Andrew's git / gitweb.git / fsck.c
?
fsck.con commit Merge branch 'jc/clean-after-sanity-tests' (1840443)
   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#include "oidset.h"
  14#include "packfile.h"
  15#include "submodule-config.h"
  16#include "config.h"
  17#include "help.h"
  18
  19static struct oidset gitmodules_found = OIDSET_INIT;
  20static struct oidset gitmodules_done = OIDSET_INIT;
  21
  22#define FSCK_FATAL -1
  23#define FSCK_INFO -2
  24
  25#define FOREACH_MSG_ID(FUNC) \
  26        /* fatal errors */ \
  27        FUNC(NUL_IN_HEADER, FATAL) \
  28        FUNC(UNTERMINATED_HEADER, FATAL) \
  29        /* errors */ \
  30        FUNC(BAD_DATE, ERROR) \
  31        FUNC(BAD_DATE_OVERFLOW, ERROR) \
  32        FUNC(BAD_EMAIL, ERROR) \
  33        FUNC(BAD_NAME, ERROR) \
  34        FUNC(BAD_OBJECT_SHA1, ERROR) \
  35        FUNC(BAD_PARENT_SHA1, ERROR) \
  36        FUNC(BAD_TAG_OBJECT, ERROR) \
  37        FUNC(BAD_TIMEZONE, ERROR) \
  38        FUNC(BAD_TREE, ERROR) \
  39        FUNC(BAD_TREE_SHA1, ERROR) \
  40        FUNC(BAD_TYPE, ERROR) \
  41        FUNC(DUPLICATE_ENTRIES, ERROR) \
  42        FUNC(MISSING_AUTHOR, ERROR) \
  43        FUNC(MISSING_COMMITTER, ERROR) \
  44        FUNC(MISSING_EMAIL, ERROR) \
  45        FUNC(MISSING_GRAFT, ERROR) \
  46        FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  47        FUNC(MISSING_OBJECT, ERROR) \
  48        FUNC(MISSING_PARENT, ERROR) \
  49        FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  50        FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  51        FUNC(MISSING_TAG, ERROR) \
  52        FUNC(MISSING_TAG_ENTRY, ERROR) \
  53        FUNC(MISSING_TAG_OBJECT, ERROR) \
  54        FUNC(MISSING_TREE, ERROR) \
  55        FUNC(MISSING_TREE_OBJECT, ERROR) \
  56        FUNC(MISSING_TYPE, ERROR) \
  57        FUNC(MISSING_TYPE_ENTRY, ERROR) \
  58        FUNC(MULTIPLE_AUTHORS, ERROR) \
  59        FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  60        FUNC(TREE_NOT_SORTED, ERROR) \
  61        FUNC(UNKNOWN_TYPE, ERROR) \
  62        FUNC(ZERO_PADDED_DATE, ERROR) \
  63        FUNC(GITMODULES_MISSING, ERROR) \
  64        FUNC(GITMODULES_BLOB, ERROR) \
  65        FUNC(GITMODULES_PARSE, ERROR) \
  66        FUNC(GITMODULES_NAME, ERROR) \
  67        FUNC(GITMODULES_SYMLINK, ERROR) \
  68        /* warnings */ \
  69        FUNC(BAD_FILEMODE, WARN) \
  70        FUNC(EMPTY_NAME, WARN) \
  71        FUNC(FULL_PATHNAME, WARN) \
  72        FUNC(HAS_DOT, WARN) \
  73        FUNC(HAS_DOTDOT, WARN) \
  74        FUNC(HAS_DOTGIT, WARN) \
  75        FUNC(NULL_SHA1, WARN) \
  76        FUNC(ZERO_PADDED_FILEMODE, WARN) \
  77        FUNC(NUL_IN_COMMIT, WARN) \
  78        /* infos (reported as warnings, but ignored by default) */ \
  79        FUNC(BAD_TAG_NAME, INFO) \
  80        FUNC(MISSING_TAGGER_ENTRY, INFO)
  81
  82#define MSG_ID(id, msg_type) FSCK_MSG_##id,
  83enum fsck_msg_id {
  84        FOREACH_MSG_ID(MSG_ID)
  85        FSCK_MSG_MAX
  86};
  87#undef MSG_ID
  88
  89#define STR(x) #x
  90#define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
  91static struct {
  92        const char *id_string;
  93        const char *downcased;
  94        const char *camelcased;
  95        int msg_type;
  96} msg_id_info[FSCK_MSG_MAX + 1] = {
  97        FOREACH_MSG_ID(MSG_ID)
  98        { NULL, NULL, NULL, -1 }
  99};
 100#undef MSG_ID
 101
 102static void prepare_msg_ids(void)
 103{
 104        int i;
 105
 106        if (msg_id_info[0].downcased)
 107                return;
 108
 109        /* convert id_string to lower case, without underscores. */
 110        for (i = 0; i < FSCK_MSG_MAX; i++) {
 111                const char *p = msg_id_info[i].id_string;
 112                int len = strlen(p);
 113                char *q = xmalloc(len);
 114
 115                msg_id_info[i].downcased = q;
 116                while (*p)
 117                        if (*p == '_')
 118                                p++;
 119                        else
 120                                *(q)++ = tolower(*(p)++);
 121                *q = '\0';
 122
 123                p = msg_id_info[i].id_string;
 124                q = xmalloc(len);
 125                msg_id_info[i].camelcased = q;
 126                while (*p) {
 127                        if (*p == '_') {
 128                                p++;
 129                                if (*p)
 130                                        *q++ = *p++;
 131                        } else {
 132                                *q++ = tolower(*p++);
 133                        }
 134                }
 135                *q = '\0';
 136        }
 137}
 138
 139static int parse_msg_id(const char *text)
 140{
 141        int i;
 142
 143        prepare_msg_ids();
 144
 145        for (i = 0; i < FSCK_MSG_MAX; i++)
 146                if (!strcmp(text, msg_id_info[i].downcased))
 147                        return i;
 148
 149        return -1;
 150}
 151
 152void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
 153{
 154        int i;
 155
 156        prepare_msg_ids();
 157
 158        for (i = 0; i < FSCK_MSG_MAX; i++)
 159                list_config_item(list, prefix, msg_id_info[i].camelcased);
 160}
 161
 162static int fsck_msg_type(enum fsck_msg_id msg_id,
 163        struct fsck_options *options)
 164{
 165        int msg_type;
 166
 167        assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 168
 169        if (options->msg_type)
 170                msg_type = options->msg_type[msg_id];
 171        else {
 172                msg_type = msg_id_info[msg_id].msg_type;
 173                if (options->strict && msg_type == FSCK_WARN)
 174                        msg_type = FSCK_ERROR;
 175        }
 176
 177        return msg_type;
 178}
 179
 180static void init_skiplist(struct fsck_options *options, const char *path)
 181{
 182        static struct oid_array skiplist = OID_ARRAY_INIT;
 183        int sorted, fd;
 184        char buffer[GIT_MAX_HEXSZ + 1];
 185        struct object_id oid;
 186
 187        if (options->skiplist)
 188                sorted = options->skiplist->sorted;
 189        else {
 190                sorted = 1;
 191                options->skiplist = &skiplist;
 192        }
 193
 194        fd = open(path, O_RDONLY);
 195        if (fd < 0)
 196                die("Could not open skip list: %s", path);
 197        for (;;) {
 198                const char *p;
 199                int result = read_in_full(fd, buffer, sizeof(buffer));
 200                if (result < 0)
 201                        die_errno("Could not read '%s'", path);
 202                if (!result)
 203                        break;
 204                if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 205                        die("Invalid SHA-1: %s", buffer);
 206                oid_array_append(&skiplist, &oid);
 207                if (sorted && skiplist.nr > 1 &&
 208                                oidcmp(&skiplist.oid[skiplist.nr - 2],
 209                                       &oid) > 0)
 210                        sorted = 0;
 211        }
 212        close(fd);
 213
 214        if (sorted)
 215                skiplist.sorted = 1;
 216}
 217
 218static int parse_msg_type(const char *str)
 219{
 220        if (!strcmp(str, "error"))
 221                return FSCK_ERROR;
 222        else if (!strcmp(str, "warn"))
 223                return FSCK_WARN;
 224        else if (!strcmp(str, "ignore"))
 225                return FSCK_IGNORE;
 226        else
 227                die("Unknown fsck message type: '%s'", str);
 228}
 229
 230int is_valid_msg_type(const char *msg_id, const char *msg_type)
 231{
 232        if (parse_msg_id(msg_id) < 0)
 233                return 0;
 234        parse_msg_type(msg_type);
 235        return 1;
 236}
 237
 238void fsck_set_msg_type(struct fsck_options *options,
 239                const char *msg_id, const char *msg_type)
 240{
 241        int id = parse_msg_id(msg_id), type;
 242
 243        if (id < 0)
 244                die("Unhandled message id: %s", msg_id);
 245        type = parse_msg_type(msg_type);
 246
 247        if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 248                die("Cannot demote %s to %s", msg_id, msg_type);
 249
 250        if (!options->msg_type) {
 251                int i;
 252                int *msg_type;
 253                ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 254                for (i = 0; i < FSCK_MSG_MAX; i++)
 255                        msg_type[i] = fsck_msg_type(i, options);
 256                options->msg_type = msg_type;
 257        }
 258
 259        options->msg_type[id] = type;
 260}
 261
 262void fsck_set_msg_types(struct fsck_options *options, const char *values)
 263{
 264        char *buf = xstrdup(values), *to_free = buf;
 265        int done = 0;
 266
 267        while (!done) {
 268                int len = strcspn(buf, " ,|"), equal;
 269
 270                done = !buf[len];
 271                if (!len) {
 272                        buf++;
 273                        continue;
 274                }
 275                buf[len] = '\0';
 276
 277                for (equal = 0;
 278                     equal < len && buf[equal] != '=' && buf[equal] != ':';
 279                     equal++)
 280                        buf[equal] = tolower(buf[equal]);
 281                buf[equal] = '\0';
 282
 283                if (!strcmp(buf, "skiplist")) {
 284                        if (equal == len)
 285                                die("skiplist requires a path");
 286                        init_skiplist(options, buf + equal + 1);
 287                        buf += len + 1;
 288                        continue;
 289                }
 290
 291                if (equal == len)
 292                        die("Missing '=': '%s'", buf);
 293
 294                fsck_set_msg_type(options, buf, buf + equal + 1);
 295                buf += len + 1;
 296        }
 297        free(to_free);
 298}
 299
 300static void append_msg_id(struct strbuf *sb, const char *msg_id)
 301{
 302        for (;;) {
 303                char c = *(msg_id)++;
 304
 305                if (!c)
 306                        break;
 307                if (c != '_')
 308                        strbuf_addch(sb, tolower(c));
 309                else {
 310                        assert(*msg_id);
 311                        strbuf_addch(sb, *(msg_id)++);
 312                }
 313        }
 314
 315        strbuf_addstr(sb, ": ");
 316}
 317
 318__attribute__((format (printf, 4, 5)))
 319static int report(struct fsck_options *options, struct object *object,
 320        enum fsck_msg_id id, const char *fmt, ...)
 321{
 322        va_list ap;
 323        struct strbuf sb = STRBUF_INIT;
 324        int msg_type = fsck_msg_type(id, options), result;
 325
 326        if (msg_type == FSCK_IGNORE)
 327                return 0;
 328
 329        if (options->skiplist && object &&
 330                        oid_array_lookup(options->skiplist, &object->oid) >= 0)
 331                return 0;
 332
 333        if (msg_type == FSCK_FATAL)
 334                msg_type = FSCK_ERROR;
 335        else if (msg_type == FSCK_INFO)
 336                msg_type = FSCK_WARN;
 337
 338        append_msg_id(&sb, msg_id_info[id].id_string);
 339
 340        va_start(ap, fmt);
 341        strbuf_vaddf(&sb, fmt, ap);
 342        result = options->error_func(options, object, msg_type, sb.buf);
 343        strbuf_release(&sb);
 344        va_end(ap);
 345
 346        return result;
 347}
 348
 349static char *get_object_name(struct fsck_options *options, struct object *obj)
 350{
 351        if (!options->object_names)
 352                return NULL;
 353        return lookup_decoration(options->object_names, obj);
 354}
 355
 356static void put_object_name(struct fsck_options *options, struct object *obj,
 357        const char *fmt, ...)
 358{
 359        va_list ap;
 360        struct strbuf buf = STRBUF_INIT;
 361        char *existing;
 362
 363        if (!options->object_names)
 364                return;
 365        existing = lookup_decoration(options->object_names, obj);
 366        if (existing)
 367                return;
 368        va_start(ap, fmt);
 369        strbuf_vaddf(&buf, fmt, ap);
 370        add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 371        va_end(ap);
 372}
 373
 374static const char *describe_object(struct fsck_options *o, struct object *obj)
 375{
 376        static struct strbuf buf = STRBUF_INIT;
 377        char *name;
 378
 379        strbuf_reset(&buf);
 380        strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 381        if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 382                strbuf_addf(&buf, " (%s)", name);
 383
 384        return buf.buf;
 385}
 386
 387static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 388{
 389        struct tree_desc desc;
 390        struct name_entry entry;
 391        int res = 0;
 392        const char *name;
 393
 394        if (parse_tree(tree))
 395                return -1;
 396
 397        name = get_object_name(options, &tree->object);
 398        if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 399                return -1;
 400        while (tree_entry_gently(&desc, &entry)) {
 401                struct object *obj;
 402                int result;
 403
 404                if (S_ISGITLINK(entry.mode))
 405                        continue;
 406
 407                if (S_ISDIR(entry.mode)) {
 408                        obj = (struct object *)lookup_tree(entry.oid);
 409                        if (name && obj)
 410                                put_object_name(options, obj, "%s%s/", name,
 411                                        entry.path);
 412                        result = options->walk(obj, OBJ_TREE, data, options);
 413                }
 414                else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 415                        obj = (struct object *)lookup_blob(entry.oid);
 416                        if (name && obj)
 417                                put_object_name(options, obj, "%s%s", name,
 418                                        entry.path);
 419                        result = options->walk(obj, OBJ_BLOB, data, options);
 420                }
 421                else {
 422                        result = error("in tree %s: entry %s has bad mode %.6o",
 423                                        describe_object(options, &tree->object), entry.path, entry.mode);
 424                }
 425                if (result < 0)
 426                        return result;
 427                if (!res)
 428                        res = result;
 429        }
 430        return res;
 431}
 432
 433static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 434{
 435        int counter = 0, generation = 0, name_prefix_len = 0;
 436        struct commit_list *parents;
 437        int res;
 438        int result;
 439        const char *name;
 440
 441        if (parse_commit(commit))
 442                return -1;
 443
 444        name = get_object_name(options, &commit->object);
 445        if (name)
 446                put_object_name(options, &get_commit_tree(commit)->object,
 447                                "%s:", name);
 448
 449        result = options->walk((struct object *)get_commit_tree(commit),
 450                               OBJ_TREE, data, options);
 451        if (result < 0)
 452                return result;
 453        res = result;
 454
 455        parents = commit->parents;
 456        if (name && parents) {
 457                int len = strlen(name), power;
 458
 459                if (len && name[len - 1] == '^') {
 460                        generation = 1;
 461                        name_prefix_len = len - 1;
 462                }
 463                else { /* parse ~<generation> suffix */
 464                        for (generation = 0, power = 1;
 465                             len && isdigit(name[len - 1]);
 466                             power *= 10)
 467                                generation += power * (name[--len] - '0');
 468                        if (power > 1 && len && name[len - 1] == '~')
 469                                name_prefix_len = len - 1;
 470                }
 471        }
 472
 473        while (parents) {
 474                if (name) {
 475                        struct object *obj = &parents->item->object;
 476
 477                        if (++counter > 1)
 478                                put_object_name(options, obj, "%s^%d",
 479                                        name, counter);
 480                        else if (generation > 0)
 481                                put_object_name(options, obj, "%.*s~%d",
 482                                        name_prefix_len, name, generation + 1);
 483                        else
 484                                put_object_name(options, obj, "%s^", name);
 485                }
 486                result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 487                if (result < 0)
 488                        return result;
 489                if (!res)
 490                        res = result;
 491                parents = parents->next;
 492        }
 493        return res;
 494}
 495
 496static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 497{
 498        char *name = get_object_name(options, &tag->object);
 499
 500        if (parse_tag(tag))
 501                return -1;
 502        if (name)
 503                put_object_name(options, tag->tagged, "%s", name);
 504        return options->walk(tag->tagged, OBJ_ANY, data, options);
 505}
 506
 507int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 508{
 509        if (!obj)
 510                return -1;
 511
 512        if (obj->type == OBJ_NONE)
 513                parse_object(&obj->oid);
 514
 515        switch (obj->type) {
 516        case OBJ_BLOB:
 517                return 0;
 518        case OBJ_TREE:
 519                return fsck_walk_tree((struct tree *)obj, data, options);
 520        case OBJ_COMMIT:
 521                return fsck_walk_commit((struct commit *)obj, data, options);
 522        case OBJ_TAG:
 523                return fsck_walk_tag((struct tag *)obj, data, options);
 524        default:
 525                error("Unknown object type for %s", describe_object(options, obj));
 526                return -1;
 527        }
 528}
 529
 530/*
 531 * The entries in a tree are ordered in the _path_ order,
 532 * which means that a directory entry is ordered by adding
 533 * a slash to the end of it.
 534 *
 535 * So a directory called "a" is ordered _after_ a file
 536 * called "a.c", because "a/" sorts after "a.c".
 537 */
 538#define TREE_UNORDERED (-1)
 539#define TREE_HAS_DUPS  (-2)
 540
 541static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 542{
 543        int len1 = strlen(name1);
 544        int len2 = strlen(name2);
 545        int len = len1 < len2 ? len1 : len2;
 546        unsigned char c1, c2;
 547        int cmp;
 548
 549        cmp = memcmp(name1, name2, len);
 550        if (cmp < 0)
 551                return 0;
 552        if (cmp > 0)
 553                return TREE_UNORDERED;
 554
 555        /*
 556         * Ok, the first <len> characters are the same.
 557         * Now we need to order the next one, but turn
 558         * a '\0' into a '/' for a directory entry.
 559         */
 560        c1 = name1[len];
 561        c2 = name2[len];
 562        if (!c1 && !c2)
 563                /*
 564                 * git-write-tree used to write out a nonsense tree that has
 565                 * entries with the same name, one blob and one tree.  Make
 566                 * sure we do not have duplicate entries.
 567                 */
 568                return TREE_HAS_DUPS;
 569        if (!c1 && S_ISDIR(mode1))
 570                c1 = '/';
 571        if (!c2 && S_ISDIR(mode2))
 572                c2 = '/';
 573        return c1 < c2 ? 0 : TREE_UNORDERED;
 574}
 575
 576static int fsck_tree(struct tree *item, struct fsck_options *options)
 577{
 578        int retval = 0;
 579        int has_null_sha1 = 0;
 580        int has_full_path = 0;
 581        int has_empty_name = 0;
 582        int has_dot = 0;
 583        int has_dotdot = 0;
 584        int has_dotgit = 0;
 585        int has_zero_pad = 0;
 586        int has_bad_modes = 0;
 587        int has_dup_entries = 0;
 588        int not_properly_sorted = 0;
 589        struct tree_desc desc;
 590        unsigned o_mode;
 591        const char *o_name;
 592
 593        if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 594                retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 595                return retval;
 596        }
 597
 598        o_mode = 0;
 599        o_name = NULL;
 600
 601        while (desc.size) {
 602                unsigned mode;
 603                const char *name;
 604                const struct object_id *oid;
 605
 606                oid = tree_entry_extract(&desc, &name, &mode);
 607
 608                has_null_sha1 |= is_null_oid(oid);
 609                has_full_path |= !!strchr(name, '/');
 610                has_empty_name |= !*name;
 611                has_dot |= !strcmp(name, ".");
 612                has_dotdot |= !strcmp(name, "..");
 613                has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 614                has_zero_pad |= *(char *)desc.buffer == '0';
 615
 616                if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 617                        if (!S_ISLNK(mode))
 618                                oidset_insert(&gitmodules_found, oid);
 619                        else
 620                                retval += report(options, &item->object,
 621                                                 FSCK_MSG_GITMODULES_SYMLINK,
 622                                                 ".gitmodules is a symbolic link");
 623                }
 624
 625                if (update_tree_entry_gently(&desc)) {
 626                        retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 627                        break;
 628                }
 629
 630                switch (mode) {
 631                /*
 632                 * Standard modes..
 633                 */
 634                case S_IFREG | 0755:
 635                case S_IFREG | 0644:
 636                case S_IFLNK:
 637                case S_IFDIR:
 638                case S_IFGITLINK:
 639                        break;
 640                /*
 641                 * This is nonstandard, but we had a few of these
 642                 * early on when we honored the full set of mode
 643                 * bits..
 644                 */
 645                case S_IFREG | 0664:
 646                        if (!options->strict)
 647                                break;
 648                        /* fallthrough */
 649                default:
 650                        has_bad_modes = 1;
 651                }
 652
 653                if (o_name) {
 654                        switch (verify_ordered(o_mode, o_name, mode, name)) {
 655                        case TREE_UNORDERED:
 656                                not_properly_sorted = 1;
 657                                break;
 658                        case TREE_HAS_DUPS:
 659                                has_dup_entries = 1;
 660                                break;
 661                        default:
 662                                break;
 663                        }
 664                }
 665
 666                o_mode = mode;
 667                o_name = name;
 668        }
 669
 670        if (has_null_sha1)
 671                retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 672        if (has_full_path)
 673                retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 674        if (has_empty_name)
 675                retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 676        if (has_dot)
 677                retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 678        if (has_dotdot)
 679                retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 680        if (has_dotgit)
 681                retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 682        if (has_zero_pad)
 683                retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 684        if (has_bad_modes)
 685                retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 686        if (has_dup_entries)
 687                retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 688        if (not_properly_sorted)
 689                retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 690        return retval;
 691}
 692
 693static int verify_headers(const void *data, unsigned long size,
 694                          struct object *obj, struct fsck_options *options)
 695{
 696        const char *buffer = (const char *)data;
 697        unsigned long i;
 698
 699        for (i = 0; i < size; i++) {
 700                switch (buffer[i]) {
 701                case '\0':
 702                        return report(options, obj,
 703                                FSCK_MSG_NUL_IN_HEADER,
 704                                "unterminated header: NUL at offset %ld", i);
 705                case '\n':
 706                        if (i + 1 < size && buffer[i + 1] == '\n')
 707                                return 0;
 708                }
 709        }
 710
 711        /*
 712         * We did not find double-LF that separates the header
 713         * and the body.  Not having a body is not a crime but
 714         * we do want to see the terminating LF for the last header
 715         * line.
 716         */
 717        if (size && buffer[size - 1] == '\n')
 718                return 0;
 719
 720        return report(options, obj,
 721                FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 722}
 723
 724static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 725{
 726        const char *p = *ident;
 727        char *end;
 728
 729        *ident = strchrnul(*ident, '\n');
 730        if (**ident == '\n')
 731                (*ident)++;
 732
 733        if (*p == '<')
 734                return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 735        p += strcspn(p, "<>\n");
 736        if (*p == '>')
 737                return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 738        if (*p != '<')
 739                return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 740        if (p[-1] != ' ')
 741                return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 742        p++;
 743        p += strcspn(p, "<>\n");
 744        if (*p != '>')
 745                return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 746        p++;
 747        if (*p != ' ')
 748                return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 749        p++;
 750        if (*p == '0' && p[1] != ' ')
 751                return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 752        if (date_overflows(parse_timestamp(p, &end, 10)))
 753                return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 754        if ((end == p || *end != ' '))
 755                return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 756        p = end + 1;
 757        if ((*p != '+' && *p != '-') ||
 758            !isdigit(p[1]) ||
 759            !isdigit(p[2]) ||
 760            !isdigit(p[3]) ||
 761            !isdigit(p[4]) ||
 762            (p[5] != '\n'))
 763                return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 764        p += 6;
 765        return 0;
 766}
 767
 768static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 769        unsigned long size, struct fsck_options *options)
 770{
 771        struct object_id tree_oid, oid;
 772        struct commit_graft *graft;
 773        unsigned parent_count, parent_line_count = 0, author_count;
 774        int err;
 775        const char *buffer_begin = buffer;
 776        const char *p;
 777
 778        if (verify_headers(buffer, size, &commit->object, options))
 779                return -1;
 780
 781        if (!skip_prefix(buffer, "tree ", &buffer))
 782                return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 783        if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 784                err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 785                if (err)
 786                        return err;
 787        }
 788        buffer = p + 1;
 789        while (skip_prefix(buffer, "parent ", &buffer)) {
 790                if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 791                        err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 792                        if (err)
 793                                return err;
 794                }
 795                buffer = p + 1;
 796                parent_line_count++;
 797        }
 798        graft = lookup_commit_graft(&commit->object.oid);
 799        parent_count = commit_list_count(commit->parents);
 800        if (graft) {
 801                if (graft->nr_parent == -1 && !parent_count)
 802                        ; /* shallow commit */
 803                else if (graft->nr_parent != parent_count) {
 804                        err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 805                        if (err)
 806                                return err;
 807                }
 808        } else {
 809                if (parent_count != parent_line_count) {
 810                        err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 811                        if (err)
 812                                return err;
 813                }
 814        }
 815        author_count = 0;
 816        while (skip_prefix(buffer, "author ", &buffer)) {
 817                author_count++;
 818                err = fsck_ident(&buffer, &commit->object, options);
 819                if (err)
 820                        return err;
 821        }
 822        if (author_count < 1)
 823                err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 824        else if (author_count > 1)
 825                err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 826        if (err)
 827                return err;
 828        if (!skip_prefix(buffer, "committer ", &buffer))
 829                return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 830        err = fsck_ident(&buffer, &commit->object, options);
 831        if (err)
 832                return err;
 833        if (!get_commit_tree(commit)) {
 834                err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
 835                if (err)
 836                        return err;
 837        }
 838        if (memchr(buffer_begin, '\0', size)) {
 839                err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 840                             "NUL byte in the commit object body");
 841                if (err)
 842                        return err;
 843        }
 844        return 0;
 845}
 846
 847static int fsck_commit(struct commit *commit, const char *data,
 848        unsigned long size, struct fsck_options *options)
 849{
 850        const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 851        int ret = fsck_commit_buffer(commit, buffer, size, options);
 852        if (!data)
 853                unuse_commit_buffer(commit, buffer);
 854        return ret;
 855}
 856
 857static int fsck_tag_buffer(struct tag *tag, const char *data,
 858        unsigned long size, struct fsck_options *options)
 859{
 860        struct object_id oid;
 861        int ret = 0;
 862        const char *buffer;
 863        char *to_free = NULL, *eol;
 864        struct strbuf sb = STRBUF_INIT;
 865        const char *p;
 866
 867        if (data)
 868                buffer = data;
 869        else {
 870                enum object_type type;
 871
 872                buffer = to_free =
 873                        read_object_file(&tag->object.oid, &type, &size);
 874                if (!buffer)
 875                        return report(options, &tag->object,
 876                                FSCK_MSG_MISSING_TAG_OBJECT,
 877                                "cannot read tag object");
 878
 879                if (type != OBJ_TAG) {
 880                        ret = report(options, &tag->object,
 881                                FSCK_MSG_TAG_OBJECT_NOT_TAG,
 882                                "expected tag got %s",
 883                            type_name(type));
 884                        goto done;
 885                }
 886        }
 887
 888        ret = verify_headers(buffer, size, &tag->object, options);
 889        if (ret)
 890                goto done;
 891
 892        if (!skip_prefix(buffer, "object ", &buffer)) {
 893                ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 894                goto done;
 895        }
 896        if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 897                ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 898                if (ret)
 899                        goto done;
 900        }
 901        buffer = p + 1;
 902
 903        if (!skip_prefix(buffer, "type ", &buffer)) {
 904                ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 905                goto done;
 906        }
 907        eol = strchr(buffer, '\n');
 908        if (!eol) {
 909                ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 910                goto done;
 911        }
 912        if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 913                ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 914        if (ret)
 915                goto done;
 916        buffer = eol + 1;
 917
 918        if (!skip_prefix(buffer, "tag ", &buffer)) {
 919                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 920                goto done;
 921        }
 922        eol = strchr(buffer, '\n');
 923        if (!eol) {
 924                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 925                goto done;
 926        }
 927        strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 928        if (check_refname_format(sb.buf, 0)) {
 929                ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 930                           "invalid 'tag' name: %.*s",
 931                           (int)(eol - buffer), buffer);
 932                if (ret)
 933                        goto done;
 934        }
 935        buffer = eol + 1;
 936
 937        if (!skip_prefix(buffer, "tagger ", &buffer)) {
 938                /* early tags do not contain 'tagger' lines; warn only */
 939                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 940                if (ret)
 941                        goto done;
 942        }
 943        else
 944                ret = fsck_ident(&buffer, &tag->object, options);
 945
 946done:
 947        strbuf_release(&sb);
 948        free(to_free);
 949        return ret;
 950}
 951
 952static int fsck_tag(struct tag *tag, const char *data,
 953        unsigned long size, struct fsck_options *options)
 954{
 955        struct object *tagged = tag->tagged;
 956
 957        if (!tagged)
 958                return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 959
 960        return fsck_tag_buffer(tag, data, size, options);
 961}
 962
 963struct fsck_gitmodules_data {
 964        struct object *obj;
 965        struct fsck_options *options;
 966        int ret;
 967};
 968
 969static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 970{
 971        struct fsck_gitmodules_data *data = vdata;
 972        const char *subsection, *key;
 973        int subsection_len;
 974        char *name;
 975
 976        if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 977            !subsection)
 978                return 0;
 979
 980        name = xmemdupz(subsection, subsection_len);
 981        if (check_submodule_name(name) < 0)
 982                data->ret |= report(data->options, data->obj,
 983                                    FSCK_MSG_GITMODULES_NAME,
 984                                    "disallowed submodule name: %s",
 985                                    name);
 986        free(name);
 987
 988        return 0;
 989}
 990
 991static int fsck_blob(struct blob *blob, const char *buf,
 992                     unsigned long size, struct fsck_options *options)
 993{
 994        struct fsck_gitmodules_data data;
 995
 996        if (!oidset_contains(&gitmodules_found, &blob->object.oid))
 997                return 0;
 998        oidset_insert(&gitmodules_done, &blob->object.oid);
 999
1000        if (!buf) {
1001                /*
1002                 * A missing buffer here is a sign that the caller found the
1003                 * blob too gigantic to load into memory. Let's just consider
1004                 * that an error.
1005                 */
1006                return report(options, &blob->object,
1007                              FSCK_MSG_GITMODULES_PARSE,
1008                              ".gitmodules too large to parse");
1009        }
1010
1011        data.obj = &blob->object;
1012        data.options = options;
1013        data.ret = 0;
1014        if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1015                                ".gitmodules", buf, size, &data))
1016                data.ret |= report(options, &blob->object,
1017                                   FSCK_MSG_GITMODULES_PARSE,
1018                                   "could not parse gitmodules blob");
1019
1020        return data.ret;
1021}
1022
1023int fsck_object(struct object *obj, void *data, unsigned long size,
1024        struct fsck_options *options)
1025{
1026        if (!obj)
1027                return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1028
1029        if (obj->type == OBJ_BLOB)
1030                return fsck_blob((struct blob *)obj, data, size, options);
1031        if (obj->type == OBJ_TREE)
1032                return fsck_tree((struct tree *) obj, options);
1033        if (obj->type == OBJ_COMMIT)
1034                return fsck_commit((struct commit *) obj, (const char *) data,
1035                        size, options);
1036        if (obj->type == OBJ_TAG)
1037                return fsck_tag((struct tag *) obj, (const char *) data,
1038                        size, options);
1039
1040        return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1041                          obj->type);
1042}
1043
1044int fsck_error_function(struct fsck_options *o,
1045        struct object *obj, int msg_type, const char *message)
1046{
1047        if (msg_type == FSCK_WARN) {
1048                warning("object %s: %s", describe_object(o, obj), message);
1049                return 0;
1050        }
1051        error("object %s: %s", describe_object(o, obj), message);
1052        return 1;
1053}
1054
1055int fsck_finish(struct fsck_options *options)
1056{
1057        int ret = 0;
1058        struct oidset_iter iter;
1059        const struct object_id *oid;
1060
1061        oidset_iter_init(&gitmodules_found, &iter);
1062        while ((oid = oidset_iter_next(&iter))) {
1063                struct blob *blob;
1064                enum object_type type;
1065                unsigned long size;
1066                char *buf;
1067
1068                if (oidset_contains(&gitmodules_done, oid))
1069                        continue;
1070
1071                blob = lookup_blob(oid);
1072                if (!blob) {
1073                        struct object *obj = lookup_unknown_object(oid->hash);
1074                        ret |= report(options, obj,
1075                                      FSCK_MSG_GITMODULES_BLOB,
1076                                      "non-blob found at .gitmodules");
1077                        continue;
1078                }
1079
1080                buf = read_object_file(oid, &type, &size);
1081                if (!buf) {
1082                        if (is_promisor_object(&blob->object.oid))
1083                                continue;
1084                        ret |= report(options, &blob->object,
1085                                      FSCK_MSG_GITMODULES_MISSING,
1086                                      "unable to read .gitmodules blob");
1087                        continue;
1088                }
1089
1090                if (type == OBJ_BLOB)
1091                        ret |= fsck_blob(blob, buf, size, options);
1092                else
1093                        ret |= report(options, &blob->object,
1094                                      FSCK_MSG_GITMODULES_BLOB,
1095                                      "non-blob found at .gitmodules");
1096                free(buf);
1097        }
1098
1099
1100        oidset_clear(&gitmodules_found);
1101        oidset_clear(&gitmodules_done);
1102        return ret;
1103}