fsck.c

   1 #include "cache.h"
   2 #include "object.h"
   3 #include "blob.h"
   4 #include "tree.h"
   5 #include "tree-walk.h"
   6 #include "commit.h"
   7 #include "tag.h"
   8 #include "fsck.h"
   9 #include "refs.h"
  10 #include "utf8.h"
  11 #include "sha1-array.h"
  12 #include "decorate.h"
  13
  14 #define FSCK_FATAL -1
  15 #define FSCK_INFO -2
  16
  17 #define FOREACH_MSG_ID(FUNC) \
  18         /* fatal errors */ \
  19         FUNC(NUL_IN_HEADER, FATAL) \
  20         FUNC(UNTERMINATED_HEADER, FATAL) \
  21         /* errors */ \
  22         FUNC(BAD_DATE, ERROR) \
  23         FUNC(BAD_DATE_OVERFLOW, ERROR) \
  24         FUNC(BAD_EMAIL, ERROR) \
  25         FUNC(BAD_NAME, ERROR) \
  26         FUNC(BAD_OBJECT_SHA1, ERROR) \
  27         FUNC(BAD_PARENT_SHA1, ERROR) \
  28         FUNC(BAD_TAG_OBJECT, ERROR) \
  29         FUNC(BAD_TIMEZONE, ERROR) \
  30         FUNC(BAD_TREE, ERROR) \
  31         FUNC(BAD_TREE_SHA1, ERROR) \
  32         FUNC(BAD_TYPE, ERROR) \
  33         FUNC(DUPLICATE_ENTRIES, ERROR) \
  34         FUNC(MISSING_AUTHOR, ERROR) \
  35         FUNC(MISSING_COMMITTER, ERROR) \
  36         FUNC(MISSING_EMAIL, ERROR) \
  37         FUNC(MISSING_GRAFT, ERROR) \
  38         FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  39         FUNC(MISSING_OBJECT, ERROR) \
  40         FUNC(MISSING_PARENT, ERROR) \
  41         FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  42         FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  43         FUNC(MISSING_TAG, ERROR) \
  44         FUNC(MISSING_TAG_ENTRY, ERROR) \
  45         FUNC(MISSING_TAG_OBJECT, ERROR) \
  46         FUNC(MISSING_TREE, ERROR) \
  47         FUNC(MISSING_TYPE, ERROR) \
  48         FUNC(MISSING_TYPE_ENTRY, ERROR) \
  49         FUNC(MULTIPLE_AUTHORS, ERROR) \
  50         FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  51         FUNC(TREE_NOT_SORTED, ERROR) \
  52         FUNC(UNKNOWN_TYPE, ERROR) \
  53         FUNC(ZERO_PADDED_DATE, ERROR) \
  54         /* warnings */ \
  55         FUNC(BAD_FILEMODE, WARN) \
  56         FUNC(EMPTY_NAME, WARN) \
  57         FUNC(FULL_PATHNAME, WARN) \
  58         FUNC(HAS_DOT, WARN) \
  59         FUNC(HAS_DOTDOT, WARN) \
  60         FUNC(HAS_DOTGIT, WARN) \
  61         FUNC(NULL_SHA1, WARN) \
  62         FUNC(ZERO_PADDED_FILEMODE, WARN) \
  63         FUNC(NUL_IN_COMMIT, WARN) \
  64         /* infos (reported as warnings, but ignored by default) */ \
  65         FUNC(BAD_TAG_NAME, INFO) \
  66         FUNC(MISSING_TAGGER_ENTRY, INFO)
  67
  68 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
  69 enum fsck_msg_id {
  70         FOREACH_MSG_ID(MSG_ID)
  71         FSCK_MSG_MAX
  72 };
  73 #undef MSG_ID
  74
  75 #define STR(x) #x
  76 #define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
  77 static struct {
  78         const char *id_string;
  79         const char *downcased;
  80         int msg_type;
  81 } msg_id_info[FSCK_MSG_MAX + 1] = {
  82         FOREACH_MSG_ID(MSG_ID)
  83         { NULL, NULL, -1 }
  84 };
  85 #undef MSG_ID
  86
  87 static int parse_msg_id(const char *text)
  88 {
  89         int i;
  90
  91         if (!msg_id_info[0].downcased) {
  92                 /* convert id_string to lower case, without underscores. */
  93                 for (i = 0; i < FSCK_MSG_MAX; i++) {
  94                         const char *p = msg_id_info[i].id_string;
  95                         int len = strlen(p);
  96                         char *q = xmalloc(len);
  97
  98                         msg_id_info[i].downcased = q;
  99                         while (*p)
 100                                 if (*p == '_')
 101                                         p++;
 102                                 else
 103                                         *(q)++ = tolower(*(p)++);
 104                         *q = '\0';
 105                 }
 106         }
 107
 108         for (i = 0; i < FSCK_MSG_MAX; i++)
 109                 if (!strcmp(text, msg_id_info[i].downcased))
 110                         return i;
 111
 112         return -1;
 113 }
 114
 115 static int fsck_msg_type(enum fsck_msg_id msg_id,
 116         struct fsck_options *options)
 117 {
 118         int msg_type;
 119
 120         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 121
 122         if (options->msg_type)
 123                 msg_type = options->msg_type[msg_id];
 124         else {
 125                 msg_type = msg_id_info[msg_id].msg_type;
 126                 if (options->strict && msg_type == FSCK_WARN)
 127                         msg_type = FSCK_ERROR;
 128         }
 129
 130         return msg_type;
 131 }
 132
 133 static void init_skiplist(struct fsck_options *options, const char *path)
 134 {
 135         static struct oid_array skiplist = OID_ARRAY_INIT;
 136         int sorted, fd;
 137         char buffer[GIT_MAX_HEXSZ + 1];
 138         struct object_id oid;
 139
 140         if (options->skiplist)
 141                 sorted = options->skiplist->sorted;
 142         else {
 143                 sorted = 1;
 144                 options->skiplist = &skiplist;
 145         }
 146
 147         fd = open(path, O_RDONLY);
 148         if (fd < 0)
 149                 die("Could not open skip list: %s", path);
 150         for (;;) {
 151                 const char *p;
 152                 int result = read_in_full(fd, buffer, sizeof(buffer));
 153                 if (result < 0)
 154                         die_errno("Could not read '%s'", path);
 155                 if (!result)
 156                         break;
 157                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 158                         die("Invalid SHA-1: %s", buffer);
 159                 oid_array_append(&skiplist, &oid);
 160                 if (sorted && skiplist.nr > 1 &&
 161                                 oidcmp(&skiplist.oid[skiplist.nr - 2],
 162                                        &oid) > 0)
 163                         sorted = 0;
 164         }
 165         close(fd);
 166
 167         if (sorted)
 168                 skiplist.sorted = 1;
 169 }
 170
 171 static int parse_msg_type(const char *str)
 172 {
 173         if (!strcmp(str, "error"))
 174                 return FSCK_ERROR;
 175         else if (!strcmp(str, "warn"))
 176                 return FSCK_WARN;
 177         else if (!strcmp(str, "ignore"))
 178                 return FSCK_IGNORE;
 179         else
 180                 die("Unknown fsck message type: '%s'", str);
 181 }
 182
 183 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 184 {
 185         if (parse_msg_id(msg_id) < 0)
 186                 return 0;
 187         parse_msg_type(msg_type);
 188         return 1;
 189 }
 190
 191 void fsck_set_msg_type(struct fsck_options *options,
 192                 const char *msg_id, const char *msg_type)
 193 {
 194         int id = parse_msg_id(msg_id), type;
 195
 196         if (id < 0)
 197                 die("Unhandled message id: %s", msg_id);
 198         type = parse_msg_type(msg_type);
 199
 200         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 201                 die("Cannot demote %s to %s", msg_id, msg_type);
 202
 203         if (!options->msg_type) {
 204                 int i;
 205                 int *msg_type;
 206                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 207                 for (i = 0; i < FSCK_MSG_MAX; i++)
 208                         msg_type[i] = fsck_msg_type(i, options);
 209                 options->msg_type = msg_type;
 210         }
 211
 212         options->msg_type[id] = type;
 213 }
 214
 215 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 216 {
 217         char *buf = xstrdup(values), *to_free = buf;
 218         int done = 0;
 219
 220         while (!done) {
 221                 int len = strcspn(buf, " ,|"), equal;
 222
 223                 done = !buf[len];
 224                 if (!len) {
 225                         buf++;
 226                         continue;
 227                 }
 228                 buf[len] = '\0';
 229
 230                 for (equal = 0;
 231                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 232                      equal++)
 233                         buf[equal] = tolower(buf[equal]);
 234                 buf[equal] = '\0';
 235
 236                 if (!strcmp(buf, "skiplist")) {
 237                         if (equal == len)
 238                                 die("skiplist requires a path");
 239                         init_skiplist(options, buf + equal + 1);
 240                         buf += len + 1;
 241                         continue;
 242                 }
 243
 244                 if (equal == len)
 245                         die("Missing '=': '%s'", buf);
 246
 247                 fsck_set_msg_type(options, buf, buf + equal + 1);
 248                 buf += len + 1;
 249         }
 250         free(to_free);
 251 }
 252
 253 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 254 {
 255         for (;;) {
 256                 char c = *(msg_id)++;
 257
 258                 if (!c)
 259                         break;
 260                 if (c != '_')
 261                         strbuf_addch(sb, tolower(c));
 262                 else {
 263                         assert(*msg_id);
 264                         strbuf_addch(sb, *(msg_id)++);
 265                 }
 266         }
 267
 268         strbuf_addstr(sb, ": ");
 269 }
 270
 271 __attribute__((format (printf, 4, 5)))
 272 static int report(struct fsck_options *options, struct object *object,
 273         enum fsck_msg_id id, const char *fmt, ...)
 274 {
 275         va_list ap;
 276         struct strbuf sb = STRBUF_INIT;
 277         int msg_type = fsck_msg_type(id, options), result;
 278
 279         if (msg_type == FSCK_IGNORE)
 280                 return 0;
 281
 282         if (options->skiplist && object &&
 283                         oid_array_lookup(options->skiplist, &object->oid) >= 0)
 284                 return 0;
 285
 286         if (msg_type == FSCK_FATAL)
 287                 msg_type = FSCK_ERROR;
 288         else if (msg_type == FSCK_INFO)
 289                 msg_type = FSCK_WARN;
 290
 291         append_msg_id(&sb, msg_id_info[id].id_string);
 292
 293         va_start(ap, fmt);
 294         strbuf_vaddf(&sb, fmt, ap);
 295         result = options->error_func(options, object, msg_type, sb.buf);
 296         strbuf_release(&sb);
 297         va_end(ap);
 298
 299         return result;
 300 }
 301
 302 static char *get_object_name(struct fsck_options *options, struct object *obj)
 303 {
 304         if (!options->object_names)
 305                 return NULL;
 306         return lookup_decoration(options->object_names, obj);
 307 }
 308
 309 static void put_object_name(struct fsck_options *options, struct object *obj,
 310         const char *fmt, ...)
 311 {
 312         va_list ap;
 313         struct strbuf buf = STRBUF_INIT;
 314         char *existing;
 315
 316         if (!options->object_names)
 317                 return;
 318         existing = lookup_decoration(options->object_names, obj);
 319         if (existing)
 320                 return;
 321         va_start(ap, fmt);
 322         strbuf_vaddf(&buf, fmt, ap);
 323         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 324         va_end(ap);
 325 }
 326
 327 static const char *describe_object(struct fsck_options *o, struct object *obj)
 328 {
 329         static struct strbuf buf = STRBUF_INIT;
 330         char *name;
 331
 332         strbuf_reset(&buf);
 333         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 334         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 335                 strbuf_addf(&buf, " (%s)", name);
 336
 337         return buf.buf;
 338 }
 339
 340 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 341 {
 342         struct tree_desc desc;
 343         struct name_entry entry;
 344         int res = 0;
 345         const char *name;
 346
 347         if (parse_tree(tree))
 348                 return -1;
 349
 350         name = get_object_name(options, &tree->object);
 351         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 352                 return -1;
 353         while (tree_entry_gently(&desc, &entry)) {
 354                 struct object *obj;
 355                 int result;
 356
 357                 if (S_ISGITLINK(entry.mode))
 358                         continue;
 359
 360                 if (S_ISDIR(entry.mode)) {
 361                         obj = (struct object *)lookup_tree(entry.oid);
 362                         if (name && obj)
 363                                 put_object_name(options, obj, "%s%s/", name,
 364                                         entry.path);
 365                         result = options->walk(obj, OBJ_TREE, data, options);
 366                 }
 367                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 368                         obj = (struct object *)lookup_blob(entry.oid);
 369                         if (name && obj)
 370                                 put_object_name(options, obj, "%s%s", name,
 371                                         entry.path);
 372                         result = options->walk(obj, OBJ_BLOB, data, options);
 373                 }
 374                 else {
 375                         result = error("in tree %s: entry %s has bad mode %.6o",
 376                                         describe_object(options, &tree->object), entry.path, entry.mode);
 377                 }
 378                 if (result < 0)
 379                         return result;
 380                 if (!res)
 381                         res = result;
 382         }
 383         return res;
 384 }
 385
 386 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 387 {
 388         int counter = 0, generation = 0, name_prefix_len = 0;
 389         struct commit_list *parents;
 390         int res;
 391         int result;
 392         const char *name;
 393
 394         if (parse_commit(commit))
 395                 return -1;
 396
 397         name = get_object_name(options, &commit->object);
 398         if (name)
 399                 put_object_name(options, &commit->tree->object, "%s:", name);
 400
 401         result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options);
 402         if (result < 0)
 403                 return result;
 404         res = result;
 405
 406         parents = commit->parents;
 407         if (name && parents) {
 408                 int len = strlen(name), power;
 409
 410                 if (len && name[len - 1] == '^') {
 411                         generation = 1;
 412                         name_prefix_len = len - 1;
 413                 }
 414                 else { /* parse ~<generation> suffix */
 415                         for (generation = 0, power = 1;
 416                              len && isdigit(name[len - 1]);
 417                              power *= 10)
 418                                 generation += power * (name[--len] - '0');
 419                         if (power > 1 && len && name[len - 1] == '~')
 420                                 name_prefix_len = len - 1;
 421                 }
 422         }
 423
 424         while (parents) {
 425                 if (name) {
 426                         struct object *obj = &parents->item->object;
 427
 428                         if (++counter > 1)
 429                                 put_object_name(options, obj, "%s^%d",
 430                                         name, counter);
 431                         else if (generation > 0)
 432                                 put_object_name(options, obj, "%.*s~%d",
 433                                         name_prefix_len, name, generation + 1);
 434                         else
 435                                 put_object_name(options, obj, "%s^", name);
 436                 }
 437                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 438                 if (result < 0)
 439                         return result;
 440                 if (!res)
 441                         res = result;
 442                 parents = parents->next;
 443         }
 444         return res;
 445 }
 446
 447 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 448 {
 449         char *name = get_object_name(options, &tag->object);
 450
 451         if (parse_tag(tag))
 452                 return -1;
 453         if (name)
 454                 put_object_name(options, tag->tagged, "%s", name);
 455         return options->walk(tag->tagged, OBJ_ANY, data, options);
 456 }
 457
 458 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 459 {
 460         if (!obj)
 461                 return -1;
 462
 463         if (obj->type == OBJ_NONE)
 464                 parse_object(&obj->oid);
 465
 466         switch (obj->type) {
 467         case OBJ_BLOB:
 468                 return 0;
 469         case OBJ_TREE:
 470                 return fsck_walk_tree((struct tree *)obj, data, options);
 471         case OBJ_COMMIT:
 472                 return fsck_walk_commit((struct commit *)obj, data, options);
 473         case OBJ_TAG:
 474                 return fsck_walk_tag((struct tag *)obj, data, options);
 475         default:
 476                 error("Unknown object type for %s", describe_object(options, obj));
 477                 return -1;
 478         }
 479 }
 480
 481 /*
 482  * The entries in a tree are ordered in the _path_ order,
 483  * which means that a directory entry is ordered by adding
 484  * a slash to the end of it.
 485  *
 486  * So a directory called "a" is ordered _after_ a file
 487  * called "a.c", because "a/" sorts after "a.c".
 488  */
 489 #define TREE_UNORDERED (-1)
 490 #define TREE_HAS_DUPS  (-2)
 491
 492 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 493 {
 494         int len1 = strlen(name1);
 495         int len2 = strlen(name2);
 496         int len = len1 < len2 ? len1 : len2;
 497         unsigned char c1, c2;
 498         int cmp;
 499
 500         cmp = memcmp(name1, name2, len);
 501         if (cmp < 0)
 502                 return 0;
 503         if (cmp > 0)
 504                 return TREE_UNORDERED;
 505
 506         /*
 507          * Ok, the first <len> characters are the same.
 508          * Now we need to order the next one, but turn
 509          * a '\0' into a '/' for a directory entry.
 510          */
 511         c1 = name1[len];
 512         c2 = name2[len];
 513         if (!c1 && !c2)
 514                 /*
 515                  * git-write-tree used to write out a nonsense tree that has
 516                  * entries with the same name, one blob and one tree.  Make
 517                  * sure we do not have duplicate entries.
 518                  */
 519                 return TREE_HAS_DUPS;
 520         if (!c1 && S_ISDIR(mode1))
 521                 c1 = '/';
 522         if (!c2 && S_ISDIR(mode2))
 523                 c2 = '/';
 524         return c1 < c2 ? 0 : TREE_UNORDERED;
 525 }
 526
 527 static int fsck_tree(struct tree *item, struct fsck_options *options)
 528 {
 529         int retval = 0;
 530         int has_null_sha1 = 0;
 531         int has_full_path = 0;
 532         int has_empty_name = 0;
 533         int has_dot = 0;
 534         int has_dotdot = 0;
 535         int has_dotgit = 0;
 536         int has_zero_pad = 0;
 537         int has_bad_modes = 0;
 538         int has_dup_entries = 0;
 539         int not_properly_sorted = 0;
 540         struct tree_desc desc;
 541         unsigned o_mode;
 542         const char *o_name;
 543
 544         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 545                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 546                 return retval;
 547         }
 548
 549         o_mode = 0;
 550         o_name = NULL;
 551
 552         while (desc.size) {
 553                 unsigned mode;
 554                 const char *name, *backslash;
 555                 const struct object_id *oid;
 556
 557                 oid = tree_entry_extract(&desc, &name, &mode);
 558
 559                 has_null_sha1 |= is_null_oid(oid);
 560                 has_full_path |= !!strchr(name, '/');
 561                 has_empty_name |= !*name;
 562                 has_dot |= !strcmp(name, ".");
 563                 has_dotdot |= !strcmp(name, "..");
 564                 has_dotgit |= (!strcmp(name, ".git") ||
 565                                is_hfs_dotgit(name) ||
 566                                is_ntfs_dotgit(name));
 567                 has_zero_pad |= *(char *)desc.buffer == '0';
 568
 569                 if ((backslash = strchr(name, '\\'))) {
 570                         while (backslash) {
 571                                 backslash++;
 572                                 has_dotgit |= is_ntfs_dotgit(backslash);
 573                                 backslash = strchr(backslash, '\\');
 574                         }
 575                 }
 576
 577                 if (update_tree_entry_gently(&desc)) {
 578                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 579                         break;
 580                 }
 581
 582                 switch (mode) {
 583                 /*
 584                  * Standard modes..
 585                  */
 586                 case S_IFREG | 0755:
 587                 case S_IFREG | 0644:
 588                 case S_IFLNK:
 589                 case S_IFDIR:
 590                 case S_IFGITLINK:
 591                         break;
 592                 /*
 593                  * This is nonstandard, but we had a few of these
 594                  * early on when we honored the full set of mode
 595                  * bits..
 596                  */
 597                 case S_IFREG | 0664:
 598                         if (!options->strict)
 599                                 break;
 600                 default:
 601                         has_bad_modes = 1;
 602                 }
 603
 604                 if (o_name) {
 605                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 606                         case TREE_UNORDERED:
 607                                 not_properly_sorted = 1;
 608                                 break;
 609                         case TREE_HAS_DUPS:
 610                                 has_dup_entries = 1;
 611                                 break;
 612                         default:
 613                                 break;
 614                         }
 615                 }
 616
 617                 o_mode = mode;
 618                 o_name = name;
 619         }
 620
 621         if (has_null_sha1)
 622                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 623         if (has_full_path)
 624                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 625         if (has_empty_name)
 626                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 627         if (has_dot)
 628                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 629         if (has_dotdot)
 630                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 631         if (has_dotgit)
 632                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 633         if (has_zero_pad)
 634                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 635         if (has_bad_modes)
 636                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 637         if (has_dup_entries)
 638                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 639         if (not_properly_sorted)
 640                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 641         return retval;
 642 }
 643
 644 static int verify_headers(const void *data, unsigned long size,
 645                           struct object *obj, struct fsck_options *options)
 646 {
 647         const char *buffer = (const char *)data;
 648         unsigned long i;
 649
 650         for (i = 0; i < size; i++) {
 651                 switch (buffer[i]) {
 652                 case '\0':
 653                         return report(options, obj,
 654                                 FSCK_MSG_NUL_IN_HEADER,
 655                                 "unterminated header: NUL at offset %ld", i);
 656                 case '\n':
 657                         if (i + 1 < size && buffer[i + 1] == '\n')
 658                                 return 0;
 659                 }
 660         }
 661
 662         /*
 663          * We did not find double-LF that separates the header
 664          * and the body.  Not having a body is not a crime but
 665          * we do want to see the terminating LF for the last header
 666          * line.
 667          */
 668         if (size && buffer[size - 1] == '\n')
 669                 return 0;
 670
 671         return report(options, obj,
 672                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 673 }
 674
 675 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 676 {
 677         const char *p = *ident;
 678         char *end;
 679
 680         *ident = strchrnul(*ident, '\n');
 681         if (**ident == '\n')
 682                 (*ident)++;
 683
 684         if (*p == '<')
 685                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 686         p += strcspn(p, "<>\n");
 687         if (*p == '>')
 688                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 689         if (*p != '<')
 690                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 691         if (p[-1] != ' ')
 692                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 693         p++;
 694         p += strcspn(p, "<>\n");
 695         if (*p != '>')
 696                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 697         p++;
 698         if (*p != ' ')
 699                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 700         p++;
 701         if (*p == '0' && p[1] != ' ')
 702                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 703         if (date_overflows(parse_timestamp(p, &end, 10)))
 704                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 705         if ((end == p || *end != ' '))
 706                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 707         p = end + 1;
 708         if ((*p != '+' && *p != '-') ||
 709             !isdigit(p[1]) ||
 710             !isdigit(p[2]) ||
 711             !isdigit(p[3]) ||
 712             !isdigit(p[4]) ||
 713             (p[5] != '\n'))
 714                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 715         p += 6;
 716         return 0;
 717 }
 718
 719 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 720         unsigned long size, struct fsck_options *options)
 721 {
 722         unsigned char tree_sha1[20], sha1[20];
 723         struct commit_graft *graft;
 724         unsigned parent_count, parent_line_count = 0, author_count;
 725         int err;
 726         const char *buffer_begin = buffer;
 727
 728         if (verify_headers(buffer, size, &commit->object, options))
 729                 return -1;
 730
 731         if (!skip_prefix(buffer, "tree ", &buffer))
 732                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 733         if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
 734                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 735                 if (err)
 736                         return err;
 737         }
 738         buffer += 41;
 739         while (skip_prefix(buffer, "parent ", &buffer)) {
 740                 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 741                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 742                         if (err)
 743                                 return err;
 744                 }
 745                 buffer += 41;
 746                 parent_line_count++;
 747         }
 748         graft = lookup_commit_graft(commit->object.oid.hash);
 749         parent_count = commit_list_count(commit->parents);
 750         if (graft) {
 751                 if (graft->nr_parent == -1 && !parent_count)
 752                         ; /* shallow commit */
 753                 else if (graft->nr_parent != parent_count) {
 754                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 755                         if (err)
 756                                 return err;
 757                 }
 758         } else {
 759                 if (parent_count != parent_line_count) {
 760                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 761                         if (err)
 762                                 return err;
 763                 }
 764         }
 765         author_count = 0;
 766         while (skip_prefix(buffer, "author ", &buffer)) {
 767                 author_count++;
 768                 err = fsck_ident(&buffer, &commit->object, options);
 769                 if (err)
 770                         return err;
 771         }
 772         if (author_count < 1)
 773                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 774         else if (author_count > 1)
 775                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 776         if (err)
 777                 return err;
 778         if (!skip_prefix(buffer, "committer ", &buffer))
 779                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 780         err = fsck_ident(&buffer, &commit->object, options);
 781         if (err)
 782                 return err;
 783         if (!commit->tree) {
 784                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
 785                 if (err)
 786                         return err;
 787         }
 788         if (memchr(buffer_begin, '\0', size)) {
 789                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 790                              "NUL byte in the commit object body");
 791                 if (err)
 792                         return err;
 793         }
 794         return 0;
 795 }
 796
 797 static int fsck_commit(struct commit *commit, const char *data,
 798         unsigned long size, struct fsck_options *options)
 799 {
 800         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 801         int ret = fsck_commit_buffer(commit, buffer, size, options);
 802         if (!data)
 803                 unuse_commit_buffer(commit, buffer);
 804         return ret;
 805 }
 806
 807 static int fsck_tag_buffer(struct tag *tag, const char *data,
 808         unsigned long size, struct fsck_options *options)
 809 {
 810         unsigned char sha1[20];
 811         int ret = 0;
 812         const char *buffer;
 813         char *to_free = NULL, *eol;
 814         struct strbuf sb = STRBUF_INIT;
 815
 816         if (data)
 817                 buffer = data;
 818         else {
 819                 enum object_type type;
 820
 821                 buffer = to_free =
 822                         read_sha1_file(tag->object.oid.hash, &type, &size);
 823                 if (!buffer)
 824                         return report(options, &tag->object,
 825                                 FSCK_MSG_MISSING_TAG_OBJECT,
 826                                 "cannot read tag object");
 827
 828                 if (type != OBJ_TAG) {
 829                         ret = report(options, &tag->object,
 830                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 831                                 "expected tag got %s",
 832                             typename(type));
 833                         goto done;
 834                 }
 835         }
 836
 837         ret = verify_headers(buffer, size, &tag->object, options);
 838         if (ret)
 839                 goto done;
 840
 841         if (!skip_prefix(buffer, "object ", &buffer)) {
 842                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 843                 goto done;
 844         }
 845         if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 846                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 847                 if (ret)
 848                         goto done;
 849         }
 850         buffer += 41;
 851
 852         if (!skip_prefix(buffer, "type ", &buffer)) {
 853                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 854                 goto done;
 855         }
 856         eol = strchr(buffer, '\n');
 857         if (!eol) {
 858                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 859                 goto done;
 860         }
 861         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 862                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 863         if (ret)
 864                 goto done;
 865         buffer = eol + 1;
 866
 867         if (!skip_prefix(buffer, "tag ", &buffer)) {
 868                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 869                 goto done;
 870         }
 871         eol = strchr(buffer, '\n');
 872         if (!eol) {
 873                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 874                 goto done;
 875         }
 876         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 877         if (check_refname_format(sb.buf, 0)) {
 878                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 879                            "invalid 'tag' name: %.*s",
 880                            (int)(eol - buffer), buffer);
 881                 if (ret)
 882                         goto done;
 883         }
 884         buffer = eol + 1;
 885
 886         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 887                 /* early tags do not contain 'tagger' lines; warn only */
 888                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 889                 if (ret)
 890                         goto done;
 891         }
 892         else
 893                 ret = fsck_ident(&buffer, &tag->object, options);
 894
 895 done:
 896         strbuf_release(&sb);
 897         free(to_free);
 898         return ret;
 899 }
 900
 901 static int fsck_tag(struct tag *tag, const char *data,
 902         unsigned long size, struct fsck_options *options)
 903 {
 904         struct object *tagged = tag->tagged;
 905
 906         if (!tagged)
 907                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 908
 909         return fsck_tag_buffer(tag, data, size, options);
 910 }
 911
 912 int fsck_object(struct object *obj, void *data, unsigned long size,
 913         struct fsck_options *options)
 914 {
 915         if (!obj)
 916                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
 917
 918         if (obj->type == OBJ_BLOB)
 919                 return 0;
 920         if (obj->type == OBJ_TREE)
 921                 return fsck_tree((struct tree *) obj, options);
 922         if (obj->type == OBJ_COMMIT)
 923                 return fsck_commit((struct commit *) obj, (const char *) data,
 924                         size, options);
 925         if (obj->type == OBJ_TAG)
 926                 return fsck_tag((struct tag *) obj, (const char *) data,
 927                         size, options);
 928
 929         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
 930                           obj->type);
 931 }
 932
 933 int fsck_error_function(struct fsck_options *o,
 934         struct object *obj, int msg_type, const char *message)
 935 {
 936         if (msg_type == FSCK_WARN) {
 937                 warning("object %s: %s", describe_object(o, obj), message);
 938                 return 0;
 939         }
 940         error("object %s: %s", describe_object(o, obj), message);
 941         return 1;
 942 }