help: add --config to list all available config
[git/git.git] / fsck.c
CommitLineData
355885d5
MK
1#include "cache.h"
2#include "object.h"
3#include "blob.h"
4#include "tree.h"
5#include "tree-walk.h"
6#include "commit.h"
7#include "tag.h"
8#include "fsck.h"
cec097be 9#include "refs.h"
a18fcc9f 10#include "utf8.h"
cd94c6f9 11#include "sha1-array.h"
7b35efd7 12#include "decorate.h"
3ac68a93 13#include "help.h"
355885d5 14
f50c4407 15#define FSCK_FATAL -1
f27d05b1 16#define FSCK_INFO -2
f50c4407 17
c99ba492 18#define FOREACH_MSG_ID(FUNC) \
f50c4407
JS
19 /* fatal errors */ \
20 FUNC(NUL_IN_HEADER, FATAL) \
21 FUNC(UNTERMINATED_HEADER, FATAL) \
c99ba492
JS
22 /* errors */ \
23 FUNC(BAD_DATE, ERROR) \
24 FUNC(BAD_DATE_OVERFLOW, ERROR) \
25 FUNC(BAD_EMAIL, ERROR) \
26 FUNC(BAD_NAME, ERROR) \
27 FUNC(BAD_OBJECT_SHA1, ERROR) \
28 FUNC(BAD_PARENT_SHA1, ERROR) \
29 FUNC(BAD_TAG_OBJECT, ERROR) \
30 FUNC(BAD_TIMEZONE, ERROR) \
31 FUNC(BAD_TREE, ERROR) \
32 FUNC(BAD_TREE_SHA1, ERROR) \
33 FUNC(BAD_TYPE, ERROR) \
34 FUNC(DUPLICATE_ENTRIES, ERROR) \
35 FUNC(MISSING_AUTHOR, ERROR) \
36 FUNC(MISSING_COMMITTER, ERROR) \
37 FUNC(MISSING_EMAIL, ERROR) \
38 FUNC(MISSING_GRAFT, ERROR) \
39 FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
40 FUNC(MISSING_OBJECT, ERROR) \
41 FUNC(MISSING_PARENT, ERROR) \
42 FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
43 FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
44 FUNC(MISSING_TAG, ERROR) \
45 FUNC(MISSING_TAG_ENTRY, ERROR) \
46 FUNC(MISSING_TAG_OBJECT, ERROR) \
47 FUNC(MISSING_TREE, ERROR) \
48 FUNC(MISSING_TYPE, ERROR) \
49 FUNC(MISSING_TYPE_ENTRY, ERROR) \
c9ad147f 50 FUNC(MULTIPLE_AUTHORS, ERROR) \
c99ba492
JS
51 FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
52 FUNC(TREE_NOT_SORTED, ERROR) \
53 FUNC(UNKNOWN_TYPE, ERROR) \
c99ba492
JS
54 FUNC(ZERO_PADDED_DATE, ERROR) \
55 /* warnings */ \
56 FUNC(BAD_FILEMODE, WARN) \
c99ba492
JS
57 FUNC(EMPTY_NAME, WARN) \
58 FUNC(FULL_PATHNAME, WARN) \
59 FUNC(HAS_DOT, WARN) \
60 FUNC(HAS_DOTDOT, WARN) \
61 FUNC(HAS_DOTGIT, WARN) \
c99ba492 62 FUNC(NULL_SHA1, WARN) \
f27d05b1 63 FUNC(ZERO_PADDED_FILEMODE, WARN) \
6d2d780f 64 FUNC(NUL_IN_COMMIT, WARN) \
f27d05b1
JS
65 /* infos (reported as warnings, but ignored by default) */ \
66 FUNC(BAD_TAG_NAME, INFO) \
67 FUNC(MISSING_TAGGER_ENTRY, INFO)
c99ba492
JS
68
69#define MSG_ID(id, msg_type) FSCK_MSG_##id,
70enum fsck_msg_id {
71 FOREACH_MSG_ID(MSG_ID)
72 FSCK_MSG_MAX
73};
74#undef MSG_ID
75
f417eed8
JS
76#define STR(x) #x
77#define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
c99ba492 78static struct {
f417eed8
JS
79 const char *id_string;
80 const char *downcased;
c99ba492
JS
81 int msg_type;
82} msg_id_info[FSCK_MSG_MAX + 1] = {
83 FOREACH_MSG_ID(MSG_ID)
f417eed8 84 { NULL, NULL, -1 }
c99ba492
JS
85};
86#undef MSG_ID
87
a46baac6 88static void prepare_msg_ids(void)
f417eed8
JS
89{
90 int i;
91
a46baac6
NTND
92 if (msg_id_info[0].downcased)
93 return;
94
95 /* convert id_string to lower case, without underscores. */
96 for (i = 0; i < FSCK_MSG_MAX; i++) {
97 const char *p = msg_id_info[i].id_string;
98 int len = strlen(p);
99 char *q = xmalloc(len);
100
101 msg_id_info[i].downcased = q;
102 while (*p)
103 if (*p == '_')
104 p++;
105 else
106 *(q)++ = tolower(*(p)++);
107 *q = '\0';
f417eed8 108 }
a46baac6
NTND
109}
110
111static int parse_msg_id(const char *text)
112{
113 int i;
114
115 prepare_msg_ids();
f417eed8
JS
116
117 for (i = 0; i < FSCK_MSG_MAX; i++)
118 if (!strcmp(text, msg_id_info[i].downcased))
119 return i;
120
121 return -1;
122}
123
3ac68a93
NTND
124void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
125{
126 int i;
127
128 prepare_msg_ids();
129
130 /* TODO: we can do better by producing camelCase names */
131 for (i = 0; i < FSCK_MSG_MAX; i++)
132 list_config_item(list, prefix, msg_id_info[i].downcased);
133}
134
c99ba492
JS
135static int fsck_msg_type(enum fsck_msg_id msg_id,
136 struct fsck_options *options)
137{
138 int msg_type;
139
0282f4dc
JS
140 assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
141
142 if (options->msg_type)
143 msg_type = options->msg_type[msg_id];
144 else {
145 msg_type = msg_id_info[msg_id].msg_type;
146 if (options->strict && msg_type == FSCK_WARN)
147 msg_type = FSCK_ERROR;
148 }
c99ba492
JS
149
150 return msg_type;
151}
152
cd94c6f9
JS
153static void init_skiplist(struct fsck_options *options, const char *path)
154{
910650d2 155 static struct oid_array skiplist = OID_ARRAY_INIT;
cd94c6f9 156 int sorted, fd;
365c27fb 157 char buffer[GIT_MAX_HEXSZ + 1];
158 struct object_id oid;
cd94c6f9
JS
159
160 if (options->skiplist)
161 sorted = options->skiplist->sorted;
162 else {
163 sorted = 1;
164 options->skiplist = &skiplist;
165 }
166
167 fd = open(path, O_RDONLY);
168 if (fd < 0)
169 die("Could not open skip list: %s", path);
170 for (;;) {
365c27fb 171 const char *p;
cd94c6f9
JS
172 int result = read_in_full(fd, buffer, sizeof(buffer));
173 if (result < 0)
174 die_errno("Could not read '%s'", path);
175 if (!result)
176 break;
365c27fb 177 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
cd94c6f9 178 die("Invalid SHA-1: %s", buffer);
910650d2 179 oid_array_append(&skiplist, &oid);
cd94c6f9 180 if (sorted && skiplist.nr > 1 &&
ee3051bd 181 oidcmp(&skiplist.oid[skiplist.nr - 2],
182 &oid) > 0)
cd94c6f9
JS
183 sorted = 0;
184 }
185 close(fd);
186
187 if (sorted)
188 skiplist.sorted = 1;
189}
190
0282f4dc
JS
191static int parse_msg_type(const char *str)
192{
193 if (!strcmp(str, "error"))
194 return FSCK_ERROR;
195 else if (!strcmp(str, "warn"))
196 return FSCK_WARN;
efaba7cc
JS
197 else if (!strcmp(str, "ignore"))
198 return FSCK_IGNORE;
0282f4dc
JS
199 else
200 die("Unknown fsck message type: '%s'", str);
201}
202
5d477a33
JS
203int is_valid_msg_type(const char *msg_id, const char *msg_type)
204{
205 if (parse_msg_id(msg_id) < 0)
206 return 0;
207 parse_msg_type(msg_type);
208 return 1;
209}
210
0282f4dc
JS
211void fsck_set_msg_type(struct fsck_options *options,
212 const char *msg_id, const char *msg_type)
213{
214 int id = parse_msg_id(msg_id), type;
215
216 if (id < 0)
217 die("Unhandled message id: %s", msg_id);
218 type = parse_msg_type(msg_type);
219
f50c4407
JS
220 if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
221 die("Cannot demote %s to %s", msg_id, msg_type);
222
0282f4dc
JS
223 if (!options->msg_type) {
224 int i;
b32fa95f
JK
225 int *msg_type;
226 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
0282f4dc
JS
227 for (i = 0; i < FSCK_MSG_MAX; i++)
228 msg_type[i] = fsck_msg_type(i, options);
229 options->msg_type = msg_type;
230 }
231
232 options->msg_type[id] = type;
233}
234
235void fsck_set_msg_types(struct fsck_options *options, const char *values)
236{
237 char *buf = xstrdup(values), *to_free = buf;
238 int done = 0;
239
240 while (!done) {
241 int len = strcspn(buf, " ,|"), equal;
242
243 done = !buf[len];
244 if (!len) {
245 buf++;
246 continue;
247 }
248 buf[len] = '\0';
249
250 for (equal = 0;
251 equal < len && buf[equal] != '=' && buf[equal] != ':';
252 equal++)
253 buf[equal] = tolower(buf[equal]);
254 buf[equal] = '\0';
255
cd94c6f9
JS
256 if (!strcmp(buf, "skiplist")) {
257 if (equal == len)
258 die("skiplist requires a path");
259 init_skiplist(options, buf + equal + 1);
260 buf += len + 1;
261 continue;
262 }
263
0282f4dc
JS
264 if (equal == len)
265 die("Missing '=': '%s'", buf);
355885d5 266
0282f4dc
JS
267 fsck_set_msg_type(options, buf, buf + equal + 1);
268 buf += len + 1;
269 }
270 free(to_free);
271}
272
71ab8fa8
JS
273static void append_msg_id(struct strbuf *sb, const char *msg_id)
274{
275 for (;;) {
276 char c = *(msg_id)++;
277
278 if (!c)
279 break;
280 if (c != '_')
281 strbuf_addch(sb, tolower(c));
282 else {
283 assert(*msg_id);
284 strbuf_addch(sb, *(msg_id)++);
285 }
286 }
287
288 strbuf_addstr(sb, ": ");
289}
290
c99ba492
JS
291__attribute__((format (printf, 4, 5)))
292static int report(struct fsck_options *options, struct object *object,
293 enum fsck_msg_id id, const char *fmt, ...)
294{
295 va_list ap;
296 struct strbuf sb = STRBUF_INIT;
297 int msg_type = fsck_msg_type(id, options), result;
298
efaba7cc
JS
299 if (msg_type == FSCK_IGNORE)
300 return 0;
301
cd94c6f9 302 if (options->skiplist && object &&
910650d2 303 oid_array_lookup(options->skiplist, &object->oid) >= 0)
cd94c6f9
JS
304 return 0;
305
f50c4407
JS
306 if (msg_type == FSCK_FATAL)
307 msg_type = FSCK_ERROR;
f27d05b1
JS
308 else if (msg_type == FSCK_INFO)
309 msg_type = FSCK_WARN;
f50c4407 310
71ab8fa8
JS
311 append_msg_id(&sb, msg_id_info[id].id_string);
312
c99ba492
JS
313 va_start(ap, fmt);
314 strbuf_vaddf(&sb, fmt, ap);
1cd772cc 315 result = options->error_func(options, object, msg_type, sb.buf);
c99ba492
JS
316 strbuf_release(&sb);
317 va_end(ap);
318
319 return result;
320}
321
7b35efd7
JS
322static char *get_object_name(struct fsck_options *options, struct object *obj)
323{
324 if (!options->object_names)
325 return NULL;
326 return lookup_decoration(options->object_names, obj);
327}
328
329static void put_object_name(struct fsck_options *options, struct object *obj,
330 const char *fmt, ...)
331{
332 va_list ap;
333 struct strbuf buf = STRBUF_INIT;
334 char *existing;
335
336 if (!options->object_names)
337 return;
338 existing = lookup_decoration(options->object_names, obj);
339 if (existing)
340 return;
341 va_start(ap, fmt);
342 strbuf_vaddf(&buf, fmt, ap);
343 add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
344 va_end(ap);
345}
346
90cf590f
JS
347static const char *describe_object(struct fsck_options *o, struct object *obj)
348{
349 static struct strbuf buf = STRBUF_INIT;
350 char *name;
351
352 strbuf_reset(&buf);
353 strbuf_addstr(&buf, oid_to_hex(&obj->oid));
354 if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
355 strbuf_addf(&buf, " (%s)", name);
356
357 return buf.buf;
358}
359
22410549 360static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
355885d5
MK
361{
362 struct tree_desc desc;
363 struct name_entry entry;
364 int res = 0;
7b35efd7 365 const char *name;
355885d5
MK
366
367 if (parse_tree(tree))
368 return -1;
369
7b35efd7 370 name = get_object_name(options, &tree->object);
8354fa3d
DT
371 if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
372 return -1;
373 while (tree_entry_gently(&desc, &entry)) {
7b35efd7 374 struct object *obj;
355885d5
MK
375 int result;
376
377 if (S_ISGITLINK(entry.mode))
378 continue;
7b35efd7
JS
379
380 if (S_ISDIR(entry.mode)) {
2720f6db
RS
381 obj = (struct object *)lookup_tree(entry.oid);
382 if (name && obj)
7b35efd7
JS
383 put_object_name(options, obj, "%s%s/", name,
384 entry.path);
385 result = options->walk(obj, OBJ_TREE, data, options);
386 }
387 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
2720f6db
RS
388 obj = (struct object *)lookup_blob(entry.oid);
389 if (name && obj)
7b35efd7
JS
390 put_object_name(options, obj, "%s%s", name,
391 entry.path);
392 result = options->walk(obj, OBJ_BLOB, data, options);
393 }
355885d5 394 else {
82247e9b 395 result = error("in tree %s: entry %s has bad mode %.6o",
90cf590f 396 describe_object(options, &tree->object), entry.path, entry.mode);
355885d5
MK
397 }
398 if (result < 0)
399 return result;
400 if (!res)
401 res = result;
402 }
403 return res;
404}
405
22410549 406static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
355885d5 407{
7b35efd7 408 int counter = 0, generation = 0, name_prefix_len = 0;
355885d5
MK
409 struct commit_list *parents;
410 int res;
411 int result;
7b35efd7 412 const char *name;
355885d5
MK
413
414 if (parse_commit(commit))
415 return -1;
416
7b35efd7
JS
417 name = get_object_name(options, &commit->object);
418 if (name)
2e27bd77
DS
419 put_object_name(options, &get_commit_tree(commit)->object,
420 "%s:", name);
7b35efd7 421
2e27bd77
DS
422 result = options->walk((struct object *)get_commit_tree(commit),
423 OBJ_TREE, data, options);
355885d5
MK
424 if (result < 0)
425 return result;
426 res = result;
427
428 parents = commit->parents;
7b35efd7
JS
429 if (name && parents) {
430 int len = strlen(name), power;
431
432 if (len && name[len - 1] == '^') {
433 generation = 1;
434 name_prefix_len = len - 1;
435 }
436 else { /* parse ~<generation> suffix */
437 for (generation = 0, power = 1;
438 len && isdigit(name[len - 1]);
439 power *= 10)
440 generation += power * (name[--len] - '0');
441 if (power > 1 && len && name[len - 1] == '~')
442 name_prefix_len = len - 1;
443 }
444 }
445
355885d5 446 while (parents) {
7b35efd7
JS
447 if (name) {
448 struct object *obj = &parents->item->object;
449
450 if (++counter > 1)
451 put_object_name(options, obj, "%s^%d",
452 name, counter);
453 else if (generation > 0)
454 put_object_name(options, obj, "%.*s~%d",
455 name_prefix_len, name, generation + 1);
456 else
457 put_object_name(options, obj, "%s^", name);
458 }
22410549 459 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
355885d5
MK
460 if (result < 0)
461 return result;
462 if (!res)
463 res = result;
464 parents = parents->next;
465 }
466 return res;
467}
468
22410549 469static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
355885d5 470{
7b35efd7
JS
471 char *name = get_object_name(options, &tag->object);
472
355885d5
MK
473 if (parse_tag(tag))
474 return -1;
7b35efd7
JS
475 if (name)
476 put_object_name(options, tag->tagged, "%s", name);
22410549 477 return options->walk(tag->tagged, OBJ_ANY, data, options);
355885d5
MK
478}
479
22410549 480int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
355885d5
MK
481{
482 if (!obj)
483 return -1;
a2b22854
JK
484
485 if (obj->type == OBJ_NONE)
c251c83d 486 parse_object(&obj->oid);
a2b22854 487
355885d5
MK
488 switch (obj->type) {
489 case OBJ_BLOB:
490 return 0;
491 case OBJ_TREE:
22410549 492 return fsck_walk_tree((struct tree *)obj, data, options);
355885d5 493 case OBJ_COMMIT:
22410549 494 return fsck_walk_commit((struct commit *)obj, data, options);
355885d5 495 case OBJ_TAG:
22410549 496 return fsck_walk_tag((struct tag *)obj, data, options);
355885d5 497 default:
90cf590f 498 error("Unknown object type for %s", describe_object(options, obj));
355885d5
MK
499 return -1;
500 }
501}
ba002f3b
MK
502
503/*
504 * The entries in a tree are ordered in the _path_ order,
505 * which means that a directory entry is ordered by adding
506 * a slash to the end of it.
507 *
508 * So a directory called "a" is ordered _after_ a file
509 * called "a.c", because "a/" sorts after "a.c".
510 */
511#define TREE_UNORDERED (-1)
512#define TREE_HAS_DUPS (-2)
513
514static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
515{
516 int len1 = strlen(name1);
517 int len2 = strlen(name2);
518 int len = len1 < len2 ? len1 : len2;
519 unsigned char c1, c2;
520 int cmp;
521
522 cmp = memcmp(name1, name2, len);
523 if (cmp < 0)
524 return 0;
525 if (cmp > 0)
526 return TREE_UNORDERED;
527
528 /*
529 * Ok, the first <len> characters are the same.
530 * Now we need to order the next one, but turn
531 * a '\0' into a '/' for a directory entry.
532 */
533 c1 = name1[len];
534 c2 = name2[len];
535 if (!c1 && !c2)
536 /*
537 * git-write-tree used to write out a nonsense tree that has
538 * entries with the same name, one blob and one tree. Make
539 * sure we do not have duplicate entries.
540 */
541 return TREE_HAS_DUPS;
542 if (!c1 && S_ISDIR(mode1))
543 c1 = '/';
544 if (!c2 && S_ISDIR(mode2))
545 c2 = '/';
546 return c1 < c2 ? 0 : TREE_UNORDERED;
547}
548
22410549 549static int fsck_tree(struct tree *item, struct fsck_options *options)
ba002f3b 550{
8354fa3d 551 int retval = 0;
c479d14a 552 int has_null_sha1 = 0;
ba002f3b
MK
553 int has_full_path = 0;
554 int has_empty_name = 0;
5d34a435
JK
555 int has_dot = 0;
556 int has_dotdot = 0;
5c17f512 557 int has_dotgit = 0;
ba002f3b
MK
558 int has_zero_pad = 0;
559 int has_bad_modes = 0;
560 int has_dup_entries = 0;
561 int not_properly_sorted = 0;
562 struct tree_desc desc;
563 unsigned o_mode;
564 const char *o_name;
ba002f3b 565
8354fa3d
DT
566 if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
567 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
568 return retval;
569 }
ba002f3b
MK
570
571 o_mode = 0;
572 o_name = NULL;
ba002f3b
MK
573
574 while (desc.size) {
575 unsigned mode;
576 const char *name;
ce6663a9 577 const struct object_id *oid;
ba002f3b 578
ce6663a9 579 oid = tree_entry_extract(&desc, &name, &mode);
ba002f3b 580
ce6663a9 581 has_null_sha1 |= is_null_oid(oid);
effd12ec
HS
582 has_full_path |= !!strchr(name, '/');
583 has_empty_name |= !*name;
584 has_dot |= !strcmp(name, ".");
585 has_dotdot |= !strcmp(name, "..");
5e519fb8
JH
586 has_dotgit |= (!strcmp(name, ".git") ||
587 is_hfs_dotgit(name) ||
588 is_ntfs_dotgit(name));
ba002f3b 589 has_zero_pad |= *(char *)desc.buffer == '0';
8354fa3d
DT
590 if (update_tree_entry_gently(&desc)) {
591 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
592 break;
593 }
ba002f3b
MK
594
595 switch (mode) {
596 /*
597 * Standard modes..
598 */
599 case S_IFREG | 0755:
600 case S_IFREG | 0644:
601 case S_IFLNK:
602 case S_IFDIR:
603 case S_IFGITLINK:
604 break;
605 /*
606 * This is nonstandard, but we had a few of these
607 * early on when we honored the full set of mode
608 * bits..
609 */
610 case S_IFREG | 0664:
22410549 611 if (!options->strict)
ba002f3b 612 break;
1cf01a34 613 /* fallthrough */
ba002f3b
MK
614 default:
615 has_bad_modes = 1;
616 }
617
618 if (o_name) {
619 switch (verify_ordered(o_mode, o_name, mode, name)) {
620 case TREE_UNORDERED:
621 not_properly_sorted = 1;
622 break;
623 case TREE_HAS_DUPS:
624 has_dup_entries = 1;
625 break;
626 default:
627 break;
628 }
629 }
630
631 o_mode = mode;
632 o_name = name;
ba002f3b
MK
633 }
634
c479d14a 635 if (has_null_sha1)
c99ba492 636 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
ba002f3b 637 if (has_full_path)
c99ba492 638 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
ba002f3b 639 if (has_empty_name)
c99ba492 640 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
5d34a435 641 if (has_dot)
c99ba492 642 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
5d34a435 643 if (has_dotdot)
c99ba492 644 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
5c17f512 645 if (has_dotgit)
c99ba492 646 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
ba002f3b 647 if (has_zero_pad)
c99ba492 648 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
ba002f3b 649 if (has_bad_modes)
c99ba492 650 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
ba002f3b 651 if (has_dup_entries)
c99ba492 652 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
ba002f3b 653 if (not_properly_sorted)
c99ba492 654 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
ba002f3b
MK
655 return retval;
656}
657
84d18c0b 658static int verify_headers(const void *data, unsigned long size,
b2f44feb 659 struct object *obj, struct fsck_options *options)
4d0d8975
JS
660{
661 const char *buffer = (const char *)data;
662 unsigned long i;
663
664 for (i = 0; i < size; i++) {
665 switch (buffer[i]) {
666 case '\0':
c99ba492
JS
667 return report(options, obj,
668 FSCK_MSG_NUL_IN_HEADER,
669 "unterminated header: NUL at offset %ld", i);
4d0d8975
JS
670 case '\n':
671 if (i + 1 < size && buffer[i + 1] == '\n')
672 return 0;
673 }
674 }
675
84d18c0b
JH
676 /*
677 * We did not find double-LF that separates the header
678 * and the body. Not having a body is not a crime but
679 * we do want to see the terminating LF for the last header
680 * line.
681 */
682 if (size && buffer[size - 1] == '\n')
683 return 0;
684
c99ba492
JS
685 return report(options, obj,
686 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
4d0d8975
JS
687}
688
22410549 689static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
daae1922 690{
e6826e33 691 const char *p = *ident;
d4b8de04
JK
692 char *end;
693
e6826e33
JS
694 *ident = strchrnul(*ident, '\n');
695 if (**ident == '\n')
696 (*ident)++;
697
698 if (*p == '<')
c99ba492 699 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
700 p += strcspn(p, "<>\n");
701 if (*p == '>')
c99ba492 702 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
e6826e33 703 if (*p != '<')
c99ba492 704 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
e6826e33 705 if (p[-1] != ' ')
c99ba492 706 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
707 p++;
708 p += strcspn(p, "<>\n");
709 if (*p != '>')
c99ba492 710 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
e6826e33
JS
711 p++;
712 if (*p != ' ')
c99ba492 713 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
e6826e33
JS
714 p++;
715 if (*p == '0' && p[1] != ' ')
c99ba492 716 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
1aeb7e75 717 if (date_overflows(parse_timestamp(p, &end, 10)))
c99ba492 718 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
e6826e33 719 if ((end == p || *end != ' '))
c99ba492 720 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
e6826e33
JS
721 p = end + 1;
722 if ((*p != '+' && *p != '-') ||
723 !isdigit(p[1]) ||
724 !isdigit(p[2]) ||
725 !isdigit(p[3]) ||
726 !isdigit(p[4]) ||
727 (p[5] != '\n'))
c99ba492 728 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
e6826e33 729 p += 6;
daae1922
JN
730 return 0;
731}
732
bc6b8fc1 733static int fsck_commit_buffer(struct commit *commit, const char *buffer,
22410549 734 unsigned long size, struct fsck_options *options)
ba002f3b 735{
ba002f3b
MK
736 unsigned char tree_sha1[20], sha1[20];
737 struct commit_graft *graft;
c9ad147f 738 unsigned parent_count, parent_line_count = 0, author_count;
daae1922 739 int err;
6d2d780f 740 const char *buffer_begin = buffer;
ba002f3b 741
b2f44feb 742 if (verify_headers(buffer, size, &commit->object, options))
4d0d8975
JS
743 return -1;
744
cf4fff57 745 if (!skip_prefix(buffer, "tree ", &buffer))
c99ba492 746 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
b3584761
JS
747 if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
748 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
749 if (err)
750 return err;
751 }
2d820a61 752 buffer += 41;
cf4fff57 753 while (skip_prefix(buffer, "parent ", &buffer)) {
b3584761
JS
754 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
755 err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
756 if (err)
757 return err;
758 }
2d820a61 759 buffer += 41;
9d02150c 760 parent_line_count++;
ba002f3b 761 }
8b65a34c 762 graft = lookup_commit_graft(&commit->object.oid);
9d02150c 763 parent_count = commit_list_count(commit->parents);
ba002f3b 764 if (graft) {
9d02150c 765 if (graft->nr_parent == -1 && !parent_count)
ba002f3b 766 ; /* shallow commit */
b3584761
JS
767 else if (graft->nr_parent != parent_count) {
768 err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
769 if (err)
770 return err;
771 }
ba002f3b 772 } else {
b3584761
JS
773 if (parent_count != parent_line_count) {
774 err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
775 if (err)
776 return err;
777 }
ba002f3b 778 }
c9ad147f
JS
779 author_count = 0;
780 while (skip_prefix(buffer, "author ", &buffer)) {
781 author_count++;
782 err = fsck_ident(&buffer, &commit->object, options);
783 if (err)
784 return err;
ba002f3b 785 }
c9ad147f
JS
786 if (author_count < 1)
787 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
788 else if (author_count > 1)
789 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
daae1922
JN
790 if (err)
791 return err;
cf4fff57 792 if (!skip_prefix(buffer, "committer ", &buffer))
c99ba492 793 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
22410549 794 err = fsck_ident(&buffer, &commit->object, options);
daae1922
JN
795 if (err)
796 return err;
2e27bd77 797 if (!get_commit_tree(commit)) {
5af29718
JH
798 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
799 if (err)
800 return err;
801 }
6d2d780f
JH
802 if (memchr(buffer_begin, '\0', size)) {
803 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
804 "NUL byte in the commit object body");
805 if (err)
806 return err;
807 }
ba002f3b
MK
808 return 0;
809}
810
90a398bb 811static int fsck_commit(struct commit *commit, const char *data,
22410549 812 unsigned long size, struct fsck_options *options)
bc6b8fc1 813{
90a398bb 814 const char *buffer = data ? data : get_commit_buffer(commit, &size);
22410549 815 int ret = fsck_commit_buffer(commit, buffer, size, options);
90a398bb
JS
816 if (!data)
817 unuse_commit_buffer(commit, buffer);
bc6b8fc1
JK
818 return ret;
819}
820
cec097be 821static int fsck_tag_buffer(struct tag *tag, const char *data,
22410549 822 unsigned long size, struct fsck_options *options)
cec097be
JS
823{
824 unsigned char sha1[20];
825 int ret = 0;
826 const char *buffer;
827 char *to_free = NULL, *eol;
828 struct strbuf sb = STRBUF_INIT;
829
830 if (data)
831 buffer = data;
832 else {
833 enum object_type type;
834
835 buffer = to_free =
b4f5aca4 836 read_object_file(&tag->object.oid, &type, &size);
cec097be 837 if (!buffer)
c99ba492
JS
838 return report(options, &tag->object,
839 FSCK_MSG_MISSING_TAG_OBJECT,
cec097be
JS
840 "cannot read tag object");
841
842 if (type != OBJ_TAG) {
c99ba492
JS
843 ret = report(options, &tag->object,
844 FSCK_MSG_TAG_OBJECT_NOT_TAG,
cec097be 845 "expected tag got %s",
debca9d2 846 type_name(type));
cec097be
JS
847 goto done;
848 }
849 }
850
8a272f29
RS
851 ret = verify_headers(buffer, size, &tag->object, options);
852 if (ret)
cec097be
JS
853 goto done;
854
855 if (!skip_prefix(buffer, "object ", &buffer)) {
c99ba492 856 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
cec097be
JS
857 goto done;
858 }
859 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
c99ba492 860 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
7d7d5b05
JS
861 if (ret)
862 goto done;
cec097be
JS
863 }
864 buffer += 41;
865
866 if (!skip_prefix(buffer, "type ", &buffer)) {
c99ba492 867 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
cec097be
JS
868 goto done;
869 }
870 eol = strchr(buffer, '\n');
871 if (!eol) {
c99ba492 872 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
cec097be
JS
873 goto done;
874 }
875 if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
c99ba492 876 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
cec097be
JS
877 if (ret)
878 goto done;
879 buffer = eol + 1;
880
881 if (!skip_prefix(buffer, "tag ", &buffer)) {
c99ba492 882 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
cec097be
JS
883 goto done;
884 }
885 eol = strchr(buffer, '\n');
886 if (!eol) {
c99ba492 887 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
cec097be
JS
888 goto done;
889 }
890 strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
f27d05b1
JS
891 if (check_refname_format(sb.buf, 0)) {
892 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
c99ba492 893 "invalid 'tag' name: %.*s",
7add4419 894 (int)(eol - buffer), buffer);
f27d05b1
JS
895 if (ret)
896 goto done;
897 }
cec097be
JS
898 buffer = eol + 1;
899
f27d05b1 900 if (!skip_prefix(buffer, "tagger ", &buffer)) {
cec097be 901 /* early tags do not contain 'tagger' lines; warn only */
f27d05b1
JS
902 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
903 if (ret)
904 goto done;
905 }
cec097be 906 else
22410549 907 ret = fsck_ident(&buffer, &tag->object, options);
cec097be
JS
908
909done:
910 strbuf_release(&sb);
911 free(to_free);
912 return ret;
913}
914
90a398bb 915static int fsck_tag(struct tag *tag, const char *data,
22410549 916 unsigned long size, struct fsck_options *options)
ba002f3b
MK
917{
918 struct object *tagged = tag->tagged;
919
920 if (!tagged)
c99ba492 921 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
cec097be 922
22410549 923 return fsck_tag_buffer(tag, data, size, options);
ba002f3b
MK
924}
925
90a398bb 926int fsck_object(struct object *obj, void *data, unsigned long size,
22410549 927 struct fsck_options *options)
ba002f3b
MK
928{
929 if (!obj)
c99ba492 930 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
ba002f3b
MK
931
932 if (obj->type == OBJ_BLOB)
933 return 0;
934 if (obj->type == OBJ_TREE)
22410549 935 return fsck_tree((struct tree *) obj, options);
ba002f3b 936 if (obj->type == OBJ_COMMIT)
90a398bb 937 return fsck_commit((struct commit *) obj, (const char *) data,
22410549 938 size, options);
ba002f3b 939 if (obj->type == OBJ_TAG)
90a398bb 940 return fsck_tag((struct tag *) obj, (const char *) data,
22410549 941 size, options);
ba002f3b 942
c99ba492 943 return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
ba002f3b
MK
944 obj->type);
945}
d6ffc8d7 946
1cd772cc
JS
947int fsck_error_function(struct fsck_options *o,
948 struct object *obj, int msg_type, const char *message)
d6ffc8d7 949{
0282f4dc 950 if (msg_type == FSCK_WARN) {
90cf590f 951 warning("object %s: %s", describe_object(o, obj), message);
0282f4dc
JS
952 return 0;
953 }
90cf590f 954 error("object %s: %s", describe_object(o, obj), message);
d6ffc8d7
MK
955 return 1;
956}