fsck.c

   1 #include "cache.h"
   2 #include "object-store.h"
   3 #include "object.h"
   4 #include "blob.h"
   5 #include "tree.h"
   6 #include "tree-walk.h"
   7 #include "commit.h"
   8 #include "tag.h"
   9 #include "fsck.h"
  10 #include "refs.h"
  11 #include "utf8.h"
  12 #include "sha1-array.h"
  13 #include "decorate.h"
  14 #include "oidset.h"
  15 #include "packfile.h"
  16 #include "submodule-config.h"
  17 #include "config.h"
  18 #include "help.h"
  19
  20 static struct oidset gitmodules_found = OIDSET_INIT;
  21 static struct oidset gitmodules_done = OIDSET_INIT;
  22
  23 #define FSCK_FATAL -1
  24 #define FSCK_INFO -2
  25
  26 #define FOREACH_MSG_ID(FUNC) \
  27         /* fatal errors */ \
  28         FUNC(NUL_IN_HEADER, FATAL) \
  29         FUNC(UNTERMINATED_HEADER, FATAL) \
  30         /* errors */ \
  31         FUNC(BAD_DATE, ERROR) \
  32         FUNC(BAD_DATE_OVERFLOW, ERROR) \
  33         FUNC(BAD_EMAIL, ERROR) \
  34         FUNC(BAD_NAME, ERROR) \
  35         FUNC(BAD_OBJECT_SHA1, ERROR) \
  36         FUNC(BAD_PARENT_SHA1, ERROR) \
  37         FUNC(BAD_TAG_OBJECT, ERROR) \
  38         FUNC(BAD_TIMEZONE, ERROR) \
  39         FUNC(BAD_TREE, ERROR) \
  40         FUNC(BAD_TREE_SHA1, ERROR) \
  41         FUNC(BAD_TYPE, ERROR) \
  42         FUNC(DUPLICATE_ENTRIES, ERROR) \
  43         FUNC(MISSING_AUTHOR, ERROR) \
  44         FUNC(MISSING_COMMITTER, ERROR) \
  45         FUNC(MISSING_EMAIL, ERROR) \
  46         FUNC(MISSING_GRAFT, ERROR) \
  47         FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  48         FUNC(MISSING_OBJECT, ERROR) \
  49         FUNC(MISSING_PARENT, ERROR) \
  50         FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  51         FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  52         FUNC(MISSING_TAG, ERROR) \
  53         FUNC(MISSING_TAG_ENTRY, ERROR) \
  54         FUNC(MISSING_TAG_OBJECT, ERROR) \
  55         FUNC(MISSING_TREE, ERROR) \
  56         FUNC(MISSING_TREE_OBJECT, ERROR) \
  57         FUNC(MISSING_TYPE, ERROR) \
  58         FUNC(MISSING_TYPE_ENTRY, ERROR) \
  59         FUNC(MULTIPLE_AUTHORS, ERROR) \
  60         FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  61         FUNC(TREE_NOT_SORTED, ERROR) \
  62         FUNC(UNKNOWN_TYPE, ERROR) \
  63         FUNC(ZERO_PADDED_DATE, ERROR) \
  64         FUNC(GITMODULES_MISSING, ERROR) \
  65         FUNC(GITMODULES_BLOB, ERROR) \
  66         FUNC(GITMODULES_PARSE, ERROR) \
  67         FUNC(GITMODULES_NAME, ERROR) \
  68         FUNC(GITMODULES_SYMLINK, ERROR) \
  69         /* warnings */ \
  70         FUNC(BAD_FILEMODE, WARN) \
  71         FUNC(EMPTY_NAME, WARN) \
  72         FUNC(FULL_PATHNAME, WARN) \
  73         FUNC(HAS_DOT, WARN) \
  74         FUNC(HAS_DOTDOT, WARN) \
  75         FUNC(HAS_DOTGIT, WARN) \
  76         FUNC(NULL_SHA1, WARN) \
  77         FUNC(ZERO_PADDED_FILEMODE, WARN) \
  78         FUNC(NUL_IN_COMMIT, WARN) \
  79         /* infos (reported as warnings, but ignored by default) */ \
  80         FUNC(BAD_TAG_NAME, INFO) \
  81         FUNC(MISSING_TAGGER_ENTRY, INFO)
  82
  83 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
  84 enum fsck_msg_id {
  85         FOREACH_MSG_ID(MSG_ID)
  86         FSCK_MSG_MAX
  87 };
  88 #undef MSG_ID
  89
  90 #define STR(x) #x
  91 #define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
  92 static struct {
  93         const char *id_string;
  94         const char *downcased;
  95         const char *camelcased;
  96         int msg_type;
  97 } msg_id_info[FSCK_MSG_MAX + 1] = {
  98         FOREACH_MSG_ID(MSG_ID)
  99         { NULL, NULL, NULL, -1 }
 100 };
 101 #undef MSG_ID
 102
 103 static void prepare_msg_ids(void)
 104 {
 105         int i;
 106
 107         if (msg_id_info[0].downcased)
 108                 return;
 109
 110         /* convert id_string to lower case, without underscores. */
 111         for (i = 0; i < FSCK_MSG_MAX; i++) {
 112                 const char *p = msg_id_info[i].id_string;
 113                 int len = strlen(p);
 114                 char *q = xmalloc(len);
 115
 116                 msg_id_info[i].downcased = q;
 117                 while (*p)
 118                         if (*p == '_')
 119                                 p++;
 120                         else
 121                                 *(q)++ = tolower(*(p)++);
 122                 *q = '\0';
 123
 124                 p = msg_id_info[i].id_string;
 125                 q = xmalloc(len);
 126                 msg_id_info[i].camelcased = q;
 127                 while (*p) {
 128                         if (*p == '_') {
 129                                 p++;
 130                                 if (*p)
 131                                         *q++ = *p++;
 132                         } else {
 133                                 *q++ = tolower(*p++);
 134                         }
 135                 }
 136                 *q = '\0';
 137         }
 138 }
 139
 140 static int parse_msg_id(const char *text)
 141 {
 142         int i;
 143
 144         prepare_msg_ids();
 145
 146         for (i = 0; i < FSCK_MSG_MAX; i++)
 147                 if (!strcmp(text, msg_id_info[i].downcased))
 148                         return i;
 149
 150         return -1;
 151 }
 152
 153 void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
 154 {
 155         int i;
 156
 157         prepare_msg_ids();
 158
 159         for (i = 0; i < FSCK_MSG_MAX; i++)
 160                 list_config_item(list, prefix, msg_id_info[i].camelcased);
 161 }
 162
 163 static int fsck_msg_type(enum fsck_msg_id msg_id,
 164         struct fsck_options *options)
 165 {
 166         int msg_type;
 167
 168         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 169
 170         if (options->msg_type)
 171                 msg_type = options->msg_type[msg_id];
 172         else {
 173                 msg_type = msg_id_info[msg_id].msg_type;
 174                 if (options->strict && msg_type == FSCK_WARN)
 175                         msg_type = FSCK_ERROR;
 176         }
 177
 178         return msg_type;
 179 }
 180
 181 static void init_skiplist(struct fsck_options *options, const char *path)
 182 {
 183         static struct oid_array skiplist = OID_ARRAY_INIT;
 184         int sorted, fd;
 185         char buffer[GIT_MAX_HEXSZ + 1];
 186         struct object_id oid;
 187
 188         if (options->skiplist)
 189                 sorted = options->skiplist->sorted;
 190         else {
 191                 sorted = 1;
 192                 options->skiplist = &skiplist;
 193         }
 194
 195         fd = open(path, O_RDONLY);
 196         if (fd < 0)
 197                 die("Could not open skip list: %s", path);
 198         for (;;) {
 199                 const char *p;
 200                 int result = read_in_full(fd, buffer, sizeof(buffer));
 201                 if (result < 0)
 202                         die_errno("Could not read '%s'", path);
 203                 if (!result)
 204                         break;
 205                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 206                         die("Invalid SHA-1: %s", buffer);
 207                 oid_array_append(&skiplist, &oid);
 208                 if (sorted && skiplist.nr > 1 &&
 209                                 oidcmp(&skiplist.oid[skiplist.nr - 2],
 210                                        &oid) > 0)
 211                         sorted = 0;
 212         }
 213         close(fd);
 214
 215         if (sorted)
 216                 skiplist.sorted = 1;
 217 }
 218
 219 static int parse_msg_type(const char *str)
 220 {
 221         if (!strcmp(str, "error"))
 222                 return FSCK_ERROR;
 223         else if (!strcmp(str, "warn"))
 224                 return FSCK_WARN;
 225         else if (!strcmp(str, "ignore"))
 226                 return FSCK_IGNORE;
 227         else
 228                 die("Unknown fsck message type: '%s'", str);
 229 }
 230
 231 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 232 {
 233         if (parse_msg_id(msg_id) < 0)
 234                 return 0;
 235         parse_msg_type(msg_type);
 236         return 1;
 237 }
 238
 239 void fsck_set_msg_type(struct fsck_options *options,
 240                 const char *msg_id, const char *msg_type)
 241 {
 242         int id = parse_msg_id(msg_id), type;
 243
 244         if (id < 0)
 245                 die("Unhandled message id: %s", msg_id);
 246         type = parse_msg_type(msg_type);
 247
 248         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 249                 die("Cannot demote %s to %s", msg_id, msg_type);
 250
 251         if (!options->msg_type) {
 252                 int i;
 253                 int *msg_type;
 254                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 255                 for (i = 0; i < FSCK_MSG_MAX; i++)
 256                         msg_type[i] = fsck_msg_type(i, options);
 257                 options->msg_type = msg_type;
 258         }
 259
 260         options->msg_type[id] = type;
 261 }
 262
 263 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 264 {
 265         char *buf = xstrdup(values), *to_free = buf;
 266         int done = 0;
 267
 268         while (!done) {
 269                 int len = strcspn(buf, " ,|"), equal;
 270
 271                 done = !buf[len];
 272                 if (!len) {
 273                         buf++;
 274                         continue;
 275                 }
 276                 buf[len] = '\0';
 277
 278                 for (equal = 0;
 279                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 280                      equal++)
 281                         buf[equal] = tolower(buf[equal]);
 282                 buf[equal] = '\0';
 283
 284                 if (!strcmp(buf, "skiplist")) {
 285                         if (equal == len)
 286                                 die("skiplist requires a path");
 287                         init_skiplist(options, buf + equal + 1);
 288                         buf += len + 1;
 289                         continue;
 290                 }
 291
 292                 if (equal == len)
 293                         die("Missing '=': '%s'", buf);
 294
 295                 fsck_set_msg_type(options, buf, buf + equal + 1);
 296                 buf += len + 1;
 297         }
 298         free(to_free);
 299 }
 300
 301 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 302 {
 303         for (;;) {
 304                 char c = *(msg_id)++;
 305
 306                 if (!c)
 307                         break;
 308                 if (c != '_')
 309                         strbuf_addch(sb, tolower(c));
 310                 else {
 311                         assert(*msg_id);
 312                         strbuf_addch(sb, *(msg_id)++);
 313                 }
 314         }
 315
 316         strbuf_addstr(sb, ": ");
 317 }
 318
 319 static int object_on_skiplist(struct fsck_options *opts, struct object *obj)
 320 {
 321         if (opts && opts->skiplist && obj)
 322                 return oid_array_lookup(opts->skiplist, &obj->oid) >= 0;
 323         return 0;
 324 }
 325
 326 __attribute__((format (printf, 4, 5)))
 327 static int report(struct fsck_options *options, struct object *object,
 328         enum fsck_msg_id id, const char *fmt, ...)
 329 {
 330         va_list ap;
 331         struct strbuf sb = STRBUF_INIT;
 332         int msg_type = fsck_msg_type(id, options), result;
 333
 334         if (msg_type == FSCK_IGNORE)
 335                 return 0;
 336
 337         if (object_on_skiplist(options, object))
 338                 return 0;
 339
 340         if (msg_type == FSCK_FATAL)
 341                 msg_type = FSCK_ERROR;
 342         else if (msg_type == FSCK_INFO)
 343                 msg_type = FSCK_WARN;
 344
 345         append_msg_id(&sb, msg_id_info[id].id_string);
 346
 347         va_start(ap, fmt);
 348         strbuf_vaddf(&sb, fmt, ap);
 349         result = options->error_func(options, object, msg_type, sb.buf);
 350         strbuf_release(&sb);
 351         va_end(ap);
 352
 353         return result;
 354 }
 355
 356 static char *get_object_name(struct fsck_options *options, struct object *obj)
 357 {
 358         if (!options->object_names)
 359                 return NULL;
 360         return lookup_decoration(options->object_names, obj);
 361 }
 362
 363 static void put_object_name(struct fsck_options *options, struct object *obj,
 364         const char *fmt, ...)
 365 {
 366         va_list ap;
 367         struct strbuf buf = STRBUF_INIT;
 368         char *existing;
 369
 370         if (!options->object_names)
 371                 return;
 372         existing = lookup_decoration(options->object_names, obj);
 373         if (existing)
 374                 return;
 375         va_start(ap, fmt);
 376         strbuf_vaddf(&buf, fmt, ap);
 377         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 378         va_end(ap);
 379 }
 380
 381 static const char *describe_object(struct fsck_options *o, struct object *obj)
 382 {
 383         static struct strbuf buf = STRBUF_INIT;
 384         char *name;
 385
 386         strbuf_reset(&buf);
 387         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 388         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 389                 strbuf_addf(&buf, " (%s)", name);
 390
 391         return buf.buf;
 392 }
 393
 394 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 395 {
 396         struct tree_desc desc;
 397         struct name_entry entry;
 398         int res = 0;
 399         const char *name;
 400
 401         if (parse_tree(tree))
 402                 return -1;
 403
 404         name = get_object_name(options, &tree->object);
 405         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 406                 return -1;
 407         while (tree_entry_gently(&desc, &entry)) {
 408                 struct object *obj;
 409                 int result;
 410
 411                 if (S_ISGITLINK(entry.mode))
 412                         continue;
 413
 414                 if (S_ISDIR(entry.mode)) {
 415                         obj = (struct object *)lookup_tree(entry.oid);
 416                         if (name && obj)
 417                                 put_object_name(options, obj, "%s%s/", name,
 418                                         entry.path);
 419                         result = options->walk(obj, OBJ_TREE, data, options);
 420                 }
 421                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 422                         obj = (struct object *)lookup_blob(entry.oid);
 423                         if (name && obj)
 424                                 put_object_name(options, obj, "%s%s", name,
 425                                         entry.path);
 426                         result = options->walk(obj, OBJ_BLOB, data, options);
 427                 }
 428                 else {
 429                         result = error("in tree %s: entry %s has bad mode %.6o",
 430                                         describe_object(options, &tree->object), entry.path, entry.mode);
 431                 }
 432                 if (result < 0)
 433                         return result;
 434                 if (!res)
 435                         res = result;
 436         }
 437         return res;
 438 }
 439
 440 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 441 {
 442         int counter = 0, generation = 0, name_prefix_len = 0;
 443         struct commit_list *parents;
 444         int res;
 445         int result;
 446         const char *name;
 447
 448         if (parse_commit(commit))
 449                 return -1;
 450
 451         name = get_object_name(options, &commit->object);
 452         if (name)
 453                 put_object_name(options, &get_commit_tree(commit)->object,
 454                                 "%s:", name);
 455
 456         result = options->walk((struct object *)get_commit_tree(commit),
 457                                OBJ_TREE, data, options);
 458         if (result < 0)
 459                 return result;
 460         res = result;
 461
 462         parents = commit->parents;
 463         if (name && parents) {
 464                 int len = strlen(name), power;
 465
 466                 if (len && name[len - 1] == '^') {
 467                         generation = 1;
 468                         name_prefix_len = len - 1;
 469                 }
 470                 else { /* parse ~<generation> suffix */
 471                         for (generation = 0, power = 1;
 472                              len && isdigit(name[len - 1]);
 473                              power *= 10)
 474                                 generation += power * (name[--len] - '0');
 475                         if (power > 1 && len && name[len - 1] == '~')
 476                                 name_prefix_len = len - 1;
 477                 }
 478         }
 479
 480         while (parents) {
 481                 if (name) {
 482                         struct object *obj = &parents->item->object;
 483
 484                         if (++counter > 1)
 485                                 put_object_name(options, obj, "%s^%d",
 486                                         name, counter);
 487                         else if (generation > 0)
 488                                 put_object_name(options, obj, "%.*s~%d",
 489                                         name_prefix_len, name, generation + 1);
 490                         else
 491                                 put_object_name(options, obj, "%s^", name);
 492                 }
 493                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 494                 if (result < 0)
 495                         return result;
 496                 if (!res)
 497                         res = result;
 498                 parents = parents->next;
 499         }
 500         return res;
 501 }
 502
 503 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 504 {
 505         char *name = get_object_name(options, &tag->object);
 506
 507         if (parse_tag(tag))
 508                 return -1;
 509         if (name)
 510                 put_object_name(options, tag->tagged, "%s", name);
 511         return options->walk(tag->tagged, OBJ_ANY, data, options);
 512 }
 513
 514 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 515 {
 516         if (!obj)
 517                 return -1;
 518
 519         if (obj->type == OBJ_NONE)
 520                 parse_object(&obj->oid);
 521
 522         switch (obj->type) {
 523         case OBJ_BLOB:
 524                 return 0;
 525         case OBJ_TREE:
 526                 return fsck_walk_tree((struct tree *)obj, data, options);
 527         case OBJ_COMMIT:
 528                 return fsck_walk_commit((struct commit *)obj, data, options);
 529         case OBJ_TAG:
 530                 return fsck_walk_tag((struct tag *)obj, data, options);
 531         default:
 532                 error("Unknown object type for %s", describe_object(options, obj));
 533                 return -1;
 534         }
 535 }
 536
 537 /*
 538  * The entries in a tree are ordered in the _path_ order,
 539  * which means that a directory entry is ordered by adding
 540  * a slash to the end of it.
 541  *
 542  * So a directory called "a" is ordered _after_ a file
 543  * called "a.c", because "a/" sorts after "a.c".
 544  */
 545 #define TREE_UNORDERED (-1)
 546 #define TREE_HAS_DUPS  (-2)
 547
 548 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 549 {
 550         int len1 = strlen(name1);
 551         int len2 = strlen(name2);
 552         int len = len1 < len2 ? len1 : len2;
 553         unsigned char c1, c2;
 554         int cmp;
 555
 556         cmp = memcmp(name1, name2, len);
 557         if (cmp < 0)
 558                 return 0;
 559         if (cmp > 0)
 560                 return TREE_UNORDERED;
 561
 562         /*
 563          * Ok, the first <len> characters are the same.
 564          * Now we need to order the next one, but turn
 565          * a '\0' into a '/' for a directory entry.
 566          */
 567         c1 = name1[len];
 568         c2 = name2[len];
 569         if (!c1 && !c2)
 570                 /*
 571                  * git-write-tree used to write out a nonsense tree that has
 572                  * entries with the same name, one blob and one tree.  Make
 573                  * sure we do not have duplicate entries.
 574                  */
 575                 return TREE_HAS_DUPS;
 576         if (!c1 && S_ISDIR(mode1))
 577                 c1 = '/';
 578         if (!c2 && S_ISDIR(mode2))
 579                 c2 = '/';
 580         return c1 < c2 ? 0 : TREE_UNORDERED;
 581 }
 582
 583 static int fsck_tree(struct tree *item, struct fsck_options *options)
 584 {
 585         int retval = 0;
 586         int has_null_sha1 = 0;
 587         int has_full_path = 0;
 588         int has_empty_name = 0;
 589         int has_dot = 0;
 590         int has_dotdot = 0;
 591         int has_dotgit = 0;
 592         int has_zero_pad = 0;
 593         int has_bad_modes = 0;
 594         int has_dup_entries = 0;
 595         int not_properly_sorted = 0;
 596         struct tree_desc desc;
 597         unsigned o_mode;
 598         const char *o_name;
 599
 600         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 601                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 602                 return retval;
 603         }
 604
 605         o_mode = 0;
 606         o_name = NULL;
 607
 608         while (desc.size) {
 609                 unsigned mode;
 610                 const char *name;
 611                 const struct object_id *oid;
 612
 613                 oid = tree_entry_extract(&desc, &name, &mode);
 614
 615                 has_null_sha1 |= is_null_oid(oid);
 616                 has_full_path |= !!strchr(name, '/');
 617                 has_empty_name |= !*name;
 618                 has_dot |= !strcmp(name, ".");
 619                 has_dotdot |= !strcmp(name, "..");
 620                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 621                 has_zero_pad |= *(char *)desc.buffer == '0';
 622
 623                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 624                         if (!S_ISLNK(mode))
 625                                 oidset_insert(&gitmodules_found, oid);
 626                         else
 627                                 retval += report(options, &item->object,
 628                                                  FSCK_MSG_GITMODULES_SYMLINK,
 629                                                  ".gitmodules is a symbolic link");
 630                 }
 631
 632                 if (update_tree_entry_gently(&desc)) {
 633                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 634                         break;
 635                 }
 636
 637                 switch (mode) {
 638                 /*
 639                  * Standard modes..
 640                  */
 641                 case S_IFREG | 0755:
 642                 case S_IFREG | 0644:
 643                 case S_IFLNK:
 644                 case S_IFDIR:
 645                 case S_IFGITLINK:
 646                         break;
 647                 /*
 648                  * This is nonstandard, but we had a few of these
 649                  * early on when we honored the full set of mode
 650                  * bits..
 651                  */
 652                 case S_IFREG | 0664:
 653                         if (!options->strict)
 654                                 break;
 655                         /* fallthrough */
 656                 default:
 657                         has_bad_modes = 1;
 658                 }
 659
 660                 if (o_name) {
 661                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 662                         case TREE_UNORDERED:
 663                                 not_properly_sorted = 1;
 664                                 break;
 665                         case TREE_HAS_DUPS:
 666                                 has_dup_entries = 1;
 667                                 break;
 668                         default:
 669                                 break;
 670                         }
 671                 }
 672
 673                 o_mode = mode;
 674                 o_name = name;
 675         }
 676
 677         if (has_null_sha1)
 678                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 679         if (has_full_path)
 680                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 681         if (has_empty_name)
 682                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 683         if (has_dot)
 684                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 685         if (has_dotdot)
 686                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 687         if (has_dotgit)
 688                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 689         if (has_zero_pad)
 690                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 691         if (has_bad_modes)
 692                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 693         if (has_dup_entries)
 694                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 695         if (not_properly_sorted)
 696                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 697         return retval;
 698 }
 699
 700 static int verify_headers(const void *data, unsigned long size,
 701                           struct object *obj, struct fsck_options *options)
 702 {
 703         const char *buffer = (const char *)data;
 704         unsigned long i;
 705
 706         for (i = 0; i < size; i++) {
 707                 switch (buffer[i]) {
 708                 case '\0':
 709                         return report(options, obj,
 710                                 FSCK_MSG_NUL_IN_HEADER,
 711                                 "unterminated header: NUL at offset %ld", i);
 712                 case '\n':
 713                         if (i + 1 < size && buffer[i + 1] == '\n')
 714                                 return 0;
 715                 }
 716         }
 717
 718         /*
 719          * We did not find double-LF that separates the header
 720          * and the body.  Not having a body is not a crime but
 721          * we do want to see the terminating LF for the last header
 722          * line.
 723          */
 724         if (size && buffer[size - 1] == '\n')
 725                 return 0;
 726
 727         return report(options, obj,
 728                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 729 }
 730
 731 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 732 {
 733         const char *p = *ident;
 734         char *end;
 735
 736         *ident = strchrnul(*ident, '\n');
 737         if (**ident == '\n')
 738                 (*ident)++;
 739
 740         if (*p == '<')
 741                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 742         p += strcspn(p, "<>\n");
 743         if (*p == '>')
 744                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 745         if (*p != '<')
 746                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 747         if (p[-1] != ' ')
 748                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 749         p++;
 750         p += strcspn(p, "<>\n");
 751         if (*p != '>')
 752                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 753         p++;
 754         if (*p != ' ')
 755                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 756         p++;
 757         if (*p == '0' && p[1] != ' ')
 758                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 759         if (date_overflows(parse_timestamp(p, &end, 10)))
 760                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 761         if ((end == p || *end != ' '))
 762                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 763         p = end + 1;
 764         if ((*p != '+' && *p != '-') ||
 765             !isdigit(p[1]) ||
 766             !isdigit(p[2]) ||
 767             !isdigit(p[3]) ||
 768             !isdigit(p[4]) ||
 769             (p[5] != '\n'))
 770                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 771         p += 6;
 772         return 0;
 773 }
 774
 775 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 776         unsigned long size, struct fsck_options *options)
 777 {
 778         struct object_id tree_oid, oid;
 779         struct commit_graft *graft;
 780         unsigned parent_count, parent_line_count = 0, author_count;
 781         int err;
 782         const char *buffer_begin = buffer;
 783         const char *p;
 784
 785         if (verify_headers(buffer, size, &commit->object, options))
 786                 return -1;
 787
 788         if (!skip_prefix(buffer, "tree ", &buffer))
 789                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 790         if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 791                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 792                 if (err)
 793                         return err;
 794         }
 795         buffer = p + 1;
 796         while (skip_prefix(buffer, "parent ", &buffer)) {
 797                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 798                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 799                         if (err)
 800                                 return err;
 801                 }
 802                 buffer = p + 1;
 803                 parent_line_count++;
 804         }
 805         graft = lookup_commit_graft(the_repository, &commit->object.oid);
 806         parent_count = commit_list_count(commit->parents);
 807         if (graft) {
 808                 if (graft->nr_parent == -1 && !parent_count)
 809                         ; /* shallow commit */
 810                 else if (graft->nr_parent != parent_count) {
 811                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 812                         if (err)
 813                                 return err;
 814                 }
 815         } else {
 816                 if (parent_count != parent_line_count) {
 817                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 818                         if (err)
 819                                 return err;
 820                 }
 821         }
 822         author_count = 0;
 823         while (skip_prefix(buffer, "author ", &buffer)) {
 824                 author_count++;
 825                 err = fsck_ident(&buffer, &commit->object, options);
 826                 if (err)
 827                         return err;
 828         }
 829         if (author_count < 1)
 830                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 831         else if (author_count > 1)
 832                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 833         if (err)
 834                 return err;
 835         if (!skip_prefix(buffer, "committer ", &buffer))
 836                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 837         err = fsck_ident(&buffer, &commit->object, options);
 838         if (err)
 839                 return err;
 840         if (!get_commit_tree(commit)) {
 841                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
 842                 if (err)
 843                         return err;
 844         }
 845         if (memchr(buffer_begin, '\0', size)) {
 846                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 847                              "NUL byte in the commit object body");
 848                 if (err)
 849                         return err;
 850         }
 851         return 0;
 852 }
 853
 854 static int fsck_commit(struct commit *commit, const char *data,
 855         unsigned long size, struct fsck_options *options)
 856 {
 857         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 858         int ret = fsck_commit_buffer(commit, buffer, size, options);
 859         if (!data)
 860                 unuse_commit_buffer(commit, buffer);
 861         return ret;
 862 }
 863
 864 static int fsck_tag_buffer(struct tag *tag, const char *data,
 865         unsigned long size, struct fsck_options *options)
 866 {
 867         struct object_id oid;
 868         int ret = 0;
 869         const char *buffer;
 870         char *to_free = NULL, *eol;
 871         struct strbuf sb = STRBUF_INIT;
 872         const char *p;
 873
 874         if (data)
 875                 buffer = data;
 876         else {
 877                 enum object_type type;
 878
 879                 buffer = to_free =
 880                         read_object_file(&tag->object.oid, &type, &size);
 881                 if (!buffer)
 882                         return report(options, &tag->object,
 883                                 FSCK_MSG_MISSING_TAG_OBJECT,
 884                                 "cannot read tag object");
 885
 886                 if (type != OBJ_TAG) {
 887                         ret = report(options, &tag->object,
 888                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 889                                 "expected tag got %s",
 890                             type_name(type));
 891                         goto done;
 892                 }
 893         }
 894
 895         ret = verify_headers(buffer, size, &tag->object, options);
 896         if (ret)
 897                 goto done;
 898
 899         if (!skip_prefix(buffer, "object ", &buffer)) {
 900                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 901                 goto done;
 902         }
 903         if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 904                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 905                 if (ret)
 906                         goto done;
 907         }
 908         buffer = p + 1;
 909
 910         if (!skip_prefix(buffer, "type ", &buffer)) {
 911                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 912                 goto done;
 913         }
 914         eol = strchr(buffer, '\n');
 915         if (!eol) {
 916                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 917                 goto done;
 918         }
 919         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 920                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 921         if (ret)
 922                 goto done;
 923         buffer = eol + 1;
 924
 925         if (!skip_prefix(buffer, "tag ", &buffer)) {
 926                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 927                 goto done;
 928         }
 929         eol = strchr(buffer, '\n');
 930         if (!eol) {
 931                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 932                 goto done;
 933         }
 934         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 935         if (check_refname_format(sb.buf, 0)) {
 936                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 937                            "invalid 'tag' name: %.*s",
 938                            (int)(eol - buffer), buffer);
 939                 if (ret)
 940                         goto done;
 941         }
 942         buffer = eol + 1;
 943
 944         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 945                 /* early tags do not contain 'tagger' lines; warn only */
 946                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 947                 if (ret)
 948                         goto done;
 949         }
 950         else
 951                 ret = fsck_ident(&buffer, &tag->object, options);
 952
 953 done:
 954         strbuf_release(&sb);
 955         free(to_free);
 956         return ret;
 957 }
 958
 959 static int fsck_tag(struct tag *tag, const char *data,
 960         unsigned long size, struct fsck_options *options)
 961 {
 962         struct object *tagged = tag->tagged;
 963
 964         if (!tagged)
 965                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 966
 967         return fsck_tag_buffer(tag, data, size, options);
 968 }
 969
 970 struct fsck_gitmodules_data {
 971         struct object *obj;
 972         struct fsck_options *options;
 973         int ret;
 974 };
 975
 976 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 977 {
 978         struct fsck_gitmodules_data *data = vdata;
 979         const char *subsection, *key;
 980         int subsection_len;
 981         char *name;
 982
 983         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 984             !subsection)
 985                 return 0;
 986
 987         name = xmemdupz(subsection, subsection_len);
 988         if (check_submodule_name(name) < 0)
 989                 data->ret |= report(data->options, data->obj,
 990                                     FSCK_MSG_GITMODULES_NAME,
 991                                     "disallowed submodule name: %s",
 992                                     name);
 993         free(name);
 994
 995         return 0;
 996 }
 997
 998 static int fsck_blob(struct blob *blob, const char *buf,
 999                      unsigned long size, struct fsck_options *options)
1000 {
1001         struct fsck_gitmodules_data data;
1002
1003         if (!oidset_contains(&gitmodules_found, &blob->object.oid))
1004                 return 0;
1005         oidset_insert(&gitmodules_done, &blob->object.oid);
1006
1007         if (object_on_skiplist(options, &blob->object))
1008                 return 0;
1009
1010         if (!buf) {
1011                 /*
1012                  * A missing buffer here is a sign that the caller found the
1013                  * blob too gigantic to load into memory. Let's just consider
1014                  * that an error.
1015                  */
1016                 return report(options, &blob->object,
1017                               FSCK_MSG_GITMODULES_PARSE,
1018                               ".gitmodules too large to parse");
1019         }
1020
1021         data.obj = &blob->object;
1022         data.options = options;
1023         data.ret = 0;
1024         if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1025                                 ".gitmodules", buf, size, &data))
1026                 data.ret |= report(options, &blob->object,
1027                                    FSCK_MSG_GITMODULES_PARSE,
1028                                    "could not parse gitmodules blob");
1029
1030         return data.ret;
1031 }
1032
1033 int fsck_object(struct object *obj, void *data, unsigned long size,
1034         struct fsck_options *options)
1035 {
1036         if (!obj)
1037                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1038
1039         if (obj->type == OBJ_BLOB)
1040                 return fsck_blob((struct blob *)obj, data, size, options);
1041         if (obj->type == OBJ_TREE)
1042                 return fsck_tree((struct tree *) obj, options);
1043         if (obj->type == OBJ_COMMIT)
1044                 return fsck_commit((struct commit *) obj, (const char *) data,
1045                         size, options);
1046         if (obj->type == OBJ_TAG)
1047                 return fsck_tag((struct tag *) obj, (const char *) data,
1048                         size, options);
1049
1050         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1051                           obj->type);
1052 }
1053
1054 int fsck_error_function(struct fsck_options *o,
1055         struct object *obj, int msg_type, const char *message)
1056 {
1057         if (msg_type == FSCK_WARN) {
1058                 warning("object %s: %s", describe_object(o, obj), message);
1059                 return 0;
1060         }
1061         error("object %s: %s", describe_object(o, obj), message);
1062         return 1;
1063 }
1064
1065 int fsck_finish(struct fsck_options *options)
1066 {
1067         int ret = 0;
1068         struct oidset_iter iter;
1069         const struct object_id *oid;
1070
1071         oidset_iter_init(&gitmodules_found, &iter);
1072         while ((oid = oidset_iter_next(&iter))) {
1073                 struct blob *blob;
1074                 enum object_type type;
1075                 unsigned long size;
1076                 char *buf;
1077
1078                 if (oidset_contains(&gitmodules_done, oid))
1079                         continue;
1080
1081                 blob = lookup_blob(oid);
1082                 if (!blob) {
1083                         struct object *obj = lookup_unknown_object(oid->hash);
1084                         ret |= report(options, obj,
1085                                       FSCK_MSG_GITMODULES_BLOB,
1086                                       "non-blob found at .gitmodules");
1087                         continue;
1088                 }
1089
1090                 buf = read_object_file(oid, &type, &size);
1091                 if (!buf) {
1092                         if (is_promisor_object(&blob->object.oid))
1093                                 continue;
1094                         ret |= report(options, &blob->object,
1095                                       FSCK_MSG_GITMODULES_MISSING,
1096                                       "unable to read .gitmodules blob");
1097                         continue;
1098                 }
1099
1100                 if (type == OBJ_BLOB)
1101                         ret |= fsck_blob(blob, buf, size, options);
1102                 else
1103                         ret |= report(options, &blob->object,
1104                                       FSCK_MSG_GITMODULES_BLOB,
1105                                       "non-blob found at .gitmodules");
1106                 free(buf);
1107         }
1108
1109
1110         oidset_clear(&gitmodules_found);
1111         oidset_clear(&gitmodules_done);
1112         return ret;
1113 }