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, &get_commit_tree(commit)->object,
 400                                 "%s:", name);
 401
 402         result = options->walk((struct object *)get_commit_tree(commit),
 403                                OBJ_TREE, data, options);
 404         if (result < 0)
 405                 return result;
 406         res = result;
 407
 408         parents = commit->parents;
 409         if (name && parents) {
 410                 int len = strlen(name), power;
 411
 412                 if (len && name[len - 1] == '^') {
 413                         generation = 1;
 414                         name_prefix_len = len - 1;
 415                 }
 416                 else { /* parse ~<generation> suffix */
 417                         for (generation = 0, power = 1;
 418                              len && isdigit(name[len - 1]);
 419                              power *= 10)
 420                                 generation += power * (name[--len] - '0');
 421                         if (power > 1 && len && name[len - 1] == '~')
 422                                 name_prefix_len = len - 1;
 423                 }
 424         }
 425
 426         while (parents) {
 427                 if (name) {
 428                         struct object *obj = &parents->item->object;
 429
 430                         if (++counter > 1)
 431                                 put_object_name(options, obj, "%s^%d",
 432                                         name, counter);
 433                         else if (generation > 0)
 434                                 put_object_name(options, obj, "%.*s~%d",
 435                                         name_prefix_len, name, generation + 1);
 436                         else
 437                                 put_object_name(options, obj, "%s^", name);
 438                 }
 439                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 440                 if (result < 0)
 441                         return result;
 442                 if (!res)
 443                         res = result;
 444                 parents = parents->next;
 445         }
 446         return res;
 447 }
 448
 449 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 450 {
 451         char *name = get_object_name(options, &tag->object);
 452
 453         if (parse_tag(tag))
 454                 return -1;
 455         if (name)
 456                 put_object_name(options, tag->tagged, "%s", name);
 457         return options->walk(tag->tagged, OBJ_ANY, data, options);
 458 }
 459
 460 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 461 {
 462         if (!obj)
 463                 return -1;
 464
 465         if (obj->type == OBJ_NONE)
 466                 parse_object(&obj->oid);
 467
 468         switch (obj->type) {
 469         case OBJ_BLOB:
 470                 return 0;
 471         case OBJ_TREE:
 472                 return fsck_walk_tree((struct tree *)obj, data, options);
 473         case OBJ_COMMIT:
 474                 return fsck_walk_commit((struct commit *)obj, data, options);
 475         case OBJ_TAG:
 476                 return fsck_walk_tag((struct tag *)obj, data, options);
 477         default:
 478                 error("Unknown object type for %s", describe_object(options, obj));
 479                 return -1;
 480         }
 481 }
 482
 483 /*
 484  * The entries in a tree are ordered in the _path_ order,
 485  * which means that a directory entry is ordered by adding
 486  * a slash to the end of it.
 487  *
 488  * So a directory called "a" is ordered _after_ a file
 489  * called "a.c", because "a/" sorts after "a.c".
 490  */
 491 #define TREE_UNORDERED (-1)
 492 #define TREE_HAS_DUPS  (-2)
 493
 494 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 495 {
 496         int len1 = strlen(name1);
 497         int len2 = strlen(name2);
 498         int len = len1 < len2 ? len1 : len2;
 499         unsigned char c1, c2;
 500         int cmp;
 501
 502         cmp = memcmp(name1, name2, len);
 503         if (cmp < 0)
 504                 return 0;
 505         if (cmp > 0)
 506                 return TREE_UNORDERED;
 507
 508         /*
 509          * Ok, the first <len> characters are the same.
 510          * Now we need to order the next one, but turn
 511          * a '\0' into a '/' for a directory entry.
 512          */
 513         c1 = name1[len];
 514         c2 = name2[len];
 515         if (!c1 && !c2)
 516                 /*
 517                  * git-write-tree used to write out a nonsense tree that has
 518                  * entries with the same name, one blob and one tree.  Make
 519                  * sure we do not have duplicate entries.
 520                  */
 521                 return TREE_HAS_DUPS;
 522         if (!c1 && S_ISDIR(mode1))
 523                 c1 = '/';
 524         if (!c2 && S_ISDIR(mode2))
 525                 c2 = '/';
 526         return c1 < c2 ? 0 : TREE_UNORDERED;
 527 }
 528
 529 static int fsck_tree(struct tree *item, struct fsck_options *options)
 530 {
 531         int retval = 0;
 532         int has_null_sha1 = 0;
 533         int has_full_path = 0;
 534         int has_empty_name = 0;
 535         int has_dot = 0;
 536         int has_dotdot = 0;
 537         int has_dotgit = 0;
 538         int has_zero_pad = 0;
 539         int has_bad_modes = 0;
 540         int has_dup_entries = 0;
 541         int not_properly_sorted = 0;
 542         struct tree_desc desc;
 543         unsigned o_mode;
 544         const char *o_name;
 545
 546         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 547                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 548                 return retval;
 549         }
 550
 551         o_mode = 0;
 552         o_name = NULL;
 553
 554         while (desc.size) {
 555                 unsigned mode;
 556                 const char *name;
 557                 const struct object_id *oid;
 558
 559                 oid = tree_entry_extract(&desc, &name, &mode);
 560
 561                 has_null_sha1 |= is_null_oid(oid);
 562                 has_full_path |= !!strchr(name, '/');
 563                 has_empty_name |= !*name;
 564                 has_dot |= !strcmp(name, ".");
 565                 has_dotdot |= !strcmp(name, "..");
 566                 has_dotgit |= (!strcmp(name, ".git") ||
 567                                is_hfs_dotgit(name) ||
 568                                is_ntfs_dotgit(name));
 569                 has_zero_pad |= *(char *)desc.buffer == '0';
 570                 if (update_tree_entry_gently(&desc)) {
 571                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 572                         break;
 573                 }
 574
 575                 switch (mode) {
 576                 /*
 577                  * Standard modes..
 578                  */
 579                 case S_IFREG | 0755:
 580                 case S_IFREG | 0644:
 581                 case S_IFLNK:
 582                 case S_IFDIR:
 583                 case S_IFGITLINK:
 584                         break;
 585                 /*
 586                  * This is nonstandard, but we had a few of these
 587                  * early on when we honored the full set of mode
 588                  * bits..
 589                  */
 590                 case S_IFREG | 0664:
 591                         if (!options->strict)
 592                                 break;
 593                         /* fallthrough */
 594                 default:
 595                         has_bad_modes = 1;
 596                 }
 597
 598                 if (o_name) {
 599                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 600                         case TREE_UNORDERED:
 601                                 not_properly_sorted = 1;
 602                                 break;
 603                         case TREE_HAS_DUPS:
 604                                 has_dup_entries = 1;
 605                                 break;
 606                         default:
 607                                 break;
 608                         }
 609                 }
 610
 611                 o_mode = mode;
 612                 o_name = name;
 613         }
 614
 615         if (has_null_sha1)
 616                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 617         if (has_full_path)
 618                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 619         if (has_empty_name)
 620                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 621         if (has_dot)
 622                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 623         if (has_dotdot)
 624                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 625         if (has_dotgit)
 626                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 627         if (has_zero_pad)
 628                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 629         if (has_bad_modes)
 630                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 631         if (has_dup_entries)
 632                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 633         if (not_properly_sorted)
 634                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 635         return retval;
 636 }
 637
 638 static int verify_headers(const void *data, unsigned long size,
 639                           struct object *obj, struct fsck_options *options)
 640 {
 641         const char *buffer = (const char *)data;
 642         unsigned long i;
 643
 644         for (i = 0; i < size; i++) {
 645                 switch (buffer[i]) {
 646                 case '\0':
 647                         return report(options, obj,
 648                                 FSCK_MSG_NUL_IN_HEADER,
 649                                 "unterminated header: NUL at offset %ld", i);
 650                 case '\n':
 651                         if (i + 1 < size && buffer[i + 1] == '\n')
 652                                 return 0;
 653                 }
 654         }
 655
 656         /*
 657          * We did not find double-LF that separates the header
 658          * and the body.  Not having a body is not a crime but
 659          * we do want to see the terminating LF for the last header
 660          * line.
 661          */
 662         if (size && buffer[size - 1] == '\n')
 663                 return 0;
 664
 665         return report(options, obj,
 666                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 667 }
 668
 669 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 670 {
 671         const char *p = *ident;
 672         char *end;
 673
 674         *ident = strchrnul(*ident, '\n');
 675         if (**ident == '\n')
 676                 (*ident)++;
 677
 678         if (*p == '<')
 679                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 680         p += strcspn(p, "<>\n");
 681         if (*p == '>')
 682                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 683         if (*p != '<')
 684                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 685         if (p[-1] != ' ')
 686                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 687         p++;
 688         p += strcspn(p, "<>\n");
 689         if (*p != '>')
 690                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 691         p++;
 692         if (*p != ' ')
 693                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 694         p++;
 695         if (*p == '0' && p[1] != ' ')
 696                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 697         if (date_overflows(parse_timestamp(p, &end, 10)))
 698                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 699         if ((end == p || *end != ' '))
 700                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 701         p = end + 1;
 702         if ((*p != '+' && *p != '-') ||
 703             !isdigit(p[1]) ||
 704             !isdigit(p[2]) ||
 705             !isdigit(p[3]) ||
 706             !isdigit(p[4]) ||
 707             (p[5] != '\n'))
 708                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 709         p += 6;
 710         return 0;
 711 }
 712
 713 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 714         unsigned long size, struct fsck_options *options)
 715 {
 716         unsigned char tree_sha1[20], sha1[20];
 717         struct commit_graft *graft;
 718         unsigned parent_count, parent_line_count = 0, author_count;
 719         int err;
 720         const char *buffer_begin = buffer;
 721
 722         if (verify_headers(buffer, size, &commit->object, options))
 723                 return -1;
 724
 725         if (!skip_prefix(buffer, "tree ", &buffer))
 726                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 727         if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
 728                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 729                 if (err)
 730                         return err;
 731         }
 732         buffer += 41;
 733         while (skip_prefix(buffer, "parent ", &buffer)) {
 734                 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 735                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 736                         if (err)
 737                                 return err;
 738                 }
 739                 buffer += 41;
 740                 parent_line_count++;
 741         }
 742         graft = lookup_commit_graft(&commit->object.oid);
 743         parent_count = commit_list_count(commit->parents);
 744         if (graft) {
 745                 if (graft->nr_parent == -1 && !parent_count)
 746                         ; /* shallow commit */
 747                 else if (graft->nr_parent != parent_count) {
 748                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 749                         if (err)
 750                                 return err;
 751                 }
 752         } else {
 753                 if (parent_count != parent_line_count) {
 754                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 755                         if (err)
 756                                 return err;
 757                 }
 758         }
 759         author_count = 0;
 760         while (skip_prefix(buffer, "author ", &buffer)) {
 761                 author_count++;
 762                 err = fsck_ident(&buffer, &commit->object, options);
 763                 if (err)
 764                         return err;
 765         }
 766         if (author_count < 1)
 767                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 768         else if (author_count > 1)
 769                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 770         if (err)
 771                 return err;
 772         if (!skip_prefix(buffer, "committer ", &buffer))
 773                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 774         err = fsck_ident(&buffer, &commit->object, options);
 775         if (err)
 776                 return err;
 777         if (!get_commit_tree(commit)) {
 778                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
 779                 if (err)
 780                         return err;
 781         }
 782         if (memchr(buffer_begin, '\0', size)) {
 783                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 784                              "NUL byte in the commit object body");
 785                 if (err)
 786                         return err;
 787         }
 788         return 0;
 789 }
 790
 791 static int fsck_commit(struct commit *commit, const char *data,
 792         unsigned long size, struct fsck_options *options)
 793 {
 794         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 795         int ret = fsck_commit_buffer(commit, buffer, size, options);
 796         if (!data)
 797                 unuse_commit_buffer(commit, buffer);
 798         return ret;
 799 }
 800
 801 static int fsck_tag_buffer(struct tag *tag, const char *data,
 802         unsigned long size, struct fsck_options *options)
 803 {
 804         unsigned char sha1[20];
 805         int ret = 0;
 806         const char *buffer;
 807         char *to_free = NULL, *eol;
 808         struct strbuf sb = STRBUF_INIT;
 809
 810         if (data)
 811                 buffer = data;
 812         else {
 813                 enum object_type type;
 814
 815                 buffer = to_free =
 816                         read_sha1_file(tag->object.oid.hash, &type, &size);
 817                 if (!buffer)
 818                         return report(options, &tag->object,
 819                                 FSCK_MSG_MISSING_TAG_OBJECT,
 820                                 "cannot read tag object");
 821
 822                 if (type != OBJ_TAG) {
 823                         ret = report(options, &tag->object,
 824                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 825                                 "expected tag got %s",
 826                             type_name(type));
 827                         goto done;
 828                 }
 829         }
 830
 831         ret = verify_headers(buffer, size, &tag->object, options);
 832         if (ret)
 833                 goto done;
 834
 835         if (!skip_prefix(buffer, "object ", &buffer)) {
 836                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 837                 goto done;
 838         }
 839         if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 840                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 841                 if (ret)
 842                         goto done;
 843         }
 844         buffer += 41;
 845
 846         if (!skip_prefix(buffer, "type ", &buffer)) {
 847                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 848                 goto done;
 849         }
 850         eol = strchr(buffer, '\n');
 851         if (!eol) {
 852                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 853                 goto done;
 854         }
 855         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 856                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 857         if (ret)
 858                 goto done;
 859         buffer = eol + 1;
 860
 861         if (!skip_prefix(buffer, "tag ", &buffer)) {
 862                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 863                 goto done;
 864         }
 865         eol = strchr(buffer, '\n');
 866         if (!eol) {
 867                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 868                 goto done;
 869         }
 870         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 871         if (check_refname_format(sb.buf, 0)) {
 872                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 873                            "invalid 'tag' name: %.*s",
 874                            (int)(eol - buffer), buffer);
 875                 if (ret)
 876                         goto done;
 877         }
 878         buffer = eol + 1;
 879
 880         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 881                 /* early tags do not contain 'tagger' lines; warn only */
 882                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 883                 if (ret)
 884                         goto done;
 885         }
 886         else
 887                 ret = fsck_ident(&buffer, &tag->object, options);
 888
 889 done:
 890         strbuf_release(&sb);
 891         free(to_free);
 892         return ret;
 893 }
 894
 895 static int fsck_tag(struct tag *tag, const char *data,
 896         unsigned long size, struct fsck_options *options)
 897 {
 898         struct object *tagged = tag->tagged;
 899
 900         if (!tagged)
 901                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 902
 903         return fsck_tag_buffer(tag, data, size, options);
 904 }
 905
 906 int fsck_object(struct object *obj, void *data, unsigned long size,
 907         struct fsck_options *options)
 908 {
 909         if (!obj)
 910                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
 911
 912         if (obj->type == OBJ_BLOB)
 913                 return 0;
 914         if (obj->type == OBJ_TREE)
 915                 return fsck_tree((struct tree *) obj, options);
 916         if (obj->type == OBJ_COMMIT)
 917                 return fsck_commit((struct commit *) obj, (const char *) data,
 918                         size, options);
 919         if (obj->type == OBJ_TAG)
 920                 return fsck_tag((struct tag *) obj, (const char *) data,
 921                         size, options);
 922
 923         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
 924                           obj->type);
 925 }
 926
 927 int fsck_error_function(struct fsck_options *o,
 928         struct object *obj, int msg_type, const char *message)
 929 {
 930         if (msg_type == FSCK_WARN) {
 931                 warning("object %s: %s", describe_object(o, obj), message);
 932                 return 0;
 933         }
 934         error("object %s: %s", describe_object(o, obj), message);
 935         return 1;
 936 }