9 #include "cache-tree.h"
10 #include "tree-walk.h"
12 #include "parse-options.h"
15 #include "streaming.h"
18 #define REACHABLE 0x0001
20 #define HAS_OBJ 0x0004
24 static int show_unreachable;
25 static int include_reflogs = 1;
26 static int check_full = 1;
27 static int connectivity_only;
28 static int check_strict;
29 static int keep_cache_objects;
30 static struct fsck_options fsck_walk_options = FSCK_OPTIONS_DEFAULT;
31 static struct fsck_options fsck_obj_options = FSCK_OPTIONS_DEFAULT;
32 static struct object_id head_oid;
33 static const char *head_points_at;
34 static int errors_found;
35 static int write_lost_and_found;
37 static int show_progress = -1;
38 static int show_dangling = 1;
39 static int name_objects;
40 #define ERROR_OBJECT 01
41 #define ERROR_REACHABLE 02
43 #define ERROR_REFS 010
45 static const char *describe_object(struct object *obj)
47 static struct strbuf buf = STRBUF_INIT;
48 char *name = name_objects ?
49 lookup_decoration(fsck_walk_options.object_names, obj) : NULL;
52 strbuf_addstr(&buf, oid_to_hex(&obj->oid));
54 strbuf_addf(&buf, " (%s)", name);
59 static int fsck_config(const char *var, const char *value, void *cb)
61 if (strcmp(var, "fsck.skiplist") == 0) {
63 struct strbuf sb = STRBUF_INIT;
65 if (git_config_pathname(&path, var, value))
67 strbuf_addf(&sb, "skiplist=%s", path);
69 fsck_set_msg_types(&fsck_obj_options, sb.buf);
74 if (skip_prefix(var, "fsck.", &var)) {
75 fsck_set_msg_type(&fsck_obj_options, var, value);
79 return git_default_config(var, value, cb);
82 static void objreport(struct object *obj, const char *msg_type,
85 fprintf(stderr, "%s in %s %s: %s\n",
86 msg_type, typename(obj->type), describe_object(obj), err);
89 static int objerror(struct object *obj, const char *err)
91 errors_found |= ERROR_OBJECT;
92 objreport(obj, "error", err);
96 static int fsck_error_func(struct fsck_options *o,
97 struct object *obj, int type, const char *message)
99 objreport(obj, (type == FSCK_WARN) ? "warning" : "error", message);
100 return (type == FSCK_WARN) ? 0 : 1;
103 static struct object_array pending;
105 static int mark_object(struct object *obj, int type, void *data, struct fsck_options *options)
107 struct object *parent = data;
110 * The only case data is NULL or type is OBJ_ANY is when
111 * mark_object_reachable() calls us. All the callers of
112 * that function has non-NULL obj hence ...
115 /* ... these references to parent->fld are safe here */
116 printf("broken link from %7s %s\n",
117 typename(parent->type), describe_object(parent));
118 printf("broken link from %7s %s\n",
119 (type == OBJ_ANY ? "unknown" : typename(type)), "unknown");
120 errors_found |= ERROR_REACHABLE;
124 if (type != OBJ_ANY && obj->type != type)
125 /* ... and the reference to parent is safe here */
126 objerror(parent, "wrong object type in link");
128 if (obj->flags & REACHABLE)
130 obj->flags |= REACHABLE;
131 if (!(obj->flags & HAS_OBJ)) {
132 if (parent && !has_object_file(&obj->oid)) {
133 printf("broken link from %7s %s\n",
134 typename(parent->type), describe_object(parent));
135 printf(" to %7s %s\n",
136 typename(obj->type), describe_object(obj));
137 errors_found |= ERROR_REACHABLE;
142 add_object_array(obj, NULL, &pending);
146 static void mark_object_reachable(struct object *obj)
148 mark_object(obj, OBJ_ANY, NULL, NULL);
151 static int traverse_one_object(struct object *obj)
154 struct tree *tree = NULL;
156 if (obj->type == OBJ_TREE) {
157 tree = (struct tree *)obj;
158 if (parse_tree(tree) < 0)
159 return 1; /* error already displayed */
161 result = fsck_walk(obj, obj, &fsck_walk_options);
163 free_tree_buffer(tree);
167 static int traverse_reachable(void)
169 struct progress *progress = NULL;
173 progress = start_progress_delay(_("Checking connectivity"), 0, 0, 2);
175 struct object_array_entry *entry;
178 entry = pending.objects + --pending.nr;
180 result |= traverse_one_object(obj);
181 display_progress(progress, ++nr);
183 stop_progress(&progress);
187 static int mark_used(struct object *obj, int type, void *data, struct fsck_options *options)
196 * Check a single reachable object
198 static void check_reachable_object(struct object *obj)
201 * We obviously want the object to be parsed,
202 * except if it was in a pack-file and we didn't
205 if (!(obj->flags & HAS_OBJ)) {
206 if (has_sha1_pack(obj->oid.hash))
207 return; /* it is in pack - forget about it */
208 if (connectivity_only && has_object_file(&obj->oid))
210 printf("missing %s %s\n", typename(obj->type),
211 describe_object(obj));
212 errors_found |= ERROR_REACHABLE;
218 * Check a single unreachable object
220 static void check_unreachable_object(struct object *obj)
223 * Missing unreachable object? Ignore it. It's not like
224 * we miss it (since it can't be reached), nor do we want
225 * to complain about it being unreachable (since it does
232 * Unreachable object that exists? Show it if asked to,
233 * since this is something that is prunable.
235 if (show_unreachable) {
236 printf("unreachable %s %s\n", typename(obj->type),
237 describe_object(obj));
242 * "!used" means that nothing at all points to it, including
243 * other unreachable objects. In other words, it's the "tip"
244 * of some set of unreachable objects, usually a commit that
247 * Such starting points are more interesting than some random
248 * set of unreachable objects, so we show them even if the user
249 * hasn't asked for _all_ unreachable objects. If you have
250 * deleted a branch by mistake, this is a prime candidate to
251 * start looking at, for example.
255 printf("dangling %s %s\n", typename(obj->type),
256 describe_object(obj));
257 if (write_lost_and_found) {
258 char *filename = git_pathdup("lost-found/%s/%s",
259 obj->type == OBJ_COMMIT ? "commit" : "other",
260 describe_object(obj));
263 if (safe_create_leading_directories_const(filename)) {
264 error("Could not create lost-found");
268 if (!(f = fopen(filename, "w")))
269 die_errno("Could not open '%s'", filename);
270 if (obj->type == OBJ_BLOB) {
271 if (stream_blob_to_fd(fileno(f), &obj->oid, NULL, 1))
272 die_errno("Could not write '%s'", filename);
274 fprintf(f, "%s\n", describe_object(obj));
276 die_errno("Could not finish '%s'",
284 * Otherwise? It's there, it's unreachable, and some other unreachable
285 * object points to it. Ignore it - it's not interesting, and we showed
286 * all the interesting cases above.
290 static void check_object(struct object *obj)
293 fprintf(stderr, "Checking %s\n", describe_object(obj));
295 if (obj->flags & REACHABLE)
296 check_reachable_object(obj);
298 check_unreachable_object(obj);
301 static void check_connectivity(void)
305 /* Traverse the pending reachable objects */
306 traverse_reachable();
308 /* Look up all the requirements, warn about missing objects.. */
309 max = get_max_object_index();
311 fprintf(stderr, "Checking connectivity (%d objects)\n", max);
313 for (i = 0; i < max; i++) {
314 struct object *obj = get_indexed_object(i);
321 static int fsck_obj(struct object *obj)
323 if (obj->flags & SEEN)
328 fprintf(stderr, "Checking %s %s\n",
329 typename(obj->type), describe_object(obj));
331 if (fsck_walk(obj, NULL, &fsck_obj_options))
332 objerror(obj, "broken links");
333 if (fsck_object(obj, NULL, 0, &fsck_obj_options))
336 if (obj->type == OBJ_TREE) {
337 struct tree *item = (struct tree *) obj;
339 free_tree_buffer(item);
342 if (obj->type == OBJ_COMMIT) {
343 struct commit *commit = (struct commit *) obj;
345 free_commit_buffer(commit);
347 if (!commit->parents && show_root)
348 printf("root %s\n", describe_object(&commit->object));
351 if (obj->type == OBJ_TAG) {
352 struct tag *tag = (struct tag *) obj;
354 if (show_tags && tag->tagged) {
355 printf("tagged %s %s", typename(tag->tagged->type),
356 describe_object(tag->tagged));
357 printf(" (%s) in %s\n", tag->tag,
358 describe_object(&tag->object));
365 static int fsck_sha1(const unsigned char *sha1)
367 struct object *obj = parse_object(sha1);
369 errors_found |= ERROR_OBJECT;
370 return error("%s: object corrupt or missing",
373 obj->flags |= HAS_OBJ;
374 return fsck_obj(obj);
377 static int fsck_obj_buffer(const unsigned char *sha1, enum object_type type,
378 unsigned long size, void *buffer, int *eaten)
381 * Note, buffer may be NULL if type is OBJ_BLOB. See
382 * verify_packfile(), data_valid variable for details.
385 obj = parse_object_buffer(sha1, type, size, buffer, eaten);
387 errors_found |= ERROR_OBJECT;
388 return error("%s: object corrupt or missing", sha1_to_hex(sha1));
390 obj->flags = HAS_OBJ;
391 return fsck_obj(obj);
394 static int default_refs;
396 static void fsck_handle_reflog_sha1(const char *refname, unsigned char *sha1,
397 unsigned long timestamp)
401 if (!is_null_sha1(sha1)) {
402 obj = lookup_object(sha1);
404 if (timestamp && name_objects)
405 add_decoration(fsck_walk_options.object_names,
407 xstrfmt("%s@{%ld}", refname, timestamp));
409 mark_object_reachable(obj);
411 error("%s: invalid reflog entry %s", refname, sha1_to_hex(sha1));
412 errors_found |= ERROR_REACHABLE;
417 static int fsck_handle_reflog_ent(unsigned char *osha1, unsigned char *nsha1,
418 const char *email, unsigned long timestamp, int tz,
419 const char *message, void *cb_data)
421 const char *refname = cb_data;
424 fprintf(stderr, "Checking reflog %s->%s\n",
425 sha1_to_hex(osha1), sha1_to_hex(nsha1));
427 fsck_handle_reflog_sha1(refname, osha1, 0);
428 fsck_handle_reflog_sha1(refname, nsha1, timestamp);
432 static int fsck_handle_reflog(const char *logname, const struct object_id *oid,
433 int flag, void *cb_data)
435 for_each_reflog_ent(logname, fsck_handle_reflog_ent, (void *)logname);
439 static int fsck_handle_ref(const char *refname, const struct object_id *oid,
440 int flag, void *cb_data)
444 obj = parse_object(oid->hash);
446 error("%s: invalid sha1 pointer %s", refname, oid_to_hex(oid));
447 errors_found |= ERROR_REACHABLE;
448 /* We'll continue with the rest despite the error.. */
451 if (obj->type != OBJ_COMMIT && is_branch(refname)) {
452 error("%s: not a commit", refname);
453 errors_found |= ERROR_REFS;
458 add_decoration(fsck_walk_options.object_names,
459 obj, xstrdup(refname));
460 mark_object_reachable(obj);
465 static void get_default_heads(void)
467 if (head_points_at && !is_null_oid(&head_oid))
468 fsck_handle_ref("HEAD", &head_oid, 0, NULL);
469 for_each_rawref(fsck_handle_ref, NULL);
471 for_each_reflog(fsck_handle_reflog, NULL);
474 * Not having any default heads isn't really fatal, but
475 * it does mean that "--unreachable" no longer makes any
476 * sense (since in this case everything will obviously
477 * be unreachable by definition.
479 * Showing dangling objects is valid, though (as those
480 * dangling objects are likely lost heads).
482 * So we just print a warning about it, and clear the
483 * "show_unreachable" flag.
486 fprintf(stderr, "notice: No default references\n");
487 show_unreachable = 0;
491 static int fsck_loose(const unsigned char *sha1, const char *path, void *data)
494 errors_found |= ERROR_OBJECT;
498 static int fsck_cruft(const char *basename, const char *path, void *data)
500 if (!starts_with(basename, "tmp_obj_"))
501 fprintf(stderr, "bad sha1 file: %s\n", path);
505 static int fsck_subdir(int nr, const char *path, void *progress)
507 display_progress(progress, nr + 1);
511 static void fsck_object_dir(const char *path)
513 struct progress *progress = NULL;
516 fprintf(stderr, "Checking object directory\n");
519 progress = start_progress(_("Checking object directories"), 256);
521 for_each_loose_file_in_objdir(path, fsck_loose, fsck_cruft, fsck_subdir,
523 display_progress(progress, 256);
524 stop_progress(&progress);
527 static int fsck_head_link(void)
529 int null_is_error = 0;
532 fprintf(stderr, "Checking HEAD link\n");
534 head_points_at = resolve_ref_unsafe("HEAD", 0, head_oid.hash, NULL);
535 if (!head_points_at) {
536 errors_found |= ERROR_REFS;
537 return error("Invalid HEAD");
539 if (!strcmp(head_points_at, "HEAD"))
542 else if (!starts_with(head_points_at, "refs/heads/")) {
543 errors_found |= ERROR_REFS;
544 return error("HEAD points to something strange (%s)",
547 if (is_null_oid(&head_oid)) {
549 errors_found |= ERROR_REFS;
550 return error("HEAD: detached HEAD points at nothing");
552 fprintf(stderr, "notice: HEAD points to an unborn branch (%s)\n",
553 head_points_at + 11);
558 static int fsck_cache_tree(struct cache_tree *it)
564 fprintf(stderr, "Checking cache tree\n");
566 if (0 <= it->entry_count) {
567 struct object *obj = parse_object(it->sha1);
569 error("%s: invalid sha1 pointer in cache-tree",
570 sha1_to_hex(it->sha1));
571 errors_found |= ERROR_REFS;
576 add_decoration(fsck_walk_options.object_names,
578 mark_object_reachable(obj);
579 if (obj->type != OBJ_TREE)
580 err |= objerror(obj, "non-tree in cache-tree");
582 for (i = 0; i < it->subtree_nr; i++)
583 err |= fsck_cache_tree(it->down[i]->cache_tree);
587 static char const * const fsck_usage[] = {
588 N_("git fsck [<options>] [<object>...]"),
592 static struct option fsck_opts[] = {
593 OPT__VERBOSE(&verbose, N_("be verbose")),
594 OPT_BOOL(0, "unreachable", &show_unreachable, N_("show unreachable objects")),
595 OPT_BOOL(0, "dangling", &show_dangling, N_("show dangling objects")),
596 OPT_BOOL(0, "tags", &show_tags, N_("report tags")),
597 OPT_BOOL(0, "root", &show_root, N_("report root nodes")),
598 OPT_BOOL(0, "cache", &keep_cache_objects, N_("make index objects head nodes")),
599 OPT_BOOL(0, "reflogs", &include_reflogs, N_("make reflogs head nodes (default)")),
600 OPT_BOOL(0, "full", &check_full, N_("also consider packs and alternate objects")),
601 OPT_BOOL(0, "connectivity-only", &connectivity_only, N_("check only connectivity")),
602 OPT_BOOL(0, "strict", &check_strict, N_("enable more strict checking")),
603 OPT_BOOL(0, "lost-found", &write_lost_and_found,
604 N_("write dangling objects in .git/lost-found")),
605 OPT_BOOL(0, "progress", &show_progress, N_("show progress")),
606 OPT_BOOL(0, "name-objects", &name_objects, N_("show verbose names for reachable objects")),
610 int cmd_fsck(int argc, const char **argv, const char *prefix)
613 struct alternate_object_database *alt;
616 check_replace_refs = 0;
618 argc = parse_options(argc, argv, prefix, fsck_opts, fsck_usage, 0);
620 fsck_walk_options.walk = mark_object;
621 fsck_obj_options.walk = mark_used;
622 fsck_obj_options.error_func = fsck_error_func;
624 fsck_obj_options.strict = 1;
626 if (show_progress == -1)
627 show_progress = isatty(2);
631 if (write_lost_and_found) {
637 fsck_walk_options.object_names =
638 xcalloc(1, sizeof(struct decoration));
640 git_config(fsck_config, NULL);
643 if (!connectivity_only) {
644 fsck_object_dir(get_object_directory());
647 for (alt = alt_odb_list; alt; alt = alt->next)
648 fsck_object_dir(alt->path);
652 struct packed_git *p;
653 uint32_t total = 0, count = 0;
654 struct progress *progress = NULL;
656 prepare_packed_git();
659 for (p = packed_git; p; p = p->next) {
660 if (open_pack_index(p))
662 total += p->num_objects;
665 progress = start_progress(_("Checking objects"), total);
667 for (p = packed_git; p; p = p->next) {
668 /* verify gives error messages itself */
669 if (verify_pack(p, fsck_obj_buffer,
671 errors_found |= ERROR_PACK;
672 count += p->num_objects;
674 stop_progress(&progress);
678 for (i = 0; i < argc; i++) {
679 const char *arg = argv[i];
680 unsigned char sha1[20];
681 if (!get_sha1(arg, sha1)) {
682 struct object *obj = lookup_object(sha1);
684 /* Error is printed by lookup_object(). */
690 add_decoration(fsck_walk_options.object_names,
692 mark_object_reachable(obj);
696 error("invalid parameter: expected sha1, got '%s'", arg);
700 * If we've not been given any explicit head information, do the
701 * default ones from .git/refs. We also consider the index file
702 * in this case (ie this implies --cache).
706 keep_cache_objects = 1;
709 if (keep_cache_objects) {
711 for (i = 0; i < active_nr; i++) {
716 mode = active_cache[i]->ce_mode;
717 if (S_ISGITLINK(mode))
719 blob = lookup_blob(active_cache[i]->oid.hash);
725 add_decoration(fsck_walk_options.object_names,
727 xstrfmt(":%s", active_cache[i]->name));
728 mark_object_reachable(obj);
730 if (active_cache_tree)
731 fsck_cache_tree(active_cache_tree);
734 check_connectivity();