dir.con commit dir.c: factor out parts of last_exclude_matching for later reuse (46aa2f9)
   1/*
   2 * This handles recursive filename detection with exclude
   3 * files, index knowledge etc..
   4 *
   5 * See Documentation/technical/api-directory-listing.txt
   6 *
   7 * Copyright (C) Linus Torvalds, 2005-2006
   8 *               Junio Hamano, 2005-2006
   9 */
  10#include "cache.h"
  11#include "dir.h"
  12#include "refs.h"
  13#include "wildmatch.h"
  14
  15struct path_simplify {
  16        int len;
  17        const char *path;
  18};
  19
  20static int read_directory_recursive(struct dir_struct *dir, const char *path, int len,
  21        int check_only, const struct path_simplify *simplify);
  22static int get_dtype(struct dirent *de, const char *path, int len);
  23
  24/* helper string functions with support for the ignore_case flag */
  25int strcmp_icase(const char *a, const char *b)
  26{
  27        return ignore_case ? strcasecmp(a, b) : strcmp(a, b);
  28}
  29
  30int strncmp_icase(const char *a, const char *b, size_t count)
  31{
  32        return ignore_case ? strncasecmp(a, b, count) : strncmp(a, b, count);
  33}
  34
  35int fnmatch_icase(const char *pattern, const char *string, int flags)
  36{
  37        return fnmatch(pattern, string, flags | (ignore_case ? FNM_CASEFOLD : 0));
  38}
  39
  40inline int git_fnmatch(const char *pattern, const char *string,
  41                       int flags, int prefix)
  42{
  43        int fnm_flags = 0;
  44        if (flags & GFNM_PATHNAME)
  45                fnm_flags |= FNM_PATHNAME;
  46        if (prefix > 0) {
  47                if (strncmp(pattern, string, prefix))
  48                        return FNM_NOMATCH;
  49                pattern += prefix;
  50                string += prefix;
  51        }
  52        if (flags & GFNM_ONESTAR) {
  53                int pattern_len = strlen(++pattern);
  54                int string_len = strlen(string);
  55                return string_len < pattern_len ||
  56                       strcmp(pattern,
  57                              string + string_len - pattern_len);
  58        }
  59        return fnmatch(pattern, string, fnm_flags);
  60}
  61
  62static int fnmatch_icase_mem(const char *pattern, int patternlen,
  63                             const char *string, int stringlen,
  64                             int flags)
  65{
  66        int match_status;
  67        struct strbuf pat_buf = STRBUF_INIT;
  68        struct strbuf str_buf = STRBUF_INIT;
  69        const char *use_pat = pattern;
  70        const char *use_str = string;
  71
  72        if (pattern[patternlen]) {
  73                strbuf_add(&pat_buf, pattern, patternlen);
  74                use_pat = pat_buf.buf;
  75        }
  76        if (string[stringlen]) {
  77                strbuf_add(&str_buf, string, stringlen);
  78                use_str = str_buf.buf;
  79        }
  80
  81        if (ignore_case)
  82                flags |= WM_CASEFOLD;
  83        match_status = wildmatch(use_pat, use_str, flags, NULL);
  84
  85        strbuf_release(&pat_buf);
  86        strbuf_release(&str_buf);
  87
  88        return match_status;
  89}
  90
  91static size_t common_prefix_len(const char **pathspec)
  92{
  93        const char *n, *first;
  94        size_t max = 0;
  95        int literal = limit_pathspec_to_literal();
  96
  97        if (!pathspec)
  98                return max;
  99
 100        first = *pathspec;
 101        while ((n = *pathspec++)) {
 102                size_t i, len = 0;
 103                for (i = 0; first == n || i < max; i++) {
 104                        char c = n[i];
 105                        if (!c || c != first[i] || (!literal && is_glob_special(c)))
 106                                break;
 107                        if (c == '/')
 108                                len = i + 1;
 109                }
 110                if (first == n || len < max) {
 111                        max = len;
 112                        if (!max)
 113                                break;
 114                }
 115        }
 116        return max;
 117}
 118
 119/*
 120 * Returns a copy of the longest leading path common among all
 121 * pathspecs.
 122 */
 123char *common_prefix(const char **pathspec)
 124{
 125        unsigned long len = common_prefix_len(pathspec);
 126
 127        return len ? xmemdupz(*pathspec, len) : NULL;
 128}
 129
 130int fill_directory(struct dir_struct *dir, const char **pathspec)
 131{
 132        size_t len;
 133
 134        /*
 135         * Calculate common prefix for the pathspec, and
 136         * use that to optimize the directory walk
 137         */
 138        len = common_prefix_len(pathspec);
 139
 140        /* Read the directory and prune it */
 141        read_directory(dir, pathspec ? *pathspec : "", len, pathspec);
 142        return len;
 143}
 144
 145int within_depth(const char *name, int namelen,
 146                        int depth, int max_depth)
 147{
 148        const char *cp = name, *cpe = name + namelen;
 149
 150        while (cp < cpe) {
 151                if (*cp++ != '/')
 152                        continue;
 153                depth++;
 154                if (depth > max_depth)
 155                        return 0;
 156        }
 157        return 1;
 158}
 159
 160/*
 161 * Does 'match' match the given name?
 162 * A match is found if
 163 *
 164 * (1) the 'match' string is leading directory of 'name', or
 165 * (2) the 'match' string is a wildcard and matches 'name', or
 166 * (3) the 'match' string is exactly the same as 'name'.
 167 *
 168 * and the return value tells which case it was.
 169 *
 170 * It returns 0 when there is no match.
 171 */
 172static int match_one(const char *match, const char *name, int namelen)
 173{
 174        int matchlen;
 175        int literal = limit_pathspec_to_literal();
 176
 177        /* If the match was just the prefix, we matched */
 178        if (!*match)
 179                return MATCHED_RECURSIVELY;
 180
 181        if (ignore_case) {
 182                for (;;) {
 183                        unsigned char c1 = tolower(*match);
 184                        unsigned char c2 = tolower(*name);
 185                        if (c1 == '\0' || (!literal && is_glob_special(c1)))
 186                                break;
 187                        if (c1 != c2)
 188                                return 0;
 189                        match++;
 190                        name++;
 191                        namelen--;
 192                }
 193        } else {
 194                for (;;) {
 195                        unsigned char c1 = *match;
 196                        unsigned char c2 = *name;
 197                        if (c1 == '\0' || (!literal && is_glob_special(c1)))
 198                                break;
 199                        if (c1 != c2)
 200                                return 0;
 201                        match++;
 202                        name++;
 203                        namelen--;
 204                }
 205        }
 206
 207        /*
 208         * If we don't match the matchstring exactly,
 209         * we need to match by fnmatch
 210         */
 211        matchlen = strlen(match);
 212        if (strncmp_icase(match, name, matchlen)) {
 213                if (literal)
 214                        return 0;
 215                return !fnmatch_icase(match, name, 0) ? MATCHED_FNMATCH : 0;
 216        }
 217
 218        if (namelen == matchlen)
 219                return MATCHED_EXACTLY;
 220        if (match[matchlen-1] == '/' || name[matchlen] == '/')
 221                return MATCHED_RECURSIVELY;
 222        return 0;
 223}
 224
 225/*
 226 * Given a name and a list of pathspecs, returns the nature of the
 227 * closest (i.e. most specific) match of the name to any of the
 228 * pathspecs.
 229 *
 230 * The caller typically calls this multiple times with the same
 231 * pathspec and seen[] array but with different name/namelen
 232 * (e.g. entries from the index) and is interested in seeing if and
 233 * how each pathspec matches all the names it calls this function
 234 * with.  A mark is left in the seen[] array for each pathspec element
 235 * indicating the closest type of match that element achieved, so if
 236 * seen[n] remains zero after multiple invocations, that means the nth
 237 * pathspec did not match any names, which could indicate that the
 238 * user mistyped the nth pathspec.
 239 */
 240int match_pathspec(const char **pathspec, const char *name, int namelen,
 241                int prefix, char *seen)
 242{
 243        int i, retval = 0;
 244
 245        if (!pathspec)
 246                return 1;
 247
 248        name += prefix;
 249        namelen -= prefix;
 250
 251        for (i = 0; pathspec[i] != NULL; i++) {
 252                int how;
 253                const char *match = pathspec[i] + prefix;
 254                if (seen && seen[i] == MATCHED_EXACTLY)
 255                        continue;
 256                how = match_one(match, name, namelen);
 257                if (how) {
 258                        if (retval < how)
 259                                retval = how;
 260                        if (seen && seen[i] < how)
 261                                seen[i] = how;
 262                }
 263        }
 264        return retval;
 265}
 266
 267/*
 268 * Does 'match' match the given name?
 269 * A match is found if
 270 *
 271 * (1) the 'match' string is leading directory of 'name', or
 272 * (2) the 'match' string is a wildcard and matches 'name', or
 273 * (3) the 'match' string is exactly the same as 'name'.
 274 *
 275 * and the return value tells which case it was.
 276 *
 277 * It returns 0 when there is no match.
 278 */
 279static int match_pathspec_item(const struct pathspec_item *item, int prefix,
 280                               const char *name, int namelen)
 281{
 282        /* name/namelen has prefix cut off by caller */
 283        const char *match = item->match + prefix;
 284        int matchlen = item->len - prefix;
 285
 286        /* If the match was just the prefix, we matched */
 287        if (!*match)
 288                return MATCHED_RECURSIVELY;
 289
 290        if (matchlen <= namelen && !strncmp(match, name, matchlen)) {
 291                if (matchlen == namelen)
 292                        return MATCHED_EXACTLY;
 293
 294                if (match[matchlen-1] == '/' || name[matchlen] == '/')
 295                        return MATCHED_RECURSIVELY;
 296        }
 297
 298        if (item->nowildcard_len < item->len &&
 299            !git_fnmatch(match, name,
 300                         item->flags & PATHSPEC_ONESTAR ? GFNM_ONESTAR : 0,
 301                         item->nowildcard_len - prefix))
 302                return MATCHED_FNMATCH;
 303
 304        return 0;
 305}
 306
 307/*
 308 * Given a name and a list of pathspecs, returns the nature of the
 309 * closest (i.e. most specific) match of the name to any of the
 310 * pathspecs.
 311 *
 312 * The caller typically calls this multiple times with the same
 313 * pathspec and seen[] array but with different name/namelen
 314 * (e.g. entries from the index) and is interested in seeing if and
 315 * how each pathspec matches all the names it calls this function
 316 * with.  A mark is left in the seen[] array for each pathspec element
 317 * indicating the closest type of match that element achieved, so if
 318 * seen[n] remains zero after multiple invocations, that means the nth
 319 * pathspec did not match any names, which could indicate that the
 320 * user mistyped the nth pathspec.
 321 */
 322int match_pathspec_depth(const struct pathspec *ps,
 323                         const char *name, int namelen,
 324                         int prefix, char *seen)
 325{
 326        int i, retval = 0;
 327
 328        if (!ps->nr) {
 329                if (!ps->recursive || ps->max_depth == -1)
 330                        return MATCHED_RECURSIVELY;
 331
 332                if (within_depth(name, namelen, 0, ps->max_depth))
 333                        return MATCHED_EXACTLY;
 334                else
 335                        return 0;
 336        }
 337
 338        name += prefix;
 339        namelen -= prefix;
 340
 341        for (i = ps->nr - 1; i >= 0; i--) {
 342                int how;
 343                if (seen && seen[i] == MATCHED_EXACTLY)
 344                        continue;
 345                how = match_pathspec_item(ps->items+i, prefix, name, namelen);
 346                if (ps->recursive && ps->max_depth != -1 &&
 347                    how && how != MATCHED_FNMATCH) {
 348                        int len = ps->items[i].len;
 349                        if (name[len] == '/')
 350                                len++;
 351                        if (within_depth(name+len, namelen-len, 0, ps->max_depth))
 352                                how = MATCHED_EXACTLY;
 353                        else
 354                                how = 0;
 355                }
 356                if (how) {
 357                        if (retval < how)
 358                                retval = how;
 359                        if (seen && seen[i] < how)
 360                                seen[i] = how;
 361                }
 362        }
 363        return retval;
 364}
 365
 366/*
 367 * Return the length of the "simple" part of a path match limiter.
 368 */
 369static int simple_length(const char *match)
 370{
 371        int len = -1;
 372
 373        for (;;) {
 374                unsigned char c = *match++;
 375                len++;
 376                if (c == '\0' || is_glob_special(c))
 377                        return len;
 378        }
 379}
 380
 381static int no_wildcard(const char *string)
 382{
 383        return string[simple_length(string)] == '\0';
 384}
 385
 386void parse_exclude_pattern(const char **pattern,
 387                           int *patternlen,
 388                           int *flags,
 389                           int *nowildcardlen)
 390{
 391        const char *p = *pattern;
 392        size_t i, len;
 393
 394        *flags = 0;
 395        if (*p == '!') {
 396                *flags |= EXC_FLAG_NEGATIVE;
 397                p++;
 398        }
 399        len = strlen(p);
 400        if (len && p[len - 1] == '/') {
 401                len--;
 402                *flags |= EXC_FLAG_MUSTBEDIR;
 403        }
 404        for (i = 0; i < len; i++) {
 405                if (p[i] == '/')
 406                        break;
 407        }
 408        if (i == len)
 409                *flags |= EXC_FLAG_NODIR;
 410        *nowildcardlen = simple_length(p);
 411        /*
 412         * we should have excluded the trailing slash from 'p' too,
 413         * but that's one more allocation. Instead just make sure
 414         * nowildcardlen does not exceed real patternlen
 415         */
 416        if (*nowildcardlen > len)
 417                *nowildcardlen = len;
 418        if (*p == '*' && no_wildcard(p + 1))
 419                *flags |= EXC_FLAG_ENDSWITH;
 420        *pattern = p;
 421        *patternlen = len;
 422}
 423
 424void add_exclude(const char *string, const char *base,
 425                 int baselen, struct exclude_list *el, int srcpos)
 426{
 427        struct exclude *x;
 428        int patternlen;
 429        int flags;
 430        int nowildcardlen;
 431
 432        parse_exclude_pattern(&string, &patternlen, &flags, &nowildcardlen);
 433        if (flags & EXC_FLAG_MUSTBEDIR) {
 434                char *s;
 435                x = xmalloc(sizeof(*x) + patternlen + 1);
 436                s = (char *)(x+1);
 437                memcpy(s, string, patternlen);
 438                s[patternlen] = '\0';
 439                x->pattern = s;
 440        } else {
 441                x = xmalloc(sizeof(*x));
 442                x->pattern = string;
 443        }
 444        x->patternlen = patternlen;
 445        x->nowildcardlen = nowildcardlen;
 446        x->base = base;
 447        x->baselen = baselen;
 448        x->flags = flags;
 449        x->srcpos = srcpos;
 450        ALLOC_GROW(el->excludes, el->nr + 1, el->alloc);
 451        el->excludes[el->nr++] = x;
 452        x->el = el;
 453}
 454
 455static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
 456{
 457        int pos, len;
 458        unsigned long sz;
 459        enum object_type type;
 460        void *data;
 461        struct index_state *istate = &the_index;
 462
 463        len = strlen(path);
 464        pos = index_name_pos(istate, path, len);
 465        if (pos < 0)
 466                return NULL;
 467        if (!ce_skip_worktree(istate->cache[pos]))
 468                return NULL;
 469        data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
 470        if (!data || type != OBJ_BLOB) {
 471                free(data);
 472                return NULL;
 473        }
 474        *size = xsize_t(sz);
 475        return data;
 476}
 477
 478/*
 479 * Frees memory within el which was allocated for exclude patterns and
 480 * the file buffer.  Does not free el itself.
 481 */
 482void clear_exclude_list(struct exclude_list *el)
 483{
 484        int i;
 485
 486        for (i = 0; i < el->nr; i++)
 487                free(el->excludes[i]);
 488        free(el->excludes);
 489        free(el->filebuf);
 490
 491        el->nr = 0;
 492        el->excludes = NULL;
 493        el->filebuf = NULL;
 494}
 495
 496int add_excludes_from_file_to_list(const char *fname,
 497                                   const char *base,
 498                                   int baselen,
 499                                   struct exclude_list *el,
 500                                   int check_index)
 501{
 502        struct stat st;
 503        int fd, i, lineno = 1;
 504        size_t size = 0;
 505        char *buf, *entry;
 506
 507        fd = open(fname, O_RDONLY);
 508        if (fd < 0 || fstat(fd, &st) < 0) {
 509                if (errno != ENOENT)
 510                        warn_on_inaccessible(fname);
 511                if (0 <= fd)
 512                        close(fd);
 513                if (!check_index ||
 514                    (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
 515                        return -1;
 516                if (size == 0) {
 517                        free(buf);
 518                        return 0;
 519                }
 520                if (buf[size-1] != '\n') {
 521                        buf = xrealloc(buf, size+1);
 522                        buf[size++] = '\n';
 523                }
 524        }
 525        else {
 526                size = xsize_t(st.st_size);
 527                if (size == 0) {
 528                        close(fd);
 529                        return 0;
 530                }
 531                buf = xmalloc(size+1);
 532                if (read_in_full(fd, buf, size) != size) {
 533                        free(buf);
 534                        close(fd);
 535                        return -1;
 536                }
 537                buf[size++] = '\n';
 538                close(fd);
 539        }
 540
 541        el->filebuf = buf;
 542        entry = buf;
 543        for (i = 0; i < size; i++) {
 544                if (buf[i] == '\n') {
 545                        if (entry != buf + i && entry[0] != '#') {
 546                                buf[i - (i && buf[i-1] == '\r')] = 0;
 547                                add_exclude(entry, base, baselen, el, lineno);
 548                        }
 549                        lineno++;
 550                        entry = buf + i + 1;
 551                }
 552        }
 553        return 0;
 554}
 555
 556struct exclude_list *add_exclude_list(struct dir_struct *dir,
 557                                      int group_type, const char *src)
 558{
 559        struct exclude_list *el;
 560        struct exclude_list_group *group;
 561
 562        group = &dir->exclude_list_group[group_type];
 563        ALLOC_GROW(group->el, group->nr + 1, group->alloc);
 564        el = &group->el[group->nr++];
 565        memset(el, 0, sizeof(*el));
 566        el->src = src;
 567        return el;
 568}
 569
 570/*
 571 * Used to set up core.excludesfile and .git/info/exclude lists.
 572 */
 573void add_excludes_from_file(struct dir_struct *dir, const char *fname)
 574{
 575        struct exclude_list *el;
 576        el = add_exclude_list(dir, EXC_FILE, fname);
 577        if (add_excludes_from_file_to_list(fname, "", 0, el, 0) < 0)
 578                die("cannot use %s as an exclude file", fname);
 579}
 580
 581/*
 582 * Loads the per-directory exclude list for the substring of base
 583 * which has a char length of baselen.
 584 */
 585static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
 586{
 587        struct exclude_list_group *group;
 588        struct exclude_list *el;
 589        struct exclude_stack *stk = NULL;
 590        int current;
 591
 592        if ((!dir->exclude_per_dir) ||
 593            (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
 594                return; /* too long a path -- ignore */
 595
 596        group = &dir->exclude_list_group[EXC_DIRS];
 597
 598        /* Pop the exclude lists from the EXCL_DIRS exclude_list_group
 599         * which originate from directories not in the prefix of the
 600         * path being checked. */
 601        while ((stk = dir->exclude_stack) != NULL) {
 602                if (stk->baselen <= baselen &&
 603                    !strncmp(dir->basebuf, base, stk->baselen))
 604                        break;
 605                el = &group->el[dir->exclude_stack->exclude_ix];
 606                dir->exclude_stack = stk->prev;
 607                free((char *)el->src); /* see strdup() below */
 608                clear_exclude_list(el);
 609                free(stk);
 610                group->nr--;
 611        }
 612
 613        /* Read from the parent directories and push them down. */
 614        current = stk ? stk->baselen : -1;
 615        while (current < baselen) {
 616                struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
 617                const char *cp;
 618
 619                if (current < 0) {
 620                        cp = base;
 621                        current = 0;
 622                }
 623                else {
 624                        cp = strchr(base + current + 1, '/');
 625                        if (!cp)
 626                                die("oops in prep_exclude");
 627                        cp++;
 628                }
 629                stk->prev = dir->exclude_stack;
 630                stk->baselen = cp - base;
 631                memcpy(dir->basebuf + current, base + current,
 632                       stk->baselen - current);
 633                strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
 634                /*
 635                 * dir->basebuf gets reused by the traversal, but we
 636                 * need fname to remain unchanged to ensure the src
 637                 * member of each struct exclude correctly
 638                 * back-references its source file.  Other invocations
 639                 * of add_exclude_list provide stable strings, so we
 640                 * strdup() and free() here in the caller.
 641                 */
 642                el = add_exclude_list(dir, EXC_DIRS, strdup(dir->basebuf));
 643                stk->exclude_ix = group->nr - 1;
 644                add_excludes_from_file_to_list(dir->basebuf,
 645                                               dir->basebuf, stk->baselen,
 646                                               el, 1);
 647                dir->exclude_stack = stk;
 648                current = stk->baselen;
 649        }
 650        dir->basebuf[baselen] = '\0';
 651}
 652
 653int match_basename(const char *basename, int basenamelen,
 654                   const char *pattern, int prefix, int patternlen,
 655                   int flags)
 656{
 657        if (prefix == patternlen) {
 658                if (patternlen == basenamelen &&
 659                    !strncmp_icase(pattern, basename, basenamelen))
 660                        return 1;
 661        } else if (flags & EXC_FLAG_ENDSWITH) {
 662                /* "*literal" matching against "fooliteral" */
 663                if (patternlen - 1 <= basenamelen &&
 664                    !strncmp_icase(pattern + 1,
 665                                   basename + basenamelen - (patternlen - 1),
 666                                   patternlen - 1))
 667                        return 1;
 668        } else {
 669                if (fnmatch_icase_mem(pattern, patternlen,
 670                                      basename, basenamelen,
 671                                      0) == 0)
 672                        return 1;
 673        }
 674        return 0;
 675}
 676
 677int match_pathname(const char *pathname, int pathlen,
 678                   const char *base, int baselen,
 679                   const char *pattern, int prefix, int patternlen,
 680                   int flags)
 681{
 682        const char *name;
 683        int namelen;
 684
 685        /*
 686         * match with FNM_PATHNAME; the pattern has base implicitly
 687         * in front of it.
 688         */
 689        if (*pattern == '/') {
 690                pattern++;
 691                patternlen--;
 692                prefix--;
 693        }
 694
 695        /*
 696         * baselen does not count the trailing slash. base[] may or
 697         * may not end with a trailing slash though.
 698         */
 699        if (pathlen < baselen + 1 ||
 700            (baselen && pathname[baselen] != '/') ||
 701            strncmp_icase(pathname, base, baselen))
 702                return 0;
 703
 704        namelen = baselen ? pathlen - baselen - 1 : pathlen;
 705        name = pathname + pathlen - namelen;
 706
 707        if (prefix) {
 708                /*
 709                 * if the non-wildcard part is longer than the
 710                 * remaining pathname, surely it cannot match.
 711                 */
 712                if (prefix > namelen)
 713                        return 0;
 714
 715                if (strncmp_icase(pattern, name, prefix))
 716                        return 0;
 717                pattern += prefix;
 718                patternlen -= prefix;
 719                name    += prefix;
 720                namelen -= prefix;
 721
 722                /*
 723                 * If the whole pattern did not have a wildcard,
 724                 * then our prefix match is all we need; we
 725                 * do not need to call fnmatch at all.
 726                 */
 727                if (!patternlen && !namelen)
 728                        return 1;
 729        }
 730
 731        return fnmatch_icase_mem(pattern, patternlen,
 732                                 name, namelen,
 733                                 WM_PATHNAME) == 0;
 734}
 735
 736/*
 737 * Scan the given exclude list in reverse to see whether pathname
 738 * should be ignored.  The first match (i.e. the last on the list), if
 739 * any, determines the fate.  Returns the exclude_list element which
 740 * matched, or NULL for undecided.
 741 */
 742static struct exclude *last_exclude_matching_from_list(const char *pathname,
 743                                                       int pathlen,
 744                                                       const char *basename,
 745                                                       int *dtype,
 746                                                       struct exclude_list *el)
 747{
 748        int i;
 749
 750        if (!el->nr)
 751                return NULL;    /* undefined */
 752
 753        for (i = el->nr - 1; 0 <= i; i--) {
 754                struct exclude *x = el->excludes[i];
 755                const char *exclude = x->pattern;
 756                int prefix = x->nowildcardlen;
 757
 758                if (x->flags & EXC_FLAG_MUSTBEDIR) {
 759                        if (*dtype == DT_UNKNOWN)
 760                                *dtype = get_dtype(NULL, pathname, pathlen);
 761                        if (*dtype != DT_DIR)
 762                                continue;
 763                }
 764
 765                if (x->flags & EXC_FLAG_NODIR) {
 766                        if (match_basename(basename,
 767                                           pathlen - (basename - pathname),
 768                                           exclude, prefix, x->patternlen,
 769                                           x->flags))
 770                                return x;
 771                        continue;
 772                }
 773
 774                assert(x->baselen == 0 || x->base[x->baselen - 1] == '/');
 775                if (match_pathname(pathname, pathlen,
 776                                   x->base, x->baselen ? x->baselen - 1 : 0,
 777                                   exclude, prefix, x->patternlen, x->flags))
 778                        return x;
 779        }
 780        return NULL; /* undecided */
 781}
 782
 783/*
 784 * Scan the list and let the last match determine the fate.
 785 * Return 1 for exclude, 0 for include and -1 for undecided.
 786 */
 787int is_excluded_from_list(const char *pathname,
 788                          int pathlen, const char *basename, int *dtype,
 789                          struct exclude_list *el)
 790{
 791        struct exclude *exclude;
 792        exclude = last_exclude_matching_from_list(pathname, pathlen, basename, dtype, el);
 793        if (exclude)
 794                return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
 795        return -1; /* undecided */
 796}
 797
 798static struct exclude *last_exclude_matching_from_lists(struct dir_struct *dir,
 799                const char *pathname, int pathlen, const char *basename,
 800                int *dtype_p)
 801{
 802        int i, j;
 803        struct exclude_list_group *group;
 804        struct exclude *exclude;
 805        for (i = EXC_CMDL; i <= EXC_FILE; i++) {
 806                group = &dir->exclude_list_group[i];
 807                for (j = group->nr - 1; j >= 0; j--) {
 808                        exclude = last_exclude_matching_from_list(
 809                                pathname, pathlen, basename, dtype_p,
 810                                &group->el[j]);
 811                        if (exclude)
 812                                return exclude;
 813                }
 814        }
 815        return NULL;
 816}
 817
 818/*
 819 * Loads the exclude lists for the directory containing pathname, then
 820 * scans all exclude lists to determine whether pathname is excluded.
 821 * Returns the exclude_list element which matched, or NULL for
 822 * undecided.
 823 */
 824static struct exclude *last_exclude_matching(struct dir_struct *dir,
 825                                             const char *pathname,
 826                                             int *dtype_p)
 827{
 828        int pathlen = strlen(pathname);
 829        const char *basename = strrchr(pathname, '/');
 830        basename = (basename) ? basename+1 : pathname;
 831
 832        prep_exclude(dir, pathname, basename-pathname);
 833
 834        return last_exclude_matching_from_lists(dir, pathname, pathlen,
 835                        basename, dtype_p);
 836}
 837
 838/*
 839 * Loads the exclude lists for the directory containing pathname, then
 840 * scans all exclude lists to determine whether pathname is excluded.
 841 * Returns 1 if true, otherwise 0.
 842 */
 843static int is_excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
 844{
 845        struct exclude *exclude =
 846                last_exclude_matching(dir, pathname, dtype_p);
 847        if (exclude)
 848                return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
 849        return 0;
 850}
 851
 852void path_exclude_check_init(struct path_exclude_check *check,
 853                             struct dir_struct *dir)
 854{
 855        check->dir = dir;
 856        check->exclude = NULL;
 857        strbuf_init(&check->path, 256);
 858}
 859
 860void path_exclude_check_clear(struct path_exclude_check *check)
 861{
 862        strbuf_release(&check->path);
 863}
 864
 865/*
 866 * For each subdirectory in name, starting with the top-most, checks
 867 * to see if that subdirectory is excluded, and if so, returns the
 868 * corresponding exclude structure.  Otherwise, checks whether name
 869 * itself (which is presumably a file) is excluded.
 870 *
 871 * A path to a directory known to be excluded is left in check->path to
 872 * optimize for repeated checks for files in the same excluded directory.
 873 */
 874struct exclude *last_exclude_matching_path(struct path_exclude_check *check,
 875                                           const char *name, int namelen,
 876                                           int *dtype)
 877{
 878        int i;
 879        struct strbuf *path = &check->path;
 880        struct exclude *exclude;
 881
 882        /*
 883         * we allow the caller to pass namelen as an optimization; it
 884         * must match the length of the name, as we eventually call
 885         * is_excluded() on the whole name string.
 886         */
 887        if (namelen < 0)
 888                namelen = strlen(name);
 889
 890        /*
 891         * If path is non-empty, and name is equal to path or a
 892         * subdirectory of path, name should be excluded, because
 893         * it's inside a directory which is already known to be
 894         * excluded and was previously left in check->path.
 895         */
 896        if (path->len &&
 897            path->len <= namelen &&
 898            !memcmp(name, path->buf, path->len) &&
 899            (!name[path->len] || name[path->len] == '/'))
 900                return check->exclude;
 901
 902        strbuf_setlen(path, 0);
 903        for (i = 0; name[i]; i++) {
 904                int ch = name[i];
 905
 906                if (ch == '/') {
 907                        int dt = DT_DIR;
 908                        exclude = last_exclude_matching(check->dir,
 909                                                        path->buf, &dt);
 910                        if (exclude) {
 911                                check->exclude = exclude;
 912                                return exclude;
 913                        }
 914                }
 915                strbuf_addch(path, ch);
 916        }
 917
 918        /* An entry in the index; cannot be a directory with subentries */
 919        strbuf_setlen(path, 0);
 920
 921        return last_exclude_matching(check->dir, name, dtype);
 922}
 923
 924/*
 925 * Is this name excluded?  This is for a caller like show_files() that
 926 * do not honor directory hierarchy and iterate through paths that are
 927 * possibly in an ignored directory.
 928 */
 929int is_path_excluded(struct path_exclude_check *check,
 930                  const char *name, int namelen, int *dtype)
 931{
 932        struct exclude *exclude =
 933                last_exclude_matching_path(check, name, namelen, dtype);
 934        if (exclude)
 935                return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
 936        return 0;
 937}
 938
 939static struct dir_entry *dir_entry_new(const char *pathname, int len)
 940{
 941        struct dir_entry *ent;
 942
 943        ent = xmalloc(sizeof(*ent) + len + 1);
 944        ent->len = len;
 945        memcpy(ent->name, pathname, len);
 946        ent->name[len] = 0;
 947        return ent;
 948}
 949
 950static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
 951{
 952        if (cache_name_exists(pathname, len, ignore_case))
 953                return NULL;
 954
 955        ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
 956        return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
 957}
 958
 959struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
 960{
 961        if (!cache_name_is_other(pathname, len))
 962                return NULL;
 963
 964        ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
 965        return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
 966}
 967
 968enum exist_status {
 969        index_nonexistent = 0,
 970        index_directory,
 971        index_gitdir
 972};
 973
 974/*
 975 * Do not use the alphabetically stored index to look up
 976 * the directory name; instead, use the case insensitive
 977 * name hash.
 978 */
 979static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
 980{
 981        struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
 982        unsigned char endchar;
 983
 984        if (!ce)
 985                return index_nonexistent;
 986        endchar = ce->name[len];
 987
 988        /*
 989         * The cache_entry structure returned will contain this dirname
 990         * and possibly additional path components.
 991         */
 992        if (endchar == '/')
 993                return index_directory;
 994
 995        /*
 996         * If there are no additional path components, then this cache_entry
 997         * represents a submodule.  Submodules, despite being directories,
 998         * are stored in the cache without a closing slash.
 999         */
1000        if (!endchar && S_ISGITLINK(ce->ce_mode))
1001                return index_gitdir;
1002
1003        /* This should never be hit, but it exists just in case. */
1004        return index_nonexistent;
1005}
1006
1007/*
1008 * The index sorts alphabetically by entry name, which
1009 * means that a gitlink sorts as '\0' at the end, while
1010 * a directory (which is defined not as an entry, but as
1011 * the files it contains) will sort with the '/' at the
1012 * end.
1013 */
1014static enum exist_status directory_exists_in_index(const char *dirname, int len)
1015{
1016        int pos;
1017
1018        if (ignore_case)
1019                return directory_exists_in_index_icase(dirname, len);
1020
1021        pos = cache_name_pos(dirname, len);
1022        if (pos < 0)
1023                pos = -pos-1;
1024        while (pos < active_nr) {
1025                struct cache_entry *ce = active_cache[pos++];
1026                unsigned char endchar;
1027
1028                if (strncmp(ce->name, dirname, len))
1029                        break;
1030                endchar = ce->name[len];
1031                if (endchar > '/')
1032                        break;
1033                if (endchar == '/')
1034                        return index_directory;
1035                if (!endchar && S_ISGITLINK(ce->ce_mode))
1036                        return index_gitdir;
1037        }
1038        return index_nonexistent;
1039}
1040
1041/*
1042 * When we find a directory when traversing the filesystem, we
1043 * have three distinct cases:
1044 *
1045 *  - ignore it
1046 *  - see it as a directory
1047 *  - recurse into it
1048 *
1049 * and which one we choose depends on a combination of existing
1050 * git index contents and the flags passed into the directory
1051 * traversal routine.
1052 *
1053 * Case 1: If we *already* have entries in the index under that
1054 * directory name, we always recurse into the directory to see
1055 * all the files.
1056 *
1057 * Case 2: If we *already* have that directory name as a gitlink,
1058 * we always continue to see it as a gitlink, regardless of whether
1059 * there is an actual git directory there or not (it might not
1060 * be checked out as a subproject!)
1061 *
1062 * Case 3: if we didn't have it in the index previously, we
1063 * have a few sub-cases:
1064 *
1065 *  (a) if "show_other_directories" is true, we show it as
1066 *      just a directory, unless "hide_empty_directories" is
1067 *      also true and the directory is empty, in which case
1068 *      we just ignore it entirely.
1069 *      if we are looking for ignored directories, look if it
1070 *      contains only ignored files to decide if it must be shown as
1071 *      ignored or not.
1072 *  (b) if it looks like a git directory, and we don't have
1073 *      'no_gitlinks' set we treat it as a gitlink, and show it
1074 *      as a directory.
1075 *  (c) otherwise, we recurse into it.
1076 */
1077enum directory_treatment {
1078        show_directory,
1079        ignore_directory,
1080        recurse_into_directory
1081};
1082
1083static enum directory_treatment treat_directory(struct dir_struct *dir,
1084        const char *dirname, int len, int exclude,
1085        const struct path_simplify *simplify)
1086{
1087        /* The "len-1" is to strip the final '/' */
1088        switch (directory_exists_in_index(dirname, len-1)) {
1089        case index_directory:
1090                return recurse_into_directory;
1091
1092        case index_gitdir:
1093                if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
1094                        return ignore_directory;
1095                return show_directory;
1096
1097        case index_nonexistent:
1098                if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
1099                        break;
1100                if (!(dir->flags & DIR_NO_GITLINKS)) {
1101                        unsigned char sha1[20];
1102                        if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
1103                                return show_directory;
1104                }
1105                return recurse_into_directory;
1106        }
1107
1108        /* This is the "show_other_directories" case */
1109
1110        /* might be a sub directory in an excluded directory */
1111        if (!exclude) {
1112                struct path_exclude_check check;
1113                int dt = DT_DIR;
1114                path_exclude_check_init(&check, dir);
1115                exclude = is_path_excluded(&check, dirname, len, &dt);
1116                path_exclude_check_clear(&check);
1117        }
1118
1119        /*
1120         * We are looking for ignored files and our directory is not ignored,
1121         * check if it contains untracked files (i.e. is listed as untracked)
1122         */
1123        if ((dir->flags & DIR_SHOW_IGNORED) && !exclude) {
1124                int ignored;
1125                dir->flags &= ~DIR_SHOW_IGNORED;
1126                ignored = read_directory_recursive(dir, dirname, len, 1, simplify);
1127                dir->flags |= DIR_SHOW_IGNORED;
1128
1129                if (ignored)
1130                        return ignore_directory;
1131        }
1132
1133        if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
1134                return show_directory;
1135        if (!read_directory_recursive(dir, dirname, len, 1, simplify))
1136                return ignore_directory;
1137        return show_directory;
1138}
1139
1140/*
1141 * Decide what to do when we find a file while traversing the
1142 * filesystem. Mostly two cases:
1143 *
1144 *  1. We are looking for ignored files
1145 *   (a) File is ignored, include it
1146 *   (b) File is in ignored path, include it
1147 *   (c) File is not ignored, exclude it
1148 *
1149 *  2. Other scenarios, include the file if not excluded
1150 *
1151 * Return 1 for exclude, 0 for include.
1152 */
1153static int treat_file(struct dir_struct *dir, struct strbuf *path, int exclude, int *dtype)
1154{
1155        struct path_exclude_check check;
1156        int exclude_file = 0;
1157
1158        /* Always exclude indexed files */
1159        if (index_name_exists(&the_index, path->buf, path->len, ignore_case))
1160                return 1;
1161
1162        if (exclude)
1163                exclude_file = !(dir->flags & DIR_SHOW_IGNORED);
1164        else if (dir->flags & DIR_SHOW_IGNORED) {
1165                path_exclude_check_init(&check, dir);
1166
1167                if (!is_path_excluded(&check, path->buf, path->len, dtype))
1168                        exclude_file = 1;
1169
1170                path_exclude_check_clear(&check);
1171        }
1172
1173        return exclude_file;
1174}
1175
1176/*
1177 * This is an inexact early pruning of any recursive directory
1178 * reading - if the path cannot possibly be in the pathspec,
1179 * return true, and we'll skip it early.
1180 */
1181static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
1182{
1183        if (simplify) {
1184                for (;;) {
1185                        const char *match = simplify->path;
1186                        int len = simplify->len;
1187
1188                        if (!match)
1189                                break;
1190                        if (len > pathlen)
1191                                len = pathlen;
1192                        if (!memcmp(path, match, len))
1193                                return 0;
1194                        simplify++;
1195                }
1196                return 1;
1197        }
1198        return 0;
1199}
1200
1201/*
1202 * This function tells us whether an excluded path matches a
1203 * list of "interesting" pathspecs. That is, whether a path matched
1204 * by any of the pathspecs could possibly be ignored by excluding
1205 * the specified path. This can happen if:
1206 *
1207 *   1. the path is mentioned explicitly in the pathspec
1208 *
1209 *   2. the path is a directory prefix of some element in the
1210 *      pathspec
1211 */
1212static int exclude_matches_pathspec(const char *path, int len,
1213                const struct path_simplify *simplify)
1214{
1215        if (simplify) {
1216                for (; simplify->path; simplify++) {
1217                        if (len == simplify->len
1218                            && !memcmp(path, simplify->path, len))
1219                                return 1;
1220                        if (len < simplify->len
1221                            && simplify->path[len] == '/'
1222                            && !memcmp(path, simplify->path, len))
1223                                return 1;
1224                }
1225        }
1226        return 0;
1227}
1228
1229static int get_index_dtype(const char *path, int len)
1230{
1231        int pos;
1232        struct cache_entry *ce;
1233
1234        ce = cache_name_exists(path, len, 0);
1235        if (ce) {
1236                if (!ce_uptodate(ce))
1237                        return DT_UNKNOWN;
1238                if (S_ISGITLINK(ce->ce_mode))
1239                        return DT_DIR;
1240                /*
1241                 * Nobody actually cares about the
1242                 * difference between DT_LNK and DT_REG
1243                 */
1244                return DT_REG;
1245        }
1246
1247        /* Try to look it up as a directory */
1248        pos = cache_name_pos(path, len);
1249        if (pos >= 0)
1250                return DT_UNKNOWN;
1251        pos = -pos-1;
1252        while (pos < active_nr) {
1253                ce = active_cache[pos++];
1254                if (strncmp(ce->name, path, len))
1255                        break;
1256                if (ce->name[len] > '/')
1257                        break;
1258                if (ce->name[len] < '/')
1259                        continue;
1260                if (!ce_uptodate(ce))
1261                        break;  /* continue? */
1262                return DT_DIR;
1263        }
1264        return DT_UNKNOWN;
1265}
1266
1267static int get_dtype(struct dirent *de, const char *path, int len)
1268{
1269        int dtype = de ? DTYPE(de) : DT_UNKNOWN;
1270        struct stat st;
1271
1272        if (dtype != DT_UNKNOWN)
1273                return dtype;
1274        dtype = get_index_dtype(path, len);
1275        if (dtype != DT_UNKNOWN)
1276                return dtype;
1277        if (lstat(path, &st))
1278                return dtype;
1279        if (S_ISREG(st.st_mode))
1280                return DT_REG;
1281        if (S_ISDIR(st.st_mode))
1282                return DT_DIR;
1283        if (S_ISLNK(st.st_mode))
1284                return DT_LNK;
1285        return dtype;
1286}
1287
1288enum path_treatment {
1289        path_ignored,
1290        path_handled,
1291        path_recurse
1292};
1293
1294static enum path_treatment treat_one_path(struct dir_struct *dir,
1295                                          struct strbuf *path,
1296                                          const struct path_simplify *simplify,
1297                                          int dtype, struct dirent *de)
1298{
1299        int exclude = is_excluded(dir, path->buf, &dtype);
1300        if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
1301            && exclude_matches_pathspec(path->buf, path->len, simplify))
1302                dir_add_ignored(dir, path->buf, path->len);
1303
1304        /*
1305         * Excluded? If we don't explicitly want to show
1306         * ignored files, ignore it
1307         */
1308        if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
1309                return path_ignored;
1310
1311        if (dtype == DT_UNKNOWN)
1312                dtype = get_dtype(de, path->buf, path->len);
1313
1314        switch (dtype) {
1315        default:
1316                return path_ignored;
1317        case DT_DIR:
1318                strbuf_addch(path, '/');
1319                switch (treat_directory(dir, path->buf, path->len, exclude, simplify)) {
1320                case show_directory:
1321                        break;
1322                case recurse_into_directory:
1323                        return path_recurse;
1324                case ignore_directory:
1325                        return path_ignored;
1326                }
1327                break;
1328        case DT_REG:
1329        case DT_LNK:
1330                switch (treat_file(dir, path, exclude, &dtype)) {
1331                case 1:
1332                        return path_ignored;
1333                default:
1334                        break;
1335                }
1336        }
1337        return path_handled;
1338}
1339
1340static enum path_treatment treat_path(struct dir_struct *dir,
1341                                      struct dirent *de,
1342                                      struct strbuf *path,
1343                                      int baselen,
1344                                      const struct path_simplify *simplify)
1345{
1346        int dtype;
1347
1348        if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
1349                return path_ignored;
1350        strbuf_setlen(path, baselen);
1351        strbuf_addstr(path, de->d_name);
1352        if (simplify_away(path->buf, path->len, simplify))
1353                return path_ignored;
1354
1355        dtype = DTYPE(de);
1356        return treat_one_path(dir, path, simplify, dtype, de);
1357}
1358
1359/*
1360 * Read a directory tree. We currently ignore anything but
1361 * directories, regular files and symlinks. That's because git
1362 * doesn't handle them at all yet. Maybe that will change some
1363 * day.
1364 *
1365 * Also, we ignore the name ".git" (even if it is not a directory).
1366 * That likely will not change.
1367 */
1368static int read_directory_recursive(struct dir_struct *dir,
1369                                    const char *base, int baselen,
1370                                    int check_only,
1371                                    const struct path_simplify *simplify)
1372{
1373        DIR *fdir;
1374        int contents = 0;
1375        struct dirent *de;
1376        struct strbuf path = STRBUF_INIT;
1377
1378        strbuf_add(&path, base, baselen);
1379
1380        fdir = opendir(path.len ? path.buf : ".");
1381        if (!fdir)
1382                goto out;
1383
1384        while ((de = readdir(fdir)) != NULL) {
1385                switch (treat_path(dir, de, &path, baselen, simplify)) {
1386                case path_recurse:
1387                        contents += read_directory_recursive(dir, path.buf,
1388                                path.len, check_only, simplify);
1389                        continue;
1390                case path_ignored:
1391                        continue;
1392                case path_handled:
1393                        break;
1394                }
1395                contents++;
1396                if (check_only)
1397                        break;
1398                dir_add_name(dir, path.buf, path.len);
1399        }
1400        closedir(fdir);
1401 out:
1402        strbuf_release(&path);
1403
1404        return contents;
1405}
1406
1407static int cmp_name(const void *p1, const void *p2)
1408{
1409        const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1410        const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1411
1412        return cache_name_compare(e1->name, e1->len,
1413                                  e2->name, e2->len);
1414}
1415
1416static struct path_simplify *create_simplify(const char **pathspec)
1417{
1418        int nr, alloc = 0;
1419        struct path_simplify *simplify = NULL;
1420
1421        if (!pathspec)
1422                return NULL;
1423
1424        for (nr = 0 ; ; nr++) {
1425                const char *match;
1426                if (nr >= alloc) {
1427                        alloc = alloc_nr(alloc);
1428                        simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1429                }
1430                match = *pathspec++;
1431                if (!match)
1432                        break;
1433                simplify[nr].path = match;
1434                simplify[nr].len = simple_length(match);
1435        }
1436        simplify[nr].path = NULL;
1437        simplify[nr].len = 0;
1438        return simplify;
1439}
1440
1441static void free_simplify(struct path_simplify *simplify)
1442{
1443        free(simplify);
1444}
1445
1446static int treat_leading_path(struct dir_struct *dir,
1447                              const char *path, int len,
1448                              const struct path_simplify *simplify)
1449{
1450        struct strbuf sb = STRBUF_INIT;
1451        int baselen, rc = 0;
1452        const char *cp;
1453        int old_flags = dir->flags;
1454
1455        while (len && path[len - 1] == '/')
1456                len--;
1457        if (!len)
1458                return 1;
1459        baselen = 0;
1460        dir->flags &= ~DIR_SHOW_OTHER_DIRECTORIES;
1461        while (1) {
1462                cp = path + baselen + !!baselen;
1463                cp = memchr(cp, '/', path + len - cp);
1464                if (!cp)
1465                        baselen = len;
1466                else
1467                        baselen = cp - path;
1468                strbuf_setlen(&sb, 0);
1469                strbuf_add(&sb, path, baselen);
1470                if (!is_directory(sb.buf))
1471                        break;
1472                if (simplify_away(sb.buf, sb.len, simplify))
1473                        break;
1474                if (treat_one_path(dir, &sb, simplify,
1475                                   DT_DIR, NULL) == path_ignored)
1476                        break; /* do not recurse into it */
1477                if (len <= baselen) {
1478                        rc = 1;
1479                        break; /* finished checking */
1480                }
1481        }
1482        strbuf_release(&sb);
1483        dir->flags = old_flags;
1484        return rc;
1485}
1486
1487int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
1488{
1489        struct path_simplify *simplify;
1490
1491        if (has_symlink_leading_path(path, len))
1492                return dir->nr;
1493
1494        simplify = create_simplify(pathspec);
1495        if (!len || treat_leading_path(dir, path, len, simplify))
1496                read_directory_recursive(dir, path, len, 0, simplify);
1497        free_simplify(simplify);
1498        qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
1499        qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
1500        return dir->nr;
1501}
1502
1503int file_exists(const char *f)
1504{
1505        struct stat sb;
1506        return lstat(f, &sb) == 0;
1507}
1508
1509/*
1510 * Given two normalized paths (a trailing slash is ok), if subdir is
1511 * outside dir, return -1.  Otherwise return the offset in subdir that
1512 * can be used as relative path to dir.
1513 */
1514int dir_inside_of(const char *subdir, const char *dir)
1515{
1516        int offset = 0;
1517
1518        assert(dir && subdir && *dir && *subdir);
1519
1520        while (*dir && *subdir && *dir == *subdir) {
1521                dir++;
1522                subdir++;
1523                offset++;
1524        }
1525
1526        /* hel[p]/me vs hel[l]/yeah */
1527        if (*dir && *subdir)
1528                return -1;
1529
1530        if (!*subdir)
1531                return !*dir ? offset : -1; /* same dir */
1532
1533        /* foo/[b]ar vs foo/[] */
1534        if (is_dir_sep(dir[-1]))
1535                return is_dir_sep(subdir[-1]) ? offset : -1;
1536
1537        /* foo[/]bar vs foo[] */
1538        return is_dir_sep(*subdir) ? offset + 1 : -1;
1539}
1540
1541int is_inside_dir(const char *dir)
1542{
1543        char cwd[PATH_MAX];
1544        if (!dir)
1545                return 0;
1546        if (!getcwd(cwd, sizeof(cwd)))
1547                die_errno("can't find the current directory");
1548        return dir_inside_of(cwd, dir) >= 0;
1549}
1550
1551int is_empty_dir(const char *path)
1552{
1553        DIR *dir = opendir(path);
1554        struct dirent *e;
1555        int ret = 1;
1556
1557        if (!dir)
1558                return 0;
1559
1560        while ((e = readdir(dir)) != NULL)
1561                if (!is_dot_or_dotdot(e->d_name)) {
1562                        ret = 0;
1563                        break;
1564                }
1565
1566        closedir(dir);
1567        return ret;
1568}
1569
1570static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
1571{
1572        DIR *dir;
1573        struct dirent *e;
1574        int ret = 0, original_len = path->len, len, kept_down = 0;
1575        int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
1576        int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
1577        unsigned char submodule_head[20];
1578
1579        if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
1580            !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
1581                /* Do not descend and nuke a nested git work tree. */
1582                if (kept_up)
1583                        *kept_up = 1;
1584                return 0;
1585        }
1586
1587        flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
1588        dir = opendir(path->buf);
1589        if (!dir) {
1590                /* an empty dir could be removed even if it is unreadble */
1591                if (!keep_toplevel)
1592                        return rmdir(path->buf);
1593                else
1594                        return -1;
1595        }
1596        if (path->buf[original_len - 1] != '/')
1597                strbuf_addch(path, '/');
1598
1599        len = path->len;
1600        while ((e = readdir(dir)) != NULL) {
1601                struct stat st;
1602                if (is_dot_or_dotdot(e->d_name))
1603                        continue;
1604
1605                strbuf_setlen(path, len);
1606                strbuf_addstr(path, e->d_name);
1607                if (lstat(path->buf, &st))
1608                        ; /* fall thru */
1609                else if (S_ISDIR(st.st_mode)) {
1610                        if (!remove_dir_recurse(path, flag, &kept_down))
1611                                continue; /* happy */
1612                } else if (!only_empty && !unlink(path->buf))
1613                        continue; /* happy, too */
1614
1615                /* path too long, stat fails, or non-directory still exists */
1616                ret = -1;
1617                break;
1618        }
1619        closedir(dir);
1620
1621        strbuf_setlen(path, original_len);
1622        if (!ret && !keep_toplevel && !kept_down)
1623                ret = rmdir(path->buf);
1624        else if (kept_up)
1625                /*
1626                 * report the uplevel that it is not an error that we
1627                 * did not rmdir() our directory.
1628                 */
1629                *kept_up = !ret;
1630        return ret;
1631}
1632
1633int remove_dir_recursively(struct strbuf *path, int flag)
1634{
1635        return remove_dir_recurse(path, flag, NULL);
1636}
1637
1638void setup_standard_excludes(struct dir_struct *dir)
1639{
1640        const char *path;
1641        char *xdg_path;
1642
1643        dir->exclude_per_dir = ".gitignore";
1644        path = git_path("info/exclude");
1645        if (!excludes_file) {
1646                home_config_paths(NULL, &xdg_path, "ignore");
1647                excludes_file = xdg_path;
1648        }
1649        if (!access_or_warn(path, R_OK))
1650                add_excludes_from_file(dir, path);
1651        if (excludes_file && !access_or_warn(excludes_file, R_OK))
1652                add_excludes_from_file(dir, excludes_file);
1653}
1654
1655int remove_path(const char *name)
1656{
1657        char *slash;
1658
1659        if (unlink(name) && errno != ENOENT && errno != ENOTDIR)
1660                return -1;
1661
1662        slash = strrchr(name, '/');
1663        if (slash) {
1664                char *dirs = xstrdup(name);
1665                slash = dirs + (slash - name);
1666                do {
1667                        *slash = '\0';
1668                } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
1669                free(dirs);
1670        }
1671        return 0;
1672}
1673
1674static int pathspec_item_cmp(const void *a_, const void *b_)
1675{
1676        struct pathspec_item *a, *b;
1677
1678        a = (struct pathspec_item *)a_;
1679        b = (struct pathspec_item *)b_;
1680        return strcmp(a->match, b->match);
1681}
1682
1683int init_pathspec(struct pathspec *pathspec, const char **paths)
1684{
1685        const char **p = paths;
1686        int i;
1687
1688        memset(pathspec, 0, sizeof(*pathspec));
1689        if (!p)
1690                return 0;
1691        while (*p)
1692                p++;
1693        pathspec->raw = paths;
1694        pathspec->nr = p - paths;
1695        if (!pathspec->nr)
1696                return 0;
1697
1698        pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1699        for (i = 0; i < pathspec->nr; i++) {
1700                struct pathspec_item *item = pathspec->items+i;
1701                const char *path = paths[i];
1702
1703                item->match = path;
1704                item->len = strlen(path);
1705                item->flags = 0;
1706                if (limit_pathspec_to_literal()) {
1707                        item->nowildcard_len = item->len;
1708                } else {
1709                        item->nowildcard_len = simple_length(path);
1710                        if (item->nowildcard_len < item->len) {
1711                                pathspec->has_wildcard = 1;
1712                                if (path[item->nowildcard_len] == '*' &&
1713                                    no_wildcard(path + item->nowildcard_len + 1))
1714                                        item->flags |= PATHSPEC_ONESTAR;
1715                        }
1716                }
1717        }
1718
1719        qsort(pathspec->items, pathspec->nr,
1720              sizeof(struct pathspec_item), pathspec_item_cmp);
1721
1722        return 0;
1723}
1724
1725void free_pathspec(struct pathspec *pathspec)
1726{
1727        free(pathspec->items);
1728        pathspec->items = NULL;
1729}
1730
1731int limit_pathspec_to_literal(void)
1732{
1733        static int flag = -1;
1734        if (flag < 0)
1735                flag = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
1736        return flag;
1737}
1738
1739/*
1740 * Frees memory within dir which was allocated for exclude lists and
1741 * the exclude_stack.  Does not free dir itself.
1742 */
1743void clear_directory(struct dir_struct *dir)
1744{
1745        int i, j;
1746        struct exclude_list_group *group;
1747        struct exclude_list *el;
1748        struct exclude_stack *stk;
1749
1750        for (i = EXC_CMDL; i <= EXC_FILE; i++) {
1751                group = &dir->exclude_list_group[i];
1752                for (j = 0; j < group->nr; j++) {
1753                        el = &group->el[j];
1754                        if (i == EXC_DIRS)
1755                                free((char *)el->src);
1756                        clear_exclude_list(el);
1757                }
1758                free(group->el);
1759        }
1760
1761        stk = dir->exclude_stack;
1762        while (stk) {
1763                struct exclude_stack *prev = stk->prev;
1764                free(stk);
1765                stk = prev;
1766        }
1767}