz);
1189
		break;
1190 1191 1192
	case CACHE_EXT_RESOLVE_UNDO:
		istate->resolve_undo = resolve_undo_read(data, sz);
		break;
1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
	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;
}

1203
int read_index(struct index_state *istate)
1204
{
1205
	return read_index_from(istate, get_index_file());
1206 1207
}

1208 1209
static void convert_from_disk(struct ondisk_cache_entry *ondisk, struct cache_entry *ce)
{
1210
	size_t len;
1211
	const char *name;
1212

K
Kjetil Barvik 已提交
1213 1214 1215 1216
	ce->ce_ctime.sec = ntohl(ondisk->ctime.sec);
	ce->ce_mtime.sec = ntohl(ondisk->mtime.sec);
	ce->ce_ctime.nsec = ntohl(ondisk->ctime.nsec);
	ce->ce_mtime.nsec = ntohl(ondisk->mtime.nsec);
1217 1218 1219 1220 1221 1222 1223 1224
	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);
1225

1226
	hashcpy(ce->sha1, ondisk->sha1);
1227 1228

	len = ce->ce_flags & CE_NAMEMASK;
1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243

	if (ce->ce_flags & CE_EXTENDED) {
		struct ondisk_cache_entry_extended *ondisk2;
		int extended_flags;
		ondisk2 = (struct ondisk_cache_entry_extended *)ondisk;
		extended_flags = ntohs(ondisk2->flags2) << 16;
		/* We do not yet understand any bit out of CE_EXTENDED_FLAGS */
		if (extended_flags & ~CE_EXTENDED_FLAGS)
			die("Unknown index entry format %08x", extended_flags);
		ce->ce_flags |= extended_flags;
		name = ondisk2->name;
	}
	else
		name = ondisk->name;

1244
	if (len == CE_NAMEMASK)
1245
		len = strlen(name);
1246 1247 1248 1249
	/*
	 * NEEDSWORK: If the original index is crafted, this copy could
	 * go unchecked.
	 */
1250
	memcpy(ce->name, name, len + 1);
1251 1252
}

1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266
static inline size_t estimate_cache_size(size_t ondisk_size, unsigned int entries)
{
	long per_entry;

	per_entry = sizeof(struct cache_entry) - sizeof(struct ondisk_cache_entry);

	/*
	 * Alignment can cause differences. This should be "alignof", but
	 * since that's a gcc'ism, just use the size of a pointer.
	 */
	per_entry += sizeof(void *);
	return ondisk_size + entries*per_entry;
}

1267
/* remember to discard_cache() before reading a different cache! */
1268
int read_index_from(struct index_state *istate, const char *path)
1269 1270 1271
{
	int fd, i;
	struct stat st;
1272
	unsigned long src_offset, dst_offset;
1273
	struct cache_header *hdr;
1274 1275
	void *mmap;
	size_t mmap_size;
1276 1277

	errno = EBUSY;
1278
	if (istate->initialized)
1279
		return istate->cache_nr;
1280

1281
	errno = ENOENT;
K
Kjetil Barvik 已提交
1282 1283
	istate->timestamp.sec = 0;
	istate->timestamp.nsec = 0;
1284
	fd = open(path, O_RDONLY);
1285 1286 1287
	if (fd < 0) {
		if (errno == ENOENT)
			return 0;
1288
		die_errno("index file open failed");
1289
	}
1290

1291
	if (fstat(fd, &st))
1292
		die_errno("cannot stat the open index");
1293 1294

	errno = EINVAL;
1295 1296
	mmap_size = xsize_t(st.st_size);
	if (mmap_size < sizeof(struct cache_header) + 20)
1297 1298
		die("index file smaller than expected");

1299
	mmap = xmmap(NULL, mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
1300
	close(fd);
1301
	if (mmap == MAP_FAILED)
1302
		die_errno("unable to map index file");
1303

1304 1305
	hdr = mmap;
	if (verify_hdr(hdr, mmap_size) < 0)
1306 1307
		goto unmap;

1308 1309 1310
	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 *));
1311

1312 1313 1314 1315 1316 1317
	/*
	 * 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
	 */
1318
	istate->alloc = xmalloc(estimate_cache_size(mmap_size, istate->cache_nr));
1319
	istate->initialized = 1;
1320 1321 1322

	src_offset = sizeof(*hdr);
	dst_offset = 0;
1323
	for (i = 0; i < istate->cache_nr; i++) {
1324
		struct ondisk_cache_entry *disk_ce;
1325 1326
		struct cache_entry *ce;

1327 1328 1329
		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);
1330
		set_index_entry(istate, i, ce);
1331 1332 1333

		src_offset += ondisk_ce_size(ce);
		dst_offset += ce_size(ce);
1334
	}
K
Kjetil Barvik 已提交
1335
	istate->timestamp.sec = st.st_mtime;
1336
	istate->timestamp.nsec = ST_MTIME_NSEC(st);
K
Kjetil Barvik 已提交
1337

1338
	while (src_offset <= mmap_size - 20 - 8) {
1339 1340 1341 1342 1343 1344
		/* 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.
		 */
1345
		uint32_t extsize;
1346
		memcpy(&extsize, (char *)mmap + src_offset + 4, 4);
1347
		extsize = ntohl(extsize);
1348
		if (read_index_extension(istate,
1349 1350
					 (const char *) mmap + src_offset,
					 (char *) mmap + src_offset + 8,
1351
					 extsize) < 0)
1352
			goto unmap;
1353 1354
		src_offset += 8;
		src_offset += extsize;
1355
	}
1356
	munmap(mmap, mmap_size);
1357
	return istate->cache_nr;
1358 1359

unmap:
1360
	munmap(mmap, mmap_size);
1361
	errno = EINVAL;
1362
	die("index file corrupt");
1363 1364
}

1365 1366
int is_index_unborn(struct index_state *istate)
{
K
Kjetil Barvik 已提交
1367
	return (!istate->cache_nr && !istate->alloc && !istate->timestamp.sec);
1368 1369
}

1370
int discard_index(struct index_state *istate)
1371
{
1372
	resolve_undo_clear_index(istate);
1373 1374
	istate->cache_nr = 0;
	istate->cache_changed = 0;
K
Kjetil Barvik 已提交
1375 1376
	istate->timestamp.sec = 0;
	istate->timestamp.nsec = 0;
1377
	istate->name_hash_initialized = 0;
1378
	free_hash(&istate->name_hash);
1379
	cache_tree_free(&(istate->cache_tree));
1380 1381
	free(istate->alloc);
	istate->alloc = NULL;
1382
	istate->initialized = 0;
1383 1384

	/* no need to throw away allocated active_cache */
1385
	return 0;
1386 1387
}

1388
int unmerged_index(const struct index_state *istate)
1389 1390 1391 1392 1393 1394 1395 1396 1397
{
	int i;
	for (i = 0; i < istate->cache_nr; i++) {
		if (ce_stage(istate->cache[i]))
			return 1;
	}
	return 0;
}

1398
#define WRITE_BUFFER_SIZE 8192
1399
static unsigned char write_buffer[WRITE_BUFFER_SIZE];
1400 1401
static unsigned long write_buffer_len;

1402
static int ce_write_flush(git_SHA_CTX *context, int fd)
1403 1404 1405
{
	unsigned int buffered = write_buffer_len;
	if (buffered) {
1406
		git_SHA1_Update(context, write_buffer, buffered);
1407
		if (write_in_full(fd, write_buffer, buffered) != buffered)
1408 1409 1410 1411 1412 1413
			return -1;
		write_buffer_len = 0;
	}
	return 0;
}

1414
static int ce_write(git_SHA_CTX *context, int fd, void *data, unsigned int len)
1415 1416 1417 1418 1419 1420 1421 1422 1423
{
	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) {
1424 1425
			write_buffer_len = buffered;
			if (ce_write_flush(context, fd))
1426 1427 1428 1429 1430
				return -1;
			buffered = 0;
		}
		write_buffer_len = buffered;
		len -= partial;
1431
		data = (char *) data + partial;
J
Junio C Hamano 已提交
1432 1433
	}
	return 0;
1434 1435
}

1436
static int write_index_ext_header(git_SHA_CTX *context, int fd,
1437
				  unsigned int ext, unsigned int sz)
1438 1439 1440
{
	ext = htonl(ext);
	sz = htonl(sz);
D
David Rientjes 已提交
1441 1442
	return ((ce_write(context, fd, &ext, 4) < 0) ||
		(ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
1443 1444
}

1445
static int ce_flush(git_SHA_CTX *context, int fd)
1446 1447
{
	unsigned int left = write_buffer_len;
1448

1449 1450
	if (left) {
		write_buffer_len = 0;
1451
		git_SHA1_Update(context, write_buffer, left);
1452
	}
1453

1454 1455
	/* Flush first if not enough space for SHA1 signature */
	if (left + 20 > WRITE_BUFFER_SIZE) {
1456
		if (write_in_full(fd, write_buffer, left) != left)
1457 1458 1459 1460
			return -1;
		left = 0;
	}

1461
	/* Append the SHA1 signature at the end */
1462
	git_SHA1_Final(write_buffer + left, context);
1463
	left += 20;
1464
	return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
1465 1466
}

J
Junio C Hamano 已提交
1467 1468 1469 1470 1471 1472 1473
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.
1474 1475 1476 1477 1478
	 *
	 * Note that this actually does not do much for gitlinks, for
	 * which ce_match_stat_basic() always goes to the actual
	 * contents.  The caller checks with is_racy_timestamp() which
	 * always says "no" for gitlinks, so we are not called for them ;-)
J
Junio C Hamano 已提交
1479 1480 1481 1482 1483 1484 1485 1486
	 */
	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)) {
1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497
		/* 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
		 *
1498
		 * but it does not.  When the second update-index runs,
1499 1500 1501 1502 1503
		 * 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
1504
		 * obtain from the filesystem next time we stat("frotz").
1505 1506 1507 1508 1509 1510 1511
		 *
		 * 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.
		 */
1512
		ce->ce_size = 0;
J
Junio C Hamano 已提交
1513 1514 1515
	}
}

1516
static int ce_write_entry(git_SHA_CTX *c, int fd, struct cache_entry *ce)
1517 1518 1519
{
	int size = ondisk_ce_size(ce);
	struct ondisk_cache_entry *ondisk = xcalloc(1, size);
1520
	char *name;
1521
	int result;
1522

K
Kjetil Barvik 已提交
1523 1524 1525 1526
	ondisk->ctime.sec = htonl(ce->ce_ctime.sec);
	ondisk->mtime.sec = htonl(ce->ce_mtime.sec);
	ondisk->ctime.nsec = htonl(ce->ce_ctime.nsec);
	ondisk->mtime.nsec = htonl(ce->ce_mtime.nsec);
1527 1528 1529 1530 1531 1532 1533 1534
	ondisk->dev  = htonl(ce->ce_dev);
	ondisk->ino  = htonl(ce->ce_ino);
	ondisk->mode = htonl(ce->ce_mode);
	ondisk->uid  = htonl(ce->ce_uid);
	ondisk->gid  = htonl(ce->ce_gid);
	ondisk->size = htonl(ce->ce_size);
	hashcpy(ondisk->sha1, ce->sha1);
	ondisk->flags = htons(ce->ce_flags);
1535 1536 1537 1538 1539 1540 1541 1542 1543
	if (ce->ce_flags & CE_EXTENDED) {
		struct ondisk_cache_entry_extended *ondisk2;
		ondisk2 = (struct ondisk_cache_entry_extended *)ondisk;
		ondisk2->flags2 = htons((ce->ce_flags & CE_EXTENDED_FLAGS) >> 16);
		name = ondisk2->name;
	}
	else
		name = ondisk->name;
	memcpy(name, ce->name, ce_namelen(ce));
1544

1545 1546 1547
	result = ce_write(c, fd, ondisk, size);
	free(ondisk);
	return result;
1548 1549
}

1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562
static int has_racy_timestamp(struct index_state *istate)
{
	int entries = istate->cache_nr;
	int i;

	for (i = 0; i < entries; i++) {
		struct cache_entry *ce = istate->cache[i];
		if (is_racy_timestamp(istate, ce))
			return 1;
	}
	return 0;
}

1563 1564 1565 1566 1567
/*
 * Opportunisticly update the index but do not complain if we can't
 */
void update_index_if_able(struct index_state *istate, struct lock_file *lockfile)
{
1568
	if ((istate->cache_changed || has_racy_timestamp(istate)) &&
1569 1570 1571 1572 1573 1574
	    !write_index(istate, lockfile->fd))
		commit_locked_index(lockfile);
	else
		rollback_lock_file(lockfile);
}

1575
int write_index(struct index_state *istate, int newfd)
1576
{
1577
	git_SHA_CTX c;
1578
	struct cache_header hdr;
1579
	int i, err, removed, extended;
1580 1581
	struct cache_entry **cache = istate->cache;
	int entries = istate->cache_nr;
1582
	struct stat st;
1583

1584
	for (i = removed = extended = 0; i < entries; i++) {
1585
		if (cache[i]->ce_flags & CE_REMOVE)
1586
			removed++;
1587

1588 1589 1590 1591 1592 1593 1594 1595
		/* reduce extended entries if possible */
		cache[i]->ce_flags &= ~CE_EXTENDED;
		if (cache[i]->ce_flags & CE_EXTENDED_FLAGS) {
			extended++;
			cache[i]->ce_flags |= CE_EXTENDED;
		}
	}

1596
	hdr.hdr_signature = htonl(CACHE_SIGNATURE);
1597 1598
	/* for extended format, increase version so older git won't try to read it */
	hdr.hdr_version = htonl(extended ? 3 : 2);
1599
	hdr.hdr_entries = htonl(entries - removed);
1600

1601
	git_SHA1_Init(&c);
1602
	if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
1603 1604 1605 1606
		return -1;

	for (i = 0; i < entries; i++) {
		struct cache_entry *ce = cache[i];
1607
		if (ce->ce_flags & CE_REMOVE)
1608
			continue;
1609
		if (!ce_uptodate(ce) && is_racy_timestamp(istate, ce))
J
Junio C Hamano 已提交
1610
			ce_smudge_racily_clean_entry(ce);
1611
		if (ce_write_entry(&c, newfd, ce) < 0)
1612 1613
			return -1;
	}
1614

1615
	/* Write extension data here */
1616
	if (istate->cache_tree) {
1617
		struct strbuf sb = STRBUF_INIT;
1618 1619 1620 1621 1622 1623

		cache_tree_write(&sb, istate->cache_tree);
		err = write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sb.len) < 0
			|| ce_write(&c, newfd, sb.buf, sb.len) < 0;
		strbuf_release(&sb);
		if (err)
1624 1625
			return -1;
	}
1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
	if (istate->resolve_undo) {
		struct strbuf sb = STRBUF_INIT;

		resolve_undo_write(&sb, istate->resolve_undo);
		err = write_index_ext_header(&c, newfd, CACHE_EXT_RESOLVE_UNDO,
					     sb.len) < 0
			|| ce_write(&c, newfd, sb.buf, sb.len) < 0;
		strbuf_release(&sb);
		if (err)
			return -1;
	}
1637 1638 1639

	if (ce_flush(&c, newfd) || fstat(newfd, &st))
		return -1;
1640 1641
	istate->timestamp.sec = (unsigned int)st.st_mtime;
	istate->timestamp.nsec = ST_MTIME_NSEC(st);
1642
	return 0;
1643
}
1644 1645 1646

/*
 * Read the index file that is potentially unmerged into given
J
Junio C Hamano 已提交
1647
 * index_state, dropping any unmerged entries.  Returns true if
1648 1649 1650 1651 1652 1653 1654
 * the index is unmerged.  Callers who want to refuse to work
 * from an unmerged state can call this and check its return value,
 * instead of calling read_cache().
 */
int read_index_unmerged(struct index_state *istate)
{
	int i;
1655
	int unmerged = 0;
1656 1657 1658 1659

	read_index(istate);
	for (i = 0; i < istate->cache_nr; i++) {
		struct cache_entry *ce = istate->cache[i];
1660 1661 1662 1663
		struct cache_entry *new_ce;
		int size, len;

		if (!ce_stage(ce))
1664
			continue;
1665 1666 1667 1668 1669
		unmerged = 1;
		len = strlen(ce->name);
		size = cache_entry_size(len);
		new_ce = xcalloc(1, size);
		memcpy(new_ce->name, ce->name, len);
1670
		new_ce->ce_flags = create_ce_flags(len, 0) | CE_CONFLICTED;
1671 1672 1673 1674 1675
		new_ce->ce_mode = ce->ce_mode;
		if (add_index_entry(istate, new_ce, 0))
			return error("%s: cannot drop to stage #0",
				     ce->name);
		i = index_name_pos(istate, new_ce->name, len);
1676
	}
1677
	return unmerged;
1678
}
1679

1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706
/*
 * Returns 1 if the path is an "other" path with respect to
 * the index; that is, the path is not mentioned in the index at all,
 * either as a file, a directory with some files in the index,
 * or as an unmerged entry.
 *
 * We helpfully remove a trailing "/" from directories so that
 * the output of read_directory can be used as-is.
 */
int index_name_is_other(const struct index_state *istate, const char *name,
		int namelen)
{
	int pos;
	if (namelen && name[namelen - 1] == '/')
		namelen--;
	pos = index_name_pos(istate, name, namelen);
	if (0 <= pos)
		return 0;	/* exact match */
	pos = -pos - 1;
	if (pos < istate->cache_nr) {
		struct cache_entry *ce = istate->cache[pos];
		if (ce_namelen(ce) == namelen &&
		    !memcmp(ce->name, name, namelen))
			return 0; /* Yup, this one exists unmerged */
	}
	return 1;
}