12 #define FOREACH_MSG_ID(FUNC) \
14 FUNC(BAD_DATE, ERROR) \
15 FUNC(BAD_DATE_OVERFLOW, ERROR) \
16 FUNC(BAD_EMAIL, ERROR) \
17 FUNC(BAD_NAME, ERROR) \
18 FUNC(BAD_OBJECT_SHA1, ERROR) \
19 FUNC(BAD_PARENT_SHA1, ERROR) \
20 FUNC(BAD_TAG_OBJECT, ERROR) \
21 FUNC(BAD_TIMEZONE, ERROR) \
22 FUNC(BAD_TREE, ERROR) \
23 FUNC(BAD_TREE_SHA1, ERROR) \
24 FUNC(BAD_TYPE, ERROR) \
25 FUNC(DUPLICATE_ENTRIES, ERROR) \
26 FUNC(MISSING_AUTHOR, ERROR) \
27 FUNC(MISSING_COMMITTER, ERROR) \
28 FUNC(MISSING_EMAIL, ERROR) \
29 FUNC(MISSING_GRAFT, ERROR) \
30 FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
31 FUNC(MISSING_OBJECT, ERROR) \
32 FUNC(MISSING_PARENT, ERROR) \
33 FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
34 FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
35 FUNC(MISSING_TAG, ERROR) \
36 FUNC(MISSING_TAG_ENTRY, ERROR) \
37 FUNC(MISSING_TAG_OBJECT, ERROR) \
38 FUNC(MISSING_TREE, ERROR) \
39 FUNC(MISSING_TYPE, ERROR) \
40 FUNC(MISSING_TYPE_ENTRY, ERROR) \
41 FUNC(MULTIPLE_AUTHORS, ERROR) \
42 FUNC(NUL_IN_HEADER, ERROR) \
43 FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
44 FUNC(TREE_NOT_SORTED, ERROR) \
45 FUNC(UNKNOWN_TYPE, ERROR) \
46 FUNC(UNTERMINATED_HEADER, ERROR) \
47 FUNC(ZERO_PADDED_DATE, ERROR) \
49 FUNC(BAD_FILEMODE, WARN) \
50 FUNC(BAD_TAG_NAME, WARN) \
51 FUNC(EMPTY_NAME, WARN) \
52 FUNC(FULL_PATHNAME, WARN) \
54 FUNC(HAS_DOTDOT, WARN) \
55 FUNC(HAS_DOTGIT, WARN) \
56 FUNC(MISSING_TAGGER_ENTRY, WARN) \
57 FUNC(NULL_SHA1, WARN) \
58 FUNC(ZERO_PADDED_FILEMODE, WARN)
60 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
62 FOREACH_MSG_ID(MSG_ID)
68 #define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
70 const char *id_string;
71 const char *downcased;
73 } msg_id_info[FSCK_MSG_MAX + 1] = {
74 FOREACH_MSG_ID(MSG_ID)
79 static int parse_msg_id(const char *text)
83 if (!msg_id_info[0].downcased) {
84 /* convert id_string to lower case, without underscores. */
85 for (i = 0; i < FSCK_MSG_MAX; i++) {
86 const char *p = msg_id_info[i].id_string;
88 char *q = xmalloc(len);
90 msg_id_info[i].downcased = q;
95 *(q)++ = tolower(*(p)++);
100 for (i = 0; i < FSCK_MSG_MAX; i++)
101 if (!strcmp(text, msg_id_info[i].downcased))
107 static int fsck_msg_type(enum fsck_msg_id msg_id,
108 struct fsck_options *options)
112 assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
114 if (options->msg_type)
115 msg_type = options->msg_type[msg_id];
117 msg_type = msg_id_info[msg_id].msg_type;
118 if (options->strict && msg_type == FSCK_WARN)
119 msg_type = FSCK_ERROR;
125 static int parse_msg_type(const char *str)
127 if (!strcmp(str, "error"))
129 else if (!strcmp(str, "warn"))
132 die("Unknown fsck message type: '%s'", str);
135 int is_valid_msg_type(const char *msg_id, const char *msg_type)
137 if (parse_msg_id(msg_id) < 0)
139 parse_msg_type(msg_type);
143 void fsck_set_msg_type(struct fsck_options *options,
144 const char *msg_id, const char *msg_type)
146 int id = parse_msg_id(msg_id), type;
149 die("Unhandled message id: %s", msg_id);
150 type = parse_msg_type(msg_type);
152 if (!options->msg_type) {
154 int *msg_type = xmalloc(sizeof(int) * FSCK_MSG_MAX);
155 for (i = 0; i < FSCK_MSG_MAX; i++)
156 msg_type[i] = fsck_msg_type(i, options);
157 options->msg_type = msg_type;
160 options->msg_type[id] = type;
163 void fsck_set_msg_types(struct fsck_options *options, const char *values)
165 char *buf = xstrdup(values), *to_free = buf;
169 int len = strcspn(buf, " ,|"), equal;
179 equal < len && buf[equal] != '=' && buf[equal] != ':';
181 buf[equal] = tolower(buf[equal]);
185 die("Missing '=': '%s'", buf);
187 fsck_set_msg_type(options, buf, buf + equal + 1);
193 static void append_msg_id(struct strbuf *sb, const char *msg_id)
196 char c = *(msg_id)++;
201 strbuf_addch(sb, tolower(c));
204 strbuf_addch(sb, *(msg_id)++);
208 strbuf_addstr(sb, ": ");
211 __attribute__((format (printf, 4, 5)))
212 static int report(struct fsck_options *options, struct object *object,
213 enum fsck_msg_id id, const char *fmt, ...)
216 struct strbuf sb = STRBUF_INIT;
217 int msg_type = fsck_msg_type(id, options), result;
219 append_msg_id(&sb, msg_id_info[id].id_string);
222 strbuf_vaddf(&sb, fmt, ap);
223 result = options->error_func(object, msg_type, sb.buf);
230 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
232 struct tree_desc desc;
233 struct name_entry entry;
236 if (parse_tree(tree))
239 init_tree_desc(&desc, tree->buffer, tree->size);
240 while (tree_entry(&desc, &entry)) {
243 if (S_ISGITLINK(entry.mode))
245 if (S_ISDIR(entry.mode))
246 result = options->walk(&lookup_tree(entry.sha1)->object, OBJ_TREE, data, options);
247 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode))
248 result = options->walk(&lookup_blob(entry.sha1)->object, OBJ_BLOB, data, options);
250 result = error("in tree %s: entry %s has bad mode %.6o",
251 sha1_to_hex(tree->object.sha1), entry.path, entry.mode);
261 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
263 struct commit_list *parents;
267 if (parse_commit(commit))
270 result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options);
275 parents = commit->parents;
277 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
282 parents = parents->next;
287 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
291 return options->walk(tag->tagged, OBJ_ANY, data, options);
294 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
302 return fsck_walk_tree((struct tree *)obj, data, options);
304 return fsck_walk_commit((struct commit *)obj, data, options);
306 return fsck_walk_tag((struct tag *)obj, data, options);
308 error("Unknown object type for %s", sha1_to_hex(obj->sha1));
314 * The entries in a tree are ordered in the _path_ order,
315 * which means that a directory entry is ordered by adding
316 * a slash to the end of it.
318 * So a directory called "a" is ordered _after_ a file
319 * called "a.c", because "a/" sorts after "a.c".
321 #define TREE_UNORDERED (-1)
322 #define TREE_HAS_DUPS (-2)
324 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
326 int len1 = strlen(name1);
327 int len2 = strlen(name2);
328 int len = len1 < len2 ? len1 : len2;
329 unsigned char c1, c2;
332 cmp = memcmp(name1, name2, len);
336 return TREE_UNORDERED;
339 * Ok, the first <len> characters are the same.
340 * Now we need to order the next one, but turn
341 * a '\0' into a '/' for a directory entry.
347 * git-write-tree used to write out a nonsense tree that has
348 * entries with the same name, one blob and one tree. Make
349 * sure we do not have duplicate entries.
351 return TREE_HAS_DUPS;
352 if (!c1 && S_ISDIR(mode1))
354 if (!c2 && S_ISDIR(mode2))
356 return c1 < c2 ? 0 : TREE_UNORDERED;
359 static int fsck_tree(struct tree *item, struct fsck_options *options)
362 int has_null_sha1 = 0;
363 int has_full_path = 0;
364 int has_empty_name = 0;
368 int has_zero_pad = 0;
369 int has_bad_modes = 0;
370 int has_dup_entries = 0;
371 int not_properly_sorted = 0;
372 struct tree_desc desc;
376 init_tree_desc(&desc, item->buffer, item->size);
384 const unsigned char *sha1;
386 sha1 = tree_entry_extract(&desc, &name, &mode);
388 has_null_sha1 |= is_null_sha1(sha1);
389 has_full_path |= !!strchr(name, '/');
390 has_empty_name |= !*name;
391 has_dot |= !strcmp(name, ".");
392 has_dotdot |= !strcmp(name, "..");
393 has_dotgit |= (!strcmp(name, ".git") ||
394 is_hfs_dotgit(name) ||
395 is_ntfs_dotgit(name));
396 has_zero_pad |= *(char *)desc.buffer == '0';
397 update_tree_entry(&desc);
410 * This is nonstandard, but we had a few of these
411 * early on when we honored the full set of mode
415 if (!options->strict)
422 switch (verify_ordered(o_mode, o_name, mode, name)) {
424 not_properly_sorted = 1;
440 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
442 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
444 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
446 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
448 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
450 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
452 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
454 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
456 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
457 if (not_properly_sorted)
458 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
462 static int require_end_of_header(const void *data, unsigned long size,
463 struct object *obj, struct fsck_options *options)
465 const char *buffer = (const char *)data;
468 for (i = 0; i < size; i++) {
471 return report(options, obj,
472 FSCK_MSG_NUL_IN_HEADER,
473 "unterminated header: NUL at offset %ld", i);
475 if (i + 1 < size && buffer[i + 1] == '\n')
480 return report(options, obj,
481 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
484 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
486 const char *p = *ident;
489 *ident = strchrnul(*ident, '\n');
494 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
495 p += strcspn(p, "<>\n");
497 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
499 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
501 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
503 p += strcspn(p, "<>\n");
505 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
508 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
510 if (*p == '0' && p[1] != ' ')
511 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
512 if (date_overflows(strtoul(p, &end, 10)))
513 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
514 if ((end == p || *end != ' '))
515 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
517 if ((*p != '+' && *p != '-') ||
523 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
528 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
529 unsigned long size, struct fsck_options *options)
531 unsigned char tree_sha1[20], sha1[20];
532 struct commit_graft *graft;
533 unsigned parent_count, parent_line_count = 0, author_count;
536 if (require_end_of_header(buffer, size, &commit->object, options))
539 if (!skip_prefix(buffer, "tree ", &buffer))
540 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
541 if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
542 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
547 while (skip_prefix(buffer, "parent ", &buffer)) {
548 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
549 err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
556 graft = lookup_commit_graft(commit->object.sha1);
557 parent_count = commit_list_count(commit->parents);
559 if (graft->nr_parent == -1 && !parent_count)
560 ; /* shallow commit */
561 else if (graft->nr_parent != parent_count) {
562 err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
567 if (parent_count != parent_line_count) {
568 err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
574 while (skip_prefix(buffer, "author ", &buffer)) {
576 err = fsck_ident(&buffer, &commit->object, options);
580 if (author_count < 1)
581 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
582 else if (author_count > 1)
583 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
586 if (!skip_prefix(buffer, "committer ", &buffer))
587 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
588 err = fsck_ident(&buffer, &commit->object, options);
592 return report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
597 static int fsck_commit(struct commit *commit, const char *data,
598 unsigned long size, struct fsck_options *options)
600 const char *buffer = data ? data : get_commit_buffer(commit, &size);
601 int ret = fsck_commit_buffer(commit, buffer, size, options);
603 unuse_commit_buffer(commit, buffer);
607 static int fsck_tag_buffer(struct tag *tag, const char *data,
608 unsigned long size, struct fsck_options *options)
610 unsigned char sha1[20];
613 char *to_free = NULL, *eol;
614 struct strbuf sb = STRBUF_INIT;
619 enum object_type type;
622 read_sha1_file(tag->object.sha1, &type, &size);
624 return report(options, &tag->object,
625 FSCK_MSG_MISSING_TAG_OBJECT,
626 "cannot read tag object");
628 if (type != OBJ_TAG) {
629 ret = report(options, &tag->object,
630 FSCK_MSG_TAG_OBJECT_NOT_TAG,
631 "expected tag got %s",
637 if (require_end_of_header(buffer, size, &tag->object, options))
640 if (!skip_prefix(buffer, "object ", &buffer)) {
641 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
644 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
645 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
650 if (!skip_prefix(buffer, "type ", &buffer)) {
651 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
654 eol = strchr(buffer, '\n');
656 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
659 if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
660 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
665 if (!skip_prefix(buffer, "tag ", &buffer)) {
666 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
669 eol = strchr(buffer, '\n');
671 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
674 strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
675 if (check_refname_format(sb.buf, 0))
676 report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
677 "invalid 'tag' name: %.*s",
678 (int)(eol - buffer), buffer);
681 if (!skip_prefix(buffer, "tagger ", &buffer))
682 /* early tags do not contain 'tagger' lines; warn only */
683 report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
685 ret = fsck_ident(&buffer, &tag->object, options);
693 static int fsck_tag(struct tag *tag, const char *data,
694 unsigned long size, struct fsck_options *options)
696 struct object *tagged = tag->tagged;
699 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
701 return fsck_tag_buffer(tag, data, size, options);
704 int fsck_object(struct object *obj, void *data, unsigned long size,
705 struct fsck_options *options)
708 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
710 if (obj->type == OBJ_BLOB)
712 if (obj->type == OBJ_TREE)
713 return fsck_tree((struct tree *) obj, options);
714 if (obj->type == OBJ_COMMIT)
715 return fsck_commit((struct commit *) obj, (const char *) data,
717 if (obj->type == OBJ_TAG)
718 return fsck_tag((struct tag *) obj, (const char *) data,
721 return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
725 int fsck_error_function(struct object *obj, int msg_type, const char *message)
727 if (msg_type == FSCK_WARN) {
728 warning("object %s: %s", sha1_to_hex(obj->sha1), message);
731 error("object %s: %s", sha1_to_hex(obj->sha1), message);