9 #include "cache-tree.h"
10 #include "tree-walk.h"
12 #include "parse-options.h"
14 #define REACHABLE 0x0001
19 static int show_unreachable
;
20 static int include_reflogs
= 1;
21 static int check_full
;
22 static int check_strict
;
23 static int keep_cache_objects
;
24 static unsigned char head_sha1
[20];
25 static int errors_found
;
26 static int write_lost_and_found
;
28 #define ERROR_OBJECT 01
29 #define ERROR_REACHABLE 02
31 #ifdef NO_D_INO_IN_DIRENT
33 #define DIRENT_SORT_HINT(de) 0
36 #define DIRENT_SORT_HINT(de) ((de)->d_ino)
39 static void objreport(struct object
*obj
, const char *severity
,
40 const char *err
, va_list params
)
42 fprintf(stderr
, "%s in %s %s: ",
43 severity
, typename(obj
->type
), sha1_to_hex(obj
->sha1
));
44 vfprintf(stderr
, err
, params
);
48 static int objerror(struct object
*obj
, const char *err
, ...)
51 va_start(params
, err
);
52 errors_found
|= ERROR_OBJECT
;
53 objreport(obj
, "error", err
, params
);
58 static int fsck_error_func(struct object
*obj
, int type
, const char *err
, ...)
61 va_start(params
, err
);
62 objreport(obj
, (type
== FSCK_WARN
) ? "warning" : "error", err
, params
);
64 return (type
== FSCK_WARN
) ? 0 : 1;
67 static int mark_object(struct object
*obj
, int type
, void *data
)
69 struct tree
*tree
= NULL
;
70 struct object
*parent
= data
;
74 printf("broken link from %7s %s\n",
75 typename(parent
->type
), sha1_to_hex(parent
->sha1
));
76 printf("broken link from %7s %s\n",
77 (type
== OBJ_ANY
? "unknown" : typename(type
)), "unknown");
78 errors_found
|= ERROR_REACHABLE
;
82 if (type
!= OBJ_ANY
&& obj
->type
!= type
)
83 objerror(parent
, "wrong object type in link");
85 if (obj
->flags
& REACHABLE
)
87 obj
->flags
|= REACHABLE
;
89 if (parent
&& !has_sha1_file(obj
->sha1
)) {
90 printf("broken link from %7s %s\n",
91 typename(parent
->type
), sha1_to_hex(parent
->sha1
));
92 printf(" to %7s %s\n",
93 typename(obj
->type
), sha1_to_hex(obj
->sha1
));
94 errors_found
|= ERROR_REACHABLE
;
99 if (obj
->type
== OBJ_TREE
) {
101 tree
= (struct tree
*)obj
;
102 if (parse_tree(tree
) < 0)
103 return 1; /* error already displayed */
105 result
= fsck_walk(obj
, mark_object
, obj
);
116 static void mark_object_reachable(struct object
*obj
)
118 mark_object(obj
, OBJ_ANY
, 0);
121 static int mark_used(struct object
*obj
, int type
, void *data
)
130 * Check a single reachable object
132 static void check_reachable_object(struct object
*obj
)
135 * We obviously want the object to be parsed,
136 * except if it was in a pack-file and we didn't
140 if (has_sha1_pack(obj
->sha1
, NULL
))
141 return; /* it is in pack - forget about it */
142 printf("missing %s %s\n", typename(obj
->type
), sha1_to_hex(obj
->sha1
));
143 errors_found
|= ERROR_REACHABLE
;
149 * Check a single unreachable object
151 static void check_unreachable_object(struct object
*obj
)
154 * Missing unreachable object? Ignore it. It's not like
155 * we miss it (since it can't be reached), nor do we want
156 * to complain about it being unreachable (since it does
163 * Unreachable object that exists? Show it if asked to,
164 * since this is something that is prunable.
166 if (show_unreachable
) {
167 printf("unreachable %s %s\n", typename(obj
->type
), sha1_to_hex(obj
->sha1
));
172 * "!used" means that nothing at all points to it, including
173 * other unreachable objects. In other words, it's the "tip"
174 * of some set of unreachable objects, usually a commit that
177 * Such starting points are more interesting than some random
178 * set of unreachable objects, so we show them even if the user
179 * hasn't asked for _all_ unreachable objects. If you have
180 * deleted a branch by mistake, this is a prime candidate to
181 * start looking at, for example.
184 printf("dangling %s %s\n", typename(obj
->type
),
185 sha1_to_hex(obj
->sha1
));
186 if (write_lost_and_found
) {
187 char *filename
= git_path("lost-found/%s/%s",
188 obj
->type
== OBJ_COMMIT
? "commit" : "other",
189 sha1_to_hex(obj
->sha1
));
192 if (safe_create_leading_directories(filename
)) {
193 error("Could not create lost-found");
196 if (!(f
= fopen(filename
, "w")))
197 die("Could not open %s", filename
);
198 if (obj
->type
== OBJ_BLOB
) {
199 enum object_type type
;
201 char *buf
= read_sha1_file(obj
->sha1
,
204 fwrite(buf
, size
, 1, f
);
208 fprintf(f
, "%s\n", sha1_to_hex(obj
->sha1
));
215 * Otherwise? It's there, it's unreachable, and some other unreachable
216 * object points to it. Ignore it - it's not interesting, and we showed
217 * all the interesting cases above.
221 static void check_object(struct object
*obj
)
224 fprintf(stderr
, "Checking %s\n", sha1_to_hex(obj
->sha1
));
226 if (obj
->flags
& REACHABLE
)
227 check_reachable_object(obj
);
229 check_unreachable_object(obj
);
232 static void check_connectivity(void)
236 /* Look up all the requirements, warn about missing objects.. */
237 max
= get_max_object_index();
239 fprintf(stderr
, "Checking connectivity (%d objects)\n", max
);
241 for (i
= 0; i
< max
; i
++) {
242 struct object
*obj
= get_indexed_object(i
);
249 static int fsck_sha1(const unsigned char *sha1
)
251 struct object
*obj
= parse_object(sha1
);
253 errors_found
|= ERROR_OBJECT
;
254 return error("%s: object corrupt or missing",
257 if (obj
->flags
& SEEN
)
262 fprintf(stderr
, "Checking %s %s\n",
263 typename(obj
->type
), sha1_to_hex(obj
->sha1
));
265 if (fsck_walk(obj
, mark_used
, 0))
266 objerror(obj
, "broken links");
267 if (fsck_object(obj
, check_strict
, fsck_error_func
))
270 if (obj
->type
== OBJ_TREE
) {
271 struct tree
*item
= (struct tree
*) obj
;
277 if (obj
->type
== OBJ_COMMIT
) {
278 struct commit
*commit
= (struct commit
*) obj
;
280 free(commit
->buffer
);
281 commit
->buffer
= NULL
;
283 if (!commit
->parents
&& show_root
)
284 printf("root %s\n", sha1_to_hex(commit
->object
.sha1
));
287 if (obj
->type
== OBJ_TAG
) {
288 struct tag
*tag
= (struct tag
*) obj
;
290 if (show_tags
&& tag
->tagged
) {
291 printf("tagged %s %s", typename(tag
->tagged
->type
), sha1_to_hex(tag
->tagged
->sha1
));
292 printf(" (%s) in %s\n", tag
->tag
, sha1_to_hex(tag
->object
.sha1
));
300 * This is the sorting chunk size: make it reasonably
301 * big so that we can sort well..
303 #define MAX_SHA1_ENTRIES (1024)
307 unsigned char sha1
[20];
312 struct sha1_entry
*entry
[MAX_SHA1_ENTRIES
];
315 static int ino_compare(const void *_a
, const void *_b
)
317 const struct sha1_entry
*a
= _a
, *b
= _b
;
318 unsigned long ino1
= a
->ino
, ino2
= b
->ino
;
319 return ino1
< ino2
? -1 : ino1
> ino2
? 1 : 0;
322 static void fsck_sha1_list(void)
324 int i
, nr
= sha1_list
.nr
;
327 qsort(sha1_list
.entry
, nr
,
328 sizeof(struct sha1_entry
*), ino_compare
);
329 for (i
= 0; i
< nr
; i
++) {
330 struct sha1_entry
*entry
= sha1_list
.entry
[i
];
331 unsigned char *sha1
= entry
->sha1
;
333 sha1_list
.entry
[i
] = NULL
;
340 static void add_sha1_list(unsigned char *sha1
, unsigned long ino
)
342 struct sha1_entry
*entry
= xmalloc(sizeof(*entry
));
346 hashcpy(entry
->sha1
, sha1
);
348 if (nr
== MAX_SHA1_ENTRIES
) {
352 sha1_list
.entry
[nr
] = entry
;
356 static void fsck_dir(int i
, char *path
)
358 DIR *dir
= opendir(path
);
365 fprintf(stderr
, "Checking directory %s\n", path
);
367 while ((de
= readdir(dir
)) != NULL
) {
369 unsigned char sha1
[20];
370 int len
= strlen(de
->d_name
);
374 if (de
->d_name
[1] != '.')
377 if (de
->d_name
[0] != '.')
381 sprintf(name
, "%02x", i
);
382 memcpy(name
+2, de
->d_name
, len
+1);
383 if (get_sha1_hex(name
, sha1
) < 0)
385 add_sha1_list(sha1
, DIRENT_SORT_HINT(de
));
388 fprintf(stderr
, "bad sha1 file: %s/%s\n", path
, de
->d_name
);
393 static int default_refs
;
395 static int fsck_handle_reflog_ent(unsigned char *osha1
, unsigned char *nsha1
,
396 const char *email
, unsigned long timestamp
, int tz
,
397 const char *message
, void *cb_data
)
402 fprintf(stderr
, "Checking reflog %s->%s\n",
403 sha1_to_hex(osha1
), sha1_to_hex(nsha1
));
405 if (!is_null_sha1(osha1
)) {
406 obj
= lookup_object(osha1
);
409 mark_object_reachable(obj
);
412 obj
= lookup_object(nsha1
);
415 mark_object_reachable(obj
);
420 static int fsck_handle_reflog(const char *logname
, const unsigned char *sha1
, int flag
, void *cb_data
)
422 for_each_reflog_ent(logname
, fsck_handle_reflog_ent
, NULL
);
426 static int is_branch(const char *refname
)
428 return !strcmp(refname
, "HEAD") || !prefixcmp(refname
, "refs/heads/");
431 static int fsck_handle_ref(const char *refname
, const unsigned char *sha1
, int flag
, void *cb_data
)
435 obj
= parse_object(sha1
);
437 error("%s: invalid sha1 pointer %s", refname
, sha1_to_hex(sha1
));
438 /* We'll continue with the rest despite the error.. */
441 if (obj
->type
!= OBJ_COMMIT
&& is_branch(refname
))
442 error("%s: not a commit", refname
);
445 mark_object_reachable(obj
);
450 static void get_default_heads(void)
452 for_each_ref(fsck_handle_ref
, NULL
);
454 for_each_reflog(fsck_handle_reflog
, NULL
);
457 * Not having any default heads isn't really fatal, but
458 * it does mean that "--unreachable" no longer makes any
459 * sense (since in this case everything will obviously
460 * be unreachable by definition.
462 * Showing dangling objects is valid, though (as those
463 * dangling objects are likely lost heads).
465 * So we just print a warning about it, and clear the
466 * "show_unreachable" flag.
469 fprintf(stderr
, "notice: No default references\n");
470 show_unreachable
= 0;
474 static void fsck_object_dir(const char *path
)
479 fprintf(stderr
, "Checking object directory\n");
481 for (i
= 0; i
< 256; i
++) {
482 static char dir
[4096];
483 sprintf(dir
, "%s/%02x", path
, i
);
489 static int fsck_head_link(void)
491 unsigned char sha1
[20];
493 int null_is_error
= 0;
494 const char *head_points_at
= resolve_ref("HEAD", sha1
, 0, &flag
);
497 fprintf(stderr
, "Checking HEAD link\n");
500 return error("Invalid HEAD");
501 if (!strcmp(head_points_at
, "HEAD"))
504 else if (prefixcmp(head_points_at
, "refs/heads/"))
505 return error("HEAD points to something strange (%s)",
507 if (is_null_sha1(sha1
)) {
509 return error("HEAD: detached HEAD points at nothing");
510 fprintf(stderr
, "notice: HEAD points to an unborn branch (%s)\n",
511 head_points_at
+ 11);
516 static int fsck_cache_tree(struct cache_tree
*it
)
522 fprintf(stderr
, "Checking cache tree\n");
524 if (0 <= it
->entry_count
) {
525 struct object
*obj
= parse_object(it
->sha1
);
527 error("%s: invalid sha1 pointer in cache-tree",
528 sha1_to_hex(it
->sha1
));
531 mark_object_reachable(obj
);
533 if (obj
->type
!= OBJ_TREE
)
534 err
|= objerror(obj
, "non-tree in cache-tree");
536 for (i
= 0; i
< it
->subtree_nr
; i
++)
537 err
|= fsck_cache_tree(it
->down
[i
]->cache_tree
);
541 static char const * const fsck_usage
[] = {
542 "git-fsck [options] [<object>...]",
546 static struct option fsck_opts
[] = {
547 OPT__VERBOSE(&verbose
),
548 OPT_BOOLEAN(0, "unreachable", &show_unreachable
, "show unreachable objects"),
549 OPT_BOOLEAN(0, "tags", &show_tags
, "report tags"),
550 OPT_BOOLEAN(0, "root", &show_root
, "report root nodes"),
551 OPT_BOOLEAN(0, "cache", &keep_cache_objects
, "make index objects head nodes"),
552 OPT_BOOLEAN(0, "reflogs", &include_reflogs
, "make reflogs head nodes (default)"),
553 OPT_BOOLEAN(0, "full", &check_full
, "also consider alternate objects"),
554 OPT_BOOLEAN(0, "strict", &check_strict
, "enable more strict checking"),
555 OPT_BOOLEAN(0, "lost-found", &write_lost_and_found
,
556 "write dangling objects in .git/lost-found"),
560 int cmd_fsck(int argc
, const char **argv
, const char *prefix
)
566 argc
= parse_options(argc
, argv
, fsck_opts
, fsck_usage
, 0);
567 if (write_lost_and_found
) {
573 fsck_object_dir(get_object_directory());
575 struct alternate_object_database
*alt
;
576 struct packed_git
*p
;
578 for (alt
= alt_odb_list
; alt
; alt
= alt
->next
) {
579 char namebuf
[PATH_MAX
];
580 int namelen
= alt
->name
- alt
->base
;
581 memcpy(namebuf
, alt
->base
, namelen
);
582 namebuf
[namelen
- 1] = 0;
583 fsck_object_dir(namebuf
);
585 prepare_packed_git();
586 for (p
= packed_git
; p
; p
= p
->next
)
587 /* verify gives error messages itself */
590 for (p
= packed_git
; p
; p
= p
->next
) {
592 if (open_pack_index(p
))
594 num
= p
->num_objects
;
595 for (j
= 0; j
< num
; j
++)
596 fsck_sha1(nth_packed_object_sha1(p
, j
));
601 for (i
= 1; i
< argc
; i
++) {
602 const char *arg
= argv
[i
];
603 if (!get_sha1(arg
, head_sha1
)) {
604 struct object
*obj
= lookup_object(head_sha1
);
606 /* Error is printed by lookup_object(). */
611 mark_object_reachable(obj
);
615 error("invalid parameter: expected sha1, got '%s'", arg
);
619 * If we've not been given any explicit head information, do the
620 * default ones from .git/refs. We also consider the index file
621 * in this case (ie this implies --cache).
625 keep_cache_objects
= 1;
628 if (keep_cache_objects
) {
630 for (i
= 0; i
< active_nr
; i
++) {
635 mode
= active_cache
[i
]->ce_mode
;
636 if (S_ISGITLINK(mode
))
638 blob
= lookup_blob(active_cache
[i
]->sha1
);
643 mark_object_reachable(obj
);
645 if (active_cache_tree
)
646 fsck_cache_tree(active_cache_tree
);
649 check_connectivity();