path.c

   1 /*
   2  * I'm tired of doing "vsnprintf()" etc just to open a
   3  * file, so here's a "return static buffer with printf"
   4  * interface for paths.
   5  *
   6  * It's obviously not thread-safe. Sue me. But it's quite
   7  * useful for doing things like
   8  *
   9  *   f = open(mkpath("%s/%s.git", base, name), O_RDONLY);
  10  *
  11  * which is what it's designed for.
  12  */
  13 #include "cache.h"
  14 #include "strbuf.h"
  15
  16 static char bad_path[] = "/bad-path/";
  17
  18 static char *get_pathname(void)
  19 {
  20         static char pathname_array[4][PATH_MAX];
  21         static int index;
  22         return pathname_array[3 & ++index];
  23 }
  24
  25 static char *cleanup_path(char *path)
  26 {
  27         /* Clean it up */
  28         if (!memcmp(path, "./", 2)) {
  29                 path += 2;
  30                 while (*path == '/')
  31                         path++;
  32         }
  33         return path;
  34 }
  35
  36 char *mksnpath(char *buf, size_t n, const char *fmt, ...)
  37 {
  38         va_list args;
  39         unsigned len;
  40
  41         va_start(args, fmt);
  42         len = vsnprintf(buf, n, fmt, args);
  43         va_end(args);
  44         if (len >= n) {
  45                 strlcpy(buf, bad_path, n);
  46                 return buf;
  47         }
  48         return cleanup_path(buf);
  49 }
  50
  51 static char *git_vsnpath(char *buf, size_t n, const char *fmt, va_list args)
  52 {
  53         const char *git_dir = get_git_dir();
  54         size_t len;
  55
  56         len = strlen(git_dir);
  57         if (n < len + 1)
  58                 goto bad;
  59         memcpy(buf, git_dir, len);
  60         if (len && !is_dir_sep(git_dir[len-1]))
  61                 buf[len++] = '/';
  62         len += vsnprintf(buf + len, n - len, fmt, args);
  63         if (len >= n)
  64                 goto bad;
  65         return cleanup_path(buf);
  66 bad:
  67         strlcpy(buf, bad_path, n);
  68         return buf;
  69 }
  70
  71 char *git_snpath(char *buf, size_t n, const char *fmt, ...)
  72 {
  73         va_list args;
  74         va_start(args, fmt);
  75         (void)git_vsnpath(buf, n, fmt, args);
  76         va_end(args);
  77         return buf;
  78 }
  79
  80 char *git_pathdup(const char *fmt, ...)
  81 {
  82         char path[PATH_MAX];
  83         va_list args;
  84         va_start(args, fmt);
  85         (void)git_vsnpath(path, sizeof(path), fmt, args);
  86         va_end(args);
  87         return xstrdup(path);
  88 }
  89
  90 char *mkpath(const char *fmt, ...)
  91 {
  92         va_list args;
  93         unsigned len;
  94         char *pathname = get_pathname();
  95
  96         va_start(args, fmt);
  97         len = vsnprintf(pathname, PATH_MAX, fmt, args);
  98         va_end(args);
  99         if (len >= PATH_MAX)
 100                 return bad_path;
 101         return cleanup_path(pathname);
 102 }
 103
 104 char *git_path(const char *fmt, ...)
 105 {
 106         const char *git_dir = get_git_dir();
 107         char *pathname = get_pathname();
 108         va_list args;
 109         unsigned len;
 110
 111         len = strlen(git_dir);
 112         if (len > PATH_MAX-100)
 113                 return bad_path;
 114         memcpy(pathname, git_dir, len);
 115         if (len && git_dir[len-1] != '/')
 116                 pathname[len++] = '/';
 117         va_start(args, fmt);
 118         len += vsnprintf(pathname + len, PATH_MAX - len, fmt, args);
 119         va_end(args);
 120         if (len >= PATH_MAX)
 121                 return bad_path;
 122         return cleanup_path(pathname);
 123 }
 124
 125 char *git_path_submodule(const char *path, const char *fmt, ...)
 126 {
 127         char *pathname = get_pathname();
 128         struct strbuf buf = STRBUF_INIT;
 129         const char *git_dir;
 130         va_list args;
 131         unsigned len;
 132
 133         len = strlen(path);
 134         if (len > PATH_MAX-100)
 135                 return bad_path;
 136
 137         strbuf_addstr(&buf, path);
 138         if (len && path[len-1] != '/')
 139                 strbuf_addch(&buf, '/');
 140         strbuf_addstr(&buf, ".git");
 141
 142         git_dir = read_gitfile_gently(buf.buf);
 143         if (git_dir) {
 144                 strbuf_reset(&buf);
 145                 strbuf_addstr(&buf, git_dir);
 146         }
 147         strbuf_addch(&buf, '/');
 148
 149         if (buf.len >= PATH_MAX)
 150                 return bad_path;
 151         memcpy(pathname, buf.buf, buf.len + 1);
 152
 153         strbuf_release(&buf);
 154         len = strlen(pathname);
 155
 156         va_start(args, fmt);
 157         len += vsnprintf(pathname + len, PATH_MAX - len, fmt, args);
 158         va_end(args);
 159         if (len >= PATH_MAX)
 160                 return bad_path;
 161         return cleanup_path(pathname);
 162 }
 163
 164 int validate_headref(const char *path)
 165 {
 166         struct stat st;
 167         char *buf, buffer[256];
 168         unsigned char sha1[20];
 169         int fd;
 170         ssize_t len;
 171
 172         if (lstat(path, &st) < 0)
 173                 return -1;
 174
 175         /* Make sure it is a "refs/.." symlink */
 176         if (S_ISLNK(st.st_mode)) {
 177                 len = readlink(path, buffer, sizeof(buffer)-1);
 178                 if (len >= 5 && !memcmp("refs/", buffer, 5))
 179                         return 0;
 180                 return -1;
 181         }
 182
 183         /*
 184          * Anything else, just open it and try to see if it is a symbolic ref.
 185          */
 186         fd = open(path, O_RDONLY);
 187         if (fd < 0)
 188                 return -1;
 189         len = read_in_full(fd, buffer, sizeof(buffer)-1);
 190         close(fd);
 191
 192         /*
 193          * Is it a symbolic ref?
 194          */
 195         if (len < 4)
 196                 return -1;
 197         if (!memcmp("ref:", buffer, 4)) {
 198                 buf = buffer + 4;
 199                 len -= 4;
 200                 while (len && isspace(*buf))
 201                         buf++, len--;
 202                 if (len >= 5 && !memcmp("refs/", buf, 5))
 203                         return 0;
 204         }
 205
 206         /*
 207          * Is this a detached HEAD?
 208          */
 209         if (!get_sha1_hex(buffer, sha1))
 210                 return 0;
 211
 212         return -1;
 213 }
 214
 215 static struct passwd *getpw_str(const char *username, size_t len)
 216 {
 217         struct passwd *pw;
 218         char *username_z = xmalloc(len + 1);
 219         memcpy(username_z, username, len);
 220         username_z[len] = '\0';
 221         pw = getpwnam(username_z);
 222         free(username_z);
 223         return pw;
 224 }
 225
 226 /*
 227  * Return a string with ~ and ~user expanded via getpw*.  If buf != NULL,
 228  * then it is a newly allocated string. Returns NULL on getpw failure or
 229  * if path is NULL.
 230  */
 231 char *expand_user_path(const char *path)
 232 {
 233         struct strbuf user_path = STRBUF_INIT;
 234         const char *first_slash = strchrnul(path, '/');
 235         const char *to_copy = path;
 236
 237         if (path == NULL)
 238                 goto return_null;
 239         if (path[0] == '~') {
 240                 const char *username = path + 1;
 241                 size_t username_len = first_slash - username;
 242                 if (username_len == 0) {
 243                         const char *home = getenv("HOME");
 244                         if (!home)
 245                                 goto return_null;
 246                         strbuf_add(&user_path, home, strlen(home));
 247                 } else {
 248                         struct passwd *pw = getpw_str(username, username_len);
 249                         if (!pw)
 250                                 goto return_null;
 251                         strbuf_add(&user_path, pw->pw_dir, strlen(pw->pw_dir));
 252                 }
 253                 to_copy = first_slash;
 254         }
 255         strbuf_add(&user_path, to_copy, strlen(to_copy));
 256         return strbuf_detach(&user_path, NULL);
 257 return_null:
 258         strbuf_release(&user_path);
 259         return NULL;
 260 }
 261
 262 /*
 263  * First, one directory to try is determined by the following algorithm.
 264  *
 265  * (0) If "strict" is given, the path is used as given and no DWIM is
 266  *     done. Otherwise:
 267  * (1) "~/path" to mean path under the running user's home directory;
 268  * (2) "~user/path" to mean path under named user's home directory;
 269  * (3) "relative/path" to mean cwd relative directory; or
 270  * (4) "/absolute/path" to mean absolute directory.
 271  *
 272  * Unless "strict" is given, we try access() for existence of "%s.git/.git",
 273  * "%s/.git", "%s.git", "%s" in this order.  The first one that exists is
 274  * what we try.
 275  *
 276  * Second, we try chdir() to that.  Upon failure, we return NULL.
 277  *
 278  * Then, we try if the current directory is a valid git repository.
 279  * Upon failure, we return NULL.
 280  *
 281  * If all goes well, we return the directory we used to chdir() (but
 282  * before ~user is expanded), avoiding getcwd() resolving symbolic
 283  * links.  User relative paths are also returned as they are given,
 284  * except DWIM suffixing.
 285  */
 286 char *enter_repo(char *path, int strict)
 287 {
 288         static char used_path[PATH_MAX];
 289         static char validated_path[PATH_MAX];
 290
 291         if (!path)
 292                 return NULL;
 293
 294         if (!strict) {
 295                 static const char *suffix[] = {
 296                         ".git/.git", "/.git", ".git", "", NULL,
 297                 };
 298                 int len = strlen(path);
 299                 int i;
 300                 while ((1 < len) && (path[len-1] == '/')) {
 301                         path[len-1] = 0;
 302                         len--;
 303                 }
 304                 if (PATH_MAX <= len)
 305                         return NULL;
 306                 if (path[0] == '~') {
 307                         char *newpath = expand_user_path(path);
 308                         if (!newpath || (PATH_MAX - 10 < strlen(newpath))) {
 309                                 free(newpath);
 310                                 return NULL;
 311                         }
 312                         /*
 313                          * Copy back into the static buffer. A pity
 314                          * since newpath was not bounded, but other
 315                          * branches of the if are limited by PATH_MAX
 316                          * anyway.
 317                          */
 318                         strcpy(used_path, newpath); free(newpath);
 319                         strcpy(validated_path, path);
 320                         path = used_path;
 321                 }
 322                 else if (PATH_MAX - 10 < len)
 323                         return NULL;
 324                 else {
 325                         path = strcpy(used_path, path);
 326                         strcpy(validated_path, path);
 327                 }
 328                 len = strlen(path);
 329                 for (i = 0; suffix[i]; i++) {
 330                         strcpy(path + len, suffix[i]);
 331                         if (!access(path, F_OK)) {
 332                                 strcat(validated_path, suffix[i]);
 333                                 break;
 334                         }
 335                 }
 336                 if (!suffix[i] || chdir(path))
 337                         return NULL;
 338                 path = validated_path;
 339         }
 340         else if (chdir(path))
 341                 return NULL;
 342
 343         if (access("objects", X_OK) == 0 && access("refs", X_OK) == 0 &&
 344             validate_headref("HEAD") == 0) {
 345                 set_git_dir(".");
 346                 check_repository_format();
 347                 return path;
 348         }
 349
 350         return NULL;
 351 }
 352
 353 int set_shared_perm(const char *path, int mode)
 354 {
 355         struct stat st;
 356         int tweak, shared, orig_mode;
 357
 358         if (!shared_repository) {
 359                 if (mode)
 360                         return chmod(path, mode & ~S_IFMT);
 361                 return 0;
 362         }
 363         if (!mode) {
 364                 if (lstat(path, &st) < 0)
 365                         return -1;
 366                 mode = st.st_mode;
 367                 orig_mode = mode;
 368         } else
 369                 orig_mode = 0;
 370         if (shared_repository < 0)
 371                 shared = -shared_repository;
 372         else
 373                 shared = shared_repository;
 374         tweak = shared;
 375
 376         if (!(mode & S_IWUSR))
 377                 tweak &= ~0222;
 378         if (mode & S_IXUSR)
 379                 /* Copy read bits to execute bits */
 380                 tweak |= (tweak & 0444) >> 2;
 381         if (shared_repository < 0)
 382                 mode = (mode & ~0777) | tweak;
 383         else
 384                 mode |= tweak;
 385
 386         if (S_ISDIR(mode)) {
 387                 /* Copy read bits to execute bits */
 388                 mode |= (shared & 0444) >> 2;
 389                 mode |= FORCE_DIR_SET_GID;
 390         }
 391
 392         if (((shared_repository < 0
 393               ? (orig_mode & (FORCE_DIR_SET_GID | 0777))
 394               : (orig_mode & mode)) != mode) &&
 395             chmod(path, (mode & ~S_IFMT)) < 0)
 396                 return -2;
 397         return 0;
 398 }
 399
 400 const char *relative_path(const char *abs, const char *base)
 401 {
 402         static char buf[PATH_MAX + 1];
 403         int i = 0, j = 0;
 404
 405         if (!base || !base[0])
 406                 return abs;
 407         while (base[i]) {
 408                 if (is_dir_sep(base[i])) {
 409                         if (!is_dir_sep(abs[j]))
 410                                 return abs;
 411                         while (is_dir_sep(base[i]))
 412                                 i++;
 413                         while (is_dir_sep(abs[j]))
 414                                 j++;
 415                         continue;
 416                 } else if (abs[j] != base[i]) {
 417                         return abs;
 418                 }
 419                 i++;
 420                 j++;
 421         }
 422         if (
 423             /* "/foo" is a prefix of "/foo" */
 424             abs[j] &&
 425             /* "/foo" is not a prefix of "/foobar" */
 426             !is_dir_sep(base[i-1]) && !is_dir_sep(abs[j])
 427            )
 428                 return abs;
 429         while (is_dir_sep(abs[j]))
 430                 j++;
 431         if (!abs[j])
 432                 strcpy(buf, ".");
 433         else
 434                 strcpy(buf, abs + j);
 435         return buf;
 436 }
 437
 438 /*
 439  * It is okay if dst == src, but they should not overlap otherwise.
 440  *
 441  * Performs the following normalizations on src, storing the result in dst:
 442  * - Ensures that components are separated by '/' (Windows only)
 443  * - Squashes sequences of '/'.
 444  * - Removes "." components.
 445  * - Removes ".." components, and the components the precede them.
 446  * Returns failure (non-zero) if a ".." component appears as first path
 447  * component anytime during the normalization. Otherwise, returns success (0).
 448  *
 449  * Note that this function is purely textual.  It does not follow symlinks,
 450  * verify the existence of the path, or make any system calls.
 451  */
 452 int normalize_path_copy(char *dst, const char *src)
 453 {
 454         char *dst0;
 455
 456         if (has_dos_drive_prefix(src)) {
 457                 *dst++ = *src++;
 458                 *dst++ = *src++;
 459         }
 460         dst0 = dst;
 461
 462         if (is_dir_sep(*src)) {
 463                 *dst++ = '/';
 464                 while (is_dir_sep(*src))
 465                         src++;
 466         }
 467
 468         for (;;) {
 469                 char c = *src;
 470
 471                 /*
 472                  * A path component that begins with . could be
 473                  * special:
 474                  * (1) "." and ends   -- ignore and terminate.
 475                  * (2) "./"           -- ignore them, eat slash and continue.
 476                  * (3) ".." and ends  -- strip one and terminate.
 477                  * (4) "../"          -- strip one, eat slash and continue.
 478                  */
 479                 if (c == '.') {
 480                         if (!src[1]) {
 481                                 /* (1) */
 482                                 src++;
 483                         } else if (is_dir_sep(src[1])) {
 484                                 /* (2) */
 485                                 src += 2;
 486                                 while (is_dir_sep(*src))
 487                                         src++;
 488                                 continue;
 489                         } else if (src[1] == '.') {
 490                                 if (!src[2]) {
 491                                         /* (3) */
 492                                         src += 2;
 493                                         goto up_one;
 494                                 } else if (is_dir_sep(src[2])) {
 495                                         /* (4) */
 496                                         src += 3;
 497                                         while (is_dir_sep(*src))
 498                                                 src++;
 499                                         goto up_one;
 500                                 }
 501                         }
 502                 }
 503
 504                 /* copy up to the next '/', and eat all '/' */
 505                 while ((c = *src++) != '\0' && !is_dir_sep(c))
 506                         *dst++ = c;
 507                 if (is_dir_sep(c)) {
 508                         *dst++ = '/';
 509                         while (is_dir_sep(c))
 510                                 c = *src++;
 511                         src--;
 512                 } else if (!c)
 513                         break;
 514                 continue;
 515
 516         up_one:
 517                 /*
 518                  * dst0..dst is prefix portion, and dst[-1] is '/';
 519                  * go up one level.
 520                  */
 521                 dst--;  /* go to trailing '/' */
 522                 if (dst <= dst0)
 523                         return -1;
 524                 /* Windows: dst[-1] cannot be backslash anymore */
 525                 while (dst0 < dst && dst[-1] != '/')
 526                         dst--;
 527         }
 528         *dst = '\0';
 529         return 0;
 530 }
 531
 532 /*
 533  * path = Canonical absolute path
 534  * prefix_list = Colon-separated list of absolute paths
 535  *
 536  * Determines, for each path in prefix_list, whether the "prefix" really
 537  * is an ancestor directory of path.  Returns the length of the longest
 538  * ancestor directory, excluding any trailing slashes, or -1 if no prefix
 539  * is an ancestor.  (Note that this means 0 is returned if prefix_list is
 540  * "/".) "/foo" is not considered an ancestor of "/foobar".  Directories
 541  * are not considered to be their own ancestors.  path must be in a
 542  * canonical form: empty components, or "." or ".." components are not
 543  * allowed.  prefix_list may be null, which is like "".
 544  */
 545 int longest_ancestor_length(const char *path, const char *prefix_list)
 546 {
 547         char buf[PATH_MAX+1];
 548         const char *ceil, *colon;
 549         int len, max_len = -1;
 550
 551         if (prefix_list == NULL || !strcmp(path, "/"))
 552                 return -1;
 553
 554         for (colon = ceil = prefix_list; *colon; ceil = colon+1) {
 555                 for (colon = ceil; *colon && *colon != PATH_SEP; colon++);
 556                 len = colon - ceil;
 557                 if (len == 0 || len > PATH_MAX || !is_absolute_path(ceil))
 558                         continue;
 559                 strlcpy(buf, ceil, len+1);
 560                 if (normalize_path_copy(buf, buf) < 0)
 561                         continue;
 562                 len = strlen(buf);
 563                 if (len > 0 && buf[len-1] == '/')
 564                         buf[--len] = '\0';
 565
 566                 if (!strncmp(path, buf, len) &&
 567                     path[len] == '/' &&
 568                     len > max_len) {
 569                         max_len = len;
 570                 }
 571         }
 572
 573         return max_len;
 574 }
 575
 576 /* strip arbitrary amount of directory separators at end of path */
 577 static inline int chomp_trailing_dir_sep(const char *path, int len)
 578 {
 579         while (len && is_dir_sep(path[len - 1]))
 580                 len--;
 581         return len;
 582 }
 583
 584 /*
 585  * If path ends with suffix (complete path components), returns the
 586  * part before suffix (sans trailing directory separators).
 587  * Otherwise returns NULL.
 588  */
 589 char *strip_path_suffix(const char *path, const char *suffix)
 590 {
 591         int path_len = strlen(path), suffix_len = strlen(suffix);
 592
 593         while (suffix_len) {
 594                 if (!path_len)
 595                         return NULL;
 596
 597                 if (is_dir_sep(path[path_len - 1])) {
 598                         if (!is_dir_sep(suffix[suffix_len - 1]))
 599                                 return NULL;
 600                         path_len = chomp_trailing_dir_sep(path, path_len);
 601                         suffix_len = chomp_trailing_dir_sep(suffix, suffix_len);
 602                 }
 603                 else if (path[--path_len] != suffix[--suffix_len])
 604                         return NULL;
 605         }
 606
 607         if (path_len && !is_dir_sep(path[path_len - 1]))
 608                 return NULL;
 609         return xstrndup(path, chomp_trailing_dir_sep(path, path_len));
 610 }
 611
 612 int daemon_avoid_alias(const char *p)
 613 {
 614         int sl, ndot;
 615
 616         /*
 617          * This resurrects the belts and suspenders paranoia check by HPA
 618          * done in <435560F7.4080006@zytor.com> thread, now enter_repo()
 619          * does not do getcwd() based path canonicalization.
 620          *
 621          * sl becomes true immediately after seeing '/' and continues to
 622          * be true as long as dots continue after that without intervening
 623          * non-dot character.
 624          */
 625         if (!p || (*p != '/' && *p != '~'))
 626                 return -1;
 627         sl = 1; ndot = 0;
 628         p++;
 629
 630         while (1) {
 631                 char ch = *p++;
 632                 if (sl) {
 633                         if (ch == '.')
 634                                 ndot++;
 635                         else if (ch == '/') {
 636                                 if (ndot < 3)
 637                                         /* reject //, /./ and /../ */
 638                                         return -1;
 639                                 ndot = 0;
 640                         }
 641                         else if (ch == 0) {
 642                                 if (0 < ndot && ndot < 3)
 643                                         /* reject /.$ and /..$ */
 644                                         return -1;
 645                                 return 0;
 646                         }
 647                         else
 648                                 sl = ndot = 0;
 649                 }
 650                 else if (ch == 0)
 651                         return 0;
 652                 else if (ch == '/') {
 653                         sl = 1;
 654                         ndot = 0;
 655                 }
 656         }
 657 }
 658
 659 int offset_1st_component(const char *path)
 660 {
 661         if (has_dos_drive_prefix(path))
 662                 return 2 + is_dir_sep(path[2]);
 663         return is_dir_sep(path[0]);
 664 }