Andrew's git / gitweb.git / fsck.c
?
fsck.con commit Merge branch 'en/rebase-i-microfixes' (b345b77)
   1#include "cache.h"
   2#include "object-store.h"
   3#include "object.h"
   4#include "blob.h"
   5#include "tree.h"
   6#include "tree-walk.h"
   7#include "commit.h"
   8#include "tag.h"
   9#include "fsck.h"
  10#include "refs.h"
  11#include "utf8.h"
  12#include "sha1-array.h"
  13#include "decorate.h"
  14#include "oidset.h"
  15#include "packfile.h"
  16#include "submodule-config.h"
  17#include "config.h"
  18#include "help.h"
  19
  20static struct oidset gitmodules_found = OIDSET_INIT;
  21static struct oidset gitmodules_done = OIDSET_INIT;
  22
  23#define FSCK_FATAL -1
  24#define FSCK_INFO -2
  25
  26#define FOREACH_MSG_ID(FUNC) \
  27        /* fatal errors */ \
  28        FUNC(NUL_IN_HEADER, FATAL) \
  29        FUNC(UNTERMINATED_HEADER, FATAL) \
  30        /* errors */ \
  31        FUNC(BAD_DATE, ERROR) \
  32        FUNC(BAD_DATE_OVERFLOW, ERROR) \
  33        FUNC(BAD_EMAIL, ERROR) \
  34        FUNC(BAD_NAME, ERROR) \
  35        FUNC(BAD_OBJECT_SHA1, ERROR) \
  36        FUNC(BAD_PARENT_SHA1, ERROR) \
  37        FUNC(BAD_TAG_OBJECT, ERROR) \
  38        FUNC(BAD_TIMEZONE, ERROR) \
  39        FUNC(BAD_TREE, ERROR) \
  40        FUNC(BAD_TREE_SHA1, ERROR) \
  41        FUNC(BAD_TYPE, ERROR) \
  42        FUNC(DUPLICATE_ENTRIES, ERROR) \
  43        FUNC(MISSING_AUTHOR, ERROR) \
  44        FUNC(MISSING_COMMITTER, ERROR) \
  45        FUNC(MISSING_EMAIL, ERROR) \
  46        FUNC(MISSING_GRAFT, ERROR) \
  47        FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  48        FUNC(MISSING_OBJECT, ERROR) \
  49        FUNC(MISSING_PARENT, ERROR) \
  50        FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  51        FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  52        FUNC(MISSING_TAG, ERROR) \
  53        FUNC(MISSING_TAG_ENTRY, ERROR) \
  54        FUNC(MISSING_TAG_OBJECT, ERROR) \
  55        FUNC(MISSING_TREE, ERROR) \
  56        FUNC(MISSING_TREE_OBJECT, ERROR) \
  57        FUNC(MISSING_TYPE, ERROR) \
  58        FUNC(MISSING_TYPE_ENTRY, ERROR) \
  59        FUNC(MULTIPLE_AUTHORS, ERROR) \
  60        FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  61        FUNC(TREE_NOT_SORTED, ERROR) \
  62        FUNC(UNKNOWN_TYPE, ERROR) \
  63        FUNC(ZERO_PADDED_DATE, ERROR) \
  64        FUNC(GITMODULES_MISSING, ERROR) \
  65        FUNC(GITMODULES_BLOB, ERROR) \
  66        FUNC(GITMODULES_PARSE, ERROR) \
  67        FUNC(GITMODULES_NAME, ERROR) \
  68        FUNC(GITMODULES_SYMLINK, ERROR) \
  69        /* warnings */ \
  70        FUNC(BAD_FILEMODE, WARN) \
  71        FUNC(EMPTY_NAME, WARN) \
  72        FUNC(FULL_PATHNAME, WARN) \
  73        FUNC(HAS_DOT, WARN) \
  74        FUNC(HAS_DOTDOT, WARN) \
  75        FUNC(HAS_DOTGIT, WARN) \
  76        FUNC(NULL_SHA1, WARN) \
  77        FUNC(ZERO_PADDED_FILEMODE, WARN) \
  78        FUNC(NUL_IN_COMMIT, WARN) \
  79        /* infos (reported as warnings, but ignored by default) */ \
  80        FUNC(BAD_TAG_NAME, INFO) \
  81        FUNC(MISSING_TAGGER_ENTRY, INFO)
  82
  83#define MSG_ID(id, msg_type) FSCK_MSG_##id,
  84enum fsck_msg_id {
  85        FOREACH_MSG_ID(MSG_ID)
  86        FSCK_MSG_MAX
  87};
  88#undef MSG_ID
  89
  90#define STR(x) #x
  91#define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
  92static struct {
  93        const char *id_string;
  94        const char *downcased;
  95        const char *camelcased;
  96        int msg_type;
  97} msg_id_info[FSCK_MSG_MAX + 1] = {
  98        FOREACH_MSG_ID(MSG_ID)
  99        { NULL, NULL, NULL, -1 }
 100};
 101#undef MSG_ID
 102
 103static void prepare_msg_ids(void)
 104{
 105        int i;
 106
 107        if (msg_id_info[0].downcased)
 108                return;
 109
 110        /* convert id_string to lower case, without underscores. */
 111        for (i = 0; i < FSCK_MSG_MAX; i++) {
 112                const char *p = msg_id_info[i].id_string;
 113                int len = strlen(p);
 114                char *q = xmalloc(len);
 115
 116                msg_id_info[i].downcased = q;
 117                while (*p)
 118                        if (*p == '_')
 119                                p++;
 120                        else
 121                                *(q)++ = tolower(*(p)++);
 122                *q = '\0';
 123
 124                p = msg_id_info[i].id_string;
 125                q = xmalloc(len);
 126                msg_id_info[i].camelcased = q;
 127                while (*p) {
 128                        if (*p == '_') {
 129                                p++;
 130                                if (*p)
 131                                        *q++ = *p++;
 132                        } else {
 133                                *q++ = tolower(*p++);
 134                        }
 135                }
 136                *q = '\0';
 137        }
 138}
 139
 140static int parse_msg_id(const char *text)
 141{
 142        int i;
 143
 144        prepare_msg_ids();
 145
 146        for (i = 0; i < FSCK_MSG_MAX; i++)
 147                if (!strcmp(text, msg_id_info[i].downcased))
 148                        return i;
 149
 150        return -1;
 151}
 152
 153void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
 154{
 155        int i;
 156
 157        prepare_msg_ids();
 158
 159        for (i = 0; i < FSCK_MSG_MAX; i++)
 160                list_config_item(list, prefix, msg_id_info[i].camelcased);
 161}
 162
 163static int fsck_msg_type(enum fsck_msg_id msg_id,
 164        struct fsck_options *options)
 165{
 166        int msg_type;
 167
 168        assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 169
 170        if (options->msg_type)
 171                msg_type = options->msg_type[msg_id];
 172        else {
 173                msg_type = msg_id_info[msg_id].msg_type;
 174                if (options->strict && msg_type == FSCK_WARN)
 175                        msg_type = FSCK_ERROR;
 176        }
 177
 178        return msg_type;
 179}
 180
 181static void init_skiplist(struct fsck_options *options, const char *path)
 182{
 183        static struct oid_array skiplist = OID_ARRAY_INIT;
 184        int sorted, fd;
 185        char buffer[GIT_MAX_HEXSZ + 1];
 186        struct object_id oid;
 187
 188        if (options->skiplist)
 189                sorted = options->skiplist->sorted;
 190        else {
 191                sorted = 1;
 192                options->skiplist = &skiplist;
 193        }
 194
 195        fd = open(path, O_RDONLY);
 196        if (fd < 0)
 197                die("Could not open skip list: %s", path);
 198        for (;;) {
 199                const char *p;
 200                int result = read_in_full(fd, buffer, sizeof(buffer));
 201                if (result < 0)
 202                        die_errno("Could not read '%s'", path);
 203                if (!result)
 204                        break;
 205                if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 206                        die("Invalid SHA-1: %s", buffer);
 207                oid_array_append(&skiplist, &oid);
 208                if (sorted && skiplist.nr > 1 &&
 209                                oidcmp(&skiplist.oid[skiplist.nr - 2],
 210                                       &oid) > 0)
 211                        sorted = 0;
 212        }
 213        close(fd);
 214
 215        if (sorted)
 216                skiplist.sorted = 1;
 217}
 218
 219static int parse_msg_type(const char *str)
 220{
 221        if (!strcmp(str, "error"))
 222                return FSCK_ERROR;
 223        else if (!strcmp(str, "warn"))
 224                return FSCK_WARN;
 225        else if (!strcmp(str, "ignore"))
 226                return FSCK_IGNORE;
 227        else
 228                die("Unknown fsck message type: '%s'", str);
 229}
 230
 231int is_valid_msg_type(const char *msg_id, const char *msg_type)
 232{
 233        if (parse_msg_id(msg_id) < 0)
 234                return 0;
 235        parse_msg_type(msg_type);
 236        return 1;
 237}
 238
 239void fsck_set_msg_type(struct fsck_options *options,
 240                const char *msg_id, const char *msg_type)
 241{
 242        int id = parse_msg_id(msg_id), type;
 243
 244        if (id < 0)
 245                die("Unhandled message id: %s", msg_id);
 246        type = parse_msg_type(msg_type);
 247
 248        if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 249                die("Cannot demote %s to %s", msg_id, msg_type);
 250
 251        if (!options->msg_type) {
 252                int i;
 253                int *msg_type;
 254                ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 255                for (i = 0; i < FSCK_MSG_MAX; i++)
 256                        msg_type[i] = fsck_msg_type(i, options);
 257                options->msg_type = msg_type;
 258        }
 259
 260        options->msg_type[id] = type;
 261}
 262
 263void fsck_set_msg_types(struct fsck_options *options, const char *values)
 264{
 265        char *buf = xstrdup(values), *to_free = buf;
 266        int done = 0;
 267
 268        while (!done) {
 269                int len = strcspn(buf, " ,|"), equal;
 270
 271                done = !buf[len];
 272                if (!len) {
 273                        buf++;
 274                        continue;
 275                }
 276                buf[len] = '\0';
 277
 278                for (equal = 0;
 279                     equal < len && buf[equal] != '=' && buf[equal] != ':';
 280                     equal++)
 281                        buf[equal] = tolower(buf[equal]);
 282                buf[equal] = '\0';
 283
 284                if (!strcmp(buf, "skiplist")) {
 285                        if (equal == len)
 286                                die("skiplist requires a path");
 287                        init_skiplist(options, buf + equal + 1);
 288                        buf += len + 1;
 289                        continue;
 290                }
 291
 292                if (equal == len)
 293                        die("Missing '=': '%s'", buf);
 294
 295                fsck_set_msg_type(options, buf, buf + equal + 1);
 296                buf += len + 1;
 297        }
 298        free(to_free);
 299}
 300
 301static void append_msg_id(struct strbuf *sb, const char *msg_id)
 302{
 303        for (;;) {
 304                char c = *(msg_id)++;
 305
 306                if (!c)
 307                        break;
 308                if (c != '_')
 309                        strbuf_addch(sb, tolower(c));
 310                else {
 311                        assert(*msg_id);
 312                        strbuf_addch(sb, *(msg_id)++);
 313                }
 314        }
 315
 316        strbuf_addstr(sb, ": ");
 317}
 318
 319__attribute__((format (printf, 4, 5)))
 320static int report(struct fsck_options *options, struct object *object,
 321        enum fsck_msg_id id, const char *fmt, ...)
 322{
 323        va_list ap;
 324        struct strbuf sb = STRBUF_INIT;
 325        int msg_type = fsck_msg_type(id, options), result;
 326
 327        if (msg_type == FSCK_IGNORE)
 328                return 0;
 329
 330        if (options->skiplist && object &&
 331                        oid_array_lookup(options->skiplist, &object->oid) >= 0)
 332                return 0;
 333
 334        if (msg_type == FSCK_FATAL)
 335                msg_type = FSCK_ERROR;
 336        else if (msg_type == FSCK_INFO)
 337                msg_type = FSCK_WARN;
 338
 339        append_msg_id(&sb, msg_id_info[id].id_string);
 340
 341        va_start(ap, fmt);
 342        strbuf_vaddf(&sb, fmt, ap);
 343        result = options->error_func(options, object, msg_type, sb.buf);
 344        strbuf_release(&sb);
 345        va_end(ap);
 346
 347        return result;
 348}
 349
 350static char *get_object_name(struct fsck_options *options, struct object *obj)
 351{
 352        if (!options->object_names)
 353                return NULL;
 354        return lookup_decoration(options->object_names, obj);
 355}
 356
 357static void put_object_name(struct fsck_options *options, struct object *obj,
 358        const char *fmt, ...)
 359{
 360        va_list ap;
 361        struct strbuf buf = STRBUF_INIT;
 362        char *existing;
 363
 364        if (!options->object_names)
 365                return;
 366        existing = lookup_decoration(options->object_names, obj);
 367        if (existing)
 368                return;
 369        va_start(ap, fmt);
 370        strbuf_vaddf(&buf, fmt, ap);
 371        add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 372        va_end(ap);
 373}
 374
 375static const char *describe_object(struct fsck_options *o, struct object *obj)
 376{
 377        static struct strbuf buf = STRBUF_INIT;
 378        char *name;
 379
 380        strbuf_reset(&buf);
 381        strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 382        if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 383                strbuf_addf(&buf, " (%s)", name);
 384
 385        return buf.buf;
 386}
 387
 388static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 389{
 390        struct tree_desc desc;
 391        struct name_entry entry;
 392        int res = 0;
 393        const char *name;
 394
 395        if (parse_tree(tree))
 396                return -1;
 397
 398        name = get_object_name(options, &tree->object);
 399        if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 400                return -1;
 401        while (tree_entry_gently(&desc, &entry)) {
 402                struct object *obj;
 403                int result;
 404
 405                if (S_ISGITLINK(entry.mode))
 406                        continue;
 407
 408                if (S_ISDIR(entry.mode)) {
 409                        obj = (struct object *)lookup_tree(entry.oid);
 410                        if (name && obj)
 411                                put_object_name(options, obj, "%s%s/", name,
 412                                        entry.path);
 413                        result = options->walk(obj, OBJ_TREE, data, options);
 414                }
 415                else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 416                        obj = (struct object *)lookup_blob(entry.oid);
 417                        if (name && obj)
 418                                put_object_name(options, obj, "%s%s", name,
 419                                        entry.path);
 420                        result = options->walk(obj, OBJ_BLOB, data, options);
 421                }
 422                else {
 423                        result = error("in tree %s: entry %s has bad mode %.6o",
 424                                        describe_object(options, &tree->object), entry.path, entry.mode);
 425                }
 426                if (result < 0)
 427                        return result;
 428                if (!res)
 429                        res = result;
 430        }
 431        return res;
 432}
 433
 434static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 435{
 436        int counter = 0, generation = 0, name_prefix_len = 0;
 437        struct commit_list *parents;
 438        int res;
 439        int result;
 440        const char *name;
 441
 442        if (parse_commit(commit))
 443                return -1;
 444
 445        name = get_object_name(options, &commit->object);
 446        if (name)
 447                put_object_name(options, &get_commit_tree(commit)->object,
 448                                "%s:", name);
 449
 450        result = options->walk((struct object *)get_commit_tree(commit),
 451                               OBJ_TREE, data, options);
 452        if (result < 0)
 453                return result;
 454        res = result;
 455
 456        parents = commit->parents;
 457        if (name && parents) {
 458                int len = strlen(name), power;
 459
 460                if (len && name[len - 1] == '^') {
 461                        generation = 1;
 462                        name_prefix_len = len - 1;
 463                }
 464                else { /* parse ~<generation> suffix */
 465                        for (generation = 0, power = 1;
 466                             len && isdigit(name[len - 1]);
 467                             power *= 10)
 468                                generation += power * (name[--len] - '0');
 469                        if (power > 1 && len && name[len - 1] == '~')
 470                                name_prefix_len = len - 1;
 471                }
 472        }
 473
 474        while (parents) {
 475                if (name) {
 476                        struct object *obj = &parents->item->object;
 477
 478                        if (++counter > 1)
 479                                put_object_name(options, obj, "%s^%d",
 480                                        name, counter);
 481                        else if (generation > 0)
 482                                put_object_name(options, obj, "%.*s~%d",
 483                                        name_prefix_len, name, generation + 1);
 484                        else
 485                                put_object_name(options, obj, "%s^", name);
 486                }
 487                result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 488                if (result < 0)
 489                        return result;
 490                if (!res)
 491                        res = result;
 492                parents = parents->next;
 493        }
 494        return res;
 495}
 496
 497static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 498{
 499        char *name = get_object_name(options, &tag->object);
 500
 501        if (parse_tag(tag))
 502                return -1;
 503        if (name)
 504                put_object_name(options, tag->tagged, "%s", name);
 505        return options->walk(tag->tagged, OBJ_ANY, data, options);
 506}
 507
 508int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 509{
 510        if (!obj)
 511                return -1;
 512
 513        if (obj->type == OBJ_NONE)
 514                parse_object(&obj->oid);
 515
 516        switch (obj->type) {
 517        case OBJ_BLOB:
 518                return 0;
 519        case OBJ_TREE:
 520                return fsck_walk_tree((struct tree *)obj, data, options);
 521        case OBJ_COMMIT:
 522                return fsck_walk_commit((struct commit *)obj, data, options);
 523        case OBJ_TAG:
 524                return fsck_walk_tag((struct tag *)obj, data, options);
 525        default:
 526                error("Unknown object type for %s", describe_object(options, obj));
 527                return -1;
 528        }
 529}
 530
 531/*
 532 * The entries in a tree are ordered in the _path_ order,
 533 * which means that a directory entry is ordered by adding
 534 * a slash to the end of it.
 535 *
 536 * So a directory called "a" is ordered _after_ a file
 537 * called "a.c", because "a/" sorts after "a.c".
 538 */
 539#define TREE_UNORDERED (-1)
 540#define TREE_HAS_DUPS  (-2)
 541
 542static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 543{
 544        int len1 = strlen(name1);
 545        int len2 = strlen(name2);
 546        int len = len1 < len2 ? len1 : len2;
 547        unsigned char c1, c2;
 548        int cmp;
 549
 550        cmp = memcmp(name1, name2, len);
 551        if (cmp < 0)
 552                return 0;
 553        if (cmp > 0)
 554                return TREE_UNORDERED;
 555
 556        /*
 557         * Ok, the first <len> characters are the same.
 558         * Now we need to order the next one, but turn
 559         * a '\0' into a '/' for a directory entry.
 560         */
 561        c1 = name1[len];
 562        c2 = name2[len];
 563        if (!c1 && !c2)
 564                /*
 565                 * git-write-tree used to write out a nonsense tree that has
 566                 * entries with the same name, one blob and one tree.  Make
 567                 * sure we do not have duplicate entries.
 568                 */
 569                return TREE_HAS_DUPS;
 570        if (!c1 && S_ISDIR(mode1))
 571                c1 = '/';
 572        if (!c2 && S_ISDIR(mode2))
 573                c2 = '/';
 574        return c1 < c2 ? 0 : TREE_UNORDERED;
 575}
 576
 577static int fsck_tree(struct tree *item, struct fsck_options *options)
 578{
 579        int retval = 0;
 580        int has_null_sha1 = 0;
 581        int has_full_path = 0;
 582        int has_empty_name = 0;
 583        int has_dot = 0;
 584        int has_dotdot = 0;
 585        int has_dotgit = 0;
 586        int has_zero_pad = 0;
 587        int has_bad_modes = 0;
 588        int has_dup_entries = 0;
 589        int not_properly_sorted = 0;
 590        struct tree_desc desc;
 591        unsigned o_mode;
 592        const char *o_name;
 593
 594        if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 595                retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 596                return retval;
 597        }
 598
 599        o_mode = 0;
 600        o_name = NULL;
 601
 602        while (desc.size) {
 603                unsigned mode;
 604                const char *name;
 605                const struct object_id *oid;
 606
 607                oid = tree_entry_extract(&desc, &name, &mode);
 608
 609                has_null_sha1 |= is_null_oid(oid);
 610                has_full_path |= !!strchr(name, '/');
 611                has_empty_name |= !*name;
 612                has_dot |= !strcmp(name, ".");
 613                has_dotdot |= !strcmp(name, "..");
 614                has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 615                has_zero_pad |= *(char *)desc.buffer == '0';
 616
 617                if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 618                        if (!S_ISLNK(mode))
 619                                oidset_insert(&gitmodules_found, oid);
 620                        else
 621                                retval += report(options, &item->object,
 622                                                 FSCK_MSG_GITMODULES_SYMLINK,
 623                                                 ".gitmodules is a symbolic link");
 624                }
 625
 626                if (update_tree_entry_gently(&desc)) {
 627                        retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 628                        break;
 629                }
 630
 631                switch (mode) {
 632                /*
 633                 * Standard modes..
 634                 */
 635                case S_IFREG | 0755:
 636                case S_IFREG | 0644:
 637                case S_IFLNK:
 638                case S_IFDIR:
 639                case S_IFGITLINK:
 640                        break;
 641                /*
 642                 * This is nonstandard, but we had a few of these
 643                 * early on when we honored the full set of mode
 644                 * bits..
 645                 */
 646                case S_IFREG | 0664:
 647                        if (!options->strict)
 648                                break;
 649                        /* fallthrough */
 650                default:
 651                        has_bad_modes = 1;
 652                }
 653
 654                if (o_name) {
 655                        switch (verify_ordered(o_mode, o_name, mode, name)) {
 656                        case TREE_UNORDERED:
 657                                not_properly_sorted = 1;
 658                                break;
 659                        case TREE_HAS_DUPS:
 660                                has_dup_entries = 1;
 661                                break;
 662                        default:
 663                                break;
 664                        }
 665                }
 666
 667                o_mode = mode;
 668                o_name = name;
 669        }
 670
 671        if (has_null_sha1)
 672                retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 673        if (has_full_path)
 674                retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 675        if (has_empty_name)
 676                retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 677        if (has_dot)
 678                retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 679        if (has_dotdot)
 680                retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 681        if (has_dotgit)
 682                retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 683        if (has_zero_pad)
 684                retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 685        if (has_bad_modes)
 686                retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 687        if (has_dup_entries)
 688                retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 689        if (not_properly_sorted)
 690                retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 691        return retval;
 692}
 693
 694static int verify_headers(const void *data, unsigned long size,
 695                          struct object *obj, struct fsck_options *options)
 696{
 697        const char *buffer = (const char *)data;
 698        unsigned long i;
 699
 700        for (i = 0; i < size; i++) {
 701                switch (buffer[i]) {
 702                case '\0':
 703                        return report(options, obj,
 704                                FSCK_MSG_NUL_IN_HEADER,
 705                                "unterminated header: NUL at offset %ld", i);
 706                case '\n':
 707                        if (i + 1 < size && buffer[i + 1] == '\n')
 708                                return 0;
 709                }
 710        }
 711
 712        /*
 713         * We did not find double-LF that separates the header
 714         * and the body.  Not having a body is not a crime but
 715         * we do want to see the terminating LF for the last header
 716         * line.
 717         */
 718        if (size && buffer[size - 1] == '\n')
 719                return 0;
 720
 721        return report(options, obj,
 722                FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 723}
 724
 725static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 726{
 727        const char *p = *ident;
 728        char *end;
 729
 730        *ident = strchrnul(*ident, '\n');
 731        if (**ident == '\n')
 732                (*ident)++;
 733
 734        if (*p == '<')
 735                return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 736        p += strcspn(p, "<>\n");
 737        if (*p == '>')
 738                return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 739        if (*p != '<')
 740                return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 741        if (p[-1] != ' ')
 742                return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 743        p++;
 744        p += strcspn(p, "<>\n");
 745        if (*p != '>')
 746                return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 747        p++;
 748        if (*p != ' ')
 749                return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 750        p++;
 751        if (*p == '0' && p[1] != ' ')
 752                return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 753        if (date_overflows(parse_timestamp(p, &end, 10)))
 754                return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 755        if ((end == p || *end != ' '))
 756                return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 757        p = end + 1;
 758        if ((*p != '+' && *p != '-') ||
 759            !isdigit(p[1]) ||
 760            !isdigit(p[2]) ||
 761            !isdigit(p[3]) ||
 762            !isdigit(p[4]) ||
 763            (p[5] != '\n'))
 764                return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 765        p += 6;
 766        return 0;
 767}
 768
 769static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 770        unsigned long size, struct fsck_options *options)
 771{
 772        struct object_id tree_oid, oid;
 773        struct commit_graft *graft;
 774        unsigned parent_count, parent_line_count = 0, author_count;
 775        int err;
 776        const char *buffer_begin = buffer;
 777        const char *p;
 778
 779        if (verify_headers(buffer, size, &commit->object, options))
 780                return -1;
 781
 782        if (!skip_prefix(buffer, "tree ", &buffer))
 783                return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 784        if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 785                err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 786                if (err)
 787                        return err;
 788        }
 789        buffer = p + 1;
 790        while (skip_prefix(buffer, "parent ", &buffer)) {
 791                if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 792                        err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 793                        if (err)
 794                                return err;
 795                }
 796                buffer = p + 1;
 797                parent_line_count++;
 798        }
 799        graft = lookup_commit_graft(the_repository, &commit->object.oid);
 800        parent_count = commit_list_count(commit->parents);
 801        if (graft) {
 802                if (graft->nr_parent == -1 && !parent_count)
 803                        ; /* shallow commit */
 804                else if (graft->nr_parent != parent_count) {
 805                        err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 806                        if (err)
 807                                return err;
 808                }
 809        } else {
 810                if (parent_count != parent_line_count) {
 811                        err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 812                        if (err)
 813                                return err;
 814                }
 815        }
 816        author_count = 0;
 817        while (skip_prefix(buffer, "author ", &buffer)) {
 818                author_count++;
 819                err = fsck_ident(&buffer, &commit->object, options);
 820                if (err)
 821                        return err;
 822        }
 823        if (author_count < 1)
 824                err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 825        else if (author_count > 1)
 826                err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 827        if (err)
 828                return err;
 829        if (!skip_prefix(buffer, "committer ", &buffer))
 830                return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 831        err = fsck_ident(&buffer, &commit->object, options);
 832        if (err)
 833                return err;
 834        if (!get_commit_tree(commit)) {
 835                err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
 836                if (err)
 837                        return err;
 838        }
 839        if (memchr(buffer_begin, '\0', size)) {
 840                err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 841                             "NUL byte in the commit object body");
 842                if (err)
 843                        return err;
 844        }
 845        return 0;
 846}
 847
 848static int fsck_commit(struct commit *commit, const char *data,
 849        unsigned long size, struct fsck_options *options)
 850{
 851        const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 852        int ret = fsck_commit_buffer(commit, buffer, size, options);
 853        if (!data)
 854                unuse_commit_buffer(commit, buffer);
 855        return ret;
 856}
 857
 858static int fsck_tag_buffer(struct tag *tag, const char *data,
 859        unsigned long size, struct fsck_options *options)
 860{
 861        struct object_id oid;
 862        int ret = 0;
 863        const char *buffer;
 864        char *to_free = NULL, *eol;
 865        struct strbuf sb = STRBUF_INIT;
 866        const char *p;
 867
 868        if (data)
 869                buffer = data;
 870        else {
 871                enum object_type type;
 872
 873                buffer = to_free =
 874                        read_object_file(&tag->object.oid, &type, &size);
 875                if (!buffer)
 876                        return report(options, &tag->object,
 877                                FSCK_MSG_MISSING_TAG_OBJECT,
 878                                "cannot read tag object");
 879
 880                if (type != OBJ_TAG) {
 881                        ret = report(options, &tag->object,
 882                                FSCK_MSG_TAG_OBJECT_NOT_TAG,
 883                                "expected tag got %s",
 884                            type_name(type));
 885                        goto done;
 886                }
 887        }
 888
 889        ret = verify_headers(buffer, size, &tag->object, options);
 890        if (ret)
 891                goto done;
 892
 893        if (!skip_prefix(buffer, "object ", &buffer)) {
 894                ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 895                goto done;
 896        }
 897        if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 898                ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 899                if (ret)
 900                        goto done;
 901        }
 902        buffer = p + 1;
 903
 904        if (!skip_prefix(buffer, "type ", &buffer)) {
 905                ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 906                goto done;
 907        }
 908        eol = strchr(buffer, '\n');
 909        if (!eol) {
 910                ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 911                goto done;
 912        }
 913        if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 914                ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 915        if (ret)
 916                goto done;
 917        buffer = eol + 1;
 918
 919        if (!skip_prefix(buffer, "tag ", &buffer)) {
 920                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 921                goto done;
 922        }
 923        eol = strchr(buffer, '\n');
 924        if (!eol) {
 925                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 926                goto done;
 927        }
 928        strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 929        if (check_refname_format(sb.buf, 0)) {
 930                ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 931                           "invalid 'tag' name: %.*s",
 932                           (int)(eol - buffer), buffer);
 933                if (ret)
 934                        goto done;
 935        }
 936        buffer = eol + 1;
 937
 938        if (!skip_prefix(buffer, "tagger ", &buffer)) {
 939                /* early tags do not contain 'tagger' lines; warn only */
 940                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 941                if (ret)
 942                        goto done;
 943        }
 944        else
 945                ret = fsck_ident(&buffer, &tag->object, options);
 946
 947done:
 948        strbuf_release(&sb);
 949        free(to_free);
 950        return ret;
 951}
 952
 953static int fsck_tag(struct tag *tag, const char *data,
 954        unsigned long size, struct fsck_options *options)
 955{
 956        struct object *tagged = tag->tagged;
 957
 958        if (!tagged)
 959                return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 960
 961        return fsck_tag_buffer(tag, data, size, options);
 962}
 963
 964struct fsck_gitmodules_data {
 965        struct object *obj;
 966        struct fsck_options *options;
 967        int ret;
 968};
 969
 970static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 971{
 972        struct fsck_gitmodules_data *data = vdata;
 973        const char *subsection, *key;
 974        int subsection_len;
 975        char *name;
 976
 977        if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 978            !subsection)
 979                return 0;
 980
 981        name = xmemdupz(subsection, subsection_len);
 982        if (check_submodule_name(name) < 0)
 983                data->ret |= report(data->options, data->obj,
 984                                    FSCK_MSG_GITMODULES_NAME,
 985                                    "disallowed submodule name: %s",
 986                                    name);
 987        free(name);
 988
 989        return 0;
 990}
 991
 992static int fsck_blob(struct blob *blob, const char *buf,
 993                     unsigned long size, struct fsck_options *options)
 994{
 995        struct fsck_gitmodules_data data;
 996
 997        if (!oidset_contains(&gitmodules_found, &blob->object.oid))
 998                return 0;
 999        oidset_insert(&gitmodules_done, &blob->object.oid);
1000
1001        if (!buf) {
1002                /*
1003                 * A missing buffer here is a sign that the caller found the
1004                 * blob too gigantic to load into memory. Let's just consider
1005                 * that an error.
1006                 */
1007                return report(options, &blob->object,
1008                              FSCK_MSG_GITMODULES_PARSE,
1009                              ".gitmodules too large to parse");
1010        }
1011
1012        data.obj = &blob->object;
1013        data.options = options;
1014        data.ret = 0;
1015        if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1016                                ".gitmodules", buf, size, &data))
1017                data.ret |= report(options, &blob->object,
1018                                   FSCK_MSG_GITMODULES_PARSE,
1019                                   "could not parse gitmodules blob");
1020
1021        return data.ret;
1022}
1023
1024int fsck_object(struct object *obj, void *data, unsigned long size,
1025        struct fsck_options *options)
1026{
1027        if (!obj)
1028                return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1029
1030        if (obj->type == OBJ_BLOB)
1031                return fsck_blob((struct blob *)obj, data, size, options);
1032        if (obj->type == OBJ_TREE)
1033                return fsck_tree((struct tree *) obj, options);
1034        if (obj->type == OBJ_COMMIT)
1035                return fsck_commit((struct commit *) obj, (const char *) data,
1036                        size, options);
1037        if (obj->type == OBJ_TAG)
1038                return fsck_tag((struct tag *) obj, (const char *) data,
1039                        size, options);
1040
1041        return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1042                          obj->type);
1043}
1044
1045int fsck_error_function(struct fsck_options *o,
1046        struct object *obj, int msg_type, const char *message)
1047{
1048        if (msg_type == FSCK_WARN) {
1049                warning("object %s: %s", describe_object(o, obj), message);
1050                return 0;
1051        }
1052        error("object %s: %s", describe_object(o, obj), message);
1053        return 1;
1054}
1055
1056int fsck_finish(struct fsck_options *options)
1057{
1058        int ret = 0;
1059        struct oidset_iter iter;
1060        const struct object_id *oid;
1061
1062        oidset_iter_init(&gitmodules_found, &iter);
1063        while ((oid = oidset_iter_next(&iter))) {
1064                struct blob *blob;
1065                enum object_type type;
1066                unsigned long size;
1067                char *buf;
1068
1069                if (oidset_contains(&gitmodules_done, oid))
1070                        continue;
1071
1072                blob = lookup_blob(oid);
1073                if (!blob) {
1074                        struct object *obj = lookup_unknown_object(oid->hash);
1075                        ret |= report(options, obj,
1076                                      FSCK_MSG_GITMODULES_BLOB,
1077                                      "non-blob found at .gitmodules");
1078                        continue;
1079                }
1080
1081                buf = read_object_file(oid, &type, &size);
1082                if (!buf) {
1083                        if (is_promisor_object(&blob->object.oid))
1084                                continue;
1085                        ret |= report(options, &blob->object,
1086                                      FSCK_MSG_GITMODULES_MISSING,
1087                                      "unable to read .gitmodules blob");
1088                        continue;
1089                }
1090
1091                if (type == OBJ_BLOB)
1092                        ret |= fsck_blob(blob, buf, size, options);
1093                else
1094                        ret |= report(options, &blob->object,
1095                                      FSCK_MSG_GITMODULES_BLOB,
1096                                      "non-blob found at .gitmodules");
1097                free(buf);
1098        }
1099
1100
1101        oidset_clear(&gitmodules_found);
1102        oidset_clear(&gitmodules_done);
1103        return ret;
1104}