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