fsck.c

   1 #include "cache.h"
   2 #include "object.h"
   3 #include "blob.h"
   4 #include "tree.h"
   5 #include "tree-walk.h"
   6 #include "commit.h"
   7 #include "tag.h"
   8 #include "fsck.h"
   9 #include "refs.h"
  10 #include "utf8.h"
  11 #include "sha1-array.h"
  12 #include "decorate.h"
  13 #include "oidset.h"
  14 #include "packfile.h"
  15 #include "submodule-config.h"
  16 #include "config.h"
  17
  18 static struct oidset gitmodules_found = OIDSET_INIT;
  19 static struct oidset gitmodules_done = OIDSET_INIT;
  20
  21 #define FSCK_FATAL -1
  22 #define FSCK_INFO -2
  23
  24 #define FOREACH_MSG_ID(FUNC) \
  25         /* fatal errors */ \
  26         FUNC(NUL_IN_HEADER, FATAL) \
  27         FUNC(UNTERMINATED_HEADER, FATAL) \
  28         /* errors */ \
  29         FUNC(BAD_DATE, ERROR) \
  30         FUNC(BAD_DATE_OVERFLOW, ERROR) \
  31         FUNC(BAD_EMAIL, ERROR) \
  32         FUNC(BAD_NAME, ERROR) \
  33         FUNC(BAD_OBJECT_SHA1, ERROR) \
  34         FUNC(BAD_PARENT_SHA1, ERROR) \
  35         FUNC(BAD_TAG_OBJECT, ERROR) \
  36         FUNC(BAD_TIMEZONE, ERROR) \
  37         FUNC(BAD_TREE, ERROR) \
  38         FUNC(BAD_TREE_SHA1, ERROR) \
  39         FUNC(BAD_TYPE, ERROR) \
  40         FUNC(DUPLICATE_ENTRIES, ERROR) \
  41         FUNC(MISSING_AUTHOR, ERROR) \
  42         FUNC(MISSING_COMMITTER, ERROR) \
  43         FUNC(MISSING_EMAIL, ERROR) \
  44         FUNC(MISSING_GRAFT, ERROR) \
  45         FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  46         FUNC(MISSING_OBJECT, ERROR) \
  47         FUNC(MISSING_PARENT, 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_TAG_OBJECT, ERROR) \
  53         FUNC(MISSING_TREE, ERROR) \
  54         FUNC(MISSING_TREE_OBJECT, ERROR) \
  55         FUNC(MISSING_TYPE, ERROR) \
  56         FUNC(MISSING_TYPE_ENTRY, ERROR) \
  57         FUNC(MULTIPLE_AUTHORS, ERROR) \
  58         FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  59         FUNC(TREE_NOT_SORTED, ERROR) \
  60         FUNC(UNKNOWN_TYPE, ERROR) \
  61         FUNC(ZERO_PADDED_DATE, ERROR) \
  62         FUNC(GITMODULES_MISSING, ERROR) \
  63         FUNC(GITMODULES_BLOB, ERROR) \
  64         FUNC(GITMODULES_PARSE, ERROR) \
  65         FUNC(GITMODULES_NAME, ERROR) \
  66         FUNC(GITMODULES_SYMLINK, ERROR) \
  67         FUNC(GITMODULES_URL, ERROR) \
  68         FUNC(GITMODULES_PATH, ERROR) \
  69         FUNC(GITMODULES_UPDATE, ERROR) \
  70         /* warnings */ \
  71         FUNC(BAD_FILEMODE, WARN) \
  72         FUNC(EMPTY_NAME, WARN) \
  73         FUNC(FULL_PATHNAME, WARN) \
  74         FUNC(HAS_DOT, WARN) \
  75         FUNC(HAS_DOTDOT, WARN) \
  76         FUNC(HAS_DOTGIT, WARN) \
  77         FUNC(NULL_SHA1, WARN) \
  78         FUNC(ZERO_PADDED_FILEMODE, WARN) \
  79         FUNC(NUL_IN_COMMIT, WARN) \
  80         /* infos (reported as warnings, but ignored by default) */ \
  81         FUNC(BAD_TAG_NAME, INFO) \
  82         FUNC(MISSING_TAGGER_ENTRY, INFO)
  83
  84 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
  85 enum fsck_msg_id {
  86         FOREACH_MSG_ID(MSG_ID)
  87         FSCK_MSG_MAX
  88 };
  89 #undef MSG_ID
  90
  91 #define STR(x) #x
  92 #define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
  93 static struct {
  94         const char *id_string;
  95         const char *downcased;
  96         int msg_type;
  97 } msg_id_info[FSCK_MSG_MAX + 1] = {
  98         FOREACH_MSG_ID(MSG_ID)
  99         { NULL, NULL, -1 }
 100 };
 101 #undef MSG_ID
 102
 103 static int parse_msg_id(const char *text)
 104 {
 105         int i;
 106
 107         if (!msg_id_info[0].downcased) {
 108                 /* convert id_string to lower case, without underscores. */
 109                 for (i = 0; i < FSCK_MSG_MAX; i++) {
 110                         const char *p = msg_id_info[i].id_string;
 111                         int len = strlen(p);
 112                         char *q = xmalloc(len);
 113
 114                         msg_id_info[i].downcased = q;
 115                         while (*p)
 116                                 if (*p == '_')
 117                                         p++;
 118                                 else
 119                                         *(q)++ = tolower(*(p)++);
 120                         *q = '\0';
 121                 }
 122         }
 123
 124         for (i = 0; i < FSCK_MSG_MAX; i++)
 125                 if (!strcmp(text, msg_id_info[i].downcased))
 126                         return i;
 127
 128         return -1;
 129 }
 130
 131 static int fsck_msg_type(enum fsck_msg_id msg_id,
 132         struct fsck_options *options)
 133 {
 134         int msg_type;
 135
 136         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 137
 138         if (options->msg_type)
 139                 msg_type = options->msg_type[msg_id];
 140         else {
 141                 msg_type = msg_id_info[msg_id].msg_type;
 142                 if (options->strict && msg_type == FSCK_WARN)
 143                         msg_type = FSCK_ERROR;
 144         }
 145
 146         return msg_type;
 147 }
 148
 149 static void init_skiplist(struct fsck_options *options, const char *path)
 150 {
 151         static struct oid_array skiplist = OID_ARRAY_INIT;
 152         int sorted, fd;
 153         char buffer[GIT_MAX_HEXSZ + 1];
 154         struct object_id oid;
 155
 156         if (options->skiplist)
 157                 sorted = options->skiplist->sorted;
 158         else {
 159                 sorted = 1;
 160                 options->skiplist = &skiplist;
 161         }
 162
 163         fd = open(path, O_RDONLY);
 164         if (fd < 0)
 165                 die("Could not open skip list: %s", path);
 166         for (;;) {
 167                 const char *p;
 168                 int result = read_in_full(fd, buffer, sizeof(buffer));
 169                 if (result < 0)
 170                         die_errno("Could not read '%s'", path);
 171                 if (!result)
 172                         break;
 173                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 174                         die("Invalid SHA-1: %s", buffer);
 175                 oid_array_append(&skiplist, &oid);
 176                 if (sorted && skiplist.nr > 1 &&
 177                                 oidcmp(&skiplist.oid[skiplist.nr - 2],
 178                                        &oid) > 0)
 179                         sorted = 0;
 180         }
 181         close(fd);
 182
 183         if (sorted)
 184                 skiplist.sorted = 1;
 185 }
 186
 187 static int parse_msg_type(const char *str)
 188 {
 189         if (!strcmp(str, "error"))
 190                 return FSCK_ERROR;
 191         else if (!strcmp(str, "warn"))
 192                 return FSCK_WARN;
 193         else if (!strcmp(str, "ignore"))
 194                 return FSCK_IGNORE;
 195         else
 196                 die("Unknown fsck message type: '%s'", str);
 197 }
 198
 199 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 200 {
 201         if (parse_msg_id(msg_id) < 0)
 202                 return 0;
 203         parse_msg_type(msg_type);
 204         return 1;
 205 }
 206
 207 void fsck_set_msg_type(struct fsck_options *options,
 208                 const char *msg_id, const char *msg_type)
 209 {
 210         int id = parse_msg_id(msg_id), type;
 211
 212         if (id < 0)
 213                 die("Unhandled message id: %s", msg_id);
 214         type = parse_msg_type(msg_type);
 215
 216         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 217                 die("Cannot demote %s to %s", msg_id, msg_type);
 218
 219         if (!options->msg_type) {
 220                 int i;
 221                 int *msg_type;
 222                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 223                 for (i = 0; i < FSCK_MSG_MAX; i++)
 224                         msg_type[i] = fsck_msg_type(i, options);
 225                 options->msg_type = msg_type;
 226         }
 227
 228         options->msg_type[id] = type;
 229 }
 230
 231 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 232 {
 233         char *buf = xstrdup(values), *to_free = buf;
 234         int done = 0;
 235
 236         while (!done) {
 237                 int len = strcspn(buf, " ,|"), equal;
 238
 239                 done = !buf[len];
 240                 if (!len) {
 241                         buf++;
 242                         continue;
 243                 }
 244                 buf[len] = '\0';
 245
 246                 for (equal = 0;
 247                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 248                      equal++)
 249                         buf[equal] = tolower(buf[equal]);
 250                 buf[equal] = '\0';
 251
 252                 if (!strcmp(buf, "skiplist")) {
 253                         if (equal == len)
 254                                 die("skiplist requires a path");
 255                         init_skiplist(options, buf + equal + 1);
 256                         buf += len + 1;
 257                         continue;
 258                 }
 259
 260                 if (equal == len)
 261                         die("Missing '=': '%s'", buf);
 262
 263                 fsck_set_msg_type(options, buf, buf + equal + 1);
 264                 buf += len + 1;
 265         }
 266         free(to_free);
 267 }
 268
 269 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 270 {
 271         for (;;) {
 272                 char c = *(msg_id)++;
 273
 274                 if (!c)
 275                         break;
 276                 if (c != '_')
 277                         strbuf_addch(sb, tolower(c));
 278                 else {
 279                         assert(*msg_id);
 280                         strbuf_addch(sb, *(msg_id)++);
 281                 }
 282         }
 283
 284         strbuf_addstr(sb, ": ");
 285 }
 286
 287 __attribute__((format (printf, 4, 5)))
 288 static int report(struct fsck_options *options, struct object *object,
 289         enum fsck_msg_id id, const char *fmt, ...)
 290 {
 291         va_list ap;
 292         struct strbuf sb = STRBUF_INIT;
 293         int msg_type = fsck_msg_type(id, options), result;
 294
 295         if (msg_type == FSCK_IGNORE)
 296                 return 0;
 297
 298         if (options->skiplist && object &&
 299                         oid_array_lookup(options->skiplist, &object->oid) >= 0)
 300                 return 0;
 301
 302         if (msg_type == FSCK_FATAL)
 303                 msg_type = FSCK_ERROR;
 304         else if (msg_type == FSCK_INFO)
 305                 msg_type = FSCK_WARN;
 306
 307         append_msg_id(&sb, msg_id_info[id].id_string);
 308
 309         va_start(ap, fmt);
 310         strbuf_vaddf(&sb, fmt, ap);
 311         result = options->error_func(options, object, msg_type, sb.buf);
 312         strbuf_release(&sb);
 313         va_end(ap);
 314
 315         return result;
 316 }
 317
 318 static char *get_object_name(struct fsck_options *options, struct object *obj)
 319 {
 320         if (!options->object_names)
 321                 return NULL;
 322         return lookup_decoration(options->object_names, obj);
 323 }
 324
 325 static void put_object_name(struct fsck_options *options, struct object *obj,
 326         const char *fmt, ...)
 327 {
 328         va_list ap;
 329         struct strbuf buf = STRBUF_INIT;
 330         char *existing;
 331
 332         if (!options->object_names)
 333                 return;
 334         existing = lookup_decoration(options->object_names, obj);
 335         if (existing)
 336                 return;
 337         va_start(ap, fmt);
 338         strbuf_vaddf(&buf, fmt, ap);
 339         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 340         va_end(ap);
 341 }
 342
 343 static const char *describe_object(struct fsck_options *o, struct object *obj)
 344 {
 345         static struct strbuf buf = STRBUF_INIT;
 346         char *name;
 347
 348         strbuf_reset(&buf);
 349         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 350         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 351                 strbuf_addf(&buf, " (%s)", name);
 352
 353         return buf.buf;
 354 }
 355
 356 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 357 {
 358         struct tree_desc desc;
 359         struct name_entry entry;
 360         int res = 0;
 361         const char *name;
 362
 363         if (parse_tree(tree))
 364                 return -1;
 365
 366         name = get_object_name(options, &tree->object);
 367         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 368                 return -1;
 369         while (tree_entry_gently(&desc, &entry)) {
 370                 struct object *obj;
 371                 int result;
 372
 373                 if (S_ISGITLINK(entry.mode))
 374                         continue;
 375
 376                 if (S_ISDIR(entry.mode)) {
 377                         obj = (struct object *)lookup_tree(entry.oid);
 378                         if (name && obj)
 379                                 put_object_name(options, obj, "%s%s/", name,
 380                                         entry.path);
 381                         result = options->walk(obj, OBJ_TREE, data, options);
 382                 }
 383                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 384                         obj = (struct object *)lookup_blob(entry.oid);
 385                         if (name && obj)
 386                                 put_object_name(options, obj, "%s%s", name,
 387                                         entry.path);
 388                         result = options->walk(obj, OBJ_BLOB, data, options);
 389                 }
 390                 else {
 391                         result = error("in tree %s: entry %s has bad mode %.6o",
 392                                         describe_object(options, &tree->object), entry.path, entry.mode);
 393                 }
 394                 if (result < 0)
 395                         return result;
 396                 if (!res)
 397                         res = result;
 398         }
 399         return res;
 400 }
 401
 402 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 403 {
 404         int counter = 0, generation = 0, name_prefix_len = 0;
 405         struct commit_list *parents;
 406         int res;
 407         int result;
 408         const char *name;
 409
 410         if (parse_commit(commit))
 411                 return -1;
 412
 413         name = get_object_name(options, &commit->object);
 414         if (name)
 415                 put_object_name(options, &commit->tree->object, "%s:", name);
 416
 417         result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options);
 418         if (result < 0)
 419                 return result;
 420         res = result;
 421
 422         parents = commit->parents;
 423         if (name && parents) {
 424                 int len = strlen(name), power;
 425
 426                 if (len && name[len - 1] == '^') {
 427                         generation = 1;
 428                         name_prefix_len = len - 1;
 429                 }
 430                 else { /* parse ~<generation> suffix */
 431                         for (generation = 0, power = 1;
 432                              len && isdigit(name[len - 1]);
 433                              power *= 10)
 434                                 generation += power * (name[--len] - '0');
 435                         if (power > 1 && len && name[len - 1] == '~')
 436                                 name_prefix_len = len - 1;
 437                 }
 438         }
 439
 440         while (parents) {
 441                 if (name) {
 442                         struct object *obj = &parents->item->object;
 443
 444                         if (++counter > 1)
 445                                 put_object_name(options, obj, "%s^%d",
 446                                         name, counter);
 447                         else if (generation > 0)
 448                                 put_object_name(options, obj, "%.*s~%d",
 449                                         name_prefix_len, name, generation + 1);
 450                         else
 451                                 put_object_name(options, obj, "%s^", name);
 452                 }
 453                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 454                 if (result < 0)
 455                         return result;
 456                 if (!res)
 457                         res = result;
 458                 parents = parents->next;
 459         }
 460         return res;
 461 }
 462
 463 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 464 {
 465         char *name = get_object_name(options, &tag->object);
 466
 467         if (parse_tag(tag))
 468                 return -1;
 469         if (name)
 470                 put_object_name(options, tag->tagged, "%s", name);
 471         return options->walk(tag->tagged, OBJ_ANY, data, options);
 472 }
 473
 474 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 475 {
 476         if (!obj)
 477                 return -1;
 478
 479         if (obj->type == OBJ_NONE)
 480                 parse_object(&obj->oid);
 481
 482         switch (obj->type) {
 483         case OBJ_BLOB:
 484                 return 0;
 485         case OBJ_TREE:
 486                 return fsck_walk_tree((struct tree *)obj, data, options);
 487         case OBJ_COMMIT:
 488                 return fsck_walk_commit((struct commit *)obj, data, options);
 489         case OBJ_TAG:
 490                 return fsck_walk_tag((struct tag *)obj, data, options);
 491         default:
 492                 error("Unknown object type for %s", describe_object(options, obj));
 493                 return -1;
 494         }
 495 }
 496
 497 /*
 498  * The entries in a tree are ordered in the _path_ order,
 499  * which means that a directory entry is ordered by adding
 500  * a slash to the end of it.
 501  *
 502  * So a directory called "a" is ordered _after_ a file
 503  * called "a.c", because "a/" sorts after "a.c".
 504  */
 505 #define TREE_UNORDERED (-1)
 506 #define TREE_HAS_DUPS  (-2)
 507
 508 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 509 {
 510         int len1 = strlen(name1);
 511         int len2 = strlen(name2);
 512         int len = len1 < len2 ? len1 : len2;
 513         unsigned char c1, c2;
 514         int cmp;
 515
 516         cmp = memcmp(name1, name2, len);
 517         if (cmp < 0)
 518                 return 0;
 519         if (cmp > 0)
 520                 return TREE_UNORDERED;
 521
 522         /*
 523          * Ok, the first <len> characters are the same.
 524          * Now we need to order the next one, but turn
 525          * a '\0' into a '/' for a directory entry.
 526          */
 527         c1 = name1[len];
 528         c2 = name2[len];
 529         if (!c1 && !c2)
 530                 /*
 531                  * git-write-tree used to write out a nonsense tree that has
 532                  * entries with the same name, one blob and one tree.  Make
 533                  * sure we do not have duplicate entries.
 534                  */
 535                 return TREE_HAS_DUPS;
 536         if (!c1 && S_ISDIR(mode1))
 537                 c1 = '/';
 538         if (!c2 && S_ISDIR(mode2))
 539                 c2 = '/';
 540         return c1 < c2 ? 0 : TREE_UNORDERED;
 541 }
 542
 543 static int fsck_tree(struct tree *item, struct fsck_options *options)
 544 {
 545         int retval = 0;
 546         int has_null_sha1 = 0;
 547         int has_full_path = 0;
 548         int has_empty_name = 0;
 549         int has_dot = 0;
 550         int has_dotdot = 0;
 551         int has_dotgit = 0;
 552         int has_zero_pad = 0;
 553         int has_bad_modes = 0;
 554         int has_dup_entries = 0;
 555         int not_properly_sorted = 0;
 556         struct tree_desc desc;
 557         unsigned o_mode;
 558         const char *o_name;
 559
 560         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 561                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 562                 return retval;
 563         }
 564
 565         o_mode = 0;
 566         o_name = NULL;
 567
 568         while (desc.size) {
 569                 unsigned mode;
 570                 const char *name, *backslash;
 571                 const struct object_id *oid;
 572
 573                 oid = tree_entry_extract(&desc, &name, &mode);
 574
 575                 has_null_sha1 |= is_null_oid(oid);
 576                 has_full_path |= !!strchr(name, '/');
 577                 has_empty_name |= !*name;
 578                 has_dot |= !strcmp(name, ".");
 579                 has_dotdot |= !strcmp(name, "..");
 580                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 581                 has_zero_pad |= *(char *)desc.buffer == '0';
 582
 583                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 584                         if (!S_ISLNK(mode))
 585                                 oidset_insert(&gitmodules_found, oid);
 586                         else
 587                                 retval += report(options, &item->object,
 588                                                  FSCK_MSG_GITMODULES_SYMLINK,
 589                                                  ".gitmodules is a symbolic link");
 590                 }
 591
 592                 if ((backslash = strchr(name, '\\'))) {
 593                         while (backslash) {
 594                                 backslash++;
 595                                 has_dotgit |= is_ntfs_dotgit(backslash);
 596                                 if (is_ntfs_dotgitmodules(backslash)) {
 597                                         if (!S_ISLNK(mode))
 598                                                 oidset_insert(&gitmodules_found, oid);
 599                                         else
 600                                                 retval += report(options, &item->object,
 601                                                                  FSCK_MSG_GITMODULES_SYMLINK,
 602                                                                  ".gitmodules is a symbolic link");
 603                                 }
 604                                 backslash = strchr(backslash, '\\');
 605                         }
 606                 }
 607
 608                 if (update_tree_entry_gently(&desc)) {
 609                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 610                         break;
 611                 }
 612
 613                 switch (mode) {
 614                 /*
 615                  * Standard modes..
 616                  */
 617                 case S_IFREG | 0755:
 618                 case S_IFREG | 0644:
 619                 case S_IFLNK:
 620                 case S_IFDIR:
 621                 case S_IFGITLINK:
 622                         break;
 623                 /*
 624                  * This is nonstandard, but we had a few of these
 625                  * early on when we honored the full set of mode
 626                  * bits..
 627                  */
 628                 case S_IFREG | 0664:
 629                         if (!options->strict)
 630                                 break;
 631                         /* fallthrough */
 632                 default:
 633                         has_bad_modes = 1;
 634                 }
 635
 636                 if (o_name) {
 637                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 638                         case TREE_UNORDERED:
 639                                 not_properly_sorted = 1;
 640                                 break;
 641                         case TREE_HAS_DUPS:
 642                                 has_dup_entries = 1;
 643                                 break;
 644                         default:
 645                                 break;
 646                         }
 647                 }
 648
 649                 o_mode = mode;
 650                 o_name = name;
 651         }
 652
 653         if (has_null_sha1)
 654                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 655         if (has_full_path)
 656                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 657         if (has_empty_name)
 658                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 659         if (has_dot)
 660                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 661         if (has_dotdot)
 662                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 663         if (has_dotgit)
 664                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 665         if (has_zero_pad)
 666                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 667         if (has_bad_modes)
 668                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 669         if (has_dup_entries)
 670                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 671         if (not_properly_sorted)
 672                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 673         return retval;
 674 }
 675
 676 static int verify_headers(const void *data, unsigned long size,
 677                           struct object *obj, struct fsck_options *options)
 678 {
 679         const char *buffer = (const char *)data;
 680         unsigned long i;
 681
 682         for (i = 0; i < size; i++) {
 683                 switch (buffer[i]) {
 684                 case '\0':
 685                         return report(options, obj,
 686                                 FSCK_MSG_NUL_IN_HEADER,
 687                                 "unterminated header: NUL at offset %ld", i);
 688                 case '\n':
 689                         if (i + 1 < size && buffer[i + 1] == '\n')
 690                                 return 0;
 691                 }
 692         }
 693
 694         /*
 695          * We did not find double-LF that separates the header
 696          * and the body.  Not having a body is not a crime but
 697          * we do want to see the terminating LF for the last header
 698          * line.
 699          */
 700         if (size && buffer[size - 1] == '\n')
 701                 return 0;
 702
 703         return report(options, obj,
 704                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 705 }
 706
 707 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 708 {
 709         const char *p = *ident;
 710         char *end;
 711
 712         *ident = strchrnul(*ident, '\n');
 713         if (**ident == '\n')
 714                 (*ident)++;
 715
 716         if (*p == '<')
 717                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 718         p += strcspn(p, "<>\n");
 719         if (*p == '>')
 720                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 721         if (*p != '<')
 722                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 723         if (p[-1] != ' ')
 724                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 725         p++;
 726         p += strcspn(p, "<>\n");
 727         if (*p != '>')
 728                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 729         p++;
 730         if (*p != ' ')
 731                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 732         p++;
 733         if (*p == '0' && p[1] != ' ')
 734                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 735         if (date_overflows(parse_timestamp(p, &end, 10)))
 736                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 737         if ((end == p || *end != ' '))
 738                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 739         p = end + 1;
 740         if ((*p != '+' && *p != '-') ||
 741             !isdigit(p[1]) ||
 742             !isdigit(p[2]) ||
 743             !isdigit(p[3]) ||
 744             !isdigit(p[4]) ||
 745             (p[5] != '\n'))
 746                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 747         p += 6;
 748         return 0;
 749 }
 750
 751 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 752         unsigned long size, struct fsck_options *options)
 753 {
 754         unsigned char tree_sha1[20], sha1[20];
 755         struct commit_graft *graft;
 756         unsigned parent_count, parent_line_count = 0, author_count;
 757         int err;
 758         const char *buffer_begin = buffer;
 759
 760         if (verify_headers(buffer, size, &commit->object, options))
 761                 return -1;
 762
 763         if (!skip_prefix(buffer, "tree ", &buffer))
 764                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 765         if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
 766                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 767                 if (err)
 768                         return err;
 769         }
 770         buffer += 41;
 771         while (skip_prefix(buffer, "parent ", &buffer)) {
 772                 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 773                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 774                         if (err)
 775                                 return err;
 776                 }
 777                 buffer += 41;
 778                 parent_line_count++;
 779         }
 780         graft = lookup_commit_graft(&commit->object.oid);
 781         parent_count = commit_list_count(commit->parents);
 782         if (graft) {
 783                 if (graft->nr_parent == -1 && !parent_count)
 784                         ; /* shallow commit */
 785                 else if (graft->nr_parent != parent_count) {
 786                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 787                         if (err)
 788                                 return err;
 789                 }
 790         } else {
 791                 if (parent_count != parent_line_count) {
 792                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 793                         if (err)
 794                                 return err;
 795                 }
 796         }
 797         author_count = 0;
 798         while (skip_prefix(buffer, "author ", &buffer)) {
 799                 author_count++;
 800                 err = fsck_ident(&buffer, &commit->object, options);
 801                 if (err)
 802                         return err;
 803         }
 804         if (author_count < 1)
 805                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 806         else if (author_count > 1)
 807                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 808         if (err)
 809                 return err;
 810         if (!skip_prefix(buffer, "committer ", &buffer))
 811                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 812         err = fsck_ident(&buffer, &commit->object, options);
 813         if (err)
 814                 return err;
 815         if (!commit->tree) {
 816                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
 817                 if (err)
 818                         return err;
 819         }
 820         if (memchr(buffer_begin, '\0', size)) {
 821                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 822                              "NUL byte in the commit object body");
 823                 if (err)
 824                         return err;
 825         }
 826         return 0;
 827 }
 828
 829 static int fsck_commit(struct commit *commit, const char *data,
 830         unsigned long size, struct fsck_options *options)
 831 {
 832         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 833         int ret = fsck_commit_buffer(commit, buffer, size, options);
 834         if (!data)
 835                 unuse_commit_buffer(commit, buffer);
 836         return ret;
 837 }
 838
 839 static int fsck_tag_buffer(struct tag *tag, const char *data,
 840         unsigned long size, struct fsck_options *options)
 841 {
 842         unsigned char sha1[20];
 843         int ret = 0;
 844         const char *buffer;
 845         char *to_free = NULL, *eol;
 846         struct strbuf sb = STRBUF_INIT;
 847
 848         if (data)
 849                 buffer = data;
 850         else {
 851                 enum object_type type;
 852
 853                 buffer = to_free =
 854                         read_sha1_file(tag->object.oid.hash, &type, &size);
 855                 if (!buffer)
 856                         return report(options, &tag->object,
 857                                 FSCK_MSG_MISSING_TAG_OBJECT,
 858                                 "cannot read tag object");
 859
 860                 if (type != OBJ_TAG) {
 861                         ret = report(options, &tag->object,
 862                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 863                                 "expected tag got %s",
 864                             type_name(type));
 865                         goto done;
 866                 }
 867         }
 868
 869         ret = verify_headers(buffer, size, &tag->object, options);
 870         if (ret)
 871                 goto done;
 872
 873         if (!skip_prefix(buffer, "object ", &buffer)) {
 874                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 875                 goto done;
 876         }
 877         if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 878                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 879                 if (ret)
 880                         goto done;
 881         }
 882         buffer += 41;
 883
 884         if (!skip_prefix(buffer, "type ", &buffer)) {
 885                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 886                 goto done;
 887         }
 888         eol = strchr(buffer, '\n');
 889         if (!eol) {
 890                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 891                 goto done;
 892         }
 893         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 894                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 895         if (ret)
 896                 goto done;
 897         buffer = eol + 1;
 898
 899         if (!skip_prefix(buffer, "tag ", &buffer)) {
 900                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 901                 goto done;
 902         }
 903         eol = strchr(buffer, '\n');
 904         if (!eol) {
 905                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 906                 goto done;
 907         }
 908         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 909         if (check_refname_format(sb.buf, 0)) {
 910                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 911                            "invalid 'tag' name: %.*s",
 912                            (int)(eol - buffer), buffer);
 913                 if (ret)
 914                         goto done;
 915         }
 916         buffer = eol + 1;
 917
 918         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 919                 /* early tags do not contain 'tagger' lines; warn only */
 920                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 921                 if (ret)
 922                         goto done;
 923         }
 924         else
 925                 ret = fsck_ident(&buffer, &tag->object, options);
 926
 927 done:
 928         strbuf_release(&sb);
 929         free(to_free);
 930         return ret;
 931 }
 932
 933 static int fsck_tag(struct tag *tag, const char *data,
 934         unsigned long size, struct fsck_options *options)
 935 {
 936         struct object *tagged = tag->tagged;
 937
 938         if (!tagged)
 939                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 940
 941         return fsck_tag_buffer(tag, data, size, options);
 942 }
 943
 944 struct fsck_gitmodules_data {
 945         struct object *obj;
 946         struct fsck_options *options;
 947         int ret;
 948 };
 949
 950 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 951 {
 952         struct fsck_gitmodules_data *data = vdata;
 953         const char *subsection, *key;
 954         int subsection_len;
 955         char *name;
 956
 957         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 958             !subsection)
 959                 return 0;
 960
 961         name = xmemdupz(subsection, subsection_len);
 962         if (check_submodule_name(name) < 0)
 963                 data->ret |= report(data->options, data->obj,
 964                                     FSCK_MSG_GITMODULES_NAME,
 965                                     "disallowed submodule name: %s",
 966                                     name);
 967         if (!strcmp(key, "url") && value &&
 968             looks_like_command_line_option(value))
 969                 data->ret |= report(data->options, data->obj,
 970                                     FSCK_MSG_GITMODULES_URL,
 971                                     "disallowed submodule url: %s",
 972                                     value);
 973         if (!strcmp(key, "path") && value &&
 974             looks_like_command_line_option(value))
 975                 data->ret |= report(data->options, data->obj,
 976                                     FSCK_MSG_GITMODULES_PATH,
 977                                     "disallowed submodule path: %s",
 978                                     value);
 979         if (!strcmp(key, "update") && value &&
 980             parse_submodule_update_type(value) == SM_UPDATE_COMMAND)
 981                 data->ret |= report(data->options, data->obj,
 982                                     FSCK_MSG_GITMODULES_UPDATE,
 983                                     "disallowed submodule update setting: %s",
 984                                     value);
 985         free(name);
 986
 987         return 0;
 988 }
 989
 990 static int fsck_blob(struct blob *blob, const char *buf,
 991                      unsigned long size, struct fsck_options *options)
 992 {
 993         struct fsck_gitmodules_data data;
 994
 995         if (!oidset_contains(&gitmodules_found, &blob->object.oid))
 996                 return 0;
 997         oidset_insert(&gitmodules_done, &blob->object.oid);
 998
 999         if (!buf) {
1000                 /*
1001                  * A missing buffer here is a sign that the caller found the
1002                  * blob too gigantic to load into memory. Let's just consider
1003                  * that an error.
1004                  */
1005                 return report(options, &blob->object,
1006                               FSCK_MSG_GITMODULES_PARSE,
1007                               ".gitmodules too large to parse");
1008         }
1009
1010         data.obj = &blob->object;
1011         data.options = options;
1012         data.ret = 0;
1013         if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1014                                 ".gitmodules", buf, size, &data))
1015                 data.ret |= report(options, &blob->object,
1016                                    FSCK_MSG_GITMODULES_PARSE,
1017                                    "could not parse gitmodules blob");
1018
1019         return data.ret;
1020 }
1021
1022 int fsck_object(struct object *obj, void *data, unsigned long size,
1023         struct fsck_options *options)
1024 {
1025         if (!obj)
1026                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1027
1028         if (obj->type == OBJ_BLOB)
1029                 return fsck_blob((struct blob *)obj, data, size, options);
1030         if (obj->type == OBJ_TREE)
1031                 return fsck_tree((struct tree *) obj, options);
1032         if (obj->type == OBJ_COMMIT)
1033                 return fsck_commit((struct commit *) obj, (const char *) data,
1034                         size, options);
1035         if (obj->type == OBJ_TAG)
1036                 return fsck_tag((struct tag *) obj, (const char *) data,
1037                         size, options);
1038
1039         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1040                           obj->type);
1041 }
1042
1043 int fsck_error_function(struct fsck_options *o,
1044         struct object *obj, int msg_type, const char *message)
1045 {
1046         if (msg_type == FSCK_WARN) {
1047                 warning("object %s: %s", describe_object(o, obj), message);
1048                 return 0;
1049         }
1050         error("object %s: %s", describe_object(o, obj), message);
1051         return 1;
1052 }
1053
1054 int fsck_finish(struct fsck_options *options)
1055 {
1056         int ret = 0;
1057         struct oidset_iter iter;
1058         const struct object_id *oid;
1059
1060         oidset_iter_init(&gitmodules_found, &iter);
1061         while ((oid = oidset_iter_next(&iter))) {
1062                 struct blob *blob;
1063                 enum object_type type;
1064                 unsigned long size;
1065                 char *buf;
1066
1067                 if (oidset_contains(&gitmodules_done, oid))
1068                         continue;
1069
1070                 blob = lookup_blob(oid);
1071                 if (!blob) {
1072                         ret |= report(options, &blob->object,
1073                                       FSCK_MSG_GITMODULES_BLOB,
1074                                       "non-blob found at .gitmodules");
1075                         continue;
1076                 }
1077
1078                 buf = read_sha1_file(oid->hash, &type, &size);
1079                 if (!buf) {
1080                         if (is_promisor_object(&blob->object.oid))
1081                                 continue;
1082                         ret |= report(options, &blob->object,
1083                                       FSCK_MSG_GITMODULES_MISSING,
1084                                       "unable to read .gitmodules blob");
1085                         continue;
1086                 }
1087
1088                 if (type == OBJ_BLOB)
1089                         ret |= fsck_blob(blob, buf, size, options);
1090                 else
1091                         ret |= report(options, &blob->object,
1092                                       FSCK_MSG_GITMODULES_BLOB,
1093                                       "non-blob found at .gitmodules");
1094                 free(buf);
1095         }
1096
1097
1098         oidset_clear(&gitmodules_found);
1099         oidset_clear(&gitmodules_done);
1100         return ret;
1101 }