fsck: use oidset instead of oid_array for skipList
[git/git.git] / fsck.c
CommitLineData
355885d5 1#include "cache.h"
cbd53a21 2#include "object-store.h"
109cd76d 3#include "repository.h"
355885d5
MK
4#include "object.h"
5#include "blob.h"
6#include "tree.h"
7#include "tree-walk.h"
8#include "commit.h"
9#include "tag.h"
10#include "fsck.h"
cec097be 11#include "refs.h"
a18fcc9f 12#include "utf8.h"
7b35efd7 13#include "decorate.h"
159e7b08 14#include "oidset.h"
27387444 15#include "packfile.h"
ed8b10f6
JK
16#include "submodule-config.h"
17#include "config.h"
3ac68a93 18#include "help.h"
159e7b08
JK
19
20static struct oidset gitmodules_found = OIDSET_INIT;
21static struct oidset gitmodules_done = OIDSET_INIT;
355885d5 22
f50c4407 23#define FSCK_FATAL -1
f27d05b1 24#define FSCK_INFO -2
f50c4407 25
c99ba492 26#define FOREACH_MSG_ID(FUNC) \
f50c4407
JS
27 /* fatal errors */ \
28 FUNC(NUL_IN_HEADER, FATAL) \
29 FUNC(UNTERMINATED_HEADER, FATAL) \
c99ba492
JS
30 /* errors */ \
31 FUNC(BAD_DATE, ERROR) \
32 FUNC(BAD_DATE_OVERFLOW, ERROR) \
33 FUNC(BAD_EMAIL, ERROR) \
34 FUNC(BAD_NAME, ERROR) \
35 FUNC(BAD_OBJECT_SHA1, ERROR) \
36 FUNC(BAD_PARENT_SHA1, ERROR) \
37 FUNC(BAD_TAG_OBJECT, ERROR) \
38 FUNC(BAD_TIMEZONE, ERROR) \
39 FUNC(BAD_TREE, ERROR) \
40 FUNC(BAD_TREE_SHA1, ERROR) \
41 FUNC(BAD_TYPE, ERROR) \
42 FUNC(DUPLICATE_ENTRIES, ERROR) \
43 FUNC(MISSING_AUTHOR, ERROR) \
44 FUNC(MISSING_COMMITTER, ERROR) \
45 FUNC(MISSING_EMAIL, ERROR) \
46 FUNC(MISSING_GRAFT, ERROR) \
47 FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
48 FUNC(MISSING_OBJECT, ERROR) \
49 FUNC(MISSING_PARENT, ERROR) \
50 FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
51 FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
52 FUNC(MISSING_TAG, ERROR) \
53 FUNC(MISSING_TAG_ENTRY, ERROR) \
54 FUNC(MISSING_TAG_OBJECT, ERROR) \
55 FUNC(MISSING_TREE, ERROR) \
159e7b08 56 FUNC(MISSING_TREE_OBJECT, ERROR) \
c99ba492
JS
57 FUNC(MISSING_TYPE, ERROR) \
58 FUNC(MISSING_TYPE_ENTRY, ERROR) \
c9ad147f 59 FUNC(MULTIPLE_AUTHORS, ERROR) \
c99ba492
JS
60 FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
61 FUNC(TREE_NOT_SORTED, ERROR) \
62 FUNC(UNKNOWN_TYPE, ERROR) \
c99ba492 63 FUNC(ZERO_PADDED_DATE, ERROR) \
159e7b08
JK
64 FUNC(GITMODULES_MISSING, ERROR) \
65 FUNC(GITMODULES_BLOB, ERROR) \
0d68764d 66 FUNC(GITMODULES_LARGE, ERROR) \
ed8b10f6 67 FUNC(GITMODULES_NAME, ERROR) \
b7b1fca1 68 FUNC(GITMODULES_SYMLINK, ERROR) \
c99ba492
JS
69 /* warnings */ \
70 FUNC(BAD_FILEMODE, WARN) \
c99ba492
JS
71 FUNC(EMPTY_NAME, WARN) \
72 FUNC(FULL_PATHNAME, WARN) \
73 FUNC(HAS_DOT, WARN) \
74 FUNC(HAS_DOTDOT, WARN) \
75 FUNC(HAS_DOTGIT, WARN) \
c99ba492 76 FUNC(NULL_SHA1, WARN) \
f27d05b1 77 FUNC(ZERO_PADDED_FILEMODE, WARN) \
6d2d780f 78 FUNC(NUL_IN_COMMIT, WARN) \
f27d05b1 79 /* infos (reported as warnings, but ignored by default) */ \
64eb14d3 80 FUNC(GITMODULES_PARSE, INFO) \
f27d05b1
JS
81 FUNC(BAD_TAG_NAME, INFO) \
82 FUNC(MISSING_TAGGER_ENTRY, INFO)
c99ba492
JS
83
84#define MSG_ID(id, msg_type) FSCK_MSG_##id,
85enum fsck_msg_id {
86 FOREACH_MSG_ID(MSG_ID)
87 FSCK_MSG_MAX
88};
89#undef MSG_ID
90
f417eed8 91#define STR(x) #x
a4a9cc19 92#define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
c99ba492 93static struct {
f417eed8
JS
94 const char *id_string;
95 const char *downcased;
a4a9cc19 96 const char *camelcased;
c99ba492
JS
97 int msg_type;
98} msg_id_info[FSCK_MSG_MAX + 1] = {
99 FOREACH_MSG_ID(MSG_ID)
a4a9cc19 100 { NULL, NULL, NULL, -1 }
c99ba492
JS
101};
102#undef MSG_ID
103
a46baac6 104static void prepare_msg_ids(void)
f417eed8
JS
105{
106 int i;
107
a46baac6
NTND
108 if (msg_id_info[0].downcased)
109 return;
110
111 /* convert id_string to lower case, without underscores. */
112 for (i = 0; i < FSCK_MSG_MAX; i++) {
113 const char *p = msg_id_info[i].id_string;
114 int len = strlen(p);
115 char *q = xmalloc(len);
116
117 msg_id_info[i].downcased = q;
118 while (*p)
119 if (*p == '_')
120 p++;
121 else
122 *(q)++ = tolower(*(p)++);
123 *q = '\0';
a4a9cc19
NTND
124
125 p = msg_id_info[i].id_string;
126 q = xmalloc(len);
127 msg_id_info[i].camelcased = q;
128 while (*p) {
129 if (*p == '_') {
130 p++;
131 if (*p)
132 *q++ = *p++;
133 } else {
134 *q++ = tolower(*p++);
135 }
f417eed8 136 }
a4a9cc19 137 *q = '\0';
f417eed8 138 }
a46baac6
NTND
139}
140
141static int parse_msg_id(const char *text)
142{
143 int i;
144
145 prepare_msg_ids();
f417eed8
JS
146
147 for (i = 0; i < FSCK_MSG_MAX; i++)
148 if (!strcmp(text, msg_id_info[i].downcased))
149 return i;
150
151 return -1;
152}
153
3ac68a93
NTND
154void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
155{
156 int i;
157
158 prepare_msg_ids();
159
3ac68a93 160 for (i = 0; i < FSCK_MSG_MAX; i++)
a4a9cc19 161 list_config_item(list, prefix, msg_id_info[i].camelcased);
3ac68a93
NTND
162}
163
c99ba492
JS
164static int fsck_msg_type(enum fsck_msg_id msg_id,
165 struct fsck_options *options)
166{
167 int msg_type;
168
0282f4dc
JS
169 assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
170
171 if (options->msg_type)
172 msg_type = options->msg_type[msg_id];
173 else {
174 msg_type = msg_id_info[msg_id].msg_type;
175 if (options->strict && msg_type == FSCK_WARN)
176 msg_type = FSCK_ERROR;
177 }
c99ba492
JS
178
179 return msg_type;
180}
181
cd94c6f9
JS
182static void init_skiplist(struct fsck_options *options, const char *path)
183{
fb895207
RS
184 FILE *fp;
185 struct strbuf sb = STRBUF_INIT;
365c27fb 186 struct object_id oid;
cd94c6f9 187
fb895207
RS
188 fp = fopen(path, "r");
189 if (!fp)
cd94c6f9 190 die("Could not open skip list: %s", path);
fb895207 191 while (!strbuf_getline(&sb, fp)) {
365c27fb 192 const char *p;
fb895207
RS
193 if (parse_oid_hex(sb.buf, &oid, &p) || *p != '\0')
194 die("Invalid SHA-1: %s", sb.buf);
3b41fb0c 195 oidset_insert(&options->skiplist, &oid);
cd94c6f9 196 }
fb895207
RS
197 if (ferror(fp))
198 die_errno("Could not read '%s'", path);
199 fclose(fp);
200 strbuf_release(&sb);
cd94c6f9
JS
201}
202
0282f4dc
JS
203static int parse_msg_type(const char *str)
204{
205 if (!strcmp(str, "error"))
206 return FSCK_ERROR;
207 else if (!strcmp(str, "warn"))
208 return FSCK_WARN;
efaba7cc
JS
209 else if (!strcmp(str, "ignore"))
210 return FSCK_IGNORE;
0282f4dc
JS
211 else
212 die("Unknown fsck message type: '%s'", str);
213}
214
5d477a33
JS
215int is_valid_msg_type(const char *msg_id, const char *msg_type)
216{
217 if (parse_msg_id(msg_id) < 0)
218 return 0;
219 parse_msg_type(msg_type);
220 return 1;
221}
222
0282f4dc
JS
223void fsck_set_msg_type(struct fsck_options *options,
224 const char *msg_id, const char *msg_type)
225{
226 int id = parse_msg_id(msg_id), type;
227
228 if (id < 0)
229 die("Unhandled message id: %s", msg_id);
230 type = parse_msg_type(msg_type);
231
f50c4407
JS
232 if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
233 die("Cannot demote %s to %s", msg_id, msg_type);
234
0282f4dc
JS
235 if (!options->msg_type) {
236 int i;
b32fa95f
JK
237 int *msg_type;
238 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
0282f4dc
JS
239 for (i = 0; i < FSCK_MSG_MAX; i++)
240 msg_type[i] = fsck_msg_type(i, options);
241 options->msg_type = msg_type;
242 }
243
244 options->msg_type[id] = type;
245}
246
247void fsck_set_msg_types(struct fsck_options *options, const char *values)
248{
249 char *buf = xstrdup(values), *to_free = buf;
250 int done = 0;
251
252 while (!done) {
253 int len = strcspn(buf, " ,|"), equal;
254
255 done = !buf[len];
256 if (!len) {
257 buf++;
258 continue;
259 }
260 buf[len] = '\0';
261
262 for (equal = 0;
263 equal < len && buf[equal] != '=' && buf[equal] != ':';
264 equal++)
265 buf[equal] = tolower(buf[equal]);
266 buf[equal] = '\0';
267
cd94c6f9
JS
268 if (!strcmp(buf, "skiplist")) {
269 if (equal == len)
270 die("skiplist requires a path");
271 init_skiplist(options, buf + equal + 1);
272 buf += len + 1;
273 continue;
274 }
275
0282f4dc
JS
276 if (equal == len)
277 die("Missing '=': '%s'", buf);
355885d5 278
0282f4dc
JS
279 fsck_set_msg_type(options, buf, buf + equal + 1);
280 buf += len + 1;
281 }
282 free(to_free);
283}
284
71ab8fa8
JS
285static void append_msg_id(struct strbuf *sb, const char *msg_id)
286{
287 for (;;) {
288 char c = *(msg_id)++;
289
290 if (!c)
291 break;
292 if (c != '_')
293 strbuf_addch(sb, tolower(c));
294 else {
295 assert(*msg_id);
296 strbuf_addch(sb, *(msg_id)++);
297 }
298 }
299
300 strbuf_addstr(sb, ": ");
301}
302
fb162877
RJ
303static int object_on_skiplist(struct fsck_options *opts, struct object *obj)
304{
3b41fb0c 305 return opts && obj && oidset_contains(&opts->skiplist, &obj->oid);
fb162877
RJ
306}
307
c99ba492
JS
308__attribute__((format (printf, 4, 5)))
309static int report(struct fsck_options *options, struct object *object,
310 enum fsck_msg_id id, const char *fmt, ...)
311{
312 va_list ap;
313 struct strbuf sb = STRBUF_INIT;
314 int msg_type = fsck_msg_type(id, options), result;
315
efaba7cc
JS
316 if (msg_type == FSCK_IGNORE)
317 return 0;
318
fb162877 319 if (object_on_skiplist(options, object))
cd94c6f9
JS
320 return 0;
321
f50c4407
JS
322 if (msg_type == FSCK_FATAL)
323 msg_type = FSCK_ERROR;
f27d05b1
JS
324 else if (msg_type == FSCK_INFO)
325 msg_type = FSCK_WARN;
f50c4407 326
71ab8fa8
JS
327 append_msg_id(&sb, msg_id_info[id].id_string);
328
c99ba492
JS
329 va_start(ap, fmt);
330 strbuf_vaddf(&sb, fmt, ap);
1cd772cc 331 result = options->error_func(options, object, msg_type, sb.buf);
c99ba492
JS
332 strbuf_release(&sb);
333 va_end(ap);
334
335 return result;
336}
337
7b35efd7
JS
338static char *get_object_name(struct fsck_options *options, struct object *obj)
339{
340 if (!options->object_names)
341 return NULL;
342 return lookup_decoration(options->object_names, obj);
343}
344
345static void put_object_name(struct fsck_options *options, struct object *obj,
346 const char *fmt, ...)
347{
348 va_list ap;
349 struct strbuf buf = STRBUF_INIT;
350 char *existing;
351
352 if (!options->object_names)
353 return;
354 existing = lookup_decoration(options->object_names, obj);
355 if (existing)
356 return;
357 va_start(ap, fmt);
358 strbuf_vaddf(&buf, fmt, ap);
359 add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
360 va_end(ap);
361}
362
90cf590f
JS
363static const char *describe_object(struct fsck_options *o, struct object *obj)
364{
365 static struct strbuf buf = STRBUF_INIT;
366 char *name;
367
368 strbuf_reset(&buf);
369 strbuf_addstr(&buf, oid_to_hex(&obj->oid));
370 if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
371 strbuf_addf(&buf, " (%s)", name);
372
373 return buf.buf;
374}
375
22410549 376static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
355885d5
MK
377{
378 struct tree_desc desc;
379 struct name_entry entry;
380 int res = 0;
7b35efd7 381 const char *name;
355885d5
MK
382
383 if (parse_tree(tree))
384 return -1;
385
7b35efd7 386 name = get_object_name(options, &tree->object);
8354fa3d
DT
387 if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
388 return -1;
389 while (tree_entry_gently(&desc, &entry)) {
7b35efd7 390 struct object *obj;
355885d5
MK
391 int result;
392
393 if (S_ISGITLINK(entry.mode))
394 continue;
7b35efd7
JS
395
396 if (S_ISDIR(entry.mode)) {
f86bcc7b 397 obj = (struct object *)lookup_tree(the_repository, entry.oid);
2720f6db 398 if (name && obj)
7b35efd7
JS
399 put_object_name(options, obj, "%s%s/", name,
400 entry.path);
401 result = options->walk(obj, OBJ_TREE, data, options);
402 }
403 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
da14a7ff 404 obj = (struct object *)lookup_blob(the_repository, entry.oid);
2720f6db 405 if (name && obj)
7b35efd7
JS
406 put_object_name(options, obj, "%s%s", name,
407 entry.path);
408 result = options->walk(obj, OBJ_BLOB, data, options);
409 }
355885d5 410 else {
82247e9b 411 result = error("in tree %s: entry %s has bad mode %.6o",
90cf590f 412 describe_object(options, &tree->object), entry.path, entry.mode);
355885d5
MK
413 }
414 if (result < 0)
415 return result;
416 if (!res)
417 res = result;
418 }
419 return res;
420}
421
22410549 422static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
355885d5 423{
7b35efd7 424 int counter = 0, generation = 0, name_prefix_len = 0;
355885d5
MK
425 struct commit_list *parents;
426 int res;
427 int result;
7b35efd7 428 const char *name;
355885d5
MK
429
430 if (parse_commit(commit))
431 return -1;
432
7b35efd7
JS
433 name = get_object_name(options, &commit->object);
434 if (name)
2e27bd77
DS
435 put_object_name(options, &get_commit_tree(commit)->object,
436 "%s:", name);
7b35efd7 437
2e27bd77
DS
438 result = options->walk((struct object *)get_commit_tree(commit),
439 OBJ_TREE, data, options);
355885d5
MK
440 if (result < 0)
441 return result;
442 res = result;
443
444 parents = commit->parents;
7b35efd7
JS
445 if (name && parents) {
446 int len = strlen(name), power;
447
448 if (len && name[len - 1] == '^') {
449 generation = 1;
450 name_prefix_len = len - 1;
451 }
452 else { /* parse ~<generation> suffix */
453 for (generation = 0, power = 1;
454 len && isdigit(name[len - 1]);
455 power *= 10)
456 generation += power * (name[--len] - '0');
457 if (power > 1 && len && name[len - 1] == '~')
458 name_prefix_len = len - 1;
459 }
460 }
461
355885d5 462 while (parents) {
7b35efd7
JS
463 if (name) {
464 struct object *obj = &parents->item->object;
465
466 if (++counter > 1)
467 put_object_name(options, obj, "%s^%d",
468 name, counter);
469 else if (generation > 0)
470 put_object_name(options, obj, "%.*s~%d",
471 name_prefix_len, name, generation + 1);
472 else
473 put_object_name(options, obj, "%s^", name);
474 }
22410549 475 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
355885d5
MK
476 if (result < 0)
477 return result;
478 if (!res)
479 res = result;
480 parents = parents->next;
481 }
482 return res;
483}
484
22410549 485static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
355885d5 486{
7b35efd7
JS
487 char *name = get_object_name(options, &tag->object);
488
355885d5
MK
489 if (parse_tag(tag))
490 return -1;
7b35efd7
JS
491 if (name)
492 put_object_name(options, tag->tagged, "%s", name);
22410549 493 return options->walk(tag->tagged, OBJ_ANY, data, options);
355885d5
MK
494}
495
22410549 496int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
355885d5
MK
497{
498 if (!obj)
499 return -1;
a2b22854
JK
500
501 if (obj->type == OBJ_NONE)
109cd76d 502 parse_object(the_repository, &obj->oid);
a2b22854 503
355885d5
MK
504 switch (obj->type) {
505 case OBJ_BLOB:
506 return 0;
507 case OBJ_TREE:
22410549 508 return fsck_walk_tree((struct tree *)obj, data, options);
355885d5 509 case OBJ_COMMIT:
22410549 510 return fsck_walk_commit((struct commit *)obj, data, options);
355885d5 511 case OBJ_TAG:
22410549 512 return fsck_walk_tag((struct tag *)obj, data, options);
355885d5 513 default:
90cf590f 514 error("Unknown object type for %s", describe_object(options, obj));
355885d5
MK
515 return -1;
516 }
517}
ba002f3b
MK
518
519/*
520 * The entries in a tree are ordered in the _path_ order,
521 * which means that a directory entry is ordered by adding
522 * a slash to the end of it.
523 *
524 * So a directory called "a" is ordered _after_ a file
525 * called "a.c", because "a/" sorts after "a.c".
526 */
527#define TREE_UNORDERED (-1)
528#define TREE_HAS_DUPS (-2)
529
530static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
531{
532 int len1 = strlen(name1);
533 int len2 = strlen(name2);
534 int len = len1 < len2 ? len1 : len2;
535 unsigned char c1, c2;
536 int cmp;
537
538 cmp = memcmp(name1, name2, len);
539 if (cmp < 0)
540 return 0;
541 if (cmp > 0)
542 return TREE_UNORDERED;
543
544 /*
545 * Ok, the first <len> characters are the same.
546 * Now we need to order the next one, but turn
547 * a '\0' into a '/' for a directory entry.
548 */
549 c1 = name1[len];
550 c2 = name2[len];
551 if (!c1 && !c2)
552 /*
553 * git-write-tree used to write out a nonsense tree that has
554 * entries with the same name, one blob and one tree. Make
555 * sure we do not have duplicate entries.
556 */
557 return TREE_HAS_DUPS;
558 if (!c1 && S_ISDIR(mode1))
559 c1 = '/';
560 if (!c2 && S_ISDIR(mode2))
561 c2 = '/';
562 return c1 < c2 ? 0 : TREE_UNORDERED;
563}
564
22410549 565static int fsck_tree(struct tree *item, struct fsck_options *options)
ba002f3b 566{
8354fa3d 567 int retval = 0;
c479d14a 568 int has_null_sha1 = 0;
ba002f3b
MK
569 int has_full_path = 0;
570 int has_empty_name = 0;
5d34a435
JK
571 int has_dot = 0;
572 int has_dotdot = 0;
5c17f512 573 int has_dotgit = 0;
ba002f3b
MK
574 int has_zero_pad = 0;
575 int has_bad_modes = 0;
576 int has_dup_entries = 0;
577 int not_properly_sorted = 0;
578 struct tree_desc desc;
579 unsigned o_mode;
580 const char *o_name;
ba002f3b 581
8354fa3d
DT
582 if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
583 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
584 return retval;
585 }
ba002f3b
MK
586
587 o_mode = 0;
588 o_name = NULL;
ba002f3b
MK
589
590 while (desc.size) {
591 unsigned mode;
592 const char *name;
ce6663a9 593 const struct object_id *oid;
ba002f3b 594
ce6663a9 595 oid = tree_entry_extract(&desc, &name, &mode);
ba002f3b 596
ce6663a9 597 has_null_sha1 |= is_null_oid(oid);
effd12ec
HS
598 has_full_path |= !!strchr(name, '/');
599 has_empty_name |= !*name;
600 has_dot |= !strcmp(name, ".");
601 has_dotdot |= !strcmp(name, "..");
ed9c3220 602 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
ba002f3b 603 has_zero_pad |= *(char *)desc.buffer == '0';
159e7b08 604
b7b1fca1
JK
605 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
606 if (!S_ISLNK(mode))
607 oidset_insert(&gitmodules_found, oid);
608 else
609 retval += report(options, &item->object,
610 FSCK_MSG_GITMODULES_SYMLINK,
611 ".gitmodules is a symbolic link");
612 }
159e7b08 613
8354fa3d
DT
614 if (update_tree_entry_gently(&desc)) {
615 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
616 break;
617 }
ba002f3b
MK
618
619 switch (mode) {
620 /*
621 * Standard modes..
622 */
623 case S_IFREG | 0755:
624 case S_IFREG | 0644:
625 case S_IFLNK:
626 case S_IFDIR:
627 case S_IFGITLINK:
628 break;
629 /*
630 * This is nonstandard, but we had a few of these
631 * early on when we honored the full set of mode
632 * bits..
633 */
634 case S_IFREG | 0664:
22410549 635 if (!options->strict)
ba002f3b 636 break;
1cf01a34 637 /* fallthrough */
ba002f3b
MK
638 default:
639 has_bad_modes = 1;
640 }
641
642 if (o_name) {
643 switch (verify_ordered(o_mode, o_name, mode, name)) {
644 case TREE_UNORDERED:
645 not_properly_sorted = 1;
646 break;
647 case TREE_HAS_DUPS:
648 has_dup_entries = 1;
649 break;
650 default:
651 break;
652 }
653 }
654
655 o_mode = mode;
656 o_name = name;
ba002f3b
MK
657 }
658
c479d14a 659 if (has_null_sha1)
c99ba492 660 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
ba002f3b 661 if (has_full_path)
c99ba492 662 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
ba002f3b 663 if (has_empty_name)
c99ba492 664 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
5d34a435 665 if (has_dot)
c99ba492 666 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
5d34a435 667 if (has_dotdot)
c99ba492 668 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
5c17f512 669 if (has_dotgit)
c99ba492 670 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
ba002f3b 671 if (has_zero_pad)
c99ba492 672 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
ba002f3b 673 if (has_bad_modes)
c99ba492 674 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
ba002f3b 675 if (has_dup_entries)
c99ba492 676 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
ba002f3b 677 if (not_properly_sorted)
c99ba492 678 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
ba002f3b
MK
679 return retval;
680}
681
84d18c0b 682static int verify_headers(const void *data, unsigned long size,
b2f44feb 683 struct object *obj, struct fsck_options *options)
4d0d8975
JS
684{
685 const char *buffer = (const char *)data;
686 unsigned long i;
687
688 for (i = 0; i < size; i++) {
689 switch (buffer[i]) {
690 case '\0':
c99ba492
JS
691 return report(options, obj,
692 FSCK_MSG_NUL_IN_HEADER,
693 "unterminated header: NUL at offset %ld", i);
4d0d8975
JS
694 case '\n':
695 if (i + 1 < size && buffer[i + 1] == '\n')
696 return 0;
697 }
698 }
699
84d18c0b
JH
700 /*
701 * We did not find double-LF that separates the header
702 * and the body. Not having a body is not a crime but
703 * we do want to see the terminating LF for the last header
704 * line.
705 */
706 if (size && buffer[size - 1] == '\n')
707 return 0;
708
c99ba492
JS
709 return report(options, obj,
710 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
4d0d8975
JS
711}
712
22410549 713static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
daae1922 714{
e6826e33 715 const char *p = *ident;
d4b8de04
JK
716 char *end;
717
e6826e33
JS
718 *ident = strchrnul(*ident, '\n');
719 if (**ident == '\n')
720 (*ident)++;
721
722 if (*p == '<')
c99ba492 723 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
724 p += strcspn(p, "<>\n");
725 if (*p == '>')
c99ba492 726 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
e6826e33 727 if (*p != '<')
c99ba492 728 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
e6826e33 729 if (p[-1] != ' ')
c99ba492 730 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
731 p++;
732 p += strcspn(p, "<>\n");
733 if (*p != '>')
c99ba492 734 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
e6826e33
JS
735 p++;
736 if (*p != ' ')
c99ba492 737 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
e6826e33
JS
738 p++;
739 if (*p == '0' && p[1] != ' ')
c99ba492 740 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
1aeb7e75 741 if (date_overflows(parse_timestamp(p, &end, 10)))
c99ba492 742 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
e6826e33 743 if ((end == p || *end != ' '))
c99ba492 744 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
e6826e33
JS
745 p = end + 1;
746 if ((*p != '+' && *p != '-') ||
747 !isdigit(p[1]) ||
748 !isdigit(p[2]) ||
749 !isdigit(p[3]) ||
750 !isdigit(p[4]) ||
751 (p[5] != '\n'))
c99ba492 752 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
e6826e33 753 p += 6;
daae1922
JN
754 return 0;
755}
756
bc6b8fc1 757static int fsck_commit_buffer(struct commit *commit, const char *buffer,
22410549 758 unsigned long size, struct fsck_options *options)
ba002f3b 759{
c54f5ca9 760 struct object_id tree_oid, oid;
ba002f3b 761 struct commit_graft *graft;
c9ad147f 762 unsigned parent_count, parent_line_count = 0, author_count;
daae1922 763 int err;
6d2d780f 764 const char *buffer_begin = buffer;
c54f5ca9 765 const char *p;
ba002f3b 766
b2f44feb 767 if (verify_headers(buffer, size, &commit->object, options))
4d0d8975
JS
768 return -1;
769
cf4fff57 770 if (!skip_prefix(buffer, "tree ", &buffer))
c99ba492 771 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
c54f5ca9 772 if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
b3584761
JS
773 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
774 if (err)
775 return err;
776 }
c54f5ca9 777 buffer = p + 1;
cf4fff57 778 while (skip_prefix(buffer, "parent ", &buffer)) {
c54f5ca9 779 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
b3584761
JS
780 err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
781 if (err)
782 return err;
783 }
c54f5ca9 784 buffer = p + 1;
9d02150c 785 parent_line_count++;
ba002f3b 786 }
1f93ecd1 787 graft = lookup_commit_graft(the_repository, &commit->object.oid);
9d02150c 788 parent_count = commit_list_count(commit->parents);
ba002f3b 789 if (graft) {
9d02150c 790 if (graft->nr_parent == -1 && !parent_count)
ba002f3b 791 ; /* shallow commit */
b3584761
JS
792 else if (graft->nr_parent != parent_count) {
793 err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
794 if (err)
795 return err;
796 }
ba002f3b 797 } else {
b3584761
JS
798 if (parent_count != parent_line_count) {
799 err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
800 if (err)
801 return err;
802 }
ba002f3b 803 }
c9ad147f
JS
804 author_count = 0;
805 while (skip_prefix(buffer, "author ", &buffer)) {
806 author_count++;
807 err = fsck_ident(&buffer, &commit->object, options);
808 if (err)
809 return err;
ba002f3b 810 }
c9ad147f
JS
811 if (author_count < 1)
812 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
813 else if (author_count > 1)
814 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
daae1922
JN
815 if (err)
816 return err;
cf4fff57 817 if (!skip_prefix(buffer, "committer ", &buffer))
c99ba492 818 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
22410549 819 err = fsck_ident(&buffer, &commit->object, options);
daae1922
JN
820 if (err)
821 return err;
2e27bd77 822 if (!get_commit_tree(commit)) {
c54f5ca9 823 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
5af29718
JH
824 if (err)
825 return err;
826 }
6d2d780f
JH
827 if (memchr(buffer_begin, '\0', size)) {
828 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
829 "NUL byte in the commit object body");
830 if (err)
831 return err;
832 }
ba002f3b
MK
833 return 0;
834}
835
90a398bb 836static int fsck_commit(struct commit *commit, const char *data,
22410549 837 unsigned long size, struct fsck_options *options)
bc6b8fc1 838{
90a398bb 839 const char *buffer = data ? data : get_commit_buffer(commit, &size);
22410549 840 int ret = fsck_commit_buffer(commit, buffer, size, options);
90a398bb
JS
841 if (!data)
842 unuse_commit_buffer(commit, buffer);
bc6b8fc1
JK
843 return ret;
844}
845
cec097be 846static int fsck_tag_buffer(struct tag *tag, const char *data,
22410549 847 unsigned long size, struct fsck_options *options)
cec097be 848{
c54f5ca9 849 struct object_id oid;
cec097be
JS
850 int ret = 0;
851 const char *buffer;
852 char *to_free = NULL, *eol;
853 struct strbuf sb = STRBUF_INIT;
c54f5ca9 854 const char *p;
cec097be
JS
855
856 if (data)
857 buffer = data;
858 else {
859 enum object_type type;
860
861 buffer = to_free =
b4f5aca4 862 read_object_file(&tag->object.oid, &type, &size);
cec097be 863 if (!buffer)
c99ba492
JS
864 return report(options, &tag->object,
865 FSCK_MSG_MISSING_TAG_OBJECT,
cec097be
JS
866 "cannot read tag object");
867
868 if (type != OBJ_TAG) {
c99ba492
JS
869 ret = report(options, &tag->object,
870 FSCK_MSG_TAG_OBJECT_NOT_TAG,
cec097be 871 "expected tag got %s",
debca9d2 872 type_name(type));
cec097be
JS
873 goto done;
874 }
875 }
876
8a272f29
RS
877 ret = verify_headers(buffer, size, &tag->object, options);
878 if (ret)
cec097be
JS
879 goto done;
880
881 if (!skip_prefix(buffer, "object ", &buffer)) {
c99ba492 882 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
cec097be
JS
883 goto done;
884 }
c54f5ca9 885 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
c99ba492 886 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
7d7d5b05
JS
887 if (ret)
888 goto done;
cec097be 889 }
c54f5ca9 890 buffer = p + 1;
cec097be
JS
891
892 if (!skip_prefix(buffer, "type ", &buffer)) {
c99ba492 893 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
cec097be
JS
894 goto done;
895 }
896 eol = strchr(buffer, '\n');
897 if (!eol) {
c99ba492 898 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
cec097be
JS
899 goto done;
900 }
901 if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
c99ba492 902 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
cec097be
JS
903 if (ret)
904 goto done;
905 buffer = eol + 1;
906
907 if (!skip_prefix(buffer, "tag ", &buffer)) {
c99ba492 908 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
cec097be
JS
909 goto done;
910 }
911 eol = strchr(buffer, '\n');
912 if (!eol) {
c99ba492 913 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
cec097be
JS
914 goto done;
915 }
916 strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
f27d05b1
JS
917 if (check_refname_format(sb.buf, 0)) {
918 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
c99ba492 919 "invalid 'tag' name: %.*s",
7add4419 920 (int)(eol - buffer), buffer);
f27d05b1
JS
921 if (ret)
922 goto done;
923 }
cec097be
JS
924 buffer = eol + 1;
925
f27d05b1 926 if (!skip_prefix(buffer, "tagger ", &buffer)) {
cec097be 927 /* early tags do not contain 'tagger' lines; warn only */
f27d05b1
JS
928 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
929 if (ret)
930 goto done;
931 }
cec097be 932 else
22410549 933 ret = fsck_ident(&buffer, &tag->object, options);
cec097be
JS
934
935done:
936 strbuf_release(&sb);
937 free(to_free);
938 return ret;
939}
940
90a398bb 941static int fsck_tag(struct tag *tag, const char *data,
22410549 942 unsigned long size, struct fsck_options *options)
ba002f3b
MK
943{
944 struct object *tagged = tag->tagged;
945
946 if (!tagged)
c99ba492 947 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
cec097be 948
22410549 949 return fsck_tag_buffer(tag, data, size, options);
ba002f3b
MK
950}
951
ed8b10f6
JK
952struct fsck_gitmodules_data {
953 struct object *obj;
954 struct fsck_options *options;
955 int ret;
956};
957
958static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
959{
960 struct fsck_gitmodules_data *data = vdata;
961 const char *subsection, *key;
962 int subsection_len;
963 char *name;
964
965 if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
966 !subsection)
967 return 0;
968
969 name = xmemdupz(subsection, subsection_len);
970 if (check_submodule_name(name) < 0)
971 data->ret |= report(data->options, data->obj,
972 FSCK_MSG_GITMODULES_NAME,
973 "disallowed submodule name: %s",
974 name);
975 free(name);
976
977 return 0;
978}
979
7ac4f3a0
JK
980static int fsck_blob(struct blob *blob, const char *buf,
981 unsigned long size, struct fsck_options *options)
982{
ed8b10f6 983 struct fsck_gitmodules_data data;
de6bd9e3 984 struct config_options config_opts = { 0 };
ed8b10f6
JK
985
986 if (!oidset_contains(&gitmodules_found, &blob->object.oid))
987 return 0;
988 oidset_insert(&gitmodules_done, &blob->object.oid);
989
fb162877
RJ
990 if (object_on_skiplist(options, &blob->object))
991 return 0;
992
ed8b10f6
JK
993 if (!buf) {
994 /*
995 * A missing buffer here is a sign that the caller found the
996 * blob too gigantic to load into memory. Let's just consider
997 * that an error.
998 */
999 return report(options, &blob->object,
0d68764d 1000 FSCK_MSG_GITMODULES_LARGE,
ed8b10f6
JK
1001 ".gitmodules too large to parse");
1002 }
1003
1004 data.obj = &blob->object;
1005 data.options = options;
1006 data.ret = 0;
de6bd9e3 1007 config_opts.error_action = CONFIG_ERROR_SILENT;
ed8b10f6 1008 if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
de6bd9e3 1009 ".gitmodules", buf, size, &data, &config_opts))
ed8b10f6
JK
1010 data.ret |= report(options, &blob->object,
1011 FSCK_MSG_GITMODULES_PARSE,
1012 "could not parse gitmodules blob");
1013
1014 return data.ret;
7ac4f3a0
JK
1015}
1016
90a398bb 1017int fsck_object(struct object *obj, void *data, unsigned long size,
22410549 1018 struct fsck_options *options)
ba002f3b
MK
1019{
1020 if (!obj)
c99ba492 1021 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
ba002f3b
MK
1022
1023 if (obj->type == OBJ_BLOB)
7ac4f3a0 1024 return fsck_blob((struct blob *)obj, data, size, options);
ba002f3b 1025 if (obj->type == OBJ_TREE)
22410549 1026 return fsck_tree((struct tree *) obj, options);
ba002f3b 1027 if (obj->type == OBJ_COMMIT)
90a398bb 1028 return fsck_commit((struct commit *) obj, (const char *) data,
22410549 1029 size, options);
ba002f3b 1030 if (obj->type == OBJ_TAG)
90a398bb 1031 return fsck_tag((struct tag *) obj, (const char *) data,
22410549 1032 size, options);
ba002f3b 1033
c99ba492 1034 return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
ba002f3b
MK
1035 obj->type);
1036}
d6ffc8d7 1037
1cd772cc
JS
1038int fsck_error_function(struct fsck_options *o,
1039 struct object *obj, int msg_type, const char *message)
d6ffc8d7 1040{
0282f4dc 1041 if (msg_type == FSCK_WARN) {
90cf590f 1042 warning("object %s: %s", describe_object(o, obj), message);
0282f4dc
JS
1043 return 0;
1044 }
90cf590f 1045 error("object %s: %s", describe_object(o, obj), message);
d6ffc8d7
MK
1046 return 1;
1047}
159e7b08
JK
1048
1049int fsck_finish(struct fsck_options *options)
1050{
1051 int ret = 0;
1052 struct oidset_iter iter;
1053 const struct object_id *oid;
1054
1055 oidset_iter_init(&gitmodules_found, &iter);
1056 while ((oid = oidset_iter_next(&iter))) {
1057 struct blob *blob;
1058 enum object_type type;
1059 unsigned long size;
1060 char *buf;
1061
1062 if (oidset_contains(&gitmodules_done, oid))
1063 continue;
1064
da14a7ff 1065 blob = lookup_blob(the_repository, oid);
159e7b08 1066 if (!blob) {
47cc9131
JK
1067 struct object *obj = lookup_unknown_object(oid->hash);
1068 ret |= report(options, obj,
159e7b08
JK
1069 FSCK_MSG_GITMODULES_BLOB,
1070 "non-blob found at .gitmodules");
1071 continue;
1072 }
1073
7913f53b 1074 buf = read_object_file(oid, &type, &size);
159e7b08 1075 if (!buf) {
27387444
JK
1076 if (is_promisor_object(&blob->object.oid))
1077 continue;
159e7b08
JK
1078 ret |= report(options, &blob->object,
1079 FSCK_MSG_GITMODULES_MISSING,
1080 "unable to read .gitmodules blob");
1081 continue;
1082 }
1083
1084 if (type == OBJ_BLOB)
1085 ret |= fsck_blob(blob, buf, size, options);
1086 else
1087 ret |= report(options, &blob->object,
1088 FSCK_MSG_GITMODULES_BLOB,
1089 "non-blob found at .gitmodules");
1090 free(buf);
1091 }
1092
1093
1094 oidset_clear(&gitmodules_found);
1095 oidset_clear(&gitmodules_done);
1096 return ret;
1097}