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 "credential.h"
  20 #include "help.h"
  21
  22 static struct oidset gitmodules_found = OIDSET_INIT;
  23 static struct oidset gitmodules_done = OIDSET_INIT;
  24
  25 #define FSCK_FATAL -1
  26 #define FSCK_INFO -2
  27
  28 #define FOREACH_MSG_ID(FUNC) \
  29         /* fatal errors */ \
  30         FUNC(NUL_IN_HEADER, FATAL) \
  31         FUNC(UNTERMINATED_HEADER, FATAL) \
  32         /* errors */ \
  33         FUNC(BAD_DATE, ERROR) \
  34         FUNC(BAD_DATE_OVERFLOW, ERROR) \
  35         FUNC(BAD_EMAIL, ERROR) \
  36         FUNC(BAD_NAME, ERROR) \
  37         FUNC(BAD_OBJECT_SHA1, ERROR) \
  38         FUNC(BAD_PARENT_SHA1, ERROR) \
  39         FUNC(BAD_TAG_OBJECT, ERROR) \
  40         FUNC(BAD_TIMEZONE, ERROR) \
  41         FUNC(BAD_TREE, ERROR) \
  42         FUNC(BAD_TREE_SHA1, ERROR) \
  43         FUNC(BAD_TYPE, ERROR) \
  44         FUNC(DUPLICATE_ENTRIES, ERROR) \
  45         FUNC(MISSING_AUTHOR, ERROR) \
  46         FUNC(MISSING_COMMITTER, ERROR) \
  47         FUNC(MISSING_EMAIL, ERROR) \
  48         FUNC(MISSING_GRAFT, ERROR) \
  49         FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  50         FUNC(MISSING_OBJECT, ERROR) \
  51         FUNC(MISSING_PARENT, ERROR) \
  52         FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  53         FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  54         FUNC(MISSING_TAG, ERROR) \
  55         FUNC(MISSING_TAG_ENTRY, ERROR) \
  56         FUNC(MISSING_TAG_OBJECT, ERROR) \
  57         FUNC(MISSING_TREE, ERROR) \
  58         FUNC(MISSING_TREE_OBJECT, ERROR) \
  59         FUNC(MISSING_TYPE, ERROR) \
  60         FUNC(MISSING_TYPE_ENTRY, ERROR) \
  61         FUNC(MULTIPLE_AUTHORS, ERROR) \
  62         FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  63         FUNC(TREE_NOT_SORTED, ERROR) \
  64         FUNC(UNKNOWN_TYPE, ERROR) \
  65         FUNC(ZERO_PADDED_DATE, ERROR) \
  66         FUNC(GITMODULES_MISSING, ERROR) \
  67         FUNC(GITMODULES_BLOB, ERROR) \
  68         FUNC(GITMODULES_LARGE, ERROR) \
  69         FUNC(GITMODULES_NAME, ERROR) \
  70         FUNC(GITMODULES_SYMLINK, ERROR) \
  71         FUNC(GITMODULES_URL, ERROR) \
  72         FUNC(GITMODULES_PATH, ERROR) \
  73         FUNC(GITMODULES_UPDATE, ERROR) \
  74         /* warnings */ \
  75         FUNC(BAD_FILEMODE, WARN) \
  76         FUNC(EMPTY_NAME, WARN) \
  77         FUNC(FULL_PATHNAME, WARN) \
  78         FUNC(HAS_DOT, WARN) \
  79         FUNC(HAS_DOTDOT, WARN) \
  80         FUNC(HAS_DOTGIT, WARN) \
  81         FUNC(NULL_SHA1, WARN) \
  82         FUNC(ZERO_PADDED_FILEMODE, WARN) \
  83         FUNC(NUL_IN_COMMIT, WARN) \
  84         /* infos (reported as warnings, but ignored by default) */ \
  85         FUNC(GITMODULES_PARSE, INFO) \
  86         FUNC(BAD_TAG_NAME, INFO) \
  87         FUNC(MISSING_TAGGER_ENTRY, INFO)
  88
  89 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
  90 enum fsck_msg_id {
  91         FOREACH_MSG_ID(MSG_ID)
  92         FSCK_MSG_MAX
  93 };
  94 #undef MSG_ID
  95
  96 #define STR(x) #x
  97 #define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
  98 static struct {
  99         const char *id_string;
 100         const char *downcased;
 101         const char *camelcased;
 102         int msg_type;
 103 } msg_id_info[FSCK_MSG_MAX + 1] = {
 104         FOREACH_MSG_ID(MSG_ID)
 105         { NULL, NULL, NULL, -1 }
 106 };
 107 #undef MSG_ID
 108
 109 static void prepare_msg_ids(void)
 110 {
 111         int i;
 112
 113         if (msg_id_info[0].downcased)
 114                 return;
 115
 116         /* convert id_string to lower case, without underscores. */
 117         for (i = 0; i < FSCK_MSG_MAX; i++) {
 118                 const char *p = msg_id_info[i].id_string;
 119                 int len = strlen(p);
 120                 char *q = xmalloc(len);
 121
 122                 msg_id_info[i].downcased = q;
 123                 while (*p)
 124                         if (*p == '_')
 125                                 p++;
 126                         else
 127                                 *(q)++ = tolower(*(p)++);
 128                 *q = '\0';
 129
 130                 p = msg_id_info[i].id_string;
 131                 q = xmalloc(len);
 132                 msg_id_info[i].camelcased = q;
 133                 while (*p) {
 134                         if (*p == '_') {
 135                                 p++;
 136                                 if (*p)
 137                                         *q++ = *p++;
 138                         } else {
 139                                 *q++ = tolower(*p++);
 140                         }
 141                 }
 142                 *q = '\0';
 143         }
 144 }
 145
 146 static int parse_msg_id(const char *text)
 147 {
 148         int i;
 149
 150         prepare_msg_ids();
 151
 152         for (i = 0; i < FSCK_MSG_MAX; i++)
 153                 if (!strcmp(text, msg_id_info[i].downcased))
 154                         return i;
 155
 156         return -1;
 157 }
 158
 159 void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
 160 {
 161         int i;
 162
 163         prepare_msg_ids();
 164
 165         for (i = 0; i < FSCK_MSG_MAX; i++)
 166                 list_config_item(list, prefix, msg_id_info[i].camelcased);
 167 }
 168
 169 static int fsck_msg_type(enum fsck_msg_id msg_id,
 170         struct fsck_options *options)
 171 {
 172         int msg_type;
 173
 174         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 175
 176         if (options->msg_type)
 177                 msg_type = options->msg_type[msg_id];
 178         else {
 179                 msg_type = msg_id_info[msg_id].msg_type;
 180                 if (options->strict && msg_type == FSCK_WARN)
 181                         msg_type = FSCK_ERROR;
 182         }
 183
 184         return msg_type;
 185 }
 186
 187 static void init_skiplist(struct fsck_options *options, const char *path)
 188 {
 189         static struct oid_array skiplist = OID_ARRAY_INIT;
 190         int sorted, fd;
 191         char buffer[GIT_MAX_HEXSZ + 1];
 192         struct object_id oid;
 193
 194         if (options->skiplist)
 195                 sorted = options->skiplist->sorted;
 196         else {
 197                 sorted = 1;
 198                 options->skiplist = &skiplist;
 199         }
 200
 201         fd = open(path, O_RDONLY);
 202         if (fd < 0)
 203                 die("Could not open skip list: %s", path);
 204         for (;;) {
 205                 const char *p;
 206                 int result = read_in_full(fd, buffer, sizeof(buffer));
 207                 if (result < 0)
 208                         die_errno("Could not read '%s'", path);
 209                 if (!result)
 210                         break;
 211                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 212                         die("Invalid SHA-1: %s", buffer);
 213                 oid_array_append(&skiplist, &oid);
 214                 if (sorted && skiplist.nr > 1 &&
 215                                 oidcmp(&skiplist.oid[skiplist.nr - 2],
 216                                        &oid) > 0)
 217                         sorted = 0;
 218         }
 219         close(fd);
 220
 221         if (sorted)
 222                 skiplist.sorted = 1;
 223 }
 224
 225 static int parse_msg_type(const char *str)
 226 {
 227         if (!strcmp(str, "error"))
 228                 return FSCK_ERROR;
 229         else if (!strcmp(str, "warn"))
 230                 return FSCK_WARN;
 231         else if (!strcmp(str, "ignore"))
 232                 return FSCK_IGNORE;
 233         else
 234                 die("Unknown fsck message type: '%s'", str);
 235 }
 236
 237 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 238 {
 239         if (parse_msg_id(msg_id) < 0)
 240                 return 0;
 241         parse_msg_type(msg_type);
 242         return 1;
 243 }
 244
 245 void fsck_set_msg_type(struct fsck_options *options,
 246                 const char *msg_id, const char *msg_type)
 247 {
 248         int id = parse_msg_id(msg_id), type;
 249
 250         if (id < 0)
 251                 die("Unhandled message id: %s", msg_id);
 252         type = parse_msg_type(msg_type);
 253
 254         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 255                 die("Cannot demote %s to %s", msg_id, msg_type);
 256
 257         if (!options->msg_type) {
 258                 int i;
 259                 int *msg_type;
 260                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 261                 for (i = 0; i < FSCK_MSG_MAX; i++)
 262                         msg_type[i] = fsck_msg_type(i, options);
 263                 options->msg_type = msg_type;
 264         }
 265
 266         options->msg_type[id] = type;
 267 }
 268
 269 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 270 {
 271         char *buf = xstrdup(values), *to_free = buf;
 272         int done = 0;
 273
 274         while (!done) {
 275                 int len = strcspn(buf, " ,|"), equal;
 276
 277                 done = !buf[len];
 278                 if (!len) {
 279                         buf++;
 280                         continue;
 281                 }
 282                 buf[len] = '\0';
 283
 284                 for (equal = 0;
 285                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 286                      equal++)
 287                         buf[equal] = tolower(buf[equal]);
 288                 buf[equal] = '\0';
 289
 290                 if (!strcmp(buf, "skiplist")) {
 291                         if (equal == len)
 292                                 die("skiplist requires a path");
 293                         init_skiplist(options, buf + equal + 1);
 294                         buf += len + 1;
 295                         continue;
 296                 }
 297
 298                 if (equal == len)
 299                         die("Missing '=': '%s'", buf);
 300
 301                 fsck_set_msg_type(options, buf, buf + equal + 1);
 302                 buf += len + 1;
 303         }
 304         free(to_free);
 305 }
 306
 307 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 308 {
 309         for (;;) {
 310                 char c = *(msg_id)++;
 311
 312                 if (!c)
 313                         break;
 314                 if (c != '_')
 315                         strbuf_addch(sb, tolower(c));
 316                 else {
 317                         assert(*msg_id);
 318                         strbuf_addch(sb, *(msg_id)++);
 319                 }
 320         }
 321
 322         strbuf_addstr(sb, ": ");
 323 }
 324
 325 static int object_on_skiplist(struct fsck_options *opts, struct object *obj)
 326 {
 327         if (opts && opts->skiplist && obj)
 328                 return oid_array_lookup(opts->skiplist, &obj->oid) >= 0;
 329         return 0;
 330 }
 331
 332 __attribute__((format (printf, 4, 5)))
 333 static int report(struct fsck_options *options, struct object *object,
 334         enum fsck_msg_id id, const char *fmt, ...)
 335 {
 336         va_list ap;
 337         struct strbuf sb = STRBUF_INIT;
 338         int msg_type = fsck_msg_type(id, options), result;
 339
 340         if (msg_type == FSCK_IGNORE)
 341                 return 0;
 342
 343         if (object_on_skiplist(options, object))
 344                 return 0;
 345
 346         if (msg_type == FSCK_FATAL)
 347                 msg_type = FSCK_ERROR;
 348         else if (msg_type == FSCK_INFO)
 349                 msg_type = FSCK_WARN;
 350
 351         append_msg_id(&sb, msg_id_info[id].id_string);
 352
 353         va_start(ap, fmt);
 354         strbuf_vaddf(&sb, fmt, ap);
 355         result = options->error_func(options, object, msg_type, sb.buf);
 356         strbuf_release(&sb);
 357         va_end(ap);
 358
 359         return result;
 360 }
 361
 362 static char *get_object_name(struct fsck_options *options, struct object *obj)
 363 {
 364         if (!options->object_names)
 365                 return NULL;
 366         return lookup_decoration(options->object_names, obj);
 367 }
 368
 369 static void put_object_name(struct fsck_options *options, struct object *obj,
 370         const char *fmt, ...)
 371 {
 372         va_list ap;
 373         struct strbuf buf = STRBUF_INIT;
 374         char *existing;
 375
 376         if (!options->object_names)
 377                 return;
 378         existing = lookup_decoration(options->object_names, obj);
 379         if (existing)
 380                 return;
 381         va_start(ap, fmt);
 382         strbuf_vaddf(&buf, fmt, ap);
 383         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 384         va_end(ap);
 385 }
 386
 387 static const char *describe_object(struct fsck_options *o, struct object *obj)
 388 {
 389         static struct strbuf buf = STRBUF_INIT;
 390         char *name;
 391
 392         strbuf_reset(&buf);
 393         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 394         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 395                 strbuf_addf(&buf, " (%s)", name);
 396
 397         return buf.buf;
 398 }
 399
 400 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 401 {
 402         struct tree_desc desc;
 403         struct name_entry entry;
 404         int res = 0;
 405         const char *name;
 406
 407         if (parse_tree(tree))
 408                 return -1;
 409
 410         name = get_object_name(options, &tree->object);
 411         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 412                 return -1;
 413         while (tree_entry_gently(&desc, &entry)) {
 414                 struct object *obj;
 415                 int result;
 416
 417                 if (S_ISGITLINK(entry.mode))
 418                         continue;
 419
 420                 if (S_ISDIR(entry.mode)) {
 421                         obj = (struct object *)lookup_tree(the_repository, entry.oid);
 422                         if (name && obj)
 423                                 put_object_name(options, obj, "%s%s/", name,
 424                                         entry.path);
 425                         result = options->walk(obj, OBJ_TREE, data, options);
 426                 }
 427                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 428                         obj = (struct object *)lookup_blob(the_repository, entry.oid);
 429                         if (name && obj)
 430                                 put_object_name(options, obj, "%s%s", name,
 431                                         entry.path);
 432                         result = options->walk(obj, OBJ_BLOB, data, options);
 433                 }
 434                 else {
 435                         result = error("in tree %s: entry %s has bad mode %.6o",
 436                                         describe_object(options, &tree->object), entry.path, entry.mode);
 437                 }
 438                 if (result < 0)
 439                         return result;
 440                 if (!res)
 441                         res = result;
 442         }
 443         return res;
 444 }
 445
 446 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 447 {
 448         int counter = 0, generation = 0, name_prefix_len = 0;
 449         struct commit_list *parents;
 450         int res;
 451         int result;
 452         const char *name;
 453
 454         if (parse_commit(commit))
 455                 return -1;
 456
 457         name = get_object_name(options, &commit->object);
 458         if (name)
 459                 put_object_name(options, &get_commit_tree(commit)->object,
 460                                 "%s:", name);
 461
 462         result = options->walk((struct object *)get_commit_tree(commit),
 463                                OBJ_TREE, data, options);
 464         if (result < 0)
 465                 return result;
 466         res = result;
 467
 468         parents = commit->parents;
 469         if (name && parents) {
 470                 int len = strlen(name), power;
 471
 472                 if (len && name[len - 1] == '^') {
 473                         generation = 1;
 474                         name_prefix_len = len - 1;
 475                 }
 476                 else { /* parse ~<generation> suffix */
 477                         for (generation = 0, power = 1;
 478                              len && isdigit(name[len - 1]);
 479                              power *= 10)
 480                                 generation += power * (name[--len] - '0');
 481                         if (power > 1 && len && name[len - 1] == '~')
 482                                 name_prefix_len = len - 1;
 483                 }
 484         }
 485
 486         while (parents) {
 487                 if (name) {
 488                         struct object *obj = &parents->item->object;
 489
 490                         if (counter++)
 491                                 put_object_name(options, obj, "%s^%d",
 492                                         name, counter);
 493                         else if (generation > 0)
 494                                 put_object_name(options, obj, "%.*s~%d",
 495                                         name_prefix_len, name, generation + 1);
 496                         else
 497                                 put_object_name(options, obj, "%s^", name);
 498                 }
 499                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 500                 if (result < 0)
 501                         return result;
 502                 if (!res)
 503                         res = result;
 504                 parents = parents->next;
 505         }
 506         return res;
 507 }
 508
 509 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 510 {
 511         char *name = get_object_name(options, &tag->object);
 512
 513         if (parse_tag(tag))
 514                 return -1;
 515         if (name)
 516                 put_object_name(options, tag->tagged, "%s", name);
 517         return options->walk(tag->tagged, OBJ_ANY, data, options);
 518 }
 519
 520 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 521 {
 522         if (!obj)
 523                 return -1;
 524
 525         if (obj->type == OBJ_NONE)
 526                 parse_object(the_repository, &obj->oid);
 527
 528         switch (obj->type) {
 529         case OBJ_BLOB:
 530                 return 0;
 531         case OBJ_TREE:
 532                 return fsck_walk_tree((struct tree *)obj, data, options);
 533         case OBJ_COMMIT:
 534                 return fsck_walk_commit((struct commit *)obj, data, options);
 535         case OBJ_TAG:
 536                 return fsck_walk_tag((struct tag *)obj, data, options);
 537         default:
 538                 error("Unknown object type for %s", describe_object(options, obj));
 539                 return -1;
 540         }
 541 }
 542
 543 /*
 544  * The entries in a tree are ordered in the _path_ order,
 545  * which means that a directory entry is ordered by adding
 546  * a slash to the end of it.
 547  *
 548  * So a directory called "a" is ordered _after_ a file
 549  * called "a.c", because "a/" sorts after "a.c".
 550  */
 551 #define TREE_UNORDERED (-1)
 552 #define TREE_HAS_DUPS  (-2)
 553
 554 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 555 {
 556         int len1 = strlen(name1);
 557         int len2 = strlen(name2);
 558         int len = len1 < len2 ? len1 : len2;
 559         unsigned char c1, c2;
 560         int cmp;
 561
 562         cmp = memcmp(name1, name2, len);
 563         if (cmp < 0)
 564                 return 0;
 565         if (cmp > 0)
 566                 return TREE_UNORDERED;
 567
 568         /*
 569          * Ok, the first <len> characters are the same.
 570          * Now we need to order the next one, but turn
 571          * a '\0' into a '/' for a directory entry.
 572          */
 573         c1 = name1[len];
 574         c2 = name2[len];
 575         if (!c1 && !c2)
 576                 /*
 577                  * git-write-tree used to write out a nonsense tree that has
 578                  * entries with the same name, one blob and one tree.  Make
 579                  * sure we do not have duplicate entries.
 580                  */
 581                 return TREE_HAS_DUPS;
 582         if (!c1 && S_ISDIR(mode1))
 583                 c1 = '/';
 584         if (!c2 && S_ISDIR(mode2))
 585                 c2 = '/';
 586         return c1 < c2 ? 0 : TREE_UNORDERED;
 587 }
 588
 589 static int fsck_tree(struct tree *item, struct fsck_options *options)
 590 {
 591         int retval = 0;
 592         int has_null_sha1 = 0;
 593         int has_full_path = 0;
 594         int has_empty_name = 0;
 595         int has_dot = 0;
 596         int has_dotdot = 0;
 597         int has_dotgit = 0;
 598         int has_zero_pad = 0;
 599         int has_bad_modes = 0;
 600         int has_dup_entries = 0;
 601         int not_properly_sorted = 0;
 602         struct tree_desc desc;
 603         unsigned o_mode;
 604         const char *o_name;
 605
 606         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 607                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 608                 return retval;
 609         }
 610
 611         o_mode = 0;
 612         o_name = NULL;
 613
 614         while (desc.size) {
 615                 unsigned mode;
 616                 const char *name, *backslash;
 617                 const struct object_id *oid;
 618
 619                 oid = tree_entry_extract(&desc, &name, &mode);
 620
 621                 has_null_sha1 |= is_null_oid(oid);
 622                 has_full_path |= !!strchr(name, '/');
 623                 has_empty_name |= !*name;
 624                 has_dot |= !strcmp(name, ".");
 625                 has_dotdot |= !strcmp(name, "..");
 626                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 627                 has_zero_pad |= *(char *)desc.buffer == '0';
 628
 629                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 630                         if (!S_ISLNK(mode))
 631                                 oidset_insert(&gitmodules_found, oid);
 632                         else
 633                                 retval += report(options, &item->object,
 634                                                  FSCK_MSG_GITMODULES_SYMLINK,
 635                                                  ".gitmodules is a symbolic link");
 636                 }
 637
 638                 if ((backslash = strchr(name, '\\'))) {
 639                         while (backslash) {
 640                                 backslash++;
 641                                 has_dotgit |= is_ntfs_dotgit(backslash);
 642                                 if (is_ntfs_dotgitmodules(backslash)) {
 643                                         if (!S_ISLNK(mode))
 644                                                 oidset_insert(&gitmodules_found, oid);
 645                                         else
 646                                                 retval += report(options, &item->object,
 647                                                                  FSCK_MSG_GITMODULES_SYMLINK,
 648                                                                  ".gitmodules is a symbolic link");
 649                                 }
 650                                 backslash = strchr(backslash, '\\');
 651                         }
 652                 }
 653
 654                 if (update_tree_entry_gently(&desc)) {
 655                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 656                         break;
 657                 }
 658
 659                 switch (mode) {
 660                 /*
 661                  * Standard modes..
 662                  */
 663                 case S_IFREG | 0755:
 664                 case S_IFREG | 0644:
 665                 case S_IFLNK:
 666                 case S_IFDIR:
 667                 case S_IFGITLINK:
 668                         break;
 669                 /*
 670                  * This is nonstandard, but we had a few of these
 671                  * early on when we honored the full set of mode
 672                  * bits..
 673                  */
 674                 case S_IFREG | 0664:
 675                         if (!options->strict)
 676                                 break;
 677                         /* fallthrough */
 678                 default:
 679                         has_bad_modes = 1;
 680                 }
 681
 682                 if (o_name) {
 683                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 684                         case TREE_UNORDERED:
 685                                 not_properly_sorted = 1;
 686                                 break;
 687                         case TREE_HAS_DUPS:
 688                                 has_dup_entries = 1;
 689                                 break;
 690                         default:
 691                                 break;
 692                         }
 693                 }
 694
 695                 o_mode = mode;
 696                 o_name = name;
 697         }
 698
 699         if (has_null_sha1)
 700                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 701         if (has_full_path)
 702                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 703         if (has_empty_name)
 704                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 705         if (has_dot)
 706                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 707         if (has_dotdot)
 708                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 709         if (has_dotgit)
 710                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 711         if (has_zero_pad)
 712                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 713         if (has_bad_modes)
 714                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 715         if (has_dup_entries)
 716                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 717         if (not_properly_sorted)
 718                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 719         return retval;
 720 }
 721
 722 static int verify_headers(const void *data, unsigned long size,
 723                           struct object *obj, struct fsck_options *options)
 724 {
 725         const char *buffer = (const char *)data;
 726         unsigned long i;
 727
 728         for (i = 0; i < size; i++) {
 729                 switch (buffer[i]) {
 730                 case '\0':
 731                         return report(options, obj,
 732                                 FSCK_MSG_NUL_IN_HEADER,
 733                                 "unterminated header: NUL at offset %ld", i);
 734                 case '\n':
 735                         if (i + 1 < size && buffer[i + 1] == '\n')
 736                                 return 0;
 737                 }
 738         }
 739
 740         /*
 741          * We did not find double-LF that separates the header
 742          * and the body.  Not having a body is not a crime but
 743          * we do want to see the terminating LF for the last header
 744          * line.
 745          */
 746         if (size && buffer[size - 1] == '\n')
 747                 return 0;
 748
 749         return report(options, obj,
 750                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 751 }
 752
 753 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 754 {
 755         const char *p = *ident;
 756         char *end;
 757
 758         *ident = strchrnul(*ident, '\n');
 759         if (**ident == '\n')
 760                 (*ident)++;
 761
 762         if (*p == '<')
 763                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 764         p += strcspn(p, "<>\n");
 765         if (*p == '>')
 766                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 767         if (*p != '<')
 768                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 769         if (p[-1] != ' ')
 770                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 771         p++;
 772         p += strcspn(p, "<>\n");
 773         if (*p != '>')
 774                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 775         p++;
 776         if (*p != ' ')
 777                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 778         p++;
 779         if (*p == '0' && p[1] != ' ')
 780                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 781         if (date_overflows(parse_timestamp(p, &end, 10)))
 782                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 783         if ((end == p || *end != ' '))
 784                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 785         p = end + 1;
 786         if ((*p != '+' && *p != '-') ||
 787             !isdigit(p[1]) ||
 788             !isdigit(p[2]) ||
 789             !isdigit(p[3]) ||
 790             !isdigit(p[4]) ||
 791             (p[5] != '\n'))
 792                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 793         p += 6;
 794         return 0;
 795 }
 796
 797 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 798         unsigned long size, struct fsck_options *options)
 799 {
 800         struct object_id tree_oid, oid;
 801         struct commit_graft *graft;
 802         unsigned parent_count, parent_line_count = 0, author_count;
 803         int err;
 804         const char *buffer_begin = buffer;
 805         const char *p;
 806
 807         if (verify_headers(buffer, size, &commit->object, options))
 808                 return -1;
 809
 810         if (!skip_prefix(buffer, "tree ", &buffer))
 811                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 812         if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 813                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 814                 if (err)
 815                         return err;
 816         }
 817         buffer = p + 1;
 818         while (skip_prefix(buffer, "parent ", &buffer)) {
 819                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 820                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 821                         if (err)
 822                                 return err;
 823                 }
 824                 buffer = p + 1;
 825                 parent_line_count++;
 826         }
 827         graft = lookup_commit_graft(the_repository, &commit->object.oid);
 828         parent_count = commit_list_count(commit->parents);
 829         if (graft) {
 830                 if (graft->nr_parent == -1 && !parent_count)
 831                         ; /* shallow commit */
 832                 else if (graft->nr_parent != parent_count) {
 833                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 834                         if (err)
 835                                 return err;
 836                 }
 837         } else {
 838                 if (parent_count != parent_line_count) {
 839                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 840                         if (err)
 841                                 return err;
 842                 }
 843         }
 844         author_count = 0;
 845         while (skip_prefix(buffer, "author ", &buffer)) {
 846                 author_count++;
 847                 err = fsck_ident(&buffer, &commit->object, options);
 848                 if (err)
 849                         return err;
 850         }
 851         if (author_count < 1)
 852                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 853         else if (author_count > 1)
 854                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 855         if (err)
 856                 return err;
 857         if (!skip_prefix(buffer, "committer ", &buffer))
 858                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 859         err = fsck_ident(&buffer, &commit->object, options);
 860         if (err)
 861                 return err;
 862         if (!get_commit_tree(commit)) {
 863                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
 864                 if (err)
 865                         return err;
 866         }
 867         if (memchr(buffer_begin, '\0', size)) {
 868                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 869                              "NUL byte in the commit object body");
 870                 if (err)
 871                         return err;
 872         }
 873         return 0;
 874 }
 875
 876 static int fsck_commit(struct commit *commit, const char *data,
 877         unsigned long size, struct fsck_options *options)
 878 {
 879         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 880         int ret = fsck_commit_buffer(commit, buffer, size, options);
 881         if (!data)
 882                 unuse_commit_buffer(commit, buffer);
 883         return ret;
 884 }
 885
 886 static int fsck_tag_buffer(struct tag *tag, const char *data,
 887         unsigned long size, struct fsck_options *options)
 888 {
 889         struct object_id oid;
 890         int ret = 0;
 891         const char *buffer;
 892         char *to_free = NULL, *eol;
 893         struct strbuf sb = STRBUF_INIT;
 894         const char *p;
 895
 896         if (data)
 897                 buffer = data;
 898         else {
 899                 enum object_type type;
 900
 901                 buffer = to_free =
 902                         read_object_file(&tag->object.oid, &type, &size);
 903                 if (!buffer)
 904                         return report(options, &tag->object,
 905                                 FSCK_MSG_MISSING_TAG_OBJECT,
 906                                 "cannot read tag object");
 907
 908                 if (type != OBJ_TAG) {
 909                         ret = report(options, &tag->object,
 910                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 911                                 "expected tag got %s",
 912                             type_name(type));
 913                         goto done;
 914                 }
 915         }
 916
 917         ret = verify_headers(buffer, size, &tag->object, options);
 918         if (ret)
 919                 goto done;
 920
 921         if (!skip_prefix(buffer, "object ", &buffer)) {
 922                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 923                 goto done;
 924         }
 925         if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 926                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 927                 if (ret)
 928                         goto done;
 929         }
 930         buffer = p + 1;
 931
 932         if (!skip_prefix(buffer, "type ", &buffer)) {
 933                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 934                 goto done;
 935         }
 936         eol = strchr(buffer, '\n');
 937         if (!eol) {
 938                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 939                 goto done;
 940         }
 941         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 942                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 943         if (ret)
 944                 goto done;
 945         buffer = eol + 1;
 946
 947         if (!skip_prefix(buffer, "tag ", &buffer)) {
 948                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 949                 goto done;
 950         }
 951         eol = strchr(buffer, '\n');
 952         if (!eol) {
 953                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 954                 goto done;
 955         }
 956         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 957         if (check_refname_format(sb.buf, 0)) {
 958                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 959                            "invalid 'tag' name: %.*s",
 960                            (int)(eol - buffer), buffer);
 961                 if (ret)
 962                         goto done;
 963         }
 964         buffer = eol + 1;
 965
 966         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 967                 /* early tags do not contain 'tagger' lines; warn only */
 968                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 969                 if (ret)
 970                         goto done;
 971         }
 972         else
 973                 ret = fsck_ident(&buffer, &tag->object, options);
 974
 975 done:
 976         strbuf_release(&sb);
 977         free(to_free);
 978         return ret;
 979 }
 980
 981 static int fsck_tag(struct tag *tag, const char *data,
 982         unsigned long size, struct fsck_options *options)
 983 {
 984         struct object *tagged = tag->tagged;
 985
 986         if (!tagged)
 987                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 988
 989         return fsck_tag_buffer(tag, data, size, options);
 990 }
 991
 992 static int check_submodule_url(const char *url)
 993 {
 994         struct credential c = CREDENTIAL_INIT;
 995         int ret;
 996
 997         if (looks_like_command_line_option(url))
 998                 return -1;
 999
1000         ret = credential_from_url_gently(&c, url, 1);
1001         credential_clear(&c);
1002         return ret;
1003 }
1004
1005 struct fsck_gitmodules_data {
1006         struct object *obj;
1007         struct fsck_options *options;
1008         int ret;
1009 };
1010
1011 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
1012 {
1013         struct fsck_gitmodules_data *data = vdata;
1014         const char *subsection, *key;
1015         int subsection_len;
1016         char *name;
1017
1018         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
1019             !subsection)
1020                 return 0;
1021
1022         name = xmemdupz(subsection, subsection_len);
1023         if (check_submodule_name(name) < 0)
1024                 data->ret |= report(data->options, data->obj,
1025                                     FSCK_MSG_GITMODULES_NAME,
1026                                     "disallowed submodule name: %s",
1027                                     name);
1028         if (!strcmp(key, "url") && value &&
1029             check_submodule_url(value) < 0)
1030                 data->ret |= report(data->options, data->obj,
1031                                     FSCK_MSG_GITMODULES_URL,
1032                                     "disallowed submodule url: %s",
1033                                     value);
1034         if (!strcmp(key, "path") && value &&
1035             looks_like_command_line_option(value))
1036                 data->ret |= report(data->options, data->obj,
1037                                     FSCK_MSG_GITMODULES_PATH,
1038                                     "disallowed submodule path: %s",
1039                                     value);
1040         if (!strcmp(key, "update") && value &&
1041             parse_submodule_update_type(value) == SM_UPDATE_COMMAND)
1042                 data->ret |= report(data->options, data->obj,
1043                                     FSCK_MSG_GITMODULES_UPDATE,
1044                                     "disallowed submodule update setting: %s",
1045                                     value);
1046         free(name);
1047
1048         return 0;
1049 }
1050
1051 static int fsck_blob(struct blob *blob, const char *buf,
1052                      unsigned long size, struct fsck_options *options)
1053 {
1054         struct fsck_gitmodules_data data;
1055         struct config_options config_opts = { 0 };
1056
1057         if (!oidset_contains(&gitmodules_found, &blob->object.oid))
1058                 return 0;
1059         oidset_insert(&gitmodules_done, &blob->object.oid);
1060
1061         if (object_on_skiplist(options, &blob->object))
1062                 return 0;
1063
1064         if (!buf) {
1065                 /*
1066                  * A missing buffer here is a sign that the caller found the
1067                  * blob too gigantic to load into memory. Let's just consider
1068                  * that an error.
1069                  */
1070                 return report(options, &blob->object,
1071                               FSCK_MSG_GITMODULES_LARGE,
1072                               ".gitmodules too large to parse");
1073         }
1074
1075         data.obj = &blob->object;
1076         data.options = options;
1077         data.ret = 0;
1078         config_opts.error_action = CONFIG_ERROR_SILENT;
1079         if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1080                                 ".gitmodules", buf, size, &data, &config_opts))
1081                 data.ret |= report(options, &blob->object,
1082                                    FSCK_MSG_GITMODULES_PARSE,
1083                                    "could not parse gitmodules blob");
1084
1085         return data.ret;
1086 }
1087
1088 int fsck_object(struct object *obj, void *data, unsigned long size,
1089         struct fsck_options *options)
1090 {
1091         if (!obj)
1092                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1093
1094         if (obj->type == OBJ_BLOB)
1095                 return fsck_blob((struct blob *)obj, data, size, options);
1096         if (obj->type == OBJ_TREE)
1097                 return fsck_tree((struct tree *) obj, options);
1098         if (obj->type == OBJ_COMMIT)
1099                 return fsck_commit((struct commit *) obj, (const char *) data,
1100                         size, options);
1101         if (obj->type == OBJ_TAG)
1102                 return fsck_tag((struct tag *) obj, (const char *) data,
1103                         size, options);
1104
1105         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1106                           obj->type);
1107 }
1108
1109 int fsck_error_function(struct fsck_options *o,
1110         struct object *obj, int msg_type, const char *message)
1111 {
1112         if (msg_type == FSCK_WARN) {
1113                 warning("object %s: %s", describe_object(o, obj), message);
1114                 return 0;
1115         }
1116         error("object %s: %s", describe_object(o, obj), message);
1117         return 1;
1118 }
1119
1120 int fsck_finish(struct fsck_options *options)
1121 {
1122         int ret = 0;
1123         struct oidset_iter iter;
1124         const struct object_id *oid;
1125
1126         oidset_iter_init(&gitmodules_found, &iter);
1127         while ((oid = oidset_iter_next(&iter))) {
1128                 struct blob *blob;
1129                 enum object_type type;
1130                 unsigned long size;
1131                 char *buf;
1132
1133                 if (oidset_contains(&gitmodules_done, oid))
1134                         continue;
1135
1136                 blob = lookup_blob(the_repository, oid);
1137                 if (!blob) {
1138                         struct object *obj = lookup_unknown_object(oid->hash);
1139                         ret |= report(options, obj,
1140                                       FSCK_MSG_GITMODULES_BLOB,
1141                                       "non-blob found at .gitmodules");
1142                         continue;
1143                 }
1144
1145                 buf = read_object_file(oid, &type, &size);
1146                 if (!buf) {
1147                         if (is_promisor_object(&blob->object.oid))
1148                                 continue;
1149                         ret |= report(options, &blob->object,
1150                                       FSCK_MSG_GITMODULES_MISSING,
1151                                       "unable to read .gitmodules blob");
1152                         continue;
1153                 }
1154
1155                 if (type == OBJ_BLOB)
1156                         ret |= fsck_blob(blob, buf, size, options);
1157                 else
1158                         ret |= report(options, &blob->object,
1159                                       FSCK_MSG_GITMODULES_BLOB,
1160                                       "non-blob found at .gitmodules");
1161                 free(buf);
1162         }
1163
1164
1165         oidset_clear(&gitmodules_found);
1166         oidset_clear(&gitmodules_done);
1167         return ret;
1168 }