tree-walk: be more specific about corrupt tree errors
[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"
355885d5 13
f50c4407 14#define FSCK_FATAL -1
f27d05b1 15#define FSCK_INFO -2
f50c4407 16
c99ba492 17#define FOREACH_MSG_ID(FUNC) \
f50c4407
JS
18 /* fatal errors */ \
19 FUNC(NUL_IN_HEADER, FATAL) \
20 FUNC(UNTERMINATED_HEADER, FATAL) \
c99ba492
JS
21 /* errors */ \
22 FUNC(BAD_DATE, ERROR) \
23 FUNC(BAD_DATE_OVERFLOW, ERROR) \
24 FUNC(BAD_EMAIL, ERROR) \
25 FUNC(BAD_NAME, ERROR) \
26 FUNC(BAD_OBJECT_SHA1, ERROR) \
27 FUNC(BAD_PARENT_SHA1, ERROR) \
28 FUNC(BAD_TAG_OBJECT, ERROR) \
29 FUNC(BAD_TIMEZONE, ERROR) \
30 FUNC(BAD_TREE, ERROR) \
31 FUNC(BAD_TREE_SHA1, ERROR) \
32 FUNC(BAD_TYPE, ERROR) \
33 FUNC(DUPLICATE_ENTRIES, ERROR) \
34 FUNC(MISSING_AUTHOR, ERROR) \
35 FUNC(MISSING_COMMITTER, ERROR) \
36 FUNC(MISSING_EMAIL, ERROR) \
37 FUNC(MISSING_GRAFT, ERROR) \
38 FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
39 FUNC(MISSING_OBJECT, ERROR) \
40 FUNC(MISSING_PARENT, ERROR) \
41 FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
42 FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
43 FUNC(MISSING_TAG, ERROR) \
44 FUNC(MISSING_TAG_ENTRY, ERROR) \
45 FUNC(MISSING_TAG_OBJECT, ERROR) \
46 FUNC(MISSING_TREE, ERROR) \
47 FUNC(MISSING_TYPE, ERROR) \
48 FUNC(MISSING_TYPE_ENTRY, ERROR) \
c9ad147f 49 FUNC(MULTIPLE_AUTHORS, ERROR) \
c99ba492
JS
50 FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
51 FUNC(TREE_NOT_SORTED, ERROR) \
52 FUNC(UNKNOWN_TYPE, ERROR) \
c99ba492
JS
53 FUNC(ZERO_PADDED_DATE, ERROR) \
54 /* warnings */ \
55 FUNC(BAD_FILEMODE, WARN) \
c99ba492
JS
56 FUNC(EMPTY_NAME, WARN) \
57 FUNC(FULL_PATHNAME, WARN) \
58 FUNC(HAS_DOT, WARN) \
59 FUNC(HAS_DOTDOT, WARN) \
60 FUNC(HAS_DOTGIT, WARN) \
c99ba492 61 FUNC(NULL_SHA1, WARN) \
f27d05b1 62 FUNC(ZERO_PADDED_FILEMODE, WARN) \
6d2d780f 63 FUNC(NUL_IN_COMMIT, WARN) \
f27d05b1
JS
64 /* infos (reported as warnings, but ignored by default) */ \
65 FUNC(BAD_TAG_NAME, INFO) \
66 FUNC(MISSING_TAGGER_ENTRY, INFO)
c99ba492
JS
67
68#define MSG_ID(id, msg_type) FSCK_MSG_##id,
69enum fsck_msg_id {
70 FOREACH_MSG_ID(MSG_ID)
71 FSCK_MSG_MAX
72};
73#undef MSG_ID
74
f417eed8
JS
75#define STR(x) #x
76#define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
c99ba492 77static struct {
f417eed8
JS
78 const char *id_string;
79 const char *downcased;
c99ba492
JS
80 int msg_type;
81} msg_id_info[FSCK_MSG_MAX + 1] = {
82 FOREACH_MSG_ID(MSG_ID)
f417eed8 83 { NULL, NULL, -1 }
c99ba492
JS
84};
85#undef MSG_ID
86
f417eed8
JS
87static int parse_msg_id(const char *text)
88{
89 int i;
90
91 if (!msg_id_info[0].downcased) {
92 /* convert id_string to lower case, without underscores. */
93 for (i = 0; i < FSCK_MSG_MAX; i++) {
94 const char *p = msg_id_info[i].id_string;
95 int len = strlen(p);
96 char *q = xmalloc(len);
97
98 msg_id_info[i].downcased = q;
99 while (*p)
100 if (*p == '_')
101 p++;
102 else
103 *(q)++ = tolower(*(p)++);
104 *q = '\0';
105 }
106 }
107
108 for (i = 0; i < FSCK_MSG_MAX; i++)
109 if (!strcmp(text, msg_id_info[i].downcased))
110 return i;
111
112 return -1;
113}
114
c99ba492
JS
115static int fsck_msg_type(enum fsck_msg_id msg_id,
116 struct fsck_options *options)
117{
118 int msg_type;
119
0282f4dc
JS
120 assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
121
122 if (options->msg_type)
123 msg_type = options->msg_type[msg_id];
124 else {
125 msg_type = msg_id_info[msg_id].msg_type;
126 if (options->strict && msg_type == FSCK_WARN)
127 msg_type = FSCK_ERROR;
128 }
c99ba492
JS
129
130 return msg_type;
131}
132
cd94c6f9
JS
133static void init_skiplist(struct fsck_options *options, const char *path)
134{
135 static struct sha1_array skiplist = SHA1_ARRAY_INIT;
136 int sorted, fd;
137 char buffer[41];
138 unsigned char sha1[20];
139
140 if (options->skiplist)
141 sorted = options->skiplist->sorted;
142 else {
143 sorted = 1;
144 options->skiplist = &skiplist;
145 }
146
147 fd = open(path, O_RDONLY);
148 if (fd < 0)
149 die("Could not open skip list: %s", path);
150 for (;;) {
151 int result = read_in_full(fd, buffer, sizeof(buffer));
152 if (result < 0)
153 die_errno("Could not read '%s'", path);
154 if (!result)
155 break;
156 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n')
157 die("Invalid SHA-1: %s", buffer);
158 sha1_array_append(&skiplist, sha1);
159 if (sorted && skiplist.nr > 1 &&
160 hashcmp(skiplist.sha1[skiplist.nr - 2],
161 sha1) > 0)
162 sorted = 0;
163 }
164 close(fd);
165
166 if (sorted)
167 skiplist.sorted = 1;
168}
169
0282f4dc
JS
170static int parse_msg_type(const char *str)
171{
172 if (!strcmp(str, "error"))
173 return FSCK_ERROR;
174 else if (!strcmp(str, "warn"))
175 return FSCK_WARN;
efaba7cc
JS
176 else if (!strcmp(str, "ignore"))
177 return FSCK_IGNORE;
0282f4dc
JS
178 else
179 die("Unknown fsck message type: '%s'", str);
180}
181
5d477a33
JS
182int is_valid_msg_type(const char *msg_id, const char *msg_type)
183{
184 if (parse_msg_id(msg_id) < 0)
185 return 0;
186 parse_msg_type(msg_type);
187 return 1;
188}
189
0282f4dc
JS
190void fsck_set_msg_type(struct fsck_options *options,
191 const char *msg_id, const char *msg_type)
192{
193 int id = parse_msg_id(msg_id), type;
194
195 if (id < 0)
196 die("Unhandled message id: %s", msg_id);
197 type = parse_msg_type(msg_type);
198
f50c4407
JS
199 if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
200 die("Cannot demote %s to %s", msg_id, msg_type);
201
0282f4dc
JS
202 if (!options->msg_type) {
203 int i;
b32fa95f
JK
204 int *msg_type;
205 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
0282f4dc
JS
206 for (i = 0; i < FSCK_MSG_MAX; i++)
207 msg_type[i] = fsck_msg_type(i, options);
208 options->msg_type = msg_type;
209 }
210
211 options->msg_type[id] = type;
212}
213
214void fsck_set_msg_types(struct fsck_options *options, const char *values)
215{
216 char *buf = xstrdup(values), *to_free = buf;
217 int done = 0;
218
219 while (!done) {
220 int len = strcspn(buf, " ,|"), equal;
221
222 done = !buf[len];
223 if (!len) {
224 buf++;
225 continue;
226 }
227 buf[len] = '\0';
228
229 for (equal = 0;
230 equal < len && buf[equal] != '=' && buf[equal] != ':';
231 equal++)
232 buf[equal] = tolower(buf[equal]);
233 buf[equal] = '\0';
234
cd94c6f9
JS
235 if (!strcmp(buf, "skiplist")) {
236 if (equal == len)
237 die("skiplist requires a path");
238 init_skiplist(options, buf + equal + 1);
239 buf += len + 1;
240 continue;
241 }
242
0282f4dc
JS
243 if (equal == len)
244 die("Missing '=': '%s'", buf);
355885d5 245
0282f4dc
JS
246 fsck_set_msg_type(options, buf, buf + equal + 1);
247 buf += len + 1;
248 }
249 free(to_free);
250}
251
71ab8fa8
JS
252static void append_msg_id(struct strbuf *sb, const char *msg_id)
253{
254 for (;;) {
255 char c = *(msg_id)++;
256
257 if (!c)
258 break;
259 if (c != '_')
260 strbuf_addch(sb, tolower(c));
261 else {
262 assert(*msg_id);
263 strbuf_addch(sb, *(msg_id)++);
264 }
265 }
266
267 strbuf_addstr(sb, ": ");
268}
269
c99ba492
JS
270__attribute__((format (printf, 4, 5)))
271static int report(struct fsck_options *options, struct object *object,
272 enum fsck_msg_id id, const char *fmt, ...)
273{
274 va_list ap;
275 struct strbuf sb = STRBUF_INIT;
276 int msg_type = fsck_msg_type(id, options), result;
277
efaba7cc
JS
278 if (msg_type == FSCK_IGNORE)
279 return 0;
280
cd94c6f9 281 if (options->skiplist && object &&
ed1c9977 282 sha1_array_lookup(options->skiplist, object->oid.hash) >= 0)
cd94c6f9
JS
283 return 0;
284
f50c4407
JS
285 if (msg_type == FSCK_FATAL)
286 msg_type = FSCK_ERROR;
f27d05b1
JS
287 else if (msg_type == FSCK_INFO)
288 msg_type = FSCK_WARN;
f50c4407 289
71ab8fa8
JS
290 append_msg_id(&sb, msg_id_info[id].id_string);
291
c99ba492
JS
292 va_start(ap, fmt);
293 strbuf_vaddf(&sb, fmt, ap);
1cd772cc 294 result = options->error_func(options, object, msg_type, sb.buf);
c99ba492
JS
295 strbuf_release(&sb);
296 va_end(ap);
297
298 return result;
299}
300
7b35efd7
JS
301static char *get_object_name(struct fsck_options *options, struct object *obj)
302{
303 if (!options->object_names)
304 return NULL;
305 return lookup_decoration(options->object_names, obj);
306}
307
308static void put_object_name(struct fsck_options *options, struct object *obj,
309 const char *fmt, ...)
310{
311 va_list ap;
312 struct strbuf buf = STRBUF_INIT;
313 char *existing;
314
315 if (!options->object_names)
316 return;
317 existing = lookup_decoration(options->object_names, obj);
318 if (existing)
319 return;
320 va_start(ap, fmt);
321 strbuf_vaddf(&buf, fmt, ap);
322 add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
323 va_end(ap);
324}
325
90cf590f
JS
326static const char *describe_object(struct fsck_options *o, struct object *obj)
327{
328 static struct strbuf buf = STRBUF_INIT;
329 char *name;
330
331 strbuf_reset(&buf);
332 strbuf_addstr(&buf, oid_to_hex(&obj->oid));
333 if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
334 strbuf_addf(&buf, " (%s)", name);
335
336 return buf.buf;
337}
338
22410549 339static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
355885d5
MK
340{
341 struct tree_desc desc;
342 struct name_entry entry;
343 int res = 0;
7b35efd7 344 const char *name;
355885d5
MK
345
346 if (parse_tree(tree))
347 return -1;
348
7b35efd7 349 name = get_object_name(options, &tree->object);
355885d5
MK
350 init_tree_desc(&desc, tree->buffer, tree->size);
351 while (tree_entry(&desc, &entry)) {
7b35efd7 352 struct object *obj;
355885d5
MK
353 int result;
354
355 if (S_ISGITLINK(entry.mode))
356 continue;
7b35efd7
JS
357
358 if (S_ISDIR(entry.mode)) {
359 obj = &lookup_tree(entry.oid->hash)->object;
360 if (name)
361 put_object_name(options, obj, "%s%s/", name,
362 entry.path);
363 result = options->walk(obj, OBJ_TREE, data, options);
364 }
365 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
366 obj = &lookup_blob(entry.oid->hash)->object;
367 if (name)
368 put_object_name(options, obj, "%s%s", name,
369 entry.path);
370 result = options->walk(obj, OBJ_BLOB, data, options);
371 }
355885d5 372 else {
82247e9b 373 result = error("in tree %s: entry %s has bad mode %.6o",
90cf590f 374 describe_object(options, &tree->object), entry.path, entry.mode);
355885d5
MK
375 }
376 if (result < 0)
377 return result;
378 if (!res)
379 res = result;
380 }
381 return res;
382}
383
22410549 384static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
355885d5 385{
7b35efd7 386 int counter = 0, generation = 0, name_prefix_len = 0;
355885d5
MK
387 struct commit_list *parents;
388 int res;
389 int result;
7b35efd7 390 const char *name;
355885d5
MK
391
392 if (parse_commit(commit))
393 return -1;
394
7b35efd7
JS
395 name = get_object_name(options, &commit->object);
396 if (name)
397 put_object_name(options, &commit->tree->object, "%s:", name);
398
22410549 399 result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options);
355885d5
MK
400 if (result < 0)
401 return result;
402 res = result;
403
404 parents = commit->parents;
7b35efd7
JS
405 if (name && parents) {
406 int len = strlen(name), power;
407
408 if (len && name[len - 1] == '^') {
409 generation = 1;
410 name_prefix_len = len - 1;
411 }
412 else { /* parse ~<generation> suffix */
413 for (generation = 0, power = 1;
414 len && isdigit(name[len - 1]);
415 power *= 10)
416 generation += power * (name[--len] - '0');
417 if (power > 1 && len && name[len - 1] == '~')
418 name_prefix_len = len - 1;
419 }
420 }
421
355885d5 422 while (parents) {
7b35efd7
JS
423 if (name) {
424 struct object *obj = &parents->item->object;
425
426 if (++counter > 1)
427 put_object_name(options, obj, "%s^%d",
428 name, counter);
429 else if (generation > 0)
430 put_object_name(options, obj, "%.*s~%d",
431 name_prefix_len, name, generation + 1);
432 else
433 put_object_name(options, obj, "%s^", name);
434 }
22410549 435 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
355885d5
MK
436 if (result < 0)
437 return result;
438 if (!res)
439 res = result;
440 parents = parents->next;
441 }
442 return res;
443}
444
22410549 445static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
355885d5 446{
7b35efd7
JS
447 char *name = get_object_name(options, &tag->object);
448
355885d5
MK
449 if (parse_tag(tag))
450 return -1;
7b35efd7
JS
451 if (name)
452 put_object_name(options, tag->tagged, "%s", name);
22410549 453 return options->walk(tag->tagged, OBJ_ANY, data, options);
355885d5
MK
454}
455
22410549 456int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
355885d5
MK
457{
458 if (!obj)
459 return -1;
460 switch (obj->type) {
461 case OBJ_BLOB:
462 return 0;
463 case OBJ_TREE:
22410549 464 return fsck_walk_tree((struct tree *)obj, data, options);
355885d5 465 case OBJ_COMMIT:
22410549 466 return fsck_walk_commit((struct commit *)obj, data, options);
355885d5 467 case OBJ_TAG:
22410549 468 return fsck_walk_tag((struct tag *)obj, data, options);
355885d5 469 default:
90cf590f 470 error("Unknown object type for %s", describe_object(options, obj));
355885d5
MK
471 return -1;
472 }
473}
ba002f3b
MK
474
475/*
476 * The entries in a tree are ordered in the _path_ order,
477 * which means that a directory entry is ordered by adding
478 * a slash to the end of it.
479 *
480 * So a directory called "a" is ordered _after_ a file
481 * called "a.c", because "a/" sorts after "a.c".
482 */
483#define TREE_UNORDERED (-1)
484#define TREE_HAS_DUPS (-2)
485
486static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
487{
488 int len1 = strlen(name1);
489 int len2 = strlen(name2);
490 int len = len1 < len2 ? len1 : len2;
491 unsigned char c1, c2;
492 int cmp;
493
494 cmp = memcmp(name1, name2, len);
495 if (cmp < 0)
496 return 0;
497 if (cmp > 0)
498 return TREE_UNORDERED;
499
500 /*
501 * Ok, the first <len> characters are the same.
502 * Now we need to order the next one, but turn
503 * a '\0' into a '/' for a directory entry.
504 */
505 c1 = name1[len];
506 c2 = name2[len];
507 if (!c1 && !c2)
508 /*
509 * git-write-tree used to write out a nonsense tree that has
510 * entries with the same name, one blob and one tree. Make
511 * sure we do not have duplicate entries.
512 */
513 return TREE_HAS_DUPS;
514 if (!c1 && S_ISDIR(mode1))
515 c1 = '/';
516 if (!c2 && S_ISDIR(mode2))
517 c2 = '/';
518 return c1 < c2 ? 0 : TREE_UNORDERED;
519}
520
22410549 521static int fsck_tree(struct tree *item, struct fsck_options *options)
ba002f3b
MK
522{
523 int retval;
c479d14a 524 int has_null_sha1 = 0;
ba002f3b
MK
525 int has_full_path = 0;
526 int has_empty_name = 0;
5d34a435
JK
527 int has_dot = 0;
528 int has_dotdot = 0;
5c17f512 529 int has_dotgit = 0;
ba002f3b
MK
530 int has_zero_pad = 0;
531 int has_bad_modes = 0;
532 int has_dup_entries = 0;
533 int not_properly_sorted = 0;
534 struct tree_desc desc;
535 unsigned o_mode;
536 const char *o_name;
ba002f3b
MK
537
538 init_tree_desc(&desc, item->buffer, item->size);
539
540 o_mode = 0;
541 o_name = NULL;
ba002f3b
MK
542
543 while (desc.size) {
544 unsigned mode;
545 const char *name;
ce6663a9 546 const struct object_id *oid;
ba002f3b 547
ce6663a9 548 oid = tree_entry_extract(&desc, &name, &mode);
ba002f3b 549
ce6663a9 550 has_null_sha1 |= is_null_oid(oid);
effd12ec
HS
551 has_full_path |= !!strchr(name, '/');
552 has_empty_name |= !*name;
553 has_dot |= !strcmp(name, ".");
554 has_dotdot |= !strcmp(name, "..");
5e519fb8
JH
555 has_dotgit |= (!strcmp(name, ".git") ||
556 is_hfs_dotgit(name) ||
557 is_ntfs_dotgit(name));
ba002f3b
MK
558 has_zero_pad |= *(char *)desc.buffer == '0';
559 update_tree_entry(&desc);
560
561 switch (mode) {
562 /*
563 * Standard modes..
564 */
565 case S_IFREG | 0755:
566 case S_IFREG | 0644:
567 case S_IFLNK:
568 case S_IFDIR:
569 case S_IFGITLINK:
570 break;
571 /*
572 * This is nonstandard, but we had a few of these
573 * early on when we honored the full set of mode
574 * bits..
575 */
576 case S_IFREG | 0664:
22410549 577 if (!options->strict)
ba002f3b
MK
578 break;
579 default:
580 has_bad_modes = 1;
581 }
582
583 if (o_name) {
584 switch (verify_ordered(o_mode, o_name, mode, name)) {
585 case TREE_UNORDERED:
586 not_properly_sorted = 1;
587 break;
588 case TREE_HAS_DUPS:
589 has_dup_entries = 1;
590 break;
591 default:
592 break;
593 }
594 }
595
596 o_mode = mode;
597 o_name = name;
ba002f3b
MK
598 }
599
600 retval = 0;
c479d14a 601 if (has_null_sha1)
c99ba492 602 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
ba002f3b 603 if (has_full_path)
c99ba492 604 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
ba002f3b 605 if (has_empty_name)
c99ba492 606 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
5d34a435 607 if (has_dot)
c99ba492 608 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
5d34a435 609 if (has_dotdot)
c99ba492 610 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
5c17f512 611 if (has_dotgit)
c99ba492 612 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
ba002f3b 613 if (has_zero_pad)
c99ba492 614 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
ba002f3b 615 if (has_bad_modes)
c99ba492 616 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
ba002f3b 617 if (has_dup_entries)
c99ba492 618 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
ba002f3b 619 if (not_properly_sorted)
c99ba492 620 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
ba002f3b
MK
621 return retval;
622}
623
84d18c0b 624static int verify_headers(const void *data, unsigned long size,
b2f44feb 625 struct object *obj, struct fsck_options *options)
4d0d8975
JS
626{
627 const char *buffer = (const char *)data;
628 unsigned long i;
629
630 for (i = 0; i < size; i++) {
631 switch (buffer[i]) {
632 case '\0':
c99ba492
JS
633 return report(options, obj,
634 FSCK_MSG_NUL_IN_HEADER,
635 "unterminated header: NUL at offset %ld", i);
4d0d8975
JS
636 case '\n':
637 if (i + 1 < size && buffer[i + 1] == '\n')
638 return 0;
639 }
640 }
641
84d18c0b
JH
642 /*
643 * We did not find double-LF that separates the header
644 * and the body. Not having a body is not a crime but
645 * we do want to see the terminating LF for the last header
646 * line.
647 */
648 if (size && buffer[size - 1] == '\n')
649 return 0;
650
c99ba492
JS
651 return report(options, obj,
652 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
4d0d8975
JS
653}
654
22410549 655static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
daae1922 656{
e6826e33 657 const char *p = *ident;
d4b8de04
JK
658 char *end;
659
e6826e33
JS
660 *ident = strchrnul(*ident, '\n');
661 if (**ident == '\n')
662 (*ident)++;
663
664 if (*p == '<')
c99ba492 665 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
666 p += strcspn(p, "<>\n");
667 if (*p == '>')
c99ba492 668 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
e6826e33 669 if (*p != '<')
c99ba492 670 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
e6826e33 671 if (p[-1] != ' ')
c99ba492 672 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
673 p++;
674 p += strcspn(p, "<>\n");
675 if (*p != '>')
c99ba492 676 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
e6826e33
JS
677 p++;
678 if (*p != ' ')
c99ba492 679 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
e6826e33
JS
680 p++;
681 if (*p == '0' && p[1] != ' ')
c99ba492 682 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
e6826e33 683 if (date_overflows(strtoul(p, &end, 10)))
c99ba492 684 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
e6826e33 685 if ((end == p || *end != ' '))
c99ba492 686 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
e6826e33
JS
687 p = end + 1;
688 if ((*p != '+' && *p != '-') ||
689 !isdigit(p[1]) ||
690 !isdigit(p[2]) ||
691 !isdigit(p[3]) ||
692 !isdigit(p[4]) ||
693 (p[5] != '\n'))
c99ba492 694 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
e6826e33 695 p += 6;
daae1922
JN
696 return 0;
697}
698
bc6b8fc1 699static int fsck_commit_buffer(struct commit *commit, const char *buffer,
22410549 700 unsigned long size, struct fsck_options *options)
ba002f3b 701{
ba002f3b
MK
702 unsigned char tree_sha1[20], sha1[20];
703 struct commit_graft *graft;
c9ad147f 704 unsigned parent_count, parent_line_count = 0, author_count;
daae1922 705 int err;
6d2d780f 706 const char *buffer_begin = buffer;
ba002f3b 707
b2f44feb 708 if (verify_headers(buffer, size, &commit->object, options))
4d0d8975
JS
709 return -1;
710
cf4fff57 711 if (!skip_prefix(buffer, "tree ", &buffer))
c99ba492 712 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
b3584761
JS
713 if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
714 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
715 if (err)
716 return err;
717 }
2d820a61 718 buffer += 41;
cf4fff57 719 while (skip_prefix(buffer, "parent ", &buffer)) {
b3584761
JS
720 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
721 err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
722 if (err)
723 return err;
724 }
2d820a61 725 buffer += 41;
9d02150c 726 parent_line_count++;
ba002f3b 727 }
ed1c9977 728 graft = lookup_commit_graft(commit->object.oid.hash);
9d02150c 729 parent_count = commit_list_count(commit->parents);
ba002f3b 730 if (graft) {
9d02150c 731 if (graft->nr_parent == -1 && !parent_count)
ba002f3b 732 ; /* shallow commit */
b3584761
JS
733 else if (graft->nr_parent != parent_count) {
734 err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
735 if (err)
736 return err;
737 }
ba002f3b 738 } else {
b3584761
JS
739 if (parent_count != parent_line_count) {
740 err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
741 if (err)
742 return err;
743 }
ba002f3b 744 }
c9ad147f
JS
745 author_count = 0;
746 while (skip_prefix(buffer, "author ", &buffer)) {
747 author_count++;
748 err = fsck_ident(&buffer, &commit->object, options);
749 if (err)
750 return err;
ba002f3b 751 }
c9ad147f
JS
752 if (author_count < 1)
753 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
754 else if (author_count > 1)
755 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
daae1922
JN
756 if (err)
757 return err;
cf4fff57 758 if (!skip_prefix(buffer, "committer ", &buffer))
c99ba492 759 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
22410549 760 err = fsck_ident(&buffer, &commit->object, options);
daae1922
JN
761 if (err)
762 return err;
5af29718
JH
763 if (!commit->tree) {
764 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
765 if (err)
766 return err;
767 }
6d2d780f
JH
768 if (memchr(buffer_begin, '\0', size)) {
769 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
770 "NUL byte in the commit object body");
771 if (err)
772 return err;
773 }
ba002f3b
MK
774 return 0;
775}
776
90a398bb 777static int fsck_commit(struct commit *commit, const char *data,
22410549 778 unsigned long size, struct fsck_options *options)
bc6b8fc1 779{
90a398bb 780 const char *buffer = data ? data : get_commit_buffer(commit, &size);
22410549 781 int ret = fsck_commit_buffer(commit, buffer, size, options);
90a398bb
JS
782 if (!data)
783 unuse_commit_buffer(commit, buffer);
bc6b8fc1
JK
784 return ret;
785}
786
cec097be 787static int fsck_tag_buffer(struct tag *tag, const char *data,
22410549 788 unsigned long size, struct fsck_options *options)
cec097be
JS
789{
790 unsigned char sha1[20];
791 int ret = 0;
792 const char *buffer;
793 char *to_free = NULL, *eol;
794 struct strbuf sb = STRBUF_INIT;
795
796 if (data)
797 buffer = data;
798 else {
799 enum object_type type;
800
801 buffer = to_free =
ed1c9977 802 read_sha1_file(tag->object.oid.hash, &type, &size);
cec097be 803 if (!buffer)
c99ba492
JS
804 return report(options, &tag->object,
805 FSCK_MSG_MISSING_TAG_OBJECT,
cec097be
JS
806 "cannot read tag object");
807
808 if (type != OBJ_TAG) {
c99ba492
JS
809 ret = report(options, &tag->object,
810 FSCK_MSG_TAG_OBJECT_NOT_TAG,
cec097be
JS
811 "expected tag got %s",
812 typename(type));
813 goto done;
814 }
815 }
816
8a272f29
RS
817 ret = verify_headers(buffer, size, &tag->object, options);
818 if (ret)
cec097be
JS
819 goto done;
820
821 if (!skip_prefix(buffer, "object ", &buffer)) {
c99ba492 822 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
cec097be
JS
823 goto done;
824 }
825 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
c99ba492 826 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
7d7d5b05
JS
827 if (ret)
828 goto done;
cec097be
JS
829 }
830 buffer += 41;
831
832 if (!skip_prefix(buffer, "type ", &buffer)) {
c99ba492 833 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
cec097be
JS
834 goto done;
835 }
836 eol = strchr(buffer, '\n');
837 if (!eol) {
c99ba492 838 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
cec097be
JS
839 goto done;
840 }
841 if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
c99ba492 842 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
cec097be
JS
843 if (ret)
844 goto done;
845 buffer = eol + 1;
846
847 if (!skip_prefix(buffer, "tag ", &buffer)) {
c99ba492 848 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
cec097be
JS
849 goto done;
850 }
851 eol = strchr(buffer, '\n');
852 if (!eol) {
c99ba492 853 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
cec097be
JS
854 goto done;
855 }
856 strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
f27d05b1
JS
857 if (check_refname_format(sb.buf, 0)) {
858 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
c99ba492 859 "invalid 'tag' name: %.*s",
7add4419 860 (int)(eol - buffer), buffer);
f27d05b1
JS
861 if (ret)
862 goto done;
863 }
cec097be
JS
864 buffer = eol + 1;
865
f27d05b1 866 if (!skip_prefix(buffer, "tagger ", &buffer)) {
cec097be 867 /* early tags do not contain 'tagger' lines; warn only */
f27d05b1
JS
868 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
869 if (ret)
870 goto done;
871 }
cec097be 872 else
22410549 873 ret = fsck_ident(&buffer, &tag->object, options);
cec097be
JS
874
875done:
876 strbuf_release(&sb);
877 free(to_free);
878 return ret;
879}
880
90a398bb 881static int fsck_tag(struct tag *tag, const char *data,
22410549 882 unsigned long size, struct fsck_options *options)
ba002f3b
MK
883{
884 struct object *tagged = tag->tagged;
885
886 if (!tagged)
c99ba492 887 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
cec097be 888
22410549 889 return fsck_tag_buffer(tag, data, size, options);
ba002f3b
MK
890}
891
90a398bb 892int fsck_object(struct object *obj, void *data, unsigned long size,
22410549 893 struct fsck_options *options)
ba002f3b
MK
894{
895 if (!obj)
c99ba492 896 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
ba002f3b
MK
897
898 if (obj->type == OBJ_BLOB)
899 return 0;
900 if (obj->type == OBJ_TREE)
22410549 901 return fsck_tree((struct tree *) obj, options);
ba002f3b 902 if (obj->type == OBJ_COMMIT)
90a398bb 903 return fsck_commit((struct commit *) obj, (const char *) data,
22410549 904 size, options);
ba002f3b 905 if (obj->type == OBJ_TAG)
90a398bb 906 return fsck_tag((struct tag *) obj, (const char *) data,
22410549 907 size, options);
ba002f3b 908
c99ba492 909 return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
ba002f3b
MK
910 obj->type);
911}
d6ffc8d7 912
1cd772cc
JS
913int fsck_error_function(struct fsck_options *o,
914 struct object *obj, int msg_type, const char *message)
d6ffc8d7 915{
0282f4dc 916 if (msg_type == FSCK_WARN) {
90cf590f 917 warning("object %s: %s", describe_object(o, obj), message);
0282f4dc
JS
918 return 0;
919 }
90cf590f 920 error("object %s: %s", describe_object(o, obj), message);
d6ffc8d7
MK
921 return 1;
922}