fsck.c

   1 #include "cache.h"
   2 #include "object-store.h"
   3 #include "object.h"
   4 #include "blob.h"
   5 #include "tree.h"
   6 #include "tree-walk.h"
   7 #include "commit.h"
   8 #include "tag.h"
   9 #include "fsck.h"
  10 #include "refs.h"
  11 #include "utf8.h"
  12 #include "sha1-array.h"
  13 #include "decorate.h"
  14 #include "oidset.h"
  15 #include "packfile.h"
  16 #include "submodule-config.h"
  17 #include "config.h"
  18 #include "help.h"
  19
  20 static struct oidset gitmodules_found = OIDSET_INIT;
  21 static struct oidset gitmodules_done = OIDSET_INIT;
  22
  23 #define FSCK_FATAL -1
  24 #define FSCK_INFO -2
  25
  26 #define FOREACH_MSG_ID(FUNC) \
  27         /* fatal errors */ \
  28         FUNC(NUL_IN_HEADER, FATAL) \
  29         FUNC(UNTERMINATED_HEADER, FATAL) \
  30         /* errors */ \
  31         FUNC(BAD_DATE, ERROR) \
  32         FUNC(BAD_DATE_OVERFLOW, ERROR) \
  33         FUNC(BAD_EMAIL, ERROR) \
  34         FUNC(BAD_NAME, ERROR) \
  35         FUNC(BAD_OBJECT_SHA1, ERROR) \
  36         FUNC(BAD_PARENT_SHA1, ERROR) \
  37         FUNC(BAD_TAG_OBJECT, ERROR) \
  38         FUNC(BAD_TIMEZONE, ERROR) \
  39         FUNC(BAD_TREE, ERROR) \
  40         FUNC(BAD_TREE_SHA1, ERROR) \
  41         FUNC(BAD_TYPE, ERROR) \
  42         FUNC(DUPLICATE_ENTRIES, ERROR) \
  43         FUNC(MISSING_AUTHOR, ERROR) \
  44         FUNC(MISSING_COMMITTER, ERROR) \
  45         FUNC(MISSING_EMAIL, ERROR) \
  46         FUNC(MISSING_GRAFT, ERROR) \
  47         FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  48         FUNC(MISSING_OBJECT, ERROR) \
  49         FUNC(MISSING_PARENT, ERROR) \
  50         FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  51         FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  52         FUNC(MISSING_TAG, ERROR) \
  53         FUNC(MISSING_TAG_ENTRY, ERROR) \
  54         FUNC(MISSING_TAG_OBJECT, ERROR) \
  55         FUNC(MISSING_TREE, ERROR) \
  56         FUNC(MISSING_TREE_OBJECT, ERROR) \
  57         FUNC(MISSING_TYPE, ERROR) \
  58         FUNC(MISSING_TYPE_ENTRY, ERROR) \
  59         FUNC(MULTIPLE_AUTHORS, ERROR) \
  60         FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  61         FUNC(TREE_NOT_SORTED, ERROR) \
  62         FUNC(UNKNOWN_TYPE, ERROR) \
  63         FUNC(ZERO_PADDED_DATE, ERROR) \
  64         FUNC(GITMODULES_MISSING, ERROR) \
  65         FUNC(GITMODULES_BLOB, ERROR) \
  66         FUNC(GITMODULES_LARGE, ERROR) \
  67         FUNC(GITMODULES_NAME, ERROR) \
  68         FUNC(GITMODULES_SYMLINK, ERROR) \
  69         /* warnings */ \
  70         FUNC(BAD_FILEMODE, WARN) \
  71         FUNC(EMPTY_NAME, WARN) \
  72         FUNC(FULL_PATHNAME, WARN) \
  73         FUNC(HAS_DOT, WARN) \
  74         FUNC(HAS_DOTDOT, WARN) \
  75         FUNC(HAS_DOTGIT, WARN) \
  76         FUNC(NULL_SHA1, WARN) \
  77         FUNC(ZERO_PADDED_FILEMODE, WARN) \
  78         FUNC(NUL_IN_COMMIT, WARN) \
  79         /* infos (reported as warnings, but ignored by default) */ \
  80         FUNC(GITMODULES_PARSE, INFO) \
  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, NULL, FSCK_##msg_type },
  93 static struct {
  94         const char *id_string;
  95         const char *downcased;
  96         const char *camelcased;
  97         int msg_type;
  98 } msg_id_info[FSCK_MSG_MAX + 1] = {
  99         FOREACH_MSG_ID(MSG_ID)
 100         { NULL, NULL, NULL, -1 }
 101 };
 102 #undef MSG_ID
 103
 104 static void prepare_msg_ids(void)
 105 {
 106         int i;
 107
 108         if (msg_id_info[0].downcased)
 109                 return;
 110
 111         /* convert id_string to lower case, without underscores. */
 112         for (i = 0; i < FSCK_MSG_MAX; i++) {
 113                 const char *p = msg_id_info[i].id_string;
 114                 int len = strlen(p);
 115                 char *q = xmalloc(len);
 116
 117                 msg_id_info[i].downcased = q;
 118                 while (*p)
 119                         if (*p == '_')
 120                                 p++;
 121                         else
 122                                 *(q)++ = tolower(*(p)++);
 123                 *q = '\0';
 124
 125                 p = msg_id_info[i].id_string;
 126                 q = xmalloc(len);
 127                 msg_id_info[i].camelcased = q;
 128                 while (*p) {
 129                         if (*p == '_') {
 130                                 p++;
 131                                 if (*p)
 132                                         *q++ = *p++;
 133                         } else {
 134                                 *q++ = tolower(*p++);
 135                         }
 136                 }
 137                 *q = '\0';
 138         }
 139 }
 140
 141 static int parse_msg_id(const char *text)
 142 {
 143         int i;
 144
 145         prepare_msg_ids();
 146
 147         for (i = 0; i < FSCK_MSG_MAX; i++)
 148                 if (!strcmp(text, msg_id_info[i].downcased))
 149                         return i;
 150
 151         return -1;
 152 }
 153
 154 void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
 155 {
 156         int i;
 157
 158         prepare_msg_ids();
 159
 160         for (i = 0; i < FSCK_MSG_MAX; i++)
 161                 list_config_item(list, prefix, msg_id_info[i].camelcased);
 162 }
 163
 164 static int fsck_msg_type(enum fsck_msg_id msg_id,
 165         struct fsck_options *options)
 166 {
 167         int msg_type;
 168
 169         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 170
 171         if (options->msg_type)
 172                 msg_type = options->msg_type[msg_id];
 173         else {
 174                 msg_type = msg_id_info[msg_id].msg_type;
 175                 if (options->strict && msg_type == FSCK_WARN)
 176                         msg_type = FSCK_ERROR;
 177         }
 178
 179         return msg_type;
 180 }
 181
 182 static void init_skiplist(struct fsck_options *options, const char *path)
 183 {
 184         static struct oid_array skiplist = OID_ARRAY_INIT;
 185         int sorted, fd;
 186         char buffer[GIT_MAX_HEXSZ + 1];
 187         struct object_id oid;
 188
 189         if (options->skiplist)
 190                 sorted = options->skiplist->sorted;
 191         else {
 192                 sorted = 1;
 193                 options->skiplist = &skiplist;
 194         }
 195
 196         fd = open(path, O_RDONLY);
 197         if (fd < 0)
 198                 die("Could not open skip list: %s", path);
 199         for (;;) {
 200                 const char *p;
 201                 int result = read_in_full(fd, buffer, sizeof(buffer));
 202                 if (result < 0)
 203                         die_errno("Could not read '%s'", path);
 204                 if (!result)
 205                         break;
 206                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 207                         die("Invalid SHA-1: %s", buffer);
 208                 oid_array_append(&skiplist, &oid);
 209                 if (sorted && skiplist.nr > 1 &&
 210                                 oidcmp(&skiplist.oid[skiplist.nr - 2],
 211                                        &oid) > 0)
 212                         sorted = 0;
 213         }
 214         close(fd);
 215
 216         if (sorted)
 217                 skiplist.sorted = 1;
 218 }
 219
 220 static int parse_msg_type(const char *str)
 221 {
 222         if (!strcmp(str, "error"))
 223                 return FSCK_ERROR;
 224         else if (!strcmp(str, "warn"))
 225                 return FSCK_WARN;
 226         else if (!strcmp(str, "ignore"))
 227                 return FSCK_IGNORE;
 228         else
 229                 die("Unknown fsck message type: '%s'", str);
 230 }
 231
 232 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 233 {
 234         if (parse_msg_id(msg_id) < 0)
 235                 return 0;
 236         parse_msg_type(msg_type);
 237         return 1;
 238 }
 239
 240 void fsck_set_msg_type(struct fsck_options *options,
 241                 const char *msg_id, const char *msg_type)
 242 {
 243         int id = parse_msg_id(msg_id), type;
 244
 245         if (id < 0)
 246                 die("Unhandled message id: %s", msg_id);
 247         type = parse_msg_type(msg_type);
 248
 249         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 250                 die("Cannot demote %s to %s", msg_id, msg_type);
 251
 252         if (!options->msg_type) {
 253                 int i;
 254                 int *msg_type;
 255                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 256                 for (i = 0; i < FSCK_MSG_MAX; i++)
 257                         msg_type[i] = fsck_msg_type(i, options);
 258                 options->msg_type = msg_type;
 259         }
 260
 261         options->msg_type[id] = type;
 262 }
 263
 264 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 265 {
 266         char *buf = xstrdup(values), *to_free = buf;
 267         int done = 0;
 268
 269         while (!done) {
 270                 int len = strcspn(buf, " ,|"), equal;
 271
 272                 done = !buf[len];
 273                 if (!len) {
 274                         buf++;
 275                         continue;
 276                 }
 277                 buf[len] = '\0';
 278
 279                 for (equal = 0;
 280                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 281                      equal++)
 282                         buf[equal] = tolower(buf[equal]);
 283                 buf[equal] = '\0';
 284
 285                 if (!strcmp(buf, "skiplist")) {
 286                         if (equal == len)
 287                                 die("skiplist requires a path");
 288                         init_skiplist(options, buf + equal + 1);
 289                         buf += len + 1;
 290                         continue;
 291                 }
 292
 293                 if (equal == len)
 294                         die("Missing '=': '%s'", buf);
 295
 296                 fsck_set_msg_type(options, buf, buf + equal + 1);
 297                 buf += len + 1;
 298         }
 299         free(to_free);
 300 }
 301
 302 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 303 {
 304         for (;;) {
 305                 char c = *(msg_id)++;
 306
 307                 if (!c)
 308                         break;
 309                 if (c != '_')
 310                         strbuf_addch(sb, tolower(c));
 311                 else {
 312                         assert(*msg_id);
 313                         strbuf_addch(sb, *(msg_id)++);
 314                 }
 315         }
 316
 317         strbuf_addstr(sb, ": ");
 318 }
 319
 320 static int object_on_skiplist(struct fsck_options *opts, struct object *obj)
 321 {
 322         if (opts && opts->skiplist && obj)
 323                 return oid_array_lookup(opts->skiplist, &obj->oid) >= 0;
 324         return 0;
 325 }
 326
 327 __attribute__((format (printf, 4, 5)))
 328 static int report(struct fsck_options *options, struct object *object,
 329         enum fsck_msg_id id, const char *fmt, ...)
 330 {
 331         va_list ap;
 332         struct strbuf sb = STRBUF_INIT;
 333         int msg_type = fsck_msg_type(id, options), result;
 334
 335         if (msg_type == FSCK_IGNORE)
 336                 return 0;
 337
 338         if (object_on_skiplist(options, object))
 339                 return 0;
 340
 341         if (msg_type == FSCK_FATAL)
 342                 msg_type = FSCK_ERROR;
 343         else if (msg_type == FSCK_INFO)
 344                 msg_type = FSCK_WARN;
 345
 346         append_msg_id(&sb, msg_id_info[id].id_string);
 347
 348         va_start(ap, fmt);
 349         strbuf_vaddf(&sb, fmt, ap);
 350         result = options->error_func(options, object, msg_type, sb.buf);
 351         strbuf_release(&sb);
 352         va_end(ap);
 353
 354         return result;
 355 }
 356
 357 static char *get_object_name(struct fsck_options *options, struct object *obj)
 358 {
 359         if (!options->object_names)
 360                 return NULL;
 361         return lookup_decoration(options->object_names, obj);
 362 }
 363
 364 static void put_object_name(struct fsck_options *options, struct object *obj,
 365         const char *fmt, ...)
 366 {
 367         va_list ap;
 368         struct strbuf buf = STRBUF_INIT;
 369         char *existing;
 370
 371         if (!options->object_names)
 372                 return;
 373         existing = lookup_decoration(options->object_names, obj);
 374         if (existing)
 375                 return;
 376         va_start(ap, fmt);
 377         strbuf_vaddf(&buf, fmt, ap);
 378         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 379         va_end(ap);
 380 }
 381
 382 static const char *describe_object(struct fsck_options *o, struct object *obj)
 383 {
 384         static struct strbuf buf = STRBUF_INIT;
 385         char *name;
 386
 387         strbuf_reset(&buf);
 388         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 389         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 390                 strbuf_addf(&buf, " (%s)", name);
 391
 392         return buf.buf;
 393 }
 394
 395 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 396 {
 397         struct tree_desc desc;
 398         struct name_entry entry;
 399         int res = 0;
 400         const char *name;
 401
 402         if (parse_tree(tree))
 403                 return -1;
 404
 405         name = get_object_name(options, &tree->object);
 406         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 407                 return -1;
 408         while (tree_entry_gently(&desc, &entry)) {
 409                 struct object *obj;
 410                 int result;
 411
 412                 if (S_ISGITLINK(entry.mode))
 413                         continue;
 414
 415                 if (S_ISDIR(entry.mode)) {
 416                         obj = (struct object *)lookup_tree(entry.oid);
 417                         if (name && obj)
 418                                 put_object_name(options, obj, "%s%s/", name,
 419                                         entry.path);
 420                         result = options->walk(obj, OBJ_TREE, data, options);
 421                 }
 422                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 423                         obj = (struct object *)lookup_blob(entry.oid);
 424                         if (name && obj)
 425                                 put_object_name(options, obj, "%s%s", name,
 426                                         entry.path);
 427                         result = options->walk(obj, OBJ_BLOB, data, options);
 428                 }
 429                 else {
 430                         result = error("in tree %s: entry %s has bad mode %.6o",
 431                                         describe_object(options, &tree->object), entry.path, entry.mode);
 432                 }
 433                 if (result < 0)
 434                         return result;
 435                 if (!res)
 436                         res = result;
 437         }
 438         return res;
 439 }
 440
 441 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 442 {
 443         int counter = 0, generation = 0, name_prefix_len = 0;
 444         struct commit_list *parents;
 445         int res;
 446         int result;
 447         const char *name;
 448
 449         if (parse_commit(commit))
 450                 return -1;
 451
 452         name = get_object_name(options, &commit->object);
 453         if (name)
 454                 put_object_name(options, &get_commit_tree(commit)->object,
 455                                 "%s:", name);
 456
 457         result = options->walk((struct object *)get_commit_tree(commit),
 458                                OBJ_TREE, data, options);
 459         if (result < 0)
 460                 return result;
 461         res = result;
 462
 463         parents = commit->parents;
 464         if (name && parents) {
 465                 int len = strlen(name), power;
 466
 467                 if (len && name[len - 1] == '^') {
 468                         generation = 1;
 469                         name_prefix_len = len - 1;
 470                 }
 471                 else { /* parse ~<generation> suffix */
 472                         for (generation = 0, power = 1;
 473                              len && isdigit(name[len - 1]);
 474                              power *= 10)
 475                                 generation += power * (name[--len] - '0');
 476                         if (power > 1 && len && name[len - 1] == '~')
 477                                 name_prefix_len = len - 1;
 478                 }
 479         }
 480
 481         while (parents) {
 482                 if (name) {
 483                         struct object *obj = &parents->item->object;
 484
 485                         if (++counter > 1)
 486                                 put_object_name(options, obj, "%s^%d",
 487                                         name, counter);
 488                         else if (generation > 0)
 489                                 put_object_name(options, obj, "%.*s~%d",
 490                                         name_prefix_len, name, generation + 1);
 491                         else
 492                                 put_object_name(options, obj, "%s^", name);
 493                 }
 494                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 495                 if (result < 0)
 496                         return result;
 497                 if (!res)
 498                         res = result;
 499                 parents = parents->next;
 500         }
 501         return res;
 502 }
 503
 504 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 505 {
 506         char *name = get_object_name(options, &tag->object);
 507
 508         if (parse_tag(tag))
 509                 return -1;
 510         if (name)
 511                 put_object_name(options, tag->tagged, "%s", name);
 512         return options->walk(tag->tagged, OBJ_ANY, data, options);
 513 }
 514
 515 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 516 {
 517         if (!obj)
 518                 return -1;
 519
 520         if (obj->type == OBJ_NONE)
 521                 parse_object(&obj->oid);
 522
 523         switch (obj->type) {
 524         case OBJ_BLOB:
 525                 return 0;
 526         case OBJ_TREE:
 527                 return fsck_walk_tree((struct tree *)obj, data, options);
 528         case OBJ_COMMIT:
 529                 return fsck_walk_commit((struct commit *)obj, data, options);
 530         case OBJ_TAG:
 531                 return fsck_walk_tag((struct tag *)obj, data, options);
 532         default:
 533                 error("Unknown object type for %s", describe_object(options, obj));
 534                 return -1;
 535         }
 536 }
 537
 538 /*
 539  * The entries in a tree are ordered in the _path_ order,
 540  * which means that a directory entry is ordered by adding
 541  * a slash to the end of it.
 542  *
 543  * So a directory called "a" is ordered _after_ a file
 544  * called "a.c", because "a/" sorts after "a.c".
 545  */
 546 #define TREE_UNORDERED (-1)
 547 #define TREE_HAS_DUPS  (-2)
 548
 549 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 550 {
 551         int len1 = strlen(name1);
 552         int len2 = strlen(name2);
 553         int len = len1 < len2 ? len1 : len2;
 554         unsigned char c1, c2;
 555         int cmp;
 556
 557         cmp = memcmp(name1, name2, len);
 558         if (cmp < 0)
 559                 return 0;
 560         if (cmp > 0)
 561                 return TREE_UNORDERED;
 562
 563         /*
 564          * Ok, the first <len> characters are the same.
 565          * Now we need to order the next one, but turn
 566          * a '\0' into a '/' for a directory entry.
 567          */
 568         c1 = name1[len];
 569         c2 = name2[len];
 570         if (!c1 && !c2)
 571                 /*
 572                  * git-write-tree used to write out a nonsense tree that has
 573                  * entries with the same name, one blob and one tree.  Make
 574                  * sure we do not have duplicate entries.
 575                  */
 576                 return TREE_HAS_DUPS;
 577         if (!c1 && S_ISDIR(mode1))
 578                 c1 = '/';
 579         if (!c2 && S_ISDIR(mode2))
 580                 c2 = '/';
 581         return c1 < c2 ? 0 : TREE_UNORDERED;
 582 }
 583
 584 static int fsck_tree(struct tree *item, struct fsck_options *options)
 585 {
 586         int retval = 0;
 587         int has_null_sha1 = 0;
 588         int has_full_path = 0;
 589         int has_empty_name = 0;
 590         int has_dot = 0;
 591         int has_dotdot = 0;
 592         int has_dotgit = 0;
 593         int has_zero_pad = 0;
 594         int has_bad_modes = 0;
 595         int has_dup_entries = 0;
 596         int not_properly_sorted = 0;
 597         struct tree_desc desc;
 598         unsigned o_mode;
 599         const char *o_name;
 600
 601         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 602                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 603                 return retval;
 604         }
 605
 606         o_mode = 0;
 607         o_name = NULL;
 608
 609         while (desc.size) {
 610                 unsigned mode;
 611                 const char *name;
 612                 const struct object_id *oid;
 613
 614                 oid = tree_entry_extract(&desc, &name, &mode);
 615
 616                 has_null_sha1 |= is_null_oid(oid);
 617                 has_full_path |= !!strchr(name, '/');
 618                 has_empty_name |= !*name;
 619                 has_dot |= !strcmp(name, ".");
 620                 has_dotdot |= !strcmp(name, "..");
 621                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 622                 has_zero_pad |= *(char *)desc.buffer == '0';
 623
 624                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 625                         if (!S_ISLNK(mode))
 626                                 oidset_insert(&gitmodules_found, oid);
 627                         else
 628                                 retval += report(options, &item->object,
 629                                                  FSCK_MSG_GITMODULES_SYMLINK,
 630                                                  ".gitmodules is a symbolic link");
 631                 }
 632
 633                 if (update_tree_entry_gently(&desc)) {
 634                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 635                         break;
 636                 }
 637
 638                 switch (mode) {
 639                 /*
 640                  * Standard modes..
 641                  */
 642                 case S_IFREG | 0755:
 643                 case S_IFREG | 0644:
 644                 case S_IFLNK:
 645                 case S_IFDIR:
 646                 case S_IFGITLINK:
 647                         break;
 648                 /*
 649                  * This is nonstandard, but we had a few of these
 650                  * early on when we honored the full set of mode
 651                  * bits..
 652                  */
 653                 case S_IFREG | 0664:
 654                         if (!options->strict)
 655                                 break;
 656                         /* fallthrough */
 657                 default:
 658                         has_bad_modes = 1;
 659                 }
 660
 661                 if (o_name) {
 662                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 663                         case TREE_UNORDERED:
 664                                 not_properly_sorted = 1;
 665                                 break;
 666                         case TREE_HAS_DUPS:
 667                                 has_dup_entries = 1;
 668                                 break;
 669                         default:
 670                                 break;
 671                         }
 672                 }
 673
 674                 o_mode = mode;
 675                 o_name = name;
 676         }
 677
 678         if (has_null_sha1)
 679                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 680         if (has_full_path)
 681                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 682         if (has_empty_name)
 683                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 684         if (has_dot)
 685                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 686         if (has_dotdot)
 687                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 688         if (has_dotgit)
 689                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 690         if (has_zero_pad)
 691                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 692         if (has_bad_modes)
 693                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 694         if (has_dup_entries)
 695                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 696         if (not_properly_sorted)
 697                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 698         return retval;
 699 }
 700
 701 static int verify_headers(const void *data, unsigned long size,
 702                           struct object *obj, struct fsck_options *options)
 703 {
 704         const char *buffer = (const char *)data;
 705         unsigned long i;
 706
 707         for (i = 0; i < size; i++) {
 708                 switch (buffer[i]) {
 709                 case '\0':
 710                         return report(options, obj,
 711                                 FSCK_MSG_NUL_IN_HEADER,
 712                                 "unterminated header: NUL at offset %ld", i);
 713                 case '\n':
 714                         if (i + 1 < size && buffer[i + 1] == '\n')
 715                                 return 0;
 716                 }
 717         }
 718
 719         /*
 720          * We did not find double-LF that separates the header
 721          * and the body.  Not having a body is not a crime but
 722          * we do want to see the terminating LF for the last header
 723          * line.
 724          */
 725         if (size && buffer[size - 1] == '\n')
 726                 return 0;
 727
 728         return report(options, obj,
 729                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 730 }
 731
 732 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 733 {
 734         const char *p = *ident;
 735         char *end;
 736
 737         *ident = strchrnul(*ident, '\n');
 738         if (**ident == '\n')
 739                 (*ident)++;
 740
 741         if (*p == '<')
 742                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 743         p += strcspn(p, "<>\n");
 744         if (*p == '>')
 745                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 746         if (*p != '<')
 747                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 748         if (p[-1] != ' ')
 749                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 750         p++;
 751         p += strcspn(p, "<>\n");
 752         if (*p != '>')
 753                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 754         p++;
 755         if (*p != ' ')
 756                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 757         p++;
 758         if (*p == '0' && p[1] != ' ')
 759                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 760         if (date_overflows(parse_timestamp(p, &end, 10)))
 761                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 762         if ((end == p || *end != ' '))
 763                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 764         p = end + 1;
 765         if ((*p != '+' && *p != '-') ||
 766             !isdigit(p[1]) ||
 767             !isdigit(p[2]) ||
 768             !isdigit(p[3]) ||
 769             !isdigit(p[4]) ||
 770             (p[5] != '\n'))
 771                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 772         p += 6;
 773         return 0;
 774 }
 775
 776 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 777         unsigned long size, struct fsck_options *options)
 778 {
 779         struct object_id tree_oid, oid;
 780         struct commit_graft *graft;
 781         unsigned parent_count, parent_line_count = 0, author_count;
 782         int err;
 783         const char *buffer_begin = buffer;
 784         const char *p;
 785
 786         if (verify_headers(buffer, size, &commit->object, options))
 787                 return -1;
 788
 789         if (!skip_prefix(buffer, "tree ", &buffer))
 790                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 791         if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 792                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 793                 if (err)
 794                         return err;
 795         }
 796         buffer = p + 1;
 797         while (skip_prefix(buffer, "parent ", &buffer)) {
 798                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 799                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 800                         if (err)
 801                                 return err;
 802                 }
 803                 buffer = p + 1;
 804                 parent_line_count++;
 805         }
 806         graft = lookup_commit_graft(the_repository, &commit->object.oid);
 807         parent_count = commit_list_count(commit->parents);
 808         if (graft) {
 809                 if (graft->nr_parent == -1 && !parent_count)
 810                         ; /* shallow commit */
 811                 else if (graft->nr_parent != parent_count) {
 812                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 813                         if (err)
 814                                 return err;
 815                 }
 816         } else {
 817                 if (parent_count != parent_line_count) {
 818                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 819                         if (err)
 820                                 return err;
 821                 }
 822         }
 823         author_count = 0;
 824         while (skip_prefix(buffer, "author ", &buffer)) {
 825                 author_count++;
 826                 err = fsck_ident(&buffer, &commit->object, options);
 827                 if (err)
 828                         return err;
 829         }
 830         if (author_count < 1)
 831                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 832         else if (author_count > 1)
 833                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 834         if (err)
 835                 return err;
 836         if (!skip_prefix(buffer, "committer ", &buffer))
 837                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 838         err = fsck_ident(&buffer, &commit->object, options);
 839         if (err)
 840                 return err;
 841         if (!get_commit_tree(commit)) {
 842                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
 843                 if (err)
 844                         return err;
 845         }
 846         if (memchr(buffer_begin, '\0', size)) {
 847                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 848                              "NUL byte in the commit object body");
 849                 if (err)
 850                         return err;
 851         }
 852         return 0;
 853 }
 854
 855 static int fsck_commit(struct commit *commit, const char *data,
 856         unsigned long size, struct fsck_options *options)
 857 {
 858         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 859         int ret = fsck_commit_buffer(commit, buffer, size, options);
 860         if (!data)
 861                 unuse_commit_buffer(commit, buffer);
 862         return ret;
 863 }
 864
 865 static int fsck_tag_buffer(struct tag *tag, const char *data,
 866         unsigned long size, struct fsck_options *options)
 867 {
 868         struct object_id oid;
 869         int ret = 0;
 870         const char *buffer;
 871         char *to_free = NULL, *eol;
 872         struct strbuf sb = STRBUF_INIT;
 873         const char *p;
 874
 875         if (data)
 876                 buffer = data;
 877         else {
 878                 enum object_type type;
 879
 880                 buffer = to_free =
 881                         read_object_file(&tag->object.oid, &type, &size);
 882                 if (!buffer)
 883                         return report(options, &tag->object,
 884                                 FSCK_MSG_MISSING_TAG_OBJECT,
 885                                 "cannot read tag object");
 886
 887                 if (type != OBJ_TAG) {
 888                         ret = report(options, &tag->object,
 889                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 890                                 "expected tag got %s",
 891                             type_name(type));
 892                         goto done;
 893                 }
 894         }
 895
 896         ret = verify_headers(buffer, size, &tag->object, options);
 897         if (ret)
 898                 goto done;
 899
 900         if (!skip_prefix(buffer, "object ", &buffer)) {
 901                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 902                 goto done;
 903         }
 904         if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 905                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 906                 if (ret)
 907                         goto done;
 908         }
 909         buffer = p + 1;
 910
 911         if (!skip_prefix(buffer, "type ", &buffer)) {
 912                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 913                 goto done;
 914         }
 915         eol = strchr(buffer, '\n');
 916         if (!eol) {
 917                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 918                 goto done;
 919         }
 920         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 921                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 922         if (ret)
 923                 goto done;
 924         buffer = eol + 1;
 925
 926         if (!skip_prefix(buffer, "tag ", &buffer)) {
 927                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 928                 goto done;
 929         }
 930         eol = strchr(buffer, '\n');
 931         if (!eol) {
 932                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 933                 goto done;
 934         }
 935         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 936         if (check_refname_format(sb.buf, 0)) {
 937                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 938                            "invalid 'tag' name: %.*s",
 939                            (int)(eol - buffer), buffer);
 940                 if (ret)
 941                         goto done;
 942         }
 943         buffer = eol + 1;
 944
 945         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 946                 /* early tags do not contain 'tagger' lines; warn only */
 947                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 948                 if (ret)
 949                         goto done;
 950         }
 951         else
 952                 ret = fsck_ident(&buffer, &tag->object, options);
 953
 954 done:
 955         strbuf_release(&sb);
 956         free(to_free);
 957         return ret;
 958 }
 959
 960 static int fsck_tag(struct tag *tag, const char *data,
 961         unsigned long size, struct fsck_options *options)
 962 {
 963         struct object *tagged = tag->tagged;
 964
 965         if (!tagged)
 966                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 967
 968         return fsck_tag_buffer(tag, data, size, options);
 969 }
 970
 971 struct fsck_gitmodules_data {
 972         struct object *obj;
 973         struct fsck_options *options;
 974         int ret;
 975 };
 976
 977 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 978 {
 979         struct fsck_gitmodules_data *data = vdata;
 980         const char *subsection, *key;
 981         int subsection_len;
 982         char *name;
 983
 984         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 985             !subsection)
 986                 return 0;
 987
 988         name = xmemdupz(subsection, subsection_len);
 989         if (check_submodule_name(name) < 0)
 990                 data->ret |= report(data->options, data->obj,
 991                                     FSCK_MSG_GITMODULES_NAME,
 992                                     "disallowed submodule name: %s",
 993                                     name);
 994         free(name);
 995
 996         return 0;
 997 }
 998
 999 static int fsck_blob(struct blob *blob, const char *buf,
1000                      unsigned long size, struct fsck_options *options)
1001 {
1002         struct fsck_gitmodules_data data;
1003         struct config_options config_opts = { 0 };
1004
1005         if (!oidset_contains(&gitmodules_found, &blob->object.oid))
1006                 return 0;
1007         oidset_insert(&gitmodules_done, &blob->object.oid);
1008
1009         if (object_on_skiplist(options, &blob->object))
1010                 return 0;
1011
1012         if (!buf) {
1013                 /*
1014                  * A missing buffer here is a sign that the caller found the
1015                  * blob too gigantic to load into memory. Let's just consider
1016                  * that an error.
1017                  */
1018                 return report(options, &blob->object,
1019                               FSCK_MSG_GITMODULES_LARGE,
1020                               ".gitmodules too large to parse");
1021         }
1022
1023         data.obj = &blob->object;
1024         data.options = options;
1025         data.ret = 0;
1026         config_opts.error_action = CONFIG_ERROR_SILENT;
1027         if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1028                                 ".gitmodules", buf, size, &data, &config_opts))
1029                 data.ret |= report(options, &blob->object,
1030                                    FSCK_MSG_GITMODULES_PARSE,
1031                                    "could not parse gitmodules blob");
1032
1033         return data.ret;
1034 }
1035
1036 int fsck_object(struct object *obj, void *data, unsigned long size,
1037         struct fsck_options *options)
1038 {
1039         if (!obj)
1040                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1041
1042         if (obj->type == OBJ_BLOB)
1043                 return fsck_blob((struct blob *)obj, data, size, options);
1044         if (obj->type == OBJ_TREE)
1045                 return fsck_tree((struct tree *) obj, options);
1046         if (obj->type == OBJ_COMMIT)
1047                 return fsck_commit((struct commit *) obj, (const char *) data,
1048                         size, options);
1049         if (obj->type == OBJ_TAG)
1050                 return fsck_tag((struct tag *) obj, (const char *) data,
1051                         size, options);
1052
1053         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1054                           obj->type);
1055 }
1056
1057 int fsck_error_function(struct fsck_options *o,
1058         struct object *obj, int msg_type, const char *message)
1059 {
1060         if (msg_type == FSCK_WARN) {
1061                 warning("object %s: %s", describe_object(o, obj), message);
1062                 return 0;
1063         }
1064         error("object %s: %s", describe_object(o, obj), message);
1065         return 1;
1066 }
1067
1068 int fsck_finish(struct fsck_options *options)
1069 {
1070         int ret = 0;
1071         struct oidset_iter iter;
1072         const struct object_id *oid;
1073
1074         oidset_iter_init(&gitmodules_found, &iter);
1075         while ((oid = oidset_iter_next(&iter))) {
1076                 struct blob *blob;
1077                 enum object_type type;
1078                 unsigned long size;
1079                 char *buf;
1080
1081                 if (oidset_contains(&gitmodules_done, oid))
1082                         continue;
1083
1084                 blob = lookup_blob(oid);
1085                 if (!blob) {
1086                         struct object *obj = lookup_unknown_object(oid->hash);
1087                         ret |= report(options, obj,
1088                                       FSCK_MSG_GITMODULES_BLOB,
1089                                       "non-blob found at .gitmodules");
1090                         continue;
1091                 }
1092
1093                 buf = read_object_file(oid, &type, &size);
1094                 if (!buf) {
1095                         if (is_promisor_object(&blob->object.oid))
1096                                 continue;
1097                         ret |= report(options, &blob->object,
1098                                       FSCK_MSG_GITMODULES_MISSING,
1099                                       "unable to read .gitmodules blob");
1100                         continue;
1101                 }
1102
1103                 if (type == OBJ_BLOB)
1104                         ret |= fsck_blob(blob, buf, size, options);
1105                 else
1106                         ret |= report(options, &blob->object,
1107                                       FSCK_MSG_GITMODULES_BLOB,
1108                                       "non-blob found at .gitmodules");
1109                 free(buf);
1110         }
1111
1112
1113         oidset_clear(&gitmodules_found);
1114         oidset_clear(&gitmodules_done);
1115         return ret;
1116 }