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 sha1_array skiplist = SHA1_ARRAY_INIT;
 136         int sorted, fd;
 137         char buffer[41];
 138         unsigned char sha1[20];
 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                 int result = read_in_full(fd, buffer, sizeof(buffer));
 152                 if (result < 0)
 153                         die_errno("Could not read '%s'", path);
 154                 if (!result)
 155                         break;
 156                 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n')
 157                         die("Invalid SHA-1: %s", buffer);
 158                 sha1_array_append(&skiplist, sha1);
 159                 if (sorted && skiplist.nr > 1 &&
 160                                 hashcmp(skiplist.sha1[skiplist.nr - 2],
 161                                         sha1) > 0)
 162                         sorted = 0;
 163         }
 164         close(fd);
 165
 166         if (sorted)
 167                 skiplist.sorted = 1;
 168 }
 169
 170 static int parse_msg_type(const char *str)
 171 {
 172         if (!strcmp(str, "error"))
 173                 return FSCK_ERROR;
 174         else if (!strcmp(str, "warn"))
 175                 return FSCK_WARN;
 176         else if (!strcmp(str, "ignore"))
 177                 return FSCK_IGNORE;
 178         else
 179                 die("Unknown fsck message type: '%s'", str);
 180 }
 181
 182 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 183 {
 184         if (parse_msg_id(msg_id) < 0)
 185                 return 0;
 186         parse_msg_type(msg_type);
 187         return 1;
 188 }
 189
 190 void fsck_set_msg_type(struct fsck_options *options,
 191                 const char *msg_id, const char *msg_type)
 192 {
 193         int id = parse_msg_id(msg_id), type;
 194
 195         if (id < 0)
 196                 die("Unhandled message id: %s", msg_id);
 197         type = parse_msg_type(msg_type);
 198
 199         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 200                 die("Cannot demote %s to %s", msg_id, msg_type);
 201
 202         if (!options->msg_type) {
 203                 int i;
 204                 int *msg_type;
 205                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 206                 for (i = 0; i < FSCK_MSG_MAX; i++)
 207                         msg_type[i] = fsck_msg_type(i, options);
 208                 options->msg_type = msg_type;
 209         }
 210
 211         options->msg_type[id] = type;
 212 }
 213
 214 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 215 {
 216         char *buf = xstrdup(values), *to_free = buf;
 217         int done = 0;
 218
 219         while (!done) {
 220                 int len = strcspn(buf, " ,|"), equal;
 221
 222                 done = !buf[len];
 223                 if (!len) {
 224                         buf++;
 225                         continue;
 226                 }
 227                 buf[len] = '\0';
 228
 229                 for (equal = 0;
 230                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 231                      equal++)
 232                         buf[equal] = tolower(buf[equal]);
 233                 buf[equal] = '\0';
 234
 235                 if (!strcmp(buf, "skiplist")) {
 236                         if (equal == len)
 237                                 die("skiplist requires a path");
 238                         init_skiplist(options, buf + equal + 1);
 239                         buf += len + 1;
 240                         continue;
 241                 }
 242
 243                 if (equal == len)
 244                         die("Missing '=': '%s'", buf);
 245
 246                 fsck_set_msg_type(options, buf, buf + equal + 1);
 247                 buf += len + 1;
 248         }
 249         free(to_free);
 250 }
 251
 252 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 253 {
 254         for (;;) {
 255                 char c = *(msg_id)++;
 256
 257                 if (!c)
 258                         break;
 259                 if (c != '_')
 260                         strbuf_addch(sb, tolower(c));
 261                 else {
 262                         assert(*msg_id);
 263                         strbuf_addch(sb, *(msg_id)++);
 264                 }
 265         }
 266
 267         strbuf_addstr(sb, ": ");
 268 }
 269
 270 __attribute__((format (printf, 4, 5)))
 271 static int report(struct fsck_options *options, struct object *object,
 272         enum fsck_msg_id id, const char *fmt, ...)
 273 {
 274         va_list ap;
 275         struct strbuf sb = STRBUF_INIT;
 276         int msg_type = fsck_msg_type(id, options), result;
 277
 278         if (msg_type == FSCK_IGNORE)
 279                 return 0;
 280
 281         if (options->skiplist && object &&
 282                         sha1_array_lookup(options->skiplist, object->oid.hash) >= 0)
 283                 return 0;
 284
 285         if (msg_type == FSCK_FATAL)
 286                 msg_type = FSCK_ERROR;
 287         else if (msg_type == FSCK_INFO)
 288                 msg_type = FSCK_WARN;
 289
 290         append_msg_id(&sb, msg_id_info[id].id_string);
 291
 292         va_start(ap, fmt);
 293         strbuf_vaddf(&sb, fmt, ap);
 294         result = options->error_func(object, msg_type, sb.buf);
 295         strbuf_release(&sb);
 296         va_end(ap);
 297
 298         return result;
 299 }
 300
 301 static char *get_object_name(struct fsck_options *options, struct object *obj)
 302 {
 303         if (!options->object_names)
 304                 return NULL;
 305         return lookup_decoration(options->object_names, obj);
 306 }
 307
 308 static void put_object_name(struct fsck_options *options, struct object *obj,
 309         const char *fmt, ...)
 310 {
 311         va_list ap;
 312         struct strbuf buf = STRBUF_INIT;
 313         char *existing;
 314
 315         if (!options->object_names)
 316                 return;
 317         existing = lookup_decoration(options->object_names, obj);
 318         if (existing)
 319                 return;
 320         va_start(ap, fmt);
 321         strbuf_vaddf(&buf, fmt, ap);
 322         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 323         va_end(ap);
 324 }
 325
 326 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 327 {
 328         struct tree_desc desc;
 329         struct name_entry entry;
 330         int res = 0;
 331         const char *name;
 332
 333         if (parse_tree(tree))
 334                 return -1;
 335
 336         name = get_object_name(options, &tree->object);
 337         init_tree_desc(&desc, tree->buffer, tree->size);
 338         while (tree_entry(&desc, &entry)) {
 339                 struct object *obj;
 340                 int result;
 341
 342                 if (S_ISGITLINK(entry.mode))
 343                         continue;
 344
 345                 if (S_ISDIR(entry.mode)) {
 346                         obj = &lookup_tree(entry.oid->hash)->object;
 347                         if (name)
 348                                 put_object_name(options, obj, "%s%s/", name,
 349                                         entry.path);
 350                         result = options->walk(obj, OBJ_TREE, data, options);
 351                 }
 352                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 353                         obj = &lookup_blob(entry.oid->hash)->object;
 354                         if (name)
 355                                 put_object_name(options, obj, "%s%s", name,
 356                                         entry.path);
 357                         result = options->walk(obj, OBJ_BLOB, data, options);
 358                 }
 359                 else {
 360                         result = error("in tree %s: entry %s has bad mode %.6o",
 361                                         oid_to_hex(&tree->object.oid), entry.path, entry.mode);
 362                 }
 363                 if (result < 0)
 364                         return result;
 365                 if (!res)
 366                         res = result;
 367         }
 368         return res;
 369 }
 370
 371 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 372 {
 373         int counter = 0, generation = 0, name_prefix_len = 0;
 374         struct commit_list *parents;
 375         int res;
 376         int result;
 377         const char *name;
 378
 379         if (parse_commit(commit))
 380                 return -1;
 381
 382         name = get_object_name(options, &commit->object);
 383         if (name)
 384                 put_object_name(options, &commit->tree->object, "%s:", name);
 385
 386         result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options);
 387         if (result < 0)
 388                 return result;
 389         res = result;
 390
 391         parents = commit->parents;
 392         if (name && parents) {
 393                 int len = strlen(name), power;
 394
 395                 if (len && name[len - 1] == '^') {
 396                         generation = 1;
 397                         name_prefix_len = len - 1;
 398                 }
 399                 else { /* parse ~<generation> suffix */
 400                         for (generation = 0, power = 1;
 401                              len && isdigit(name[len - 1]);
 402                              power *= 10)
 403                                 generation += power * (name[--len] - '0');
 404                         if (power > 1 && len && name[len - 1] == '~')
 405                                 name_prefix_len = len - 1;
 406                 }
 407         }
 408
 409         while (parents) {
 410                 if (name) {
 411                         struct object *obj = &parents->item->object;
 412
 413                         if (++counter > 1)
 414                                 put_object_name(options, obj, "%s^%d",
 415                                         name, counter);
 416                         else if (generation > 0)
 417                                 put_object_name(options, obj, "%.*s~%d",
 418                                         name_prefix_len, name, generation + 1);
 419                         else
 420                                 put_object_name(options, obj, "%s^", name);
 421                 }
 422                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 423                 if (result < 0)
 424                         return result;
 425                 if (!res)
 426                         res = result;
 427                 parents = parents->next;
 428         }
 429         return res;
 430 }
 431
 432 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 433 {
 434         char *name = get_object_name(options, &tag->object);
 435
 436         if (parse_tag(tag))
 437                 return -1;
 438         if (name)
 439                 put_object_name(options, tag->tagged, "%s", name);
 440         return options->walk(tag->tagged, OBJ_ANY, data, options);
 441 }
 442
 443 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 444 {
 445         if (!obj)
 446                 return -1;
 447         switch (obj->type) {
 448         case OBJ_BLOB:
 449                 return 0;
 450         case OBJ_TREE:
 451                 return fsck_walk_tree((struct tree *)obj, data, options);
 452         case OBJ_COMMIT:
 453                 return fsck_walk_commit((struct commit *)obj, data, options);
 454         case OBJ_TAG:
 455                 return fsck_walk_tag((struct tag *)obj, data, options);
 456         default:
 457                 error("Unknown object type for %s", oid_to_hex(&obj->oid));
 458                 return -1;
 459         }
 460 }
 461
 462 /*
 463  * The entries in a tree are ordered in the _path_ order,
 464  * which means that a directory entry is ordered by adding
 465  * a slash to the end of it.
 466  *
 467  * So a directory called "a" is ordered _after_ a file
 468  * called "a.c", because "a/" sorts after "a.c".
 469  */
 470 #define TREE_UNORDERED (-1)
 471 #define TREE_HAS_DUPS  (-2)
 472
 473 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 474 {
 475         int len1 = strlen(name1);
 476         int len2 = strlen(name2);
 477         int len = len1 < len2 ? len1 : len2;
 478         unsigned char c1, c2;
 479         int cmp;
 480
 481         cmp = memcmp(name1, name2, len);
 482         if (cmp < 0)
 483                 return 0;
 484         if (cmp > 0)
 485                 return TREE_UNORDERED;
 486
 487         /*
 488          * Ok, the first <len> characters are the same.
 489          * Now we need to order the next one, but turn
 490          * a '\0' into a '/' for a directory entry.
 491          */
 492         c1 = name1[len];
 493         c2 = name2[len];
 494         if (!c1 && !c2)
 495                 /*
 496                  * git-write-tree used to write out a nonsense tree that has
 497                  * entries with the same name, one blob and one tree.  Make
 498                  * sure we do not have duplicate entries.
 499                  */
 500                 return TREE_HAS_DUPS;
 501         if (!c1 && S_ISDIR(mode1))
 502                 c1 = '/';
 503         if (!c2 && S_ISDIR(mode2))
 504                 c2 = '/';
 505         return c1 < c2 ? 0 : TREE_UNORDERED;
 506 }
 507
 508 static int fsck_tree(struct tree *item, struct fsck_options *options)
 509 {
 510         int retval;
 511         int has_null_sha1 = 0;
 512         int has_full_path = 0;
 513         int has_empty_name = 0;
 514         int has_dot = 0;
 515         int has_dotdot = 0;
 516         int has_dotgit = 0;
 517         int has_zero_pad = 0;
 518         int has_bad_modes = 0;
 519         int has_dup_entries = 0;
 520         int not_properly_sorted = 0;
 521         struct tree_desc desc;
 522         unsigned o_mode;
 523         const char *o_name;
 524
 525         init_tree_desc(&desc, item->buffer, item->size);
 526
 527         o_mode = 0;
 528         o_name = NULL;
 529
 530         while (desc.size) {
 531                 unsigned mode;
 532                 const char *name;
 533                 const struct object_id *oid;
 534
 535                 oid = tree_entry_extract(&desc, &name, &mode);
 536
 537                 has_null_sha1 |= is_null_oid(oid);
 538                 has_full_path |= !!strchr(name, '/');
 539                 has_empty_name |= !*name;
 540                 has_dot |= !strcmp(name, ".");
 541                 has_dotdot |= !strcmp(name, "..");
 542                 has_dotgit |= (!strcmp(name, ".git") ||
 543                                is_hfs_dotgit(name) ||
 544                                is_ntfs_dotgit(name));
 545                 has_zero_pad |= *(char *)desc.buffer == '0';
 546                 update_tree_entry(&desc);
 547
 548                 switch (mode) {
 549                 /*
 550                  * Standard modes..
 551                  */
 552                 case S_IFREG | 0755:
 553                 case S_IFREG | 0644:
 554                 case S_IFLNK:
 555                 case S_IFDIR:
 556                 case S_IFGITLINK:
 557                         break;
 558                 /*
 559                  * This is nonstandard, but we had a few of these
 560                  * early on when we honored the full set of mode
 561                  * bits..
 562                  */
 563                 case S_IFREG | 0664:
 564                         if (!options->strict)
 565                                 break;
 566                 default:
 567                         has_bad_modes = 1;
 568                 }
 569
 570                 if (o_name) {
 571                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 572                         case TREE_UNORDERED:
 573                                 not_properly_sorted = 1;
 574                                 break;
 575                         case TREE_HAS_DUPS:
 576                                 has_dup_entries = 1;
 577                                 break;
 578                         default:
 579                                 break;
 580                         }
 581                 }
 582
 583                 o_mode = mode;
 584                 o_name = name;
 585         }
 586
 587         retval = 0;
 588         if (has_null_sha1)
 589                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 590         if (has_full_path)
 591                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 592         if (has_empty_name)
 593                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 594         if (has_dot)
 595                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 596         if (has_dotdot)
 597                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 598         if (has_dotgit)
 599                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 600         if (has_zero_pad)
 601                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 602         if (has_bad_modes)
 603                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 604         if (has_dup_entries)
 605                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 606         if (not_properly_sorted)
 607                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 608         return retval;
 609 }
 610
 611 static int verify_headers(const void *data, unsigned long size,
 612                           struct object *obj, struct fsck_options *options)
 613 {
 614         const char *buffer = (const char *)data;
 615         unsigned long i;
 616
 617         for (i = 0; i < size; i++) {
 618                 switch (buffer[i]) {
 619                 case '\0':
 620                         return report(options, obj,
 621                                 FSCK_MSG_NUL_IN_HEADER,
 622                                 "unterminated header: NUL at offset %ld", i);
 623                 case '\n':
 624                         if (i + 1 < size && buffer[i + 1] == '\n')
 625                                 return 0;
 626                 }
 627         }
 628
 629         /*
 630          * We did not find double-LF that separates the header
 631          * and the body.  Not having a body is not a crime but
 632          * we do want to see the terminating LF for the last header
 633          * line.
 634          */
 635         if (size && buffer[size - 1] == '\n')
 636                 return 0;
 637
 638         return report(options, obj,
 639                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 640 }
 641
 642 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 643 {
 644         const char *p = *ident;
 645         char *end;
 646
 647         *ident = strchrnul(*ident, '\n');
 648         if (**ident == '\n')
 649                 (*ident)++;
 650
 651         if (*p == '<')
 652                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 653         p += strcspn(p, "<>\n");
 654         if (*p == '>')
 655                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 656         if (*p != '<')
 657                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 658         if (p[-1] != ' ')
 659                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 660         p++;
 661         p += strcspn(p, "<>\n");
 662         if (*p != '>')
 663                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 664         p++;
 665         if (*p != ' ')
 666                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 667         p++;
 668         if (*p == '0' && p[1] != ' ')
 669                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 670         if (date_overflows(strtoul(p, &end, 10)))
 671                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 672         if ((end == p || *end != ' '))
 673                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 674         p = end + 1;
 675         if ((*p != '+' && *p != '-') ||
 676             !isdigit(p[1]) ||
 677             !isdigit(p[2]) ||
 678             !isdigit(p[3]) ||
 679             !isdigit(p[4]) ||
 680             (p[5] != '\n'))
 681                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 682         p += 6;
 683         return 0;
 684 }
 685
 686 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 687         unsigned long size, struct fsck_options *options)
 688 {
 689         unsigned char tree_sha1[20], sha1[20];
 690         struct commit_graft *graft;
 691         unsigned parent_count, parent_line_count = 0, author_count;
 692         int err;
 693         const char *buffer_begin = buffer;
 694
 695         if (verify_headers(buffer, size, &commit->object, options))
 696                 return -1;
 697
 698         if (!skip_prefix(buffer, "tree ", &buffer))
 699                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 700         if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
 701                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 702                 if (err)
 703                         return err;
 704         }
 705         buffer += 41;
 706         while (skip_prefix(buffer, "parent ", &buffer)) {
 707                 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 708                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 709                         if (err)
 710                                 return err;
 711                 }
 712                 buffer += 41;
 713                 parent_line_count++;
 714         }
 715         graft = lookup_commit_graft(commit->object.oid.hash);
 716         parent_count = commit_list_count(commit->parents);
 717         if (graft) {
 718                 if (graft->nr_parent == -1 && !parent_count)
 719                         ; /* shallow commit */
 720                 else if (graft->nr_parent != parent_count) {
 721                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 722                         if (err)
 723                                 return err;
 724                 }
 725         } else {
 726                 if (parent_count != parent_line_count) {
 727                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 728                         if (err)
 729                                 return err;
 730                 }
 731         }
 732         author_count = 0;
 733         while (skip_prefix(buffer, "author ", &buffer)) {
 734                 author_count++;
 735                 err = fsck_ident(&buffer, &commit->object, options);
 736                 if (err)
 737                         return err;
 738         }
 739         if (author_count < 1)
 740                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 741         else if (author_count > 1)
 742                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 743         if (err)
 744                 return err;
 745         if (!skip_prefix(buffer, "committer ", &buffer))
 746                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 747         err = fsck_ident(&buffer, &commit->object, options);
 748         if (err)
 749                 return err;
 750         if (!commit->tree) {
 751                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
 752                 if (err)
 753                         return err;
 754         }
 755         if (memchr(buffer_begin, '\0', size)) {
 756                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 757                              "NUL byte in the commit object body");
 758                 if (err)
 759                         return err;
 760         }
 761         return 0;
 762 }
 763
 764 static int fsck_commit(struct commit *commit, const char *data,
 765         unsigned long size, struct fsck_options *options)
 766 {
 767         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 768         int ret = fsck_commit_buffer(commit, buffer, size, options);
 769         if (!data)
 770                 unuse_commit_buffer(commit, buffer);
 771         return ret;
 772 }
 773
 774 static int fsck_tag_buffer(struct tag *tag, const char *data,
 775         unsigned long size, struct fsck_options *options)
 776 {
 777         unsigned char sha1[20];
 778         int ret = 0;
 779         const char *buffer;
 780         char *to_free = NULL, *eol;
 781         struct strbuf sb = STRBUF_INIT;
 782
 783         if (data)
 784                 buffer = data;
 785         else {
 786                 enum object_type type;
 787
 788                 buffer = to_free =
 789                         read_sha1_file(tag->object.oid.hash, &type, &size);
 790                 if (!buffer)
 791                         return report(options, &tag->object,
 792                                 FSCK_MSG_MISSING_TAG_OBJECT,
 793                                 "cannot read tag object");
 794
 795                 if (type != OBJ_TAG) {
 796                         ret = report(options, &tag->object,
 797                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 798                                 "expected tag got %s",
 799                             typename(type));
 800                         goto done;
 801                 }
 802         }
 803
 804         ret = verify_headers(buffer, size, &tag->object, options);
 805         if (ret)
 806                 goto done;
 807
 808         if (!skip_prefix(buffer, "object ", &buffer)) {
 809                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 810                 goto done;
 811         }
 812         if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 813                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 814                 if (ret)
 815                         goto done;
 816         }
 817         buffer += 41;
 818
 819         if (!skip_prefix(buffer, "type ", &buffer)) {
 820                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 821                 goto done;
 822         }
 823         eol = strchr(buffer, '\n');
 824         if (!eol) {
 825                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 826                 goto done;
 827         }
 828         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 829                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 830         if (ret)
 831                 goto done;
 832         buffer = eol + 1;
 833
 834         if (!skip_prefix(buffer, "tag ", &buffer)) {
 835                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 836                 goto done;
 837         }
 838         eol = strchr(buffer, '\n');
 839         if (!eol) {
 840                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 841                 goto done;
 842         }
 843         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 844         if (check_refname_format(sb.buf, 0)) {
 845                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 846                            "invalid 'tag' name: %.*s",
 847                            (int)(eol - buffer), buffer);
 848                 if (ret)
 849                         goto done;
 850         }
 851         buffer = eol + 1;
 852
 853         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 854                 /* early tags do not contain 'tagger' lines; warn only */
 855                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 856                 if (ret)
 857                         goto done;
 858         }
 859         else
 860                 ret = fsck_ident(&buffer, &tag->object, options);
 861
 862 done:
 863         strbuf_release(&sb);
 864         free(to_free);
 865         return ret;
 866 }
 867
 868 static int fsck_tag(struct tag *tag, const char *data,
 869         unsigned long size, struct fsck_options *options)
 870 {
 871         struct object *tagged = tag->tagged;
 872
 873         if (!tagged)
 874                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 875
 876         return fsck_tag_buffer(tag, data, size, options);
 877 }
 878
 879 int fsck_object(struct object *obj, void *data, unsigned long size,
 880         struct fsck_options *options)
 881 {
 882         if (!obj)
 883                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
 884
 885         if (obj->type == OBJ_BLOB)
 886                 return 0;
 887         if (obj->type == OBJ_TREE)
 888                 return fsck_tree((struct tree *) obj, options);
 889         if (obj->type == OBJ_COMMIT)
 890                 return fsck_commit((struct commit *) obj, (const char *) data,
 891                         size, options);
 892         if (obj->type == OBJ_TAG)
 893                 return fsck_tag((struct tag *) obj, (const char *) data,
 894                         size, options);
 895
 896         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
 897                           obj->type);
 898 }
 899
 900 int fsck_error_function(struct object *obj, int msg_type, const char *message)
 901 {
 902         if (msg_type == FSCK_WARN) {
 903                 warning("object %s: %s", oid_to_hex(&obj->oid), message);
 904                 return 0;
 905         }
 906         error("object %s: %s", oid_to_hex(&obj->oid), message);
 907         return 1;
 908 }