Merge branch 'rs/describe-unique-abbrev' into maint
[git.git] / builtin / fsck.c
blob92ce775a74d88036b024b24d74e3987da521ef41
1 #include "builtin.h"
2 #include "cache.h"
3 #include "config.h"
4 #include "commit.h"
5 #include "tree.h"
6 #include "blob.h"
7 #include "tag.h"
8 #include "refs.h"
9 #include "pack.h"
10 #include "cache-tree.h"
11 #include "tree-walk.h"
12 #include "fsck.h"
13 #include "parse-options.h"
14 #include "dir.h"
15 #include "progress.h"
16 #include "streaming.h"
17 #include "decorate.h"
18 #include "packfile.h"
20 #define REACHABLE 0x0001
21 #define SEEN 0x0002
22 #define HAS_OBJ 0x0004
23 /* This flag is set if something points to this object. */
24 #define USED 0x0008
26 static int show_root;
27 static int show_tags;
28 static int show_unreachable;
29 static int include_reflogs = 1;
30 static int check_full = 1;
31 static int connectivity_only;
32 static int check_strict;
33 static int keep_cache_objects;
34 static struct fsck_options fsck_walk_options = FSCK_OPTIONS_DEFAULT;
35 static struct fsck_options fsck_obj_options = FSCK_OPTIONS_DEFAULT;
36 static struct object_id head_oid;
37 static const char *head_points_at;
38 static int errors_found;
39 static int write_lost_and_found;
40 static int verbose;
41 static int show_progress = -1;
42 static int show_dangling = 1;
43 static int name_objects;
44 #define ERROR_OBJECT 01
45 #define ERROR_REACHABLE 02
46 #define ERROR_PACK 04
47 #define ERROR_REFS 010
49 static const char *describe_object(struct object *obj)
51 static struct strbuf buf = STRBUF_INIT;
52 char *name = name_objects ?
53 lookup_decoration(fsck_walk_options.object_names, obj) : NULL;
55 strbuf_reset(&buf);
56 strbuf_addstr(&buf, oid_to_hex(&obj->oid));
57 if (name)
58 strbuf_addf(&buf, " (%s)", name);
60 return buf.buf;
63 static const char *printable_type(struct object *obj)
65 const char *ret;
67 if (obj->type == OBJ_NONE) {
68 enum object_type type = sha1_object_info(obj->oid.hash, NULL);
69 if (type > 0)
70 object_as_type(obj, type, 0);
73 ret = typename(obj->type);
74 if (!ret)
75 ret = "unknown";
77 return ret;
80 static int fsck_config(const char *var, const char *value, void *cb)
82 if (strcmp(var, "fsck.skiplist") == 0) {
83 const char *path;
84 struct strbuf sb = STRBUF_INIT;
86 if (git_config_pathname(&path, var, value))
87 return 1;
88 strbuf_addf(&sb, "skiplist=%s", path);
89 free((char *)path);
90 fsck_set_msg_types(&fsck_obj_options, sb.buf);
91 strbuf_release(&sb);
92 return 0;
95 if (skip_prefix(var, "fsck.", &var)) {
96 fsck_set_msg_type(&fsck_obj_options, var, value);
97 return 0;
100 return git_default_config(var, value, cb);
103 static void objreport(struct object *obj, const char *msg_type,
104 const char *err)
106 fprintf(stderr, "%s in %s %s: %s\n",
107 msg_type, printable_type(obj), describe_object(obj), err);
110 static int objerror(struct object *obj, const char *err)
112 errors_found |= ERROR_OBJECT;
113 objreport(obj, "error", err);
114 return -1;
117 static int fsck_error_func(struct fsck_options *o,
118 struct object *obj, int type, const char *message)
120 objreport(obj, (type == FSCK_WARN) ? "warning" : "error", message);
121 return (type == FSCK_WARN) ? 0 : 1;
124 static struct object_array pending;
126 static int mark_object(struct object *obj, int type, void *data, struct fsck_options *options)
128 struct object *parent = data;
131 * The only case data is NULL or type is OBJ_ANY is when
132 * mark_object_reachable() calls us. All the callers of
133 * that function has non-NULL obj hence ...
135 if (!obj) {
136 /* ... these references to parent->fld are safe here */
137 printf("broken link from %7s %s\n",
138 printable_type(parent), describe_object(parent));
139 printf("broken link from %7s %s\n",
140 (type == OBJ_ANY ? "unknown" : typename(type)), "unknown");
141 errors_found |= ERROR_REACHABLE;
142 return 1;
145 if (type != OBJ_ANY && obj->type != type)
146 /* ... and the reference to parent is safe here */
147 objerror(parent, "wrong object type in link");
149 if (obj->flags & REACHABLE)
150 return 0;
151 obj->flags |= REACHABLE;
152 if (!(obj->flags & HAS_OBJ)) {
153 if (parent && !has_object_file(&obj->oid)) {
154 printf("broken link from %7s %s\n",
155 printable_type(parent), describe_object(parent));
156 printf(" to %7s %s\n",
157 printable_type(obj), describe_object(obj));
158 errors_found |= ERROR_REACHABLE;
160 return 1;
163 add_object_array(obj, NULL, &pending);
164 return 0;
167 static void mark_object_reachable(struct object *obj)
169 mark_object(obj, OBJ_ANY, NULL, NULL);
172 static int traverse_one_object(struct object *obj)
174 int result = fsck_walk(obj, obj, &fsck_walk_options);
176 if (obj->type == OBJ_TREE) {
177 struct tree *tree = (struct tree *)obj;
178 free_tree_buffer(tree);
180 return result;
183 static int traverse_reachable(void)
185 struct progress *progress = NULL;
186 unsigned int nr = 0;
187 int result = 0;
188 if (show_progress)
189 progress = start_delayed_progress(_("Checking connectivity"), 0);
190 while (pending.nr) {
191 result |= traverse_one_object(object_array_pop(&pending));
192 display_progress(progress, ++nr);
194 stop_progress(&progress);
195 return !!result;
198 static int mark_used(struct object *obj, int type, void *data, struct fsck_options *options)
200 if (!obj)
201 return 1;
202 obj->flags |= USED;
203 return 0;
207 * Check a single reachable object
209 static void check_reachable_object(struct object *obj)
212 * We obviously want the object to be parsed,
213 * except if it was in a pack-file and we didn't
214 * do a full fsck
216 if (!(obj->flags & HAS_OBJ)) {
217 if (has_sha1_pack(obj->oid.hash))
218 return; /* it is in pack - forget about it */
219 printf("missing %s %s\n", printable_type(obj),
220 describe_object(obj));
221 errors_found |= ERROR_REACHABLE;
222 return;
227 * Check a single unreachable object
229 static void check_unreachable_object(struct object *obj)
232 * Missing unreachable object? Ignore it. It's not like
233 * we miss it (since it can't be reached), nor do we want
234 * to complain about it being unreachable (since it does
235 * not exist).
237 if (!(obj->flags & HAS_OBJ))
238 return;
241 * Unreachable object that exists? Show it if asked to,
242 * since this is something that is prunable.
244 if (show_unreachable) {
245 printf("unreachable %s %s\n", printable_type(obj),
246 describe_object(obj));
247 return;
251 * "!USED" means that nothing at all points to it, including
252 * other unreachable objects. In other words, it's the "tip"
253 * of some set of unreachable objects, usually a commit that
254 * got dropped.
256 * Such starting points are more interesting than some random
257 * set of unreachable objects, so we show them even if the user
258 * hasn't asked for _all_ unreachable objects. If you have
259 * deleted a branch by mistake, this is a prime candidate to
260 * start looking at, for example.
262 if (!(obj->flags & USED)) {
263 if (show_dangling)
264 printf("dangling %s %s\n", printable_type(obj),
265 describe_object(obj));
266 if (write_lost_and_found) {
267 char *filename = git_pathdup("lost-found/%s/%s",
268 obj->type == OBJ_COMMIT ? "commit" : "other",
269 describe_object(obj));
270 FILE *f;
272 if (safe_create_leading_directories_const(filename)) {
273 error("Could not create lost-found");
274 free(filename);
275 return;
277 f = xfopen(filename, "w");
278 if (obj->type == OBJ_BLOB) {
279 if (stream_blob_to_fd(fileno(f), &obj->oid, NULL, 1))
280 die_errno("Could not write '%s'", filename);
281 } else
282 fprintf(f, "%s\n", describe_object(obj));
283 if (fclose(f))
284 die_errno("Could not finish '%s'",
285 filename);
286 free(filename);
288 return;
292 * Otherwise? It's there, it's unreachable, and some other unreachable
293 * object points to it. Ignore it - it's not interesting, and we showed
294 * all the interesting cases above.
298 static void check_object(struct object *obj)
300 if (verbose)
301 fprintf(stderr, "Checking %s\n", describe_object(obj));
303 if (obj->flags & REACHABLE)
304 check_reachable_object(obj);
305 else
306 check_unreachable_object(obj);
309 static void check_connectivity(void)
311 int i, max;
313 /* Traverse the pending reachable objects */
314 traverse_reachable();
316 /* Look up all the requirements, warn about missing objects.. */
317 max = get_max_object_index();
318 if (verbose)
319 fprintf(stderr, "Checking connectivity (%d objects)\n", max);
321 for (i = 0; i < max; i++) {
322 struct object *obj = get_indexed_object(i);
324 if (obj)
325 check_object(obj);
329 static int fsck_obj(struct object *obj)
331 int err;
333 if (obj->flags & SEEN)
334 return 0;
335 obj->flags |= SEEN;
337 if (verbose)
338 fprintf(stderr, "Checking %s %s\n",
339 printable_type(obj), describe_object(obj));
341 if (fsck_walk(obj, NULL, &fsck_obj_options))
342 objerror(obj, "broken links");
343 err = fsck_object(obj, NULL, 0, &fsck_obj_options);
344 if (err)
345 goto out;
347 if (obj->type == OBJ_COMMIT) {
348 struct commit *commit = (struct commit *) obj;
350 if (!commit->parents && show_root)
351 printf("root %s\n", describe_object(&commit->object));
354 if (obj->type == OBJ_TAG) {
355 struct tag *tag = (struct tag *) obj;
357 if (show_tags && tag->tagged) {
358 printf("tagged %s %s", printable_type(tag->tagged),
359 describe_object(tag->tagged));
360 printf(" (%s) in %s\n", tag->tag,
361 describe_object(&tag->object));
365 out:
366 if (obj->type == OBJ_TREE)
367 free_tree_buffer((struct tree *)obj);
368 if (obj->type == OBJ_COMMIT)
369 free_commit_buffer((struct commit *)obj);
370 return err;
373 static int fsck_obj_buffer(const struct object_id *oid, enum object_type type,
374 unsigned long size, void *buffer, int *eaten)
377 * Note, buffer may be NULL if type is OBJ_BLOB. See
378 * verify_packfile(), data_valid variable for details.
380 struct object *obj;
381 obj = parse_object_buffer(oid, type, size, buffer, eaten);
382 if (!obj) {
383 errors_found |= ERROR_OBJECT;
384 return error("%s: object corrupt or missing", oid_to_hex(oid));
386 obj->flags &= ~(REACHABLE | SEEN);
387 obj->flags |= HAS_OBJ;
388 return fsck_obj(obj);
391 static int default_refs;
393 static void fsck_handle_reflog_oid(const char *refname, struct object_id *oid,
394 timestamp_t timestamp)
396 struct object *obj;
398 if (!is_null_oid(oid)) {
399 obj = lookup_object(oid->hash);
400 if (obj && (obj->flags & HAS_OBJ)) {
401 if (timestamp && name_objects)
402 add_decoration(fsck_walk_options.object_names,
403 obj,
404 xstrfmt("%s@{%"PRItime"}", refname, timestamp));
405 obj->flags |= USED;
406 mark_object_reachable(obj);
407 } else {
408 error("%s: invalid reflog entry %s", refname, oid_to_hex(oid));
409 errors_found |= ERROR_REACHABLE;
414 static int fsck_handle_reflog_ent(struct object_id *ooid, struct object_id *noid,
415 const char *email, timestamp_t timestamp, int tz,
416 const char *message, void *cb_data)
418 const char *refname = cb_data;
420 if (verbose)
421 fprintf(stderr, "Checking reflog %s->%s\n",
422 oid_to_hex(ooid), oid_to_hex(noid));
424 fsck_handle_reflog_oid(refname, ooid, 0);
425 fsck_handle_reflog_oid(refname, noid, timestamp);
426 return 0;
429 static int fsck_handle_reflog(const char *logname, const struct object_id *oid,
430 int flag, void *cb_data)
432 for_each_reflog_ent(logname, fsck_handle_reflog_ent, (void *)logname);
433 return 0;
436 static int fsck_handle_ref(const char *refname, const struct object_id *oid,
437 int flag, void *cb_data)
439 struct object *obj;
441 obj = parse_object(oid);
442 if (!obj) {
443 error("%s: invalid sha1 pointer %s", refname, oid_to_hex(oid));
444 errors_found |= ERROR_REACHABLE;
445 /* We'll continue with the rest despite the error.. */
446 return 0;
448 if (obj->type != OBJ_COMMIT && is_branch(refname)) {
449 error("%s: not a commit", refname);
450 errors_found |= ERROR_REFS;
452 default_refs++;
453 obj->flags |= USED;
454 if (name_objects)
455 add_decoration(fsck_walk_options.object_names,
456 obj, xstrdup(refname));
457 mark_object_reachable(obj);
459 return 0;
462 static void get_default_heads(void)
464 if (head_points_at && !is_null_oid(&head_oid))
465 fsck_handle_ref("HEAD", &head_oid, 0, NULL);
466 for_each_rawref(fsck_handle_ref, NULL);
467 if (include_reflogs)
468 for_each_reflog(fsck_handle_reflog, NULL);
471 * Not having any default heads isn't really fatal, but
472 * it does mean that "--unreachable" no longer makes any
473 * sense (since in this case everything will obviously
474 * be unreachable by definition.
476 * Showing dangling objects is valid, though (as those
477 * dangling objects are likely lost heads).
479 * So we just print a warning about it, and clear the
480 * "show_unreachable" flag.
482 if (!default_refs) {
483 fprintf(stderr, "notice: No default references\n");
484 show_unreachable = 0;
488 static struct object *parse_loose_object(const struct object_id *oid,
489 const char *path)
491 struct object *obj;
492 void *contents;
493 enum object_type type;
494 unsigned long size;
495 int eaten;
497 if (read_loose_object(path, oid->hash, &type, &size, &contents) < 0)
498 return NULL;
500 if (!contents && type != OBJ_BLOB)
501 die("BUG: read_loose_object streamed a non-blob");
503 obj = parse_object_buffer(oid, type, size, contents, &eaten);
505 if (!eaten)
506 free(contents);
507 return obj;
510 static int fsck_loose(const struct object_id *oid, const char *path, void *data)
512 struct object *obj = parse_loose_object(oid, path);
514 if (!obj) {
515 errors_found |= ERROR_OBJECT;
516 error("%s: object corrupt or missing: %s",
517 oid_to_hex(oid), path);
518 return 0; /* keep checking other objects */
521 obj->flags &= ~(REACHABLE | SEEN);
522 obj->flags |= HAS_OBJ;
523 if (fsck_obj(obj))
524 errors_found |= ERROR_OBJECT;
525 return 0;
528 static int fsck_cruft(const char *basename, const char *path, void *data)
530 if (!starts_with(basename, "tmp_obj_"))
531 fprintf(stderr, "bad sha1 file: %s\n", path);
532 return 0;
535 static int fsck_subdir(unsigned int nr, const char *path, void *progress)
537 display_progress(progress, nr + 1);
538 return 0;
541 static void fsck_object_dir(const char *path)
543 struct progress *progress = NULL;
545 if (verbose)
546 fprintf(stderr, "Checking object directory\n");
548 if (show_progress)
549 progress = start_progress(_("Checking object directories"), 256);
551 for_each_loose_file_in_objdir(path, fsck_loose, fsck_cruft, fsck_subdir,
552 progress);
553 display_progress(progress, 256);
554 stop_progress(&progress);
557 static int fsck_head_link(void)
559 int null_is_error = 0;
561 if (verbose)
562 fprintf(stderr, "Checking HEAD link\n");
564 head_points_at = resolve_ref_unsafe("HEAD", 0, &head_oid, NULL);
565 if (!head_points_at) {
566 errors_found |= ERROR_REFS;
567 return error("Invalid HEAD");
569 if (!strcmp(head_points_at, "HEAD"))
570 /* detached HEAD */
571 null_is_error = 1;
572 else if (!starts_with(head_points_at, "refs/heads/")) {
573 errors_found |= ERROR_REFS;
574 return error("HEAD points to something strange (%s)",
575 head_points_at);
577 if (is_null_oid(&head_oid)) {
578 if (null_is_error) {
579 errors_found |= ERROR_REFS;
580 return error("HEAD: detached HEAD points at nothing");
582 fprintf(stderr, "notice: HEAD points to an unborn branch (%s)\n",
583 head_points_at + 11);
585 return 0;
588 static int fsck_cache_tree(struct cache_tree *it)
590 int i;
591 int err = 0;
593 if (verbose)
594 fprintf(stderr, "Checking cache tree\n");
596 if (0 <= it->entry_count) {
597 struct object *obj = parse_object(&it->oid);
598 if (!obj) {
599 error("%s: invalid sha1 pointer in cache-tree",
600 oid_to_hex(&it->oid));
601 errors_found |= ERROR_REFS;
602 return 1;
604 obj->flags |= USED;
605 if (name_objects)
606 add_decoration(fsck_walk_options.object_names,
607 obj, xstrdup(":"));
608 mark_object_reachable(obj);
609 if (obj->type != OBJ_TREE)
610 err |= objerror(obj, "non-tree in cache-tree");
612 for (i = 0; i < it->subtree_nr; i++)
613 err |= fsck_cache_tree(it->down[i]->cache_tree);
614 return err;
617 static void mark_object_for_connectivity(const struct object_id *oid)
619 struct object *obj = lookup_unknown_object(oid->hash);
620 obj->flags |= HAS_OBJ;
623 static int mark_loose_for_connectivity(const struct object_id *oid,
624 const char *path,
625 void *data)
627 mark_object_for_connectivity(oid);
628 return 0;
631 static int mark_packed_for_connectivity(const struct object_id *oid,
632 struct packed_git *pack,
633 uint32_t pos,
634 void *data)
636 mark_object_for_connectivity(oid);
637 return 0;
640 static char const * const fsck_usage[] = {
641 N_("git fsck [<options>] [<object>...]"),
642 NULL
645 static struct option fsck_opts[] = {
646 OPT__VERBOSE(&verbose, N_("be verbose")),
647 OPT_BOOL(0, "unreachable", &show_unreachable, N_("show unreachable objects")),
648 OPT_BOOL(0, "dangling", &show_dangling, N_("show dangling objects")),
649 OPT_BOOL(0, "tags", &show_tags, N_("report tags")),
650 OPT_BOOL(0, "root", &show_root, N_("report root nodes")),
651 OPT_BOOL(0, "cache", &keep_cache_objects, N_("make index objects head nodes")),
652 OPT_BOOL(0, "reflogs", &include_reflogs, N_("make reflogs head nodes (default)")),
653 OPT_BOOL(0, "full", &check_full, N_("also consider packs and alternate objects")),
654 OPT_BOOL(0, "connectivity-only", &connectivity_only, N_("check only connectivity")),
655 OPT_BOOL(0, "strict", &check_strict, N_("enable more strict checking")),
656 OPT_BOOL(0, "lost-found", &write_lost_and_found,
657 N_("write dangling objects in .git/lost-found")),
658 OPT_BOOL(0, "progress", &show_progress, N_("show progress")),
659 OPT_BOOL(0, "name-objects", &name_objects, N_("show verbose names for reachable objects")),
660 OPT_END(),
663 int cmd_fsck(int argc, const char **argv, const char *prefix)
665 int i;
666 struct alternate_object_database *alt;
668 errors_found = 0;
669 check_replace_refs = 0;
671 argc = parse_options(argc, argv, prefix, fsck_opts, fsck_usage, 0);
673 fsck_walk_options.walk = mark_object;
674 fsck_obj_options.walk = mark_used;
675 fsck_obj_options.error_func = fsck_error_func;
676 if (check_strict)
677 fsck_obj_options.strict = 1;
679 if (show_progress == -1)
680 show_progress = isatty(2);
681 if (verbose)
682 show_progress = 0;
684 if (write_lost_and_found) {
685 check_full = 1;
686 include_reflogs = 0;
689 if (name_objects)
690 fsck_walk_options.object_names =
691 xcalloc(1, sizeof(struct decoration));
693 git_config(fsck_config, NULL);
695 fsck_head_link();
696 if (connectivity_only) {
697 for_each_loose_object(mark_loose_for_connectivity, NULL, 0);
698 for_each_packed_object(mark_packed_for_connectivity, NULL, 0);
699 } else {
700 fsck_object_dir(get_object_directory());
702 prepare_alt_odb();
703 for (alt = alt_odb_list; alt; alt = alt->next)
704 fsck_object_dir(alt->path);
706 if (check_full) {
707 struct packed_git *p;
708 uint32_t total = 0, count = 0;
709 struct progress *progress = NULL;
711 prepare_packed_git();
713 if (show_progress) {
714 for (p = packed_git; p; p = p->next) {
715 if (open_pack_index(p))
716 continue;
717 total += p->num_objects;
720 progress = start_progress(_("Checking objects"), total);
722 for (p = packed_git; p; p = p->next) {
723 /* verify gives error messages itself */
724 if (verify_pack(p, fsck_obj_buffer,
725 progress, count))
726 errors_found |= ERROR_PACK;
727 count += p->num_objects;
729 stop_progress(&progress);
733 for (i = 0; i < argc; i++) {
734 const char *arg = argv[i];
735 struct object_id oid;
736 if (!get_oid(arg, &oid)) {
737 struct object *obj = lookup_object(oid.hash);
739 if (!obj || !(obj->flags & HAS_OBJ)) {
740 error("%s: object missing", oid_to_hex(&oid));
741 errors_found |= ERROR_OBJECT;
742 continue;
745 obj->flags |= USED;
746 if (name_objects)
747 add_decoration(fsck_walk_options.object_names,
748 obj, xstrdup(arg));
749 mark_object_reachable(obj);
750 continue;
752 error("invalid parameter: expected sha1, got '%s'", arg);
753 errors_found |= ERROR_OBJECT;
757 * If we've not been given any explicit head information, do the
758 * default ones from .git/refs. We also consider the index file
759 * in this case (ie this implies --cache).
761 if (!argc) {
762 get_default_heads();
763 keep_cache_objects = 1;
766 if (keep_cache_objects) {
767 verify_index_checksum = 1;
768 verify_ce_order = 1;
769 read_cache();
770 for (i = 0; i < active_nr; i++) {
771 unsigned int mode;
772 struct blob *blob;
773 struct object *obj;
775 mode = active_cache[i]->ce_mode;
776 if (S_ISGITLINK(mode))
777 continue;
778 blob = lookup_blob(&active_cache[i]->oid);
779 if (!blob)
780 continue;
781 obj = &blob->object;
782 obj->flags |= USED;
783 if (name_objects)
784 add_decoration(fsck_walk_options.object_names,
785 obj,
786 xstrfmt(":%s", active_cache[i]->name));
787 mark_object_reachable(obj);
789 if (active_cache_tree)
790 fsck_cache_tree(active_cache_tree);
793 check_connectivity();
794 return errors_found;