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(NUL_IN_HEADER, ERROR) \
42 FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
43 FUNC(TREE_NOT_SORTED, ERROR) \
44 FUNC(UNKNOWN_TYPE, ERROR) \
45 FUNC(UNTERMINATED_HEADER, ERROR) \
46 FUNC(ZERO_PADDED_DATE, ERROR) \
48 FUNC(BAD_FILEMODE, WARN) \
49 FUNC(BAD_TAG_NAME, WARN) \
50 FUNC(EMPTY_NAME, WARN) \
51 FUNC(FULL_PATHNAME, WARN) \
53 FUNC(HAS_DOTDOT, WARN) \
54 FUNC(HAS_DOTGIT, WARN) \
55 FUNC(MISSING_TAGGER_ENTRY, WARN) \
56 FUNC(NULL_SHA1, WARN) \
57 FUNC(ZERO_PADDED_FILEMODE, WARN)
59 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
61 FOREACH_MSG_ID(MSG_ID)
67 #define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
69 const char *id_string;
70 const char *downcased;
72 } msg_id_info[FSCK_MSG_MAX + 1] = {
73 FOREACH_MSG_ID(MSG_ID)
78 static int parse_msg_id(const char *text)
82 if (!msg_id_info[0].downcased) {
83 /* convert id_string to lower case, without underscores. */
84 for (i = 0; i < FSCK_MSG_MAX; i++) {
85 const char *p = msg_id_info[i].id_string;
87 char *q = xmalloc(len);
89 msg_id_info[i].downcased = q;
94 *(q)++ = tolower(*(p)++);
99 for (i = 0; i < FSCK_MSG_MAX; i++)
100 if (!strcmp(text, msg_id_info[i].downcased))
106 static int fsck_msg_type(enum fsck_msg_id msg_id,
107 struct fsck_options *options)
111 assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
113 if (options->msg_type)
114 msg_type = options->msg_type[msg_id];
116 msg_type = msg_id_info[msg_id].msg_type;
117 if (options->strict && msg_type == FSCK_WARN)
118 msg_type = FSCK_ERROR;
124 static int parse_msg_type(const char *str)
126 if (!strcmp(str, "error"))
128 else if (!strcmp(str, "warn"))
131 die("Unknown fsck message type: '%s'", str);
134 int is_valid_msg_type(const char *msg_id, const char *msg_type)
136 if (parse_msg_id(msg_id) < 0)
138 parse_msg_type(msg_type);
142 void fsck_set_msg_type(struct fsck_options *options,
143 const char *msg_id, const char *msg_type)
145 int id = parse_msg_id(msg_id), type;
148 die("Unhandled message id: %s", msg_id);
149 type = parse_msg_type(msg_type);
151 if (!options->msg_type) {
153 int *msg_type = xmalloc(sizeof(int) * FSCK_MSG_MAX);
154 for (i = 0; i < FSCK_MSG_MAX; i++)
155 msg_type[i] = fsck_msg_type(i, options);
156 options->msg_type = msg_type;
159 options->msg_type[id] = type;
162 void fsck_set_msg_types(struct fsck_options *options, const char *values)
164 char *buf = xstrdup(values), *to_free = buf;
168 int len = strcspn(buf, " ,|"), equal;
178 equal < len && buf[equal] != '=' && buf[equal] != ':';
180 buf[equal] = tolower(buf[equal]);
184 die("Missing '=': '%s'", buf);
186 fsck_set_msg_type(options, buf, buf + equal + 1);
192 static void append_msg_id(struct strbuf *sb, const char *msg_id)
195 char c = *(msg_id)++;
200 strbuf_addch(sb, tolower(c));
203 strbuf_addch(sb, *(msg_id)++);
207 strbuf_addstr(sb, ": ");
210 __attribute__((format (printf, 4, 5)))
211 static int report(struct fsck_options *options, struct object *object,
212 enum fsck_msg_id id, const char *fmt, ...)
215 struct strbuf sb = STRBUF_INIT;
216 int msg_type = fsck_msg_type(id, options), result;
218 append_msg_id(&sb, msg_id_info[id].id_string);
221 strbuf_vaddf(&sb, fmt, ap);
222 result = options->error_func(object, msg_type, sb.buf);
229 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
231 struct tree_desc desc;
232 struct name_entry entry;
235 if (parse_tree(tree))
238 init_tree_desc(&desc, tree->buffer, tree->size);
239 while (tree_entry(&desc, &entry)) {
242 if (S_ISGITLINK(entry.mode))
244 if (S_ISDIR(entry.mode))
245 result = options->walk(&lookup_tree(entry.sha1)->object, OBJ_TREE, data, options);
246 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode))
247 result = options->walk(&lookup_blob(entry.sha1)->object, OBJ_BLOB, data, options);
249 result = error("in tree %s: entry %s has bad mode %.6o",
250 sha1_to_hex(tree->object.sha1), entry.path, entry.mode);
260 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
262 struct commit_list *parents;
266 if (parse_commit(commit))
269 result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options);
274 parents = commit->parents;
276 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
281 parents = parents->next;
286 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
290 return options->walk(tag->tagged, OBJ_ANY, data, options);
293 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
301 return fsck_walk_tree((struct tree *)obj, data, options);
303 return fsck_walk_commit((struct commit *)obj, data, options);
305 return fsck_walk_tag((struct tag *)obj, data, options);
307 error("Unknown object type for %s", sha1_to_hex(obj->sha1));
313 * The entries in a tree are ordered in the _path_ order,
314 * which means that a directory entry is ordered by adding
315 * a slash to the end of it.
317 * So a directory called "a" is ordered _after_ a file
318 * called "a.c", because "a/" sorts after "a.c".
320 #define TREE_UNORDERED (-1)
321 #define TREE_HAS_DUPS (-2)
323 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
325 int len1 = strlen(name1);
326 int len2 = strlen(name2);
327 int len = len1 < len2 ? len1 : len2;
328 unsigned char c1, c2;
331 cmp = memcmp(name1, name2, len);
335 return TREE_UNORDERED;
338 * Ok, the first <len> characters are the same.
339 * Now we need to order the next one, but turn
340 * a '\0' into a '/' for a directory entry.
346 * git-write-tree used to write out a nonsense tree that has
347 * entries with the same name, one blob and one tree. Make
348 * sure we do not have duplicate entries.
350 return TREE_HAS_DUPS;
351 if (!c1 && S_ISDIR(mode1))
353 if (!c2 && S_ISDIR(mode2))
355 return c1 < c2 ? 0 : TREE_UNORDERED;
358 static int fsck_tree(struct tree *item, struct fsck_options *options)
361 int has_null_sha1 = 0;
362 int has_full_path = 0;
363 int has_empty_name = 0;
367 int has_zero_pad = 0;
368 int has_bad_modes = 0;
369 int has_dup_entries = 0;
370 int not_properly_sorted = 0;
371 struct tree_desc desc;
375 init_tree_desc(&desc, item->buffer, item->size);
383 const unsigned char *sha1;
385 sha1 = tree_entry_extract(&desc, &name, &mode);
387 has_null_sha1 |= is_null_sha1(sha1);
388 has_full_path |= !!strchr(name, '/');
389 has_empty_name |= !*name;
390 has_dot |= !strcmp(name, ".");
391 has_dotdot |= !strcmp(name, "..");
392 has_dotgit |= (!strcmp(name, ".git") ||
393 is_hfs_dotgit(name) ||
394 is_ntfs_dotgit(name));
395 has_zero_pad |= *(char *)desc.buffer == '0';
396 update_tree_entry(&desc);
409 * This is nonstandard, but we had a few of these
410 * early on when we honored the full set of mode
414 if (!options->strict)
421 switch (verify_ordered(o_mode, o_name, mode, name)) {
423 not_properly_sorted = 1;
439 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
441 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
443 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
445 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
447 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
449 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
451 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
453 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
455 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
456 if (not_properly_sorted)
457 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
461 static int require_end_of_header(const void *data, unsigned long size,
462 struct object *obj, struct fsck_options *options)
464 const char *buffer = (const char *)data;
467 for (i = 0; i < size; i++) {
470 return report(options, obj,
471 FSCK_MSG_NUL_IN_HEADER,
472 "unterminated header: NUL at offset %ld", i);
474 if (i + 1 < size && buffer[i + 1] == '\n')
479 return report(options, obj,
480 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
483 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
485 const char *p = *ident;
488 *ident = strchrnul(*ident, '\n');
493 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
494 p += strcspn(p, "<>\n");
496 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
498 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
500 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
502 p += strcspn(p, "<>\n");
504 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
507 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
509 if (*p == '0' && p[1] != ' ')
510 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
511 if (date_overflows(strtoul(p, &end, 10)))
512 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
513 if ((end == p || *end != ' '))
514 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
516 if ((*p != '+' && *p != '-') ||
522 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
527 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
528 unsigned long size, struct fsck_options *options)
530 unsigned char tree_sha1[20], sha1[20];
531 struct commit_graft *graft;
532 unsigned parent_count, parent_line_count = 0;
535 if (require_end_of_header(buffer, size, &commit->object, options))
538 if (!skip_prefix(buffer, "tree ", &buffer))
539 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
540 if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n')
541 return report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
543 while (skip_prefix(buffer, "parent ", &buffer)) {
544 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n')
545 return report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
549 graft = lookup_commit_graft(commit->object.sha1);
550 parent_count = commit_list_count(commit->parents);
552 if (graft->nr_parent == -1 && !parent_count)
553 ; /* shallow commit */
554 else if (graft->nr_parent != parent_count)
555 return report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
557 if (parent_count != parent_line_count)
558 return report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
560 if (!skip_prefix(buffer, "author ", &buffer))
561 return report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
562 err = fsck_ident(&buffer, &commit->object, options);
565 if (!skip_prefix(buffer, "committer ", &buffer))
566 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
567 err = fsck_ident(&buffer, &commit->object, options);
571 return report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
576 static int fsck_commit(struct commit *commit, const char *data,
577 unsigned long size, struct fsck_options *options)
579 const char *buffer = data ? data : get_commit_buffer(commit, &size);
580 int ret = fsck_commit_buffer(commit, buffer, size, options);
582 unuse_commit_buffer(commit, buffer);
586 static int fsck_tag_buffer(struct tag *tag, const char *data,
587 unsigned long size, struct fsck_options *options)
589 unsigned char sha1[20];
592 char *to_free = NULL, *eol;
593 struct strbuf sb = STRBUF_INIT;
598 enum object_type type;
601 read_sha1_file(tag->object.sha1, &type, &size);
603 return report(options, &tag->object,
604 FSCK_MSG_MISSING_TAG_OBJECT,
605 "cannot read tag object");
607 if (type != OBJ_TAG) {
608 ret = report(options, &tag->object,
609 FSCK_MSG_TAG_OBJECT_NOT_TAG,
610 "expected tag got %s",
616 if (require_end_of_header(buffer, size, &tag->object, options))
619 if (!skip_prefix(buffer, "object ", &buffer)) {
620 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
623 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
624 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
629 if (!skip_prefix(buffer, "type ", &buffer)) {
630 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
633 eol = strchr(buffer, '\n');
635 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
638 if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
639 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
644 if (!skip_prefix(buffer, "tag ", &buffer)) {
645 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
648 eol = strchr(buffer, '\n');
650 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
653 strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
654 if (check_refname_format(sb.buf, 0))
655 report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
656 "invalid 'tag' name: %.*s",
657 (int)(eol - buffer), buffer);
660 if (!skip_prefix(buffer, "tagger ", &buffer))
661 /* early tags do not contain 'tagger' lines; warn only */
662 report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
664 ret = fsck_ident(&buffer, &tag->object, options);
672 static int fsck_tag(struct tag *tag, const char *data,
673 unsigned long size, struct fsck_options *options)
675 struct object *tagged = tag->tagged;
678 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
680 return fsck_tag_buffer(tag, data, size, options);
683 int fsck_object(struct object *obj, void *data, unsigned long size,
684 struct fsck_options *options)
687 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
689 if (obj->type == OBJ_BLOB)
691 if (obj->type == OBJ_TREE)
692 return fsck_tree((struct tree *) obj, options);
693 if (obj->type == OBJ_COMMIT)
694 return fsck_commit((struct commit *) obj, (const char *) data,
696 if (obj->type == OBJ_TAG)
697 return fsck_tag((struct tag *) obj, (const char *) data,
700 return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
704 int fsck_error_function(struct object *obj, int msg_type, const char *message)
706 if (msg_type == FSCK_WARN) {
707 warning("object %s: %s", sha1_to_hex(obj->sha1), message);
710 error("object %s: %s", sha1_to_hex(obj->sha1), message);