fsck.c

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