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 "sha1-array.h"
  14 #include "decorate.h"
  15 #include "oidset.h"
  16 #include "packfile.h"
  17 #include "submodule-config.h"
  18 #include "config.h"
  19 #include "help.h"
  20
  21 static struct oidset gitmodules_found = OIDSET_INIT;
  22 static struct oidset gitmodules_done = OIDSET_INIT;
  23
  24 #define FSCK_FATAL -1
  25 #define FSCK_INFO -2
  26
  27 #define FOREACH_MSG_ID(FUNC) \
  28         /* fatal errors */ \
  29         FUNC(NUL_IN_HEADER, FATAL) \
  30         FUNC(UNTERMINATED_HEADER, FATAL) \
  31         /* errors */ \
  32         FUNC(BAD_DATE, ERROR) \
  33         FUNC(BAD_DATE_OVERFLOW, ERROR) \
  34         FUNC(BAD_EMAIL, ERROR) \
  35         FUNC(BAD_NAME, ERROR) \
  36         FUNC(BAD_OBJECT_SHA1, ERROR) \
  37         FUNC(BAD_PARENT_SHA1, ERROR) \
  38         FUNC(BAD_TAG_OBJECT, ERROR) \
  39         FUNC(BAD_TIMEZONE, ERROR) \
  40         FUNC(BAD_TREE, ERROR) \
  41         FUNC(BAD_TREE_SHA1, ERROR) \
  42         FUNC(BAD_TYPE, ERROR) \
  43         FUNC(DUPLICATE_ENTRIES, ERROR) \
  44         FUNC(MISSING_AUTHOR, ERROR) \
  45         FUNC(MISSING_COMMITTER, ERROR) \
  46         FUNC(MISSING_EMAIL, ERROR) \
  47         FUNC(MISSING_GRAFT, ERROR) \
  48         FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  49         FUNC(MISSING_OBJECT, ERROR) \
  50         FUNC(MISSING_PARENT, ERROR) \
  51         FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  52         FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  53         FUNC(MISSING_TAG, ERROR) \
  54         FUNC(MISSING_TAG_ENTRY, ERROR) \
  55         FUNC(MISSING_TAG_OBJECT, ERROR) \
  56         FUNC(MISSING_TREE, ERROR) \
  57         FUNC(MISSING_TREE_OBJECT, ERROR) \
  58         FUNC(MISSING_TYPE, ERROR) \
  59         FUNC(MISSING_TYPE_ENTRY, ERROR) \
  60         FUNC(MULTIPLE_AUTHORS, ERROR) \
  61         FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  62         FUNC(TREE_NOT_SORTED, ERROR) \
  63         FUNC(UNKNOWN_TYPE, ERROR) \
  64         FUNC(ZERO_PADDED_DATE, ERROR) \
  65         FUNC(GITMODULES_MISSING, ERROR) \
  66         FUNC(GITMODULES_BLOB, ERROR) \
  67         FUNC(GITMODULES_LARGE, ERROR) \
  68         FUNC(GITMODULES_NAME, ERROR) \
  69         FUNC(GITMODULES_SYMLINK, ERROR) \
  70         FUNC(GITMODULES_URL, ERROR) \
  71         FUNC(GITMODULES_PATH, ERROR) \
  72         /* warnings */ \
  73         FUNC(BAD_FILEMODE, WARN) \
  74         FUNC(EMPTY_NAME, WARN) \
  75         FUNC(FULL_PATHNAME, WARN) \
  76         FUNC(HAS_DOT, WARN) \
  77         FUNC(HAS_DOTDOT, WARN) \
  78         FUNC(HAS_DOTGIT, WARN) \
  79         FUNC(NULL_SHA1, WARN) \
  80         FUNC(ZERO_PADDED_FILEMODE, WARN) \
  81         FUNC(NUL_IN_COMMIT, WARN) \
  82         /* infos (reported as warnings, but ignored by default) */ \
  83         FUNC(GITMODULES_PARSE, INFO) \
  84         FUNC(BAD_TAG_NAME, INFO) \
  85         FUNC(MISSING_TAGGER_ENTRY, INFO)
  86
  87 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
  88 enum fsck_msg_id {
  89         FOREACH_MSG_ID(MSG_ID)
  90         FSCK_MSG_MAX
  91 };
  92 #undef MSG_ID
  93
  94 #define STR(x) #x
  95 #define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
  96 static struct {
  97         const char *id_string;
  98         const char *downcased;
  99         const char *camelcased;
 100         int msg_type;
 101 } msg_id_info[FSCK_MSG_MAX + 1] = {
 102         FOREACH_MSG_ID(MSG_ID)
 103         { NULL, NULL, NULL, -1 }
 104 };
 105 #undef MSG_ID
 106
 107 static void prepare_msg_ids(void)
 108 {
 109         int i;
 110
 111         if (msg_id_info[0].downcased)
 112                 return;
 113
 114         /* convert id_string to lower case, without underscores. */
 115         for (i = 0; i < FSCK_MSG_MAX; i++) {
 116                 const char *p = msg_id_info[i].id_string;
 117                 int len = strlen(p);
 118                 char *q = xmalloc(len);
 119
 120                 msg_id_info[i].downcased = q;
 121                 while (*p)
 122                         if (*p == '_')
 123                                 p++;
 124                         else
 125                                 *(q)++ = tolower(*(p)++);
 126                 *q = '\0';
 127
 128                 p = msg_id_info[i].id_string;
 129                 q = xmalloc(len);
 130                 msg_id_info[i].camelcased = q;
 131                 while (*p) {
 132                         if (*p == '_') {
 133                                 p++;
 134                                 if (*p)
 135                                         *q++ = *p++;
 136                         } else {
 137                                 *q++ = tolower(*p++);
 138                         }
 139                 }
 140                 *q = '\0';
 141         }
 142 }
 143
 144 static int parse_msg_id(const char *text)
 145 {
 146         int i;
 147
 148         prepare_msg_ids();
 149
 150         for (i = 0; i < FSCK_MSG_MAX; i++)
 151                 if (!strcmp(text, msg_id_info[i].downcased))
 152                         return i;
 153
 154         return -1;
 155 }
 156
 157 void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
 158 {
 159         int i;
 160
 161         prepare_msg_ids();
 162
 163         for (i = 0; i < FSCK_MSG_MAX; i++)
 164                 list_config_item(list, prefix, msg_id_info[i].camelcased);
 165 }
 166
 167 static int fsck_msg_type(enum fsck_msg_id msg_id,
 168         struct fsck_options *options)
 169 {
 170         int msg_type;
 171
 172         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 173
 174         if (options->msg_type)
 175                 msg_type = options->msg_type[msg_id];
 176         else {
 177                 msg_type = msg_id_info[msg_id].msg_type;
 178                 if (options->strict && msg_type == FSCK_WARN)
 179                         msg_type = FSCK_ERROR;
 180         }
 181
 182         return msg_type;
 183 }
 184
 185 static void init_skiplist(struct fsck_options *options, const char *path)
 186 {
 187         static struct oid_array skiplist = OID_ARRAY_INIT;
 188         int sorted, fd;
 189         char buffer[GIT_MAX_HEXSZ + 1];
 190         struct object_id oid;
 191
 192         if (options->skiplist)
 193                 sorted = options->skiplist->sorted;
 194         else {
 195                 sorted = 1;
 196                 options->skiplist = &skiplist;
 197         }
 198
 199         fd = open(path, O_RDONLY);
 200         if (fd < 0)
 201                 die("Could not open skip list: %s", path);
 202         for (;;) {
 203                 const char *p;
 204                 int result = read_in_full(fd, buffer, sizeof(buffer));
 205                 if (result < 0)
 206                         die_errno("Could not read '%s'", path);
 207                 if (!result)
 208                         break;
 209                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 210                         die("Invalid SHA-1: %s", buffer);
 211                 oid_array_append(&skiplist, &oid);
 212                 if (sorted && skiplist.nr > 1 &&
 213                                 oidcmp(&skiplist.oid[skiplist.nr - 2],
 214                                        &oid) > 0)
 215                         sorted = 0;
 216         }
 217         close(fd);
 218
 219         if (sorted)
 220                 skiplist.sorted = 1;
 221 }
 222
 223 static int parse_msg_type(const char *str)
 224 {
 225         if (!strcmp(str, "error"))
 226                 return FSCK_ERROR;
 227         else if (!strcmp(str, "warn"))
 228                 return FSCK_WARN;
 229         else if (!strcmp(str, "ignore"))
 230                 return FSCK_IGNORE;
 231         else
 232                 die("Unknown fsck message type: '%s'", str);
 233 }
 234
 235 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 236 {
 237         if (parse_msg_id(msg_id) < 0)
 238                 return 0;
 239         parse_msg_type(msg_type);
 240         return 1;
 241 }
 242
 243 void fsck_set_msg_type(struct fsck_options *options,
 244                 const char *msg_id, const char *msg_type)
 245 {
 246         int id = parse_msg_id(msg_id), type;
 247
 248         if (id < 0)
 249                 die("Unhandled message id: %s", msg_id);
 250         type = parse_msg_type(msg_type);
 251
 252         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 253                 die("Cannot demote %s to %s", msg_id, msg_type);
 254
 255         if (!options->msg_type) {
 256                 int i;
 257                 int *msg_type;
 258                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 259                 for (i = 0; i < FSCK_MSG_MAX; i++)
 260                         msg_type[i] = fsck_msg_type(i, options);
 261                 options->msg_type = msg_type;
 262         }
 263
 264         options->msg_type[id] = type;
 265 }
 266
 267 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 268 {
 269         char *buf = xstrdup(values), *to_free = buf;
 270         int done = 0;
 271
 272         while (!done) {
 273                 int len = strcspn(buf, " ,|"), equal;
 274
 275                 done = !buf[len];
 276                 if (!len) {
 277                         buf++;
 278                         continue;
 279                 }
 280                 buf[len] = '\0';
 281
 282                 for (equal = 0;
 283                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 284                      equal++)
 285                         buf[equal] = tolower(buf[equal]);
 286                 buf[equal] = '\0';
 287
 288                 if (!strcmp(buf, "skiplist")) {
 289                         if (equal == len)
 290                                 die("skiplist requires a path");
 291                         init_skiplist(options, buf + equal + 1);
 292                         buf += len + 1;
 293                         continue;
 294                 }
 295
 296                 if (equal == len)
 297                         die("Missing '=': '%s'", buf);
 298
 299                 fsck_set_msg_type(options, buf, buf + equal + 1);
 300                 buf += len + 1;
 301         }
 302         free(to_free);
 303 }
 304
 305 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 306 {
 307         for (;;) {
 308                 char c = *(msg_id)++;
 309
 310                 if (!c)
 311                         break;
 312                 if (c != '_')
 313                         strbuf_addch(sb, tolower(c));
 314                 else {
 315                         assert(*msg_id);
 316                         strbuf_addch(sb, *(msg_id)++);
 317                 }
 318         }
 319
 320         strbuf_addstr(sb, ": ");
 321 }
 322
 323 static int object_on_skiplist(struct fsck_options *opts, struct object *obj)
 324 {
 325         if (opts && opts->skiplist && obj)
 326                 return oid_array_lookup(opts->skiplist, &obj->oid) >= 0;
 327         return 0;
 328 }
 329
 330 __attribute__((format (printf, 4, 5)))
 331 static int report(struct fsck_options *options, struct object *object,
 332         enum fsck_msg_id id, const char *fmt, ...)
 333 {
 334         va_list ap;
 335         struct strbuf sb = STRBUF_INIT;
 336         int msg_type = fsck_msg_type(id, options), result;
 337
 338         if (msg_type == FSCK_IGNORE)
 339                 return 0;
 340
 341         if (object_on_skiplist(options, object))
 342                 return 0;
 343
 344         if (msg_type == FSCK_FATAL)
 345                 msg_type = FSCK_ERROR;
 346         else if (msg_type == FSCK_INFO)
 347                 msg_type = FSCK_WARN;
 348
 349         append_msg_id(&sb, msg_id_info[id].id_string);
 350
 351         va_start(ap, fmt);
 352         strbuf_vaddf(&sb, fmt, ap);
 353         result = options->error_func(options, object, msg_type, sb.buf);
 354         strbuf_release(&sb);
 355         va_end(ap);
 356
 357         return result;
 358 }
 359
 360 static char *get_object_name(struct fsck_options *options, struct object *obj)
 361 {
 362         if (!options->object_names)
 363                 return NULL;
 364         return lookup_decoration(options->object_names, obj);
 365 }
 366
 367 static void put_object_name(struct fsck_options *options, struct object *obj,
 368         const char *fmt, ...)
 369 {
 370         va_list ap;
 371         struct strbuf buf = STRBUF_INIT;
 372         char *existing;
 373
 374         if (!options->object_names)
 375                 return;
 376         existing = lookup_decoration(options->object_names, obj);
 377         if (existing)
 378                 return;
 379         va_start(ap, fmt);
 380         strbuf_vaddf(&buf, fmt, ap);
 381         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 382         va_end(ap);
 383 }
 384
 385 static const char *describe_object(struct fsck_options *o, struct object *obj)
 386 {
 387         static struct strbuf buf = STRBUF_INIT;
 388         char *name;
 389
 390         strbuf_reset(&buf);
 391         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 392         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 393                 strbuf_addf(&buf, " (%s)", name);
 394
 395         return buf.buf;
 396 }
 397
 398 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 399 {
 400         struct tree_desc desc;
 401         struct name_entry entry;
 402         int res = 0;
 403         const char *name;
 404
 405         if (parse_tree(tree))
 406                 return -1;
 407
 408         name = get_object_name(options, &tree->object);
 409         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 410                 return -1;
 411         while (tree_entry_gently(&desc, &entry)) {
 412                 struct object *obj;
 413                 int result;
 414
 415                 if (S_ISGITLINK(entry.mode))
 416                         continue;
 417
 418                 if (S_ISDIR(entry.mode)) {
 419                         obj = (struct object *)lookup_tree(the_repository, entry.oid);
 420                         if (name && obj)
 421                                 put_object_name(options, obj, "%s%s/", name,
 422                                         entry.path);
 423                         result = options->walk(obj, OBJ_TREE, data, options);
 424                 }
 425                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 426                         obj = (struct object *)lookup_blob(the_repository, entry.oid);
 427                         if (name && obj)
 428                                 put_object_name(options, obj, "%s%s", name,
 429                                         entry.path);
 430                         result = options->walk(obj, OBJ_BLOB, data, options);
 431                 }
 432                 else {
 433                         result = error("in tree %s: entry %s has bad mode %.6o",
 434                                         describe_object(options, &tree->object), entry.path, entry.mode);
 435                 }
 436                 if (result < 0)
 437                         return result;
 438                 if (!res)
 439                         res = result;
 440         }
 441         return res;
 442 }
 443
 444 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 445 {
 446         int counter = 0, generation = 0, name_prefix_len = 0;
 447         struct commit_list *parents;
 448         int res;
 449         int result;
 450         const char *name;
 451
 452         if (parse_commit(commit))
 453                 return -1;
 454
 455         name = get_object_name(options, &commit->object);
 456         if (name)
 457                 put_object_name(options, &get_commit_tree(commit)->object,
 458                                 "%s:", name);
 459
 460         result = options->walk((struct object *)get_commit_tree(commit),
 461                                OBJ_TREE, data, options);
 462         if (result < 0)
 463                 return result;
 464         res = result;
 465
 466         parents = commit->parents;
 467         if (name && parents) {
 468                 int len = strlen(name), power;
 469
 470                 if (len && name[len - 1] == '^') {
 471                         generation = 1;
 472                         name_prefix_len = len - 1;
 473                 }
 474                 else { /* parse ~<generation> suffix */
 475                         for (generation = 0, power = 1;
 476                              len && isdigit(name[len - 1]);
 477                              power *= 10)
 478                                 generation += power * (name[--len] - '0');
 479                         if (power > 1 && len && name[len - 1] == '~')
 480                                 name_prefix_len = len - 1;
 481                 }
 482         }
 483
 484         while (parents) {
 485                 if (name) {
 486                         struct object *obj = &parents->item->object;
 487
 488                         if (++counter > 1)
 489                                 put_object_name(options, obj, "%s^%d",
 490                                         name, counter);
 491                         else if (generation > 0)
 492                                 put_object_name(options, obj, "%.*s~%d",
 493                                         name_prefix_len, name, generation + 1);
 494                         else
 495                                 put_object_name(options, obj, "%s^", name);
 496                 }
 497                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 498                 if (result < 0)
 499                         return result;
 500                 if (!res)
 501                         res = result;
 502                 parents = parents->next;
 503         }
 504         return res;
 505 }
 506
 507 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 508 {
 509         char *name = get_object_name(options, &tag->object);
 510
 511         if (parse_tag(tag))
 512                 return -1;
 513         if (name)
 514                 put_object_name(options, tag->tagged, "%s", name);
 515         return options->walk(tag->tagged, OBJ_ANY, data, options);
 516 }
 517
 518 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 519 {
 520         if (!obj)
 521                 return -1;
 522
 523         if (obj->type == OBJ_NONE)
 524                 parse_object(the_repository, &obj->oid);
 525
 526         switch (obj->type) {
 527         case OBJ_BLOB:
 528                 return 0;
 529         case OBJ_TREE:
 530                 return fsck_walk_tree((struct tree *)obj, data, options);
 531         case OBJ_COMMIT:
 532                 return fsck_walk_commit((struct commit *)obj, data, options);
 533         case OBJ_TAG:
 534                 return fsck_walk_tag((struct tag *)obj, data, options);
 535         default:
 536                 error("Unknown object type for %s", describe_object(options, obj));
 537                 return -1;
 538         }
 539 }
 540
 541 /*
 542  * The entries in a tree are ordered in the _path_ order,
 543  * which means that a directory entry is ordered by adding
 544  * a slash to the end of it.
 545  *
 546  * So a directory called "a" is ordered _after_ a file
 547  * called "a.c", because "a/" sorts after "a.c".
 548  */
 549 #define TREE_UNORDERED (-1)
 550 #define TREE_HAS_DUPS  (-2)
 551
 552 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 553 {
 554         int len1 = strlen(name1);
 555         int len2 = strlen(name2);
 556         int len = len1 < len2 ? len1 : len2;
 557         unsigned char c1, c2;
 558         int cmp;
 559
 560         cmp = memcmp(name1, name2, len);
 561         if (cmp < 0)
 562                 return 0;
 563         if (cmp > 0)
 564                 return TREE_UNORDERED;
 565
 566         /*
 567          * Ok, the first <len> characters are the same.
 568          * Now we need to order the next one, but turn
 569          * a '\0' into a '/' for a directory entry.
 570          */
 571         c1 = name1[len];
 572         c2 = name2[len];
 573         if (!c1 && !c2)
 574                 /*
 575                  * git-write-tree used to write out a nonsense tree that has
 576                  * entries with the same name, one blob and one tree.  Make
 577                  * sure we do not have duplicate entries.
 578                  */
 579                 return TREE_HAS_DUPS;
 580         if (!c1 && S_ISDIR(mode1))
 581                 c1 = '/';
 582         if (!c2 && S_ISDIR(mode2))
 583                 c2 = '/';
 584         return c1 < c2 ? 0 : TREE_UNORDERED;
 585 }
 586
 587 static int fsck_tree(struct tree *item, struct fsck_options *options)
 588 {
 589         int retval = 0;
 590         int has_null_sha1 = 0;
 591         int has_full_path = 0;
 592         int has_empty_name = 0;
 593         int has_dot = 0;
 594         int has_dotdot = 0;
 595         int has_dotgit = 0;
 596         int has_zero_pad = 0;
 597         int has_bad_modes = 0;
 598         int has_dup_entries = 0;
 599         int not_properly_sorted = 0;
 600         struct tree_desc desc;
 601         unsigned o_mode;
 602         const char *o_name;
 603
 604         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 605                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 606                 return retval;
 607         }
 608
 609         o_mode = 0;
 610         o_name = NULL;
 611
 612         while (desc.size) {
 613                 unsigned mode;
 614                 const char *name;
 615                 const struct object_id *oid;
 616
 617                 oid = tree_entry_extract(&desc, &name, &mode);
 618
 619                 has_null_sha1 |= is_null_oid(oid);
 620                 has_full_path |= !!strchr(name, '/');
 621                 has_empty_name |= !*name;
 622                 has_dot |= !strcmp(name, ".");
 623                 has_dotdot |= !strcmp(name, "..");
 624                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 625                 has_zero_pad |= *(char *)desc.buffer == '0';
 626
 627                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 628                         if (!S_ISLNK(mode))
 629                                 oidset_insert(&gitmodules_found, oid);
 630                         else
 631                                 retval += report(options, &item->object,
 632                                                  FSCK_MSG_GITMODULES_SYMLINK,
 633                                                  ".gitmodules is a symbolic link");
 634                 }
 635
 636                 if (update_tree_entry_gently(&desc)) {
 637                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 638                         break;
 639                 }
 640
 641                 switch (mode) {
 642                 /*
 643                  * Standard modes..
 644                  */
 645                 case S_IFREG | 0755:
 646                 case S_IFREG | 0644:
 647                 case S_IFLNK:
 648                 case S_IFDIR:
 649                 case S_IFGITLINK:
 650                         break;
 651                 /*
 652                  * This is nonstandard, but we had a few of these
 653                  * early on when we honored the full set of mode
 654                  * bits..
 655                  */
 656                 case S_IFREG | 0664:
 657                         if (!options->strict)
 658                                 break;
 659                         /* fallthrough */
 660                 default:
 661                         has_bad_modes = 1;
 662                 }
 663
 664                 if (o_name) {
 665                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 666                         case TREE_UNORDERED:
 667                                 not_properly_sorted = 1;
 668                                 break;
 669                         case TREE_HAS_DUPS:
 670                                 has_dup_entries = 1;
 671                                 break;
 672                         default:
 673                                 break;
 674                         }
 675                 }
 676
 677                 o_mode = mode;
 678                 o_name = name;
 679         }
 680
 681         if (has_null_sha1)
 682                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 683         if (has_full_path)
 684                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 685         if (has_empty_name)
 686                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 687         if (has_dot)
 688                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 689         if (has_dotdot)
 690                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 691         if (has_dotgit)
 692                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 693         if (has_zero_pad)
 694                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 695         if (has_bad_modes)
 696                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 697         if (has_dup_entries)
 698                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 699         if (not_properly_sorted)
 700                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 701         return retval;
 702 }
 703
 704 static int verify_headers(const void *data, unsigned long size,
 705                           struct object *obj, struct fsck_options *options)
 706 {
 707         const char *buffer = (const char *)data;
 708         unsigned long i;
 709
 710         for (i = 0; i < size; i++) {
 711                 switch (buffer[i]) {
 712                 case '\0':
 713                         return report(options, obj,
 714                                 FSCK_MSG_NUL_IN_HEADER,
 715                                 "unterminated header: NUL at offset %ld", i);
 716                 case '\n':
 717                         if (i + 1 < size && buffer[i + 1] == '\n')
 718                                 return 0;
 719                 }
 720         }
 721
 722         /*
 723          * We did not find double-LF that separates the header
 724          * and the body.  Not having a body is not a crime but
 725          * we do want to see the terminating LF for the last header
 726          * line.
 727          */
 728         if (size && buffer[size - 1] == '\n')
 729                 return 0;
 730
 731         return report(options, obj,
 732                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 733 }
 734
 735 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 736 {
 737         const char *p = *ident;
 738         char *end;
 739
 740         *ident = strchrnul(*ident, '\n');
 741         if (**ident == '\n')
 742                 (*ident)++;
 743
 744         if (*p == '<')
 745                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 746         p += strcspn(p, "<>\n");
 747         if (*p == '>')
 748                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 749         if (*p != '<')
 750                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 751         if (p[-1] != ' ')
 752                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 753         p++;
 754         p += strcspn(p, "<>\n");
 755         if (*p != '>')
 756                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 757         p++;
 758         if (*p != ' ')
 759                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 760         p++;
 761         if (*p == '0' && p[1] != ' ')
 762                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 763         if (date_overflows(parse_timestamp(p, &end, 10)))
 764                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 765         if ((end == p || *end != ' '))
 766                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 767         p = end + 1;
 768         if ((*p != '+' && *p != '-') ||
 769             !isdigit(p[1]) ||
 770             !isdigit(p[2]) ||
 771             !isdigit(p[3]) ||
 772             !isdigit(p[4]) ||
 773             (p[5] != '\n'))
 774                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 775         p += 6;
 776         return 0;
 777 }
 778
 779 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 780         unsigned long size, struct fsck_options *options)
 781 {
 782         struct object_id tree_oid, oid;
 783         struct commit_graft *graft;
 784         unsigned parent_count, parent_line_count = 0, author_count;
 785         int err;
 786         const char *buffer_begin = buffer;
 787         const char *p;
 788
 789         if (verify_headers(buffer, size, &commit->object, options))
 790                 return -1;
 791
 792         if (!skip_prefix(buffer, "tree ", &buffer))
 793                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 794         if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 795                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 796                 if (err)
 797                         return err;
 798         }
 799         buffer = p + 1;
 800         while (skip_prefix(buffer, "parent ", &buffer)) {
 801                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 802                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 803                         if (err)
 804                                 return err;
 805                 }
 806                 buffer = p + 1;
 807                 parent_line_count++;
 808         }
 809         graft = lookup_commit_graft(the_repository, &commit->object.oid);
 810         parent_count = commit_list_count(commit->parents);
 811         if (graft) {
 812                 if (graft->nr_parent == -1 && !parent_count)
 813                         ; /* shallow commit */
 814                 else if (graft->nr_parent != parent_count) {
 815                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 816                         if (err)
 817                                 return err;
 818                 }
 819         } else {
 820                 if (parent_count != parent_line_count) {
 821                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 822                         if (err)
 823                                 return err;
 824                 }
 825         }
 826         author_count = 0;
 827         while (skip_prefix(buffer, "author ", &buffer)) {
 828                 author_count++;
 829                 err = fsck_ident(&buffer, &commit->object, options);
 830                 if (err)
 831                         return err;
 832         }
 833         if (author_count < 1)
 834                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 835         else if (author_count > 1)
 836                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 837         if (err)
 838                 return err;
 839         if (!skip_prefix(buffer, "committer ", &buffer))
 840                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 841         err = fsck_ident(&buffer, &commit->object, options);
 842         if (err)
 843                 return err;
 844         if (!get_commit_tree(commit)) {
 845                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
 846                 if (err)
 847                         return err;
 848         }
 849         if (memchr(buffer_begin, '\0', size)) {
 850                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 851                              "NUL byte in the commit object body");
 852                 if (err)
 853                         return err;
 854         }
 855         return 0;
 856 }
 857
 858 static int fsck_commit(struct commit *commit, const char *data,
 859         unsigned long size, struct fsck_options *options)
 860 {
 861         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 862         int ret = fsck_commit_buffer(commit, buffer, size, options);
 863         if (!data)
 864                 unuse_commit_buffer(commit, buffer);
 865         return ret;
 866 }
 867
 868 static int fsck_tag_buffer(struct tag *tag, const char *data,
 869         unsigned long size, struct fsck_options *options)
 870 {
 871         struct object_id oid;
 872         int ret = 0;
 873         const char *buffer;
 874         char *to_free = NULL, *eol;
 875         struct strbuf sb = STRBUF_INIT;
 876         const char *p;
 877
 878         if (data)
 879                 buffer = data;
 880         else {
 881                 enum object_type type;
 882
 883                 buffer = to_free =
 884                         read_object_file(&tag->object.oid, &type, &size);
 885                 if (!buffer)
 886                         return report(options, &tag->object,
 887                                 FSCK_MSG_MISSING_TAG_OBJECT,
 888                                 "cannot read tag object");
 889
 890                 if (type != OBJ_TAG) {
 891                         ret = report(options, &tag->object,
 892                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 893                                 "expected tag got %s",
 894                             type_name(type));
 895                         goto done;
 896                 }
 897         }
 898
 899         ret = verify_headers(buffer, size, &tag->object, options);
 900         if (ret)
 901                 goto done;
 902
 903         if (!skip_prefix(buffer, "object ", &buffer)) {
 904                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 905                 goto done;
 906         }
 907         if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 908                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 909                 if (ret)
 910                         goto done;
 911         }
 912         buffer = p + 1;
 913
 914         if (!skip_prefix(buffer, "type ", &buffer)) {
 915                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 916                 goto done;
 917         }
 918         eol = strchr(buffer, '\n');
 919         if (!eol) {
 920                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 921                 goto done;
 922         }
 923         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 924                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 925         if (ret)
 926                 goto done;
 927         buffer = eol + 1;
 928
 929         if (!skip_prefix(buffer, "tag ", &buffer)) {
 930                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 931                 goto done;
 932         }
 933         eol = strchr(buffer, '\n');
 934         if (!eol) {
 935                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 936                 goto done;
 937         }
 938         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 939         if (check_refname_format(sb.buf, 0)) {
 940                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 941                            "invalid 'tag' name: %.*s",
 942                            (int)(eol - buffer), buffer);
 943                 if (ret)
 944                         goto done;
 945         }
 946         buffer = eol + 1;
 947
 948         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 949                 /* early tags do not contain 'tagger' lines; warn only */
 950                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 951                 if (ret)
 952                         goto done;
 953         }
 954         else
 955                 ret = fsck_ident(&buffer, &tag->object, options);
 956
 957 done:
 958         strbuf_release(&sb);
 959         free(to_free);
 960         return ret;
 961 }
 962
 963 static int fsck_tag(struct tag *tag, const char *data,
 964         unsigned long size, struct fsck_options *options)
 965 {
 966         struct object *tagged = tag->tagged;
 967
 968         if (!tagged)
 969                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 970
 971         return fsck_tag_buffer(tag, data, size, options);
 972 }
 973
 974 struct fsck_gitmodules_data {
 975         struct object *obj;
 976         struct fsck_options *options;
 977         int ret;
 978 };
 979
 980 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 981 {
 982         struct fsck_gitmodules_data *data = vdata;
 983         const char *subsection, *key;
 984         int subsection_len;
 985         char *name;
 986
 987         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 988             !subsection)
 989                 return 0;
 990
 991         name = xmemdupz(subsection, subsection_len);
 992         if (check_submodule_name(name) < 0)
 993                 data->ret |= report(data->options, data->obj,
 994                                     FSCK_MSG_GITMODULES_NAME,
 995                                     "disallowed submodule name: %s",
 996                                     name);
 997         if (!strcmp(key, "url") && value &&
 998             looks_like_command_line_option(value))
 999                 data->ret |= report(data->options, data->obj,
1000                                     FSCK_MSG_GITMODULES_URL,
1001                                     "disallowed submodule url: %s",
1002                                     value);
1003         if (!strcmp(key, "path") && value &&
1004             looks_like_command_line_option(value))
1005                 data->ret |= report(data->options, data->obj,
1006                                     FSCK_MSG_GITMODULES_PATH,
1007                                     "disallowed submodule path: %s",
1008                                     value);
1009         free(name);
1010
1011         return 0;
1012 }
1013
1014 static int fsck_blob(struct blob *blob, const char *buf,
1015                      unsigned long size, struct fsck_options *options)
1016 {
1017         struct fsck_gitmodules_data data;
1018         struct config_options config_opts = { 0 };
1019
1020         if (!oidset_contains(&gitmodules_found, &blob->object.oid))
1021                 return 0;
1022         oidset_insert(&gitmodules_done, &blob->object.oid);
1023
1024         if (object_on_skiplist(options, &blob->object))
1025                 return 0;
1026
1027         if (!buf) {
1028                 /*
1029                  * A missing buffer here is a sign that the caller found the
1030                  * blob too gigantic to load into memory. Let's just consider
1031                  * that an error.
1032                  */
1033                 return report(options, &blob->object,
1034                               FSCK_MSG_GITMODULES_LARGE,
1035                               ".gitmodules too large to parse");
1036         }
1037
1038         data.obj = &blob->object;
1039         data.options = options;
1040         data.ret = 0;
1041         config_opts.error_action = CONFIG_ERROR_SILENT;
1042         if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1043                                 ".gitmodules", buf, size, &data, &config_opts))
1044                 data.ret |= report(options, &blob->object,
1045                                    FSCK_MSG_GITMODULES_PARSE,
1046                                    "could not parse gitmodules blob");
1047
1048         return data.ret;
1049 }
1050
1051 int fsck_object(struct object *obj, void *data, unsigned long size,
1052         struct fsck_options *options)
1053 {
1054         if (!obj)
1055                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1056
1057         if (obj->type == OBJ_BLOB)
1058                 return fsck_blob((struct blob *)obj, data, size, options);
1059         if (obj->type == OBJ_TREE)
1060                 return fsck_tree((struct tree *) obj, options);
1061         if (obj->type == OBJ_COMMIT)
1062                 return fsck_commit((struct commit *) obj, (const char *) data,
1063                         size, options);
1064         if (obj->type == OBJ_TAG)
1065                 return fsck_tag((struct tag *) obj, (const char *) data,
1066                         size, options);
1067
1068         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1069                           obj->type);
1070 }
1071
1072 int fsck_error_function(struct fsck_options *o,
1073         struct object *obj, int msg_type, const char *message)
1074 {
1075         if (msg_type == FSCK_WARN) {
1076                 warning("object %s: %s", describe_object(o, obj), message);
1077                 return 0;
1078         }
1079         error("object %s: %s", describe_object(o, obj), message);
1080         return 1;
1081 }
1082
1083 int fsck_finish(struct fsck_options *options)
1084 {
1085         int ret = 0;
1086         struct oidset_iter iter;
1087         const struct object_id *oid;
1088
1089         oidset_iter_init(&gitmodules_found, &iter);
1090         while ((oid = oidset_iter_next(&iter))) {
1091                 struct blob *blob;
1092                 enum object_type type;
1093                 unsigned long size;
1094                 char *buf;
1095
1096                 if (oidset_contains(&gitmodules_done, oid))
1097                         continue;
1098
1099                 blob = lookup_blob(the_repository, oid);
1100                 if (!blob) {
1101                         struct object *obj = lookup_unknown_object(oid->hash);
1102                         ret |= report(options, obj,
1103                                       FSCK_MSG_GITMODULES_BLOB,
1104                                       "non-blob found at .gitmodules");
1105                         continue;
1106                 }
1107
1108                 buf = read_object_file(oid, &type, &size);
1109                 if (!buf) {
1110                         if (is_promisor_object(&blob->object.oid))
1111                                 continue;
1112                         ret |= report(options, &blob->object,
1113                                       FSCK_MSG_GITMODULES_MISSING,
1114                                       "unable to read .gitmodules blob");
1115                         continue;
1116                 }
1117
1118                 if (type == OBJ_BLOB)
1119                         ret |= fsck_blob(blob, buf, size, options);
1120                 else
1121                         ret |= report(options, &blob->object,
1122                                       FSCK_MSG_GITMODULES_BLOB,
1123                                       "non-blob found at .gitmodules");
1124                 free(buf);
1125         }
1126
1127
1128         oidset_clear(&gitmodules_found);
1129         oidset_clear(&gitmodules_done);
1130         return ret;
1131 }