Convert struct object to object_id
[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"
355885d5 12
f50c4407 13#define FSCK_FATAL -1
f27d05b1 14#define FSCK_INFO -2
f50c4407 15
c99ba492 16#define FOREACH_MSG_ID(FUNC) \
f50c4407
JS
17 /* fatal errors */ \
18 FUNC(NUL_IN_HEADER, FATAL) \
19 FUNC(UNTERMINATED_HEADER, FATAL) \
c99ba492
JS
20 /* errors */ \
21 FUNC(BAD_DATE, ERROR) \
22 FUNC(BAD_DATE_OVERFLOW, ERROR) \
23 FUNC(BAD_EMAIL, ERROR) \
24 FUNC(BAD_NAME, ERROR) \
25 FUNC(BAD_OBJECT_SHA1, ERROR) \
26 FUNC(BAD_PARENT_SHA1, ERROR) \
27 FUNC(BAD_TAG_OBJECT, ERROR) \
28 FUNC(BAD_TIMEZONE, ERROR) \
29 FUNC(BAD_TREE, ERROR) \
30 FUNC(BAD_TREE_SHA1, ERROR) \
31 FUNC(BAD_TYPE, ERROR) \
32 FUNC(DUPLICATE_ENTRIES, ERROR) \
33 FUNC(MISSING_AUTHOR, ERROR) \
34 FUNC(MISSING_COMMITTER, ERROR) \
35 FUNC(MISSING_EMAIL, ERROR) \
36 FUNC(MISSING_GRAFT, ERROR) \
37 FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
38 FUNC(MISSING_OBJECT, ERROR) \
39 FUNC(MISSING_PARENT, ERROR) \
40 FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
41 FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
42 FUNC(MISSING_TAG, ERROR) \
43 FUNC(MISSING_TAG_ENTRY, ERROR) \
44 FUNC(MISSING_TAG_OBJECT, ERROR) \
45 FUNC(MISSING_TREE, ERROR) \
46 FUNC(MISSING_TYPE, ERROR) \
47 FUNC(MISSING_TYPE_ENTRY, ERROR) \
c9ad147f 48 FUNC(MULTIPLE_AUTHORS, ERROR) \
c99ba492
JS
49 FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
50 FUNC(TREE_NOT_SORTED, ERROR) \
51 FUNC(UNKNOWN_TYPE, ERROR) \
c99ba492
JS
52 FUNC(ZERO_PADDED_DATE, ERROR) \
53 /* warnings */ \
54 FUNC(BAD_FILEMODE, WARN) \
c99ba492
JS
55 FUNC(EMPTY_NAME, WARN) \
56 FUNC(FULL_PATHNAME, WARN) \
57 FUNC(HAS_DOT, WARN) \
58 FUNC(HAS_DOTDOT, WARN) \
59 FUNC(HAS_DOTGIT, WARN) \
c99ba492 60 FUNC(NULL_SHA1, WARN) \
f27d05b1
JS
61 FUNC(ZERO_PADDED_FILEMODE, WARN) \
62 /* infos (reported as warnings, but ignored by default) */ \
63 FUNC(BAD_TAG_NAME, INFO) \
64 FUNC(MISSING_TAGGER_ENTRY, INFO)
c99ba492
JS
65
66#define MSG_ID(id, msg_type) FSCK_MSG_##id,
67enum fsck_msg_id {
68 FOREACH_MSG_ID(MSG_ID)
69 FSCK_MSG_MAX
70};
71#undef MSG_ID
72
f417eed8
JS
73#define STR(x) #x
74#define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
c99ba492 75static struct {
f417eed8
JS
76 const char *id_string;
77 const char *downcased;
c99ba492
JS
78 int msg_type;
79} msg_id_info[FSCK_MSG_MAX + 1] = {
80 FOREACH_MSG_ID(MSG_ID)
f417eed8 81 { NULL, NULL, -1 }
c99ba492
JS
82};
83#undef MSG_ID
84
f417eed8
JS
85static int parse_msg_id(const char *text)
86{
87 int i;
88
89 if (!msg_id_info[0].downcased) {
90 /* convert id_string to lower case, without underscores. */
91 for (i = 0; i < FSCK_MSG_MAX; i++) {
92 const char *p = msg_id_info[i].id_string;
93 int len = strlen(p);
94 char *q = xmalloc(len);
95
96 msg_id_info[i].downcased = q;
97 while (*p)
98 if (*p == '_')
99 p++;
100 else
101 *(q)++ = tolower(*(p)++);
102 *q = '\0';
103 }
104 }
105
106 for (i = 0; i < FSCK_MSG_MAX; i++)
107 if (!strcmp(text, msg_id_info[i].downcased))
108 return i;
109
110 return -1;
111}
112
c99ba492
JS
113static int fsck_msg_type(enum fsck_msg_id msg_id,
114 struct fsck_options *options)
115{
116 int msg_type;
117
0282f4dc
JS
118 assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
119
120 if (options->msg_type)
121 msg_type = options->msg_type[msg_id];
122 else {
123 msg_type = msg_id_info[msg_id].msg_type;
124 if (options->strict && msg_type == FSCK_WARN)
125 msg_type = FSCK_ERROR;
126 }
c99ba492
JS
127
128 return msg_type;
129}
130
cd94c6f9
JS
131static void init_skiplist(struct fsck_options *options, const char *path)
132{
133 static struct sha1_array skiplist = SHA1_ARRAY_INIT;
134 int sorted, fd;
135 char buffer[41];
136 unsigned char sha1[20];
137
138 if (options->skiplist)
139 sorted = options->skiplist->sorted;
140 else {
141 sorted = 1;
142 options->skiplist = &skiplist;
143 }
144
145 fd = open(path, O_RDONLY);
146 if (fd < 0)
147 die("Could not open skip list: %s", path);
148 for (;;) {
149 int result = read_in_full(fd, buffer, sizeof(buffer));
150 if (result < 0)
151 die_errno("Could not read '%s'", path);
152 if (!result)
153 break;
154 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n')
155 die("Invalid SHA-1: %s", buffer);
156 sha1_array_append(&skiplist, sha1);
157 if (sorted && skiplist.nr > 1 &&
158 hashcmp(skiplist.sha1[skiplist.nr - 2],
159 sha1) > 0)
160 sorted = 0;
161 }
162 close(fd);
163
164 if (sorted)
165 skiplist.sorted = 1;
166}
167
0282f4dc
JS
168static int parse_msg_type(const char *str)
169{
170 if (!strcmp(str, "error"))
171 return FSCK_ERROR;
172 else if (!strcmp(str, "warn"))
173 return FSCK_WARN;
efaba7cc
JS
174 else if (!strcmp(str, "ignore"))
175 return FSCK_IGNORE;
0282f4dc
JS
176 else
177 die("Unknown fsck message type: '%s'", str);
178}
179
5d477a33
JS
180int is_valid_msg_type(const char *msg_id, const char *msg_type)
181{
182 if (parse_msg_id(msg_id) < 0)
183 return 0;
184 parse_msg_type(msg_type);
185 return 1;
186}
187
0282f4dc
JS
188void fsck_set_msg_type(struct fsck_options *options,
189 const char *msg_id, const char *msg_type)
190{
191 int id = parse_msg_id(msg_id), type;
192
193 if (id < 0)
194 die("Unhandled message id: %s", msg_id);
195 type = parse_msg_type(msg_type);
196
f50c4407
JS
197 if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
198 die("Cannot demote %s to %s", msg_id, msg_type);
199
0282f4dc
JS
200 if (!options->msg_type) {
201 int i;
202 int *msg_type = xmalloc(sizeof(int) * FSCK_MSG_MAX);
203 for (i = 0; i < FSCK_MSG_MAX; i++)
204 msg_type[i] = fsck_msg_type(i, options);
205 options->msg_type = msg_type;
206 }
207
208 options->msg_type[id] = type;
209}
210
211void fsck_set_msg_types(struct fsck_options *options, const char *values)
212{
213 char *buf = xstrdup(values), *to_free = buf;
214 int done = 0;
215
216 while (!done) {
217 int len = strcspn(buf, " ,|"), equal;
218
219 done = !buf[len];
220 if (!len) {
221 buf++;
222 continue;
223 }
224 buf[len] = '\0';
225
226 for (equal = 0;
227 equal < len && buf[equal] != '=' && buf[equal] != ':';
228 equal++)
229 buf[equal] = tolower(buf[equal]);
230 buf[equal] = '\0';
231
cd94c6f9
JS
232 if (!strcmp(buf, "skiplist")) {
233 if (equal == len)
234 die("skiplist requires a path");
235 init_skiplist(options, buf + equal + 1);
236 buf += len + 1;
237 continue;
238 }
239
0282f4dc
JS
240 if (equal == len)
241 die("Missing '=': '%s'", buf);
355885d5 242
0282f4dc
JS
243 fsck_set_msg_type(options, buf, buf + equal + 1);
244 buf += len + 1;
245 }
246 free(to_free);
247}
248
71ab8fa8
JS
249static void append_msg_id(struct strbuf *sb, const char *msg_id)
250{
251 for (;;) {
252 char c = *(msg_id)++;
253
254 if (!c)
255 break;
256 if (c != '_')
257 strbuf_addch(sb, tolower(c));
258 else {
259 assert(*msg_id);
260 strbuf_addch(sb, *(msg_id)++);
261 }
262 }
263
264 strbuf_addstr(sb, ": ");
265}
266
c99ba492
JS
267__attribute__((format (printf, 4, 5)))
268static int report(struct fsck_options *options, struct object *object,
269 enum fsck_msg_id id, const char *fmt, ...)
270{
271 va_list ap;
272 struct strbuf sb = STRBUF_INIT;
273 int msg_type = fsck_msg_type(id, options), result;
274
efaba7cc
JS
275 if (msg_type == FSCK_IGNORE)
276 return 0;
277
cd94c6f9 278 if (options->skiplist && object &&
7999b2cf 279 sha1_array_lookup(options->skiplist, get_object_hash(*object)) >= 0)
cd94c6f9
JS
280 return 0;
281
f50c4407
JS
282 if (msg_type == FSCK_FATAL)
283 msg_type = FSCK_ERROR;
f27d05b1
JS
284 else if (msg_type == FSCK_INFO)
285 msg_type = FSCK_WARN;
f50c4407 286
71ab8fa8
JS
287 append_msg_id(&sb, msg_id_info[id].id_string);
288
c99ba492
JS
289 va_start(ap, fmt);
290 strbuf_vaddf(&sb, fmt, ap);
291 result = options->error_func(object, msg_type, sb.buf);
292 strbuf_release(&sb);
293 va_end(ap);
294
295 return result;
296}
297
22410549 298static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
355885d5
MK
299{
300 struct tree_desc desc;
301 struct name_entry entry;
302 int res = 0;
303
304 if (parse_tree(tree))
305 return -1;
306
307 init_tree_desc(&desc, tree->buffer, tree->size);
308 while (tree_entry(&desc, &entry)) {
309 int result;
310
311 if (S_ISGITLINK(entry.mode))
312 continue;
313 if (S_ISDIR(entry.mode))
22410549 314 result = options->walk(&lookup_tree(entry.sha1)->object, OBJ_TREE, data, options);
355885d5 315 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode))
22410549 316 result = options->walk(&lookup_blob(entry.sha1)->object, OBJ_BLOB, data, options);
355885d5 317 else {
82247e9b 318 result = error("in tree %s: entry %s has bad mode %.6o",
f2fd0760 319 oid_to_hex(&tree->object.oid), entry.path, entry.mode);
355885d5
MK
320 }
321 if (result < 0)
322 return result;
323 if (!res)
324 res = result;
325 }
326 return res;
327}
328
22410549 329static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
355885d5
MK
330{
331 struct commit_list *parents;
332 int res;
333 int result;
334
335 if (parse_commit(commit))
336 return -1;
337
22410549 338 result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options);
355885d5
MK
339 if (result < 0)
340 return result;
341 res = result;
342
343 parents = commit->parents;
344 while (parents) {
22410549 345 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
355885d5
MK
346 if (result < 0)
347 return result;
348 if (!res)
349 res = result;
350 parents = parents->next;
351 }
352 return res;
353}
354
22410549 355static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
355885d5
MK
356{
357 if (parse_tag(tag))
358 return -1;
22410549 359 return options->walk(tag->tagged, OBJ_ANY, data, options);
355885d5
MK
360}
361
22410549 362int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
355885d5
MK
363{
364 if (!obj)
365 return -1;
366 switch (obj->type) {
367 case OBJ_BLOB:
368 return 0;
369 case OBJ_TREE:
22410549 370 return fsck_walk_tree((struct tree *)obj, data, options);
355885d5 371 case OBJ_COMMIT:
22410549 372 return fsck_walk_commit((struct commit *)obj, data, options);
355885d5 373 case OBJ_TAG:
22410549 374 return fsck_walk_tag((struct tag *)obj, data, options);
355885d5 375 default:
f2fd0760 376 error("Unknown object type for %s", oid_to_hex(&obj->oid));
355885d5
MK
377 return -1;
378 }
379}
ba002f3b
MK
380
381/*
382 * The entries in a tree are ordered in the _path_ order,
383 * which means that a directory entry is ordered by adding
384 * a slash to the end of it.
385 *
386 * So a directory called "a" is ordered _after_ a file
387 * called "a.c", because "a/" sorts after "a.c".
388 */
389#define TREE_UNORDERED (-1)
390#define TREE_HAS_DUPS (-2)
391
392static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
393{
394 int len1 = strlen(name1);
395 int len2 = strlen(name2);
396 int len = len1 < len2 ? len1 : len2;
397 unsigned char c1, c2;
398 int cmp;
399
400 cmp = memcmp(name1, name2, len);
401 if (cmp < 0)
402 return 0;
403 if (cmp > 0)
404 return TREE_UNORDERED;
405
406 /*
407 * Ok, the first <len> characters are the same.
408 * Now we need to order the next one, but turn
409 * a '\0' into a '/' for a directory entry.
410 */
411 c1 = name1[len];
412 c2 = name2[len];
413 if (!c1 && !c2)
414 /*
415 * git-write-tree used to write out a nonsense tree that has
416 * entries with the same name, one blob and one tree. Make
417 * sure we do not have duplicate entries.
418 */
419 return TREE_HAS_DUPS;
420 if (!c1 && S_ISDIR(mode1))
421 c1 = '/';
422 if (!c2 && S_ISDIR(mode2))
423 c2 = '/';
424 return c1 < c2 ? 0 : TREE_UNORDERED;
425}
426
22410549 427static int fsck_tree(struct tree *item, struct fsck_options *options)
ba002f3b
MK
428{
429 int retval;
c479d14a 430 int has_null_sha1 = 0;
ba002f3b
MK
431 int has_full_path = 0;
432 int has_empty_name = 0;
5d34a435
JK
433 int has_dot = 0;
434 int has_dotdot = 0;
5c17f512 435 int has_dotgit = 0;
ba002f3b
MK
436 int has_zero_pad = 0;
437 int has_bad_modes = 0;
438 int has_dup_entries = 0;
439 int not_properly_sorted = 0;
440 struct tree_desc desc;
441 unsigned o_mode;
442 const char *o_name;
ba002f3b
MK
443
444 init_tree_desc(&desc, item->buffer, item->size);
445
446 o_mode = 0;
447 o_name = NULL;
ba002f3b
MK
448
449 while (desc.size) {
450 unsigned mode;
451 const char *name;
c479d14a 452 const unsigned char *sha1;
ba002f3b 453
c479d14a 454 sha1 = tree_entry_extract(&desc, &name, &mode);
ba002f3b 455
effd12ec
HS
456 has_null_sha1 |= is_null_sha1(sha1);
457 has_full_path |= !!strchr(name, '/');
458 has_empty_name |= !*name;
459 has_dot |= !strcmp(name, ".");
460 has_dotdot |= !strcmp(name, "..");
5e519fb8
JH
461 has_dotgit |= (!strcmp(name, ".git") ||
462 is_hfs_dotgit(name) ||
463 is_ntfs_dotgit(name));
ba002f3b
MK
464 has_zero_pad |= *(char *)desc.buffer == '0';
465 update_tree_entry(&desc);
466
467 switch (mode) {
468 /*
469 * Standard modes..
470 */
471 case S_IFREG | 0755:
472 case S_IFREG | 0644:
473 case S_IFLNK:
474 case S_IFDIR:
475 case S_IFGITLINK:
476 break;
477 /*
478 * This is nonstandard, but we had a few of these
479 * early on when we honored the full set of mode
480 * bits..
481 */
482 case S_IFREG | 0664:
22410549 483 if (!options->strict)
ba002f3b
MK
484 break;
485 default:
486 has_bad_modes = 1;
487 }
488
489 if (o_name) {
490 switch (verify_ordered(o_mode, o_name, mode, name)) {
491 case TREE_UNORDERED:
492 not_properly_sorted = 1;
493 break;
494 case TREE_HAS_DUPS:
495 has_dup_entries = 1;
496 break;
497 default:
498 break;
499 }
500 }
501
502 o_mode = mode;
503 o_name = name;
ba002f3b
MK
504 }
505
506 retval = 0;
c479d14a 507 if (has_null_sha1)
c99ba492 508 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
ba002f3b 509 if (has_full_path)
c99ba492 510 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
ba002f3b 511 if (has_empty_name)
c99ba492 512 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
5d34a435 513 if (has_dot)
c99ba492 514 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
5d34a435 515 if (has_dotdot)
c99ba492 516 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
5c17f512 517 if (has_dotgit)
c99ba492 518 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
ba002f3b 519 if (has_zero_pad)
c99ba492 520 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
ba002f3b 521 if (has_bad_modes)
c99ba492 522 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
ba002f3b 523 if (has_dup_entries)
c99ba492 524 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
ba002f3b 525 if (not_properly_sorted)
c99ba492 526 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
ba002f3b
MK
527 return retval;
528}
529
84d18c0b 530static int verify_headers(const void *data, unsigned long size,
b2f44feb 531 struct object *obj, struct fsck_options *options)
4d0d8975
JS
532{
533 const char *buffer = (const char *)data;
534 unsigned long i;
535
536 for (i = 0; i < size; i++) {
537 switch (buffer[i]) {
538 case '\0':
c99ba492
JS
539 return report(options, obj,
540 FSCK_MSG_NUL_IN_HEADER,
541 "unterminated header: NUL at offset %ld", i);
4d0d8975
JS
542 case '\n':
543 if (i + 1 < size && buffer[i + 1] == '\n')
544 return 0;
545 }
546 }
547
84d18c0b
JH
548 /*
549 * We did not find double-LF that separates the header
550 * and the body. Not having a body is not a crime but
551 * we do want to see the terminating LF for the last header
552 * line.
553 */
554 if (size && buffer[size - 1] == '\n')
555 return 0;
556
c99ba492
JS
557 return report(options, obj,
558 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
4d0d8975
JS
559}
560
22410549 561static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
daae1922 562{
e6826e33 563 const char *p = *ident;
d4b8de04
JK
564 char *end;
565
e6826e33
JS
566 *ident = strchrnul(*ident, '\n');
567 if (**ident == '\n')
568 (*ident)++;
569
570 if (*p == '<')
c99ba492 571 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
572 p += strcspn(p, "<>\n");
573 if (*p == '>')
c99ba492 574 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
e6826e33 575 if (*p != '<')
c99ba492 576 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
e6826e33 577 if (p[-1] != ' ')
c99ba492 578 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
579 p++;
580 p += strcspn(p, "<>\n");
581 if (*p != '>')
c99ba492 582 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
e6826e33
JS
583 p++;
584 if (*p != ' ')
c99ba492 585 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
e6826e33
JS
586 p++;
587 if (*p == '0' && p[1] != ' ')
c99ba492 588 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
e6826e33 589 if (date_overflows(strtoul(p, &end, 10)))
c99ba492 590 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
e6826e33 591 if ((end == p || *end != ' '))
c99ba492 592 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
e6826e33
JS
593 p = end + 1;
594 if ((*p != '+' && *p != '-') ||
595 !isdigit(p[1]) ||
596 !isdigit(p[2]) ||
597 !isdigit(p[3]) ||
598 !isdigit(p[4]) ||
599 (p[5] != '\n'))
c99ba492 600 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
e6826e33 601 p += 6;
daae1922
JN
602 return 0;
603}
604
bc6b8fc1 605static int fsck_commit_buffer(struct commit *commit, const char *buffer,
22410549 606 unsigned long size, struct fsck_options *options)
ba002f3b 607{
ba002f3b
MK
608 unsigned char tree_sha1[20], sha1[20];
609 struct commit_graft *graft;
c9ad147f 610 unsigned parent_count, parent_line_count = 0, author_count;
daae1922 611 int err;
ba002f3b 612
b2f44feb 613 if (verify_headers(buffer, size, &commit->object, options))
4d0d8975
JS
614 return -1;
615
cf4fff57 616 if (!skip_prefix(buffer, "tree ", &buffer))
c99ba492 617 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
b3584761
JS
618 if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
619 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
620 if (err)
621 return err;
622 }
2d820a61 623 buffer += 41;
cf4fff57 624 while (skip_prefix(buffer, "parent ", &buffer)) {
b3584761
JS
625 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
626 err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
627 if (err)
628 return err;
629 }
2d820a61 630 buffer += 41;
9d02150c 631 parent_line_count++;
ba002f3b 632 }
7999b2cf 633 graft = lookup_commit_graft(get_object_hash(commit->object));
9d02150c 634 parent_count = commit_list_count(commit->parents);
ba002f3b 635 if (graft) {
9d02150c 636 if (graft->nr_parent == -1 && !parent_count)
ba002f3b 637 ; /* shallow commit */
b3584761
JS
638 else if (graft->nr_parent != parent_count) {
639 err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
640 if (err)
641 return err;
642 }
ba002f3b 643 } else {
b3584761
JS
644 if (parent_count != parent_line_count) {
645 err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
646 if (err)
647 return err;
648 }
ba002f3b 649 }
c9ad147f
JS
650 author_count = 0;
651 while (skip_prefix(buffer, "author ", &buffer)) {
652 author_count++;
653 err = fsck_ident(&buffer, &commit->object, options);
654 if (err)
655 return err;
ba002f3b 656 }
c9ad147f
JS
657 if (author_count < 1)
658 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
659 else if (author_count > 1)
660 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
daae1922
JN
661 if (err)
662 return err;
cf4fff57 663 if (!skip_prefix(buffer, "committer ", &buffer))
c99ba492 664 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
22410549 665 err = fsck_ident(&buffer, &commit->object, options);
daae1922
JN
666 if (err)
667 return err;
ba002f3b 668 if (!commit->tree)
c99ba492 669 return report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
ba002f3b
MK
670
671 return 0;
672}
673
90a398bb 674static int fsck_commit(struct commit *commit, const char *data,
22410549 675 unsigned long size, struct fsck_options *options)
bc6b8fc1 676{
90a398bb 677 const char *buffer = data ? data : get_commit_buffer(commit, &size);
22410549 678 int ret = fsck_commit_buffer(commit, buffer, size, options);
90a398bb
JS
679 if (!data)
680 unuse_commit_buffer(commit, buffer);
bc6b8fc1
JK
681 return ret;
682}
683
cec097be 684static int fsck_tag_buffer(struct tag *tag, const char *data,
22410549 685 unsigned long size, struct fsck_options *options)
cec097be
JS
686{
687 unsigned char sha1[20];
688 int ret = 0;
689 const char *buffer;
690 char *to_free = NULL, *eol;
691 struct strbuf sb = STRBUF_INIT;
692
693 if (data)
694 buffer = data;
695 else {
696 enum object_type type;
697
698 buffer = to_free =
7999b2cf 699 read_sha1_file(get_object_hash(tag->object), &type, &size);
cec097be 700 if (!buffer)
c99ba492
JS
701 return report(options, &tag->object,
702 FSCK_MSG_MISSING_TAG_OBJECT,
cec097be
JS
703 "cannot read tag object");
704
705 if (type != OBJ_TAG) {
c99ba492
JS
706 ret = report(options, &tag->object,
707 FSCK_MSG_TAG_OBJECT_NOT_TAG,
cec097be
JS
708 "expected tag got %s",
709 typename(type));
710 goto done;
711 }
712 }
713
b2f44feb 714 if (verify_headers(buffer, size, &tag->object, options))
cec097be
JS
715 goto done;
716
717 if (!skip_prefix(buffer, "object ", &buffer)) {
c99ba492 718 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
cec097be
JS
719 goto done;
720 }
721 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
c99ba492 722 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
7d7d5b05
JS
723 if (ret)
724 goto done;
cec097be
JS
725 }
726 buffer += 41;
727
728 if (!skip_prefix(buffer, "type ", &buffer)) {
c99ba492 729 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
cec097be
JS
730 goto done;
731 }
732 eol = strchr(buffer, '\n');
733 if (!eol) {
c99ba492 734 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
cec097be
JS
735 goto done;
736 }
737 if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
c99ba492 738 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
cec097be
JS
739 if (ret)
740 goto done;
741 buffer = eol + 1;
742
743 if (!skip_prefix(buffer, "tag ", &buffer)) {
c99ba492 744 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
cec097be
JS
745 goto done;
746 }
747 eol = strchr(buffer, '\n');
748 if (!eol) {
c99ba492 749 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
cec097be
JS
750 goto done;
751 }
752 strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
f27d05b1
JS
753 if (check_refname_format(sb.buf, 0)) {
754 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
c99ba492 755 "invalid 'tag' name: %.*s",
7add4419 756 (int)(eol - buffer), buffer);
f27d05b1
JS
757 if (ret)
758 goto done;
759 }
cec097be
JS
760 buffer = eol + 1;
761
f27d05b1 762 if (!skip_prefix(buffer, "tagger ", &buffer)) {
cec097be 763 /* early tags do not contain 'tagger' lines; warn only */
f27d05b1
JS
764 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
765 if (ret)
766 goto done;
767 }
cec097be 768 else
22410549 769 ret = fsck_ident(&buffer, &tag->object, options);
cec097be
JS
770
771done:
772 strbuf_release(&sb);
773 free(to_free);
774 return ret;
775}
776
90a398bb 777static int fsck_tag(struct tag *tag, const char *data,
22410549 778 unsigned long size, struct fsck_options *options)
ba002f3b
MK
779{
780 struct object *tagged = tag->tagged;
781
782 if (!tagged)
c99ba492 783 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
cec097be 784
22410549 785 return fsck_tag_buffer(tag, data, size, options);
ba002f3b
MK
786}
787
90a398bb 788int fsck_object(struct object *obj, void *data, unsigned long size,
22410549 789 struct fsck_options *options)
ba002f3b
MK
790{
791 if (!obj)
c99ba492 792 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
ba002f3b
MK
793
794 if (obj->type == OBJ_BLOB)
795 return 0;
796 if (obj->type == OBJ_TREE)
22410549 797 return fsck_tree((struct tree *) obj, options);
ba002f3b 798 if (obj->type == OBJ_COMMIT)
90a398bb 799 return fsck_commit((struct commit *) obj, (const char *) data,
22410549 800 size, options);
ba002f3b 801 if (obj->type == OBJ_TAG)
90a398bb 802 return fsck_tag((struct tag *) obj, (const char *) data,
22410549 803 size, options);
ba002f3b 804
c99ba492 805 return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
ba002f3b
MK
806 obj->type);
807}
d6ffc8d7 808
c99ba492 809int fsck_error_function(struct object *obj, int msg_type, const char *message)
d6ffc8d7 810{
0282f4dc 811 if (msg_type == FSCK_WARN) {
f2fd0760 812 warning("object %s: %s", oid_to_hex(&obj->oid), message);
0282f4dc
JS
813 return 0;
814 }
f2fd0760 815 error("object %s: %s", oid_to_hex(&obj->oid), message);
d6ffc8d7
MK
816 return 1;
817}