Andrew's git / gitweb.git / fsck.c
?
fsck.con commit fsck: check skiplist for object in fsck_blob() (fb16287)
   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
  18static struct oidset gitmodules_found = OIDSET_INIT;
  19static struct oidset gitmodules_done = OIDSET_INIT;
  20
  21#define FSCK_FATAL -1
  22#define FSCK_INFO -2
  23
  24#define FOREACH_MSG_ID(FUNC) \
  25        /* fatal errors */ \
  26        FUNC(NUL_IN_HEADER, FATAL) \
  27        FUNC(UNTERMINATED_HEADER, FATAL) \
  28        /* errors */ \
  29        FUNC(BAD_DATE, ERROR) \
  30        FUNC(BAD_DATE_OVERFLOW, ERROR) \
  31        FUNC(BAD_EMAIL, ERROR) \
  32        FUNC(BAD_NAME, ERROR) \
  33        FUNC(BAD_OBJECT_SHA1, ERROR) \
  34        FUNC(BAD_PARENT_SHA1, ERROR) \
  35        FUNC(BAD_TAG_OBJECT, ERROR) \
  36        FUNC(BAD_TIMEZONE, ERROR) \
  37        FUNC(BAD_TREE, ERROR) \
  38        FUNC(BAD_TREE_SHA1, ERROR) \
  39        FUNC(BAD_TYPE, ERROR) \
  40        FUNC(DUPLICATE_ENTRIES, ERROR) \
  41        FUNC(MISSING_AUTHOR, ERROR) \
  42        FUNC(MISSING_COMMITTER, ERROR) \
  43        FUNC(MISSING_EMAIL, ERROR) \
  44        FUNC(MISSING_GRAFT, ERROR) \
  45        FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  46        FUNC(MISSING_OBJECT, ERROR) \
  47        FUNC(MISSING_PARENT, ERROR) \
  48        FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  49        FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  50        FUNC(MISSING_TAG, ERROR) \
  51        FUNC(MISSING_TAG_ENTRY, ERROR) \
  52        FUNC(MISSING_TAG_OBJECT, ERROR) \
  53        FUNC(MISSING_TREE, ERROR) \
  54        FUNC(MISSING_TREE_OBJECT, ERROR) \
  55        FUNC(MISSING_TYPE, ERROR) \
  56        FUNC(MISSING_TYPE_ENTRY, ERROR) \
  57        FUNC(MULTIPLE_AUTHORS, ERROR) \
  58        FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  59        FUNC(TREE_NOT_SORTED, ERROR) \
  60        FUNC(UNKNOWN_TYPE, ERROR) \
  61        FUNC(ZERO_PADDED_DATE, ERROR) \
  62        FUNC(GITMODULES_MISSING, ERROR) \
  63        FUNC(GITMODULES_BLOB, ERROR) \
  64        FUNC(GITMODULES_PARSE, ERROR) \
  65        FUNC(GITMODULES_NAME, ERROR) \
  66        FUNC(GITMODULES_SYMLINK, ERROR) \
  67        /* warnings */ \
  68        FUNC(BAD_FILEMODE, WARN) \
  69        FUNC(EMPTY_NAME, WARN) \
  70        FUNC(FULL_PATHNAME, WARN) \
  71        FUNC(HAS_DOT, WARN) \
  72        FUNC(HAS_DOTDOT, WARN) \
  73        FUNC(HAS_DOTGIT, WARN) \
  74        FUNC(NULL_SHA1, WARN) \
  75        FUNC(ZERO_PADDED_FILEMODE, WARN) \
  76        FUNC(NUL_IN_COMMIT, WARN) \
  77        /* infos (reported as warnings, but ignored by default) */ \
  78        FUNC(BAD_TAG_NAME, INFO) \
  79        FUNC(MISSING_TAGGER_ENTRY, INFO)
  80
  81#define MSG_ID(id, msg_type) FSCK_MSG_##id,
  82enum fsck_msg_id {
  83        FOREACH_MSG_ID(MSG_ID)
  84        FSCK_MSG_MAX
  85};
  86#undef MSG_ID
  87
  88#define STR(x) #x
  89#define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
  90static struct {
  91        const char *id_string;
  92        const char *downcased;
  93        int msg_type;
  94} msg_id_info[FSCK_MSG_MAX + 1] = {
  95        FOREACH_MSG_ID(MSG_ID)
  96        { NULL, NULL, -1 }
  97};
  98#undef MSG_ID
  99
 100static int parse_msg_id(const char *text)
 101{
 102        int i;
 103
 104        if (!msg_id_info[0].downcased) {
 105                /* convert id_string to lower case, without underscores. */
 106                for (i = 0; i < FSCK_MSG_MAX; i++) {
 107                        const char *p = msg_id_info[i].id_string;
 108                        int len = strlen(p);
 109                        char *q = xmalloc(len);
 110
 111                        msg_id_info[i].downcased = q;
 112                        while (*p)
 113                                if (*p == '_')
 114                                        p++;
 115                                else
 116                                        *(q)++ = tolower(*(p)++);
 117                        *q = '\0';
 118                }
 119        }
 120
 121        for (i = 0; i < FSCK_MSG_MAX; i++)
 122                if (!strcmp(text, msg_id_info[i].downcased))
 123                        return i;
 124
 125        return -1;
 126}
 127
 128static int fsck_msg_type(enum fsck_msg_id msg_id,
 129        struct fsck_options *options)
 130{
 131        int msg_type;
 132
 133        assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 134
 135        if (options->msg_type)
 136                msg_type = options->msg_type[msg_id];
 137        else {
 138                msg_type = msg_id_info[msg_id].msg_type;
 139                if (options->strict && msg_type == FSCK_WARN)
 140                        msg_type = FSCK_ERROR;
 141        }
 142
 143        return msg_type;
 144}
 145
 146static void init_skiplist(struct fsck_options *options, const char *path)
 147{
 148        static struct oid_array skiplist = OID_ARRAY_INIT;
 149        int sorted, fd;
 150        char buffer[GIT_MAX_HEXSZ + 1];
 151        struct object_id oid;
 152
 153        if (options->skiplist)
 154                sorted = options->skiplist->sorted;
 155        else {
 156                sorted = 1;
 157                options->skiplist = &skiplist;
 158        }
 159
 160        fd = open(path, O_RDONLY);
 161        if (fd < 0)
 162                die("Could not open skip list: %s", path);
 163        for (;;) {
 164                const char *p;
 165                int result = read_in_full(fd, buffer, sizeof(buffer));
 166                if (result < 0)
 167                        die_errno("Could not read '%s'", path);
 168                if (!result)
 169                        break;
 170                if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 171                        die("Invalid SHA-1: %s", buffer);
 172                oid_array_append(&skiplist, &oid);
 173                if (sorted && skiplist.nr > 1 &&
 174                                oidcmp(&skiplist.oid[skiplist.nr - 2],
 175                                       &oid) > 0)
 176                        sorted = 0;
 177        }
 178        close(fd);
 179
 180        if (sorted)
 181                skiplist.sorted = 1;
 182}
 183
 184static int parse_msg_type(const char *str)
 185{
 186        if (!strcmp(str, "error"))
 187                return FSCK_ERROR;
 188        else if (!strcmp(str, "warn"))
 189                return FSCK_WARN;
 190        else if (!strcmp(str, "ignore"))
 191                return FSCK_IGNORE;
 192        else
 193                die("Unknown fsck message type: '%s'", str);
 194}
 195
 196int is_valid_msg_type(const char *msg_id, const char *msg_type)
 197{
 198        if (parse_msg_id(msg_id) < 0)
 199                return 0;
 200        parse_msg_type(msg_type);
 201        return 1;
 202}
 203
 204void fsck_set_msg_type(struct fsck_options *options,
 205                const char *msg_id, const char *msg_type)
 206{
 207        int id = parse_msg_id(msg_id), type;
 208
 209        if (id < 0)
 210                die("Unhandled message id: %s", msg_id);
 211        type = parse_msg_type(msg_type);
 212
 213        if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 214                die("Cannot demote %s to %s", msg_id, msg_type);
 215
 216        if (!options->msg_type) {
 217                int i;
 218                int *msg_type;
 219                ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 220                for (i = 0; i < FSCK_MSG_MAX; i++)
 221                        msg_type[i] = fsck_msg_type(i, options);
 222                options->msg_type = msg_type;
 223        }
 224
 225        options->msg_type[id] = type;
 226}
 227
 228void fsck_set_msg_types(struct fsck_options *options, const char *values)
 229{
 230        char *buf = xstrdup(values), *to_free = buf;
 231        int done = 0;
 232
 233        while (!done) {
 234                int len = strcspn(buf, " ,|"), equal;
 235
 236                done = !buf[len];
 237                if (!len) {
 238                        buf++;
 239                        continue;
 240                }
 241                buf[len] = '\0';
 242
 243                for (equal = 0;
 244                     equal < len && buf[equal] != '=' && buf[equal] != ':';
 245                     equal++)
 246                        buf[equal] = tolower(buf[equal]);
 247                buf[equal] = '\0';
 248
 249                if (!strcmp(buf, "skiplist")) {
 250                        if (equal == len)
 251                                die("skiplist requires a path");
 252                        init_skiplist(options, buf + equal + 1);
 253                        buf += len + 1;
 254                        continue;
 255                }
 256
 257                if (equal == len)
 258                        die("Missing '=': '%s'", buf);
 259
 260                fsck_set_msg_type(options, buf, buf + equal + 1);
 261                buf += len + 1;
 262        }
 263        free(to_free);
 264}
 265
 266static void append_msg_id(struct strbuf *sb, const char *msg_id)
 267{
 268        for (;;) {
 269                char c = *(msg_id)++;
 270
 271                if (!c)
 272                        break;
 273                if (c != '_')
 274                        strbuf_addch(sb, tolower(c));
 275                else {
 276                        assert(*msg_id);
 277                        strbuf_addch(sb, *(msg_id)++);
 278                }
 279        }
 280
 281        strbuf_addstr(sb, ": ");
 282}
 283
 284static int object_on_skiplist(struct fsck_options *opts, struct object *obj)
 285{
 286        if (opts && opts->skiplist && obj)
 287                return oid_array_lookup(opts->skiplist, &obj->oid) >= 0;
 288        return 0;
 289}
 290
 291__attribute__((format (printf, 4, 5)))
 292static int report(struct fsck_options *options, struct object *object,
 293        enum fsck_msg_id id, const char *fmt, ...)
 294{
 295        va_list ap;
 296        struct strbuf sb = STRBUF_INIT;
 297        int msg_type = fsck_msg_type(id, options), result;
 298
 299        if (msg_type == FSCK_IGNORE)
 300                return 0;
 301
 302        if (object_on_skiplist(options, object))
 303                return 0;
 304
 305        if (msg_type == FSCK_FATAL)
 306                msg_type = FSCK_ERROR;
 307        else if (msg_type == FSCK_INFO)
 308                msg_type = FSCK_WARN;
 309
 310        append_msg_id(&sb, msg_id_info[id].id_string);
 311
 312        va_start(ap, fmt);
 313        strbuf_vaddf(&sb, fmt, ap);
 314        result = options->error_func(options, object, msg_type, sb.buf);
 315        strbuf_release(&sb);
 316        va_end(ap);
 317
 318        return result;
 319}
 320
 321static char *get_object_name(struct fsck_options *options, struct object *obj)
 322{
 323        if (!options->object_names)
 324                return NULL;
 325        return lookup_decoration(options->object_names, obj);
 326}
 327
 328static void put_object_name(struct fsck_options *options, struct object *obj,
 329        const char *fmt, ...)
 330{
 331        va_list ap;
 332        struct strbuf buf = STRBUF_INIT;
 333        char *existing;
 334
 335        if (!options->object_names)
 336                return;
 337        existing = lookup_decoration(options->object_names, obj);
 338        if (existing)
 339                return;
 340        va_start(ap, fmt);
 341        strbuf_vaddf(&buf, fmt, ap);
 342        add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 343        va_end(ap);
 344}
 345
 346static const char *describe_object(struct fsck_options *o, struct object *obj)
 347{
 348        static struct strbuf buf = STRBUF_INIT;
 349        char *name;
 350
 351        strbuf_reset(&buf);
 352        strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 353        if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 354                strbuf_addf(&buf, " (%s)", name);
 355
 356        return buf.buf;
 357}
 358
 359static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 360{
 361        struct tree_desc desc;
 362        struct name_entry entry;
 363        int res = 0;
 364        const char *name;
 365
 366        if (parse_tree(tree))
 367                return -1;
 368
 369        name = get_object_name(options, &tree->object);
 370        if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 371                return -1;
 372        while (tree_entry_gently(&desc, &entry)) {
 373                struct object *obj;
 374                int result;
 375
 376                if (S_ISGITLINK(entry.mode))
 377                        continue;
 378
 379                if (S_ISDIR(entry.mode)) {
 380                        obj = (struct object *)lookup_tree(entry.oid);
 381                        if (name && obj)
 382                                put_object_name(options, obj, "%s%s/", name,
 383                                        entry.path);
 384                        result = options->walk(obj, OBJ_TREE, data, options);
 385                }
 386                else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 387                        obj = (struct object *)lookup_blob(entry.oid);
 388                        if (name && obj)
 389                                put_object_name(options, obj, "%s%s", name,
 390                                        entry.path);
 391                        result = options->walk(obj, OBJ_BLOB, data, options);
 392                }
 393                else {
 394                        result = error("in tree %s: entry %s has bad mode %.6o",
 395                                        describe_object(options, &tree->object), entry.path, entry.mode);
 396                }
 397                if (result < 0)
 398                        return result;
 399                if (!res)
 400                        res = result;
 401        }
 402        return res;
 403}
 404
 405static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 406{
 407        int counter = 0, generation = 0, name_prefix_len = 0;
 408        struct commit_list *parents;
 409        int res;
 410        int result;
 411        const char *name;
 412
 413        if (parse_commit(commit))
 414                return -1;
 415
 416        name = get_object_name(options, &commit->object);
 417        if (name)
 418                put_object_name(options, &commit->tree->object, "%s:", name);
 419
 420        result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options);
 421        if (result < 0)
 422                return result;
 423        res = result;
 424
 425        parents = commit->parents;
 426        if (name && parents) {
 427                int len = strlen(name), power;
 428
 429                if (len && name[len - 1] == '^') {
 430                        generation = 1;
 431                        name_prefix_len = len - 1;
 432                }
 433                else { /* parse ~<generation> suffix */
 434                        for (generation = 0, power = 1;
 435                             len && isdigit(name[len - 1]);
 436                             power *= 10)
 437                                generation += power * (name[--len] - '0');
 438                        if (power > 1 && len && name[len - 1] == '~')
 439                                name_prefix_len = len - 1;
 440                }
 441        }
 442
 443        while (parents) {
 444                if (name) {
 445                        struct object *obj = &parents->item->object;
 446
 447                        if (++counter > 1)
 448                                put_object_name(options, obj, "%s^%d",
 449                                        name, counter);
 450                        else if (generation > 0)
 451                                put_object_name(options, obj, "%.*s~%d",
 452                                        name_prefix_len, name, generation + 1);
 453                        else
 454                                put_object_name(options, obj, "%s^", name);
 455                }
 456                result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 457                if (result < 0)
 458                        return result;
 459                if (!res)
 460                        res = result;
 461                parents = parents->next;
 462        }
 463        return res;
 464}
 465
 466static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 467{
 468        char *name = get_object_name(options, &tag->object);
 469
 470        if (parse_tag(tag))
 471                return -1;
 472        if (name)
 473                put_object_name(options, tag->tagged, "%s", name);
 474        return options->walk(tag->tagged, OBJ_ANY, data, options);
 475}
 476
 477int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 478{
 479        if (!obj)
 480                return -1;
 481
 482        if (obj->type == OBJ_NONE)
 483                parse_object(&obj->oid);
 484
 485        switch (obj->type) {
 486        case OBJ_BLOB:
 487                return 0;
 488        case OBJ_TREE:
 489                return fsck_walk_tree((struct tree *)obj, data, options);
 490        case OBJ_COMMIT:
 491                return fsck_walk_commit((struct commit *)obj, data, options);
 492        case OBJ_TAG:
 493                return fsck_walk_tag((struct tag *)obj, data, options);
 494        default:
 495                error("Unknown object type for %s", describe_object(options, obj));
 496                return -1;
 497        }
 498}
 499
 500/*
 501 * The entries in a tree are ordered in the _path_ order,
 502 * which means that a directory entry is ordered by adding
 503 * a slash to the end of it.
 504 *
 505 * So a directory called "a" is ordered _after_ a file
 506 * called "a.c", because "a/" sorts after "a.c".
 507 */
 508#define TREE_UNORDERED (-1)
 509#define TREE_HAS_DUPS  (-2)
 510
 511static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 512{
 513        int len1 = strlen(name1);
 514        int len2 = strlen(name2);
 515        int len = len1 < len2 ? len1 : len2;
 516        unsigned char c1, c2;
 517        int cmp;
 518
 519        cmp = memcmp(name1, name2, len);
 520        if (cmp < 0)
 521                return 0;
 522        if (cmp > 0)
 523                return TREE_UNORDERED;
 524
 525        /*
 526         * Ok, the first <len> characters are the same.
 527         * Now we need to order the next one, but turn
 528         * a '\0' into a '/' for a directory entry.
 529         */
 530        c1 = name1[len];
 531        c2 = name2[len];
 532        if (!c1 && !c2)
 533                /*
 534                 * git-write-tree used to write out a nonsense tree that has
 535                 * entries with the same name, one blob and one tree.  Make
 536                 * sure we do not have duplicate entries.
 537                 */
 538                return TREE_HAS_DUPS;
 539        if (!c1 && S_ISDIR(mode1))
 540                c1 = '/';
 541        if (!c2 && S_ISDIR(mode2))
 542                c2 = '/';
 543        return c1 < c2 ? 0 : TREE_UNORDERED;
 544}
 545
 546static int fsck_tree(struct tree *item, struct fsck_options *options)
 547{
 548        int retval = 0;
 549        int has_null_sha1 = 0;
 550        int has_full_path = 0;
 551        int has_empty_name = 0;
 552        int has_dot = 0;
 553        int has_dotdot = 0;
 554        int has_dotgit = 0;
 555        int has_zero_pad = 0;
 556        int has_bad_modes = 0;
 557        int has_dup_entries = 0;
 558        int not_properly_sorted = 0;
 559        struct tree_desc desc;
 560        unsigned o_mode;
 561        const char *o_name;
 562
 563        if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 564                retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 565                return retval;
 566        }
 567
 568        o_mode = 0;
 569        o_name = NULL;
 570
 571        while (desc.size) {
 572                unsigned mode;
 573                const char *name;
 574                const struct object_id *oid;
 575
 576                oid = tree_entry_extract(&desc, &name, &mode);
 577
 578                has_null_sha1 |= is_null_oid(oid);
 579                has_full_path |= !!strchr(name, '/');
 580                has_empty_name |= !*name;
 581                has_dot |= !strcmp(name, ".");
 582                has_dotdot |= !strcmp(name, "..");
 583                has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 584                has_zero_pad |= *(char *)desc.buffer == '0';
 585
 586                if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 587                        if (!S_ISLNK(mode))
 588                                oidset_insert(&gitmodules_found, oid);
 589                        else
 590                                retval += report(options, &item->object,
 591                                                 FSCK_MSG_GITMODULES_SYMLINK,
 592                                                 ".gitmodules is a symbolic link");
 593                }
 594
 595                if (update_tree_entry_gently(&desc)) {
 596                        retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 597                        break;
 598                }
 599
 600                switch (mode) {
 601                /*
 602                 * Standard modes..
 603                 */
 604                case S_IFREG | 0755:
 605                case S_IFREG | 0644:
 606                case S_IFLNK:
 607                case S_IFDIR:
 608                case S_IFGITLINK:
 609                        break;
 610                /*
 611                 * This is nonstandard, but we had a few of these
 612                 * early on when we honored the full set of mode
 613                 * bits..
 614                 */
 615                case S_IFREG | 0664:
 616                        if (!options->strict)
 617                                break;
 618                        /* fallthrough */
 619                default:
 620                        has_bad_modes = 1;
 621                }
 622
 623                if (o_name) {
 624                        switch (verify_ordered(o_mode, o_name, mode, name)) {
 625                        case TREE_UNORDERED:
 626                                not_properly_sorted = 1;
 627                                break;
 628                        case TREE_HAS_DUPS:
 629                                has_dup_entries = 1;
 630                                break;
 631                        default:
 632                                break;
 633                        }
 634                }
 635
 636                o_mode = mode;
 637                o_name = name;
 638        }
 639
 640        if (has_null_sha1)
 641                retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 642        if (has_full_path)
 643                retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 644        if (has_empty_name)
 645                retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 646        if (has_dot)
 647                retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 648        if (has_dotdot)
 649                retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 650        if (has_dotgit)
 651                retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 652        if (has_zero_pad)
 653                retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 654        if (has_bad_modes)
 655                retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 656        if (has_dup_entries)
 657                retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 658        if (not_properly_sorted)
 659                retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 660        return retval;
 661}
 662
 663static int verify_headers(const void *data, unsigned long size,
 664                          struct object *obj, struct fsck_options *options)
 665{
 666        const char *buffer = (const char *)data;
 667        unsigned long i;
 668
 669        for (i = 0; i < size; i++) {
 670                switch (buffer[i]) {
 671                case '\0':
 672                        return report(options, obj,
 673                                FSCK_MSG_NUL_IN_HEADER,
 674                                "unterminated header: NUL at offset %ld", i);
 675                case '\n':
 676                        if (i + 1 < size && buffer[i + 1] == '\n')
 677                                return 0;
 678                }
 679        }
 680
 681        /*
 682         * We did not find double-LF that separates the header
 683         * and the body.  Not having a body is not a crime but
 684         * we do want to see the terminating LF for the last header
 685         * line.
 686         */
 687        if (size && buffer[size - 1] == '\n')
 688                return 0;
 689
 690        return report(options, obj,
 691                FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 692}
 693
 694static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 695{
 696        const char *p = *ident;
 697        char *end;
 698
 699        *ident = strchrnul(*ident, '\n');
 700        if (**ident == '\n')
 701                (*ident)++;
 702
 703        if (*p == '<')
 704                return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 705        p += strcspn(p, "<>\n");
 706        if (*p == '>')
 707                return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 708        if (*p != '<')
 709                return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 710        if (p[-1] != ' ')
 711                return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 712        p++;
 713        p += strcspn(p, "<>\n");
 714        if (*p != '>')
 715                return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 716        p++;
 717        if (*p != ' ')
 718                return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 719        p++;
 720        if (*p == '0' && p[1] != ' ')
 721                return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 722        if (date_overflows(parse_timestamp(p, &end, 10)))
 723                return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 724        if ((end == p || *end != ' '))
 725                return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 726        p = end + 1;
 727        if ((*p != '+' && *p != '-') ||
 728            !isdigit(p[1]) ||
 729            !isdigit(p[2]) ||
 730            !isdigit(p[3]) ||
 731            !isdigit(p[4]) ||
 732            (p[5] != '\n'))
 733                return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 734        p += 6;
 735        return 0;
 736}
 737
 738static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 739        unsigned long size, struct fsck_options *options)
 740{
 741        unsigned char tree_sha1[20], sha1[20];
 742        struct commit_graft *graft;
 743        unsigned parent_count, parent_line_count = 0, author_count;
 744        int err;
 745        const char *buffer_begin = buffer;
 746
 747        if (verify_headers(buffer, size, &commit->object, options))
 748                return -1;
 749
 750        if (!skip_prefix(buffer, "tree ", &buffer))
 751                return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 752        if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') {
 753                err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 754                if (err)
 755                        return err;
 756        }
 757        buffer += 41;
 758        while (skip_prefix(buffer, "parent ", &buffer)) {
 759                if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 760                        err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 761                        if (err)
 762                                return err;
 763                }
 764                buffer += 41;
 765                parent_line_count++;
 766        }
 767        graft = lookup_commit_graft(&commit->object.oid);
 768        parent_count = commit_list_count(commit->parents);
 769        if (graft) {
 770                if (graft->nr_parent == -1 && !parent_count)
 771                        ; /* shallow commit */
 772                else if (graft->nr_parent != parent_count) {
 773                        err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 774                        if (err)
 775                                return err;
 776                }
 777        } else {
 778                if (parent_count != parent_line_count) {
 779                        err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 780                        if (err)
 781                                return err;
 782                }
 783        }
 784        author_count = 0;
 785        while (skip_prefix(buffer, "author ", &buffer)) {
 786                author_count++;
 787                err = fsck_ident(&buffer, &commit->object, options);
 788                if (err)
 789                        return err;
 790        }
 791        if (author_count < 1)
 792                err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 793        else if (author_count > 1)
 794                err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 795        if (err)
 796                return err;
 797        if (!skip_prefix(buffer, "committer ", &buffer))
 798                return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 799        err = fsck_ident(&buffer, &commit->object, options);
 800        if (err)
 801                return err;
 802        if (!commit->tree) {
 803                err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
 804                if (err)
 805                        return err;
 806        }
 807        if (memchr(buffer_begin, '\0', size)) {
 808                err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 809                             "NUL byte in the commit object body");
 810                if (err)
 811                        return err;
 812        }
 813        return 0;
 814}
 815
 816static int fsck_commit(struct commit *commit, const char *data,
 817        unsigned long size, struct fsck_options *options)
 818{
 819        const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 820        int ret = fsck_commit_buffer(commit, buffer, size, options);
 821        if (!data)
 822                unuse_commit_buffer(commit, buffer);
 823        return ret;
 824}
 825
 826static int fsck_tag_buffer(struct tag *tag, const char *data,
 827        unsigned long size, struct fsck_options *options)
 828{
 829        unsigned char sha1[20];
 830        int ret = 0;
 831        const char *buffer;
 832        char *to_free = NULL, *eol;
 833        struct strbuf sb = STRBUF_INIT;
 834
 835        if (data)
 836                buffer = data;
 837        else {
 838                enum object_type type;
 839
 840                buffer = to_free =
 841                        read_sha1_file(tag->object.oid.hash, &type, &size);
 842                if (!buffer)
 843                        return report(options, &tag->object,
 844                                FSCK_MSG_MISSING_TAG_OBJECT,
 845                                "cannot read tag object");
 846
 847                if (type != OBJ_TAG) {
 848                        ret = report(options, &tag->object,
 849                                FSCK_MSG_TAG_OBJECT_NOT_TAG,
 850                                "expected tag got %s",
 851                            type_name(type));
 852                        goto done;
 853                }
 854        }
 855
 856        ret = verify_headers(buffer, size, &tag->object, options);
 857        if (ret)
 858                goto done;
 859
 860        if (!skip_prefix(buffer, "object ", &buffer)) {
 861                ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 862                goto done;
 863        }
 864        if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 865                ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 866                if (ret)
 867                        goto done;
 868        }
 869        buffer += 41;
 870
 871        if (!skip_prefix(buffer, "type ", &buffer)) {
 872                ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 873                goto done;
 874        }
 875        eol = strchr(buffer, '\n');
 876        if (!eol) {
 877                ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 878                goto done;
 879        }
 880        if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 881                ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 882        if (ret)
 883                goto done;
 884        buffer = eol + 1;
 885
 886        if (!skip_prefix(buffer, "tag ", &buffer)) {
 887                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 888                goto done;
 889        }
 890        eol = strchr(buffer, '\n');
 891        if (!eol) {
 892                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 893                goto done;
 894        }
 895        strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 896        if (check_refname_format(sb.buf, 0)) {
 897                ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 898                           "invalid 'tag' name: %.*s",
 899                           (int)(eol - buffer), buffer);
 900                if (ret)
 901                        goto done;
 902        }
 903        buffer = eol + 1;
 904
 905        if (!skip_prefix(buffer, "tagger ", &buffer)) {
 906                /* early tags do not contain 'tagger' lines; warn only */
 907                ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 908                if (ret)
 909                        goto done;
 910        }
 911        else
 912                ret = fsck_ident(&buffer, &tag->object, options);
 913
 914done:
 915        strbuf_release(&sb);
 916        free(to_free);
 917        return ret;
 918}
 919
 920static int fsck_tag(struct tag *tag, const char *data,
 921        unsigned long size, struct fsck_options *options)
 922{
 923        struct object *tagged = tag->tagged;
 924
 925        if (!tagged)
 926                return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 927
 928        return fsck_tag_buffer(tag, data, size, options);
 929}
 930
 931struct fsck_gitmodules_data {
 932        struct object *obj;
 933        struct fsck_options *options;
 934        int ret;
 935};
 936
 937static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
 938{
 939        struct fsck_gitmodules_data *data = vdata;
 940        const char *subsection, *key;
 941        int subsection_len;
 942        char *name;
 943
 944        if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
 945            !subsection)
 946                return 0;
 947
 948        name = xmemdupz(subsection, subsection_len);
 949        if (check_submodule_name(name) < 0)
 950                data->ret |= report(data->options, data->obj,
 951                                    FSCK_MSG_GITMODULES_NAME,
 952                                    "disallowed submodule name: %s",
 953                                    name);
 954        free(name);
 955
 956        return 0;
 957}
 958
 959static int fsck_blob(struct blob *blob, const char *buf,
 960                     unsigned long size, struct fsck_options *options)
 961{
 962        struct fsck_gitmodules_data data;
 963
 964        if (!oidset_contains(&gitmodules_found, &blob->object.oid))
 965                return 0;
 966        oidset_insert(&gitmodules_done, &blob->object.oid);
 967
 968        if (object_on_skiplist(options, &blob->object))
 969                return 0;
 970
 971        if (!buf) {
 972                /*
 973                 * A missing buffer here is a sign that the caller found the
 974                 * blob too gigantic to load into memory. Let's just consider
 975                 * that an error.
 976                 */
 977                return report(options, &blob->object,
 978                              FSCK_MSG_GITMODULES_PARSE,
 979                              ".gitmodules too large to parse");
 980        }
 981
 982        data.obj = &blob->object;
 983        data.options = options;
 984        data.ret = 0;
 985        if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
 986                                ".gitmodules", buf, size, &data))
 987                data.ret |= report(options, &blob->object,
 988                                   FSCK_MSG_GITMODULES_PARSE,
 989                                   "could not parse gitmodules blob");
 990
 991        return data.ret;
 992}
 993
 994int fsck_object(struct object *obj, void *data, unsigned long size,
 995        struct fsck_options *options)
 996{
 997        if (!obj)
 998                return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
 999
1000        if (obj->type == OBJ_BLOB)
1001                return fsck_blob((struct blob *)obj, data, size, options);
1002        if (obj->type == OBJ_TREE)
1003                return fsck_tree((struct tree *) obj, options);
1004        if (obj->type == OBJ_COMMIT)
1005                return fsck_commit((struct commit *) obj, (const char *) data,
1006                        size, options);
1007        if (obj->type == OBJ_TAG)
1008                return fsck_tag((struct tag *) obj, (const char *) data,
1009                        size, options);
1010
1011        return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1012                          obj->type);
1013}
1014
1015int fsck_error_function(struct fsck_options *o,
1016        struct object *obj, int msg_type, const char *message)
1017{
1018        if (msg_type == FSCK_WARN) {
1019                warning("object %s: %s", describe_object(o, obj), message);
1020                return 0;
1021        }
1022        error("object %s: %s", describe_object(o, obj), message);
1023        return 1;
1024}
1025
1026int fsck_finish(struct fsck_options *options)
1027{
1028        int ret = 0;
1029        struct oidset_iter iter;
1030        const struct object_id *oid;
1031
1032        oidset_iter_init(&gitmodules_found, &iter);
1033        while ((oid = oidset_iter_next(&iter))) {
1034                struct blob *blob;
1035                enum object_type type;
1036                unsigned long size;
1037                char *buf;
1038
1039                if (oidset_contains(&gitmodules_done, oid))
1040                        continue;
1041
1042                blob = lookup_blob(oid);
1043                if (!blob) {
1044                        struct object *obj = lookup_unknown_object(oid->hash);
1045                        ret |= report(options, obj,
1046                                      FSCK_MSG_GITMODULES_BLOB,
1047                                      "non-blob found at .gitmodules");
1048                        continue;
1049                }
1050
1051                buf = read_sha1_file(oid->hash, &type, &size);
1052                if (!buf) {
1053                        if (is_promisor_object(&blob->object.oid))
1054                                continue;
1055                        ret |= report(options, &blob->object,
1056                                      FSCK_MSG_GITMODULES_MISSING,
1057                                      "unable to read .gitmodules blob");
1058                        continue;
1059                }
1060
1061                if (type == OBJ_BLOB)
1062                        ret |= fsck_blob(blob, buf, size, options);
1063                else
1064                        ret |= report(options, &blob->object,
1065                                      FSCK_MSG_GITMODULES_BLOB,
1066                                      "non-blob found at .gitmodules");
1067                free(buf);
1068        }
1069
1070
1071        oidset_clear(&gitmodules_found);
1072        oidset_clear(&gitmodules_done);
1073        return ret;
1074}