object: convert lookup_unknown_object() to use object_id
[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) \
a124133e 69 FUNC(GITMODULES_URL, ERROR) \
1a7fd1fb 70 FUNC(GITMODULES_PATH, ERROR) \
c99ba492
JS
71 /* warnings */ \
72 FUNC(BAD_FILEMODE, WARN) \
c99ba492
JS
73 FUNC(EMPTY_NAME, WARN) \
74 FUNC(FULL_PATHNAME, WARN) \
75 FUNC(HAS_DOT, WARN) \
76 FUNC(HAS_DOTDOT, WARN) \
77 FUNC(HAS_DOTGIT, WARN) \
c99ba492 78 FUNC(NULL_SHA1, WARN) \
f27d05b1 79 FUNC(ZERO_PADDED_FILEMODE, WARN) \
6d2d780f 80 FUNC(NUL_IN_COMMIT, WARN) \
f27d05b1 81 /* infos (reported as warnings, but ignored by default) */ \
64eb14d3 82 FUNC(GITMODULES_PARSE, INFO) \
f27d05b1
JS
83 FUNC(BAD_TAG_NAME, INFO) \
84 FUNC(MISSING_TAGGER_ENTRY, INFO)
c99ba492
JS
85
86#define MSG_ID(id, msg_type) FSCK_MSG_##id,
87enum fsck_msg_id {
88 FOREACH_MSG_ID(MSG_ID)
89 FSCK_MSG_MAX
90};
91#undef MSG_ID
92
f417eed8 93#define STR(x) #x
a4a9cc19 94#define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
c99ba492 95static struct {
f417eed8
JS
96 const char *id_string;
97 const char *downcased;
a4a9cc19 98 const char *camelcased;
c99ba492
JS
99 int msg_type;
100} msg_id_info[FSCK_MSG_MAX + 1] = {
101 FOREACH_MSG_ID(MSG_ID)
a4a9cc19 102 { NULL, NULL, NULL, -1 }
c99ba492
JS
103};
104#undef MSG_ID
105
a46baac6 106static void prepare_msg_ids(void)
f417eed8
JS
107{
108 int i;
109
a46baac6
NTND
110 if (msg_id_info[0].downcased)
111 return;
112
113 /* convert id_string to lower case, without underscores. */
114 for (i = 0; i < FSCK_MSG_MAX; i++) {
115 const char *p = msg_id_info[i].id_string;
116 int len = strlen(p);
117 char *q = xmalloc(len);
118
119 msg_id_info[i].downcased = q;
120 while (*p)
121 if (*p == '_')
122 p++;
123 else
124 *(q)++ = tolower(*(p)++);
125 *q = '\0';
a4a9cc19
NTND
126
127 p = msg_id_info[i].id_string;
128 q = xmalloc(len);
129 msg_id_info[i].camelcased = q;
130 while (*p) {
131 if (*p == '_') {
132 p++;
133 if (*p)
134 *q++ = *p++;
135 } else {
136 *q++ = tolower(*p++);
137 }
f417eed8 138 }
a4a9cc19 139 *q = '\0';
f417eed8 140 }
a46baac6
NTND
141}
142
143static int parse_msg_id(const char *text)
144{
145 int i;
146
147 prepare_msg_ids();
f417eed8
JS
148
149 for (i = 0; i < FSCK_MSG_MAX; i++)
150 if (!strcmp(text, msg_id_info[i].downcased))
151 return i;
152
153 return -1;
154}
155
3ac68a93
NTND
156void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
157{
158 int i;
159
160 prepare_msg_ids();
161
3ac68a93 162 for (i = 0; i < FSCK_MSG_MAX; i++)
a4a9cc19 163 list_config_item(list, prefix, msg_id_info[i].camelcased);
3ac68a93
NTND
164}
165
c99ba492
JS
166static int fsck_msg_type(enum fsck_msg_id msg_id,
167 struct fsck_options *options)
168{
169 int msg_type;
170
0282f4dc
JS
171 assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
172
173 if (options->msg_type)
174 msg_type = options->msg_type[msg_id];
175 else {
176 msg_type = msg_id_info[msg_id].msg_type;
177 if (options->strict && msg_type == FSCK_WARN)
178 msg_type = FSCK_ERROR;
179 }
c99ba492
JS
180
181 return msg_type;
182}
183
cd94c6f9
JS
184static void init_skiplist(struct fsck_options *options, const char *path)
185{
fb895207
RS
186 FILE *fp;
187 struct strbuf sb = STRBUF_INIT;
365c27fb 188 struct object_id oid;
cd94c6f9 189
fb895207
RS
190 fp = fopen(path, "r");
191 if (!fp)
cd94c6f9 192 die("Could not open skip list: %s", path);
fb895207 193 while (!strbuf_getline(&sb, fp)) {
365c27fb 194 const char *p;
371a6550
ÆAB
195 const char *hash;
196
197 /*
198 * Allow trailing comments, leading whitespace
199 * (including before commits), and empty or whitespace
200 * only lines.
201 */
202 hash = strchr(sb.buf, '#');
203 if (hash)
204 strbuf_setlen(&sb, hash - sb.buf);
205 strbuf_trim(&sb);
206 if (!sb.len)
207 continue;
cd94c6f9 208
fb895207
RS
209 if (parse_oid_hex(sb.buf, &oid, &p) || *p != '\0')
210 die("Invalid SHA-1: %s", sb.buf);
3b41fb0c 211 oidset_insert(&options->skiplist, &oid);
cd94c6f9 212 }
fb895207
RS
213 if (ferror(fp))
214 die_errno("Could not read '%s'", path);
215 fclose(fp);
216 strbuf_release(&sb);
cd94c6f9
JS
217}
218
0282f4dc
JS
219static int parse_msg_type(const char *str)
220{
221 if (!strcmp(str, "error"))
222 return FSCK_ERROR;
223 else if (!strcmp(str, "warn"))
224 return FSCK_WARN;
efaba7cc
JS
225 else if (!strcmp(str, "ignore"))
226 return FSCK_IGNORE;
0282f4dc
JS
227 else
228 die("Unknown fsck message type: '%s'", str);
229}
230
5d477a33
JS
231int is_valid_msg_type(const char *msg_id, const char *msg_type)
232{
233 if (parse_msg_id(msg_id) < 0)
234 return 0;
235 parse_msg_type(msg_type);
236 return 1;
237}
238
0282f4dc
JS
239void fsck_set_msg_type(struct fsck_options *options,
240 const char *msg_id, const char *msg_type)
241{
242 int id = parse_msg_id(msg_id), type;
243
244 if (id < 0)
245 die("Unhandled message id: %s", msg_id);
246 type = parse_msg_type(msg_type);
247
f50c4407
JS
248 if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
249 die("Cannot demote %s to %s", msg_id, msg_type);
250
0282f4dc
JS
251 if (!options->msg_type) {
252 int i;
b32fa95f
JK
253 int *msg_type;
254 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
0282f4dc
JS
255 for (i = 0; i < FSCK_MSG_MAX; i++)
256 msg_type[i] = fsck_msg_type(i, options);
257 options->msg_type = msg_type;
258 }
259
260 options->msg_type[id] = type;
261}
262
263void fsck_set_msg_types(struct fsck_options *options, const char *values)
264{
265 char *buf = xstrdup(values), *to_free = buf;
266 int done = 0;
267
268 while (!done) {
269 int len = strcspn(buf, " ,|"), equal;
270
271 done = !buf[len];
272 if (!len) {
273 buf++;
274 continue;
275 }
276 buf[len] = '\0';
277
278 for (equal = 0;
279 equal < len && buf[equal] != '=' && buf[equal] != ':';
280 equal++)
281 buf[equal] = tolower(buf[equal]);
282 buf[equal] = '\0';
283
cd94c6f9
JS
284 if (!strcmp(buf, "skiplist")) {
285 if (equal == len)
286 die("skiplist requires a path");
287 init_skiplist(options, buf + equal + 1);
288 buf += len + 1;
289 continue;
290 }
291
0282f4dc
JS
292 if (equal == len)
293 die("Missing '=': '%s'", buf);
355885d5 294
0282f4dc
JS
295 fsck_set_msg_type(options, buf, buf + equal + 1);
296 buf += len + 1;
297 }
298 free(to_free);
299}
300
71ab8fa8
JS
301static void append_msg_id(struct strbuf *sb, const char *msg_id)
302{
303 for (;;) {
304 char c = *(msg_id)++;
305
306 if (!c)
307 break;
308 if (c != '_')
309 strbuf_addch(sb, tolower(c));
310 else {
311 assert(*msg_id);
312 strbuf_addch(sb, *(msg_id)++);
313 }
314 }
315
316 strbuf_addstr(sb, ": ");
317}
318
fb162877
RJ
319static int object_on_skiplist(struct fsck_options *opts, struct object *obj)
320{
3b41fb0c 321 return opts && obj && oidset_contains(&opts->skiplist, &obj->oid);
fb162877
RJ
322}
323
c99ba492
JS
324__attribute__((format (printf, 4, 5)))
325static int report(struct fsck_options *options, struct object *object,
326 enum fsck_msg_id id, const char *fmt, ...)
327{
328 va_list ap;
329 struct strbuf sb = STRBUF_INIT;
330 int msg_type = fsck_msg_type(id, options), result;
331
efaba7cc
JS
332 if (msg_type == FSCK_IGNORE)
333 return 0;
334
fb162877 335 if (object_on_skiplist(options, object))
cd94c6f9
JS
336 return 0;
337
f50c4407
JS
338 if (msg_type == FSCK_FATAL)
339 msg_type = FSCK_ERROR;
f27d05b1
JS
340 else if (msg_type == FSCK_INFO)
341 msg_type = FSCK_WARN;
f50c4407 342
71ab8fa8
JS
343 append_msg_id(&sb, msg_id_info[id].id_string);
344
c99ba492
JS
345 va_start(ap, fmt);
346 strbuf_vaddf(&sb, fmt, ap);
1cd772cc 347 result = options->error_func(options, object, msg_type, sb.buf);
c99ba492
JS
348 strbuf_release(&sb);
349 va_end(ap);
350
351 return result;
352}
353
7b35efd7
JS
354static char *get_object_name(struct fsck_options *options, struct object *obj)
355{
356 if (!options->object_names)
357 return NULL;
358 return lookup_decoration(options->object_names, obj);
359}
360
361static void put_object_name(struct fsck_options *options, struct object *obj,
362 const char *fmt, ...)
363{
364 va_list ap;
365 struct strbuf buf = STRBUF_INIT;
366 char *existing;
367
368 if (!options->object_names)
369 return;
370 existing = lookup_decoration(options->object_names, obj);
371 if (existing)
372 return;
373 va_start(ap, fmt);
374 strbuf_vaddf(&buf, fmt, ap);
375 add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
376 va_end(ap);
377}
378
90cf590f
JS
379static const char *describe_object(struct fsck_options *o, struct object *obj)
380{
381 static struct strbuf buf = STRBUF_INIT;
382 char *name;
383
384 strbuf_reset(&buf);
385 strbuf_addstr(&buf, oid_to_hex(&obj->oid));
386 if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
387 strbuf_addf(&buf, " (%s)", name);
388
389 return buf.buf;
390}
391
22410549 392static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
355885d5
MK
393{
394 struct tree_desc desc;
395 struct name_entry entry;
396 int res = 0;
7b35efd7 397 const char *name;
355885d5
MK
398
399 if (parse_tree(tree))
400 return -1;
401
7b35efd7 402 name = get_object_name(options, &tree->object);
8354fa3d
DT
403 if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
404 return -1;
405 while (tree_entry_gently(&desc, &entry)) {
7b35efd7 406 struct object *obj;
355885d5
MK
407 int result;
408
409 if (S_ISGITLINK(entry.mode))
410 continue;
7b35efd7
JS
411
412 if (S_ISDIR(entry.mode)) {
ea82b2a0 413 obj = (struct object *)lookup_tree(the_repository, &entry.oid);
2720f6db 414 if (name && obj)
7b35efd7
JS
415 put_object_name(options, obj, "%s%s/", name,
416 entry.path);
417 result = options->walk(obj, OBJ_TREE, data, options);
418 }
419 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
ea82b2a0 420 obj = (struct object *)lookup_blob(the_repository, &entry.oid);
2720f6db 421 if (name && obj)
7b35efd7
JS
422 put_object_name(options, obj, "%s%s", name,
423 entry.path);
424 result = options->walk(obj, OBJ_BLOB, data, options);
425 }
355885d5 426 else {
82247e9b 427 result = error("in tree %s: entry %s has bad mode %.6o",
90cf590f 428 describe_object(options, &tree->object), entry.path, entry.mode);
355885d5
MK
429 }
430 if (result < 0)
431 return result;
432 if (!res)
433 res = result;
434 }
435 return res;
436}
437
22410549 438static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
355885d5 439{
7b35efd7 440 int counter = 0, generation = 0, name_prefix_len = 0;
355885d5
MK
441 struct commit_list *parents;
442 int res;
443 int result;
7b35efd7 444 const char *name;
355885d5
MK
445
446 if (parse_commit(commit))
447 return -1;
448
7b35efd7
JS
449 name = get_object_name(options, &commit->object);
450 if (name)
2e27bd77
DS
451 put_object_name(options, &get_commit_tree(commit)->object,
452 "%s:", name);
7b35efd7 453
2e27bd77
DS
454 result = options->walk((struct object *)get_commit_tree(commit),
455 OBJ_TREE, data, options);
355885d5
MK
456 if (result < 0)
457 return result;
458 res = result;
459
460 parents = commit->parents;
7b35efd7
JS
461 if (name && parents) {
462 int len = strlen(name), power;
463
464 if (len && name[len - 1] == '^') {
465 generation = 1;
466 name_prefix_len = len - 1;
467 }
468 else { /* parse ~<generation> suffix */
469 for (generation = 0, power = 1;
470 len && isdigit(name[len - 1]);
471 power *= 10)
472 generation += power * (name[--len] - '0');
473 if (power > 1 && len && name[len - 1] == '~')
474 name_prefix_len = len - 1;
475 }
476 }
477
355885d5 478 while (parents) {
7b35efd7
JS
479 if (name) {
480 struct object *obj = &parents->item->object;
481
b84c7838 482 if (counter++)
7b35efd7
JS
483 put_object_name(options, obj, "%s^%d",
484 name, counter);
485 else if (generation > 0)
486 put_object_name(options, obj, "%.*s~%d",
487 name_prefix_len, name, generation + 1);
488 else
489 put_object_name(options, obj, "%s^", name);
490 }
22410549 491 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
355885d5
MK
492 if (result < 0)
493 return result;
494 if (!res)
495 res = result;
496 parents = parents->next;
497 }
498 return res;
499}
500
22410549 501static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
355885d5 502{
7b35efd7
JS
503 char *name = get_object_name(options, &tag->object);
504
355885d5
MK
505 if (parse_tag(tag))
506 return -1;
7b35efd7
JS
507 if (name)
508 put_object_name(options, tag->tagged, "%s", name);
22410549 509 return options->walk(tag->tagged, OBJ_ANY, data, options);
355885d5
MK
510}
511
22410549 512int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
355885d5
MK
513{
514 if (!obj)
515 return -1;
a2b22854
JK
516
517 if (obj->type == OBJ_NONE)
109cd76d 518 parse_object(the_repository, &obj->oid);
a2b22854 519
355885d5
MK
520 switch (obj->type) {
521 case OBJ_BLOB:
522 return 0;
523 case OBJ_TREE:
22410549 524 return fsck_walk_tree((struct tree *)obj, data, options);
355885d5 525 case OBJ_COMMIT:
22410549 526 return fsck_walk_commit((struct commit *)obj, data, options);
355885d5 527 case OBJ_TAG:
22410549 528 return fsck_walk_tag((struct tag *)obj, data, options);
355885d5 529 default:
90cf590f 530 error("Unknown object type for %s", describe_object(options, obj));
355885d5
MK
531 return -1;
532 }
533}
ba002f3b
MK
534
535/*
536 * The entries in a tree are ordered in the _path_ order,
537 * which means that a directory entry is ordered by adding
538 * a slash to the end of it.
539 *
540 * So a directory called "a" is ordered _after_ a file
541 * called "a.c", because "a/" sorts after "a.c".
542 */
543#define TREE_UNORDERED (-1)
544#define TREE_HAS_DUPS (-2)
545
546static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
547{
548 int len1 = strlen(name1);
549 int len2 = strlen(name2);
550 int len = len1 < len2 ? len1 : len2;
551 unsigned char c1, c2;
552 int cmp;
553
554 cmp = memcmp(name1, name2, len);
555 if (cmp < 0)
556 return 0;
557 if (cmp > 0)
558 return TREE_UNORDERED;
559
560 /*
561 * Ok, the first <len> characters are the same.
562 * Now we need to order the next one, but turn
563 * a '\0' into a '/' for a directory entry.
564 */
565 c1 = name1[len];
566 c2 = name2[len];
567 if (!c1 && !c2)
568 /*
569 * git-write-tree used to write out a nonsense tree that has
570 * entries with the same name, one blob and one tree. Make
571 * sure we do not have duplicate entries.
572 */
573 return TREE_HAS_DUPS;
574 if (!c1 && S_ISDIR(mode1))
575 c1 = '/';
576 if (!c2 && S_ISDIR(mode2))
577 c2 = '/';
578 return c1 < c2 ? 0 : TREE_UNORDERED;
579}
580
22410549 581static int fsck_tree(struct tree *item, struct fsck_options *options)
ba002f3b 582{
8354fa3d 583 int retval = 0;
c479d14a 584 int has_null_sha1 = 0;
ba002f3b
MK
585 int has_full_path = 0;
586 int has_empty_name = 0;
5d34a435
JK
587 int has_dot = 0;
588 int has_dotdot = 0;
5c17f512 589 int has_dotgit = 0;
ba002f3b
MK
590 int has_zero_pad = 0;
591 int has_bad_modes = 0;
592 int has_dup_entries = 0;
593 int not_properly_sorted = 0;
594 struct tree_desc desc;
595 unsigned o_mode;
596 const char *o_name;
ba002f3b 597
8354fa3d
DT
598 if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
599 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
600 return retval;
601 }
ba002f3b
MK
602
603 o_mode = 0;
604 o_name = NULL;
ba002f3b
MK
605
606 while (desc.size) {
5ec1e728 607 unsigned short mode;
ba002f3b 608 const char *name;
ce6663a9 609 const struct object_id *oid;
ba002f3b 610
ce6663a9 611 oid = tree_entry_extract(&desc, &name, &mode);
ba002f3b 612
ce6663a9 613 has_null_sha1 |= is_null_oid(oid);
effd12ec
HS
614 has_full_path |= !!strchr(name, '/');
615 has_empty_name |= !*name;
616 has_dot |= !strcmp(name, ".");
617 has_dotdot |= !strcmp(name, "..");
ed9c3220 618 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
ba002f3b 619 has_zero_pad |= *(char *)desc.buffer == '0';
159e7b08 620
b7b1fca1
JK
621 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
622 if (!S_ISLNK(mode))
623 oidset_insert(&gitmodules_found, oid);
624 else
625 retval += report(options, &item->object,
626 FSCK_MSG_GITMODULES_SYMLINK,
627 ".gitmodules is a symbolic link");
628 }
159e7b08 629
8354fa3d
DT
630 if (update_tree_entry_gently(&desc)) {
631 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
632 break;
633 }
ba002f3b
MK
634
635 switch (mode) {
636 /*
637 * Standard modes..
638 */
639 case S_IFREG | 0755:
640 case S_IFREG | 0644:
641 case S_IFLNK:
642 case S_IFDIR:
643 case S_IFGITLINK:
644 break;
645 /*
646 * This is nonstandard, but we had a few of these
647 * early on when we honored the full set of mode
648 * bits..
649 */
650 case S_IFREG | 0664:
22410549 651 if (!options->strict)
ba002f3b 652 break;
1cf01a34 653 /* fallthrough */
ba002f3b
MK
654 default:
655 has_bad_modes = 1;
656 }
657
658 if (o_name) {
659 switch (verify_ordered(o_mode, o_name, mode, name)) {
660 case TREE_UNORDERED:
661 not_properly_sorted = 1;
662 break;
663 case TREE_HAS_DUPS:
664 has_dup_entries = 1;
665 break;
666 default:
667 break;
668 }
669 }
670
671 o_mode = mode;
672 o_name = name;
ba002f3b
MK
673 }
674
c479d14a 675 if (has_null_sha1)
c99ba492 676 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
ba002f3b 677 if (has_full_path)
c99ba492 678 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
ba002f3b 679 if (has_empty_name)
c99ba492 680 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
5d34a435 681 if (has_dot)
c99ba492 682 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
5d34a435 683 if (has_dotdot)
c99ba492 684 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
5c17f512 685 if (has_dotgit)
c99ba492 686 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
ba002f3b 687 if (has_zero_pad)
c99ba492 688 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
ba002f3b 689 if (has_bad_modes)
c99ba492 690 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
ba002f3b 691 if (has_dup_entries)
c99ba492 692 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
ba002f3b 693 if (not_properly_sorted)
c99ba492 694 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
ba002f3b
MK
695 return retval;
696}
697
84d18c0b 698static int verify_headers(const void *data, unsigned long size,
b2f44feb 699 struct object *obj, struct fsck_options *options)
4d0d8975
JS
700{
701 const char *buffer = (const char *)data;
702 unsigned long i;
703
704 for (i = 0; i < size; i++) {
705 switch (buffer[i]) {
706 case '\0':
c99ba492
JS
707 return report(options, obj,
708 FSCK_MSG_NUL_IN_HEADER,
709 "unterminated header: NUL at offset %ld", i);
4d0d8975
JS
710 case '\n':
711 if (i + 1 < size && buffer[i + 1] == '\n')
712 return 0;
713 }
714 }
715
84d18c0b
JH
716 /*
717 * We did not find double-LF that separates the header
718 * and the body. Not having a body is not a crime but
719 * we do want to see the terminating LF for the last header
720 * line.
721 */
722 if (size && buffer[size - 1] == '\n')
723 return 0;
724
c99ba492
JS
725 return report(options, obj,
726 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
4d0d8975
JS
727}
728
22410549 729static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
daae1922 730{
e6826e33 731 const char *p = *ident;
d4b8de04
JK
732 char *end;
733
e6826e33
JS
734 *ident = strchrnul(*ident, '\n');
735 if (**ident == '\n')
736 (*ident)++;
737
738 if (*p == '<')
c99ba492 739 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
740 p += strcspn(p, "<>\n");
741 if (*p == '>')
c99ba492 742 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
e6826e33 743 if (*p != '<')
c99ba492 744 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
e6826e33 745 if (p[-1] != ' ')
c99ba492 746 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
e6826e33
JS
747 p++;
748 p += strcspn(p, "<>\n");
749 if (*p != '>')
c99ba492 750 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
e6826e33
JS
751 p++;
752 if (*p != ' ')
c99ba492 753 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
e6826e33
JS
754 p++;
755 if (*p == '0' && p[1] != ' ')
c99ba492 756 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
1aeb7e75 757 if (date_overflows(parse_timestamp(p, &end, 10)))
c99ba492 758 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
e6826e33 759 if ((end == p || *end != ' '))
c99ba492 760 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
e6826e33
JS
761 p = end + 1;
762 if ((*p != '+' && *p != '-') ||
763 !isdigit(p[1]) ||
764 !isdigit(p[2]) ||
765 !isdigit(p[3]) ||
766 !isdigit(p[4]) ||
767 (p[5] != '\n'))
c99ba492 768 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
e6826e33 769 p += 6;
daae1922
JN
770 return 0;
771}
772
bc6b8fc1 773static int fsck_commit_buffer(struct commit *commit, const char *buffer,
22410549 774 unsigned long size, struct fsck_options *options)
ba002f3b 775{
c54f5ca9 776 struct object_id tree_oid, oid;
ba002f3b 777 struct commit_graft *graft;
c9ad147f 778 unsigned parent_count, parent_line_count = 0, author_count;
daae1922 779 int err;
6d2d780f 780 const char *buffer_begin = buffer;
c54f5ca9 781 const char *p;
ba002f3b 782
b2f44feb 783 if (verify_headers(buffer, size, &commit->object, options))
4d0d8975
JS
784 return -1;
785
cf4fff57 786 if (!skip_prefix(buffer, "tree ", &buffer))
c99ba492 787 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
c54f5ca9 788 if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
b3584761
JS
789 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
790 if (err)
791 return err;
792 }
c54f5ca9 793 buffer = p + 1;
cf4fff57 794 while (skip_prefix(buffer, "parent ", &buffer)) {
c54f5ca9 795 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
b3584761
JS
796 err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
797 if (err)
798 return err;
799 }
c54f5ca9 800 buffer = p + 1;
9d02150c 801 parent_line_count++;
ba002f3b 802 }
1f93ecd1 803 graft = lookup_commit_graft(the_repository, &commit->object.oid);
9d02150c 804 parent_count = commit_list_count(commit->parents);
ba002f3b 805 if (graft) {
9d02150c 806 if (graft->nr_parent == -1 && !parent_count)
ba002f3b 807 ; /* shallow commit */
b3584761
JS
808 else if (graft->nr_parent != parent_count) {
809 err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
810 if (err)
811 return err;
812 }
ba002f3b 813 } else {
b3584761
JS
814 if (parent_count != parent_line_count) {
815 err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
816 if (err)
817 return err;
818 }
ba002f3b 819 }
c9ad147f
JS
820 author_count = 0;
821 while (skip_prefix(buffer, "author ", &buffer)) {
822 author_count++;
823 err = fsck_ident(&buffer, &commit->object, options);
824 if (err)
825 return err;
ba002f3b 826 }
c9ad147f
JS
827 if (author_count < 1)
828 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
829 else if (author_count > 1)
830 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
daae1922
JN
831 if (err)
832 return err;
cf4fff57 833 if (!skip_prefix(buffer, "committer ", &buffer))
c99ba492 834 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
22410549 835 err = fsck_ident(&buffer, &commit->object, options);
daae1922
JN
836 if (err)
837 return err;
2e27bd77 838 if (!get_commit_tree(commit)) {
c54f5ca9 839 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
5af29718
JH
840 if (err)
841 return err;
842 }
6d2d780f
JH
843 if (memchr(buffer_begin, '\0', size)) {
844 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
845 "NUL byte in the commit object body");
846 if (err)
847 return err;
848 }
ba002f3b
MK
849 return 0;
850}
851
90a398bb 852static int fsck_commit(struct commit *commit, const char *data,
22410549 853 unsigned long size, struct fsck_options *options)
bc6b8fc1 854{
90a398bb 855 const char *buffer = data ? data : get_commit_buffer(commit, &size);
22410549 856 int ret = fsck_commit_buffer(commit, buffer, size, options);
90a398bb
JS
857 if (!data)
858 unuse_commit_buffer(commit, buffer);
bc6b8fc1
JK
859 return ret;
860}
861
cec097be 862static int fsck_tag_buffer(struct tag *tag, const char *data,
22410549 863 unsigned long size, struct fsck_options *options)
cec097be 864{
c54f5ca9 865 struct object_id oid;
cec097be
JS
866 int ret = 0;
867 const char *buffer;
868 char *to_free = NULL, *eol;
869 struct strbuf sb = STRBUF_INIT;
c54f5ca9 870 const char *p;
cec097be
JS
871
872 if (data)
873 buffer = data;
874 else {
875 enum object_type type;
876
877 buffer = to_free =
b4f5aca4 878 read_object_file(&tag->object.oid, &type, &size);
cec097be 879 if (!buffer)
c99ba492
JS
880 return report(options, &tag->object,
881 FSCK_MSG_MISSING_TAG_OBJECT,
cec097be
JS
882 "cannot read tag object");
883
884 if (type != OBJ_TAG) {
c99ba492
JS
885 ret = report(options, &tag->object,
886 FSCK_MSG_TAG_OBJECT_NOT_TAG,
cec097be 887 "expected tag got %s",
debca9d2 888 type_name(type));
cec097be
JS
889 goto done;
890 }
891 }
892
8a272f29
RS
893 ret = verify_headers(buffer, size, &tag->object, options);
894 if (ret)
cec097be
JS
895 goto done;
896
897 if (!skip_prefix(buffer, "object ", &buffer)) {
c99ba492 898 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
cec097be
JS
899 goto done;
900 }
c54f5ca9 901 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
c99ba492 902 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
7d7d5b05
JS
903 if (ret)
904 goto done;
cec097be 905 }
c54f5ca9 906 buffer = p + 1;
cec097be
JS
907
908 if (!skip_prefix(buffer, "type ", &buffer)) {
c99ba492 909 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
cec097be
JS
910 goto done;
911 }
912 eol = strchr(buffer, '\n');
913 if (!eol) {
c99ba492 914 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
cec097be
JS
915 goto done;
916 }
917 if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
c99ba492 918 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
cec097be
JS
919 if (ret)
920 goto done;
921 buffer = eol + 1;
922
923 if (!skip_prefix(buffer, "tag ", &buffer)) {
c99ba492 924 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
cec097be
JS
925 goto done;
926 }
927 eol = strchr(buffer, '\n');
928 if (!eol) {
c99ba492 929 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
cec097be
JS
930 goto done;
931 }
932 strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
f27d05b1
JS
933 if (check_refname_format(sb.buf, 0)) {
934 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
c99ba492 935 "invalid 'tag' name: %.*s",
7add4419 936 (int)(eol - buffer), buffer);
f27d05b1
JS
937 if (ret)
938 goto done;
939 }
cec097be
JS
940 buffer = eol + 1;
941
f27d05b1 942 if (!skip_prefix(buffer, "tagger ", &buffer)) {
cec097be 943 /* early tags do not contain 'tagger' lines; warn only */
f27d05b1
JS
944 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
945 if (ret)
946 goto done;
947 }
cec097be 948 else
22410549 949 ret = fsck_ident(&buffer, &tag->object, options);
cec097be
JS
950
951done:
952 strbuf_release(&sb);
953 free(to_free);
954 return ret;
955}
956
90a398bb 957static int fsck_tag(struct tag *tag, const char *data,
22410549 958 unsigned long size, struct fsck_options *options)
ba002f3b
MK
959{
960 struct object *tagged = tag->tagged;
961
962 if (!tagged)
c99ba492 963 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
cec097be 964
22410549 965 return fsck_tag_buffer(tag, data, size, options);
ba002f3b
MK
966}
967
ed8b10f6
JK
968struct fsck_gitmodules_data {
969 struct object *obj;
970 struct fsck_options *options;
971 int ret;
972};
973
974static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
975{
976 struct fsck_gitmodules_data *data = vdata;
977 const char *subsection, *key;
978 int subsection_len;
979 char *name;
980
981 if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
982 !subsection)
983 return 0;
984
985 name = xmemdupz(subsection, subsection_len);
986 if (check_submodule_name(name) < 0)
987 data->ret |= report(data->options, data->obj,
988 FSCK_MSG_GITMODULES_NAME,
989 "disallowed submodule name: %s",
990 name);
a124133e
JK
991 if (!strcmp(key, "url") && value &&
992 looks_like_command_line_option(value))
993 data->ret |= report(data->options, data->obj,
994 FSCK_MSG_GITMODULES_URL,
995 "disallowed submodule url: %s",
996 value);
1a7fd1fb
JK
997 if (!strcmp(key, "path") && value &&
998 looks_like_command_line_option(value))
999 data->ret |= report(data->options, data->obj,
1000 FSCK_MSG_GITMODULES_PATH,
1001 "disallowed submodule path: %s",
1002 value);
ed8b10f6
JK
1003 free(name);
1004
1005 return 0;
1006}
1007
7ac4f3a0
JK
1008static int fsck_blob(struct blob *blob, const char *buf,
1009 unsigned long size, struct fsck_options *options)
1010{
ed8b10f6 1011 struct fsck_gitmodules_data data;
de6bd9e3 1012 struct config_options config_opts = { 0 };
ed8b10f6
JK
1013
1014 if (!oidset_contains(&gitmodules_found, &blob->object.oid))
1015 return 0;
1016 oidset_insert(&gitmodules_done, &blob->object.oid);
1017
fb162877
RJ
1018 if (object_on_skiplist(options, &blob->object))
1019 return 0;
1020
ed8b10f6
JK
1021 if (!buf) {
1022 /*
1023 * A missing buffer here is a sign that the caller found the
1024 * blob too gigantic to load into memory. Let's just consider
1025 * that an error.
1026 */
1027 return report(options, &blob->object,
0d68764d 1028 FSCK_MSG_GITMODULES_LARGE,
ed8b10f6
JK
1029 ".gitmodules too large to parse");
1030 }
1031
1032 data.obj = &blob->object;
1033 data.options = options;
1034 data.ret = 0;
de6bd9e3 1035 config_opts.error_action = CONFIG_ERROR_SILENT;
ed8b10f6 1036 if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
de6bd9e3 1037 ".gitmodules", buf, size, &data, &config_opts))
ed8b10f6
JK
1038 data.ret |= report(options, &blob->object,
1039 FSCK_MSG_GITMODULES_PARSE,
1040 "could not parse gitmodules blob");
1041
1042 return data.ret;
7ac4f3a0
JK
1043}
1044
90a398bb 1045int fsck_object(struct object *obj, void *data, unsigned long size,
22410549 1046 struct fsck_options *options)
ba002f3b
MK
1047{
1048 if (!obj)
c99ba492 1049 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
ba002f3b
MK
1050
1051 if (obj->type == OBJ_BLOB)
7ac4f3a0 1052 return fsck_blob((struct blob *)obj, data, size, options);
ba002f3b 1053 if (obj->type == OBJ_TREE)
22410549 1054 return fsck_tree((struct tree *) obj, options);
ba002f3b 1055 if (obj->type == OBJ_COMMIT)
90a398bb 1056 return fsck_commit((struct commit *) obj, (const char *) data,
22410549 1057 size, options);
ba002f3b 1058 if (obj->type == OBJ_TAG)
90a398bb 1059 return fsck_tag((struct tag *) obj, (const char *) data,
22410549 1060 size, options);
ba002f3b 1061
c99ba492 1062 return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
ba002f3b
MK
1063 obj->type);
1064}
d6ffc8d7 1065
1cd772cc
JS
1066int fsck_error_function(struct fsck_options *o,
1067 struct object *obj, int msg_type, const char *message)
d6ffc8d7 1068{
0282f4dc 1069 if (msg_type == FSCK_WARN) {
90cf590f 1070 warning("object %s: %s", describe_object(o, obj), message);
0282f4dc
JS
1071 return 0;
1072 }
90cf590f 1073 error("object %s: %s", describe_object(o, obj), message);
d6ffc8d7
MK
1074 return 1;
1075}
159e7b08
JK
1076
1077int fsck_finish(struct fsck_options *options)
1078{
1079 int ret = 0;
1080 struct oidset_iter iter;
1081 const struct object_id *oid;
1082
1083 oidset_iter_init(&gitmodules_found, &iter);
1084 while ((oid = oidset_iter_next(&iter))) {
1085 struct blob *blob;
1086 enum object_type type;
1087 unsigned long size;
1088 char *buf;
1089
1090 if (oidset_contains(&gitmodules_done, oid))
1091 continue;
1092
da14a7ff 1093 blob = lookup_blob(the_repository, oid);
159e7b08 1094 if (!blob) {
0ebbcf70 1095 struct object *obj = lookup_unknown_object(oid);
47cc9131 1096 ret |= report(options, obj,
159e7b08
JK
1097 FSCK_MSG_GITMODULES_BLOB,
1098 "non-blob found at .gitmodules");
1099 continue;
1100 }
1101
7913f53b 1102 buf = read_object_file(oid, &type, &size);
159e7b08 1103 if (!buf) {
27387444
JK
1104 if (is_promisor_object(&blob->object.oid))
1105 continue;
159e7b08
JK
1106 ret |= report(options, &blob->object,
1107 FSCK_MSG_GITMODULES_MISSING,
1108 "unable to read .gitmodules blob");
1109 continue;
1110 }
1111
1112 if (type == OBJ_BLOB)
1113 ret |= fsck_blob(blob, buf, size, options);
1114 else
1115 ret |= report(options, &blob->object,
1116 FSCK_MSG_GITMODULES_BLOB,
1117 "non-blob found at .gitmodules");
1118 free(buf);
1119 }
1120
1121
1122 oidset_clear(&gitmodules_found);
1123 oidset_clear(&gitmodules_done);
1124 return ret;
1125}