Merge branch 'jk/pack-objects-with-bitmap-fix'
[git/debian.git] / cache-tree.c
blob5ce51468f0b402fd2121e2c8b51ff12703d431ba
1 #include "cache.h"
2 #include "lockfile.h"
3 #include "tree.h"
4 #include "tree-walk.h"
5 #include "cache-tree.h"
6 #include "object-store.h"
7 #include "replace-object.h"
9 #ifndef DEBUG
10 #define DEBUG 0
11 #endif
13 struct cache_tree *cache_tree(void)
15 struct cache_tree *it = xcalloc(1, sizeof(struct cache_tree));
16 it->entry_count = -1;
17 return it;
20 void cache_tree_free(struct cache_tree **it_p)
22 int i;
23 struct cache_tree *it = *it_p;
25 if (!it)
26 return;
27 for (i = 0; i < it->subtree_nr; i++)
28 if (it->down[i]) {
29 cache_tree_free(&it->down[i]->cache_tree);
30 free(it->down[i]);
32 free(it->down);
33 free(it);
34 *it_p = NULL;
37 static int subtree_name_cmp(const char *one, int onelen,
38 const char *two, int twolen)
40 if (onelen < twolen)
41 return -1;
42 if (twolen < onelen)
43 return 1;
44 return memcmp(one, two, onelen);
47 static int subtree_pos(struct cache_tree *it, const char *path, int pathlen)
49 struct cache_tree_sub **down = it->down;
50 int lo, hi;
51 lo = 0;
52 hi = it->subtree_nr;
53 while (lo < hi) {
54 int mi = lo + (hi - lo) / 2;
55 struct cache_tree_sub *mdl = down[mi];
56 int cmp = subtree_name_cmp(path, pathlen,
57 mdl->name, mdl->namelen);
58 if (!cmp)
59 return mi;
60 if (cmp < 0)
61 hi = mi;
62 else
63 lo = mi + 1;
65 return -lo-1;
68 static struct cache_tree_sub *find_subtree(struct cache_tree *it,
69 const char *path,
70 int pathlen,
71 int create)
73 struct cache_tree_sub *down;
74 int pos = subtree_pos(it, path, pathlen);
75 if (0 <= pos)
76 return it->down[pos];
77 if (!create)
78 return NULL;
80 pos = -pos-1;
81 ALLOC_GROW(it->down, it->subtree_nr + 1, it->subtree_alloc);
82 it->subtree_nr++;
84 FLEX_ALLOC_MEM(down, name, path, pathlen);
85 down->cache_tree = NULL;
86 down->namelen = pathlen;
88 if (pos < it->subtree_nr)
89 MOVE_ARRAY(it->down + pos + 1, it->down + pos,
90 it->subtree_nr - pos - 1);
91 it->down[pos] = down;
92 return down;
95 struct cache_tree_sub *cache_tree_sub(struct cache_tree *it, const char *path)
97 int pathlen = strlen(path);
98 return find_subtree(it, path, pathlen, 1);
101 static int do_invalidate_path(struct cache_tree *it, const char *path)
103 /* a/b/c
104 * ==> invalidate self
105 * ==> find "a", have it invalidate "b/c"
107 * ==> invalidate self
108 * ==> if "a" exists as a subtree, remove it.
110 const char *slash;
111 int namelen;
112 struct cache_tree_sub *down;
114 #if DEBUG
115 fprintf(stderr, "cache-tree invalidate <%s>\n", path);
116 #endif
118 if (!it)
119 return 0;
120 slash = strchrnul(path, '/');
121 namelen = slash - path;
122 it->entry_count = -1;
123 if (!*slash) {
124 int pos;
125 pos = subtree_pos(it, path, namelen);
126 if (0 <= pos) {
127 cache_tree_free(&it->down[pos]->cache_tree);
128 free(it->down[pos]);
129 /* 0 1 2 3 4 5
130 * ^ ^subtree_nr = 6
131 * pos
132 * move 4 and 5 up one place (2 entries)
133 * 2 = 6 - 3 - 1 = subtree_nr - pos - 1
135 MOVE_ARRAY(it->down + pos, it->down + pos + 1,
136 it->subtree_nr - pos - 1);
137 it->subtree_nr--;
139 return 1;
141 down = find_subtree(it, path, namelen, 0);
142 if (down)
143 do_invalidate_path(down->cache_tree, slash + 1);
144 return 1;
147 void cache_tree_invalidate_path(struct index_state *istate, const char *path)
149 if (do_invalidate_path(istate->cache_tree, path))
150 istate->cache_changed |= CACHE_TREE_CHANGED;
153 static int verify_cache(struct cache_entry **cache,
154 int entries, int flags)
156 int i, funny;
157 int silent = flags & WRITE_TREE_SILENT;
159 /* Verify that the tree is merged */
160 funny = 0;
161 for (i = 0; i < entries; i++) {
162 const struct cache_entry *ce = cache[i];
163 if (ce_stage(ce)) {
164 if (silent)
165 return -1;
166 if (10 < ++funny) {
167 fprintf(stderr, "...\n");
168 break;
170 fprintf(stderr, "%s: unmerged (%s)\n",
171 ce->name, oid_to_hex(&ce->oid));
174 if (funny)
175 return -1;
177 /* Also verify that the cache does not have path and path/file
178 * at the same time. At this point we know the cache has only
179 * stage 0 entries.
181 funny = 0;
182 for (i = 0; i < entries - 1; i++) {
183 /* path/file always comes after path because of the way
184 * the cache is sorted. Also path can appear only once,
185 * which means conflicting one would immediately follow.
187 const char *this_name = cache[i]->name;
188 const char *next_name = cache[i+1]->name;
189 int this_len = strlen(this_name);
190 if (this_len < strlen(next_name) &&
191 strncmp(this_name, next_name, this_len) == 0 &&
192 next_name[this_len] == '/') {
193 if (10 < ++funny) {
194 fprintf(stderr, "...\n");
195 break;
197 fprintf(stderr, "You have both %s and %s\n",
198 this_name, next_name);
201 if (funny)
202 return -1;
203 return 0;
206 static void discard_unused_subtrees(struct cache_tree *it)
208 struct cache_tree_sub **down = it->down;
209 int nr = it->subtree_nr;
210 int dst, src;
211 for (dst = src = 0; src < nr; src++) {
212 struct cache_tree_sub *s = down[src];
213 if (s->used)
214 down[dst++] = s;
215 else {
216 cache_tree_free(&s->cache_tree);
217 free(s);
218 it->subtree_nr--;
223 int cache_tree_fully_valid(struct cache_tree *it)
225 int i;
226 if (!it)
227 return 0;
228 if (it->entry_count < 0 || !has_sha1_file(it->oid.hash))
229 return 0;
230 for (i = 0; i < it->subtree_nr; i++) {
231 if (!cache_tree_fully_valid(it->down[i]->cache_tree))
232 return 0;
234 return 1;
237 static int update_one(struct cache_tree *it,
238 struct cache_entry **cache,
239 int entries,
240 const char *base,
241 int baselen,
242 int *skip_count,
243 int flags)
245 struct strbuf buffer;
246 int missing_ok = flags & WRITE_TREE_MISSING_OK;
247 int dryrun = flags & WRITE_TREE_DRY_RUN;
248 int repair = flags & WRITE_TREE_REPAIR;
249 int to_invalidate = 0;
250 int i;
252 assert(!(dryrun && repair));
254 *skip_count = 0;
256 if (0 <= it->entry_count && has_sha1_file(it->oid.hash))
257 return it->entry_count;
260 * We first scan for subtrees and update them; we start by
261 * marking existing subtrees -- the ones that are unmarked
262 * should not be in the result.
264 for (i = 0; i < it->subtree_nr; i++)
265 it->down[i]->used = 0;
268 * Find the subtrees and update them.
270 i = 0;
271 while (i < entries) {
272 const struct cache_entry *ce = cache[i];
273 struct cache_tree_sub *sub;
274 const char *path, *slash;
275 int pathlen, sublen, subcnt, subskip;
277 path = ce->name;
278 pathlen = ce_namelen(ce);
279 if (pathlen <= baselen || memcmp(base, path, baselen))
280 break; /* at the end of this level */
282 slash = strchr(path + baselen, '/');
283 if (!slash) {
284 i++;
285 continue;
288 * a/bbb/c (base = a/, slash = /c)
289 * ==>
290 * path+baselen = bbb/c, sublen = 3
292 sublen = slash - (path + baselen);
293 sub = find_subtree(it, path + baselen, sublen, 1);
294 if (!sub->cache_tree)
295 sub->cache_tree = cache_tree();
296 subcnt = update_one(sub->cache_tree,
297 cache + i, entries - i,
298 path,
299 baselen + sublen + 1,
300 &subskip,
301 flags);
302 if (subcnt < 0)
303 return subcnt;
304 if (!subcnt)
305 die("index cache-tree records empty sub-tree");
306 i += subcnt;
307 sub->count = subcnt; /* to be used in the next loop */
308 *skip_count += subskip;
309 sub->used = 1;
312 discard_unused_subtrees(it);
315 * Then write out the tree object for this level.
317 strbuf_init(&buffer, 8192);
319 i = 0;
320 while (i < entries) {
321 const struct cache_entry *ce = cache[i];
322 struct cache_tree_sub *sub = NULL;
323 const char *path, *slash;
324 int pathlen, entlen;
325 const struct object_id *oid;
326 unsigned mode;
327 int expected_missing = 0;
328 int contains_ita = 0;
330 path = ce->name;
331 pathlen = ce_namelen(ce);
332 if (pathlen <= baselen || memcmp(base, path, baselen))
333 break; /* at the end of this level */
335 slash = strchr(path + baselen, '/');
336 if (slash) {
337 entlen = slash - (path + baselen);
338 sub = find_subtree(it, path + baselen, entlen, 0);
339 if (!sub)
340 die("cache-tree.c: '%.*s' in '%s' not found",
341 entlen, path + baselen, path);
342 i += sub->count;
343 oid = &sub->cache_tree->oid;
344 mode = S_IFDIR;
345 contains_ita = sub->cache_tree->entry_count < 0;
346 if (contains_ita) {
347 to_invalidate = 1;
348 expected_missing = 1;
351 else {
352 oid = &ce->oid;
353 mode = ce->ce_mode;
354 entlen = pathlen - baselen;
355 i++;
358 if (is_null_oid(oid) ||
359 (mode != S_IFGITLINK && !missing_ok && !has_object_file(oid))) {
360 strbuf_release(&buffer);
361 if (expected_missing)
362 return -1;
363 return error("invalid object %06o %s for '%.*s'",
364 mode, oid_to_hex(oid), entlen+baselen, path);
368 * CE_REMOVE entries are removed before the index is
369 * written to disk. Skip them to remain consistent
370 * with the future on-disk index.
372 if (ce->ce_flags & CE_REMOVE) {
373 *skip_count = *skip_count + 1;
374 continue;
378 * CE_INTENT_TO_ADD entries exist on on-disk index but
379 * they are not part of generated trees. Invalidate up
380 * to root to force cache-tree users to read elsewhere.
382 if (!sub && ce_intent_to_add(ce)) {
383 to_invalidate = 1;
384 continue;
388 * "sub" can be an empty tree if all subentries are i-t-a.
390 if (contains_ita && is_empty_tree_oid(oid))
391 continue;
393 strbuf_grow(&buffer, entlen + 100);
394 strbuf_addf(&buffer, "%o %.*s%c", mode, entlen, path + baselen, '\0');
395 strbuf_add(&buffer, oid->hash, the_hash_algo->rawsz);
397 #if DEBUG
398 fprintf(stderr, "cache-tree update-one %o %.*s\n",
399 mode, entlen, path + baselen);
400 #endif
403 if (repair) {
404 struct object_id oid;
405 hash_object_file(buffer.buf, buffer.len, tree_type, &oid);
406 if (has_object_file(&oid))
407 oidcpy(&it->oid, &oid);
408 else
409 to_invalidate = 1;
410 } else if (dryrun) {
411 hash_object_file(buffer.buf, buffer.len, tree_type, &it->oid);
412 } else if (write_object_file(buffer.buf, buffer.len, tree_type,
413 &it->oid)) {
414 strbuf_release(&buffer);
415 return -1;
418 strbuf_release(&buffer);
419 it->entry_count = to_invalidate ? -1 : i - *skip_count;
420 #if DEBUG
421 fprintf(stderr, "cache-tree update-one (%d ent, %d subtree) %s\n",
422 it->entry_count, it->subtree_nr,
423 oid_to_hex(&it->oid));
424 #endif
425 return i;
428 int cache_tree_update(struct index_state *istate, int flags)
430 struct cache_tree *it = istate->cache_tree;
431 struct cache_entry **cache = istate->cache;
432 int entries = istate->cache_nr;
433 int skip, i = verify_cache(cache, entries, flags);
435 if (i)
436 return i;
437 trace_performance_enter();
438 i = update_one(it, cache, entries, "", 0, &skip, flags);
439 trace_performance_leave("cache_tree_update");
440 if (i < 0)
441 return i;
442 istate->cache_changed |= CACHE_TREE_CHANGED;
443 return 0;
446 static void write_one(struct strbuf *buffer, struct cache_tree *it,
447 const char *path, int pathlen)
449 int i;
451 /* One "cache-tree" entry consists of the following:
452 * path (NUL terminated)
453 * entry_count, subtree_nr ("%d %d\n")
454 * tree-sha1 (missing if invalid)
455 * subtree_nr "cache-tree" entries for subtrees.
457 strbuf_grow(buffer, pathlen + 100);
458 strbuf_add(buffer, path, pathlen);
459 strbuf_addf(buffer, "%c%d %d\n", 0, it->entry_count, it->subtree_nr);
461 #if DEBUG
462 if (0 <= it->entry_count)
463 fprintf(stderr, "cache-tree <%.*s> (%d ent, %d subtree) %s\n",
464 pathlen, path, it->entry_count, it->subtree_nr,
465 oid_to_hex(&it->oid));
466 else
467 fprintf(stderr, "cache-tree <%.*s> (%d subtree) invalid\n",
468 pathlen, path, it->subtree_nr);
469 #endif
471 if (0 <= it->entry_count) {
472 strbuf_add(buffer, it->oid.hash, the_hash_algo->rawsz);
474 for (i = 0; i < it->subtree_nr; i++) {
475 struct cache_tree_sub *down = it->down[i];
476 if (i) {
477 struct cache_tree_sub *prev = it->down[i-1];
478 if (subtree_name_cmp(down->name, down->namelen,
479 prev->name, prev->namelen) <= 0)
480 die("fatal - unsorted cache subtree");
482 write_one(buffer, down->cache_tree, down->name, down->namelen);
486 void cache_tree_write(struct strbuf *sb, struct cache_tree *root)
488 write_one(sb, root, "", 0);
491 static struct cache_tree *read_one(const char **buffer, unsigned long *size_p)
493 const char *buf = *buffer;
494 unsigned long size = *size_p;
495 const char *cp;
496 char *ep;
497 struct cache_tree *it;
498 int i, subtree_nr;
499 const unsigned rawsz = the_hash_algo->rawsz;
501 it = NULL;
502 /* skip name, but make sure name exists */
503 while (size && *buf) {
504 size--;
505 buf++;
507 if (!size)
508 goto free_return;
509 buf++; size--;
510 it = cache_tree();
512 cp = buf;
513 it->entry_count = strtol(cp, &ep, 10);
514 if (cp == ep)
515 goto free_return;
516 cp = ep;
517 subtree_nr = strtol(cp, &ep, 10);
518 if (cp == ep)
519 goto free_return;
520 while (size && *buf && *buf != '\n') {
521 size--;
522 buf++;
524 if (!size)
525 goto free_return;
526 buf++; size--;
527 if (0 <= it->entry_count) {
528 if (size < rawsz)
529 goto free_return;
530 oidread(&it->oid, (const unsigned char *)buf);
531 buf += rawsz;
532 size -= rawsz;
535 #if DEBUG
536 if (0 <= it->entry_count)
537 fprintf(stderr, "cache-tree <%s> (%d ent, %d subtree) %s\n",
538 *buffer, it->entry_count, subtree_nr,
539 oid_to_hex(&it->oid));
540 else
541 fprintf(stderr, "cache-tree <%s> (%d subtrees) invalid\n",
542 *buffer, subtree_nr);
543 #endif
546 * Just a heuristic -- we do not add directories that often but
547 * we do not want to have to extend it immediately when we do,
548 * hence +2.
550 it->subtree_alloc = subtree_nr + 2;
551 it->down = xcalloc(it->subtree_alloc, sizeof(struct cache_tree_sub *));
552 for (i = 0; i < subtree_nr; i++) {
553 /* read each subtree */
554 struct cache_tree *sub;
555 struct cache_tree_sub *subtree;
556 const char *name = buf;
558 sub = read_one(&buf, &size);
559 if (!sub)
560 goto free_return;
561 subtree = cache_tree_sub(it, name);
562 subtree->cache_tree = sub;
564 if (subtree_nr != it->subtree_nr)
565 die("cache-tree: internal error");
566 *buffer = buf;
567 *size_p = size;
568 return it;
570 free_return:
571 cache_tree_free(&it);
572 return NULL;
575 struct cache_tree *cache_tree_read(const char *buffer, unsigned long size)
577 if (buffer[0])
578 return NULL; /* not the whole tree */
579 return read_one(&buffer, &size);
582 static struct cache_tree *cache_tree_find(struct cache_tree *it, const char *path)
584 if (!it)
585 return NULL;
586 while (*path) {
587 const char *slash;
588 struct cache_tree_sub *sub;
590 slash = strchrnul(path, '/');
592 * Between path and slash is the name of the subtree
593 * to look for.
595 sub = find_subtree(it, path, slash - path, 0);
596 if (!sub)
597 return NULL;
598 it = sub->cache_tree;
600 path = slash;
601 while (*path == '/')
602 path++;
604 return it;
607 int write_index_as_tree(struct object_id *oid, struct index_state *index_state, const char *index_path, int flags, const char *prefix)
609 int entries, was_valid;
610 struct lock_file lock_file = LOCK_INIT;
611 int ret = 0;
613 hold_lock_file_for_update(&lock_file, index_path, LOCK_DIE_ON_ERROR);
615 entries = read_index_from(index_state, index_path, get_git_dir());
616 if (entries < 0) {
617 ret = WRITE_TREE_UNREADABLE_INDEX;
618 goto out;
620 if (flags & WRITE_TREE_IGNORE_CACHE_TREE)
621 cache_tree_free(&index_state->cache_tree);
623 if (!index_state->cache_tree)
624 index_state->cache_tree = cache_tree();
626 was_valid = cache_tree_fully_valid(index_state->cache_tree);
627 if (!was_valid) {
628 if (cache_tree_update(index_state, flags) < 0) {
629 ret = WRITE_TREE_UNMERGED_INDEX;
630 goto out;
632 write_locked_index(index_state, &lock_file, COMMIT_LOCK);
633 /* Not being able to write is fine -- we are only interested
634 * in updating the cache-tree part, and if the next caller
635 * ends up using the old index with unupdated cache-tree part
636 * it misses the work we did here, but that is just a
637 * performance penalty and not a big deal.
641 if (prefix) {
642 struct cache_tree *subtree;
643 subtree = cache_tree_find(index_state->cache_tree, prefix);
644 if (!subtree) {
645 ret = WRITE_TREE_PREFIX_ERROR;
646 goto out;
648 oidcpy(oid, &subtree->oid);
650 else
651 oidcpy(oid, &index_state->cache_tree->oid);
653 out:
654 rollback_lock_file(&lock_file);
655 return ret;
658 static void prime_cache_tree_rec(struct cache_tree *it, struct tree *tree)
660 struct tree_desc desc;
661 struct name_entry entry;
662 int cnt;
664 oidcpy(&it->oid, &tree->object.oid);
665 init_tree_desc(&desc, tree->buffer, tree->size);
666 cnt = 0;
667 while (tree_entry(&desc, &entry)) {
668 if (!S_ISDIR(entry.mode))
669 cnt++;
670 else {
671 struct cache_tree_sub *sub;
672 struct tree *subtree = lookup_tree(the_repository,
673 entry.oid);
674 if (!subtree->object.parsed)
675 parse_tree(subtree);
676 sub = cache_tree_sub(it, entry.path);
677 sub->cache_tree = cache_tree();
678 prime_cache_tree_rec(sub->cache_tree, subtree);
679 cnt += sub->cache_tree->entry_count;
682 it->entry_count = cnt;
685 void prime_cache_tree(struct index_state *istate, struct tree *tree)
687 cache_tree_free(&istate->cache_tree);
688 istate->cache_tree = cache_tree();
689 prime_cache_tree_rec(istate->cache_tree, tree);
690 istate->cache_changed |= CACHE_TREE_CHANGED;
694 * find the cache_tree that corresponds to the current level without
695 * exploding the full path into textual form. The root of the
696 * cache tree is given as "root", and our current level is "info".
697 * (1) When at root level, info->prev is NULL, so it is "root" itself.
698 * (2) Otherwise, find the cache_tree that corresponds to one level
699 * above us, and find ourselves in there.
701 static struct cache_tree *find_cache_tree_from_traversal(struct cache_tree *root,
702 struct traverse_info *info)
704 struct cache_tree *our_parent;
706 if (!info->prev)
707 return root;
708 our_parent = find_cache_tree_from_traversal(root, info->prev);
709 return cache_tree_find(our_parent, info->name.path);
712 int cache_tree_matches_traversal(struct cache_tree *root,
713 struct name_entry *ent,
714 struct traverse_info *info)
716 struct cache_tree *it;
718 it = find_cache_tree_from_traversal(root, info);
719 it = cache_tree_find(it, ent->path);
720 if (it && it->entry_count > 0 && oideq(ent->oid, &it->oid))
721 return it->entry_count;
722 return 0;
725 static void verify_one(struct index_state *istate,
726 struct cache_tree *it,
727 struct strbuf *path)
729 int i, pos, len = path->len;
730 struct strbuf tree_buf = STRBUF_INIT;
731 struct object_id new_oid;
733 for (i = 0; i < it->subtree_nr; i++) {
734 strbuf_addf(path, "%s/", it->down[i]->name);
735 verify_one(istate, it->down[i]->cache_tree, path);
736 strbuf_setlen(path, len);
739 if (it->entry_count < 0 ||
740 /* no verification on tests (t7003) that replace trees */
741 lookup_replace_object(the_repository, &it->oid) != &it->oid)
742 return;
744 if (path->len) {
745 pos = index_name_pos(istate, path->buf, path->len);
746 pos = -pos - 1;
747 } else {
748 pos = 0;
751 i = 0;
752 while (i < it->entry_count) {
753 struct cache_entry *ce = istate->cache[pos + i];
754 const char *slash;
755 struct cache_tree_sub *sub = NULL;
756 const struct object_id *oid;
757 const char *name;
758 unsigned mode;
759 int entlen;
761 if (ce->ce_flags & (CE_STAGEMASK | CE_INTENT_TO_ADD | CE_REMOVE))
762 BUG("%s with flags 0x%x should not be in cache-tree",
763 ce->name, ce->ce_flags);
764 name = ce->name + path->len;
765 slash = strchr(name, '/');
766 if (slash) {
767 entlen = slash - name;
768 sub = find_subtree(it, ce->name + path->len, entlen, 0);
769 if (!sub || sub->cache_tree->entry_count < 0)
770 BUG("bad subtree '%.*s'", entlen, name);
771 oid = &sub->cache_tree->oid;
772 mode = S_IFDIR;
773 i += sub->cache_tree->entry_count;
774 } else {
775 oid = &ce->oid;
776 mode = ce->ce_mode;
777 entlen = ce_namelen(ce) - path->len;
778 i++;
780 strbuf_addf(&tree_buf, "%o %.*s%c", mode, entlen, name, '\0');
781 strbuf_add(&tree_buf, oid->hash, the_hash_algo->rawsz);
783 hash_object_file(tree_buf.buf, tree_buf.len, tree_type, &new_oid);
784 if (oidcmp(&new_oid, &it->oid))
785 BUG("cache-tree for path %.*s does not match. "
786 "Expected %s got %s", len, path->buf,
787 oid_to_hex(&new_oid), oid_to_hex(&it->oid));
788 strbuf_setlen(path, len);
789 strbuf_release(&tree_buf);
792 void cache_tree_verify(struct index_state *istate)
794 struct strbuf path = STRBUF_INIT;
796 if (!istate->cache_tree)
797 return;
798 verify_one(istate, istate->cache_tree, &path);
799 strbuf_release(&path);