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 /*
 292  * Return the length of the "simple" part of a path match limiter.
 293  */
 294 static int simple_length(const char *match)
 295 {
 296         int len = -1;
 297
 298         for (;;) {
 299                 unsigned char c = *match++;
 300                 len++;
 301                 if (c == '\0' || is_glob_special(c))
 302                         return len;
 303         }
 304 }
 305
 306 static int no_wildcard(const char *string)
 307 {
 308         return string[simple_length(string)] == '\0';
 309 }
 310
 311 void add_exclude(const char *string, const char *base,
 312                  int baselen, struct exclude_list *which)
 313 {
 314         struct exclude *x;
 315         size_t len;
 316         int to_exclude = 1;
 317         int flags = 0;
 318
 319         if (*string == '!') {
 320                 to_exclude = 0;
 321                 string++;
 322         }
 323         len = strlen(string);
 324         if (len && string[len - 1] == '/') {
 325                 char *s;
 326                 x = xmalloc(sizeof(*x) + len);
 327                 s = (char *)(x+1);
 328                 memcpy(s, string, len - 1);
 329                 s[len - 1] = '\0';
 330                 string = s;
 331                 x->pattern = s;
 332                 flags = EXC_FLAG_MUSTBEDIR;
 333         } else {
 334                 x = xmalloc(sizeof(*x));
 335                 x->pattern = string;
 336         }
 337         x->to_exclude = to_exclude;
 338         x->patternlen = strlen(string);
 339         x->base = base;
 340         x->baselen = baselen;
 341         x->flags = flags;
 342         if (!strchr(string, '/'))
 343                 x->flags |= EXC_FLAG_NODIR;
 344         x->nowildcardlen = simple_length(string);
 345         if (*string == '*' && no_wildcard(string+1))
 346                 x->flags |= EXC_FLAG_ENDSWITH;
 347         ALLOC_GROW(which->excludes, which->nr + 1, which->alloc);
 348         which->excludes[which->nr++] = x;
 349 }
 350
 351 static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
 352 {
 353         int pos, len;
 354         unsigned long sz;
 355         enum object_type type;
 356         void *data;
 357         struct index_state *istate = &the_index;
 358
 359         len = strlen(path);
 360         pos = index_name_pos(istate, path, len);
 361         if (pos < 0)
 362                 return NULL;
 363         if (!ce_skip_worktree(istate->cache[pos]))
 364                 return NULL;
 365         data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
 366         if (!data || type != OBJ_BLOB) {
 367                 free(data);
 368                 return NULL;
 369         }
 370         *size = xsize_t(sz);
 371         return data;
 372 }
 373
 374 void free_excludes(struct exclude_list *el)
 375 {
 376         int i;
 377
 378         for (i = 0; i < el->nr; i++)
 379                 free(el->excludes[i]);
 380         free(el->excludes);
 381
 382         el->nr = 0;
 383         el->excludes = NULL;
 384 }
 385
 386 int add_excludes_from_file_to_list(const char *fname,
 387                                    const char *base,
 388                                    int baselen,
 389                                    char **buf_p,
 390                                    struct exclude_list *which,
 391                                    int check_index)
 392 {
 393         struct stat st;
 394         int fd, i;
 395         size_t size = 0;
 396         char *buf, *entry;
 397
 398         fd = open(fname, O_RDONLY);
 399         if (fd < 0 || fstat(fd, &st) < 0) {
 400                 if (0 <= fd)
 401                         close(fd);
 402                 if (!check_index ||
 403                     (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
 404                         return -1;
 405                 if (size == 0) {
 406                         free(buf);
 407                         return 0;
 408                 }
 409                 if (buf[size-1] != '\n') {
 410                         buf = xrealloc(buf, size+1);
 411                         buf[size++] = '\n';
 412                 }
 413         }
 414         else {
 415                 size = xsize_t(st.st_size);
 416                 if (size == 0) {
 417                         close(fd);
 418                         return 0;
 419                 }
 420                 buf = xmalloc(size+1);
 421                 if (read_in_full(fd, buf, size) != size) {
 422                         free(buf);
 423                         close(fd);
 424                         return -1;
 425                 }
 426                 buf[size++] = '\n';
 427                 close(fd);
 428         }
 429
 430         if (buf_p)
 431                 *buf_p = buf;
 432         entry = buf;
 433         for (i = 0; i < size; i++) {
 434                 if (buf[i] == '\n') {
 435                         if (entry != buf + i && entry[0] != '#') {
 436                                 buf[i - (i && buf[i-1] == '\r')] = 0;
 437                                 add_exclude(entry, base, baselen, which);
 438                         }
 439                         entry = buf + i + 1;
 440                 }
 441         }
 442         return 0;
 443 }
 444
 445 void add_excludes_from_file(struct dir_struct *dir, const char *fname)
 446 {
 447         if (add_excludes_from_file_to_list(fname, "", 0, NULL,
 448                                            &dir->exclude_list[EXC_FILE], 0) < 0)
 449                 die("cannot use %s as an exclude file", fname);
 450 }
 451
 452 static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
 453 {
 454         struct exclude_list *el;
 455         struct exclude_stack *stk = NULL;
 456         int current;
 457
 458         if ((!dir->exclude_per_dir) ||
 459             (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
 460                 return; /* too long a path -- ignore */
 461
 462         /* Pop the ones that are not the prefix of the path being checked. */
 463         el = &dir->exclude_list[EXC_DIRS];
 464         while ((stk = dir->exclude_stack) != NULL) {
 465                 if (stk->baselen <= baselen &&
 466                     !strncmp(dir->basebuf, base, stk->baselen))
 467                         break;
 468                 dir->exclude_stack = stk->prev;
 469                 while (stk->exclude_ix < el->nr)
 470                         free(el->excludes[--el->nr]);
 471                 free(stk->filebuf);
 472                 free(stk);
 473         }
 474
 475         /* Read from the parent directories and push them down. */
 476         current = stk ? stk->baselen : -1;
 477         while (current < baselen) {
 478                 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
 479                 const char *cp;
 480
 481                 if (current < 0) {
 482                         cp = base;
 483                         current = 0;
 484                 }
 485                 else {
 486                         cp = strchr(base + current + 1, '/');
 487                         if (!cp)
 488                                 die("oops in prep_exclude");
 489                         cp++;
 490                 }
 491                 stk->prev = dir->exclude_stack;
 492                 stk->baselen = cp - base;
 493                 stk->exclude_ix = el->nr;
 494                 memcpy(dir->basebuf + current, base + current,
 495                        stk->baselen - current);
 496                 strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
 497                 add_excludes_from_file_to_list(dir->basebuf,
 498                                                dir->basebuf, stk->baselen,
 499                                                &stk->filebuf, el, 1);
 500                 dir->exclude_stack = stk;
 501                 current = stk->baselen;
 502         }
 503         dir->basebuf[baselen] = '\0';
 504 }
 505
 506 /* Scan the list and let the last match determine the fate.
 507  * Return 1 for exclude, 0 for include and -1 for undecided.
 508  */
 509 int excluded_from_list(const char *pathname,
 510                        int pathlen, const char *basename, int *dtype,
 511                        struct exclude_list *el)
 512 {
 513         int i;
 514
 515         if (!el->nr)
 516                 return -1;      /* undefined */
 517
 518         for (i = el->nr - 1; 0 <= i; i--) {
 519                 struct exclude *x = el->excludes[i];
 520                 const char *name, *exclude = x->pattern;
 521                 int to_exclude = x->to_exclude;
 522                 int namelen, prefix = x->nowildcardlen;
 523
 524                 if (x->flags & EXC_FLAG_MUSTBEDIR) {
 525                         if (*dtype == DT_UNKNOWN)
 526                                 *dtype = get_dtype(NULL, pathname, pathlen);
 527                         if (*dtype != DT_DIR)
 528                                 continue;
 529                 }
 530
 531                 if (x->flags & EXC_FLAG_NODIR) {
 532                         /* match basename */
 533                         if (prefix == x->patternlen) {
 534                                 if (!strcmp_icase(exclude, basename))
 535                                         return to_exclude;
 536                         } else if (x->flags & EXC_FLAG_ENDSWITH) {
 537                                 if (x->patternlen - 1 <= pathlen &&
 538                                     !strcmp_icase(exclude + 1, pathname + pathlen - x->patternlen + 1))
 539                                         return to_exclude;
 540                         } else {
 541                                 if (fnmatch_icase(exclude, basename, 0) == 0)
 542                                         return to_exclude;
 543                         }
 544                         continue;
 545                 }
 546
 547                 /* match with FNM_PATHNAME:
 548                  * exclude has base (baselen long) implicitly in front of it.
 549                  */
 550                 if (*exclude == '/') {
 551                         exclude++;
 552                         prefix--;
 553                 }
 554
 555                 if (pathlen < x->baselen ||
 556                     (x->baselen && pathname[x->baselen-1] != '/') ||
 557                     strncmp_icase(pathname, x->base, x->baselen))
 558                         continue;
 559
 560                 namelen = x->baselen ? pathlen - x->baselen : pathlen;
 561                 name = pathname + pathlen  - namelen;
 562
 563                 /* if the non-wildcard part is longer than the
 564                    remaining pathname, surely it cannot match */
 565                 if (prefix > namelen)
 566                         continue;
 567
 568                 if (prefix) {
 569                         if (strncmp_icase(exclude, name, prefix))
 570                                 continue;
 571                         exclude += prefix;
 572                         name    += prefix;
 573                         namelen -= prefix;
 574                 }
 575
 576                 if (!namelen || !fnmatch_icase(exclude, name, FNM_PATHNAME))
 577                         return to_exclude;
 578         }
 579         return -1; /* undecided */
 580 }
 581
 582 static int excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
 583 {
 584         int pathlen = strlen(pathname);
 585         int st;
 586         const char *basename = strrchr(pathname, '/');
 587         basename = (basename) ? basename+1 : pathname;
 588
 589         prep_exclude(dir, pathname, basename-pathname);
 590         for (st = EXC_CMDL; st <= EXC_FILE; st++) {
 591                 switch (excluded_from_list(pathname, pathlen, basename,
 592                                            dtype_p, &dir->exclude_list[st])) {
 593                 case 0:
 594                         return 0;
 595                 case 1:
 596                         return 1;
 597                 }
 598         }
 599         return 0;
 600 }
 601
 602 void path_exclude_check_init(struct path_exclude_check *check,
 603                              struct dir_struct *dir)
 604 {
 605         check->dir = dir;
 606         strbuf_init(&check->path, 256);
 607 }
 608
 609 void path_exclude_check_clear(struct path_exclude_check *check)
 610 {
 611         strbuf_release(&check->path);
 612 }
 613
 614 /*
 615  * Is this name excluded?  This is for a caller like show_files() that
 616  * do not honor directory hierarchy and iterate through paths that are
 617  * possibly in an ignored directory.
 618  *
 619  * A path to a directory known to be excluded is left in check->path to
 620  * optimize for repeated checks for files in the same excluded directory.
 621  */
 622 int path_excluded(struct path_exclude_check *check,
 623                   const char *name, int namelen, int *dtype)
 624 {
 625         int i;
 626         struct strbuf *path = &check->path;
 627
 628         /*
 629          * we allow the caller to pass namelen as an optimization; it
 630          * must match the length of the name, as we eventually call
 631          * excluded() on the whole name string.
 632          */
 633         if (namelen < 0)
 634                 namelen = strlen(name);
 635
 636         if (path->len &&
 637             path->len <= namelen &&
 638             !memcmp(name, path->buf, path->len) &&
 639             (!name[path->len] || name[path->len] == '/'))
 640                 return 1;
 641
 642         strbuf_setlen(path, 0);
 643         for (i = 0; name[i]; i++) {
 644                 int ch = name[i];
 645
 646                 if (ch == '/') {
 647                         int dt = DT_DIR;
 648                         if (excluded(check->dir, path->buf, &dt))
 649                                 return 1;
 650                 }
 651                 strbuf_addch(path, ch);
 652         }
 653
 654         /* An entry in the index; cannot be a directory with subentries */
 655         strbuf_setlen(path, 0);
 656
 657         return excluded(check->dir, name, dtype);
 658 }
 659
 660 static struct dir_entry *dir_entry_new(const char *pathname, int len)
 661 {
 662         struct dir_entry *ent;
 663
 664         ent = xmalloc(sizeof(*ent) + len + 1);
 665         ent->len = len;
 666         memcpy(ent->name, pathname, len);
 667         ent->name[len] = 0;
 668         return ent;
 669 }
 670
 671 static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
 672 {
 673         if (cache_name_exists(pathname, len, ignore_case))
 674                 return NULL;
 675
 676         ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
 677         return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
 678 }
 679
 680 struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
 681 {
 682         if (!cache_name_is_other(pathname, len))
 683                 return NULL;
 684
 685         ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
 686         return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
 687 }
 688
 689 enum exist_status {
 690         index_nonexistent = 0,
 691         index_directory,
 692         index_gitdir
 693 };
 694
 695 /*
 696  * Do not use the alphabetically stored index to look up
 697  * the directory name; instead, use the case insensitive
 698  * name hash.
 699  */
 700 static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
 701 {
 702         struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
 703         unsigned char endchar;
 704
 705         if (!ce)
 706                 return index_nonexistent;
 707         endchar = ce->name[len];
 708
 709         /*
 710          * The cache_entry structure returned will contain this dirname
 711          * and possibly additional path components.
 712          */
 713         if (endchar == '/')
 714                 return index_directory;
 715
 716         /*
 717          * If there are no additional path components, then this cache_entry
 718          * represents a submodule.  Submodules, despite being directories,
 719          * are stored in the cache without a closing slash.
 720          */
 721         if (!endchar && S_ISGITLINK(ce->ce_mode))
 722                 return index_gitdir;
 723
 724         /* This should never be hit, but it exists just in case. */
 725         return index_nonexistent;
 726 }
 727
 728 /*
 729  * The index sorts alphabetically by entry name, which
 730  * means that a gitlink sorts as '\0' at the end, while
 731  * a directory (which is defined not as an entry, but as
 732  * the files it contains) will sort with the '/' at the
 733  * end.
 734  */
 735 static enum exist_status directory_exists_in_index(const char *dirname, int len)
 736 {
 737         int pos;
 738
 739         if (ignore_case)
 740                 return directory_exists_in_index_icase(dirname, len);
 741
 742         pos = cache_name_pos(dirname, len);
 743         if (pos < 0)
 744                 pos = -pos-1;
 745         while (pos < active_nr) {
 746                 struct cache_entry *ce = active_cache[pos++];
 747                 unsigned char endchar;
 748
 749                 if (strncmp(ce->name, dirname, len))
 750                         break;
 751                 endchar = ce->name[len];
 752                 if (endchar > '/')
 753                         break;
 754                 if (endchar == '/')
 755                         return index_directory;
 756                 if (!endchar && S_ISGITLINK(ce->ce_mode))
 757                         return index_gitdir;
 758         }
 759         return index_nonexistent;
 760 }
 761
 762 /*
 763  * When we find a directory when traversing the filesystem, we
 764  * have three distinct cases:
 765  *
 766  *  - ignore it
 767  *  - see it as a directory
 768  *  - recurse into it
 769  *
 770  * and which one we choose depends on a combination of existing
 771  * git index contents and the flags passed into the directory
 772  * traversal routine.
 773  *
 774  * Case 1: If we *already* have entries in the index under that
 775  * directory name, we always recurse into the directory to see
 776  * all the files.
 777  *
 778  * Case 2: If we *already* have that directory name as a gitlink,
 779  * we always continue to see it as a gitlink, regardless of whether
 780  * there is an actual git directory there or not (it might not
 781  * be checked out as a subproject!)
 782  *
 783  * Case 3: if we didn't have it in the index previously, we
 784  * have a few sub-cases:
 785  *
 786  *  (a) if "show_other_directories" is true, we show it as
 787  *      just a directory, unless "hide_empty_directories" is
 788  *      also true and the directory is empty, in which case
 789  *      we just ignore it entirely.
 790  *  (b) if it looks like a git directory, and we don't have
 791  *      'no_gitlinks' set we treat it as a gitlink, and show it
 792  *      as a directory.
 793  *  (c) otherwise, we recurse into it.
 794  */
 795 enum directory_treatment {
 796         show_directory,
 797         ignore_directory,
 798         recurse_into_directory
 799 };
 800
 801 static enum directory_treatment treat_directory(struct dir_struct *dir,
 802         const char *dirname, int len,
 803         const struct path_simplify *simplify)
 804 {
 805         /* The "len-1" is to strip the final '/' */
 806         switch (directory_exists_in_index(dirname, len-1)) {
 807         case index_directory:
 808                 return recurse_into_directory;
 809
 810         case index_gitdir:
 811                 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 812                         return ignore_directory;
 813                 return show_directory;
 814
 815         case index_nonexistent:
 816                 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 817                         break;
 818                 if (!(dir->flags & DIR_NO_GITLINKS)) {
 819                         unsigned char sha1[20];
 820                         if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
 821                                 return show_directory;
 822                 }
 823                 return recurse_into_directory;
 824         }
 825
 826         /* This is the "show_other_directories" case */
 827         if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
 828                 return show_directory;
 829         if (!read_directory_recursive(dir, dirname, len, 1, simplify))
 830                 return ignore_directory;
 831         return show_directory;
 832 }
 833
 834 /*
 835  * This is an inexact early pruning of any recursive directory
 836  * reading - if the path cannot possibly be in the pathspec,
 837  * return true, and we'll skip it early.
 838  */
 839 static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
 840 {
 841         if (simplify) {
 842                 for (;;) {
 843                         const char *match = simplify->path;
 844                         int len = simplify->len;
 845
 846                         if (!match)
 847                                 break;
 848                         if (len > pathlen)
 849                                 len = pathlen;
 850                         if (!memcmp(path, match, len))
 851                                 return 0;
 852                         simplify++;
 853                 }
 854                 return 1;
 855         }
 856         return 0;
 857 }
 858
 859 /*
 860  * This function tells us whether an excluded path matches a
 861  * list of "interesting" pathspecs. That is, whether a path matched
 862  * by any of the pathspecs could possibly be ignored by excluding
 863  * the specified path. This can happen if:
 864  *
 865  *   1. the path is mentioned explicitly in the pathspec
 866  *
 867  *   2. the path is a directory prefix of some element in the
 868  *      pathspec
 869  */
 870 static int exclude_matches_pathspec(const char *path, int len,
 871                 const struct path_simplify *simplify)
 872 {
 873         if (simplify) {
 874                 for (; simplify->path; simplify++) {
 875                         if (len == simplify->len
 876                             && !memcmp(path, simplify->path, len))
 877                                 return 1;
 878                         if (len < simplify->len
 879                             && simplify->path[len] == '/'
 880                             && !memcmp(path, simplify->path, len))
 881                                 return 1;
 882                 }
 883         }
 884         return 0;
 885 }
 886
 887 static int get_index_dtype(const char *path, int len)
 888 {
 889         int pos;
 890         struct cache_entry *ce;
 891
 892         ce = cache_name_exists(path, len, 0);
 893         if (ce) {
 894                 if (!ce_uptodate(ce))
 895                         return DT_UNKNOWN;
 896                 if (S_ISGITLINK(ce->ce_mode))
 897                         return DT_DIR;
 898                 /*
 899                  * Nobody actually cares about the
 900                  * difference between DT_LNK and DT_REG
 901                  */
 902                 return DT_REG;
 903         }
 904
 905         /* Try to look it up as a directory */
 906         pos = cache_name_pos(path, len);
 907         if (pos >= 0)
 908                 return DT_UNKNOWN;
 909         pos = -pos-1;
 910         while (pos < active_nr) {
 911                 ce = active_cache[pos++];
 912                 if (strncmp(ce->name, path, len))
 913                         break;
 914                 if (ce->name[len] > '/')
 915                         break;
 916                 if (ce->name[len] < '/')
 917                         continue;
 918                 if (!ce_uptodate(ce))
 919                         break;  /* continue? */
 920                 return DT_DIR;
 921         }
 922         return DT_UNKNOWN;
 923 }
 924
 925 static int get_dtype(struct dirent *de, const char *path, int len)
 926 {
 927         int dtype = de ? DTYPE(de) : DT_UNKNOWN;
 928         struct stat st;
 929
 930         if (dtype != DT_UNKNOWN)
 931                 return dtype;
 932         dtype = get_index_dtype(path, len);
 933         if (dtype != DT_UNKNOWN)
 934                 return dtype;
 935         if (lstat(path, &st))
 936                 return dtype;
 937         if (S_ISREG(st.st_mode))
 938                 return DT_REG;
 939         if (S_ISDIR(st.st_mode))
 940                 return DT_DIR;
 941         if (S_ISLNK(st.st_mode))
 942                 return DT_LNK;
 943         return dtype;
 944 }
 945
 946 enum path_treatment {
 947         path_ignored,
 948         path_handled,
 949         path_recurse
 950 };
 951
 952 static enum path_treatment treat_one_path(struct dir_struct *dir,
 953                                           struct strbuf *path,
 954                                           const struct path_simplify *simplify,
 955                                           int dtype, struct dirent *de)
 956 {
 957         int exclude = excluded(dir, path->buf, &dtype);
 958         if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
 959             && exclude_matches_pathspec(path->buf, path->len, simplify))
 960                 dir_add_ignored(dir, path->buf, path->len);
 961
 962         /*
 963          * Excluded? If we don't explicitly want to show
 964          * ignored files, ignore it
 965          */
 966         if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
 967                 return path_ignored;
 968
 969         if (dtype == DT_UNKNOWN)
 970                 dtype = get_dtype(de, path->buf, path->len);
 971
 972         /*
 973          * Do we want to see just the ignored files?
 974          * We still need to recurse into directories,
 975          * even if we don't ignore them, since the
 976          * directory may contain files that we do..
 977          */
 978         if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) {
 979                 if (dtype != DT_DIR)
 980                         return path_ignored;
 981         }
 982
 983         switch (dtype) {
 984         default:
 985                 return path_ignored;
 986         case DT_DIR:
 987                 strbuf_addch(path, '/');
 988                 switch (treat_directory(dir, path->buf, path->len, simplify)) {
 989                 case show_directory:
 990                         if (exclude != !!(dir->flags
 991                                           & DIR_SHOW_IGNORED))
 992                                 return path_ignored;
 993                         break;
 994                 case recurse_into_directory:
 995                         return path_recurse;
 996                 case ignore_directory:
 997                         return path_ignored;
 998                 }
 999                 break;
1000         case DT_REG:
1001         case DT_LNK:
1002                 break;
1003         }
1004         return path_handled;
1005 }
1006
1007 static enum path_treatment treat_path(struct dir_struct *dir,
1008                                       struct dirent *de,
1009                                       struct strbuf *path,
1010                                       int baselen,
1011                                       const struct path_simplify *simplify)
1012 {
1013         int dtype;
1014
1015         if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
1016                 return path_ignored;
1017         strbuf_setlen(path, baselen);
1018         strbuf_addstr(path, de->d_name);
1019         if (simplify_away(path->buf, path->len, simplify))
1020                 return path_ignored;
1021
1022         dtype = DTYPE(de);
1023         return treat_one_path(dir, path, simplify, dtype, de);
1024 }
1025
1026 /*
1027  * Read a directory tree. We currently ignore anything but
1028  * directories, regular files and symlinks. That's because git
1029  * doesn't handle them at all yet. Maybe that will change some
1030  * day.
1031  *
1032  * Also, we ignore the name ".git" (even if it is not a directory).
1033  * That likely will not change.
1034  */
1035 static int read_directory_recursive(struct dir_struct *dir,
1036                                     const char *base, int baselen,
1037                                     int check_only,
1038                                     const struct path_simplify *simplify)
1039 {
1040         DIR *fdir;
1041         int contents = 0;
1042         struct dirent *de;
1043         struct strbuf path = STRBUF_INIT;
1044
1045         strbuf_add(&path, base, baselen);
1046
1047         fdir = opendir(path.len ? path.buf : ".");
1048         if (!fdir)
1049                 goto out;
1050
1051         while ((de = readdir(fdir)) != NULL) {
1052                 switch (treat_path(dir, de, &path, baselen, simplify)) {
1053                 case path_recurse:
1054                         contents += read_directory_recursive(dir, path.buf,
1055                                                              path.len, 0,
1056                                                              simplify);
1057                         continue;
1058                 case path_ignored:
1059                         continue;
1060                 case path_handled:
1061                         break;
1062                 }
1063                 contents++;
1064                 if (check_only)
1065                         break;
1066                 dir_add_name(dir, path.buf, path.len);
1067         }
1068         closedir(fdir);
1069  out:
1070         strbuf_release(&path);
1071
1072         return contents;
1073 }
1074
1075 static int cmp_name(const void *p1, const void *p2)
1076 {
1077         const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1078         const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1079
1080         return cache_name_compare(e1->name, e1->len,
1081                                   e2->name, e2->len);
1082 }
1083
1084 static struct path_simplify *create_simplify(const char **pathspec)
1085 {
1086         int nr, alloc = 0;
1087         struct path_simplify *simplify = NULL;
1088
1089         if (!pathspec)
1090                 return NULL;
1091
1092         for (nr = 0 ; ; nr++) {
1093                 const char *match;
1094                 if (nr >= alloc) {
1095                         alloc = alloc_nr(alloc);
1096                         simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1097                 }
1098                 match = *pathspec++;
1099                 if (!match)
1100                         break;
1101                 simplify[nr].path = match;
1102                 simplify[nr].len = simple_length(match);
1103         }
1104         simplify[nr].path = NULL;
1105         simplify[nr].len = 0;
1106         return simplify;
1107 }
1108
1109 static void free_simplify(struct path_simplify *simplify)
1110 {
1111         free(simplify);
1112 }
1113
1114 static int treat_leading_path(struct dir_struct *dir,
1115                               const char *path, int len,
1116                               const struct path_simplify *simplify)
1117 {
1118         struct strbuf sb = STRBUF_INIT;
1119         int baselen, rc = 0;
1120         const char *cp;
1121
1122         while (len && path[len - 1] == '/')
1123                 len--;
1124         if (!len)
1125                 return 1;
1126         baselen = 0;
1127         while (1) {
1128                 cp = path + baselen + !!baselen;
1129                 cp = memchr(cp, '/', path + len - cp);
1130                 if (!cp)
1131                         baselen = len;
1132                 else
1133                         baselen = cp - path;
1134                 strbuf_setlen(&sb, 0);
1135                 strbuf_add(&sb, path, baselen);
1136                 if (!is_directory(sb.buf))
1137                         break;
1138                 if (simplify_away(sb.buf, sb.len, simplify))
1139                         break;
1140                 if (treat_one_path(dir, &sb, simplify,
1141                                    DT_DIR, NULL) == path_ignored)
1142                         break; /* do not recurse into it */
1143                 if (len <= baselen) {
1144                         rc = 1;
1145                         break; /* finished checking */
1146                 }
1147         }
1148         strbuf_release(&sb);
1149         return rc;
1150 }
1151
1152 int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
1153 {
1154         struct path_simplify *simplify;
1155
1156         if (has_symlink_leading_path(path, len))
1157                 return dir->nr;
1158
1159         simplify = create_simplify(pathspec);
1160         if (!len || treat_leading_path(dir, path, len, simplify))
1161                 read_directory_recursive(dir, path, len, 0, simplify);
1162         free_simplify(simplify);
1163         qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
1164         qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
1165         return dir->nr;
1166 }
1167
1168 int file_exists(const char *f)
1169 {
1170         struct stat sb;
1171         return lstat(f, &sb) == 0;
1172 }
1173
1174 /*
1175  * Given two normalized paths (a trailing slash is ok), if subdir is
1176  * outside dir, return -1.  Otherwise return the offset in subdir that
1177  * can be used as relative path to dir.
1178  */
1179 int dir_inside_of(const char *subdir, const char *dir)
1180 {
1181         int offset = 0;
1182
1183         assert(dir && subdir && *dir && *subdir);
1184
1185         while (*dir && *subdir && *dir == *subdir) {
1186                 dir++;
1187                 subdir++;
1188                 offset++;
1189         }
1190
1191         /* hel[p]/me vs hel[l]/yeah */
1192         if (*dir && *subdir)
1193                 return -1;
1194
1195         if (!*subdir)
1196                 return !*dir ? offset : -1; /* same dir */
1197
1198         /* foo/[b]ar vs foo/[] */
1199         if (is_dir_sep(dir[-1]))
1200                 return is_dir_sep(subdir[-1]) ? offset : -1;
1201
1202         /* foo[/]bar vs foo[] */
1203         return is_dir_sep(*subdir) ? offset + 1 : -1;
1204 }
1205
1206 int is_inside_dir(const char *dir)
1207 {
1208         char cwd[PATH_MAX];
1209         if (!dir)
1210                 return 0;
1211         if (!getcwd(cwd, sizeof(cwd)))
1212                 die_errno("can't find the current directory");
1213         return dir_inside_of(cwd, dir) >= 0;
1214 }
1215
1216 int is_empty_dir(const char *path)
1217 {
1218         DIR *dir = opendir(path);
1219         struct dirent *e;
1220         int ret = 1;
1221
1222         if (!dir)
1223                 return 0;
1224
1225         while ((e = readdir(dir)) != NULL)
1226                 if (!is_dot_or_dotdot(e->d_name)) {
1227                         ret = 0;
1228                         break;
1229                 }
1230
1231         closedir(dir);
1232         return ret;
1233 }
1234
1235 static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
1236 {
1237         DIR *dir;
1238         struct dirent *e;
1239         int ret = 0, original_len = path->len, len, kept_down = 0;
1240         int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
1241         int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
1242         unsigned char submodule_head[20];
1243
1244         if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
1245             !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
1246                 /* Do not descend and nuke a nested git work tree. */
1247                 if (kept_up)
1248                         *kept_up = 1;
1249                 return 0;
1250         }
1251
1252         flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
1253         dir = opendir(path->buf);
1254         if (!dir) {
1255                 /* an empty dir could be removed even if it is unreadble */
1256                 if (!keep_toplevel)
1257                         return rmdir(path->buf);
1258                 else
1259                         return -1;
1260         }
1261         if (path->buf[original_len - 1] != '/')
1262                 strbuf_addch(path, '/');
1263
1264         len = path->len;
1265         while ((e = readdir(dir)) != NULL) {
1266                 struct stat st;
1267                 if (is_dot_or_dotdot(e->d_name))
1268                         continue;
1269
1270                 strbuf_setlen(path, len);
1271                 strbuf_addstr(path, e->d_name);
1272                 if (lstat(path->buf, &st))
1273                         ; /* fall thru */
1274                 else if (S_ISDIR(st.st_mode)) {
1275                         if (!remove_dir_recurse(path, flag, &kept_down))
1276                                 continue; /* happy */
1277                 } else if (!only_empty && !unlink(path->buf))
1278                         continue; /* happy, too */
1279
1280                 /* path too long, stat fails, or non-directory still exists */
1281                 ret = -1;
1282                 break;
1283         }
1284         closedir(dir);
1285
1286         strbuf_setlen(path, original_len);
1287         if (!ret && !keep_toplevel && !kept_down)
1288                 ret = rmdir(path->buf);
1289         else if (kept_up)
1290                 /*
1291                  * report the uplevel that it is not an error that we
1292                  * did not rmdir() our directory.
1293                  */
1294                 *kept_up = !ret;
1295         return ret;
1296 }
1297
1298 int remove_dir_recursively(struct strbuf *path, int flag)
1299 {
1300         return remove_dir_recurse(path, flag, NULL);
1301 }
1302
1303 void setup_standard_excludes(struct dir_struct *dir)
1304 {
1305         const char *path;
1306         char *xdg_path;
1307
1308         dir->exclude_per_dir = ".gitignore";
1309         path = git_path("info/exclude");
1310         if (!excludes_file) {
1311                 home_config_paths(NULL, &xdg_path, "ignore");
1312                 excludes_file = xdg_path;
1313         }
1314         if (!access(path, R_OK))
1315                 add_excludes_from_file(dir, path);
1316         if (excludes_file && !access(excludes_file, R_OK))
1317                 add_excludes_from_file(dir, excludes_file);
1318 }
1319
1320 int remove_path(const char *name)
1321 {
1322         char *slash;
1323
1324         if (unlink(name) && errno != ENOENT)
1325                 return -1;
1326
1327         slash = strrchr(name, '/');
1328         if (slash) {
1329                 char *dirs = xstrdup(name);
1330                 slash = dirs + (slash - name);
1331                 do {
1332                         *slash = '\0';
1333                 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
1334                 free(dirs);
1335         }
1336         return 0;
1337 }
1338
1339 static int pathspec_item_cmp(const void *a_, const void *b_)
1340 {
1341         struct pathspec_item *a, *b;
1342
1343         a = (struct pathspec_item *)a_;
1344         b = (struct pathspec_item *)b_;
1345         return strcmp(a->match, b->match);
1346 }
1347
1348 int init_pathspec(struct pathspec *pathspec, const char **paths)
1349 {
1350         const char **p = paths;
1351         int i;
1352
1353         memset(pathspec, 0, sizeof(*pathspec));
1354         if (!p)
1355                 return 0;
1356         while (*p)
1357                 p++;
1358         pathspec->raw = paths;
1359         pathspec->nr = p - paths;
1360         if (!pathspec->nr)
1361                 return 0;
1362
1363         pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1364         for (i = 0; i < pathspec->nr; i++) {
1365                 struct pathspec_item *item = pathspec->items+i;
1366                 const char *path = paths[i];
1367
1368                 item->match = path;
1369                 item->len = strlen(path);
1370                 item->use_wildcard = !no_wildcard(path);
1371                 if (item->use_wildcard)
1372                         pathspec->has_wildcard = 1;
1373         }
1374
1375         qsort(pathspec->items, pathspec->nr,
1376               sizeof(struct pathspec_item), pathspec_item_cmp);
1377
1378         return 0;
1379 }
1380
1381 void free_pathspec(struct pathspec *pathspec)
1382 {
1383         free(pathspec->items);
1384         pathspec->items = NULL;
1385 }