]>
Commit | Line | Data |
---|---|---|
1 | #define USE_THE_INDEX_COMPATIBILITY_MACROS | |
2 | #include "builtin.h" | |
3 | #include "cache.h" | |
4 | #include "repository.h" | |
5 | #include "config.h" | |
6 | #include "commit.h" | |
7 | #include "tree.h" | |
8 | #include "blob.h" | |
9 | #include "tag.h" | |
10 | #include "refs.h" | |
11 | #include "pack.h" | |
12 | #include "cache-tree.h" | |
13 | #include "tree-walk.h" | |
14 | #include "fsck.h" | |
15 | #include "parse-options.h" | |
16 | #include "dir.h" | |
17 | #include "progress.h" | |
18 | #include "streaming.h" | |
19 | #include "decorate.h" | |
20 | #include "packfile.h" | |
21 | #include "object-store.h" | |
22 | #include "run-command.h" | |
23 | #include "worktree.h" | |
24 | ||
25 | #define REACHABLE 0x0001 | |
26 | #define SEEN 0x0002 | |
27 | #define HAS_OBJ 0x0004 | |
28 | /* This flag is set if something points to this object. */ | |
29 | #define USED 0x0008 | |
30 | ||
31 | static int show_root; | |
32 | static int show_tags; | |
33 | static int show_unreachable; | |
34 | static int include_reflogs = 1; | |
35 | static int check_full = 1; | |
36 | static int connectivity_only; | |
37 | static int check_strict; | |
38 | static int keep_cache_objects; | |
39 | static struct fsck_options fsck_walk_options = FSCK_OPTIONS_DEFAULT; | |
40 | static struct fsck_options fsck_obj_options = FSCK_OPTIONS_DEFAULT; | |
41 | static int errors_found; | |
42 | static int write_lost_and_found; | |
43 | static int verbose; | |
44 | static int show_progress = -1; | |
45 | static int show_dangling = 1; | |
46 | static int name_objects; | |
47 | #define ERROR_OBJECT 01 | |
48 | #define ERROR_REACHABLE 02 | |
49 | #define ERROR_PACK 04 | |
50 | #define ERROR_REFS 010 | |
51 | #define ERROR_COMMIT_GRAPH 020 | |
52 | ||
53 | static const char *describe_object(const struct object_id *oid) | |
54 | { | |
55 | return fsck_describe_object(&fsck_walk_options, oid); | |
56 | } | |
57 | ||
58 | static const char *printable_type(const struct object_id *oid, | |
59 | enum object_type type) | |
60 | { | |
61 | const char *ret; | |
62 | ||
63 | if (type == OBJ_NONE) | |
64 | type = oid_object_info(the_repository, oid, NULL); | |
65 | ||
66 | ret = type_name(type); | |
67 | if (!ret) | |
68 | ret = _("unknown"); | |
69 | ||
70 | return ret; | |
71 | } | |
72 | ||
73 | static int fsck_config(const char *var, const char *value, void *cb) | |
74 | { | |
75 | if (strcmp(var, "fsck.skiplist") == 0) { | |
76 | const char *path; | |
77 | struct strbuf sb = STRBUF_INIT; | |
78 | ||
79 | if (git_config_pathname(&path, var, value)) | |
80 | return 1; | |
81 | strbuf_addf(&sb, "skiplist=%s", path); | |
82 | free((char *)path); | |
83 | fsck_set_msg_types(&fsck_obj_options, sb.buf); | |
84 | strbuf_release(&sb); | |
85 | return 0; | |
86 | } | |
87 | ||
88 | if (skip_prefix(var, "fsck.", &var)) { | |
89 | fsck_set_msg_type(&fsck_obj_options, var, value); | |
90 | return 0; | |
91 | } | |
92 | ||
93 | return git_default_config(var, value, cb); | |
94 | } | |
95 | ||
96 | static int objerror(struct object *obj, const char *err) | |
97 | { | |
98 | errors_found |= ERROR_OBJECT; | |
99 | /* TRANSLATORS: e.g. error in tree 01bfda: <more explanation> */ | |
100 | fprintf_ln(stderr, _("error in %s %s: %s"), | |
101 | printable_type(&obj->oid, obj->type), | |
102 | describe_object(&obj->oid), err); | |
103 | return -1; | |
104 | } | |
105 | ||
106 | static int fsck_error_func(struct fsck_options *o, | |
107 | const struct object_id *oid, | |
108 | enum object_type object_type, | |
109 | int msg_type, const char *message) | |
110 | { | |
111 | switch (msg_type) { | |
112 | case FSCK_WARN: | |
113 | /* TRANSLATORS: e.g. warning in tree 01bfda: <more explanation> */ | |
114 | fprintf_ln(stderr, _("warning in %s %s: %s"), | |
115 | printable_type(oid, object_type), | |
116 | describe_object(oid), message); | |
117 | return 0; | |
118 | case FSCK_ERROR: | |
119 | /* TRANSLATORS: e.g. error in tree 01bfda: <more explanation> */ | |
120 | fprintf_ln(stderr, _("error in %s %s: %s"), | |
121 | printable_type(oid, object_type), | |
122 | describe_object(oid), message); | |
123 | return 1; | |
124 | default: | |
125 | BUG("%d (FSCK_IGNORE?) should never trigger this callback", | |
126 | msg_type); | |
127 | } | |
128 | } | |
129 | ||
130 | static struct object_array pending; | |
131 | ||
132 | static int mark_object(struct object *obj, int type, void *data, struct fsck_options *options) | |
133 | { | |
134 | struct object *parent = data; | |
135 | ||
136 | /* | |
137 | * The only case data is NULL or type is OBJ_ANY is when | |
138 | * mark_object_reachable() calls us. All the callers of | |
139 | * that function has non-NULL obj hence ... | |
140 | */ | |
141 | if (!obj) { | |
142 | /* ... these references to parent->fld are safe here */ | |
143 | printf_ln(_("broken link from %7s %s"), | |
144 | printable_type(&parent->oid, parent->type), | |
145 | describe_object(&parent->oid)); | |
146 | printf_ln(_("broken link from %7s %s"), | |
147 | (type == OBJ_ANY ? _("unknown") : type_name(type)), | |
148 | _("unknown")); | |
149 | errors_found |= ERROR_REACHABLE; | |
150 | return 1; | |
151 | } | |
152 | ||
153 | if (type != OBJ_ANY && obj->type != type) | |
154 | /* ... and the reference to parent is safe here */ | |
155 | objerror(parent, _("wrong object type in link")); | |
156 | ||
157 | if (obj->flags & REACHABLE) | |
158 | return 0; | |
159 | obj->flags |= REACHABLE; | |
160 | ||
161 | if (is_promisor_object(&obj->oid)) | |
162 | /* | |
163 | * Further recursion does not need to be performed on this | |
164 | * object since it is a promisor object (so it does not need to | |
165 | * be added to "pending"). | |
166 | */ | |
167 | return 0; | |
168 | ||
169 | if (!(obj->flags & HAS_OBJ)) { | |
170 | if (parent && !has_object_file(&obj->oid)) { | |
171 | printf_ln(_("broken link from %7s %s\n" | |
172 | " to %7s %s"), | |
173 | printable_type(&parent->oid, parent->type), | |
174 | describe_object(&parent->oid), | |
175 | printable_type(&obj->oid, obj->type), | |
176 | describe_object(&obj->oid)); | |
177 | errors_found |= ERROR_REACHABLE; | |
178 | } | |
179 | return 1; | |
180 | } | |
181 | ||
182 | add_object_array(obj, NULL, &pending); | |
183 | return 0; | |
184 | } | |
185 | ||
186 | static void mark_object_reachable(struct object *obj) | |
187 | { | |
188 | mark_object(obj, OBJ_ANY, NULL, NULL); | |
189 | } | |
190 | ||
191 | static int traverse_one_object(struct object *obj) | |
192 | { | |
193 | int result = fsck_walk(obj, obj, &fsck_walk_options); | |
194 | ||
195 | if (obj->type == OBJ_TREE) { | |
196 | struct tree *tree = (struct tree *)obj; | |
197 | free_tree_buffer(tree); | |
198 | } | |
199 | return result; | |
200 | } | |
201 | ||
202 | static int traverse_reachable(void) | |
203 | { | |
204 | struct progress *progress = NULL; | |
205 | unsigned int nr = 0; | |
206 | int result = 0; | |
207 | if (show_progress) | |
208 | progress = start_delayed_progress(_("Checking connectivity"), 0); | |
209 | while (pending.nr) { | |
210 | result |= traverse_one_object(object_array_pop(&pending)); | |
211 | display_progress(progress, ++nr); | |
212 | } | |
213 | stop_progress(&progress); | |
214 | return !!result; | |
215 | } | |
216 | ||
217 | static int mark_used(struct object *obj, int type, void *data, struct fsck_options *options) | |
218 | { | |
219 | if (!obj) | |
220 | return 1; | |
221 | obj->flags |= USED; | |
222 | return 0; | |
223 | } | |
224 | ||
225 | static void mark_unreachable_referents(const struct object_id *oid) | |
226 | { | |
227 | struct fsck_options options = FSCK_OPTIONS_DEFAULT; | |
228 | struct object *obj = lookup_object(the_repository, oid); | |
229 | ||
230 | if (!obj || !(obj->flags & HAS_OBJ)) | |
231 | return; /* not part of our original set */ | |
232 | if (obj->flags & REACHABLE) | |
233 | return; /* reachable objects already traversed */ | |
234 | ||
235 | /* | |
236 | * Avoid passing OBJ_NONE to fsck_walk, which will parse the object | |
237 | * (and we want to avoid parsing blobs). | |
238 | */ | |
239 | if (obj->type == OBJ_NONE) { | |
240 | enum object_type type = oid_object_info(the_repository, | |
241 | &obj->oid, NULL); | |
242 | if (type > 0) | |
243 | object_as_type(the_repository, obj, type, 0); | |
244 | } | |
245 | ||
246 | options.walk = mark_used; | |
247 | fsck_walk(obj, NULL, &options); | |
248 | } | |
249 | ||
250 | static int mark_loose_unreachable_referents(const struct object_id *oid, | |
251 | const char *path, | |
252 | void *data) | |
253 | { | |
254 | mark_unreachable_referents(oid); | |
255 | return 0; | |
256 | } | |
257 | ||
258 | static int mark_packed_unreachable_referents(const struct object_id *oid, | |
259 | struct packed_git *pack, | |
260 | uint32_t pos, | |
261 | void *data) | |
262 | { | |
263 | mark_unreachable_referents(oid); | |
264 | return 0; | |
265 | } | |
266 | ||
267 | /* | |
268 | * Check a single reachable object | |
269 | */ | |
270 | static void check_reachable_object(struct object *obj) | |
271 | { | |
272 | /* | |
273 | * We obviously want the object to be parsed, | |
274 | * except if it was in a pack-file and we didn't | |
275 | * do a full fsck | |
276 | */ | |
277 | if (!(obj->flags & HAS_OBJ)) { | |
278 | if (is_promisor_object(&obj->oid)) | |
279 | return; | |
280 | if (has_object_pack(&obj->oid)) | |
281 | return; /* it is in pack - forget about it */ | |
282 | printf_ln(_("missing %s %s"), | |
283 | printable_type(&obj->oid, obj->type), | |
284 | describe_object(&obj->oid)); | |
285 | errors_found |= ERROR_REACHABLE; | |
286 | return; | |
287 | } | |
288 | } | |
289 | ||
290 | /* | |
291 | * Check a single unreachable object | |
292 | */ | |
293 | static void check_unreachable_object(struct object *obj) | |
294 | { | |
295 | /* | |
296 | * Missing unreachable object? Ignore it. It's not like | |
297 | * we miss it (since it can't be reached), nor do we want | |
298 | * to complain about it being unreachable (since it does | |
299 | * not exist). | |
300 | */ | |
301 | if (!(obj->flags & HAS_OBJ)) | |
302 | return; | |
303 | ||
304 | /* | |
305 | * Unreachable object that exists? Show it if asked to, | |
306 | * since this is something that is prunable. | |
307 | */ | |
308 | if (show_unreachable) { | |
309 | printf_ln(_("unreachable %s %s"), | |
310 | printable_type(&obj->oid, obj->type), | |
311 | describe_object(&obj->oid)); | |
312 | return; | |
313 | } | |
314 | ||
315 | /* | |
316 | * "!USED" means that nothing at all points to it, including | |
317 | * other unreachable objects. In other words, it's the "tip" | |
318 | * of some set of unreachable objects, usually a commit that | |
319 | * got dropped. | |
320 | * | |
321 | * Such starting points are more interesting than some random | |
322 | * set of unreachable objects, so we show them even if the user | |
323 | * hasn't asked for _all_ unreachable objects. If you have | |
324 | * deleted a branch by mistake, this is a prime candidate to | |
325 | * start looking at, for example. | |
326 | */ | |
327 | if (!(obj->flags & USED)) { | |
328 | if (show_dangling) | |
329 | printf_ln(_("dangling %s %s"), | |
330 | printable_type(&obj->oid, obj->type), | |
331 | describe_object(&obj->oid)); | |
332 | if (write_lost_and_found) { | |
333 | char *filename = git_pathdup("lost-found/%s/%s", | |
334 | obj->type == OBJ_COMMIT ? "commit" : "other", | |
335 | describe_object(&obj->oid)); | |
336 | FILE *f; | |
337 | ||
338 | if (safe_create_leading_directories_const(filename)) { | |
339 | error(_("could not create lost-found")); | |
340 | free(filename); | |
341 | return; | |
342 | } | |
343 | f = xfopen(filename, "w"); | |
344 | if (obj->type == OBJ_BLOB) { | |
345 | if (stream_blob_to_fd(fileno(f), &obj->oid, NULL, 1)) | |
346 | die_errno(_("could not write '%s'"), filename); | |
347 | } else | |
348 | fprintf(f, "%s\n", describe_object(&obj->oid)); | |
349 | if (fclose(f)) | |
350 | die_errno(_("could not finish '%s'"), | |
351 | filename); | |
352 | free(filename); | |
353 | } | |
354 | return; | |
355 | } | |
356 | ||
357 | /* | |
358 | * Otherwise? It's there, it's unreachable, and some other unreachable | |
359 | * object points to it. Ignore it - it's not interesting, and we showed | |
360 | * all the interesting cases above. | |
361 | */ | |
362 | } | |
363 | ||
364 | static void check_object(struct object *obj) | |
365 | { | |
366 | if (verbose) | |
367 | fprintf_ln(stderr, _("Checking %s"), describe_object(&obj->oid)); | |
368 | ||
369 | if (obj->flags & REACHABLE) | |
370 | check_reachable_object(obj); | |
371 | else | |
372 | check_unreachable_object(obj); | |
373 | } | |
374 | ||
375 | static void check_connectivity(void) | |
376 | { | |
377 | int i, max; | |
378 | ||
379 | /* Traverse the pending reachable objects */ | |
380 | traverse_reachable(); | |
381 | ||
382 | /* | |
383 | * With --connectivity-only, we won't have actually opened and marked | |
384 | * unreachable objects with USED. Do that now to make --dangling, etc | |
385 | * accurate. | |
386 | */ | |
387 | if (connectivity_only && (show_dangling || write_lost_and_found)) { | |
388 | /* | |
389 | * Even though we already have a "struct object" for each of | |
390 | * these in memory, we must not iterate over the internal | |
391 | * object hash as we do below. Our loop would potentially | |
392 | * resize the hash, making our iteration invalid. | |
393 | * | |
394 | * Instead, we'll just go back to the source list of objects, | |
395 | * and ignore any that weren't present in our earlier | |
396 | * traversal. | |
397 | */ | |
398 | for_each_loose_object(mark_loose_unreachable_referents, NULL, 0); | |
399 | for_each_packed_object(mark_packed_unreachable_referents, NULL, 0); | |
400 | } | |
401 | ||
402 | /* Look up all the requirements, warn about missing objects.. */ | |
403 | max = get_max_object_index(); | |
404 | if (verbose) | |
405 | fprintf_ln(stderr, _("Checking connectivity (%d objects)"), max); | |
406 | ||
407 | for (i = 0; i < max; i++) { | |
408 | struct object *obj = get_indexed_object(i); | |
409 | ||
410 | if (obj) | |
411 | check_object(obj); | |
412 | } | |
413 | } | |
414 | ||
415 | static int fsck_obj(struct object *obj, void *buffer, unsigned long size) | |
416 | { | |
417 | int err; | |
418 | ||
419 | if (obj->flags & SEEN) | |
420 | return 0; | |
421 | obj->flags |= SEEN; | |
422 | ||
423 | if (verbose) | |
424 | fprintf_ln(stderr, _("Checking %s %s"), | |
425 | printable_type(&obj->oid, obj->type), | |
426 | describe_object(&obj->oid)); | |
427 | ||
428 | if (fsck_walk(obj, NULL, &fsck_obj_options)) | |
429 | objerror(obj, _("broken links")); | |
430 | err = fsck_object(obj, buffer, size, &fsck_obj_options); | |
431 | if (err) | |
432 | goto out; | |
433 | ||
434 | if (obj->type == OBJ_COMMIT) { | |
435 | struct commit *commit = (struct commit *) obj; | |
436 | ||
437 | if (!commit->parents && show_root) | |
438 | printf_ln(_("root %s"), | |
439 | describe_object(&commit->object.oid)); | |
440 | } | |
441 | ||
442 | if (obj->type == OBJ_TAG) { | |
443 | struct tag *tag = (struct tag *) obj; | |
444 | ||
445 | if (show_tags && tag->tagged) { | |
446 | printf_ln(_("tagged %s %s (%s) in %s"), | |
447 | printable_type(&tag->tagged->oid, tag->tagged->type), | |
448 | describe_object(&tag->tagged->oid), | |
449 | tag->tag, | |
450 | describe_object(&tag->object.oid)); | |
451 | } | |
452 | } | |
453 | ||
454 | out: | |
455 | if (obj->type == OBJ_TREE) | |
456 | free_tree_buffer((struct tree *)obj); | |
457 | if (obj->type == OBJ_COMMIT) | |
458 | free_commit_buffer(the_repository->parsed_objects, | |
459 | (struct commit *)obj); | |
460 | return err; | |
461 | } | |
462 | ||
463 | static int fsck_obj_buffer(const struct object_id *oid, enum object_type type, | |
464 | unsigned long size, void *buffer, int *eaten) | |
465 | { | |
466 | /* | |
467 | * Note, buffer may be NULL if type is OBJ_BLOB. See | |
468 | * verify_packfile(), data_valid variable for details. | |
469 | */ | |
470 | struct object *obj; | |
471 | obj = parse_object_buffer(the_repository, oid, type, size, buffer, | |
472 | eaten); | |
473 | if (!obj) { | |
474 | errors_found |= ERROR_OBJECT; | |
475 | return error(_("%s: object corrupt or missing"), | |
476 | oid_to_hex(oid)); | |
477 | } | |
478 | obj->flags &= ~(REACHABLE | SEEN); | |
479 | obj->flags |= HAS_OBJ; | |
480 | return fsck_obj(obj, buffer, size); | |
481 | } | |
482 | ||
483 | static int default_refs; | |
484 | ||
485 | static void fsck_handle_reflog_oid(const char *refname, struct object_id *oid, | |
486 | timestamp_t timestamp) | |
487 | { | |
488 | struct object *obj; | |
489 | ||
490 | if (!is_null_oid(oid)) { | |
491 | obj = lookup_object(the_repository, oid); | |
492 | if (obj && (obj->flags & HAS_OBJ)) { | |
493 | if (timestamp) | |
494 | fsck_put_object_name(&fsck_walk_options, oid, | |
495 | "%s@{%"PRItime"}", | |
496 | refname, timestamp); | |
497 | obj->flags |= USED; | |
498 | mark_object_reachable(obj); | |
499 | } else if (!is_promisor_object(oid)) { | |
500 | error(_("%s: invalid reflog entry %s"), | |
501 | refname, oid_to_hex(oid)); | |
502 | errors_found |= ERROR_REACHABLE; | |
503 | } | |
504 | } | |
505 | } | |
506 | ||
507 | static int fsck_handle_reflog_ent(struct object_id *ooid, struct object_id *noid, | |
508 | const char *email, timestamp_t timestamp, int tz, | |
509 | const char *message, void *cb_data) | |
510 | { | |
511 | const char *refname = cb_data; | |
512 | ||
513 | if (verbose) | |
514 | fprintf_ln(stderr, _("Checking reflog %s->%s"), | |
515 | oid_to_hex(ooid), oid_to_hex(noid)); | |
516 | ||
517 | fsck_handle_reflog_oid(refname, ooid, 0); | |
518 | fsck_handle_reflog_oid(refname, noid, timestamp); | |
519 | return 0; | |
520 | } | |
521 | ||
522 | static int fsck_handle_reflog(const char *logname, const struct object_id *oid, | |
523 | int flag, void *cb_data) | |
524 | { | |
525 | struct strbuf refname = STRBUF_INIT; | |
526 | ||
527 | strbuf_worktree_ref(cb_data, &refname, logname); | |
528 | for_each_reflog_ent(refname.buf, fsck_handle_reflog_ent, refname.buf); | |
529 | strbuf_release(&refname); | |
530 | return 0; | |
531 | } | |
532 | ||
533 | static int fsck_handle_ref(const char *refname, const struct object_id *oid, | |
534 | int flag, void *cb_data) | |
535 | { | |
536 | struct object *obj; | |
537 | ||
538 | obj = parse_object(the_repository, oid); | |
539 | if (!obj) { | |
540 | if (is_promisor_object(oid)) { | |
541 | /* | |
542 | * Increment default_refs anyway, because this is a | |
543 | * valid ref. | |
544 | */ | |
545 | default_refs++; | |
546 | return 0; | |
547 | } | |
548 | error(_("%s: invalid sha1 pointer %s"), | |
549 | refname, oid_to_hex(oid)); | |
550 | errors_found |= ERROR_REACHABLE; | |
551 | /* We'll continue with the rest despite the error.. */ | |
552 | return 0; | |
553 | } | |
554 | if (obj->type != OBJ_COMMIT && is_branch(refname)) { | |
555 | error(_("%s: not a commit"), refname); | |
556 | errors_found |= ERROR_REFS; | |
557 | } | |
558 | default_refs++; | |
559 | obj->flags |= USED; | |
560 | fsck_put_object_name(&fsck_walk_options, | |
561 | oid, "%s", refname); | |
562 | mark_object_reachable(obj); | |
563 | ||
564 | return 0; | |
565 | } | |
566 | ||
567 | static int fsck_head_link(const char *head_ref_name, | |
568 | const char **head_points_at, | |
569 | struct object_id *head_oid); | |
570 | ||
571 | static void get_default_heads(void) | |
572 | { | |
573 | struct worktree **worktrees, **p; | |
574 | const char *head_points_at; | |
575 | struct object_id head_oid; | |
576 | ||
577 | for_each_rawref(fsck_handle_ref, NULL); | |
578 | ||
579 | worktrees = get_worktrees(0); | |
580 | for (p = worktrees; *p; p++) { | |
581 | struct worktree *wt = *p; | |
582 | struct strbuf ref = STRBUF_INIT; | |
583 | ||
584 | strbuf_worktree_ref(wt, &ref, "HEAD"); | |
585 | fsck_head_link(ref.buf, &head_points_at, &head_oid); | |
586 | if (head_points_at && !is_null_oid(&head_oid)) | |
587 | fsck_handle_ref(ref.buf, &head_oid, 0, NULL); | |
588 | strbuf_release(&ref); | |
589 | ||
590 | if (include_reflogs) | |
591 | refs_for_each_reflog(get_worktree_ref_store(wt), | |
592 | fsck_handle_reflog, wt); | |
593 | } | |
594 | free_worktrees(worktrees); | |
595 | ||
596 | /* | |
597 | * Not having any default heads isn't really fatal, but | |
598 | * it does mean that "--unreachable" no longer makes any | |
599 | * sense (since in this case everything will obviously | |
600 | * be unreachable by definition. | |
601 | * | |
602 | * Showing dangling objects is valid, though (as those | |
603 | * dangling objects are likely lost heads). | |
604 | * | |
605 | * So we just print a warning about it, and clear the | |
606 | * "show_unreachable" flag. | |
607 | */ | |
608 | if (!default_refs) { | |
609 | fprintf_ln(stderr, _("notice: No default references")); | |
610 | show_unreachable = 0; | |
611 | } | |
612 | } | |
613 | ||
614 | static int fsck_loose(const struct object_id *oid, const char *path, void *data) | |
615 | { | |
616 | struct object *obj; | |
617 | enum object_type type; | |
618 | unsigned long size; | |
619 | void *contents; | |
620 | int eaten; | |
621 | ||
622 | if (read_loose_object(path, oid, &type, &size, &contents) < 0) { | |
623 | errors_found |= ERROR_OBJECT; | |
624 | error(_("%s: object corrupt or missing: %s"), | |
625 | oid_to_hex(oid), path); | |
626 | return 0; /* keep checking other objects */ | |
627 | } | |
628 | ||
629 | if (!contents && type != OBJ_BLOB) | |
630 | BUG("read_loose_object streamed a non-blob"); | |
631 | ||
632 | obj = parse_object_buffer(the_repository, oid, type, size, | |
633 | contents, &eaten); | |
634 | ||
635 | if (!obj) { | |
636 | errors_found |= ERROR_OBJECT; | |
637 | error(_("%s: object could not be parsed: %s"), | |
638 | oid_to_hex(oid), path); | |
639 | if (!eaten) | |
640 | free(contents); | |
641 | return 0; /* keep checking other objects */ | |
642 | } | |
643 | ||
644 | obj->flags &= ~(REACHABLE | SEEN); | |
645 | obj->flags |= HAS_OBJ; | |
646 | if (fsck_obj(obj, contents, size)) | |
647 | errors_found |= ERROR_OBJECT; | |
648 | ||
649 | if (!eaten) | |
650 | free(contents); | |
651 | return 0; /* keep checking other objects, even if we saw an error */ | |
652 | } | |
653 | ||
654 | static int fsck_cruft(const char *basename, const char *path, void *data) | |
655 | { | |
656 | if (!starts_with(basename, "tmp_obj_")) | |
657 | fprintf_ln(stderr, _("bad sha1 file: %s"), path); | |
658 | return 0; | |
659 | } | |
660 | ||
661 | static int fsck_subdir(unsigned int nr, const char *path, void *progress) | |
662 | { | |
663 | display_progress(progress, nr + 1); | |
664 | return 0; | |
665 | } | |
666 | ||
667 | static void fsck_object_dir(const char *path) | |
668 | { | |
669 | struct progress *progress = NULL; | |
670 | ||
671 | if (verbose) | |
672 | fprintf_ln(stderr, _("Checking object directory")); | |
673 | ||
674 | if (show_progress) | |
675 | progress = start_progress(_("Checking object directories"), 256); | |
676 | ||
677 | for_each_loose_file_in_objdir(path, fsck_loose, fsck_cruft, fsck_subdir, | |
678 | progress); | |
679 | display_progress(progress, 256); | |
680 | stop_progress(&progress); | |
681 | } | |
682 | ||
683 | static int fsck_head_link(const char *head_ref_name, | |
684 | const char **head_points_at, | |
685 | struct object_id *head_oid) | |
686 | { | |
687 | int null_is_error = 0; | |
688 | ||
689 | if (verbose) | |
690 | fprintf_ln(stderr, _("Checking %s link"), head_ref_name); | |
691 | ||
692 | *head_points_at = resolve_ref_unsafe(head_ref_name, 0, head_oid, NULL); | |
693 | if (!*head_points_at) { | |
694 | errors_found |= ERROR_REFS; | |
695 | return error(_("invalid %s"), head_ref_name); | |
696 | } | |
697 | if (!strcmp(*head_points_at, head_ref_name)) | |
698 | /* detached HEAD */ | |
699 | null_is_error = 1; | |
700 | else if (!starts_with(*head_points_at, "refs/heads/")) { | |
701 | errors_found |= ERROR_REFS; | |
702 | return error(_("%s points to something strange (%s)"), | |
703 | head_ref_name, *head_points_at); | |
704 | } | |
705 | if (is_null_oid(head_oid)) { | |
706 | if (null_is_error) { | |
707 | errors_found |= ERROR_REFS; | |
708 | return error(_("%s: detached HEAD points at nothing"), | |
709 | head_ref_name); | |
710 | } | |
711 | fprintf_ln(stderr, | |
712 | _("notice: %s points to an unborn branch (%s)"), | |
713 | head_ref_name, *head_points_at + 11); | |
714 | } | |
715 | return 0; | |
716 | } | |
717 | ||
718 | static int fsck_cache_tree(struct cache_tree *it) | |
719 | { | |
720 | int i; | |
721 | int err = 0; | |
722 | ||
723 | if (verbose) | |
724 | fprintf_ln(stderr, _("Checking cache tree")); | |
725 | ||
726 | if (0 <= it->entry_count) { | |
727 | struct object *obj = parse_object(the_repository, &it->oid); | |
728 | if (!obj) { | |
729 | error(_("%s: invalid sha1 pointer in cache-tree"), | |
730 | oid_to_hex(&it->oid)); | |
731 | errors_found |= ERROR_REFS; | |
732 | return 1; | |
733 | } | |
734 | obj->flags |= USED; | |
735 | fsck_put_object_name(&fsck_walk_options, &it->oid, ":"); | |
736 | mark_object_reachable(obj); | |
737 | if (obj->type != OBJ_TREE) | |
738 | err |= objerror(obj, _("non-tree in cache-tree")); | |
739 | } | |
740 | for (i = 0; i < it->subtree_nr; i++) | |
741 | err |= fsck_cache_tree(it->down[i]->cache_tree); | |
742 | return err; | |
743 | } | |
744 | ||
745 | static void mark_object_for_connectivity(const struct object_id *oid) | |
746 | { | |
747 | struct object *obj = lookup_unknown_object(oid); | |
748 | obj->flags |= HAS_OBJ; | |
749 | } | |
750 | ||
751 | static int mark_loose_for_connectivity(const struct object_id *oid, | |
752 | const char *path, | |
753 | void *data) | |
754 | { | |
755 | mark_object_for_connectivity(oid); | |
756 | return 0; | |
757 | } | |
758 | ||
759 | static int mark_packed_for_connectivity(const struct object_id *oid, | |
760 | struct packed_git *pack, | |
761 | uint32_t pos, | |
762 | void *data) | |
763 | { | |
764 | mark_object_for_connectivity(oid); | |
765 | return 0; | |
766 | } | |
767 | ||
768 | static char const * const fsck_usage[] = { | |
769 | N_("git fsck [<options>] [<object>...]"), | |
770 | NULL | |
771 | }; | |
772 | ||
773 | static struct option fsck_opts[] = { | |
774 | OPT__VERBOSE(&verbose, N_("be verbose")), | |
775 | OPT_BOOL(0, "unreachable", &show_unreachable, N_("show unreachable objects")), | |
776 | OPT_BOOL(0, "dangling", &show_dangling, N_("show dangling objects")), | |
777 | OPT_BOOL(0, "tags", &show_tags, N_("report tags")), | |
778 | OPT_BOOL(0, "root", &show_root, N_("report root nodes")), | |
779 | OPT_BOOL(0, "cache", &keep_cache_objects, N_("make index objects head nodes")), | |
780 | OPT_BOOL(0, "reflogs", &include_reflogs, N_("make reflogs head nodes (default)")), | |
781 | OPT_BOOL(0, "full", &check_full, N_("also consider packs and alternate objects")), | |
782 | OPT_BOOL(0, "connectivity-only", &connectivity_only, N_("check only connectivity")), | |
783 | OPT_BOOL(0, "strict", &check_strict, N_("enable more strict checking")), | |
784 | OPT_BOOL(0, "lost-found", &write_lost_and_found, | |
785 | N_("write dangling objects in .git/lost-found")), | |
786 | OPT_BOOL(0, "progress", &show_progress, N_("show progress")), | |
787 | OPT_BOOL(0, "name-objects", &name_objects, N_("show verbose names for reachable objects")), | |
788 | OPT_END(), | |
789 | }; | |
790 | ||
791 | int cmd_fsck(int argc, const char **argv, const char *prefix) | |
792 | { | |
793 | int i; | |
794 | struct object_directory *odb; | |
795 | ||
796 | /* fsck knows how to handle missing promisor objects */ | |
797 | fetch_if_missing = 0; | |
798 | ||
799 | errors_found = 0; | |
800 | read_replace_refs = 0; | |
801 | ||
802 | argc = parse_options(argc, argv, prefix, fsck_opts, fsck_usage, 0); | |
803 | ||
804 | fsck_walk_options.walk = mark_object; | |
805 | fsck_obj_options.walk = mark_used; | |
806 | fsck_obj_options.error_func = fsck_error_func; | |
807 | if (check_strict) | |
808 | fsck_obj_options.strict = 1; | |
809 | ||
810 | if (show_progress == -1) | |
811 | show_progress = isatty(2); | |
812 | if (verbose) | |
813 | show_progress = 0; | |
814 | ||
815 | if (write_lost_and_found) { | |
816 | check_full = 1; | |
817 | include_reflogs = 0; | |
818 | } | |
819 | ||
820 | if (name_objects) | |
821 | fsck_enable_object_names(&fsck_walk_options); | |
822 | ||
823 | git_config(fsck_config, NULL); | |
824 | ||
825 | if (connectivity_only) { | |
826 | for_each_loose_object(mark_loose_for_connectivity, NULL, 0); | |
827 | for_each_packed_object(mark_packed_for_connectivity, NULL, 0); | |
828 | } else { | |
829 | prepare_alt_odb(the_repository); | |
830 | for (odb = the_repository->objects->odb; odb; odb = odb->next) | |
831 | fsck_object_dir(odb->path); | |
832 | ||
833 | if (check_full) { | |
834 | struct packed_git *p; | |
835 | uint32_t total = 0, count = 0; | |
836 | struct progress *progress = NULL; | |
837 | ||
838 | if (show_progress) { | |
839 | for (p = get_all_packs(the_repository); p; | |
840 | p = p->next) { | |
841 | if (open_pack_index(p)) | |
842 | continue; | |
843 | total += p->num_objects; | |
844 | } | |
845 | ||
846 | progress = start_progress(_("Checking objects"), total); | |
847 | } | |
848 | for (p = get_all_packs(the_repository); p; | |
849 | p = p->next) { | |
850 | /* verify gives error messages itself */ | |
851 | if (verify_pack(the_repository, | |
852 | p, fsck_obj_buffer, | |
853 | progress, count)) | |
854 | errors_found |= ERROR_PACK; | |
855 | count += p->num_objects; | |
856 | } | |
857 | stop_progress(&progress); | |
858 | } | |
859 | ||
860 | if (fsck_finish(&fsck_obj_options)) | |
861 | errors_found |= ERROR_OBJECT; | |
862 | } | |
863 | ||
864 | for (i = 0; i < argc; i++) { | |
865 | const char *arg = argv[i]; | |
866 | struct object_id oid; | |
867 | if (!get_oid(arg, &oid)) { | |
868 | struct object *obj = lookup_object(the_repository, | |
869 | &oid); | |
870 | ||
871 | if (!obj || !(obj->flags & HAS_OBJ)) { | |
872 | if (is_promisor_object(&oid)) | |
873 | continue; | |
874 | error(_("%s: object missing"), oid_to_hex(&oid)); | |
875 | errors_found |= ERROR_OBJECT; | |
876 | continue; | |
877 | } | |
878 | ||
879 | obj->flags |= USED; | |
880 | fsck_put_object_name(&fsck_walk_options, &oid, | |
881 | "%s", arg); | |
882 | mark_object_reachable(obj); | |
883 | continue; | |
884 | } | |
885 | error(_("invalid parameter: expected sha1, got '%s'"), arg); | |
886 | errors_found |= ERROR_OBJECT; | |
887 | } | |
888 | ||
889 | /* | |
890 | * If we've not been given any explicit head information, do the | |
891 | * default ones from .git/refs. We also consider the index file | |
892 | * in this case (ie this implies --cache). | |
893 | */ | |
894 | if (!argc) { | |
895 | get_default_heads(); | |
896 | keep_cache_objects = 1; | |
897 | } | |
898 | ||
899 | if (keep_cache_objects) { | |
900 | verify_index_checksum = 1; | |
901 | verify_ce_order = 1; | |
902 | read_cache(); | |
903 | for (i = 0; i < active_nr; i++) { | |
904 | unsigned int mode; | |
905 | struct blob *blob; | |
906 | struct object *obj; | |
907 | ||
908 | mode = active_cache[i]->ce_mode; | |
909 | if (S_ISGITLINK(mode)) | |
910 | continue; | |
911 | blob = lookup_blob(the_repository, | |
912 | &active_cache[i]->oid); | |
913 | if (!blob) | |
914 | continue; | |
915 | obj = &blob->object; | |
916 | obj->flags |= USED; | |
917 | fsck_put_object_name(&fsck_walk_options, &obj->oid, | |
918 | ":%s", active_cache[i]->name); | |
919 | mark_object_reachable(obj); | |
920 | } | |
921 | if (active_cache_tree) | |
922 | fsck_cache_tree(active_cache_tree); | |
923 | } | |
924 | ||
925 | check_connectivity(); | |
926 | ||
927 | if (!git_config_get_bool("core.commitgraph", &i) && i) { | |
928 | struct child_process commit_graph_verify = CHILD_PROCESS_INIT; | |
929 | const char *verify_argv[] = { "commit-graph", "verify", NULL, NULL, NULL }; | |
930 | ||
931 | prepare_alt_odb(the_repository); | |
932 | for (odb = the_repository->objects->odb; odb; odb = odb->next) { | |
933 | child_process_init(&commit_graph_verify); | |
934 | commit_graph_verify.argv = verify_argv; | |
935 | commit_graph_verify.git_cmd = 1; | |
936 | verify_argv[2] = "--object-dir"; | |
937 | verify_argv[3] = odb->path; | |
938 | if (run_command(&commit_graph_verify)) | |
939 | errors_found |= ERROR_COMMIT_GRAPH; | |
940 | } | |
941 | } | |
942 | ||
943 | if (!git_config_get_bool("core.multipackindex", &i) && i) { | |
944 | struct child_process midx_verify = CHILD_PROCESS_INIT; | |
945 | const char *midx_argv[] = { "multi-pack-index", "verify", NULL, NULL, NULL }; | |
946 | ||
947 | prepare_alt_odb(the_repository); | |
948 | for (odb = the_repository->objects->odb; odb; odb = odb->next) { | |
949 | child_process_init(&midx_verify); | |
950 | midx_verify.argv = midx_argv; | |
951 | midx_verify.git_cmd = 1; | |
952 | midx_argv[2] = "--object-dir"; | |
953 | midx_argv[3] = odb->path; | |
954 | if (run_command(&midx_verify)) | |
955 | errors_found |= ERROR_COMMIT_GRAPH; | |
956 | } | |
957 | } | |
958 | ||
959 | return errors_found; | |
960 | } |