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,
  18         const char *path, const char *base, int baselen,
  19         int check_only, const struct path_simplify *simplify);
  20
  21 int common_prefix(const char **pathspec)
  22 {
  23         const char *path, *slash, *next;
  24         int prefix;
  25
  26         if (!pathspec)
  27                 return 0;
  28
  29         path = *pathspec;
  30         slash = strrchr(path, '/');
  31         if (!slash)
  32                 return 0;
  33
  34         prefix = slash - path + 1;
  35         while ((next = *++pathspec) != NULL) {
  36                 int len = strlen(next);
  37                 if (len >= prefix && !memcmp(path, next, prefix))
  38                         continue;
  39                 len = prefix - 1;
  40                 for (;;) {
  41                         if (!len)
  42                                 return 0;
  43                         if (next[--len] != '/')
  44                                 continue;
  45                         if (memcmp(path, next, len+1))
  46                                 continue;
  47                         prefix = len + 1;
  48                         break;
  49                 }
  50         }
  51         return prefix;
  52 }
  53
  54 /*
  55  * Does 'match' matches the given name?
  56  * A match is found if
  57  *
  58  * (1) the 'match' string is leading directory of 'name', or
  59  * (2) the 'match' string is a wildcard and matches 'name', or
  60  * (3) the 'match' string is exactly the same as 'name'.
  61  *
  62  * and the return value tells which case it was.
  63  *
  64  * It returns 0 when there is no match.
  65  */
  66 static int match_one(const char *match, const char *name, int namelen)
  67 {
  68         int matchlen;
  69
  70         /* If the match was just the prefix, we matched */
  71         matchlen = strlen(match);
  72         if (!matchlen)
  73                 return MATCHED_RECURSIVELY;
  74
  75         /*
  76          * If we don't match the matchstring exactly,
  77          * we need to match by fnmatch
  78          */
  79         if (strncmp(match, name, matchlen))
  80                 return !fnmatch(match, name, 0) ? MATCHED_FNMATCH : 0;
  81
  82         if (!name[matchlen])
  83                 return MATCHED_EXACTLY;
  84         if (match[matchlen-1] == '/' || name[matchlen] == '/')
  85                 return MATCHED_RECURSIVELY;
  86         return 0;
  87 }
  88
  89 /*
  90  * Given a name and a list of pathspecs, see if the name matches
  91  * any of the pathspecs.  The caller is also interested in seeing
  92  * all pathspec matches some names it calls this function with
  93  * (otherwise the user could have mistyped the unmatched pathspec),
  94  * and a mark is left in seen[] array for pathspec element that
  95  * actually matched anything.
  96  */
  97 int match_pathspec(const char **pathspec, const char *name, int namelen, int prefix, char *seen)
  98 {
  99         int retval;
 100         const char *match;
 101
 102         name += prefix;
 103         namelen -= prefix;
 104
 105         for (retval = 0; (match = *pathspec++) != NULL; seen++) {
 106                 int how;
 107                 if (retval && *seen == MATCHED_EXACTLY)
 108                         continue;
 109                 match += prefix;
 110                 how = match_one(match, name, namelen);
 111                 if (how) {
 112                         if (retval < how)
 113                                 retval = how;
 114                         if (*seen < how)
 115                                 *seen = how;
 116                 }
 117         }
 118         return retval;
 119 }
 120
 121 void add_exclude(const char *string, const char *base,
 122                  int baselen, struct exclude_list *which)
 123 {
 124         struct exclude *x = xmalloc(sizeof (*x));
 125
 126         x->pattern = string;
 127         x->base = base;
 128         x->baselen = baselen;
 129         if (which->nr == which->alloc) {
 130                 which->alloc = alloc_nr(which->alloc);
 131                 which->excludes = xrealloc(which->excludes,
 132                                            which->alloc * sizeof(x));
 133         }
 134         which->excludes[which->nr++] = x;
 135 }
 136
 137 static int add_excludes_from_file_1(const char *fname,
 138                                     const char *base,
 139                                     int baselen,
 140                                     struct exclude_list *which)
 141 {
 142         struct stat st;
 143         int fd, i;
 144         size_t size;
 145         char *buf, *entry;
 146
 147         fd = open(fname, O_RDONLY);
 148         if (fd < 0 || fstat(fd, &st) < 0)
 149                 goto err;
 150         size = xsize_t(st.st_size);
 151         if (size == 0) {
 152                 close(fd);
 153                 return 0;
 154         }
 155         buf = xmalloc(size+1);
 156         if (read_in_full(fd, buf, size) != size)
 157                 goto err;
 158         close(fd);
 159
 160         buf[size++] = '\n';
 161         entry = buf;
 162         for (i = 0; i < size; i++) {
 163                 if (buf[i] == '\n') {
 164                         if (entry != buf + i && entry[0] != '#') {
 165                                 buf[i - (i && buf[i-1] == '\r')] = 0;
 166                                 add_exclude(entry, base, baselen, which);
 167                         }
 168                         entry = buf + i + 1;
 169                 }
 170         }
 171         return 0;
 172
 173  err:
 174         if (0 <= fd)
 175                 close(fd);
 176         return -1;
 177 }
 178
 179 void add_excludes_from_file(struct dir_struct *dir, const char *fname)
 180 {
 181         if (add_excludes_from_file_1(fname, "", 0,
 182                                      &dir->exclude_list[EXC_FILE]) < 0)
 183                 die("cannot use %s as an exclude file", fname);
 184 }
 185
 186 int push_exclude_per_directory(struct dir_struct *dir, const char *base, int baselen)
 187 {
 188         char exclude_file[PATH_MAX];
 189         struct exclude_list *el = &dir->exclude_list[EXC_DIRS];
 190         int current_nr = el->nr;
 191
 192         if (dir->exclude_per_dir) {
 193                 memcpy(exclude_file, base, baselen);
 194                 strcpy(exclude_file + baselen, dir->exclude_per_dir);
 195                 add_excludes_from_file_1(exclude_file, base, baselen, el);
 196         }
 197         return current_nr;
 198 }
 199
 200 void pop_exclude_per_directory(struct dir_struct *dir, int stk)
 201 {
 202         struct exclude_list *el = &dir->exclude_list[EXC_DIRS];
 203
 204         while (stk < el->nr)
 205                 free(el->excludes[--el->nr]);
 206 }
 207
 208 /* Scan the list and let the last match determines the fate.
 209  * Return 1 for exclude, 0 for include and -1 for undecided.
 210  */
 211 static int excluded_1(const char *pathname,
 212                       int pathlen,
 213                       struct exclude_list *el)
 214 {
 215         int i;
 216
 217         if (el->nr) {
 218                 for (i = el->nr - 1; 0 <= i; i--) {
 219                         struct exclude *x = el->excludes[i];
 220                         const char *exclude = x->pattern;
 221                         int to_exclude = 1;
 222
 223                         if (*exclude == '!') {
 224                                 to_exclude = 0;
 225                                 exclude++;
 226                         }
 227
 228                         if (!strchr(exclude, '/')) {
 229                                 /* match basename */
 230                                 const char *basename = strrchr(pathname, '/');
 231                                 basename = (basename) ? basename+1 : pathname;
 232                                 if (fnmatch(exclude, basename, 0) == 0)
 233                                         return to_exclude;
 234                         }
 235                         else {
 236                                 /* match with FNM_PATHNAME:
 237                                  * exclude has base (baselen long) implicitly
 238                                  * in front of it.
 239                                  */
 240                                 int baselen = x->baselen;
 241                                 if (*exclude == '/')
 242                                         exclude++;
 243
 244                                 if (pathlen < baselen ||
 245                                     (baselen && pathname[baselen-1] != '/') ||
 246                                     strncmp(pathname, x->base, baselen))
 247                                     continue;
 248
 249                                 if (fnmatch(exclude, pathname+baselen,
 250                                             FNM_PATHNAME) == 0)
 251                                         return to_exclude;
 252                         }
 253                 }
 254         }
 255         return -1; /* undecided */
 256 }
 257
 258 int excluded(struct dir_struct *dir, const char *pathname)
 259 {
 260         int pathlen = strlen(pathname);
 261         int st;
 262
 263         for (st = EXC_CMDL; st <= EXC_FILE; st++) {
 264                 switch (excluded_1(pathname, pathlen, &dir->exclude_list[st])) {
 265                 case 0:
 266                         return 0;
 267                 case 1:
 268                         return 1;
 269                 }
 270         }
 271         return 0;
 272 }
 273
 274 struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
 275 {
 276         struct dir_entry *ent;
 277
 278         if (cache_name_pos(pathname, len) >= 0)
 279                 return NULL;
 280
 281         if (dir->nr == dir->alloc) {
 282                 int alloc = alloc_nr(dir->alloc);
 283                 dir->alloc = alloc;
 284                 dir->entries = xrealloc(dir->entries, alloc*sizeof(ent));
 285         }
 286         ent = xmalloc(sizeof(*ent) + len + 1);
 287         ent->ignored = ent->ignored_dir = 0;
 288         ent->len = len;
 289         memcpy(ent->name, pathname, len);
 290         ent->name[len] = 0;
 291         dir->entries[dir->nr++] = ent;
 292         return ent;
 293 }
 294
 295 enum exist_status {
 296         index_nonexistent = 0,
 297         index_directory,
 298         index_gitdir,
 299 };
 300
 301 /*
 302  * The index sorts alphabetically by entry name, which
 303  * means that a gitlink sorts as '\0' at the end, while
 304  * a directory (which is defined not as an entry, but as
 305  * the files it contains) will sort with the '/' at the
 306  * end.
 307  */
 308 static enum exist_status directory_exists_in_index(const char *dirname, int len)
 309 {
 310         int pos = cache_name_pos(dirname, len);
 311         if (pos < 0)
 312                 pos = -pos-1;
 313         while (pos < active_nr) {
 314                 struct cache_entry *ce = active_cache[pos++];
 315                 unsigned char endchar;
 316
 317                 if (strncmp(ce->name, dirname, len))
 318                         break;
 319                 endchar = ce->name[len];
 320                 if (endchar > '/')
 321                         break;
 322                 if (endchar == '/')
 323                         return index_directory;
 324                 if (!endchar && S_ISDIRLNK(ntohl(ce->ce_mode)))
 325                         return index_gitdir;
 326         }
 327         return index_nonexistent;
 328 }
 329
 330 /*
 331  * When we find a directory when traversing the filesystem, we
 332  * have three distinct cases:
 333  *
 334  *  - ignore it
 335  *  - see it as a directory
 336  *  - recurse into it
 337  *
 338  * and which one we choose depends on a combination of existing
 339  * git index contents and the flags passed into the directory
 340  * traversal routine.
 341  *
 342  * Case 1: If we *already* have entries in the index under that
 343  * directory name, we always recurse into the directory to see
 344  * all the files.
 345  *
 346  * Case 2: If we *already* have that directory name as a gitlink,
 347  * we always continue to see it as a gitlink, regardless of whether
 348  * there is an actual git directory there or not (it might not
 349  * be checked out as a subproject!)
 350  *
 351  * Case 3: if we didn't have it in the index previously, we
 352  * have a few sub-cases:
 353  *
 354  *  (a) if "show_other_directories" is true, we show it as
 355  *      just a directory, unless "hide_empty_directories" is
 356  *      also true and the directory is empty, in which case
 357  *      we just ignore it entirely.
 358  *  (b) if it looks like a git directory, and we don't have
 359  *      'no_dirlinks' set we treat it as a gitlink, and show it
 360  *      as a directory.
 361  *  (c) otherwise, we recurse into it.
 362  */
 363 enum directory_treatment {
 364         show_directory,
 365         ignore_directory,
 366         recurse_into_directory,
 367 };
 368
 369 static enum directory_treatment treat_directory(struct dir_struct *dir,
 370         const char *dirname, int len,
 371         const struct path_simplify *simplify)
 372 {
 373         /* The "len-1" is to strip the final '/' */
 374         switch (directory_exists_in_index(dirname, len-1)) {
 375         case index_directory:
 376                 return recurse_into_directory;
 377
 378         case index_gitdir:
 379                 if (dir->show_other_directories)
 380                         return ignore_directory;
 381                 return show_directory;
 382
 383         case index_nonexistent:
 384                 if (dir->show_other_directories)
 385                         break;
 386                 if (!dir->no_dirlinks) {
 387                         unsigned char sha1[20];
 388                         if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
 389                                 return show_directory;
 390                 }
 391                 return recurse_into_directory;
 392         }
 393
 394         /* This is the "show_other_directories" case */
 395         if (!dir->hide_empty_directories)
 396                 return show_directory;
 397         if (!read_directory_recursive(dir, dirname, dirname, len, 1, simplify))
 398                 return ignore_directory;
 399         return show_directory;
 400 }
 401
 402 /*
 403  * This is an inexact early pruning of any recursive directory
 404  * reading - if the path cannot possibly be in the pathspec,
 405  * return true, and we'll skip it early.
 406  */
 407 static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
 408 {
 409         if (simplify) {
 410                 for (;;) {
 411                         const char *match = simplify->path;
 412                         int len = simplify->len;
 413
 414                         if (!match)
 415                                 break;
 416                         if (len > pathlen)
 417                                 len = pathlen;
 418                         if (!memcmp(path, match, len))
 419                                 return 0;
 420                         simplify++;
 421                 }
 422                 return 1;
 423         }
 424         return 0;
 425 }
 426
 427 /*
 428  * Read a directory tree. We currently ignore anything but
 429  * directories, regular files and symlinks. That's because git
 430  * doesn't handle them at all yet. Maybe that will change some
 431  * day.
 432  *
 433  * Also, we ignore the name ".git" (even if it is not a directory).
 434  * That likely will not change.
 435  */
 436 static int read_directory_recursive(struct dir_struct *dir, const char *path, const char *base, int baselen, int check_only, const struct path_simplify *simplify)
 437 {
 438         DIR *fdir = opendir(path);
 439         int contents = 0;
 440
 441         if (fdir) {
 442                 int exclude_stk;
 443                 struct dirent *de;
 444                 char fullname[PATH_MAX + 1];
 445                 memcpy(fullname, base, baselen);
 446
 447                 exclude_stk = push_exclude_per_directory(dir, base, baselen);
 448
 449                 while ((de = readdir(fdir)) != NULL) {
 450                         int len;
 451
 452                         if ((de->d_name[0] == '.') &&
 453                             (de->d_name[1] == 0 ||
 454                              !strcmp(de->d_name + 1, ".") ||
 455                              !strcmp(de->d_name + 1, "git")))
 456                                 continue;
 457                         len = strlen(de->d_name);
 458                         /* Ignore overly long pathnames! */
 459                         if (len + baselen + 8 > sizeof(fullname))
 460                                 continue;
 461                         memcpy(fullname + baselen, de->d_name, len+1);
 462                         if (simplify_away(fullname, baselen + len, simplify))
 463                                 continue;
 464                         if (excluded(dir, fullname) != dir->show_ignored) {
 465                                 if (!dir->show_ignored || DTYPE(de) != DT_DIR) {
 466                                         continue;
 467                                 }
 468                         }
 469
 470                         switch (DTYPE(de)) {
 471                         struct stat st;
 472                         default:
 473                                 continue;
 474                         case DT_UNKNOWN:
 475                                 if (lstat(fullname, &st))
 476                                         continue;
 477                                 if (S_ISREG(st.st_mode) || S_ISLNK(st.st_mode))
 478                                         break;
 479                                 if (!S_ISDIR(st.st_mode))
 480                                         continue;
 481                                 /* fallthrough */
 482                         case DT_DIR:
 483                                 memcpy(fullname + baselen + len, "/", 2);
 484                                 len++;
 485                                 switch (treat_directory(dir, fullname, baselen + len, simplify)) {
 486                                 case show_directory:
 487                                         break;
 488                                 case recurse_into_directory:
 489                                         contents += read_directory_recursive(dir,
 490                                                 fullname, fullname, baselen + len, 0, simplify);
 491                                         continue;
 492                                 case ignore_directory:
 493                                         continue;
 494                                 }
 495                                 break;
 496                         case DT_REG:
 497                         case DT_LNK:
 498                                 break;
 499                         }
 500                         contents++;
 501                         if (check_only)
 502                                 goto exit_early;
 503                         else
 504                                 dir_add_name(dir, fullname, baselen + len);
 505                 }
 506 exit_early:
 507                 closedir(fdir);
 508
 509                 pop_exclude_per_directory(dir, exclude_stk);
 510         }
 511
 512         return contents;
 513 }
 514
 515 static int cmp_name(const void *p1, const void *p2)
 516 {
 517         const struct dir_entry *e1 = *(const struct dir_entry **)p1;
 518         const struct dir_entry *e2 = *(const struct dir_entry **)p2;
 519
 520         return cache_name_compare(e1->name, e1->len,
 521                                   e2->name, e2->len);
 522 }
 523
 524 /*
 525  * Return the length of the "simple" part of a path match limiter.
 526  */
 527 static int simple_length(const char *match)
 528 {
 529         const char special[256] = {
 530                 [0] = 1, ['?'] = 1,
 531                 ['\\'] = 1, ['*'] = 1,
 532                 ['['] = 1
 533         };
 534         int len = -1;
 535
 536         for (;;) {
 537                 unsigned char c = *match++;
 538                 len++;
 539                 if (special[c])
 540                         return len;
 541         }
 542 }
 543
 544 static struct path_simplify *create_simplify(const char **pathspec)
 545 {
 546         int nr, alloc = 0;
 547         struct path_simplify *simplify = NULL;
 548
 549         if (!pathspec)
 550                 return NULL;
 551
 552         for (nr = 0 ; ; nr++) {
 553                 const char *match;
 554                 if (nr >= alloc) {
 555                         alloc = alloc_nr(alloc);
 556                         simplify = xrealloc(simplify, alloc * sizeof(*simplify));
 557                 }
 558                 match = *pathspec++;
 559                 if (!match)
 560                         break;
 561                 simplify[nr].path = match;
 562                 simplify[nr].len = simple_length(match);
 563         }
 564         simplify[nr].path = NULL;
 565         simplify[nr].len = 0;
 566         return simplify;
 567 }
 568
 569 static void free_simplify(struct path_simplify *simplify)
 570 {
 571         if (simplify)
 572                 free(simplify);
 573 }
 574
 575 int read_directory(struct dir_struct *dir, const char *path, const char *base, int baselen, const char **pathspec)
 576 {
 577         struct path_simplify *simplify = create_simplify(pathspec);
 578
 579         /*
 580          * Make sure to do the per-directory exclude for all the
 581          * directories leading up to our base.
 582          */
 583         if (baselen) {
 584                 if (dir->exclude_per_dir) {
 585                         char *p, *pp = xmalloc(baselen+1);
 586                         memcpy(pp, base, baselen+1);
 587                         p = pp;
 588                         while (1) {
 589                                 char save = *p;
 590                                 *p = 0;
 591                                 push_exclude_per_directory(dir, pp, p-pp);
 592                                 *p++ = save;
 593                                 if (!save)
 594                                         break;
 595                                 p = strchr(p, '/');
 596                                 if (p)
 597                                         p++;
 598                                 else
 599                                         p = pp + baselen;
 600                         }
 601                         free(pp);
 602                 }
 603         }
 604
 605         read_directory_recursive(dir, path, base, baselen, 0, simplify);
 606         free_simplify(simplify);
 607         qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
 608         return dir->nr;
 609 }
 610
 611 int
 612 file_exists(const char *f)
 613 {
 614   struct stat sb;
 615   return stat(f, &sb) == 0;
 616 }