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, msg_type, sb.buf);
 309         strbuf_release(&sb);
 310         va_end(ap);
 311
 312         return result;
 313 }
 314
 315 static char *get_object_name(struct fsck_options *options, struct object *obj)
 316 {
 317         if (!options->object_names)
 318                 return NULL;
 319         return lookup_decoration(options->object_names, obj);
 320 }
 321
 322 static void put_object_name(struct fsck_options *options, struct object *obj,
 323         const char *fmt, ...)
 324 {
 325         va_list ap;
 326         struct strbuf buf = STRBUF_INIT;
 327         char *existing;
 328
 329         if (!options->object_names)
 330                 return;
 331         existing = lookup_decoration(options->object_names, obj);
 332         if (existing)
 333                 return;
 334         va_start(ap, fmt);
 335         strbuf_vaddf(&buf, fmt, ap);
 336         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 337         va_end(ap);
 338 }
 339
 340 static const char *describe_object(struct fsck_options *o, struct object *obj)
 341 {
 342         static struct strbuf buf = STRBUF_INIT;
 343         char *name;
 344
 345         strbuf_reset(&buf);
 346         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 347         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 348                 strbuf_addf(&buf, " (%s)", name);
 349
 350         return buf.buf;
 351 }
 352
 353 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 354 {
 355         struct tree_desc desc;
 356         struct name_entry entry;
 357         int res = 0;
 358         const char *name;
 359
 360         if (parse_tree(tree))
 361                 return -1;
 362
 363         name = get_object_name(options, &tree->object);
 364         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 365                 return -1;
 366         while (tree_entry_gently(&desc, &entry)) {
 367                 struct object *obj;
 368                 int result;
 369
 370                 if (S_ISGITLINK(entry.mode))
 371                         continue;
 372
 373                 if (S_ISDIR(entry.mode)) {
 374                         obj = (struct object *)lookup_tree(the_repository, &entry.oid);
 375                         if (name && obj)
 376                                 put_object_name(options, obj, "%s%s/", name,
 377                                         entry.path);
 378                         result = options->walk(obj, OBJ_TREE, data, options);
 379                 }
 380                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 381                         obj = (struct object *)lookup_blob(the_repository, &entry.oid);
 382                         if (name && obj)
 383                                 put_object_name(options, obj, "%s%s", name,
 384                                         entry.path);
 385                         result = options->walk(obj, OBJ_BLOB, data, options);
 386                 }
 387                 else {
 388                         result = error("in tree %s: entry %s has bad mode %.6o",
 389                                         describe_object(options, &tree->object), entry.path, entry.mode);
 390                 }
 391                 if (result < 0)
 392                         return result;
 393                 if (!res)
 394                         res = result;
 395         }
 396         return res;
 397 }
 398
 399 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 400 {
 401         int counter = 0, generation = 0, name_prefix_len = 0;
 402         struct commit_list *parents;
 403         int res;
 404         int result;
 405         const char *name;
 406
 407         if (parse_commit(commit))
 408                 return -1;
 409
 410         name = get_object_name(options, &commit->object);
 411         if (name)
 412                 put_object_name(options, &get_commit_tree(commit)->object,
 413                                 "%s:", name);
 414
 415         result = options->walk((struct object *)get_commit_tree(commit),
 416                                OBJ_TREE, data, options);
 417         if (result < 0)
 418                 return result;
 419         res = result;
 420
 421         parents = commit->parents;
 422         if (name && parents) {
 423                 int len = strlen(name), power;
 424
 425                 if (len && name[len - 1] == '^') {
 426                         generation = 1;
 427                         name_prefix_len = len - 1;
 428                 }
 429                 else { /* parse ~<generation> suffix */
 430                         for (generation = 0, power = 1;
 431                              len && isdigit(name[len - 1]);
 432                              power *= 10)
 433                                 generation += power * (name[--len] - '0');
 434                         if (power > 1 && len && name[len - 1] == '~')
 435                                 name_prefix_len = len - 1;
 436                 }
 437         }
 438
 439         while (parents) {
 440                 if (name) {
 441                         struct object *obj = &parents->item->object;
 442
 443                         if (counter++)
 444                                 put_object_name(options, obj, "%s^%d",
 445                                         name, counter);
 446                         else if (generation > 0)
 447                                 put_object_name(options, obj, "%.*s~%d",
 448                                         name_prefix_len, name, generation + 1);
 449                         else
 450                                 put_object_name(options, obj, "%s^", name);
 451                 }
 452                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 453                 if (result < 0)
 454                         return result;
 455                 if (!res)
 456                         res = result;
 457                 parents = parents->next;
 458         }
 459         return res;
 460 }
 461
 462 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 463 {
 464         char *name = get_object_name(options, &tag->object);
 465
 466         if (parse_tag(tag))
 467                 return -1;
 468         if (name)
 469                 put_object_name(options, tag->tagged, "%s", name);
 470         return options->walk(tag->tagged, OBJ_ANY, data, options);
 471 }
 472
 473 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 474 {
 475         if (!obj)
 476                 return -1;
 477
 478         if (obj->type == OBJ_NONE)
 479                 parse_object(the_repository, &obj->oid);
 480
 481         switch (obj->type) {
 482         case OBJ_BLOB:
 483                 return 0;
 484         case OBJ_TREE:
 485                 return fsck_walk_tree((struct tree *)obj, data, options);
 486         case OBJ_COMMIT:
 487                 return fsck_walk_commit((struct commit *)obj, data, options);
 488         case OBJ_TAG:
 489                 return fsck_walk_tag((struct tag *)obj, data, options);
 490         default:
 491                 error("Unknown object type for %s", describe_object(options, obj));
 492                 return -1;
 493         }
 494 }
 495
 496 /*
 497  * The entries in a tree are ordered in the _path_ order,
 498  * which means that a directory entry is ordered by adding
 499  * a slash to the end of it.
 500  *
 501  * So a directory called "a" is ordered _after_ a file
 502  * called "a.c", because "a/" sorts after "a.c".
 503  */
 504 #define TREE_UNORDERED (-1)
 505 #define TREE_HAS_DUPS  (-2)
 506
 507 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 508 {
 509         int len1 = strlen(name1);
 510         int len2 = strlen(name2);
 511         int len = len1 < len2 ? len1 : len2;
 512         unsigned char c1, c2;
 513         int cmp;
 514
 515         cmp = memcmp(name1, name2, len);
 516         if (cmp < 0)
 517                 return 0;
 518         if (cmp > 0)
 519                 return TREE_UNORDERED;
 520
 521         /*
 522          * Ok, the first <len> characters are the same.
 523          * Now we need to order the next one, but turn
 524          * a '\0' into a '/' for a directory entry.
 525          */
 526         c1 = name1[len];
 527         c2 = name2[len];
 528         if (!c1 && !c2)
 529                 /*
 530                  * git-write-tree used to write out a nonsense tree that has
 531                  * entries with the same name, one blob and one tree.  Make
 532                  * sure we do not have duplicate entries.
 533                  */
 534                 return TREE_HAS_DUPS;
 535         if (!c1 && S_ISDIR(mode1))
 536                 c1 = '/';
 537         if (!c2 && S_ISDIR(mode2))
 538                 c2 = '/';
 539         return c1 < c2 ? 0 : TREE_UNORDERED;
 540 }
 541
 542 static int fsck_tree(struct tree *item,
 543                      const char *buffer, unsigned long size,
 544                      struct fsck_options *options)
 545 {
 546         int retval = 0;
 547         int has_null_sha1 = 0;
 548         int has_full_path = 0;
 549         int has_empty_name = 0;
 550         int has_dot = 0;
 551         int has_dotdot = 0;
 552         int has_dotgit = 0;
 553         int has_zero_pad = 0;
 554         int has_bad_modes = 0;
 555         int has_dup_entries = 0;
 556         int not_properly_sorted = 0;
 557         struct tree_desc desc;
 558         unsigned o_mode;
 559         const char *o_name;
 560
 561         if (init_tree_desc_gently(&desc, buffer, size)) {
 562                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 563                 return retval;
 564         }
 565
 566         o_mode = 0;
 567         o_name = NULL;
 568
 569         while (desc.size) {
 570                 unsigned short mode;
 571                 const char *name;
 572                 const struct object_id *oid;
 573
 574                 oid = tree_entry_extract(&desc, &name, &mode);
 575
 576                 has_null_sha1 |= is_null_oid(oid);
 577                 has_full_path |= !!strchr(name, '/');
 578                 has_empty_name |= !*name;
 579                 has_dot |= !strcmp(name, ".");
 580                 has_dotdot |= !strcmp(name, "..");
 581                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 582                 has_zero_pad |= *(char *)desc.buffer == '0';
 583
 584                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 585                         if (!S_ISLNK(mode))
 586                                 oidset_insert(&gitmodules_found, oid);
 587                         else
 588                                 retval += report(options, &item->object,
 589                                                  FSCK_MSG_GITMODULES_SYMLINK,
 590                                                  ".gitmodules is a symbolic link");
 591                 }
 592
 593                 if (update_tree_entry_gently(&desc)) {
 594                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 595                         break;
 596                 }
 597
 598                 switch (mode) {
 599                 /*
 600                  * Standard modes..
 601                  */
 602                 case S_IFREG | 0755:
 603                 case S_IFREG | 0644:
 604                 case S_IFLNK:
 605                 case S_IFDIR:
 606                 case S_IFGITLINK:
 607                         break;
 608                 /*
 609                  * This is nonstandard, but we had a few of these
 610                  * early on when we honored the full set of mode
 611                  * bits..
 612                  */
 613                 case S_IFREG | 0664:
 614                         if (!options->strict)
 615                                 break;
 616                         /* fallthrough */
 617                 default:
 618                         has_bad_modes = 1;
 619                 }
 620
 621                 if (o_name) {
 622                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 623                         case TREE_UNORDERED:
 624                                 not_properly_sorted = 1;
 625                                 break;
 626                         case TREE_HAS_DUPS:
 627                                 has_dup_entries = 1;
 628                                 break;
 629                         default:
 630                                 break;
 631                         }
 632                 }
 633
 634                 o_mode = mode;
 635                 o_name = name;
 636         }
 637
 638         if (has_null_sha1)
 639                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 640         if (has_full_path)
 641                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 642         if (has_empty_name)
 643                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 644         if (has_dot)
 645                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 646         if (has_dotdot)
 647                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 648         if (has_dotgit)
 649                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 650         if (has_zero_pad)
 651                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 652         if (has_bad_modes)
 653                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 654         if (has_dup_entries)
 655                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 656         if (not_properly_sorted)
 657                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 658         return retval;
 659 }
 660
 661 static int verify_headers(const void *data, unsigned long size,
 662                           struct object *obj, struct fsck_options *options)
 663 {
 664         const char *buffer = (const char *)data;
 665         unsigned long i;
 666
 667         for (i = 0; i < size; i++) {
 668                 switch (buffer[i]) {
 669                 case '\0':
 670                         return report(options, obj,
 671                                 FSCK_MSG_NUL_IN_HEADER,
 672                                 "unterminated header: NUL at offset %ld", i);
 673                 case '\n':
 674                         if (i + 1 < size && buffer[i + 1] == '\n')
 675                                 return 0;
 676                 }
 677         }
 678
 679         /*
 680          * We did not find double-LF that separates the header
 681          * and the body.  Not having a body is not a crime but
 682          * we do want to see the terminating LF for the last header
 683          * line.
 684          */
 685         if (size && buffer[size - 1] == '\n')
 686                 return 0;
 687
 688         return report(options, obj,
 689                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 690 }
 691
 692 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 693 {
 694         const char *p = *ident;
 695         char *end;
 696
 697         *ident = strchrnul(*ident, '\n');
 698         if (**ident == '\n')
 699                 (*ident)++;
 700
 701         if (*p == '<')
 702                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 703         p += strcspn(p, "<>\n");
 704         if (*p == '>')
 705                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 706         if (*p != '<')
 707                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 708         if (p[-1] != ' ')
 709                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 710         p++;
 711         p += strcspn(p, "<>\n");
 712         if (*p != '>')
 713                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 714         p++;
 715         if (*p != ' ')
 716                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 717         p++;
 718         if (*p == '0' && p[1] != ' ')
 719                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 720         if (date_overflows(parse_timestamp(p, &end, 10)))
 721                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 722         if ((end == p || *end != ' '))
 723                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 724         p = end + 1;
 725         if ((*p != '+' && *p != '-') ||
 726             !isdigit(p[1]) ||
 727             !isdigit(p[2]) ||
 728             !isdigit(p[3]) ||
 729             !isdigit(p[4]) ||
 730             (p[5] != '\n'))
 731                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 732         p += 6;
 733         return 0;
 734 }
 735
 736 static int fsck_commit(struct commit *commit, const char *buffer,
 737                        unsigned long size, struct fsck_options *options)
 738 {
 739         struct object_id tree_oid, oid;
 740         unsigned author_count;
 741         int err;
 742         const char *buffer_begin = buffer;
 743         const char *p;
 744
 745         if (verify_headers(buffer, size, &commit->object, options))
 746                 return -1;
 747
 748         if (!skip_prefix(buffer, "tree ", &buffer))
 749                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 750         if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 751                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 752                 if (err)
 753                         return err;
 754         }
 755         buffer = p + 1;
 756         while (skip_prefix(buffer, "parent ", &buffer)) {
 757                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 758                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 759                         if (err)
 760                                 return err;
 761                 }
 762                 buffer = p + 1;
 763         }
 764         author_count = 0;
 765         while (skip_prefix(buffer, "author ", &buffer)) {
 766                 author_count++;
 767                 err = fsck_ident(&buffer, &commit->object, options);
 768                 if (err)
 769                         return err;
 770         }
 771         if (author_count < 1)
 772                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 773         else if (author_count > 1)
 774                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 775         if (err)
 776                 return err;
 777         if (!skip_prefix(buffer, "committer ", &buffer))
 778                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 779         err = fsck_ident(&buffer, &commit->object, options);
 780         if (err)
 781                 return err;
 782         if (memchr(buffer_begin, '\0', size)) {
 783                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 784                              "NUL byte in the commit object body");
 785                 if (err)
 786                         return err;
 787         }
 788         return 0;
 789 }
 790
 791 static int fsck_tag(struct tag *tag, const char *buffer,
 792                     unsigned long size, struct fsck_options *options)
 793 {
 794         struct object_id oid;
 795         int ret = 0;
 796         char *eol;
 797         struct strbuf sb = STRBUF_INIT;
 798         const char *p;
 799
 800         ret = verify_headers(buffer, size, &tag->object, options);
 801         if (ret)
 802                 goto done;
 803
 804         if (!skip_prefix(buffer, "object ", &buffer)) {
 805                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 806                 goto done;
 807         }
 808         if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 809                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 810                 if (ret)
 811                         goto done;
 812         }
 813         buffer = p + 1;
 814
 815         if (!skip_prefix(buffer, "type ", &buffer)) {
 816                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 817                 goto done;
 818         }
 819         eol = strchr(buffer, '\n');
 820         if (!eol) {
 821                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 822                 goto done;
 823         }
 824         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 825                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 826         if (ret)
 827                 goto done;
 828         buffer = eol + 1;
 829
 830         if (!skip_prefix(buffer, "tag ", &buffer)) {
 831                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 832                 goto done;
 833         }
 834         eol = strchr(buffer, '\n');
 835         if (!eol) {
 836                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 837                 goto done;
 838         }
 839         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 840         if (check_refname_format(sb.buf, 0)) {
 841                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 842                            "invalid 'tag' name: %.*s",
 843                            (int)(eol - buffer), buffer);
 844                 if (ret)
 845                         goto done;
 846         }
 847         buffer = eol + 1;
 848
 849         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 850                 /* early tags do not contain 'tagger' lines; warn only */
 851                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 852                 if (ret)
 853                         goto done;
 854         }
 855         else
 856                 ret = fsck_ident(&buffer, &tag->object, options);
 857
 858 done:
 859         strbuf_release(&sb);
 860         return ret;
 861 }
 862
 863 struct fsck_gitmodules_data {
 864         struct object *obj;
 865         struct fsck_options *options;
 866         int ret;
 867 };
 868
 869 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 870 {
 871         struct fsck_gitmodules_data *data = vdata;
 872         const char *subsection, *key;
 873         int subsection_len;
 874         char *name;
 875
 876         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 877             !subsection)
 878                 return 0;
 879
 880         name = xmemdupz(subsection, subsection_len);
 881         if (check_submodule_name(name) < 0)
 882                 data->ret |= report(data->options, data->obj,
 883                                     FSCK_MSG_GITMODULES_NAME,
 884                                     "disallowed submodule name: %s",
 885                                     name);
 886         if (!strcmp(key, "url") && value &&
 887             looks_like_command_line_option(value))
 888                 data->ret |= report(data->options, data->obj,
 889                                     FSCK_MSG_GITMODULES_URL,
 890                                     "disallowed submodule url: %s",
 891                                     value);
 892         if (!strcmp(key, "path") && value &&
 893             looks_like_command_line_option(value))
 894                 data->ret |= report(data->options, data->obj,
 895                                     FSCK_MSG_GITMODULES_PATH,
 896                                     "disallowed submodule path: %s",
 897                                     value);
 898         free(name);
 899
 900         return 0;
 901 }
 902
 903 static int fsck_blob(struct blob *blob, const char *buf,
 904                      unsigned long size, struct fsck_options *options)
 905 {
 906         struct fsck_gitmodules_data data;
 907         struct config_options config_opts = { 0 };
 908
 909         if (!oidset_contains(&gitmodules_found, &blob->object.oid))
 910                 return 0;
 911         oidset_insert(&gitmodules_done, &blob->object.oid);
 912
 913         if (object_on_skiplist(options, &blob->object))
 914                 return 0;
 915
 916         if (!buf) {
 917                 /*
 918                  * A missing buffer here is a sign that the caller found the
 919                  * blob too gigantic to load into memory. Let's just consider
 920                  * that an error.
 921                  */
 922                 return report(options, &blob->object,
 923                               FSCK_MSG_GITMODULES_LARGE,
 924                               ".gitmodules too large to parse");
 925         }
 926
 927         data.obj = &blob->object;
 928         data.options = options;
 929         data.ret = 0;
 930         config_opts.error_action = CONFIG_ERROR_SILENT;
 931         if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
 932                                 ".gitmodules", buf, size, &data, &config_opts))
 933                 data.ret |= report(options, &blob->object,
 934                                    FSCK_MSG_GITMODULES_PARSE,
 935                                    "could not parse gitmodules blob");
 936
 937         return data.ret;
 938 }
 939
 940 int fsck_object(struct object *obj, void *data, unsigned long size,
 941         struct fsck_options *options)
 942 {
 943         if (!obj)
 944                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
 945
 946         if (obj->type == OBJ_BLOB)
 947                 return fsck_blob((struct blob *)obj, data, size, options);
 948         if (obj->type == OBJ_TREE)
 949                 return fsck_tree((struct tree *) obj, data, size, options);
 950         if (obj->type == OBJ_COMMIT)
 951                 return fsck_commit((struct commit *) obj, (const char *) data,
 952                         size, options);
 953         if (obj->type == OBJ_TAG)
 954                 return fsck_tag((struct tag *) obj, (const char *) data,
 955                         size, options);
 956
 957         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
 958                           obj->type);
 959 }
 960
 961 int fsck_error_function(struct fsck_options *o,
 962         struct object *obj, int msg_type, const char *message)
 963 {
 964         if (msg_type == FSCK_WARN) {
 965                 warning("object %s: %s", describe_object(o, obj), message);
 966                 return 0;
 967         }
 968         error("object %s: %s", describe_object(o, obj), message);
 969         return 1;
 970 }
 971
 972 int fsck_finish(struct fsck_options *options)
 973 {
 974         int ret = 0;
 975         struct oidset_iter iter;
 976         const struct object_id *oid;
 977
 978         oidset_iter_init(&gitmodules_found, &iter);
 979         while ((oid = oidset_iter_next(&iter))) {
 980                 struct blob *blob;
 981                 enum object_type type;
 982                 unsigned long size;
 983                 char *buf;
 984
 985                 if (oidset_contains(&gitmodules_done, oid))
 986                         continue;
 987
 988                 blob = lookup_blob(the_repository, oid);
 989                 if (!blob) {
 990                         struct object *obj = lookup_unknown_object(oid);
 991                         ret |= report(options, obj,
 992                                       FSCK_MSG_GITMODULES_BLOB,
 993                                       "non-blob found at .gitmodules");
 994                         continue;
 995                 }
 996
 997                 buf = read_object_file(oid, &type, &size);
 998                 if (!buf) {
 999                         if (is_promisor_object(&blob->object.oid))
1000                                 continue;
1001                         ret |= report(options, &blob->object,
1002                                       FSCK_MSG_GITMODULES_MISSING,
1003                                       "unable to read .gitmodules blob");
1004                         continue;
1005                 }
1006
1007                 if (type == OBJ_BLOB)
1008                         ret |= fsck_blob(blob, buf, size, options);
1009                 else
1010                         ret |= report(options, &blob->object,
1011                                       FSCK_MSG_GITMODULES_BLOB,
1012                                       "non-blob found at .gitmodules");
1013                 free(buf);
1014         }
1015
1016
1017         oidset_clear(&gitmodules_found);
1018         oidset_clear(&gitmodules_done);
1019         return ret;
1020 }