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