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 /* helper string functions with support for the ignore_case flag */
  22 int strcmp_icase(const char *a, const char *b)
  23 {
  24         return ignore_case ? strcasecmp(a, b) : strcmp(a, b);
  25 }
  26
  27 int strncmp_icase(const char *a, const char *b, size_t count)
  28 {
  29         return ignore_case ? strncasecmp(a, b, count) : strncmp(a, b, count);
  30 }
  31
  32 int fnmatch_icase(const char *pattern, const char *string, int flags)
  33 {
  34         return fnmatch(pattern, string, flags | (ignore_case ? FNM_CASEFOLD : 0));
  35 }
  36
  37 static size_t common_prefix_len(const char **pathspec)
  38 {
  39         const char *n, *first;
  40         size_t max = 0;
  41
  42         if (!pathspec)
  43                 return max;
  44
  45         first = *pathspec;
  46         while ((n = *pathspec++)) {
  47                 size_t i, len = 0;
  48                 for (i = 0; first == n || i < max; i++) {
  49                         char c = n[i];
  50                         if (!c || c != first[i] || is_glob_special(c))
  51                                 break;
  52                         if (c == '/')
  53                                 len = i + 1;
  54                 }
  55                 if (first == n || len < max) {
  56                         max = len;
  57                         if (!max)
  58                                 break;
  59                 }
  60         }
  61         return max;
  62 }
  63
  64 /*
  65  * Returns a copy of the longest leading path common among all
  66  * pathspecs.
  67  */
  68 char *common_prefix(const char **pathspec)
  69 {
  70         unsigned long len = common_prefix_len(pathspec);
  71
  72         return len ? xmemdupz(*pathspec, len) : NULL;
  73 }
  74
  75 int fill_directory(struct dir_struct *dir, const char **pathspec)
  76 {
  77         size_t len;
  78
  79         /*
  80          * Calculate common prefix for the pathspec, and
  81          * use that to optimize the directory walk
  82          */
  83         len = common_prefix_len(pathspec);
  84
  85         /* Read the directory and prune it */
  86         read_directory(dir, pathspec ? *pathspec : "", len, pathspec);
  87         return len;
  88 }
  89
  90 int within_depth(const char *name, int namelen,
  91                         int depth, int max_depth)
  92 {
  93         const char *cp = name, *cpe = name + namelen;
  94
  95         while (cp < cpe) {
  96                 if (*cp++ != '/')
  97                         continue;
  98                 depth++;
  99                 if (depth > max_depth)
 100                         return 0;
 101         }
 102         return 1;
 103 }
 104
 105 /*
 106  * Does 'match' match the given name?
 107  * A match is found if
 108  *
 109  * (1) the 'match' string is leading directory of 'name', or
 110  * (2) the 'match' string is a wildcard and matches 'name', or
 111  * (3) the 'match' string is exactly the same as 'name'.
 112  *
 113  * and the return value tells which case it was.
 114  *
 115  * It returns 0 when there is no match.
 116  */
 117 static int match_one(const char *match, const char *name, int namelen)
 118 {
 119         int matchlen;
 120
 121         /* If the match was just the prefix, we matched */
 122         if (!*match)
 123                 return MATCHED_RECURSIVELY;
 124
 125         if (ignore_case) {
 126                 for (;;) {
 127                         unsigned char c1 = tolower(*match);
 128                         unsigned char c2 = tolower(*name);
 129                         if (c1 == '\0' || is_glob_special(c1))
 130                                 break;
 131                         if (c1 != c2)
 132                                 return 0;
 133                         match++;
 134                         name++;
 135                         namelen--;
 136                 }
 137         } else {
 138                 for (;;) {
 139                         unsigned char c1 = *match;
 140                         unsigned char c2 = *name;
 141                         if (c1 == '\0' || is_glob_special(c1))
 142                                 break;
 143                         if (c1 != c2)
 144                                 return 0;
 145                         match++;
 146                         name++;
 147                         namelen--;
 148                 }
 149         }
 150
 151
 152         /*
 153          * If we don't match the matchstring exactly,
 154          * we need to match by fnmatch
 155          */
 156         matchlen = strlen(match);
 157         if (strncmp_icase(match, name, matchlen))
 158                 return !fnmatch_icase(match, name, 0) ? MATCHED_FNMATCH : 0;
 159
 160         if (namelen == matchlen)
 161                 return MATCHED_EXACTLY;
 162         if (match[matchlen-1] == '/' || name[matchlen] == '/')
 163                 return MATCHED_RECURSIVELY;
 164         return 0;
 165 }
 166
 167 /*
 168  * Given a name and a list of pathspecs, see if the name matches
 169  * any of the pathspecs.  The caller is also interested in seeing
 170  * all pathspec matches some names it calls this function with
 171  * (otherwise the user could have mistyped the unmatched pathspec),
 172  * and a mark is left in seen[] array for pathspec element that
 173  * actually matched anything.
 174  */
 175 int match_pathspec(const char **pathspec, const char *name, int namelen,
 176                 int prefix, char *seen)
 177 {
 178         int i, retval = 0;
 179
 180         if (!pathspec)
 181                 return 1;
 182
 183         name += prefix;
 184         namelen -= prefix;
 185
 186         for (i = 0; pathspec[i] != NULL; i++) {
 187                 int how;
 188                 const char *match = pathspec[i] + prefix;
 189                 if (seen && seen[i] == MATCHED_EXACTLY)
 190                         continue;
 191                 how = match_one(match, name, namelen);
 192                 if (how) {
 193                         if (retval < how)
 194                                 retval = how;
 195                         if (seen && seen[i] < how)
 196                                 seen[i] = how;
 197                 }
 198         }
 199         return retval;
 200 }
 201
 202 /*
 203  * Does 'match' match the given name?
 204  * A match is found if
 205  *
 206  * (1) the 'match' string is leading directory of 'name', or
 207  * (2) the 'match' string is a wildcard and matches 'name', or
 208  * (3) the 'match' string is exactly the same as 'name'.
 209  *
 210  * and the return value tells which case it was.
 211  *
 212  * It returns 0 when there is no match.
 213  */
 214 static int match_pathspec_item(const struct pathspec_item *item, int prefix,
 215                                const char *name, int namelen)
 216 {
 217         /* name/namelen has prefix cut off by caller */
 218         const char *match = item->match + prefix;
 219         int matchlen = item->len - prefix;
 220
 221         /* If the match was just the prefix, we matched */
 222         if (!*match)
 223                 return MATCHED_RECURSIVELY;
 224
 225         if (matchlen <= namelen && !strncmp(match, name, matchlen)) {
 226                 if (matchlen == namelen)
 227                         return MATCHED_EXACTLY;
 228
 229                 if (match[matchlen-1] == '/' || name[matchlen] == '/')
 230                         return MATCHED_RECURSIVELY;
 231         }
 232
 233         if (item->use_wildcard && !fnmatch(match, name, 0))
 234                 return MATCHED_FNMATCH;
 235
 236         return 0;
 237 }
 238
 239 /*
 240  * Given a name and a list of pathspecs, see if the name matches
 241  * any of the pathspecs.  The caller is also interested in seeing
 242  * all pathspec matches some names it calls this function with
 243  * (otherwise the user could have mistyped the unmatched pathspec),
 244  * and a mark is left in seen[] array for pathspec element that
 245  * actually matched anything.
 246  */
 247 int match_pathspec_depth(const struct pathspec *ps,
 248                          const char *name, int namelen,
 249                          int prefix, char *seen)
 250 {
 251         int i, retval = 0;
 252
 253         if (!ps->nr) {
 254                 if (!ps->recursive || ps->max_depth == -1)
 255                         return MATCHED_RECURSIVELY;
 256
 257                 if (within_depth(name, namelen, 0, ps->max_depth))
 258                         return MATCHED_EXACTLY;
 259                 else
 260                         return 0;
 261         }
 262
 263         name += prefix;
 264         namelen -= prefix;
 265
 266         for (i = ps->nr - 1; i >= 0; i--) {
 267                 int how;
 268                 if (seen && seen[i] == MATCHED_EXACTLY)
 269                         continue;
 270                 how = match_pathspec_item(ps->items+i, prefix, name, namelen);
 271                 if (ps->recursive && ps->max_depth != -1 &&
 272                     how && how != MATCHED_FNMATCH) {
 273                         int len = ps->items[i].len;
 274                         if (name[len] == '/')
 275                                 len++;
 276                         if (within_depth(name+len, namelen-len, 0, ps->max_depth))
 277                                 how = MATCHED_EXACTLY;
 278                         else
 279                                 how = 0;
 280                 }
 281                 if (how) {
 282                         if (retval < how)
 283                                 retval = how;
 284                         if (seen && seen[i] < how)
 285                                 seen[i] = how;
 286                 }
 287         }
 288         return retval;
 289 }
 290
 291 static int no_wildcard(const char *string)
 292 {
 293         return string[strcspn(string, "*?[{\\")] == '\0';
 294 }
 295
 296 void add_exclude(const char *string, const char *base,
 297                  int baselen, struct exclude_list *which)
 298 {
 299         struct exclude *x;
 300         size_t len;
 301         int to_exclude = 1;
 302         int flags = 0;
 303
 304         if (*string == '!') {
 305                 to_exclude = 0;
 306                 string++;
 307         }
 308         len = strlen(string);
 309         if (len && string[len - 1] == '/') {
 310                 char *s;
 311                 x = xmalloc(sizeof(*x) + len);
 312                 s = (char *)(x+1);
 313                 memcpy(s, string, len - 1);
 314                 s[len - 1] = '\0';
 315                 string = s;
 316                 x->pattern = s;
 317                 flags = EXC_FLAG_MUSTBEDIR;
 318         } else {
 319                 x = xmalloc(sizeof(*x));
 320                 x->pattern = string;
 321         }
 322         x->to_exclude = to_exclude;
 323         x->patternlen = strlen(string);
 324         x->base = base;
 325         x->baselen = baselen;
 326         x->flags = flags;
 327         if (!strchr(string, '/'))
 328                 x->flags |= EXC_FLAG_NODIR;
 329         if (no_wildcard(string))
 330                 x->flags |= EXC_FLAG_NOWILDCARD;
 331         if (*string == '*' && no_wildcard(string+1))
 332                 x->flags |= EXC_FLAG_ENDSWITH;
 333         ALLOC_GROW(which->excludes, which->nr + 1, which->alloc);
 334         which->excludes[which->nr++] = x;
 335 }
 336
 337 static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
 338 {
 339         int pos, len;
 340         unsigned long sz;
 341         enum object_type type;
 342         void *data;
 343         struct index_state *istate = &the_index;
 344
 345         len = strlen(path);
 346         pos = index_name_pos(istate, path, len);
 347         if (pos < 0)
 348                 return NULL;
 349         if (!ce_skip_worktree(istate->cache[pos]))
 350                 return NULL;
 351         data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
 352         if (!data || type != OBJ_BLOB) {
 353                 free(data);
 354                 return NULL;
 355         }
 356         *size = xsize_t(sz);
 357         return data;
 358 }
 359
 360 void free_excludes(struct exclude_list *el)
 361 {
 362         int i;
 363
 364         for (i = 0; i < el->nr; i++)
 365                 free(el->excludes[i]);
 366         free(el->excludes);
 367
 368         el->nr = 0;
 369         el->excludes = NULL;
 370 }
 371
 372 int add_excludes_from_file_to_list(const char *fname,
 373                                    const char *base,
 374                                    int baselen,
 375                                    char **buf_p,
 376                                    struct exclude_list *which,
 377                                    int check_index)
 378 {
 379         struct stat st;
 380         int fd, i;
 381         size_t size = 0;
 382         char *buf, *entry;
 383
 384         fd = open(fname, O_RDONLY);
 385         if (fd < 0 || fstat(fd, &st) < 0) {
 386                 if (0 <= fd)
 387                         close(fd);
 388                 if (!check_index ||
 389                     (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
 390                         return -1;
 391                 if (size == 0) {
 392                         free(buf);
 393                         return 0;
 394                 }
 395                 if (buf[size-1] != '\n') {
 396                         buf = xrealloc(buf, size+1);
 397                         buf[size++] = '\n';
 398                 }
 399         }
 400         else {
 401                 size = xsize_t(st.st_size);
 402                 if (size == 0) {
 403                         close(fd);
 404                         return 0;
 405                 }
 406                 buf = xmalloc(size+1);
 407                 if (read_in_full(fd, buf, size) != size) {
 408                         free(buf);
 409                         close(fd);
 410                         return -1;
 411                 }
 412                 buf[size++] = '\n';
 413                 close(fd);
 414         }
 415
 416         if (buf_p)
 417                 *buf_p = buf;
 418         entry = buf;
 419         for (i = 0; i < size; i++) {
 420                 if (buf[i] == '\n') {
 421                         if (entry != buf + i && entry[0] != '#') {
 422                                 buf[i - (i && buf[i-1] == '\r')] = 0;
 423                                 add_exclude(entry, base, baselen, which);
 424                         }
 425                         entry = buf + i + 1;
 426                 }
 427         }
 428         return 0;
 429 }
 430
 431 void add_excludes_from_file(struct dir_struct *dir, const char *fname)
 432 {
 433         if (add_excludes_from_file_to_list(fname, "", 0, NULL,
 434                                            &dir->exclude_list[EXC_FILE], 0) < 0)
 435                 die("cannot use %s as an exclude file", fname);
 436 }
 437
 438 static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
 439 {
 440         struct exclude_list *el;
 441         struct exclude_stack *stk = NULL;
 442         int current;
 443
 444         if ((!dir->exclude_per_dir) ||
 445             (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
 446                 return; /* too long a path -- ignore */
 447
 448         /* Pop the ones that are not the prefix of the path being checked. */
 449         el = &dir->exclude_list[EXC_DIRS];
 450         while ((stk = dir->exclude_stack) != NULL) {
 451                 if (stk->baselen <= baselen &&
 452                     !strncmp(dir->basebuf, base, stk->baselen))
 453                         break;
 454                 dir->exclude_stack = stk->prev;
 455                 while (stk->exclude_ix < el->nr)
 456                         free(el->excludes[--el->nr]);
 457                 free(stk->filebuf);
 458                 free(stk);
 459         }
 460
 461         /* Read from the parent directories and push them down. */
 462         current = stk ? stk->baselen : -1;
 463         while (current < baselen) {
 464                 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
 465                 const char *cp;
 466
 467                 if (current < 0) {
 468                         cp = base;
 469                         current = 0;
 470                 }
 471                 else {
 472                         cp = strchr(base + current + 1, '/');
 473                         if (!cp)
 474                                 die("oops in prep_exclude");
 475                         cp++;
 476                 }
 477                 stk->prev = dir->exclude_stack;
 478                 stk->baselen = cp - base;
 479                 stk->exclude_ix = el->nr;
 480                 memcpy(dir->basebuf + current, base + current,
 481                        stk->baselen - current);
 482                 strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
 483                 add_excludes_from_file_to_list(dir->basebuf,
 484                                                dir->basebuf, stk->baselen,
 485                                                &stk->filebuf, el, 1);
 486                 dir->exclude_stack = stk;
 487                 current = stk->baselen;
 488         }
 489         dir->basebuf[baselen] = '\0';
 490 }
 491
 492 /* Scan the list and let the last match determine the fate.
 493  * Return 1 for exclude, 0 for include and -1 for undecided.
 494  */
 495 int excluded_from_list(const char *pathname,
 496                        int pathlen, const char *basename, int *dtype,
 497                        struct exclude_list *el)
 498 {
 499         int i;
 500
 501         if (el->nr) {
 502                 for (i = el->nr - 1; 0 <= i; i--) {
 503                         struct exclude *x = el->excludes[i];
 504                         const char *exclude = x->pattern;
 505                         int to_exclude = x->to_exclude;
 506
 507                         if (x->flags & EXC_FLAG_MUSTBEDIR) {
 508                                 if (*dtype == DT_UNKNOWN)
 509                                         *dtype = get_dtype(NULL, pathname, pathlen);
 510                                 if (*dtype != DT_DIR)
 511                                         continue;
 512                         }
 513
 514                         if (x->flags & EXC_FLAG_NODIR) {
 515                                 /* match basename */
 516                                 if (x->flags & EXC_FLAG_NOWILDCARD) {
 517                                         if (!strcmp_icase(exclude, basename))
 518                                                 return to_exclude;
 519                                 } else if (x->flags & EXC_FLAG_ENDSWITH) {
 520                                         if (x->patternlen - 1 <= pathlen &&
 521                                             !strcmp_icase(exclude + 1, pathname + pathlen - x->patternlen + 1))
 522                                                 return to_exclude;
 523                                 } else {
 524                                         if (fnmatch_icase(exclude, basename, 0) == 0)
 525                                                 return to_exclude;
 526                                 }
 527                         }
 528                         else {
 529                                 /* match with FNM_PATHNAME:
 530                                  * exclude has base (baselen long) implicitly
 531                                  * in front of it.
 532                                  */
 533                                 int baselen = x->baselen;
 534                                 if (*exclude == '/')
 535                                         exclude++;
 536
 537                                 if (pathlen < baselen ||
 538                                     (baselen && pathname[baselen-1] != '/') ||
 539                                     strncmp_icase(pathname, x->base, baselen))
 540                                     continue;
 541
 542                                 if (x->flags & EXC_FLAG_NOWILDCARD) {
 543                                         if (!strcmp_icase(exclude, pathname + baselen))
 544                                                 return to_exclude;
 545                                 } else {
 546                                         if (fnmatch_icase(exclude, pathname+baselen,
 547                                                     FNM_PATHNAME) == 0)
 548                                             return to_exclude;
 549                                 }
 550                         }
 551                 }
 552         }
 553         return -1; /* undecided */
 554 }
 555
 556 int excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
 557 {
 558         int pathlen = strlen(pathname);
 559         int st;
 560         const char *basename = strrchr(pathname, '/');
 561         basename = (basename) ? basename+1 : pathname;
 562
 563         prep_exclude(dir, pathname, basename-pathname);
 564         for (st = EXC_CMDL; st <= EXC_FILE; st++) {
 565                 switch (excluded_from_list(pathname, pathlen, basename,
 566                                            dtype_p, &dir->exclude_list[st])) {
 567                 case 0:
 568                         return 0;
 569                 case 1:
 570                         return 1;
 571                 }
 572         }
 573         return 0;
 574 }
 575
 576 static struct dir_entry *dir_entry_new(const char *pathname, int len)
 577 {
 578         struct dir_entry *ent;
 579
 580         ent = xmalloc(sizeof(*ent) + len + 1);
 581         ent->len = len;
 582         memcpy(ent->name, pathname, len);
 583         ent->name[len] = 0;
 584         return ent;
 585 }
 586
 587 static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
 588 {
 589         if (cache_name_exists(pathname, len, ignore_case))
 590                 return NULL;
 591
 592         ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
 593         return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
 594 }
 595
 596 struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
 597 {
 598         if (!cache_name_is_other(pathname, len))
 599                 return NULL;
 600
 601         ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
 602         return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
 603 }
 604
 605 enum exist_status {
 606         index_nonexistent = 0,
 607         index_directory,
 608         index_gitdir
 609 };
 610
 611 /*
 612  * Do not use the alphabetically stored index to look up
 613  * the directory name; instead, use the case insensitive
 614  * name hash.
 615  */
 616 static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
 617 {
 618         struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
 619         unsigned char endchar;
 620
 621         if (!ce)
 622                 return index_nonexistent;
 623         endchar = ce->name[len];
 624
 625         /*
 626          * The cache_entry structure returned will contain this dirname
 627          * and possibly additional path components.
 628          */
 629         if (endchar == '/')
 630                 return index_directory;
 631
 632         /*
 633          * If there are no additional path components, then this cache_entry
 634          * represents a submodule.  Submodules, despite being directories,
 635          * are stored in the cache without a closing slash.
 636          */
 637         if (!endchar && S_ISGITLINK(ce->ce_mode))
 638                 return index_gitdir;
 639
 640         /* This should never be hit, but it exists just in case. */
 641         return index_nonexistent;
 642 }
 643
 644 /*
 645  * The index sorts alphabetically by entry name, which
 646  * means that a gitlink sorts as '\0' at the end, while
 647  * a directory (which is defined not as an entry, but as
 648  * the files it contains) will sort with the '/' at the
 649  * end.
 650  */
 651 static enum exist_status directory_exists_in_index(const char *dirname, int len)
 652 {
 653         int pos;
 654
 655         if (ignore_case)
 656                 return directory_exists_in_index_icase(dirname, len);
 657
 658         pos = cache_name_pos(dirname, len);
 659         if (pos < 0)
 660                 pos = -pos-1;
 661         while (pos < active_nr) {
 662                 struct cache_entry *ce = active_cache[pos++];
 663                 unsigned char endchar;
 664
 665                 if (strncmp(ce->name, dirname, len))
 666                         break;
 667                 endchar = ce->name[len];
 668                 if (endchar > '/')
 669                         break;
 670                 if (endchar == '/')
 671                         return index_directory;
 672                 if (!endchar && S_ISGITLINK(ce->ce_mode))
 673                         return index_gitdir;
 674         }
 675         return index_nonexistent;
 676 }
 677
 678 /*
 679  * When we find a directory when traversing the filesystem, we
 680  * have three distinct cases:
 681  *
 682  *  - ignore it
 683  *  - see it as a directory
 684  *  - recurse into it
 685  *
 686  * and which one we choose depends on a combination of existing
 687  * git index contents and the flags passed into the directory
 688  * traversal routine.
 689  *
 690  * Case 1: If we *already* have entries in the index under that
 691  * directory name, we always recurse into the directory to see
 692  * all the files.
 693  *
 694  * Case 2: If we *already* have that directory name as a gitlink,
 695  * we always continue to see it as a gitlink, regardless of whether
 696  * there is an actual git directory there or not (it might not
 697  * be checked out as a subproject!)
 698  *
 699  * Case 3: if we didn't have it in the index previously, we
 700  * have a few sub-cases:
 701  *
 702  *  (a) if "show_other_directories" is true, we show it as
 703  *      just a directory, unless "hide_empty_directories" is
 704  *      also true and the directory is empty, in which case
 705  *      we just ignore it entirely.
 706  *  (b) if it looks like a git directory, and we don't have
 707  *      'no_gitlinks' set we treat it as a gitlink, and show it
 708  *      as a directory.
 709  *  (c) otherwise, we recurse into it.
 710  */
 711 enum directory_treatment {
 712         show_directory,
 713         ignore_directory,
 714         recurse_into_directory
 715 };
 716
 717 static enum directory_treatment treat_directory(struct dir_struct *dir,
 718         const char *dirname, int len,
 719         const struct path_simplify *simplify)
 720 {
 721         /* The "len-1" is to strip the final '/' */
 722         switch (directory_exists_in_index(dirname, len-1)) {
 723         case index_directory:
 724                 return recurse_into_directory;
 725
 726         case index_gitdir:
 727                 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 728                         return ignore_directory;
 729                 return show_directory;
 730
 731         case index_nonexistent:
 732                 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 733                         break;
 734                 if (!(dir->flags & DIR_NO_GITLINKS)) {
 735                         unsigned char sha1[20];
 736                         if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
 737                                 return show_directory;
 738                 }
 739                 return recurse_into_directory;
 740         }
 741
 742         /* This is the "show_other_directories" case */
 743         if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
 744                 return show_directory;
 745         if (!read_directory_recursive(dir, dirname, len, 1, simplify))
 746                 return ignore_directory;
 747         return show_directory;
 748 }
 749
 750 /*
 751  * This is an inexact early pruning of any recursive directory
 752  * reading - if the path cannot possibly be in the pathspec,
 753  * return true, and we'll skip it early.
 754  */
 755 static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
 756 {
 757         if (simplify) {
 758                 for (;;) {
 759                         const char *match = simplify->path;
 760                         int len = simplify->len;
 761
 762                         if (!match)
 763                                 break;
 764                         if (len > pathlen)
 765                                 len = pathlen;
 766                         if (!memcmp(path, match, len))
 767                                 return 0;
 768                         simplify++;
 769                 }
 770                 return 1;
 771         }
 772         return 0;
 773 }
 774
 775 /*
 776  * This function tells us whether an excluded path matches a
 777  * list of "interesting" pathspecs. That is, whether a path matched
 778  * by any of the pathspecs could possibly be ignored by excluding
 779  * the specified path. This can happen if:
 780  *
 781  *   1. the path is mentioned explicitly in the pathspec
 782  *
 783  *   2. the path is a directory prefix of some element in the
 784  *      pathspec
 785  */
 786 static int exclude_matches_pathspec(const char *path, int len,
 787                 const struct path_simplify *simplify)
 788 {
 789         if (simplify) {
 790                 for (; simplify->path; simplify++) {
 791                         if (len == simplify->len
 792                             && !memcmp(path, simplify->path, len))
 793                                 return 1;
 794                         if (len < simplify->len
 795                             && simplify->path[len] == '/'
 796                             && !memcmp(path, simplify->path, len))
 797                                 return 1;
 798                 }
 799         }
 800         return 0;
 801 }
 802
 803 static int get_index_dtype(const char *path, int len)
 804 {
 805         int pos;
 806         struct cache_entry *ce;
 807
 808         ce = cache_name_exists(path, len, 0);
 809         if (ce) {
 810                 if (!ce_uptodate(ce))
 811                         return DT_UNKNOWN;
 812                 if (S_ISGITLINK(ce->ce_mode))
 813                         return DT_DIR;
 814                 /*
 815                  * Nobody actually cares about the
 816                  * difference between DT_LNK and DT_REG
 817                  */
 818                 return DT_REG;
 819         }
 820
 821         /* Try to look it up as a directory */
 822         pos = cache_name_pos(path, len);
 823         if (pos >= 0)
 824                 return DT_UNKNOWN;
 825         pos = -pos-1;
 826         while (pos < active_nr) {
 827                 ce = active_cache[pos++];
 828                 if (strncmp(ce->name, path, len))
 829                         break;
 830                 if (ce->name[len] > '/')
 831                         break;
 832                 if (ce->name[len] < '/')
 833                         continue;
 834                 if (!ce_uptodate(ce))
 835                         break;  /* continue? */
 836                 return DT_DIR;
 837         }
 838         return DT_UNKNOWN;
 839 }
 840
 841 static int get_dtype(struct dirent *de, const char *path, int len)
 842 {
 843         int dtype = de ? DTYPE(de) : DT_UNKNOWN;
 844         struct stat st;
 845
 846         if (dtype != DT_UNKNOWN)
 847                 return dtype;
 848         dtype = get_index_dtype(path, len);
 849         if (dtype != DT_UNKNOWN)
 850                 return dtype;
 851         if (lstat(path, &st))
 852                 return dtype;
 853         if (S_ISREG(st.st_mode))
 854                 return DT_REG;
 855         if (S_ISDIR(st.st_mode))
 856                 return DT_DIR;
 857         if (S_ISLNK(st.st_mode))
 858                 return DT_LNK;
 859         return dtype;
 860 }
 861
 862 enum path_treatment {
 863         path_ignored,
 864         path_handled,
 865         path_recurse
 866 };
 867
 868 static enum path_treatment treat_one_path(struct dir_struct *dir,
 869                                           struct strbuf *path,
 870                                           const struct path_simplify *simplify,
 871                                           int dtype, struct dirent *de)
 872 {
 873         int exclude = excluded(dir, path->buf, &dtype);
 874         if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
 875             && exclude_matches_pathspec(path->buf, path->len, simplify))
 876                 dir_add_ignored(dir, path->buf, path->len);
 877
 878         /*
 879          * Excluded? If we don't explicitly want to show
 880          * ignored files, ignore it
 881          */
 882         if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
 883                 return path_ignored;
 884
 885         if (dtype == DT_UNKNOWN)
 886                 dtype = get_dtype(de, path->buf, path->len);
 887
 888         /*
 889          * Do we want to see just the ignored files?
 890          * We still need to recurse into directories,
 891          * even if we don't ignore them, since the
 892          * directory may contain files that we do..
 893          */
 894         if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) {
 895                 if (dtype != DT_DIR)
 896                         return path_ignored;
 897         }
 898
 899         switch (dtype) {
 900         default:
 901                 return path_ignored;
 902         case DT_DIR:
 903                 strbuf_addch(path, '/');
 904                 switch (treat_directory(dir, path->buf, path->len, simplify)) {
 905                 case show_directory:
 906                         if (exclude != !!(dir->flags
 907                                           & DIR_SHOW_IGNORED))
 908                                 return path_ignored;
 909                         break;
 910                 case recurse_into_directory:
 911                         return path_recurse;
 912                 case ignore_directory:
 913                         return path_ignored;
 914                 }
 915                 break;
 916         case DT_REG:
 917         case DT_LNK:
 918                 break;
 919         }
 920         return path_handled;
 921 }
 922
 923 static enum path_treatment treat_path(struct dir_struct *dir,
 924                                       struct dirent *de,
 925                                       struct strbuf *path,
 926                                       int baselen,
 927                                       const struct path_simplify *simplify)
 928 {
 929         int dtype;
 930
 931         if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
 932                 return path_ignored;
 933         strbuf_setlen(path, baselen);
 934         strbuf_addstr(path, de->d_name);
 935         if (simplify_away(path->buf, path->len, simplify))
 936                 return path_ignored;
 937
 938         dtype = DTYPE(de);
 939         return treat_one_path(dir, path, simplify, dtype, de);
 940 }
 941
 942 /*
 943  * Read a directory tree. We currently ignore anything but
 944  * directories, regular files and symlinks. That's because git
 945  * doesn't handle them at all yet. Maybe that will change some
 946  * day.
 947  *
 948  * Also, we ignore the name ".git" (even if it is not a directory).
 949  * That likely will not change.
 950  */
 951 static int read_directory_recursive(struct dir_struct *dir,
 952                                     const char *base, int baselen,
 953                                     int check_only,
 954                                     const struct path_simplify *simplify)
 955 {
 956         DIR *fdir;
 957         int contents = 0;
 958         struct dirent *de;
 959         struct strbuf path = STRBUF_INIT;
 960
 961         strbuf_add(&path, base, baselen);
 962
 963         fdir = opendir(path.len ? path.buf : ".");
 964         if (!fdir)
 965                 goto out;
 966
 967         while ((de = readdir(fdir)) != NULL) {
 968                 switch (treat_path(dir, de, &path, baselen, simplify)) {
 969                 case path_recurse:
 970                         contents += read_directory_recursive(dir, path.buf,
 971                                                              path.len, 0,
 972                                                              simplify);
 973                         continue;
 974                 case path_ignored:
 975                         continue;
 976                 case path_handled:
 977                         break;
 978                 }
 979                 contents++;
 980                 if (check_only)
 981                         break;
 982                 dir_add_name(dir, path.buf, path.len);
 983         }
 984         closedir(fdir);
 985  out:
 986         strbuf_release(&path);
 987
 988         return contents;
 989 }
 990
 991 static int cmp_name(const void *p1, const void *p2)
 992 {
 993         const struct dir_entry *e1 = *(const struct dir_entry **)p1;
 994         const struct dir_entry *e2 = *(const struct dir_entry **)p2;
 995
 996         return cache_name_compare(e1->name, e1->len,
 997                                   e2->name, e2->len);
 998 }
 999
1000 /*
1001  * Return the length of the "simple" part of a path match limiter.
1002  */
1003 static int simple_length(const char *match)
1004 {
1005         int len = -1;
1006
1007         for (;;) {
1008                 unsigned char c = *match++;
1009                 len++;
1010                 if (c == '\0' || is_glob_special(c))
1011                         return len;
1012         }
1013 }
1014
1015 static struct path_simplify *create_simplify(const char **pathspec)
1016 {
1017         int nr, alloc = 0;
1018         struct path_simplify *simplify = NULL;
1019
1020         if (!pathspec)
1021                 return NULL;
1022
1023         for (nr = 0 ; ; nr++) {
1024                 const char *match;
1025                 if (nr >= alloc) {
1026                         alloc = alloc_nr(alloc);
1027                         simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1028                 }
1029                 match = *pathspec++;
1030                 if (!match)
1031                         break;
1032                 simplify[nr].path = match;
1033                 simplify[nr].len = simple_length(match);
1034         }
1035         simplify[nr].path = NULL;
1036         simplify[nr].len = 0;
1037         return simplify;
1038 }
1039
1040 static void free_simplify(struct path_simplify *simplify)
1041 {
1042         free(simplify);
1043 }
1044
1045 static int treat_leading_path(struct dir_struct *dir,
1046                               const char *path, int len,
1047                               const struct path_simplify *simplify)
1048 {
1049         struct strbuf sb = STRBUF_INIT;
1050         int baselen, rc = 0;
1051         const char *cp;
1052
1053         while (len && path[len - 1] == '/')
1054                 len--;
1055         if (!len)
1056                 return 1;
1057         baselen = 0;
1058         while (1) {
1059                 cp = path + baselen + !!baselen;
1060                 cp = memchr(cp, '/', path + len - cp);
1061                 if (!cp)
1062                         baselen = len;
1063                 else
1064                         baselen = cp - path;
1065                 strbuf_setlen(&sb, 0);
1066                 strbuf_add(&sb, path, baselen);
1067                 if (!is_directory(sb.buf))
1068                         break;
1069                 if (simplify_away(sb.buf, sb.len, simplify))
1070                         break;
1071                 if (treat_one_path(dir, &sb, simplify,
1072                                    DT_DIR, NULL) == path_ignored)
1073                         break; /* do not recurse into it */
1074                 if (len <= baselen) {
1075                         rc = 1;
1076                         break; /* finished checking */
1077                 }
1078         }
1079         strbuf_release(&sb);
1080         return rc;
1081 }
1082
1083 int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
1084 {
1085         struct path_simplify *simplify;
1086
1087         if (has_symlink_leading_path(path, len))
1088                 return dir->nr;
1089
1090         simplify = create_simplify(pathspec);
1091         if (!len || treat_leading_path(dir, path, len, simplify))
1092                 read_directory_recursive(dir, path, len, 0, simplify);
1093         free_simplify(simplify);
1094         qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
1095         qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
1096         return dir->nr;
1097 }
1098
1099 int file_exists(const char *f)
1100 {
1101         struct stat sb;
1102         return lstat(f, &sb) == 0;
1103 }
1104
1105 /*
1106  * Given two normalized paths (a trailing slash is ok), if subdir is
1107  * outside dir, return -1.  Otherwise return the offset in subdir that
1108  * can be used as relative path to dir.
1109  */
1110 int dir_inside_of(const char *subdir, const char *dir)
1111 {
1112         int offset = 0;
1113
1114         assert(dir && subdir && *dir && *subdir);
1115
1116         while (*dir && *subdir && *dir == *subdir) {
1117                 dir++;
1118                 subdir++;
1119                 offset++;
1120         }
1121
1122         /* hel[p]/me vs hel[l]/yeah */
1123         if (*dir && *subdir)
1124                 return -1;
1125
1126         if (!*subdir)
1127                 return !*dir ? offset : -1; /* same dir */
1128
1129         /* foo/[b]ar vs foo/[] */
1130         if (is_dir_sep(dir[-1]))
1131                 return is_dir_sep(subdir[-1]) ? offset : -1;
1132
1133         /* foo[/]bar vs foo[] */
1134         return is_dir_sep(*subdir) ? offset + 1 : -1;
1135 }
1136
1137 int is_inside_dir(const char *dir)
1138 {
1139         char cwd[PATH_MAX];
1140         if (!dir)
1141                 return 0;
1142         if (!getcwd(cwd, sizeof(cwd)))
1143                 die_errno("can't find the current directory");
1144         return dir_inside_of(cwd, dir) >= 0;
1145 }
1146
1147 int is_empty_dir(const char *path)
1148 {
1149         DIR *dir = opendir(path);
1150         struct dirent *e;
1151         int ret = 1;
1152
1153         if (!dir)
1154                 return 0;
1155
1156         while ((e = readdir(dir)) != NULL)
1157                 if (!is_dot_or_dotdot(e->d_name)) {
1158                         ret = 0;
1159                         break;
1160                 }
1161
1162         closedir(dir);
1163         return ret;
1164 }
1165
1166 static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
1167 {
1168         DIR *dir;
1169         struct dirent *e;
1170         int ret = 0, original_len = path->len, len, kept_down = 0;
1171         int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
1172         int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
1173         unsigned char submodule_head[20];
1174
1175         if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
1176             !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
1177                 /* Do not descend and nuke a nested git work tree. */
1178                 if (kept_up)
1179                         *kept_up = 1;
1180                 return 0;
1181         }
1182
1183         flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
1184         dir = opendir(path->buf);
1185         if (!dir) {
1186                 /* an empty dir could be removed even if it is unreadble */
1187                 if (!keep_toplevel)
1188                         return rmdir(path->buf);
1189                 else
1190                         return -1;
1191         }
1192         if (path->buf[original_len - 1] != '/')
1193                 strbuf_addch(path, '/');
1194
1195         len = path->len;
1196         while ((e = readdir(dir)) != NULL) {
1197                 struct stat st;
1198                 if (is_dot_or_dotdot(e->d_name))
1199                         continue;
1200
1201                 strbuf_setlen(path, len);
1202                 strbuf_addstr(path, e->d_name);
1203                 if (lstat(path->buf, &st))
1204                         ; /* fall thru */
1205                 else if (S_ISDIR(st.st_mode)) {
1206                         if (!remove_dir_recurse(path, flag, &kept_down))
1207                                 continue; /* happy */
1208                 } else if (!only_empty && !unlink(path->buf))
1209                         continue; /* happy, too */
1210
1211                 /* path too long, stat fails, or non-directory still exists */
1212                 ret = -1;
1213                 break;
1214         }
1215         closedir(dir);
1216
1217         strbuf_setlen(path, original_len);
1218         if (!ret && !keep_toplevel && !kept_down)
1219                 ret = rmdir(path->buf);
1220         else if (kept_up)
1221                 /*
1222                  * report the uplevel that it is not an error that we
1223                  * did not rmdir() our directory.
1224                  */
1225                 *kept_up = !ret;
1226         return ret;
1227 }
1228
1229 int remove_dir_recursively(struct strbuf *path, int flag)
1230 {
1231         return remove_dir_recurse(path, flag, NULL);
1232 }
1233
1234 void setup_standard_excludes(struct dir_struct *dir)
1235 {
1236         const char *path;
1237
1238         dir->exclude_per_dir = ".gitignore";
1239         path = git_path("info/exclude");
1240         if (!access(path, R_OK))
1241                 add_excludes_from_file(dir, path);
1242         if (excludes_file && !access(excludes_file, R_OK))
1243                 add_excludes_from_file(dir, excludes_file);
1244 }
1245
1246 int remove_path(const char *name)
1247 {
1248         char *slash;
1249
1250         if (unlink(name) && errno != ENOENT)
1251                 return -1;
1252
1253         slash = strrchr(name, '/');
1254         if (slash) {
1255                 char *dirs = xstrdup(name);
1256                 slash = dirs + (slash - name);
1257                 do {
1258                         *slash = '\0';
1259                 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
1260                 free(dirs);
1261         }
1262         return 0;
1263 }
1264
1265 static int pathspec_item_cmp(const void *a_, const void *b_)
1266 {
1267         struct pathspec_item *a, *b;
1268
1269         a = (struct pathspec_item *)a_;
1270         b = (struct pathspec_item *)b_;
1271         return strcmp(a->match, b->match);
1272 }
1273
1274 int init_pathspec(struct pathspec *pathspec, const char **paths)
1275 {
1276         const char **p = paths;
1277         int i;
1278
1279         memset(pathspec, 0, sizeof(*pathspec));
1280         if (!p)
1281                 return 0;
1282         while (*p)
1283                 p++;
1284         pathspec->raw = paths;
1285         pathspec->nr = p - paths;
1286         if (!pathspec->nr)
1287                 return 0;
1288
1289         pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1290         for (i = 0; i < pathspec->nr; i++) {
1291                 struct pathspec_item *item = pathspec->items+i;
1292                 const char *path = paths[i];
1293
1294                 item->match = path;
1295                 item->len = strlen(path);
1296                 item->use_wildcard = !no_wildcard(path);
1297                 if (item->use_wildcard)
1298                         pathspec->has_wildcard = 1;
1299         }
1300
1301         qsort(pathspec->items, pathspec->nr,
1302               sizeof(struct pathspec_item), pathspec_item_cmp);
1303
1304         return 0;
1305 }
1306
1307 void free_pathspec(struct pathspec *pathspec)
1308 {
1309         free(pathspec->items);
1310         pathspec->items = NULL;
1311 }