dir.c

   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
  12 struct path_simplify {
  13         int len;
  14         const char *path;
  15 };
  16
  17 static int read_directory_recursive(struct dir_struct *dir, const char *path, int len,
  18         int check_only, const struct path_simplify *simplify);
  19 static int get_dtype(struct dirent *de, const char *path, int len);
  20
  21 static 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
  54 int 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  */
  86 static 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  */
 130 int 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
 157 static int no_wildcard(const char *string)
 158 {
 159         return string[strcspn(string, "*?[{\\")] == '\0';
 160 }
 161
 162 void 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
 203 static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
 204 {
 205         int pos, len;
 206         unsigned long sz;
 207         enum object_type type;
 208         void *data;
 209         struct index_state *istate = &the_index;
 210
 211         len = strlen(path);
 212         pos = index_name_pos(istate, path, len);
 213         if (pos < 0)
 214                 return NULL;
 215         if (!ce_skip_worktree(istate->cache[pos]))
 216                 return NULL;
 217         data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
 218         if (!data || type != OBJ_BLOB) {
 219                 free(data);
 220                 return NULL;
 221         }
 222         *size = xsize_t(sz);
 223         return data;
 224 }
 225
 226 int add_excludes_from_file_to_list(const char *fname,
 227                                    const char *base,
 228                                    int baselen,
 229                                    char **buf_p,
 230                                    struct exclude_list *which,
 231                                    int check_index)
 232 {
 233         struct stat st;
 234         int fd, i;
 235         size_t size;
 236         char *buf, *entry;
 237
 238         fd = open(fname, O_RDONLY);
 239         if (fd < 0 || fstat(fd, &st) < 0) {
 240                 if (0 <= fd)
 241                         close(fd);
 242                 if (!check_index ||
 243                     (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
 244                         return -1;
 245         }
 246         else {
 247                 size = xsize_t(st.st_size);
 248                 if (size == 0) {
 249                         close(fd);
 250                         return 0;
 251                 }
 252                 buf = xmalloc(size);
 253                 if (read_in_full(fd, buf, size) != size) {
 254                         close(fd);
 255                         return -1;
 256                 }
 257                 close(fd);
 258         }
 259
 260         if (buf_p)
 261                 *buf_p = buf;
 262         entry = buf;
 263         for (i = 0; i <= size; i++) {
 264                 if (i == size || buf[i] == '\n') {
 265                         if (entry != buf + i && entry[0] != '#') {
 266                                 buf[i - (i && buf[i-1] == '\r')] = 0;
 267                                 add_exclude(entry, base, baselen, which);
 268                         }
 269                         entry = buf + i + 1;
 270                 }
 271         }
 272         return 0;
 273 }
 274
 275 void add_excludes_from_file(struct dir_struct *dir, const char *fname)
 276 {
 277         if (add_excludes_from_file_to_list(fname, "", 0, NULL,
 278                                            &dir->exclude_list[EXC_FILE], 0) < 0)
 279                 die("cannot use %s as an exclude file", fname);
 280 }
 281
 282 static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
 283 {
 284         struct exclude_list *el;
 285         struct exclude_stack *stk = NULL;
 286         int current;
 287
 288         if ((!dir->exclude_per_dir) ||
 289             (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
 290                 return; /* too long a path -- ignore */
 291
 292         /* Pop the ones that are not the prefix of the path being checked. */
 293         el = &dir->exclude_list[EXC_DIRS];
 294         while ((stk = dir->exclude_stack) != NULL) {
 295                 if (stk->baselen <= baselen &&
 296                     !strncmp(dir->basebuf, base, stk->baselen))
 297                         break;
 298                 dir->exclude_stack = stk->prev;
 299                 while (stk->exclude_ix < el->nr)
 300                         free(el->excludes[--el->nr]);
 301                 free(stk->filebuf);
 302                 free(stk);
 303         }
 304
 305         /* Read from the parent directories and push them down. */
 306         current = stk ? stk->baselen : -1;
 307         while (current < baselen) {
 308                 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
 309                 const char *cp;
 310
 311                 if (current < 0) {
 312                         cp = base;
 313                         current = 0;
 314                 }
 315                 else {
 316                         cp = strchr(base + current + 1, '/');
 317                         if (!cp)
 318                                 die("oops in prep_exclude");
 319                         cp++;
 320                 }
 321                 stk->prev = dir->exclude_stack;
 322                 stk->baselen = cp - base;
 323                 stk->exclude_ix = el->nr;
 324                 memcpy(dir->basebuf + current, base + current,
 325                        stk->baselen - current);
 326                 strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
 327                 add_excludes_from_file_to_list(dir->basebuf,
 328                                                dir->basebuf, stk->baselen,
 329                                                &stk->filebuf, el, 1);
 330                 dir->exclude_stack = stk;
 331                 current = stk->baselen;
 332         }
 333         dir->basebuf[baselen] = '\0';
 334 }
 335
 336 /* Scan the list and let the last match determine the fate.
 337  * Return 1 for exclude, 0 for include and -1 for undecided.
 338  */
 339 int excluded_from_list(const char *pathname,
 340                        int pathlen, const char *basename, int *dtype,
 341                        struct exclude_list *el)
 342 {
 343         int i;
 344
 345         if (el->nr) {
 346                 for (i = el->nr - 1; 0 <= i; i--) {
 347                         struct exclude *x = el->excludes[i];
 348                         const char *exclude = x->pattern;
 349                         int to_exclude = x->to_exclude;
 350
 351                         if (x->flags & EXC_FLAG_MUSTBEDIR) {
 352                                 if (!dtype) {
 353                                         if (!prefixcmp(pathname, exclude))
 354                                                 return to_exclude;
 355                                         else
 356                                                 continue;
 357                                 }
 358                                 if (*dtype == DT_UNKNOWN)
 359                                         *dtype = get_dtype(NULL, pathname, pathlen);
 360                                 if (*dtype != DT_DIR)
 361                                         continue;
 362                         }
 363
 364                         if (x->flags & EXC_FLAG_NODIR) {
 365                                 /* match basename */
 366                                 if (x->flags & EXC_FLAG_NOWILDCARD) {
 367                                         if (!strcmp(exclude, basename))
 368                                                 return to_exclude;
 369                                 } else if (x->flags & EXC_FLAG_ENDSWITH) {
 370                                         if (x->patternlen - 1 <= pathlen &&
 371                                             !strcmp(exclude + 1, pathname + pathlen - x->patternlen + 1))
 372                                                 return to_exclude;
 373                                 } else {
 374                                         if (fnmatch(exclude, basename, 0) == 0)
 375                                                 return to_exclude;
 376                                 }
 377                         }
 378                         else {
 379                                 /* match with FNM_PATHNAME:
 380                                  * exclude has base (baselen long) implicitly
 381                                  * in front of it.
 382                                  */
 383                                 int baselen = x->baselen;
 384                                 if (*exclude == '/')
 385                                         exclude++;
 386
 387                                 if (pathlen < baselen ||
 388                                     (baselen && pathname[baselen-1] != '/') ||
 389                                     strncmp(pathname, x->base, baselen))
 390                                     continue;
 391
 392                                 if (x->flags & EXC_FLAG_NOWILDCARD) {
 393                                         if (!strcmp(exclude, pathname + baselen))
 394                                                 return to_exclude;
 395                                 } else {
 396                                         if (fnmatch(exclude, pathname+baselen,
 397                                                     FNM_PATHNAME) == 0)
 398                                             return to_exclude;
 399                                 }
 400                         }
 401                 }
 402         }
 403         return -1; /* undecided */
 404 }
 405
 406 int excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
 407 {
 408         int pathlen = strlen(pathname);
 409         int st;
 410         const char *basename = strrchr(pathname, '/');
 411         basename = (basename) ? basename+1 : pathname;
 412
 413         prep_exclude(dir, pathname, basename-pathname);
 414         for (st = EXC_CMDL; st <= EXC_FILE; st++) {
 415                 switch (excluded_from_list(pathname, pathlen, basename,
 416                                            dtype_p, &dir->exclude_list[st])) {
 417                 case 0:
 418                         return 0;
 419                 case 1:
 420                         return 1;
 421                 }
 422         }
 423         return 0;
 424 }
 425
 426 static struct dir_entry *dir_entry_new(const char *pathname, int len)
 427 {
 428         struct dir_entry *ent;
 429
 430         ent = xmalloc(sizeof(*ent) + len + 1);
 431         ent->len = len;
 432         memcpy(ent->name, pathname, len);
 433         ent->name[len] = 0;
 434         return ent;
 435 }
 436
 437 static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
 438 {
 439         if (cache_name_exists(pathname, len, ignore_case))
 440                 return NULL;
 441
 442         ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
 443         return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
 444 }
 445
 446 static struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
 447 {
 448         if (!cache_name_is_other(pathname, len))
 449                 return NULL;
 450
 451         ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
 452         return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
 453 }
 454
 455 enum exist_status {
 456         index_nonexistent = 0,
 457         index_directory,
 458         index_gitdir,
 459 };
 460
 461 /*
 462  * The index sorts alphabetically by entry name, which
 463  * means that a gitlink sorts as '\0' at the end, while
 464  * a directory (which is defined not as an entry, but as
 465  * the files it contains) will sort with the '/' at the
 466  * end.
 467  */
 468 static enum exist_status directory_exists_in_index(const char *dirname, int len)
 469 {
 470         int pos = cache_name_pos(dirname, len);
 471         if (pos < 0)
 472                 pos = -pos-1;
 473         while (pos < active_nr) {
 474                 struct cache_entry *ce = active_cache[pos++];
 475                 unsigned char endchar;
 476
 477                 if (strncmp(ce->name, dirname, len))
 478                         break;
 479                 endchar = ce->name[len];
 480                 if (endchar > '/')
 481                         break;
 482                 if (endchar == '/')
 483                         return index_directory;
 484                 if (!endchar && S_ISGITLINK(ce->ce_mode))
 485                         return index_gitdir;
 486         }
 487         return index_nonexistent;
 488 }
 489
 490 /*
 491  * When we find a directory when traversing the filesystem, we
 492  * have three distinct cases:
 493  *
 494  *  - ignore it
 495  *  - see it as a directory
 496  *  - recurse into it
 497  *
 498  * and which one we choose depends on a combination of existing
 499  * git index contents and the flags passed into the directory
 500  * traversal routine.
 501  *
 502  * Case 1: If we *already* have entries in the index under that
 503  * directory name, we always recurse into the directory to see
 504  * all the files.
 505  *
 506  * Case 2: If we *already* have that directory name as a gitlink,
 507  * we always continue to see it as a gitlink, regardless of whether
 508  * there is an actual git directory there or not (it might not
 509  * be checked out as a subproject!)
 510  *
 511  * Case 3: if we didn't have it in the index previously, we
 512  * have a few sub-cases:
 513  *
 514  *  (a) if "show_other_directories" is true, we show it as
 515  *      just a directory, unless "hide_empty_directories" is
 516  *      also true and the directory is empty, in which case
 517  *      we just ignore it entirely.
 518  *  (b) if it looks like a git directory, and we don't have
 519  *      'no_gitlinks' set we treat it as a gitlink, and show it
 520  *      as a directory.
 521  *  (c) otherwise, we recurse into it.
 522  */
 523 enum directory_treatment {
 524         show_directory,
 525         ignore_directory,
 526         recurse_into_directory,
 527 };
 528
 529 static enum directory_treatment treat_directory(struct dir_struct *dir,
 530         const char *dirname, int len,
 531         const struct path_simplify *simplify)
 532 {
 533         /* The "len-1" is to strip the final '/' */
 534         switch (directory_exists_in_index(dirname, len-1)) {
 535         case index_directory:
 536                 return recurse_into_directory;
 537
 538         case index_gitdir:
 539                 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 540                         return ignore_directory;
 541                 return show_directory;
 542
 543         case index_nonexistent:
 544                 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 545                         break;
 546                 if (!(dir->flags & DIR_NO_GITLINKS)) {
 547                         unsigned char sha1[20];
 548                         if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
 549                                 return show_directory;
 550                 }
 551                 return recurse_into_directory;
 552         }
 553
 554         /* This is the "show_other_directories" case */
 555         if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
 556                 return show_directory;
 557         if (!read_directory_recursive(dir, dirname, len, 1, simplify))
 558                 return ignore_directory;
 559         return show_directory;
 560 }
 561
 562 /*
 563  * This is an inexact early pruning of any recursive directory
 564  * reading - if the path cannot possibly be in the pathspec,
 565  * return true, and we'll skip it early.
 566  */
 567 static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
 568 {
 569         if (simplify) {
 570                 for (;;) {
 571                         const char *match = simplify->path;
 572                         int len = simplify->len;
 573
 574                         if (!match)
 575                                 break;
 576                         if (len > pathlen)
 577                                 len = pathlen;
 578                         if (!memcmp(path, match, len))
 579                                 return 0;
 580                         simplify++;
 581                 }
 582                 return 1;
 583         }
 584         return 0;
 585 }
 586
 587 static int in_pathspec(const char *path, int len, const struct path_simplify *simplify)
 588 {
 589         if (simplify) {
 590                 for (; simplify->path; simplify++) {
 591                         if (len == simplify->len
 592                             && !memcmp(path, simplify->path, len))
 593                                 return 1;
 594                 }
 595         }
 596         return 0;
 597 }
 598
 599 static int get_index_dtype(const char *path, int len)
 600 {
 601         int pos;
 602         struct cache_entry *ce;
 603
 604         ce = cache_name_exists(path, len, 0);
 605         if (ce) {
 606                 if (!ce_uptodate(ce))
 607                         return DT_UNKNOWN;
 608                 if (S_ISGITLINK(ce->ce_mode))
 609                         return DT_DIR;
 610                 /*
 611                  * Nobody actually cares about the
 612                  * difference between DT_LNK and DT_REG
 613                  */
 614                 return DT_REG;
 615         }
 616
 617         /* Try to look it up as a directory */
 618         pos = cache_name_pos(path, len);
 619         if (pos >= 0)
 620                 return DT_UNKNOWN;
 621         pos = -pos-1;
 622         while (pos < active_nr) {
 623                 ce = active_cache[pos++];
 624                 if (strncmp(ce->name, path, len))
 625                         break;
 626                 if (ce->name[len] > '/')
 627                         break;
 628                 if (ce->name[len] < '/')
 629                         continue;
 630                 if (!ce_uptodate(ce))
 631                         break;  /* continue? */
 632                 return DT_DIR;
 633         }
 634         return DT_UNKNOWN;
 635 }
 636
 637 static int get_dtype(struct dirent *de, const char *path, int len)
 638 {
 639         int dtype = de ? DTYPE(de) : DT_UNKNOWN;
 640         struct stat st;
 641
 642         if (dtype != DT_UNKNOWN)
 643                 return dtype;
 644         dtype = get_index_dtype(path, len);
 645         if (dtype != DT_UNKNOWN)
 646                 return dtype;
 647         if (lstat(path, &st))
 648                 return dtype;
 649         if (S_ISREG(st.st_mode))
 650                 return DT_REG;
 651         if (S_ISDIR(st.st_mode))
 652                 return DT_DIR;
 653         if (S_ISLNK(st.st_mode))
 654                 return DT_LNK;
 655         return dtype;
 656 }
 657
 658 enum path_treatment {
 659         path_ignored,
 660         path_handled,
 661         path_recurse,
 662 };
 663
 664 static enum path_treatment treat_one_path(struct dir_struct *dir,
 665                                           char *path, int *len,
 666                                           const struct path_simplify *simplify,
 667                                           int dtype, struct dirent *de)
 668 {
 669         int exclude = excluded(dir, path, &dtype);
 670         if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
 671             && in_pathspec(path, *len, simplify))
 672                 dir_add_ignored(dir, path, *len);
 673
 674         /*
 675          * Excluded? If we don't explicitly want to show
 676          * ignored files, ignore it
 677          */
 678         if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
 679                 return path_ignored;
 680
 681         if (dtype == DT_UNKNOWN)
 682                 dtype = get_dtype(de, path, *len);
 683
 684         /*
 685          * Do we want to see just the ignored files?
 686          * We still need to recurse into directories,
 687          * even if we don't ignore them, since the
 688          * directory may contain files that we do..
 689          */
 690         if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) {
 691                 if (dtype != DT_DIR)
 692                         return path_ignored;
 693         }
 694
 695         switch (dtype) {
 696         default:
 697                 return path_ignored;
 698         case DT_DIR:
 699                 memcpy(path + *len, "/", 2);
 700                 (*len)++;
 701                 switch (treat_directory(dir, path, *len, simplify)) {
 702                 case show_directory:
 703                         if (exclude != !!(dir->flags
 704                                           & DIR_SHOW_IGNORED))
 705                                 return path_ignored;
 706                         break;
 707                 case recurse_into_directory:
 708                         return path_recurse;
 709                 case ignore_directory:
 710                         return path_ignored;
 711                 }
 712                 break;
 713         case DT_REG:
 714         case DT_LNK:
 715                 break;
 716         }
 717         return path_handled;
 718 }
 719
 720 static enum path_treatment treat_path(struct dir_struct *dir,
 721                                       struct dirent *de,
 722                                       char *path, int path_max,
 723                                       int baselen,
 724                                       const struct path_simplify *simplify,
 725                                       int *len)
 726 {
 727         int dtype;
 728
 729         if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
 730                 return path_ignored;
 731         *len = strlen(de->d_name);
 732         /* Ignore overly long pathnames! */
 733         if (*len + baselen + 8 > path_max)
 734                 return path_ignored;
 735         memcpy(path + baselen, de->d_name, *len + 1);
 736         *len += baselen;
 737         if (simplify_away(path, *len, simplify))
 738                 return path_ignored;
 739
 740         dtype = DTYPE(de);
 741         return treat_one_path(dir, path, len, simplify, dtype, de);
 742 }
 743
 744 /*
 745  * Read a directory tree. We currently ignore anything but
 746  * directories, regular files and symlinks. That's because git
 747  * doesn't handle them at all yet. Maybe that will change some
 748  * day.
 749  *
 750  * Also, we ignore the name ".git" (even if it is not a directory).
 751  * That likely will not change.
 752  */
 753 static int read_directory_recursive(struct dir_struct *dir,
 754                                     const char *base, int baselen,
 755                                     int check_only,
 756                                     const struct path_simplify *simplify)
 757 {
 758         DIR *fdir = opendir(*base ? base : ".");
 759         int contents = 0;
 760
 761         if (fdir) {
 762                 struct dirent *de;
 763                 char path[PATH_MAX + 1];
 764                 memcpy(path, base, baselen);
 765
 766                 while ((de = readdir(fdir)) != NULL) {
 767                         int len;
 768                         switch (treat_path(dir, de, path, sizeof(path),
 769                                            baselen, simplify, &len)) {
 770                         case path_recurse:
 771                                 contents += read_directory_recursive
 772                                         (dir, path, len, 0, simplify);
 773                                 continue;
 774                         case path_ignored:
 775                                 continue;
 776                         case path_handled:
 777                                 break;
 778                         }
 779                         contents++;
 780                         if (check_only)
 781                                 goto exit_early;
 782                         else
 783                                 dir_add_name(dir, path, len);
 784                 }
 785 exit_early:
 786                 closedir(fdir);
 787         }
 788
 789         return contents;
 790 }
 791
 792 static int cmp_name(const void *p1, const void *p2)
 793 {
 794         const struct dir_entry *e1 = *(const struct dir_entry **)p1;
 795         const struct dir_entry *e2 = *(const struct dir_entry **)p2;
 796
 797         return cache_name_compare(e1->name, e1->len,
 798                                   e2->name, e2->len);
 799 }
 800
 801 /*
 802  * Return the length of the "simple" part of a path match limiter.
 803  */
 804 static int simple_length(const char *match)
 805 {
 806         int len = -1;
 807
 808         for (;;) {
 809                 unsigned char c = *match++;
 810                 len++;
 811                 if (c == '\0' || is_glob_special(c))
 812                         return len;
 813         }
 814 }
 815
 816 static struct path_simplify *create_simplify(const char **pathspec)
 817 {
 818         int nr, alloc = 0;
 819         struct path_simplify *simplify = NULL;
 820
 821         if (!pathspec)
 822                 return NULL;
 823
 824         for (nr = 0 ; ; nr++) {
 825                 const char *match;
 826                 if (nr >= alloc) {
 827                         alloc = alloc_nr(alloc);
 828                         simplify = xrealloc(simplify, alloc * sizeof(*simplify));
 829                 }
 830                 match = *pathspec++;
 831                 if (!match)
 832                         break;
 833                 simplify[nr].path = match;
 834                 simplify[nr].len = simple_length(match);
 835         }
 836         simplify[nr].path = NULL;
 837         simplify[nr].len = 0;
 838         return simplify;
 839 }
 840
 841 static void free_simplify(struct path_simplify *simplify)
 842 {
 843         free(simplify);
 844 }
 845
 846 static int treat_leading_path(struct dir_struct *dir,
 847                               const char *path, int len,
 848                               const struct path_simplify *simplify)
 849 {
 850         char pathbuf[PATH_MAX];
 851         int baselen, blen;
 852         const char *cp;
 853
 854         while (len && path[len - 1] == '/')
 855                 len--;
 856         if (!len)
 857                 return 1;
 858         baselen = 0;
 859         while (1) {
 860                 cp = path + baselen + !!baselen;
 861                 cp = memchr(cp, '/', path + len - cp);
 862                 if (!cp)
 863                         baselen = len;
 864                 else
 865                         baselen = cp - path;
 866                 memcpy(pathbuf, path, baselen);
 867                 pathbuf[baselen] = '\0';
 868                 if (!is_directory(pathbuf))
 869                         return 0;
 870                 if (simplify_away(pathbuf, baselen, simplify))
 871                         return 0;
 872                 blen = baselen;
 873                 if (treat_one_path(dir, pathbuf, &blen, simplify,
 874                                    DT_DIR, NULL) == path_ignored)
 875                         return 0; /* do not recurse into it */
 876                 if (len <= baselen)
 877                         return 1; /* finished checking */
 878         }
 879 }
 880
 881 int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
 882 {
 883         struct path_simplify *simplify;
 884
 885         if (has_symlink_leading_path(path, len))
 886                 return dir->nr;
 887
 888         simplify = create_simplify(pathspec);
 889         if (!len || treat_leading_path(dir, path, len, simplify))
 890                 read_directory_recursive(dir, path, len, 0, simplify);
 891         free_simplify(simplify);
 892         qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
 893         qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
 894         return dir->nr;
 895 }
 896
 897 int file_exists(const char *f)
 898 {
 899         struct stat sb;
 900         return lstat(f, &sb) == 0;
 901 }
 902
 903 /*
 904  * get_relative_cwd() gets the prefix of the current working directory
 905  * relative to 'dir'.  If we are not inside 'dir', it returns NULL.
 906  *
 907  * As a convenience, it also returns NULL if 'dir' is already NULL.  The
 908  * reason for this behaviour is that it is natural for functions returning
 909  * directory names to return NULL to say "this directory does not exist"
 910  * or "this directory is invalid".  These cases are usually handled the
 911  * same as if the cwd is not inside 'dir' at all, so get_relative_cwd()
 912  * returns NULL for both of them.
 913  *
 914  * Most notably, get_relative_cwd(buffer, size, get_git_work_tree())
 915  * unifies the handling of "outside work tree" with "no work tree at all".
 916  */
 917 char *get_relative_cwd(char *buffer, int size, const char *dir)
 918 {
 919         char *cwd = buffer;
 920
 921         if (!dir)
 922                 return NULL;
 923         if (!getcwd(buffer, size))
 924                 die_errno("can't find the current directory");
 925
 926         if (!is_absolute_path(dir))
 927                 dir = make_absolute_path(dir);
 928
 929         while (*dir && *dir == *cwd) {
 930                 dir++;
 931                 cwd++;
 932         }
 933         if (*dir)
 934                 return NULL;
 935         if (*cwd == '/')
 936                 return cwd + 1;
 937         return cwd;
 938 }
 939
 940 int is_inside_dir(const char *dir)
 941 {
 942         char buffer[PATH_MAX];
 943         return get_relative_cwd(buffer, sizeof(buffer), dir) != NULL;
 944 }
 945
 946 int is_empty_dir(const char *path)
 947 {
 948         DIR *dir = opendir(path);
 949         struct dirent *e;
 950         int ret = 1;
 951
 952         if (!dir)
 953                 return 0;
 954
 955         while ((e = readdir(dir)) != NULL)
 956                 if (!is_dot_or_dotdot(e->d_name)) {
 957                         ret = 0;
 958                         break;
 959                 }
 960
 961         closedir(dir);
 962         return ret;
 963 }
 964
 965 int remove_dir_recursively(struct strbuf *path, int flag)
 966 {
 967         DIR *dir;
 968         struct dirent *e;
 969         int ret = 0, original_len = path->len, len;
 970         int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
 971         unsigned char submodule_head[20];
 972
 973         if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
 974             !resolve_gitlink_ref(path->buf, "HEAD", submodule_head))
 975                 /* Do not descend and nuke a nested git work tree. */
 976                 return 0;
 977
 978         dir = opendir(path->buf);
 979         if (!dir)
 980                 return -1;
 981         if (path->buf[original_len - 1] != '/')
 982                 strbuf_addch(path, '/');
 983
 984         len = path->len;
 985         while ((e = readdir(dir)) != NULL) {
 986                 struct stat st;
 987                 if (is_dot_or_dotdot(e->d_name))
 988                         continue;
 989
 990                 strbuf_setlen(path, len);
 991                 strbuf_addstr(path, e->d_name);
 992                 if (lstat(path->buf, &st))
 993                         ; /* fall thru */
 994                 else if (S_ISDIR(st.st_mode)) {
 995                         if (!remove_dir_recursively(path, only_empty))
 996                                 continue; /* happy */
 997                 } else if (!only_empty && !unlink(path->buf))
 998                         continue; /* happy, too */
 999
1000                 /* path too long, stat fails, or non-directory still exists */
1001                 ret = -1;
1002                 break;
1003         }
1004         closedir(dir);
1005
1006         strbuf_setlen(path, original_len);
1007         if (!ret)
1008                 ret = rmdir(path->buf);
1009         return ret;
1010 }
1011
1012 void setup_standard_excludes(struct dir_struct *dir)
1013 {
1014         const char *path;
1015
1016         dir->exclude_per_dir = ".gitignore";
1017         path = git_path("info/exclude");
1018         if (!access(path, R_OK))
1019                 add_excludes_from_file(dir, path);
1020         if (excludes_file && !access(excludes_file, R_OK))
1021                 add_excludes_from_file(dir, excludes_file);
1022 }
1023
1024 int remove_path(const char *name)
1025 {
1026         char *slash;
1027
1028         if (unlink(name) && errno != ENOENT)
1029                 return -1;
1030
1031         slash = strrchr(name, '/');
1032         if (slash) {
1033                 char *dirs = xstrdup(name);
1034                 slash = dirs + (slash - name);
1035                 do {
1036                         *slash = '\0';
1037                 } while (rmdir(dirs) && (slash = strrchr(dirs, '/')));
1038                 free(dirs);
1039         }
1040         return 0;
1041 }
1042