dir.con commit Merge branch 'maint' (e7e5548)
   1/*
   2 * This handles recursive filename detection with exclude
   3 * files, index knowledge etc..
   4 *
   5 * Copyright (C) Linus Torvalds, 2005-2006
   6 *               Junio Hamano, 2005-2006
   7 */
   8#include "cache.h"
   9#include "dir.h"
  10#include "refs.h"
  11
  12struct path_simplify {
  13        int len;
  14        const char *path;
  15};
  16
  17static int read_directory_recursive(struct dir_struct *dir, const char *path, int len,
  18        int check_only, const struct path_simplify *simplify);
  19static int get_dtype(struct dirent *de, const char *path, int len);
  20
  21static int common_prefix(const char **pathspec)
  22{
  23        const char *path, *slash, *next;
  24        int prefix;
  25
  26        if (!pathspec)
  27                return 0;
  28
  29        path = *pathspec;
  30        slash = strrchr(path, '/');
  31        if (!slash)
  32                return 0;
  33
  34        prefix = slash - path + 1;
  35        while ((next = *++pathspec) != NULL) {
  36                int len = strlen(next);
  37                if (len >= prefix && !memcmp(path, next, prefix))
  38                        continue;
  39                len = prefix - 1;
  40                for (;;) {
  41                        if (!len)
  42                                return 0;
  43                        if (next[--len] != '/')
  44                                continue;
  45                        if (memcmp(path, next, len+1))
  46                                continue;
  47                        prefix = len + 1;
  48                        break;
  49                }
  50        }
  51        return prefix;
  52}
  53
  54int fill_directory(struct dir_struct *dir, const char **pathspec)
  55{
  56        const char *path;
  57        int len;
  58
  59        /*
  60         * Calculate common prefix for the pathspec, and
  61         * use that to optimize the directory walk
  62         */
  63        len = common_prefix(pathspec);
  64        path = "";
  65
  66        if (len)
  67                path = xmemdupz(*pathspec, len);
  68
  69        /* Read the directory and prune it */
  70        read_directory(dir, path, len, pathspec);
  71        return len;
  72}
  73
  74/*
  75 * Does 'match' match the given name?
  76 * A match is found if
  77 *
  78 * (1) the 'match' string is leading directory of 'name', or
  79 * (2) the 'match' string is a wildcard and matches 'name', or
  80 * (3) the 'match' string is exactly the same as 'name'.
  81 *
  82 * and the return value tells which case it was.
  83 *
  84 * It returns 0 when there is no match.
  85 */
  86static int match_one(const char *match, const char *name, int namelen)
  87{
  88        int matchlen;
  89
  90        /* If the match was just the prefix, we matched */
  91        if (!*match)
  92                return MATCHED_RECURSIVELY;
  93
  94        for (;;) {
  95                unsigned char c1 = *match;
  96                unsigned char c2 = *name;
  97                if (c1 == '\0' || is_glob_special(c1))
  98                        break;
  99                if (c1 != c2)
 100                        return 0;
 101                match++;
 102                name++;
 103                namelen--;
 104        }
 105
 106
 107        /*
 108         * If we don't match the matchstring exactly,
 109         * we need to match by fnmatch
 110         */
 111        matchlen = strlen(match);
 112        if (strncmp(match, name, matchlen))
 113                return !fnmatch(match, name, 0) ? MATCHED_FNMATCH : 0;
 114
 115        if (namelen == matchlen)
 116                return MATCHED_EXACTLY;
 117        if (match[matchlen-1] == '/' || name[matchlen] == '/')
 118                return MATCHED_RECURSIVELY;
 119        return 0;
 120}
 121
 122/*
 123 * Given a name and a list of pathspecs, see if the name matches
 124 * any of the pathspecs.  The caller is also interested in seeing
 125 * all pathspec matches some names it calls this function with
 126 * (otherwise the user could have mistyped the unmatched pathspec),
 127 * and a mark is left in seen[] array for pathspec element that
 128 * actually matched anything.
 129 */
 130int match_pathspec(const char **pathspec, const char *name, int namelen,
 131                int prefix, char *seen)
 132{
 133        int i, retval = 0;
 134
 135        if (!pathspec)
 136                return 1;
 137
 138        name += prefix;
 139        namelen -= prefix;
 140
 141        for (i = 0; pathspec[i] != NULL; i++) {
 142                int how;
 143                const char *match = pathspec[i] + prefix;
 144                if (seen && seen[i] == MATCHED_EXACTLY)
 145                        continue;
 146                how = match_one(match, name, namelen);
 147                if (how) {
 148                        if (retval < how)
 149                                retval = how;
 150                        if (seen && seen[i] < how)
 151                                seen[i] = how;
 152                }
 153        }
 154        return retval;
 155}
 156
 157static int no_wildcard(const char *string)
 158{
 159        return string[strcspn(string, "*?[{\\")] == '\0';
 160}
 161
 162void add_exclude(const char *string, const char *base,
 163                 int baselen, struct exclude_list *which)
 164{
 165        struct exclude *x;
 166        size_t len;
 167        int to_exclude = 1;
 168        int flags = 0;
 169
 170        if (*string == '!') {
 171                to_exclude = 0;
 172                string++;
 173        }
 174        len = strlen(string);
 175        if (len && string[len - 1] == '/') {
 176                char *s;
 177                x = xmalloc(sizeof(*x) + len);
 178                s = (char *)(x+1);
 179                memcpy(s, string, len - 1);
 180                s[len - 1] = '\0';
 181                string = s;
 182                x->pattern = s;
 183                flags = EXC_FLAG_MUSTBEDIR;
 184        } else {
 185                x = xmalloc(sizeof(*x));
 186                x->pattern = string;
 187        }
 188        x->to_exclude = to_exclude;
 189        x->patternlen = strlen(string);
 190        x->base = base;
 191        x->baselen = baselen;
 192        x->flags = flags;
 193        if (!strchr(string, '/'))
 194                x->flags |= EXC_FLAG_NODIR;
 195        if (no_wildcard(string))
 196                x->flags |= EXC_FLAG_NOWILDCARD;
 197        if (*string == '*' && no_wildcard(string+1))
 198                x->flags |= EXC_FLAG_ENDSWITH;
 199        ALLOC_GROW(which->excludes, which->nr + 1, which->alloc);
 200        which->excludes[which->nr++] = x;
 201}
 202
 203static int add_excludes_from_file_1(const char *fname,
 204                                    const char *base,
 205                                    int baselen,
 206                                    char **buf_p,
 207                                    struct exclude_list *which)
 208{
 209        struct stat st;
 210        int fd, i;
 211        size_t size;
 212        char *buf, *entry;
 213
 214        fd = open(fname, O_RDONLY);
 215        if (fd < 0 || fstat(fd, &st) < 0)
 216                goto err;
 217        size = xsize_t(st.st_size);
 218        if (size == 0) {
 219                close(fd);
 220                return 0;
 221        }
 222        buf = xmalloc(size+1);
 223        if (read_in_full(fd, buf, size) != size)
 224        {
 225                free(buf);
 226                goto err;
 227        }
 228        close(fd);
 229
 230        if (buf_p)
 231                *buf_p = buf;
 232        buf[size++] = '\n';
 233        entry = buf;
 234        for (i = 0; i < size; i++) {
 235                if (buf[i] == '\n') {
 236                        if (entry != buf + i && entry[0] != '#') {
 237                                buf[i - (i && buf[i-1] == '\r')] = 0;
 238                                add_exclude(entry, base, baselen, which);
 239                        }
 240                        entry = buf + i + 1;
 241                }
 242        }
 243        return 0;
 244
 245 err:
 246        if (0 <= fd)
 247                close(fd);
 248        return -1;
 249}
 250
 251void add_excludes_from_file(struct dir_struct *dir, const char *fname)
 252{
 253        if (add_excludes_from_file_1(fname, "", 0, NULL,
 254                                     &dir->exclude_list[EXC_FILE]) < 0)
 255                die("cannot use %s as an exclude file", fname);
 256}
 257
 258static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
 259{
 260        struct exclude_list *el;
 261        struct exclude_stack *stk = NULL;
 262        int current;
 263
 264        if ((!dir->exclude_per_dir) ||
 265            (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
 266                return; /* too long a path -- ignore */
 267
 268        /* Pop the ones that are not the prefix of the path being checked. */
 269        el = &dir->exclude_list[EXC_DIRS];
 270        while ((stk = dir->exclude_stack) != NULL) {
 271                if (stk->baselen <= baselen &&
 272                    !strncmp(dir->basebuf, base, stk->baselen))
 273                        break;
 274                dir->exclude_stack = stk->prev;
 275                while (stk->exclude_ix < el->nr)
 276                        free(el->excludes[--el->nr]);
 277                free(stk->filebuf);
 278                free(stk);
 279        }
 280
 281        /* Read from the parent directories and push them down. */
 282        current = stk ? stk->baselen : -1;
 283        while (current < baselen) {
 284                struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
 285                const char *cp;
 286
 287                if (current < 0) {
 288                        cp = base;
 289                        current = 0;
 290                }
 291                else {
 292                        cp = strchr(base + current + 1, '/');
 293                        if (!cp)
 294                                die("oops in prep_exclude");
 295                        cp++;
 296                }
 297                stk->prev = dir->exclude_stack;
 298                stk->baselen = cp - base;
 299                stk->exclude_ix = el->nr;
 300                memcpy(dir->basebuf + current, base + current,
 301                       stk->baselen - current);
 302                strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
 303                add_excludes_from_file_1(dir->basebuf,
 304                                         dir->basebuf, stk->baselen,
 305                                         &stk->filebuf, el);
 306                dir->exclude_stack = stk;
 307                current = stk->baselen;
 308        }
 309        dir->basebuf[baselen] = '\0';
 310}
 311
 312/* Scan the list and let the last match determine the fate.
 313 * Return 1 for exclude, 0 for include and -1 for undecided.
 314 */
 315static int excluded_1(const char *pathname,
 316                      int pathlen, const char *basename, int *dtype,
 317                      struct exclude_list *el)
 318{
 319        int i;
 320
 321        if (el->nr) {
 322                for (i = el->nr - 1; 0 <= i; i--) {
 323                        struct exclude *x = el->excludes[i];
 324                        const char *exclude = x->pattern;
 325                        int to_exclude = x->to_exclude;
 326
 327                        if (x->flags & EXC_FLAG_MUSTBEDIR) {
 328                                if (*dtype == DT_UNKNOWN)
 329                                        *dtype = get_dtype(NULL, pathname, pathlen);
 330                                if (*dtype != DT_DIR)
 331                                        continue;
 332                        }
 333
 334                        if (x->flags & EXC_FLAG_NODIR) {
 335                                /* match basename */
 336                                if (x->flags & EXC_FLAG_NOWILDCARD) {
 337                                        if (!strcmp(exclude, basename))
 338                                                return to_exclude;
 339                                } else if (x->flags & EXC_FLAG_ENDSWITH) {
 340                                        if (x->patternlen - 1 <= pathlen &&
 341                                            !strcmp(exclude + 1, pathname + pathlen - x->patternlen + 1))
 342                                                return to_exclude;
 343                                } else {
 344                                        if (fnmatch(exclude, basename, 0) == 0)
 345                                                return to_exclude;
 346                                }
 347                        }
 348                        else {
 349                                /* match with FNM_PATHNAME:
 350                                 * exclude has base (baselen long) implicitly
 351                                 * in front of it.
 352                                 */
 353                                int baselen = x->baselen;
 354                                if (*exclude == '/')
 355                                        exclude++;
 356
 357                                if (pathlen < baselen ||
 358                                    (baselen && pathname[baselen-1] != '/') ||
 359                                    strncmp(pathname, x->base, baselen))
 360                                    continue;
 361
 362                                if (x->flags & EXC_FLAG_NOWILDCARD) {
 363                                        if (!strcmp(exclude, pathname + baselen))
 364                                                return to_exclude;
 365                                } else {
 366                                        if (fnmatch(exclude, pathname+baselen,
 367                                                    FNM_PATHNAME) == 0)
 368                                            return to_exclude;
 369                                }
 370                        }
 371                }
 372        }
 373        return -1; /* undecided */
 374}
 375
 376int excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
 377{
 378        int pathlen = strlen(pathname);
 379        int st;
 380        const char *basename = strrchr(pathname, '/');
 381        basename = (basename) ? basename+1 : pathname;
 382
 383        prep_exclude(dir, pathname, basename-pathname);
 384        for (st = EXC_CMDL; st <= EXC_FILE; st++) {
 385                switch (excluded_1(pathname, pathlen, basename,
 386                                   dtype_p, &dir->exclude_list[st])) {
 387                case 0:
 388                        return 0;
 389                case 1:
 390                        return 1;
 391                }
 392        }
 393        return 0;
 394}
 395
 396static struct dir_entry *dir_entry_new(const char *pathname, int len)
 397{
 398        struct dir_entry *ent;
 399
 400        ent = xmalloc(sizeof(*ent) + len + 1);
 401        ent->len = len;
 402        memcpy(ent->name, pathname, len);
 403        ent->name[len] = 0;
 404        return ent;
 405}
 406
 407static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
 408{
 409        if (cache_name_exists(pathname, len, ignore_case))
 410                return NULL;
 411
 412        ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
 413        return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
 414}
 415
 416static struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
 417{
 418        if (!cache_name_is_other(pathname, len))
 419                return NULL;
 420
 421        ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
 422        return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
 423}
 424
 425enum exist_status {
 426        index_nonexistent = 0,
 427        index_directory,
 428        index_gitdir,
 429};
 430
 431/*
 432 * The index sorts alphabetically by entry name, which
 433 * means that a gitlink sorts as '\0' at the end, while
 434 * a directory (which is defined not as an entry, but as
 435 * the files it contains) will sort with the '/' at the
 436 * end.
 437 */
 438static enum exist_status directory_exists_in_index(const char *dirname, int len)
 439{
 440        int pos = cache_name_pos(dirname, len);
 441        if (pos < 0)
 442                pos = -pos-1;
 443        while (pos < active_nr) {
 444                struct cache_entry *ce = active_cache[pos++];
 445                unsigned char endchar;
 446
 447                if (strncmp(ce->name, dirname, len))
 448                        break;
 449                endchar = ce->name[len];
 450                if (endchar > '/')
 451                        break;
 452                if (endchar == '/')
 453                        return index_directory;
 454                if (!endchar && S_ISGITLINK(ce->ce_mode))
 455                        return index_gitdir;
 456        }
 457        return index_nonexistent;
 458}
 459
 460/*
 461 * When we find a directory when traversing the filesystem, we
 462 * have three distinct cases:
 463 *
 464 *  - ignore it
 465 *  - see it as a directory
 466 *  - recurse into it
 467 *
 468 * and which one we choose depends on a combination of existing
 469 * git index contents and the flags passed into the directory
 470 * traversal routine.
 471 *
 472 * Case 1: If we *already* have entries in the index under that
 473 * directory name, we always recurse into the directory to see
 474 * all the files.
 475 *
 476 * Case 2: If we *already* have that directory name as a gitlink,
 477 * we always continue to see it as a gitlink, regardless of whether
 478 * there is an actual git directory there or not (it might not
 479 * be checked out as a subproject!)
 480 *
 481 * Case 3: if we didn't have it in the index previously, we
 482 * have a few sub-cases:
 483 *
 484 *  (a) if "show_other_directories" is true, we show it as
 485 *      just a directory, unless "hide_empty_directories" is
 486 *      also true and the directory is empty, in which case
 487 *      we just ignore it entirely.
 488 *  (b) if it looks like a git directory, and we don't have
 489 *      'no_gitlinks' set we treat it as a gitlink, and show it
 490 *      as a directory.
 491 *  (c) otherwise, we recurse into it.
 492 */
 493enum directory_treatment {
 494        show_directory,
 495        ignore_directory,
 496        recurse_into_directory,
 497};
 498
 499static enum directory_treatment treat_directory(struct dir_struct *dir,
 500        const char *dirname, int len,
 501        const struct path_simplify *simplify)
 502{
 503        /* The "len-1" is to strip the final '/' */
 504        switch (directory_exists_in_index(dirname, len-1)) {
 505        case index_directory:
 506                return recurse_into_directory;
 507
 508        case index_gitdir:
 509                if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 510                        return ignore_directory;
 511                return show_directory;
 512
 513        case index_nonexistent:
 514                if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 515                        break;
 516                if (!(dir->flags & DIR_NO_GITLINKS)) {
 517                        unsigned char sha1[20];
 518                        if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
 519                                return show_directory;
 520                }
 521                return recurse_into_directory;
 522        }
 523
 524        /* This is the "show_other_directories" case */
 525        if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
 526                return show_directory;
 527        if (!read_directory_recursive(dir, dirname, len, 1, simplify))
 528                return ignore_directory;
 529        return show_directory;
 530}
 531
 532/*
 533 * This is an inexact early pruning of any recursive directory
 534 * reading - if the path cannot possibly be in the pathspec,
 535 * return true, and we'll skip it early.
 536 */
 537static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
 538{
 539        if (simplify) {
 540                for (;;) {
 541                        const char *match = simplify->path;
 542                        int len = simplify->len;
 543
 544                        if (!match)
 545                                break;
 546                        if (len > pathlen)
 547                                len = pathlen;
 548                        if (!memcmp(path, match, len))
 549                                return 0;
 550                        simplify++;
 551                }
 552                return 1;
 553        }
 554        return 0;
 555}
 556
 557static int in_pathspec(const char *path, int len, const struct path_simplify *simplify)
 558{
 559        if (simplify) {
 560                for (; simplify->path; simplify++) {
 561                        if (len == simplify->len
 562                            && !memcmp(path, simplify->path, len))
 563                                return 1;
 564                }
 565        }
 566        return 0;
 567}
 568
 569static int get_index_dtype(const char *path, int len)
 570{
 571        int pos;
 572        struct cache_entry *ce;
 573
 574        ce = cache_name_exists(path, len, 0);
 575        if (ce) {
 576                if (!ce_uptodate(ce))
 577                        return DT_UNKNOWN;
 578                if (S_ISGITLINK(ce->ce_mode))
 579                        return DT_DIR;
 580                /*
 581                 * Nobody actually cares about the
 582                 * difference between DT_LNK and DT_REG
 583                 */
 584                return DT_REG;
 585        }
 586
 587        /* Try to look it up as a directory */
 588        pos = cache_name_pos(path, len);
 589        if (pos >= 0)
 590                return DT_UNKNOWN;
 591        pos = -pos-1;
 592        while (pos < active_nr) {
 593                ce = active_cache[pos++];
 594                if (strncmp(ce->name, path, len))
 595                        break;
 596                if (ce->name[len] > '/')
 597                        break;
 598                if (ce->name[len] < '/')
 599                        continue;
 600                if (!ce_uptodate(ce))
 601                        break;  /* continue? */
 602                return DT_DIR;
 603        }
 604        return DT_UNKNOWN;
 605}
 606
 607static int get_dtype(struct dirent *de, const char *path, int len)
 608{
 609        int dtype = de ? DTYPE(de) : DT_UNKNOWN;
 610        struct stat st;
 611
 612        if (dtype != DT_UNKNOWN)
 613                return dtype;
 614        dtype = get_index_dtype(path, len);
 615        if (dtype != DT_UNKNOWN)
 616                return dtype;
 617        if (lstat(path, &st))
 618                return dtype;
 619        if (S_ISREG(st.st_mode))
 620                return DT_REG;
 621        if (S_ISDIR(st.st_mode))
 622                return DT_DIR;
 623        if (S_ISLNK(st.st_mode))
 624                return DT_LNK;
 625        return dtype;
 626}
 627
 628/*
 629 * Read a directory tree. We currently ignore anything but
 630 * directories, regular files and symlinks. That's because git
 631 * doesn't handle them at all yet. Maybe that will change some
 632 * day.
 633 *
 634 * Also, we ignore the name ".git" (even if it is not a directory).
 635 * That likely will not change.
 636 */
 637static int read_directory_recursive(struct dir_struct *dir, const char *base, int baselen, int check_only, const struct path_simplify *simplify)
 638{
 639        DIR *fdir = opendir(*base ? base : ".");
 640        int contents = 0;
 641
 642        if (fdir) {
 643                struct dirent *de;
 644                char path[PATH_MAX + 1];
 645                memcpy(path, base, baselen);
 646
 647                while ((de = readdir(fdir)) != NULL) {
 648                        int len, dtype;
 649                        int exclude;
 650
 651                        if (is_dot_or_dotdot(de->d_name) ||
 652                             !strcmp(de->d_name, ".git"))
 653                                continue;
 654                        len = strlen(de->d_name);
 655                        /* Ignore overly long pathnames! */
 656                        if (len + baselen + 8 > sizeof(path))
 657                                continue;
 658                        memcpy(path + baselen, de->d_name, len+1);
 659                        len = baselen + len;
 660                        if (simplify_away(path, len, simplify))
 661                                continue;
 662
 663                        dtype = DTYPE(de);
 664                        exclude = excluded(dir, path, &dtype);
 665                        if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
 666                            && in_pathspec(path, len, simplify))
 667                                dir_add_ignored(dir, path,len);
 668
 669                        /*
 670                         * Excluded? If we don't explicitly want to show
 671                         * ignored files, ignore it
 672                         */
 673                        if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
 674                                continue;
 675
 676                        if (dtype == DT_UNKNOWN)
 677                                dtype = get_dtype(de, path, len);
 678
 679                        /*
 680                         * Do we want to see just the ignored files?
 681                         * We still need to recurse into directories,
 682                         * even if we don't ignore them, since the
 683                         * directory may contain files that we do..
 684                         */
 685                        if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) {
 686                                if (dtype != DT_DIR)
 687                                        continue;
 688                        }
 689
 690                        switch (dtype) {
 691                        default:
 692                                continue;
 693                        case DT_DIR:
 694                                memcpy(path + len, "/", 2);
 695                                len++;
 696                                switch (treat_directory(dir, path, len, simplify)) {
 697                                case show_directory:
 698                                        if (exclude != !!(dir->flags
 699                                                        & DIR_SHOW_IGNORED))
 700                                                continue;
 701                                        break;
 702                                case recurse_into_directory:
 703                                        contents += read_directory_recursive(dir,
 704                                                path, len, 0, simplify);
 705                                        continue;
 706                                case ignore_directory:
 707                                        continue;
 708                                }
 709                                break;
 710                        case DT_REG:
 711                        case DT_LNK:
 712                                break;
 713                        }
 714                        contents++;
 715                        if (check_only)
 716                                goto exit_early;
 717                        else
 718                                dir_add_name(dir, path, len);
 719                }
 720exit_early:
 721                closedir(fdir);
 722        }
 723
 724        return contents;
 725}
 726
 727static int cmp_name(const void *p1, const void *p2)
 728{
 729        const struct dir_entry *e1 = *(const struct dir_entry **)p1;
 730        const struct dir_entry *e2 = *(const struct dir_entry **)p2;
 731
 732        return cache_name_compare(e1->name, e1->len,
 733                                  e2->name, e2->len);
 734}
 735
 736/*
 737 * Return the length of the "simple" part of a path match limiter.
 738 */
 739static int simple_length(const char *match)
 740{
 741        int len = -1;
 742
 743        for (;;) {
 744                unsigned char c = *match++;
 745                len++;
 746                if (c == '\0' || is_glob_special(c))
 747                        return len;
 748        }
 749}
 750
 751static struct path_simplify *create_simplify(const char **pathspec)
 752{
 753        int nr, alloc = 0;
 754        struct path_simplify *simplify = NULL;
 755
 756        if (!pathspec)
 757                return NULL;
 758
 759        for (nr = 0 ; ; nr++) {
 760                const char *match;
 761                if (nr >= alloc) {
 762                        alloc = alloc_nr(alloc);
 763                        simplify = xrealloc(simplify, alloc * sizeof(*simplify));
 764                }
 765                match = *pathspec++;
 766                if (!match)
 767                        break;
 768                simplify[nr].path = match;
 769                simplify[nr].len = simple_length(match);
 770        }
 771        simplify[nr].path = NULL;
 772        simplify[nr].len = 0;
 773        return simplify;
 774}
 775
 776static void free_simplify(struct path_simplify *simplify)
 777{
 778        free(simplify);
 779}
 780
 781int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
 782{
 783        struct path_simplify *simplify;
 784
 785        if (has_symlink_leading_path(path, len))
 786                return dir->nr;
 787
 788        simplify = create_simplify(pathspec);
 789        read_directory_recursive(dir, path, len, 0, simplify);
 790        free_simplify(simplify);
 791        qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
 792        qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
 793        return dir->nr;
 794}
 795
 796int file_exists(const char *f)
 797{
 798        struct stat sb;
 799        return lstat(f, &sb) == 0;
 800}
 801
 802/*
 803 * get_relative_cwd() gets the prefix of the current working directory
 804 * relative to 'dir'.  If we are not inside 'dir', it returns NULL.
 805 *
 806 * As a convenience, it also returns NULL if 'dir' is already NULL.  The
 807 * reason for this behaviour is that it is natural for functions returning
 808 * directory names to return NULL to say "this directory does not exist"
 809 * or "this directory is invalid".  These cases are usually handled the
 810 * same as if the cwd is not inside 'dir' at all, so get_relative_cwd()
 811 * returns NULL for both of them.
 812 *
 813 * Most notably, get_relative_cwd(buffer, size, get_git_work_tree())
 814 * unifies the handling of "outside work tree" with "no work tree at all".
 815 */
 816char *get_relative_cwd(char *buffer, int size, const char *dir)
 817{
 818        char *cwd = buffer;
 819
 820        if (!dir)
 821                return NULL;
 822        if (!getcwd(buffer, size))
 823                die_errno("can't find the current directory");
 824
 825        if (!is_absolute_path(dir))
 826                dir = make_absolute_path(dir);
 827
 828        while (*dir && *dir == *cwd) {
 829                dir++;
 830                cwd++;
 831        }
 832        if (*dir)
 833                return NULL;
 834        if (*cwd == '/')
 835                return cwd + 1;
 836        return cwd;
 837}
 838
 839int is_inside_dir(const char *dir)
 840{
 841        char buffer[PATH_MAX];
 842        return get_relative_cwd(buffer, sizeof(buffer), dir) != NULL;
 843}
 844
 845int is_empty_dir(const char *path)
 846{
 847        DIR *dir = opendir(path);
 848        struct dirent *e;
 849        int ret = 1;
 850
 851        if (!dir)
 852                return 0;
 853
 854        while ((e = readdir(dir)) != NULL)
 855                if (!is_dot_or_dotdot(e->d_name)) {
 856                        ret = 0;
 857                        break;
 858                }
 859
 860        closedir(dir);
 861        return ret;
 862}
 863
 864int remove_dir_recursively(struct strbuf *path, int flag)
 865{
 866        DIR *dir;
 867        struct dirent *e;
 868        int ret = 0, original_len = path->len, len;
 869        int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
 870        unsigned char submodule_head[20];
 871
 872        if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
 873            !resolve_gitlink_ref(path->buf, "HEAD", submodule_head))
 874                /* Do not descend and nuke a nested git work tree. */
 875                return 0;
 876
 877        dir = opendir(path->buf);
 878        if (!dir)
 879                return -1;
 880        if (path->buf[original_len - 1] != '/')
 881                strbuf_addch(path, '/');
 882
 883        len = path->len;
 884        while ((e = readdir(dir)) != NULL) {
 885                struct stat st;
 886                if (is_dot_or_dotdot(e->d_name))
 887                        continue;
 888
 889                strbuf_setlen(path, len);
 890                strbuf_addstr(path, e->d_name);
 891                if (lstat(path->buf, &st))
 892                        ; /* fall thru */
 893                else if (S_ISDIR(st.st_mode)) {
 894                        if (!remove_dir_recursively(path, only_empty))
 895                                continue; /* happy */
 896                } else if (!only_empty && !unlink(path->buf))
 897                        continue; /* happy, too */
 898
 899                /* path too long, stat fails, or non-directory still exists */
 900                ret = -1;
 901                break;
 902        }
 903        closedir(dir);
 904
 905        strbuf_setlen(path, original_len);
 906        if (!ret)
 907                ret = rmdir(path->buf);
 908        return ret;
 909}
 910
 911void setup_standard_excludes(struct dir_struct *dir)
 912{
 913        const char *path;
 914
 915        dir->exclude_per_dir = ".gitignore";
 916        path = git_path("info/exclude");
 917        if (!access(path, R_OK))
 918                add_excludes_from_file(dir, path);
 919        if (excludes_file && !access(excludes_file, R_OK))
 920                add_excludes_from_file(dir, excludes_file);
 921}
 922
 923int remove_path(const char *name)
 924{
 925        char *slash;
 926
 927        if (unlink(name) && errno != ENOENT)
 928                return -1;
 929
 930        slash = strrchr(name, '/');
 931        if (slash) {
 932                char *dirs = xstrdup(name);
 933                slash = dirs + (slash - name);
 934                do {
 935                        *slash = '\0';
 936                } while (rmdir(dirs) && (slash = strrchr(dirs, '/')));
 937                free(dirs);
 938        }
 939        return 0;
 940}
 941