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