fsck.c

   1 #include "cache.h"
   2 #include "object-store.h"
   3 #include "repository.h"
   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"
  11 #include "refs.h"
  12 #include "utf8.h"
  13 #include "decorate.h"
  14 #include "oidset.h"
  15 #include "packfile.h"
  16 #include "submodule-config.h"
  17 #include "config.h"
  18 #include "help.h"
  19
  20 static struct oidset gitmodules_found = OIDSET_INIT;
  21 static struct oidset gitmodules_done = OIDSET_INIT;
  22
  23 #define FSCK_FATAL -1
  24 #define FSCK_INFO -2
  25
  26 #define FOREACH_MSG_ID(FUNC) \
  27         /* fatal errors */ \
  28         FUNC(NUL_IN_HEADER, FATAL) \
  29         FUNC(UNTERMINATED_HEADER, FATAL) \
  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) \
  56         FUNC(MISSING_TREE_OBJECT, ERROR) \
  57         FUNC(MISSING_TYPE, ERROR) \
  58         FUNC(MISSING_TYPE_ENTRY, ERROR) \
  59         FUNC(MULTIPLE_AUTHORS, ERROR) \
  60         FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  61         FUNC(TREE_NOT_SORTED, ERROR) \
  62         FUNC(UNKNOWN_TYPE, ERROR) \
  63         FUNC(ZERO_PADDED_DATE, ERROR) \
  64         FUNC(GITMODULES_MISSING, ERROR) \
  65         FUNC(GITMODULES_BLOB, ERROR) \
  66         FUNC(GITMODULES_LARGE, ERROR) \
  67         FUNC(GITMODULES_NAME, ERROR) \
  68         FUNC(GITMODULES_SYMLINK, ERROR) \
  69         FUNC(GITMODULES_URL, ERROR) \
  70         FUNC(GITMODULES_PATH, ERROR) \
  71         FUNC(GITMODULES_UPDATE, ERROR) \
  72         /* warnings */ \
  73         FUNC(BAD_FILEMODE, WARN) \
  74         FUNC(EMPTY_NAME, WARN) \
  75         FUNC(FULL_PATHNAME, WARN) \
  76         FUNC(HAS_DOT, WARN) \
  77         FUNC(HAS_DOTDOT, WARN) \
  78         FUNC(HAS_DOTGIT, WARN) \
  79         FUNC(NULL_SHA1, WARN) \
  80         FUNC(ZERO_PADDED_FILEMODE, WARN) \
  81         FUNC(NUL_IN_COMMIT, WARN) \
  82         /* infos (reported as warnings, but ignored by default) */ \
  83         FUNC(GITMODULES_PARSE, INFO) \
  84         FUNC(BAD_TAG_NAME, INFO) \
  85         FUNC(MISSING_TAGGER_ENTRY, INFO)
  86
  87 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
  88 enum fsck_msg_id {
  89         FOREACH_MSG_ID(MSG_ID)
  90         FSCK_MSG_MAX
  91 };
  92 #undef MSG_ID
  93
  94 #define STR(x) #x
  95 #define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
  96 static struct {
  97         const char *id_string;
  98         const char *downcased;
  99         const char *camelcased;
 100         int msg_type;
 101 } msg_id_info[FSCK_MSG_MAX + 1] = {
 102         FOREACH_MSG_ID(MSG_ID)
 103         { NULL, NULL, NULL, -1 }
 104 };
 105 #undef MSG_ID
 106
 107 static void prepare_msg_ids(void)
 108 {
 109         int i;
 110
 111         if (msg_id_info[0].downcased)
 112                 return;
 113
 114         /* convert id_string to lower case, without underscores. */
 115         for (i = 0; i < FSCK_MSG_MAX; i++) {
 116                 const char *p = msg_id_info[i].id_string;
 117                 int len = strlen(p);
 118                 char *q = xmalloc(len);
 119
 120                 msg_id_info[i].downcased = q;
 121                 while (*p)
 122                         if (*p == '_')
 123                                 p++;
 124                         else
 125                                 *(q)++ = tolower(*(p)++);
 126                 *q = '\0';
 127
 128                 p = msg_id_info[i].id_string;
 129                 q = xmalloc(len);
 130                 msg_id_info[i].camelcased = q;
 131                 while (*p) {
 132                         if (*p == '_') {
 133                                 p++;
 134                                 if (*p)
 135                                         *q++ = *p++;
 136                         } else {
 137                                 *q++ = tolower(*p++);
 138                         }
 139                 }
 140                 *q = '\0';
 141         }
 142 }
 143
 144 static int parse_msg_id(const char *text)
 145 {
 146         int i;
 147
 148         prepare_msg_ids();
 149
 150         for (i = 0; i < FSCK_MSG_MAX; i++)
 151                 if (!strcmp(text, msg_id_info[i].downcased))
 152                         return i;
 153
 154         return -1;
 155 }
 156
 157 void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
 158 {
 159         int i;
 160
 161         prepare_msg_ids();
 162
 163         for (i = 0; i < FSCK_MSG_MAX; i++)
 164                 list_config_item(list, prefix, msg_id_info[i].camelcased);
 165 }
 166
 167 static int fsck_msg_type(enum fsck_msg_id msg_id,
 168         struct fsck_options *options)
 169 {
 170         int msg_type;
 171
 172         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 173
 174         if (options->msg_type)
 175                 msg_type = options->msg_type[msg_id];
 176         else {
 177                 msg_type = msg_id_info[msg_id].msg_type;
 178                 if (options->strict && msg_type == FSCK_WARN)
 179                         msg_type = FSCK_ERROR;
 180         }
 181
 182         return msg_type;
 183 }
 184
 185 static void init_skiplist(struct fsck_options *options, const char *path)
 186 {
 187         FILE *fp;
 188         struct strbuf sb = STRBUF_INIT;
 189         struct object_id oid;
 190
 191         fp = fopen(path, "r");
 192         if (!fp)
 193                 die("Could not open skip list: %s", path);
 194         while (!strbuf_getline(&sb, fp)) {
 195                 const char *p;
 196                 const char *hash;
 197
 198                 /*
 199                  * Allow trailing comments, leading whitespace
 200                  * (including before commits), and empty or whitespace
 201                  * only lines.
 202                  */
 203                 hash = strchr(sb.buf, '#');
 204                 if (hash)
 205                         strbuf_setlen(&sb, hash - sb.buf);
 206                 strbuf_trim(&sb);
 207                 if (!sb.len)
 208                         continue;
 209
 210                 if (parse_oid_hex(sb.buf, &oid, &p) || *p != '\0')
 211                         die("Invalid SHA-1: %s", sb.buf);
 212                 oidset_insert(&options->skiplist, &oid);
 213         }
 214         if (ferror(fp))
 215                 die_errno("Could not read '%s'", path);
 216         fclose(fp);
 217         strbuf_release(&sb);
 218 }
 219
 220 static int parse_msg_type(const char *str)
 221 {
 222         if (!strcmp(str, "error"))
 223                 return FSCK_ERROR;
 224         else if (!strcmp(str, "warn"))
 225                 return FSCK_WARN;
 226         else if (!strcmp(str, "ignore"))
 227                 return FSCK_IGNORE;
 228         else
 229                 die("Unknown fsck message type: '%s'", str);
 230 }
 231
 232 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 233 {
 234         if (parse_msg_id(msg_id) < 0)
 235                 return 0;
 236         parse_msg_type(msg_type);
 237         return 1;
 238 }
 239
 240 void fsck_set_msg_type(struct fsck_options *options,
 241                 const char *msg_id, const char *msg_type)
 242 {
 243         int id = parse_msg_id(msg_id), type;
 244
 245         if (id < 0)
 246                 die("Unhandled message id: %s", msg_id);
 247         type = parse_msg_type(msg_type);
 248
 249         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 250                 die("Cannot demote %s to %s", msg_id, msg_type);
 251
 252         if (!options->msg_type) {
 253                 int i;
 254                 int *msg_type;
 255                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 256                 for (i = 0; i < FSCK_MSG_MAX; i++)
 257                         msg_type[i] = fsck_msg_type(i, options);
 258                 options->msg_type = msg_type;
 259         }
 260
 261         options->msg_type[id] = type;
 262 }
 263
 264 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 265 {
 266         char *buf = xstrdup(values), *to_free = buf;
 267         int done = 0;
 268
 269         while (!done) {
 270                 int len = strcspn(buf, " ,|"), equal;
 271
 272                 done = !buf[len];
 273                 if (!len) {
 274                         buf++;
 275                         continue;
 276                 }
 277                 buf[len] = '\0';
 278
 279                 for (equal = 0;
 280                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 281                      equal++)
 282                         buf[equal] = tolower(buf[equal]);
 283                 buf[equal] = '\0';
 284
 285                 if (!strcmp(buf, "skiplist")) {
 286                         if (equal == len)
 287                                 die("skiplist requires a path");
 288                         init_skiplist(options, buf + equal + 1);
 289                         buf += len + 1;
 290                         continue;
 291                 }
 292
 293                 if (equal == len)
 294                         die("Missing '=': '%s'", buf);
 295
 296                 fsck_set_msg_type(options, buf, buf + equal + 1);
 297                 buf += len + 1;
 298         }
 299         free(to_free);
 300 }
 301
 302 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 303 {
 304         for (;;) {
 305                 char c = *(msg_id)++;
 306
 307                 if (!c)
 308                         break;
 309                 if (c != '_')
 310                         strbuf_addch(sb, tolower(c));
 311                 else {
 312                         assert(*msg_id);
 313                         strbuf_addch(sb, *(msg_id)++);
 314                 }
 315         }
 316
 317         strbuf_addstr(sb, ": ");
 318 }
 319
 320 static int object_on_skiplist(struct fsck_options *opts, struct object *obj)
 321 {
 322         return opts && obj && oidset_contains(&opts->skiplist, &obj->oid);
 323 }
 324
 325 __attribute__((format (printf, 4, 5)))
 326 static int report(struct fsck_options *options, struct object *object,
 327         enum fsck_msg_id id, const char *fmt, ...)
 328 {
 329         va_list ap;
 330         struct strbuf sb = STRBUF_INIT;
 331         int msg_type = fsck_msg_type(id, options), result;
 332
 333         if (msg_type == FSCK_IGNORE)
 334                 return 0;
 335
 336         if (object_on_skiplist(options, object))
 337                 return 0;
 338
 339         if (msg_type == FSCK_FATAL)
 340                 msg_type = FSCK_ERROR;
 341         else if (msg_type == FSCK_INFO)
 342                 msg_type = FSCK_WARN;
 343
 344         append_msg_id(&sb, msg_id_info[id].id_string);
 345
 346         va_start(ap, fmt);
 347         strbuf_vaddf(&sb, fmt, ap);
 348         result = options->error_func(options, object, msg_type, sb.buf);
 349         strbuf_release(&sb);
 350         va_end(ap);
 351
 352         return result;
 353 }
 354
 355 static char *get_object_name(struct fsck_options *options, struct object *obj)
 356 {
 357         if (!options->object_names)
 358                 return NULL;
 359         return lookup_decoration(options->object_names, obj);
 360 }
 361
 362 static void put_object_name(struct fsck_options *options, struct object *obj,
 363         const char *fmt, ...)
 364 {
 365         va_list ap;
 366         struct strbuf buf = STRBUF_INIT;
 367         char *existing;
 368
 369         if (!options->object_names)
 370                 return;
 371         existing = lookup_decoration(options->object_names, obj);
 372         if (existing)
 373                 return;
 374         va_start(ap, fmt);
 375         strbuf_vaddf(&buf, fmt, ap);
 376         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 377         va_end(ap);
 378 }
 379
 380 static const char *describe_object(struct fsck_options *o, struct object *obj)
 381 {
 382         static struct strbuf buf = STRBUF_INIT;
 383         char *name;
 384
 385         strbuf_reset(&buf);
 386         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 387         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 388                 strbuf_addf(&buf, " (%s)", name);
 389
 390         return buf.buf;
 391 }
 392
 393 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 394 {
 395         struct tree_desc desc;
 396         struct name_entry entry;
 397         int res = 0;
 398         const char *name;
 399
 400         if (parse_tree(tree))
 401                 return -1;
 402
 403         name = get_object_name(options, &tree->object);
 404         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 405                 return -1;
 406         while (tree_entry_gently(&desc, &entry)) {
 407                 struct object *obj;
 408                 int result;
 409
 410                 if (S_ISGITLINK(entry.mode))
 411                         continue;
 412
 413                 if (S_ISDIR(entry.mode)) {
 414                         obj = (struct object *)lookup_tree(the_repository, entry.oid);
 415                         if (name && obj)
 416                                 put_object_name(options, obj, "%s%s/", name,
 417                                         entry.path);
 418                         result = options->walk(obj, OBJ_TREE, data, options);
 419                 }
 420                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 421                         obj = (struct object *)lookup_blob(the_repository, entry.oid);
 422                         if (name && obj)
 423                                 put_object_name(options, obj, "%s%s", name,
 424                                         entry.path);
 425                         result = options->walk(obj, OBJ_BLOB, data, options);
 426                 }
 427                 else {
 428                         result = error("in tree %s: entry %s has bad mode %.6o",
 429                                         describe_object(options, &tree->object), entry.path, entry.mode);
 430                 }
 431                 if (result < 0)
 432                         return result;
 433                 if (!res)
 434                         res = result;
 435         }
 436         return res;
 437 }
 438
 439 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 440 {
 441         int counter = 0, generation = 0, name_prefix_len = 0;
 442         struct commit_list *parents;
 443         int res;
 444         int result;
 445         const char *name;
 446
 447         if (parse_commit(commit))
 448                 return -1;
 449
 450         name = get_object_name(options, &commit->object);
 451         if (name)
 452                 put_object_name(options, &get_commit_tree(commit)->object,
 453                                 "%s:", name);
 454
 455         result = options->walk((struct object *)get_commit_tree(commit),
 456                                OBJ_TREE, data, options);
 457         if (result < 0)
 458                 return result;
 459         res = result;
 460
 461         parents = commit->parents;
 462         if (name && parents) {
 463                 int len = strlen(name), power;
 464
 465                 if (len && name[len - 1] == '^') {
 466                         generation = 1;
 467                         name_prefix_len = len - 1;
 468                 }
 469                 else { /* parse ~<generation> suffix */
 470                         for (generation = 0, power = 1;
 471                              len && isdigit(name[len - 1]);
 472                              power *= 10)
 473                                 generation += power * (name[--len] - '0');
 474                         if (power > 1 && len && name[len - 1] == '~')
 475                                 name_prefix_len = len - 1;
 476                 }
 477         }
 478
 479         while (parents) {
 480                 if (name) {
 481                         struct object *obj = &parents->item->object;
 482
 483                         if (counter++)
 484                                 put_object_name(options, obj, "%s^%d",
 485                                         name, counter);
 486                         else if (generation > 0)
 487                                 put_object_name(options, obj, "%.*s~%d",
 488                                         name_prefix_len, name, generation + 1);
 489                         else
 490                                 put_object_name(options, obj, "%s^", name);
 491                 }
 492                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 493                 if (result < 0)
 494                         return result;
 495                 if (!res)
 496                         res = result;
 497                 parents = parents->next;
 498         }
 499         return res;
 500 }
 501
 502 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 503 {
 504         char *name = get_object_name(options, &tag->object);
 505
 506         if (parse_tag(tag))
 507                 return -1;
 508         if (name)
 509                 put_object_name(options, tag->tagged, "%s", name);
 510         return options->walk(tag->tagged, OBJ_ANY, data, options);
 511 }
 512
 513 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 514 {
 515         if (!obj)
 516                 return -1;
 517
 518         if (obj->type == OBJ_NONE)
 519                 parse_object(the_repository, &obj->oid);
 520
 521         switch (obj->type) {
 522         case OBJ_BLOB:
 523                 return 0;
 524         case OBJ_TREE:
 525                 return fsck_walk_tree((struct tree *)obj, data, options);
 526         case OBJ_COMMIT:
 527                 return fsck_walk_commit((struct commit *)obj, data, options);
 528         case OBJ_TAG:
 529                 return fsck_walk_tag((struct tag *)obj, data, options);
 530         default:
 531                 error("Unknown object type for %s", describe_object(options, obj));
 532                 return -1;
 533         }
 534 }
 535
 536 /*
 537  * The entries in a tree are ordered in the _path_ order,
 538  * which means that a directory entry is ordered by adding
 539  * a slash to the end of it.
 540  *
 541  * So a directory called "a" is ordered _after_ a file
 542  * called "a.c", because "a/" sorts after "a.c".
 543  */
 544 #define TREE_UNORDERED (-1)
 545 #define TREE_HAS_DUPS  (-2)
 546
 547 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 548 {
 549         int len1 = strlen(name1);
 550         int len2 = strlen(name2);
 551         int len = len1 < len2 ? len1 : len2;
 552         unsigned char c1, c2;
 553         int cmp;
 554
 555         cmp = memcmp(name1, name2, len);
 556         if (cmp < 0)
 557                 return 0;
 558         if (cmp > 0)
 559                 return TREE_UNORDERED;
 560
 561         /*
 562          * Ok, the first <len> characters are the same.
 563          * Now we need to order the next one, but turn
 564          * a '\0' into a '/' for a directory entry.
 565          */
 566         c1 = name1[len];
 567         c2 = name2[len];
 568         if (!c1 && !c2)
 569                 /*
 570                  * git-write-tree used to write out a nonsense tree that has
 571                  * entries with the same name, one blob and one tree.  Make
 572                  * sure we do not have duplicate entries.
 573                  */
 574                 return TREE_HAS_DUPS;
 575         if (!c1 && S_ISDIR(mode1))
 576                 c1 = '/';
 577         if (!c2 && S_ISDIR(mode2))
 578                 c2 = '/';
 579         return c1 < c2 ? 0 : TREE_UNORDERED;
 580 }
 581
 582 static int fsck_tree(struct tree *item, struct fsck_options *options)
 583 {
 584         int retval = 0;
 585         int has_null_sha1 = 0;
 586         int has_full_path = 0;
 587         int has_empty_name = 0;
 588         int has_dot = 0;
 589         int has_dotdot = 0;
 590         int has_dotgit = 0;
 591         int has_zero_pad = 0;
 592         int has_bad_modes = 0;
 593         int has_dup_entries = 0;
 594         int not_properly_sorted = 0;
 595         struct tree_desc desc;
 596         unsigned o_mode;
 597         const char *o_name;
 598
 599         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 600                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 601                 return retval;
 602         }
 603
 604         o_mode = 0;
 605         o_name = NULL;
 606
 607         while (desc.size) {
 608                 unsigned mode;
 609                 const char *name, *backslash;
 610                 const struct object_id *oid;
 611
 612                 oid = tree_entry_extract(&desc, &name, &mode);
 613
 614                 has_null_sha1 |= is_null_oid(oid);
 615                 has_full_path |= !!strchr(name, '/');
 616                 has_empty_name |= !*name;
 617                 has_dot |= !strcmp(name, ".");
 618                 has_dotdot |= !strcmp(name, "..");
 619                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 620                 has_zero_pad |= *(char *)desc.buffer == '0';
 621
 622                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 623                         if (!S_ISLNK(mode))
 624                                 oidset_insert(&gitmodules_found, oid);
 625                         else
 626                                 retval += report(options, &item->object,
 627                                                  FSCK_MSG_GITMODULES_SYMLINK,
 628                                                  ".gitmodules is a symbolic link");
 629                 }
 630
 631                 if ((backslash = strchr(name, '\\'))) {
 632                         while (backslash) {
 633                                 backslash++;
 634                                 has_dotgit |= is_ntfs_dotgit(backslash);
 635                                 if (is_ntfs_dotgitmodules(backslash)) {
 636                                         if (!S_ISLNK(mode))
 637                                                 oidset_insert(&gitmodules_found, oid);
 638                                         else
 639                                                 retval += report(options, &item->object,
 640                                                                  FSCK_MSG_GITMODULES_SYMLINK,
 641                                                                  ".gitmodules is a symbolic link");
 642                                 }
 643                                 backslash = strchr(backslash, '\\');
 644                         }
 645                 }
 646
 647                 if (update_tree_entry_gently(&desc)) {
 648                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 649                         break;
 650                 }
 651
 652                 switch (mode) {
 653                 /*
 654                  * Standard modes..
 655                  */
 656                 case S_IFREG | 0755:
 657                 case S_IFREG | 0644:
 658                 case S_IFLNK:
 659                 case S_IFDIR:
 660                 case S_IFGITLINK:
 661                         break;
 662                 /*
 663                  * This is nonstandard, but we had a few of these
 664                  * early on when we honored the full set of mode
 665                  * bits..
 666                  */
 667                 case S_IFREG | 0664:
 668                         if (!options->strict)
 669                                 break;
 670                         /* fallthrough */
 671                 default:
 672                         has_bad_modes = 1;
 673                 }
 674
 675                 if (o_name) {
 676                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 677                         case TREE_UNORDERED:
 678                                 not_properly_sorted = 1;
 679                                 break;
 680                         case TREE_HAS_DUPS:
 681                                 has_dup_entries = 1;
 682                                 break;
 683                         default:
 684                                 break;
 685                         }
 686                 }
 687
 688                 o_mode = mode;
 689                 o_name = name;
 690         }
 691
 692         if (has_null_sha1)
 693                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 694         if (has_full_path)
 695                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 696         if (has_empty_name)
 697                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 698         if (has_dot)
 699                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 700         if (has_dotdot)
 701                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 702         if (has_dotgit)
 703                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 704         if (has_zero_pad)
 705                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 706         if (has_bad_modes)
 707                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 708         if (has_dup_entries)
 709                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 710         if (not_properly_sorted)
 711                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 712         return retval;
 713 }
 714
 715 static int verify_headers(const void *data, unsigned long size,
 716                           struct object *obj, struct fsck_options *options)
 717 {
 718         const char *buffer = (const char *)data;
 719         unsigned long i;
 720
 721         for (i = 0; i < size; i++) {
 722                 switch (buffer[i]) {
 723                 case '\0':
 724                         return report(options, obj,
 725                                 FSCK_MSG_NUL_IN_HEADER,
 726                                 "unterminated header: NUL at offset %ld", i);
 727                 case '\n':
 728                         if (i + 1 < size && buffer[i + 1] == '\n')
 729                                 return 0;
 730                 }
 731         }
 732
 733         /*
 734          * We did not find double-LF that separates the header
 735          * and the body.  Not having a body is not a crime but
 736          * we do want to see the terminating LF for the last header
 737          * line.
 738          */
 739         if (size && buffer[size - 1] == '\n')
 740                 return 0;
 741
 742         return report(options, obj,
 743                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 744 }
 745
 746 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 747 {
 748         const char *p = *ident;
 749         char *end;
 750
 751         *ident = strchrnul(*ident, '\n');
 752         if (**ident == '\n')
 753                 (*ident)++;
 754
 755         if (*p == '<')
 756                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 757         p += strcspn(p, "<>\n");
 758         if (*p == '>')
 759                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 760         if (*p != '<')
 761                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 762         if (p[-1] != ' ')
 763                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 764         p++;
 765         p += strcspn(p, "<>\n");
 766         if (*p != '>')
 767                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 768         p++;
 769         if (*p != ' ')
 770                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 771         p++;
 772         if (*p == '0' && p[1] != ' ')
 773                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 774         if (date_overflows(parse_timestamp(p, &end, 10)))
 775                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 776         if ((end == p || *end != ' '))
 777                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 778         p = end + 1;
 779         if ((*p != '+' && *p != '-') ||
 780             !isdigit(p[1]) ||
 781             !isdigit(p[2]) ||
 782             !isdigit(p[3]) ||
 783             !isdigit(p[4]) ||
 784             (p[5] != '\n'))
 785                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 786         p += 6;
 787         return 0;
 788 }
 789
 790 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 791         unsigned long size, struct fsck_options *options)
 792 {
 793         struct object_id tree_oid, oid;
 794         struct commit_graft *graft;
 795         unsigned parent_count, parent_line_count = 0, author_count;
 796         int err;
 797         const char *buffer_begin = buffer;
 798         const char *p;
 799
 800         if (verify_headers(buffer, size, &commit->object, options))
 801                 return -1;
 802
 803         if (!skip_prefix(buffer, "tree ", &buffer))
 804                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 805         if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 806                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 807                 if (err)
 808                         return err;
 809         }
 810         buffer = p + 1;
 811         while (skip_prefix(buffer, "parent ", &buffer)) {
 812                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 813                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 814                         if (err)
 815                                 return err;
 816                 }
 817                 buffer = p + 1;
 818                 parent_line_count++;
 819         }
 820         graft = lookup_commit_graft(the_repository, &commit->object.oid);
 821         parent_count = commit_list_count(commit->parents);
 822         if (graft) {
 823                 if (graft->nr_parent == -1 && !parent_count)
 824                         ; /* shallow commit */
 825                 else if (graft->nr_parent != parent_count) {
 826                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 827                         if (err)
 828                                 return err;
 829                 }
 830         } else {
 831                 if (parent_count != parent_line_count) {
 832                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 833                         if (err)
 834                                 return err;
 835                 }
 836         }
 837         author_count = 0;
 838         while (skip_prefix(buffer, "author ", &buffer)) {
 839                 author_count++;
 840                 err = fsck_ident(&buffer, &commit->object, options);
 841                 if (err)
 842                         return err;
 843         }
 844         if (author_count < 1)
 845                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 846         else if (author_count > 1)
 847                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 848         if (err)
 849                 return err;
 850         if (!skip_prefix(buffer, "committer ", &buffer))
 851                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 852         err = fsck_ident(&buffer, &commit->object, options);
 853         if (err)
 854                 return err;
 855         if (!get_commit_tree(commit)) {
 856                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
 857                 if (err)
 858                         return err;
 859         }
 860         if (memchr(buffer_begin, '\0', size)) {
 861                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 862                              "NUL byte in the commit object body");
 863                 if (err)
 864                         return err;
 865         }
 866         return 0;
 867 }
 868
 869 static int fsck_commit(struct commit *commit, const char *data,
 870         unsigned long size, struct fsck_options *options)
 871 {
 872         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 873         int ret = fsck_commit_buffer(commit, buffer, size, options);
 874         if (!data)
 875                 unuse_commit_buffer(commit, buffer);
 876         return ret;
 877 }
 878
 879 static int fsck_tag_buffer(struct tag *tag, const char *data,
 880         unsigned long size, struct fsck_options *options)
 881 {
 882         struct object_id oid;
 883         int ret = 0;
 884         const char *buffer;
 885         char *to_free = NULL, *eol;
 886         struct strbuf sb = STRBUF_INIT;
 887         const char *p;
 888
 889         if (data)
 890                 buffer = data;
 891         else {
 892                 enum object_type type;
 893
 894                 buffer = to_free =
 895                         read_object_file(&tag->object.oid, &type, &size);
 896                 if (!buffer)
 897                         return report(options, &tag->object,
 898                                 FSCK_MSG_MISSING_TAG_OBJECT,
 899                                 "cannot read tag object");
 900
 901                 if (type != OBJ_TAG) {
 902                         ret = report(options, &tag->object,
 903                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 904                                 "expected tag got %s",
 905                             type_name(type));
 906                         goto done;
 907                 }
 908         }
 909
 910         ret = verify_headers(buffer, size, &tag->object, options);
 911         if (ret)
 912                 goto done;
 913
 914         if (!skip_prefix(buffer, "object ", &buffer)) {
 915                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 916                 goto done;
 917         }
 918         if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 919                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 920                 if (ret)
 921                         goto done;
 922         }
 923         buffer = p + 1;
 924
 925         if (!skip_prefix(buffer, "type ", &buffer)) {
 926                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 927                 goto done;
 928         }
 929         eol = strchr(buffer, '\n');
 930         if (!eol) {
 931                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 932                 goto done;
 933         }
 934         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 935                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 936         if (ret)
 937                 goto done;
 938         buffer = eol + 1;
 939
 940         if (!skip_prefix(buffer, "tag ", &buffer)) {
 941                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 942                 goto done;
 943         }
 944         eol = strchr(buffer, '\n');
 945         if (!eol) {
 946                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 947                 goto done;
 948         }
 949         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 950         if (check_refname_format(sb.buf, 0)) {
 951                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 952                            "invalid 'tag' name: %.*s",
 953                            (int)(eol - buffer), buffer);
 954                 if (ret)
 955                         goto done;
 956         }
 957         buffer = eol + 1;
 958
 959         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 960                 /* early tags do not contain 'tagger' lines; warn only */
 961                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 962                 if (ret)
 963                         goto done;
 964         }
 965         else
 966                 ret = fsck_ident(&buffer, &tag->object, options);
 967
 968 done:
 969         strbuf_release(&sb);
 970         free(to_free);
 971         return ret;
 972 }
 973
 974 static int fsck_tag(struct tag *tag, const char *data,
 975         unsigned long size, struct fsck_options *options)
 976 {
 977         struct object *tagged = tag->tagged;
 978
 979         if (!tagged)
 980                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 981
 982         return fsck_tag_buffer(tag, data, size, options);
 983 }
 984
 985 struct fsck_gitmodules_data {
 986         struct object *obj;
 987         struct fsck_options *options;
 988         int ret;
 989 };
 990
 991 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 992 {
 993         struct fsck_gitmodules_data *data = vdata;
 994         const char *subsection, *key;
 995         int subsection_len;
 996         char *name;
 997
 998         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 999             !subsection)
1000                 return 0;
1001
1002         name = xmemdupz(subsection, subsection_len);
1003         if (check_submodule_name(name) < 0)
1004                 data->ret |= report(data->options, data->obj,
1005                                     FSCK_MSG_GITMODULES_NAME,
1006                                     "disallowed submodule name: %s",
1007                                     name);
1008         if (!strcmp(key, "url") && value &&
1009             looks_like_command_line_option(value))
1010                 data->ret |= report(data->options, data->obj,
1011                                     FSCK_MSG_GITMODULES_URL,
1012                                     "disallowed submodule url: %s",
1013                                     value);
1014         if (!strcmp(key, "path") && value &&
1015             looks_like_command_line_option(value))
1016                 data->ret |= report(data->options, data->obj,
1017                                     FSCK_MSG_GITMODULES_PATH,
1018                                     "disallowed submodule path: %s",
1019                                     value);
1020         if (!strcmp(key, "update") && value &&
1021             parse_submodule_update_type(value) == SM_UPDATE_COMMAND)
1022                 data->ret |= report(data->options, data->obj,
1023                                     FSCK_MSG_GITMODULES_UPDATE,
1024                                     "disallowed submodule update setting: %s",
1025                                     value);
1026         free(name);
1027
1028         return 0;
1029 }
1030
1031 static int fsck_blob(struct blob *blob, const char *buf,
1032                      unsigned long size, struct fsck_options *options)
1033 {
1034         struct fsck_gitmodules_data data;
1035         struct config_options config_opts = { 0 };
1036
1037         if (!oidset_contains(&gitmodules_found, &blob->object.oid))
1038                 return 0;
1039         oidset_insert(&gitmodules_done, &blob->object.oid);
1040
1041         if (object_on_skiplist(options, &blob->object))
1042                 return 0;
1043
1044         if (!buf) {
1045                 /*
1046                  * A missing buffer here is a sign that the caller found the
1047                  * blob too gigantic to load into memory. Let's just consider
1048                  * that an error.
1049                  */
1050                 return report(options, &blob->object,
1051                               FSCK_MSG_GITMODULES_LARGE,
1052                               ".gitmodules too large to parse");
1053         }
1054
1055         data.obj = &blob->object;
1056         data.options = options;
1057         data.ret = 0;
1058         config_opts.error_action = CONFIG_ERROR_SILENT;
1059         if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1060                                 ".gitmodules", buf, size, &data, &config_opts))
1061                 data.ret |= report(options, &blob->object,
1062                                    FSCK_MSG_GITMODULES_PARSE,
1063                                    "could not parse gitmodules blob");
1064
1065         return data.ret;
1066 }
1067
1068 int fsck_object(struct object *obj, void *data, unsigned long size,
1069         struct fsck_options *options)
1070 {
1071         if (!obj)
1072                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1073
1074         if (obj->type == OBJ_BLOB)
1075                 return fsck_blob((struct blob *)obj, data, size, options);
1076         if (obj->type == OBJ_TREE)
1077                 return fsck_tree((struct tree *) obj, options);
1078         if (obj->type == OBJ_COMMIT)
1079                 return fsck_commit((struct commit *) obj, (const char *) data,
1080                         size, options);
1081         if (obj->type == OBJ_TAG)
1082                 return fsck_tag((struct tag *) obj, (const char *) data,
1083                         size, options);
1084
1085         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1086                           obj->type);
1087 }
1088
1089 int fsck_error_function(struct fsck_options *o,
1090         struct object *obj, int msg_type, const char *message)
1091 {
1092         if (msg_type == FSCK_WARN) {
1093                 warning("object %s: %s", describe_object(o, obj), message);
1094                 return 0;
1095         }
1096         error("object %s: %s", describe_object(o, obj), message);
1097         return 1;
1098 }
1099
1100 int fsck_finish(struct fsck_options *options)
1101 {
1102         int ret = 0;
1103         struct oidset_iter iter;
1104         const struct object_id *oid;
1105
1106         oidset_iter_init(&gitmodules_found, &iter);
1107         while ((oid = oidset_iter_next(&iter))) {
1108                 struct blob *blob;
1109                 enum object_type type;
1110                 unsigned long size;
1111                 char *buf;
1112
1113                 if (oidset_contains(&gitmodules_done, oid))
1114                         continue;
1115
1116                 blob = lookup_blob(the_repository, oid);
1117                 if (!blob) {
1118                         struct object *obj = lookup_unknown_object(oid->hash);
1119                         ret |= report(options, obj,
1120                                       FSCK_MSG_GITMODULES_BLOB,
1121                                       "non-blob found at .gitmodules");
1122                         continue;
1123                 }
1124
1125                 buf = read_object_file(oid, &type, &size);
1126                 if (!buf) {
1127                         if (is_promisor_object(&blob->object.oid))
1128                                 continue;
1129                         ret |= report(options, &blob->object,
1130                                       FSCK_MSG_GITMODULES_MISSING,
1131                                       "unable to read .gitmodules blob");
1132                         continue;
1133                 }
1134
1135                 if (type == OBJ_BLOB)
1136                         ret |= fsck_blob(blob, buf, size, options);
1137                 else
1138                         ret |= report(options, &blob->object,
1139                                       FSCK_MSG_GITMODULES_BLOB,
1140                                       "non-blob found at .gitmodules");
1141                 free(buf);
1142         }
1143
1144
1145         oidset_clear(&gitmodules_found);
1146         oidset_clear(&gitmodules_done);
1147         return ret;
1148 }