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 static 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 void path_exclude_check_init(struct path_exclude_check *check,
 577                              struct dir_struct *dir)
 578 {
 579         check->dir = dir;
 580         strbuf_init(&check->path, 256);
 581 }
 582
 583 void path_exclude_check_clear(struct path_exclude_check *check)
 584 {
 585         strbuf_release(&check->path);
 586 }
 587
 588 /*
 589  * Is this name excluded?  This is for a caller like show_files() that
 590  * do not honor directory hierarchy and iterate through paths that are
 591  * possibly in an ignored directory.
 592  *
 593  * A path to a directory known to be excluded is left in check->path to
 594  * optimize for repeated checks for files in the same excluded directory.
 595  */
 596 int path_excluded(struct path_exclude_check *check,
 597                   const char *name, int namelen, int *dtype)
 598 {
 599         int i;
 600         struct strbuf *path = &check->path;
 601
 602         /*
 603          * we allow the caller to pass namelen as an optimization; it
 604          * must match the length of the name, as we eventually call
 605          * excluded() on the whole name string.
 606          */
 607         if (namelen < 0)
 608                 namelen = strlen(name);
 609
 610         if (path->len &&
 611             path->len <= namelen &&
 612             !memcmp(name, path->buf, path->len) &&
 613             (!name[path->len] || name[path->len] == '/'))
 614                 return 1;
 615
 616         strbuf_setlen(path, 0);
 617         for (i = 0; name[i]; i++) {
 618                 int ch = name[i];
 619
 620                 if (ch == '/') {
 621                         int dt = DT_DIR;
 622                         if (excluded(check->dir, path->buf, &dt))
 623                                 return 1;
 624                 }
 625                 strbuf_addch(path, ch);
 626         }
 627
 628         /* An entry in the index; cannot be a directory with subentries */
 629         strbuf_setlen(path, 0);
 630
 631         return excluded(check->dir, name, dtype);
 632 }
 633
 634 static struct dir_entry *dir_entry_new(const char *pathname, int len)
 635 {
 636         struct dir_entry *ent;
 637
 638         ent = xmalloc(sizeof(*ent) + len + 1);
 639         ent->len = len;
 640         memcpy(ent->name, pathname, len);
 641         ent->name[len] = 0;
 642         return ent;
 643 }
 644
 645 static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
 646 {
 647         if (cache_name_exists(pathname, len, ignore_case))
 648                 return NULL;
 649
 650         ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
 651         return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
 652 }
 653
 654 struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
 655 {
 656         if (!cache_name_is_other(pathname, len))
 657                 return NULL;
 658
 659         ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
 660         return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
 661 }
 662
 663 enum exist_status {
 664         index_nonexistent = 0,
 665         index_directory,
 666         index_gitdir
 667 };
 668
 669 /*
 670  * Do not use the alphabetically stored index to look up
 671  * the directory name; instead, use the case insensitive
 672  * name hash.
 673  */
 674 static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
 675 {
 676         struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
 677         unsigned char endchar;
 678
 679         if (!ce)
 680                 return index_nonexistent;
 681         endchar = ce->name[len];
 682
 683         /*
 684          * The cache_entry structure returned will contain this dirname
 685          * and possibly additional path components.
 686          */
 687         if (endchar == '/')
 688                 return index_directory;
 689
 690         /*
 691          * If there are no additional path components, then this cache_entry
 692          * represents a submodule.  Submodules, despite being directories,
 693          * are stored in the cache without a closing slash.
 694          */
 695         if (!endchar && S_ISGITLINK(ce->ce_mode))
 696                 return index_gitdir;
 697
 698         /* This should never be hit, but it exists just in case. */
 699         return index_nonexistent;
 700 }
 701
 702 /*
 703  * The index sorts alphabetically by entry name, which
 704  * means that a gitlink sorts as '\0' at the end, while
 705  * a directory (which is defined not as an entry, but as
 706  * the files it contains) will sort with the '/' at the
 707  * end.
 708  */
 709 static enum exist_status directory_exists_in_index(const char *dirname, int len)
 710 {
 711         int pos;
 712
 713         if (ignore_case)
 714                 return directory_exists_in_index_icase(dirname, len);
 715
 716         pos = cache_name_pos(dirname, len);
 717         if (pos < 0)
 718                 pos = -pos-1;
 719         while (pos < active_nr) {
 720                 struct cache_entry *ce = active_cache[pos++];
 721                 unsigned char endchar;
 722
 723                 if (strncmp(ce->name, dirname, len))
 724                         break;
 725                 endchar = ce->name[len];
 726                 if (endchar > '/')
 727                         break;
 728                 if (endchar == '/')
 729                         return index_directory;
 730                 if (!endchar && S_ISGITLINK(ce->ce_mode))
 731                         return index_gitdir;
 732         }
 733         return index_nonexistent;
 734 }
 735
 736 /*
 737  * When we find a directory when traversing the filesystem, we
 738  * have three distinct cases:
 739  *
 740  *  - ignore it
 741  *  - see it as a directory
 742  *  - recurse into it
 743  *
 744  * and which one we choose depends on a combination of existing
 745  * git index contents and the flags passed into the directory
 746  * traversal routine.
 747  *
 748  * Case 1: If we *already* have entries in the index under that
 749  * directory name, we always recurse into the directory to see
 750  * all the files.
 751  *
 752  * Case 2: If we *already* have that directory name as a gitlink,
 753  * we always continue to see it as a gitlink, regardless of whether
 754  * there is an actual git directory there or not (it might not
 755  * be checked out as a subproject!)
 756  *
 757  * Case 3: if we didn't have it in the index previously, we
 758  * have a few sub-cases:
 759  *
 760  *  (a) if "show_other_directories" is true, we show it as
 761  *      just a directory, unless "hide_empty_directories" is
 762  *      also true and the directory is empty, in which case
 763  *      we just ignore it entirely.
 764  *  (b) if it looks like a git directory, and we don't have
 765  *      'no_gitlinks' set we treat it as a gitlink, and show it
 766  *      as a directory.
 767  *  (c) otherwise, we recurse into it.
 768  */
 769 enum directory_treatment {
 770         show_directory,
 771         ignore_directory,
 772         recurse_into_directory
 773 };
 774
 775 static enum directory_treatment treat_directory(struct dir_struct *dir,
 776         const char *dirname, int len,
 777         const struct path_simplify *simplify)
 778 {
 779         /* The "len-1" is to strip the final '/' */
 780         switch (directory_exists_in_index(dirname, len-1)) {
 781         case index_directory:
 782                 return recurse_into_directory;
 783
 784         case index_gitdir:
 785                 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 786                         return ignore_directory;
 787                 return show_directory;
 788
 789         case index_nonexistent:
 790                 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
 791                         break;
 792                 if (!(dir->flags & DIR_NO_GITLINKS)) {
 793                         unsigned char sha1[20];
 794                         if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
 795                                 return show_directory;
 796                 }
 797                 return recurse_into_directory;
 798         }
 799
 800         /* This is the "show_other_directories" case */
 801         if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
 802                 return show_directory;
 803         if (!read_directory_recursive(dir, dirname, len, 1, simplify))
 804                 return ignore_directory;
 805         return show_directory;
 806 }
 807
 808 /*
 809  * This is an inexact early pruning of any recursive directory
 810  * reading - if the path cannot possibly be in the pathspec,
 811  * return true, and we'll skip it early.
 812  */
 813 static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
 814 {
 815         if (simplify) {
 816                 for (;;) {
 817                         const char *match = simplify->path;
 818                         int len = simplify->len;
 819
 820                         if (!match)
 821                                 break;
 822                         if (len > pathlen)
 823                                 len = pathlen;
 824                         if (!memcmp(path, match, len))
 825                                 return 0;
 826                         simplify++;
 827                 }
 828                 return 1;
 829         }
 830         return 0;
 831 }
 832
 833 /*
 834  * This function tells us whether an excluded path matches a
 835  * list of "interesting" pathspecs. That is, whether a path matched
 836  * by any of the pathspecs could possibly be ignored by excluding
 837  * the specified path. This can happen if:
 838  *
 839  *   1. the path is mentioned explicitly in the pathspec
 840  *
 841  *   2. the path is a directory prefix of some element in the
 842  *      pathspec
 843  */
 844 static int exclude_matches_pathspec(const char *path, int len,
 845                 const struct path_simplify *simplify)
 846 {
 847         if (simplify) {
 848                 for (; simplify->path; simplify++) {
 849                         if (len == simplify->len
 850                             && !memcmp(path, simplify->path, len))
 851                                 return 1;
 852                         if (len < simplify->len
 853                             && simplify->path[len] == '/'
 854                             && !memcmp(path, simplify->path, len))
 855                                 return 1;
 856                 }
 857         }
 858         return 0;
 859 }
 860
 861 static int get_index_dtype(const char *path, int len)
 862 {
 863         int pos;
 864         struct cache_entry *ce;
 865
 866         ce = cache_name_exists(path, len, 0);
 867         if (ce) {
 868                 if (!ce_uptodate(ce))
 869                         return DT_UNKNOWN;
 870                 if (S_ISGITLINK(ce->ce_mode))
 871                         return DT_DIR;
 872                 /*
 873                  * Nobody actually cares about the
 874                  * difference between DT_LNK and DT_REG
 875                  */
 876                 return DT_REG;
 877         }
 878
 879         /* Try to look it up as a directory */
 880         pos = cache_name_pos(path, len);
 881         if (pos >= 0)
 882                 return DT_UNKNOWN;
 883         pos = -pos-1;
 884         while (pos < active_nr) {
 885                 ce = active_cache[pos++];
 886                 if (strncmp(ce->name, path, len))
 887                         break;
 888                 if (ce->name[len] > '/')
 889                         break;
 890                 if (ce->name[len] < '/')
 891                         continue;
 892                 if (!ce_uptodate(ce))
 893                         break;  /* continue? */
 894                 return DT_DIR;
 895         }
 896         return DT_UNKNOWN;
 897 }
 898
 899 static int get_dtype(struct dirent *de, const char *path, int len)
 900 {
 901         int dtype = de ? DTYPE(de) : DT_UNKNOWN;
 902         struct stat st;
 903
 904         if (dtype != DT_UNKNOWN)
 905                 return dtype;
 906         dtype = get_index_dtype(path, len);
 907         if (dtype != DT_UNKNOWN)
 908                 return dtype;
 909         if (lstat(path, &st))
 910                 return dtype;
 911         if (S_ISREG(st.st_mode))
 912                 return DT_REG;
 913         if (S_ISDIR(st.st_mode))
 914                 return DT_DIR;
 915         if (S_ISLNK(st.st_mode))
 916                 return DT_LNK;
 917         return dtype;
 918 }
 919
 920 enum path_treatment {
 921         path_ignored,
 922         path_handled,
 923         path_recurse
 924 };
 925
 926 static enum path_treatment treat_one_path(struct dir_struct *dir,
 927                                           struct strbuf *path,
 928                                           const struct path_simplify *simplify,
 929                                           int dtype, struct dirent *de)
 930 {
 931         int exclude = excluded(dir, path->buf, &dtype);
 932         if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
 933             && exclude_matches_pathspec(path->buf, path->len, simplify))
 934                 dir_add_ignored(dir, path->buf, path->len);
 935
 936         /*
 937          * Excluded? If we don't explicitly want to show
 938          * ignored files, ignore it
 939          */
 940         if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
 941                 return path_ignored;
 942
 943         if (dtype == DT_UNKNOWN)
 944                 dtype = get_dtype(de, path->buf, path->len);
 945
 946         /*
 947          * Do we want to see just the ignored files?
 948          * We still need to recurse into directories,
 949          * even if we don't ignore them, since the
 950          * directory may contain files that we do..
 951          */
 952         if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) {
 953                 if (dtype != DT_DIR)
 954                         return path_ignored;
 955         }
 956
 957         switch (dtype) {
 958         default:
 959                 return path_ignored;
 960         case DT_DIR:
 961                 strbuf_addch(path, '/');
 962                 switch (treat_directory(dir, path->buf, path->len, simplify)) {
 963                 case show_directory:
 964                         if (exclude != !!(dir->flags
 965                                           & DIR_SHOW_IGNORED))
 966                                 return path_ignored;
 967                         break;
 968                 case recurse_into_directory:
 969                         return path_recurse;
 970                 case ignore_directory:
 971                         return path_ignored;
 972                 }
 973                 break;
 974         case DT_REG:
 975         case DT_LNK:
 976                 break;
 977         }
 978         return path_handled;
 979 }
 980
 981 static enum path_treatment treat_path(struct dir_struct *dir,
 982                                       struct dirent *de,
 983                                       struct strbuf *path,
 984                                       int baselen,
 985                                       const struct path_simplify *simplify)
 986 {
 987         int dtype;
 988
 989         if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
 990                 return path_ignored;
 991         strbuf_setlen(path, baselen);
 992         strbuf_addstr(path, de->d_name);
 993         if (simplify_away(path->buf, path->len, simplify))
 994                 return path_ignored;
 995
 996         dtype = DTYPE(de);
 997         return treat_one_path(dir, path, simplify, dtype, de);
 998 }
 999
1000 /*
1001  * Read a directory tree. We currently ignore anything but
1002  * directories, regular files and symlinks. That's because git
1003  * doesn't handle them at all yet. Maybe that will change some
1004  * day.
1005  *
1006  * Also, we ignore the name ".git" (even if it is not a directory).
1007  * That likely will not change.
1008  */
1009 static int read_directory_recursive(struct dir_struct *dir,
1010                                     const char *base, int baselen,
1011                                     int check_only,
1012                                     const struct path_simplify *simplify)
1013 {
1014         DIR *fdir;
1015         int contents = 0;
1016         struct dirent *de;
1017         struct strbuf path = STRBUF_INIT;
1018
1019         strbuf_add(&path, base, baselen);
1020
1021         fdir = opendir(path.len ? path.buf : ".");
1022         if (!fdir)
1023                 goto out;
1024
1025         while ((de = readdir(fdir)) != NULL) {
1026                 switch (treat_path(dir, de, &path, baselen, simplify)) {
1027                 case path_recurse:
1028                         contents += read_directory_recursive(dir, path.buf,
1029                                                              path.len, 0,
1030                                                              simplify);
1031                         continue;
1032                 case path_ignored:
1033                         continue;
1034                 case path_handled:
1035                         break;
1036                 }
1037                 contents++;
1038                 if (check_only)
1039                         break;
1040                 dir_add_name(dir, path.buf, path.len);
1041         }
1042         closedir(fdir);
1043  out:
1044         strbuf_release(&path);
1045
1046         return contents;
1047 }
1048
1049 static int cmp_name(const void *p1, const void *p2)
1050 {
1051         const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1052         const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1053
1054         return cache_name_compare(e1->name, e1->len,
1055                                   e2->name, e2->len);
1056 }
1057
1058 /*
1059  * Return the length of the "simple" part of a path match limiter.
1060  */
1061 static int simple_length(const char *match)
1062 {
1063         int len = -1;
1064
1065         for (;;) {
1066                 unsigned char c = *match++;
1067                 len++;
1068                 if (c == '\0' || is_glob_special(c))
1069                         return len;
1070         }
1071 }
1072
1073 static struct path_simplify *create_simplify(const char **pathspec)
1074 {
1075         int nr, alloc = 0;
1076         struct path_simplify *simplify = NULL;
1077
1078         if (!pathspec)
1079                 return NULL;
1080
1081         for (nr = 0 ; ; nr++) {
1082                 const char *match;
1083                 if (nr >= alloc) {
1084                         alloc = alloc_nr(alloc);
1085                         simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1086                 }
1087                 match = *pathspec++;
1088                 if (!match)
1089                         break;
1090                 simplify[nr].path = match;
1091                 simplify[nr].len = simple_length(match);
1092         }
1093         simplify[nr].path = NULL;
1094         simplify[nr].len = 0;
1095         return simplify;
1096 }
1097
1098 static void free_simplify(struct path_simplify *simplify)
1099 {
1100         free(simplify);
1101 }
1102
1103 static int treat_leading_path(struct dir_struct *dir,
1104                               const char *path, int len,
1105                               const struct path_simplify *simplify)
1106 {
1107         struct strbuf sb = STRBUF_INIT;
1108         int baselen, rc = 0;
1109         const char *cp;
1110
1111         while (len && path[len - 1] == '/')
1112                 len--;
1113         if (!len)
1114                 return 1;
1115         baselen = 0;
1116         while (1) {
1117                 cp = path + baselen + !!baselen;
1118                 cp = memchr(cp, '/', path + len - cp);
1119                 if (!cp)
1120                         baselen = len;
1121                 else
1122                         baselen = cp - path;
1123                 strbuf_setlen(&sb, 0);
1124                 strbuf_add(&sb, path, baselen);
1125                 if (!is_directory(sb.buf))
1126                         break;
1127                 if (simplify_away(sb.buf, sb.len, simplify))
1128                         break;
1129                 if (treat_one_path(dir, &sb, simplify,
1130                                    DT_DIR, NULL) == path_ignored)
1131                         break; /* do not recurse into it */
1132                 if (len <= baselen) {
1133                         rc = 1;
1134                         break; /* finished checking */
1135                 }
1136         }
1137         strbuf_release(&sb);
1138         return rc;
1139 }
1140
1141 int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
1142 {
1143         struct path_simplify *simplify;
1144
1145         if (has_symlink_leading_path(path, len))
1146                 return dir->nr;
1147
1148         simplify = create_simplify(pathspec);
1149         if (!len || treat_leading_path(dir, path, len, simplify))
1150                 read_directory_recursive(dir, path, len, 0, simplify);
1151         free_simplify(simplify);
1152         qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
1153         qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
1154         return dir->nr;
1155 }
1156
1157 int file_exists(const char *f)
1158 {
1159         struct stat sb;
1160         return lstat(f, &sb) == 0;
1161 }
1162
1163 /*
1164  * Given two normalized paths (a trailing slash is ok), if subdir is
1165  * outside dir, return -1.  Otherwise return the offset in subdir that
1166  * can be used as relative path to dir.
1167  */
1168 int dir_inside_of(const char *subdir, const char *dir)
1169 {
1170         int offset = 0;
1171
1172         assert(dir && subdir && *dir && *subdir);
1173
1174         while (*dir && *subdir && *dir == *subdir) {
1175                 dir++;
1176                 subdir++;
1177                 offset++;
1178         }
1179
1180         /* hel[p]/me vs hel[l]/yeah */
1181         if (*dir && *subdir)
1182                 return -1;
1183
1184         if (!*subdir)
1185                 return !*dir ? offset : -1; /* same dir */
1186
1187         /* foo/[b]ar vs foo/[] */
1188         if (is_dir_sep(dir[-1]))
1189                 return is_dir_sep(subdir[-1]) ? offset : -1;
1190
1191         /* foo[/]bar vs foo[] */
1192         return is_dir_sep(*subdir) ? offset + 1 : -1;
1193 }
1194
1195 int is_inside_dir(const char *dir)
1196 {
1197         char cwd[PATH_MAX];
1198         if (!dir)
1199                 return 0;
1200         if (!getcwd(cwd, sizeof(cwd)))
1201                 die_errno("can't find the current directory");
1202         return dir_inside_of(cwd, dir) >= 0;
1203 }
1204
1205 int is_empty_dir(const char *path)
1206 {
1207         DIR *dir = opendir(path);
1208         struct dirent *e;
1209         int ret = 1;
1210
1211         if (!dir)
1212                 return 0;
1213
1214         while ((e = readdir(dir)) != NULL)
1215                 if (!is_dot_or_dotdot(e->d_name)) {
1216                         ret = 0;
1217                         break;
1218                 }
1219
1220         closedir(dir);
1221         return ret;
1222 }
1223
1224 static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
1225 {
1226         DIR *dir;
1227         struct dirent *e;
1228         int ret = 0, original_len = path->len, len, kept_down = 0;
1229         int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
1230         int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
1231         unsigned char submodule_head[20];
1232
1233         if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
1234             !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
1235                 /* Do not descend and nuke a nested git work tree. */
1236                 if (kept_up)
1237                         *kept_up = 1;
1238                 return 0;
1239         }
1240
1241         flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
1242         dir = opendir(path->buf);
1243         if (!dir) {
1244                 /* an empty dir could be removed even if it is unreadble */
1245                 if (!keep_toplevel)
1246                         return rmdir(path->buf);
1247                 else
1248                         return -1;
1249         }
1250         if (path->buf[original_len - 1] != '/')
1251                 strbuf_addch(path, '/');
1252
1253         len = path->len;
1254         while ((e = readdir(dir)) != NULL) {
1255                 struct stat st;
1256                 if (is_dot_or_dotdot(e->d_name))
1257                         continue;
1258
1259                 strbuf_setlen(path, len);
1260                 strbuf_addstr(path, e->d_name);
1261                 if (lstat(path->buf, &st))
1262                         ; /* fall thru */
1263                 else if (S_ISDIR(st.st_mode)) {
1264                         if (!remove_dir_recurse(path, flag, &kept_down))
1265                                 continue; /* happy */
1266                 } else if (!only_empty && !unlink(path->buf))
1267                         continue; /* happy, too */
1268
1269                 /* path too long, stat fails, or non-directory still exists */
1270                 ret = -1;
1271                 break;
1272         }
1273         closedir(dir);
1274
1275         strbuf_setlen(path, original_len);
1276         if (!ret && !keep_toplevel && !kept_down)
1277                 ret = rmdir(path->buf);
1278         else if (kept_up)
1279                 /*
1280                  * report the uplevel that it is not an error that we
1281                  * did not rmdir() our directory.
1282                  */
1283                 *kept_up = !ret;
1284         return ret;
1285 }
1286
1287 int remove_dir_recursively(struct strbuf *path, int flag)
1288 {
1289         return remove_dir_recurse(path, flag, NULL);
1290 }
1291
1292 void setup_standard_excludes(struct dir_struct *dir)
1293 {
1294         const char *path;
1295
1296         dir->exclude_per_dir = ".gitignore";
1297         path = git_path("info/exclude");
1298         if (!access(path, R_OK))
1299                 add_excludes_from_file(dir, path);
1300         if (excludes_file && !access(excludes_file, R_OK))
1301                 add_excludes_from_file(dir, excludes_file);
1302 }
1303
1304 int remove_path(const char *name)
1305 {
1306         char *slash;
1307
1308         if (unlink(name) && errno != ENOENT)
1309                 return -1;
1310
1311         slash = strrchr(name, '/');
1312         if (slash) {
1313                 char *dirs = xstrdup(name);
1314                 slash = dirs + (slash - name);
1315                 do {
1316                         *slash = '\0';
1317                 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
1318                 free(dirs);
1319         }
1320         return 0;
1321 }
1322
1323 static int pathspec_item_cmp(const void *a_, const void *b_)
1324 {
1325         struct pathspec_item *a, *b;
1326
1327         a = (struct pathspec_item *)a_;
1328         b = (struct pathspec_item *)b_;
1329         return strcmp(a->match, b->match);
1330 }
1331
1332 int init_pathspec(struct pathspec *pathspec, const char **paths)
1333 {
1334         const char **p = paths;
1335         int i;
1336
1337         memset(pathspec, 0, sizeof(*pathspec));
1338         if (!p)
1339                 return 0;
1340         while (*p)
1341                 p++;
1342         pathspec->raw = paths;
1343         pathspec->nr = p - paths;
1344         if (!pathspec->nr)
1345                 return 0;
1346
1347         pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1348         for (i = 0; i < pathspec->nr; i++) {
1349                 struct pathspec_item *item = pathspec->items+i;
1350                 const char *path = paths[i];
1351
1352                 item->match = path;
1353                 item->len = strlen(path);
1354                 item->use_wildcard = !no_wildcard(path);
1355                 if (item->use_wildcard)
1356                         pathspec->has_wildcard = 1;
1357         }
1358
1359         qsort(pathspec->items, pathspec->nr,
1360               sizeof(struct pathspec_item), pathspec_item_cmp);
1361
1362         return 0;
1363 }
1364
1365 void free_pathspec(struct pathspec *pathspec)
1366 {
1367         free(pathspec->items);
1368         pathspec->items = NULL;
1369 }