t5520: verify that `pull --rebase` shows the helpful advice when failing
[git/raj.git] / builtin / fsck.c
blobc6d17e63fd832d5b37ea93e523cfb774b563c452
1 #include "builtin.h"
2 #include "cache.h"
3 #include "commit.h"
4 #include "tree.h"
5 #include "blob.h"
6 #include "tag.h"
7 #include "refs.h"
8 #include "pack.h"
9 #include "cache-tree.h"
10 #include "tree-walk.h"
11 #include "fsck.h"
12 #include "parse-options.h"
13 #include "dir.h"
14 #include "progress.h"
15 #include "streaming.h"
16 #include "decorate.h"
18 #define REACHABLE 0x0001
19 #define SEEN 0x0002
20 #define HAS_OBJ 0x0004
22 static int show_root;
23 static int show_tags;
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;
36 static int verbose;
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
42 #define ERROR_PACK 04
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;
51 strbuf_reset(&buf);
52 strbuf_addstr(&buf, oid_to_hex(&obj->oid));
53 if (name)
54 strbuf_addf(&buf, " (%s)", name);
56 return buf.buf;
59 static int fsck_config(const char *var, const char *value, void *cb)
61 if (strcmp(var, "fsck.skiplist") == 0) {
62 const char *path;
63 struct strbuf sb = STRBUF_INIT;
65 if (git_config_pathname(&path, var, value))
66 return 1;
67 strbuf_addf(&sb, "skiplist=%s", path);
68 free((char *)path);
69 fsck_set_msg_types(&fsck_obj_options, sb.buf);
70 strbuf_release(&sb);
71 return 0;
74 if (skip_prefix(var, "fsck.", &var)) {
75 fsck_set_msg_type(&fsck_obj_options, var, value);
76 return 0;
79 return git_default_config(var, value, cb);
82 static void objreport(struct object *obj, const char *msg_type,
83 const char *err)
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);
93 return -1;
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 ...
114 if (!obj) {
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;
121 return 1;
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)
129 return 0;
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;
139 return 1;
142 add_object_array(obj, NULL, &pending);
143 return 0;
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)
153 int result;
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);
162 if (tree)
163 free_tree_buffer(tree);
164 return result;
167 static int traverse_reachable(void)
169 struct progress *progress = NULL;
170 unsigned int nr = 0;
171 int result = 0;
172 if (show_progress)
173 progress = start_progress_delay(_("Checking connectivity"), 0, 0, 2);
174 while (pending.nr) {
175 struct object_array_entry *entry;
176 struct object *obj;
178 entry = pending.objects + --pending.nr;
179 obj = entry->item;
180 result |= traverse_one_object(obj);
181 display_progress(progress, ++nr);
183 stop_progress(&progress);
184 return !!result;
187 static int mark_used(struct object *obj, int type, void *data, struct fsck_options *options)
189 if (!obj)
190 return 1;
191 obj->used = 1;
192 return 0;
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
203 * do a full fsck
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))
209 return;
210 printf("missing %s %s\n", typename(obj->type),
211 describe_object(obj));
212 errors_found |= ERROR_REACHABLE;
213 return;
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
226 * not exist).
228 if (!obj->parsed)
229 return;
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));
238 return;
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
245 * got dropped.
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.
253 if (!obj->used) {
254 if (show_dangling)
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));
261 FILE *f;
263 if (safe_create_leading_directories_const(filename)) {
264 error("Could not create lost-found");
265 free(filename);
266 return;
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.hash, NULL, 1))
272 die_errno("Could not write '%s'", filename);
273 } else
274 fprintf(f, "%s\n", describe_object(obj));
275 if (fclose(f))
276 die_errno("Could not finish '%s'",
277 filename);
278 free(filename);
280 return;
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)
292 if (verbose)
293 fprintf(stderr, "Checking %s\n", describe_object(obj));
295 if (obj->flags & REACHABLE)
296 check_reachable_object(obj);
297 else
298 check_unreachable_object(obj);
301 static void check_connectivity(void)
303 int i, max;
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();
310 if (verbose)
311 fprintf(stderr, "Checking connectivity (%d objects)\n", max);
313 for (i = 0; i < max; i++) {
314 struct object *obj = get_indexed_object(i);
316 if (obj)
317 check_object(obj);
321 static int fsck_obj(struct object *obj)
323 if (obj->flags & SEEN)
324 return 0;
325 obj->flags |= SEEN;
327 if (verbose)
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))
334 return -1;
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));
362 return 0;
365 static int fsck_sha1(const unsigned char *sha1)
367 struct object *obj = parse_object(sha1);
368 if (!obj) {
369 errors_found |= ERROR_OBJECT;
370 return error("%s: object corrupt or missing",
371 sha1_to_hex(sha1));
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)
380 struct object *obj;
381 obj = parse_object_buffer(sha1, type, size, buffer, eaten);
382 if (!obj) {
383 errors_found |= ERROR_OBJECT;
384 return error("%s: object corrupt or missing", sha1_to_hex(sha1));
386 obj->flags = HAS_OBJ;
387 return fsck_obj(obj);
390 static int default_refs;
392 static void fsck_handle_reflog_sha1(const char *refname, unsigned char *sha1,
393 unsigned long timestamp)
395 struct object *obj;
397 if (!is_null_sha1(sha1)) {
398 obj = lookup_object(sha1);
399 if (obj) {
400 if (timestamp && name_objects)
401 add_decoration(fsck_walk_options.object_names,
402 obj,
403 xstrfmt("%s@{%ld}", refname, timestamp));
404 obj->used = 1;
405 mark_object_reachable(obj);
406 } else {
407 error("%s: invalid reflog entry %s", refname, sha1_to_hex(sha1));
408 errors_found |= ERROR_REACHABLE;
413 static int fsck_handle_reflog_ent(unsigned char *osha1, unsigned char *nsha1,
414 const char *email, unsigned long timestamp, int tz,
415 const char *message, void *cb_data)
417 const char *refname = cb_data;
419 if (verbose)
420 fprintf(stderr, "Checking reflog %s->%s\n",
421 sha1_to_hex(osha1), sha1_to_hex(nsha1));
423 fsck_handle_reflog_sha1(refname, osha1, 0);
424 fsck_handle_reflog_sha1(refname, nsha1, timestamp);
425 return 0;
428 static int fsck_handle_reflog(const char *logname, const struct object_id *oid,
429 int flag, void *cb_data)
431 for_each_reflog_ent(logname, fsck_handle_reflog_ent, (void *)logname);
432 return 0;
435 static int fsck_handle_ref(const char *refname, const struct object_id *oid,
436 int flag, void *cb_data)
438 struct object *obj;
440 obj = parse_object(oid->hash);
441 if (!obj) {
442 error("%s: invalid sha1 pointer %s", refname, oid_to_hex(oid));
443 errors_found |= ERROR_REACHABLE;
444 /* We'll continue with the rest despite the error.. */
445 return 0;
447 if (obj->type != OBJ_COMMIT && is_branch(refname)) {
448 error("%s: not a commit", refname);
449 errors_found |= ERROR_REFS;
451 default_refs++;
452 obj->used = 1;
453 if (name_objects)
454 add_decoration(fsck_walk_options.object_names,
455 obj, xstrdup(refname));
456 mark_object_reachable(obj);
458 return 0;
461 static void get_default_heads(void)
463 if (head_points_at && !is_null_oid(&head_oid))
464 fsck_handle_ref("HEAD", &head_oid, 0, NULL);
465 for_each_rawref(fsck_handle_ref, NULL);
466 if (include_reflogs)
467 for_each_reflog(fsck_handle_reflog, NULL);
470 * Not having any default heads isn't really fatal, but
471 * it does mean that "--unreachable" no longer makes any
472 * sense (since in this case everything will obviously
473 * be unreachable by definition.
475 * Showing dangling objects is valid, though (as those
476 * dangling objects are likely lost heads).
478 * So we just print a warning about it, and clear the
479 * "show_unreachable" flag.
481 if (!default_refs) {
482 fprintf(stderr, "notice: No default references\n");
483 show_unreachable = 0;
487 static int fsck_loose(const unsigned char *sha1, const char *path, void *data)
489 if (fsck_sha1(sha1))
490 errors_found |= ERROR_OBJECT;
491 return 0;
494 static int fsck_cruft(const char *basename, const char *path, void *data)
496 if (!starts_with(basename, "tmp_obj_"))
497 fprintf(stderr, "bad sha1 file: %s\n", path);
498 return 0;
501 static int fsck_subdir(int nr, const char *path, void *progress)
503 display_progress(progress, nr + 1);
504 return 0;
507 static void fsck_object_dir(const char *path)
509 struct progress *progress = NULL;
511 if (verbose)
512 fprintf(stderr, "Checking object directory\n");
514 if (show_progress)
515 progress = start_progress(_("Checking object directories"), 256);
517 for_each_loose_file_in_objdir(path, fsck_loose, fsck_cruft, fsck_subdir,
518 progress);
519 display_progress(progress, 256);
520 stop_progress(&progress);
523 static int fsck_head_link(void)
525 int null_is_error = 0;
527 if (verbose)
528 fprintf(stderr, "Checking HEAD link\n");
530 head_points_at = resolve_ref_unsafe("HEAD", 0, head_oid.hash, NULL);
531 if (!head_points_at) {
532 errors_found |= ERROR_REFS;
533 return error("Invalid HEAD");
535 if (!strcmp(head_points_at, "HEAD"))
536 /* detached HEAD */
537 null_is_error = 1;
538 else if (!starts_with(head_points_at, "refs/heads/")) {
539 errors_found |= ERROR_REFS;
540 return error("HEAD points to something strange (%s)",
541 head_points_at);
543 if (is_null_oid(&head_oid)) {
544 if (null_is_error) {
545 errors_found |= ERROR_REFS;
546 return error("HEAD: detached HEAD points at nothing");
548 fprintf(stderr, "notice: HEAD points to an unborn branch (%s)\n",
549 head_points_at + 11);
551 return 0;
554 static int fsck_cache_tree(struct cache_tree *it)
556 int i;
557 int err = 0;
559 if (verbose)
560 fprintf(stderr, "Checking cache tree\n");
562 if (0 <= it->entry_count) {
563 struct object *obj = parse_object(it->sha1);
564 if (!obj) {
565 error("%s: invalid sha1 pointer in cache-tree",
566 sha1_to_hex(it->sha1));
567 errors_found |= ERROR_REFS;
568 return 1;
570 obj->used = 1;
571 if (name_objects)
572 add_decoration(fsck_walk_options.object_names,
573 obj, xstrdup(":"));
574 mark_object_reachable(obj);
575 if (obj->type != OBJ_TREE)
576 err |= objerror(obj, "non-tree in cache-tree");
578 for (i = 0; i < it->subtree_nr; i++)
579 err |= fsck_cache_tree(it->down[i]->cache_tree);
580 return err;
583 static char const * const fsck_usage[] = {
584 N_("git fsck [<options>] [<object>...]"),
585 NULL
588 static struct option fsck_opts[] = {
589 OPT__VERBOSE(&verbose, N_("be verbose")),
590 OPT_BOOL(0, "unreachable", &show_unreachable, N_("show unreachable objects")),
591 OPT_BOOL(0, "dangling", &show_dangling, N_("show dangling objects")),
592 OPT_BOOL(0, "tags", &show_tags, N_("report tags")),
593 OPT_BOOL(0, "root", &show_root, N_("report root nodes")),
594 OPT_BOOL(0, "cache", &keep_cache_objects, N_("make index objects head nodes")),
595 OPT_BOOL(0, "reflogs", &include_reflogs, N_("make reflogs head nodes (default)")),
596 OPT_BOOL(0, "full", &check_full, N_("also consider packs and alternate objects")),
597 OPT_BOOL(0, "connectivity-only", &connectivity_only, N_("check only connectivity")),
598 OPT_BOOL(0, "strict", &check_strict, N_("enable more strict checking")),
599 OPT_BOOL(0, "lost-found", &write_lost_and_found,
600 N_("write dangling objects in .git/lost-found")),
601 OPT_BOOL(0, "progress", &show_progress, N_("show progress")),
602 OPT_BOOL(0, "name-objects", &name_objects, N_("show verbose names for reachable objects")),
603 OPT_END(),
606 int cmd_fsck(int argc, const char **argv, const char *prefix)
608 int i, heads;
609 struct alternate_object_database *alt;
611 errors_found = 0;
612 check_replace_refs = 0;
614 argc = parse_options(argc, argv, prefix, fsck_opts, fsck_usage, 0);
616 fsck_walk_options.walk = mark_object;
617 fsck_obj_options.walk = mark_used;
618 fsck_obj_options.error_func = fsck_error_func;
619 if (check_strict)
620 fsck_obj_options.strict = 1;
622 if (show_progress == -1)
623 show_progress = isatty(2);
624 if (verbose)
625 show_progress = 0;
627 if (write_lost_and_found) {
628 check_full = 1;
629 include_reflogs = 0;
632 if (name_objects)
633 fsck_walk_options.object_names =
634 xcalloc(1, sizeof(struct decoration));
636 git_config(fsck_config, NULL);
638 fsck_head_link();
639 if (!connectivity_only) {
640 fsck_object_dir(get_object_directory());
642 prepare_alt_odb();
643 for (alt = alt_odb_list; alt; alt = alt->next) {
644 /* directory name, minus trailing slash */
645 size_t namelen = alt->name - alt->base - 1;
646 struct strbuf name = STRBUF_INIT;
647 strbuf_add(&name, alt->base, namelen);
648 fsck_object_dir(name.buf);
649 strbuf_release(&name);
653 if (check_full) {
654 struct packed_git *p;
655 uint32_t total = 0, count = 0;
656 struct progress *progress = NULL;
658 prepare_packed_git();
660 if (show_progress) {
661 for (p = packed_git; p; p = p->next) {
662 if (open_pack_index(p))
663 continue;
664 total += p->num_objects;
667 progress = start_progress(_("Checking objects"), total);
669 for (p = packed_git; p; p = p->next) {
670 /* verify gives error messages itself */
671 if (verify_pack(p, fsck_obj_buffer,
672 progress, count))
673 errors_found |= ERROR_PACK;
674 count += p->num_objects;
676 stop_progress(&progress);
679 heads = 0;
680 for (i = 0; i < argc; i++) {
681 const char *arg = argv[i];
682 unsigned char sha1[20];
683 if (!get_sha1(arg, sha1)) {
684 struct object *obj = lookup_object(sha1);
686 /* Error is printed by lookup_object(). */
687 if (!obj)
688 continue;
690 obj->used = 1;
691 if (name_objects)
692 add_decoration(fsck_walk_options.object_names,
693 obj, xstrdup(arg));
694 mark_object_reachable(obj);
695 heads++;
696 continue;
698 error("invalid parameter: expected sha1, got '%s'", arg);
702 * If we've not been given any explicit head information, do the
703 * default ones from .git/refs. We also consider the index file
704 * in this case (ie this implies --cache).
706 if (!heads) {
707 get_default_heads();
708 keep_cache_objects = 1;
711 if (keep_cache_objects) {
712 read_cache();
713 for (i = 0; i < active_nr; i++) {
714 unsigned int mode;
715 struct blob *blob;
716 struct object *obj;
718 mode = active_cache[i]->ce_mode;
719 if (S_ISGITLINK(mode))
720 continue;
721 blob = lookup_blob(active_cache[i]->sha1);
722 if (!blob)
723 continue;
724 obj = &blob->object;
725 obj->used = 1;
726 if (name_objects)
727 add_decoration(fsck_walk_options.object_names,
728 obj,
729 xstrfmt(":%s", active_cache[i]->name));
730 mark_object_reachable(obj);
732 if (active_cache_tree)
733 fsck_cache_tree(active_cache_tree);
736 check_connectivity();
737 return errors_found;