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
 126 /* git_mkstemp() - create tmp file honoring TMPDIR variable */
 127 int git_mkstemp(char *path, size_t len, const char *template)
 128 {
 129         const char *tmp;
 130         size_t n;
 131
 132         tmp = getenv("TMPDIR");
 133         if (!tmp)
 134                 tmp = "/tmp";
 135         n = snprintf(path, len, "%s/%s", tmp, template);
 136         if (len <= n) {
 137                 errno = ENAMETOOLONG;
 138                 return -1;
 139         }
 140         return mkstemp(path);
 141 }
 142
 143 /* git_mkstemps() - create tmp file with suffix honoring TMPDIR variable. */
 144 int git_mkstemps(char *path, size_t len, const char *template, int suffix_len)
 145 {
 146         const char *tmp;
 147         size_t n;
 148
 149         tmp = getenv("TMPDIR");
 150         if (!tmp)
 151                 tmp = "/tmp";
 152         n = snprintf(path, len, "%s/%s", tmp, template);
 153         if (len <= n) {
 154                 errno = ENAMETOOLONG;
 155                 return -1;
 156         }
 157         return mkstemps(path, suffix_len);
 158 }
 159
 160 /* Adapted from libiberty's mkstemp.c. */
 161
 162 #undef TMP_MAX
 163 #define TMP_MAX 16384
 164
 165 int git_mkstemps_mode(char *pattern, int suffix_len, int mode)
 166 {
 167         static const char letters[] =
 168                 "abcdefghijklmnopqrstuvwxyz"
 169                 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 170                 "0123456789";
 171         static const int num_letters = 62;
 172         uint64_t value;
 173         struct timeval tv;
 174         char *template;
 175         size_t len;
 176         int fd, count;
 177
 178         len = strlen(pattern);
 179
 180         if (len < 6 + suffix_len) {
 181                 errno = EINVAL;
 182                 return -1;
 183         }
 184
 185         if (strncmp(&pattern[len - 6 - suffix_len], "XXXXXX", 6)) {
 186                 errno = EINVAL;
 187                 return -1;
 188         }
 189
 190         /*
 191          * Replace pattern's XXXXXX characters with randomness.
 192          * Try TMP_MAX different filenames.
 193          */
 194         gettimeofday(&tv, NULL);
 195         value = ((size_t)(tv.tv_usec << 16)) ^ tv.tv_sec ^ getpid();
 196         template = &pattern[len - 6 - suffix_len];
 197         for (count = 0; count < TMP_MAX; ++count) {
 198                 uint64_t v = value;
 199                 /* Fill in the random bits. */
 200                 template[0] = letters[v % num_letters]; v /= num_letters;
 201                 template[1] = letters[v % num_letters]; v /= num_letters;
 202                 template[2] = letters[v % num_letters]; v /= num_letters;
 203                 template[3] = letters[v % num_letters]; v /= num_letters;
 204                 template[4] = letters[v % num_letters]; v /= num_letters;
 205                 template[5] = letters[v % num_letters]; v /= num_letters;
 206
 207                 fd = open(pattern, O_CREAT | O_EXCL | O_RDWR, mode);
 208                 if (fd > 0)
 209                         return fd;
 210                 /*
 211                  * Fatal error (EPERM, ENOSPC etc).
 212                  * It doesn't make sense to loop.
 213                  */
 214                 if (errno != EEXIST)
 215                         break;
 216                 /*
 217                  * This is a random value.  It is only necessary that
 218                  * the next TMP_MAX values generated by adding 7777 to
 219                  * VALUE are different with (module 2^32).
 220                  */
 221                 value += 7777;
 222         }
 223         /* We return the null string if we can't find a unique file name.  */
 224         pattern[0] = '\0';
 225         return -1;
 226 }
 227
 228 int git_mkstemp_mode(char *pattern, int mode)
 229 {
 230         /* mkstemp is just mkstemps with no suffix */
 231         return git_mkstemps_mode(pattern, 0, mode);
 232 }
 233
 234 int gitmkstemps(char *pattern, int suffix_len)
 235 {
 236         return git_mkstemps_mode(pattern, suffix_len, 0600);
 237 }
 238
 239 int validate_headref(const char *path)
 240 {
 241         struct stat st;
 242         char *buf, buffer[256];
 243         unsigned char sha1[20];
 244         int fd;
 245         ssize_t len;
 246
 247         if (lstat(path, &st) < 0)
 248                 return -1;
 249
 250         /* Make sure it is a "refs/.." symlink */
 251         if (S_ISLNK(st.st_mode)) {
 252                 len = readlink(path, buffer, sizeof(buffer)-1);
 253                 if (len >= 5 && !memcmp("refs/", buffer, 5))
 254                         return 0;
 255                 return -1;
 256         }
 257
 258         /*
 259          * Anything else, just open it and try to see if it is a symbolic ref.
 260          */
 261         fd = open(path, O_RDONLY);
 262         if (fd < 0)
 263                 return -1;
 264         len = read_in_full(fd, buffer, sizeof(buffer)-1);
 265         close(fd);
 266
 267         /*
 268          * Is it a symbolic ref?
 269          */
 270         if (len < 4)
 271                 return -1;
 272         if (!memcmp("ref:", buffer, 4)) {
 273                 buf = buffer + 4;
 274                 len -= 4;
 275                 while (len && isspace(*buf))
 276                         buf++, len--;
 277                 if (len >= 5 && !memcmp("refs/", buf, 5))
 278                         return 0;
 279         }
 280
 281         /*
 282          * Is this a detached HEAD?
 283          */
 284         if (!get_sha1_hex(buffer, sha1))
 285                 return 0;
 286
 287         return -1;
 288 }
 289
 290 static struct passwd *getpw_str(const char *username, size_t len)
 291 {
 292         struct passwd *pw;
 293         char *username_z = xmalloc(len + 1);
 294         memcpy(username_z, username, len);
 295         username_z[len] = '\0';
 296         pw = getpwnam(username_z);
 297         free(username_z);
 298         return pw;
 299 }
 300
 301 /*
 302  * Return a string with ~ and ~user expanded via getpw*.  If buf != NULL,
 303  * then it is a newly allocated string. Returns NULL on getpw failure or
 304  * if path is NULL.
 305  */
 306 char *expand_user_path(const char *path)
 307 {
 308         struct strbuf user_path = STRBUF_INIT;
 309         const char *first_slash = strchrnul(path, '/');
 310         const char *to_copy = path;
 311
 312         if (path == NULL)
 313                 goto return_null;
 314         if (path[0] == '~') {
 315                 const char *username = path + 1;
 316                 size_t username_len = first_slash - username;
 317                 if (username_len == 0) {
 318                         const char *home = getenv("HOME");
 319                         if (!home)
 320                                 goto return_null;
 321                         strbuf_add(&user_path, home, strlen(home));
 322                 } else {
 323                         struct passwd *pw = getpw_str(username, username_len);
 324                         if (!pw)
 325                                 goto return_null;
 326                         strbuf_add(&user_path, pw->pw_dir, strlen(pw->pw_dir));
 327                 }
 328                 to_copy = first_slash;
 329         }
 330         strbuf_add(&user_path, to_copy, strlen(to_copy));
 331         return strbuf_detach(&user_path, NULL);
 332 return_null:
 333         strbuf_release(&user_path);
 334         return NULL;
 335 }
 336
 337 /*
 338  * First, one directory to try is determined by the following algorithm.
 339  *
 340  * (0) If "strict" is given, the path is used as given and no DWIM is
 341  *     done. Otherwise:
 342  * (1) "~/path" to mean path under the running user's home directory;
 343  * (2) "~user/path" to mean path under named user's home directory;
 344  * (3) "relative/path" to mean cwd relative directory; or
 345  * (4) "/absolute/path" to mean absolute directory.
 346  *
 347  * Unless "strict" is given, we try access() for existence of "%s.git/.git",
 348  * "%s/.git", "%s.git", "%s" in this order.  The first one that exists is
 349  * what we try.
 350  *
 351  * Second, we try chdir() to that.  Upon failure, we return NULL.
 352  *
 353  * Then, we try if the current directory is a valid git repository.
 354  * Upon failure, we return NULL.
 355  *
 356  * If all goes well, we return the directory we used to chdir() (but
 357  * before ~user is expanded), avoiding getcwd() resolving symbolic
 358  * links.  User relative paths are also returned as they are given,
 359  * except DWIM suffixing.
 360  */
 361 char *enter_repo(char *path, int strict)
 362 {
 363         static char used_path[PATH_MAX];
 364         static char validated_path[PATH_MAX];
 365
 366         if (!path)
 367                 return NULL;
 368
 369         if (!strict) {
 370                 static const char *suffix[] = {
 371                         ".git/.git", "/.git", ".git", "", NULL,
 372                 };
 373                 int len = strlen(path);
 374                 int i;
 375                 while ((1 < len) && (path[len-1] == '/')) {
 376                         path[len-1] = 0;
 377                         len--;
 378                 }
 379                 if (PATH_MAX <= len)
 380                         return NULL;
 381                 if (path[0] == '~') {
 382                         char *newpath = expand_user_path(path);
 383                         if (!newpath || (PATH_MAX - 10 < strlen(newpath))) {
 384                                 free(newpath);
 385                                 return NULL;
 386                         }
 387                         /*
 388                          * Copy back into the static buffer. A pity
 389                          * since newpath was not bounded, but other
 390                          * branches of the if are limited by PATH_MAX
 391                          * anyway.
 392                          */
 393                         strcpy(used_path, newpath); free(newpath);
 394                         strcpy(validated_path, path);
 395                         path = used_path;
 396                 }
 397                 else if (PATH_MAX - 10 < len)
 398                         return NULL;
 399                 else {
 400                         path = strcpy(used_path, path);
 401                         strcpy(validated_path, path);
 402                 }
 403                 len = strlen(path);
 404                 for (i = 0; suffix[i]; i++) {
 405                         strcpy(path + len, suffix[i]);
 406                         if (!access(path, F_OK)) {
 407                                 strcat(validated_path, suffix[i]);
 408                                 break;
 409                         }
 410                 }
 411                 if (!suffix[i] || chdir(path))
 412                         return NULL;
 413                 path = validated_path;
 414         }
 415         else if (chdir(path))
 416                 return NULL;
 417
 418         if (access("objects", X_OK) == 0 && access("refs", X_OK) == 0 &&
 419             validate_headref("HEAD") == 0) {
 420                 set_git_dir(".");
 421                 check_repository_format();
 422                 return path;
 423         }
 424
 425         return NULL;
 426 }
 427
 428 int set_shared_perm(const char *path, int mode)
 429 {
 430         struct stat st;
 431         int tweak, shared, orig_mode;
 432
 433         if (!shared_repository) {
 434                 if (mode)
 435                         return chmod(path, mode & ~S_IFMT);
 436                 return 0;
 437         }
 438         if (!mode) {
 439                 if (lstat(path, &st) < 0)
 440                         return -1;
 441                 mode = st.st_mode;
 442                 orig_mode = mode;
 443         } else
 444                 orig_mode = 0;
 445         if (shared_repository < 0)
 446                 shared = -shared_repository;
 447         else
 448                 shared = shared_repository;
 449         tweak = shared;
 450
 451         if (!(mode & S_IWUSR))
 452                 tweak &= ~0222;
 453         if (mode & S_IXUSR)
 454                 /* Copy read bits to execute bits */
 455                 tweak |= (tweak & 0444) >> 2;
 456         if (shared_repository < 0)
 457                 mode = (mode & ~0777) | tweak;
 458         else
 459                 mode |= tweak;
 460
 461         if (S_ISDIR(mode)) {
 462                 /* Copy read bits to execute bits */
 463                 mode |= (shared & 0444) >> 2;
 464                 mode |= FORCE_DIR_SET_GID;
 465         }
 466
 467         if (((shared_repository < 0
 468               ? (orig_mode & (FORCE_DIR_SET_GID | 0777))
 469               : (orig_mode & mode)) != mode) &&
 470             chmod(path, (mode & ~S_IFMT)) < 0)
 471                 return -2;
 472         return 0;
 473 }
 474
 475 const char *make_relative_path(const char *abs, const char *base)
 476 {
 477         static char buf[PATH_MAX + 1];
 478         int i = 0, j = 0;
 479
 480         if (!base || !base[0])
 481                 return abs;
 482         while (base[i]) {
 483                 if (is_dir_sep(base[i])) {
 484                         if (!is_dir_sep(abs[j]))
 485                                 return abs;
 486                         while (is_dir_sep(base[i]))
 487                                 i++;
 488                         while (is_dir_sep(abs[j]))
 489                                 j++;
 490                         continue;
 491                 } else if (abs[j] != base[i]) {
 492                         return abs;
 493                 }
 494                 i++;
 495                 j++;
 496         }
 497         if (
 498             /* "/foo" is a prefix of "/foo" */
 499             abs[j] &&
 500             /* "/foo" is not a prefix of "/foobar" */
 501             !is_dir_sep(base[i-1]) && !is_dir_sep(abs[j])
 502            )
 503                 return abs;
 504         while (is_dir_sep(abs[j]))
 505                 j++;
 506         if (!abs[j])
 507                 strcpy(buf, ".");
 508         else
 509                 strcpy(buf, abs + j);
 510         return buf;
 511 }
 512
 513 /*
 514  * It is okay if dst == src, but they should not overlap otherwise.
 515  *
 516  * Performs the following normalizations on src, storing the result in dst:
 517  * - Ensures that components are separated by '/' (Windows only)
 518  * - Squashes sequences of '/'.
 519  * - Removes "." components.
 520  * - Removes ".." components, and the components the precede them.
 521  * Returns failure (non-zero) if a ".." component appears as first path
 522  * component anytime during the normalization. Otherwise, returns success (0).
 523  *
 524  * Note that this function is purely textual.  It does not follow symlinks,
 525  * verify the existence of the path, or make any system calls.
 526  */
 527 int normalize_path_copy(char *dst, const char *src)
 528 {
 529         char *dst0;
 530
 531         if (has_dos_drive_prefix(src)) {
 532                 *dst++ = *src++;
 533                 *dst++ = *src++;
 534         }
 535         dst0 = dst;
 536
 537         if (is_dir_sep(*src)) {
 538                 *dst++ = '/';
 539                 while (is_dir_sep(*src))
 540                         src++;
 541         }
 542
 543         for (;;) {
 544                 char c = *src;
 545
 546                 /*
 547                  * A path component that begins with . could be
 548                  * special:
 549                  * (1) "." and ends   -- ignore and terminate.
 550                  * (2) "./"           -- ignore them, eat slash and continue.
 551                  * (3) ".." and ends  -- strip one and terminate.
 552                  * (4) "../"          -- strip one, eat slash and continue.
 553                  */
 554                 if (c == '.') {
 555                         if (!src[1]) {
 556                                 /* (1) */
 557                                 src++;
 558                         } else if (is_dir_sep(src[1])) {
 559                                 /* (2) */
 560                                 src += 2;
 561                                 while (is_dir_sep(*src))
 562                                         src++;
 563                                 continue;
 564                         } else if (src[1] == '.') {
 565                                 if (!src[2]) {
 566                                         /* (3) */
 567                                         src += 2;
 568                                         goto up_one;
 569                                 } else if (is_dir_sep(src[2])) {
 570                                         /* (4) */
 571                                         src += 3;
 572                                         while (is_dir_sep(*src))
 573                                                 src++;
 574                                         goto up_one;
 575                                 }
 576                         }
 577                 }
 578
 579                 /* copy up to the next '/', and eat all '/' */
 580                 while ((c = *src++) != '\0' && !is_dir_sep(c))
 581                         *dst++ = c;
 582                 if (is_dir_sep(c)) {
 583                         *dst++ = '/';
 584                         while (is_dir_sep(c))
 585                                 c = *src++;
 586                         src--;
 587                 } else if (!c)
 588                         break;
 589                 continue;
 590
 591         up_one:
 592                 /*
 593                  * dst0..dst is prefix portion, and dst[-1] is '/';
 594                  * go up one level.
 595                  */
 596                 dst--;  /* go to trailing '/' */
 597                 if (dst <= dst0)
 598                         return -1;
 599                 /* Windows: dst[-1] cannot be backslash anymore */
 600                 while (dst0 < dst && dst[-1] != '/')
 601                         dst--;
 602         }
 603         *dst = '\0';
 604         return 0;
 605 }
 606
 607 /*
 608  * path = Canonical absolute path
 609  * prefix_list = Colon-separated list of absolute paths
 610  *
 611  * Determines, for each path in prefix_list, whether the "prefix" really
 612  * is an ancestor directory of path.  Returns the length of the longest
 613  * ancestor directory, excluding any trailing slashes, or -1 if no prefix
 614  * is an ancestor.  (Note that this means 0 is returned if prefix_list is
 615  * "/".) "/foo" is not considered an ancestor of "/foobar".  Directories
 616  * are not considered to be their own ancestors.  path must be in a
 617  * canonical form: empty components, or "." or ".." components are not
 618  * allowed.  prefix_list may be null, which is like "".
 619  */
 620 int longest_ancestor_length(const char *path, const char *prefix_list)
 621 {
 622         char buf[PATH_MAX+1];
 623         const char *ceil, *colon;
 624         int len, max_len = -1;
 625
 626         if (prefix_list == NULL || !strcmp(path, "/"))
 627                 return -1;
 628
 629         for (colon = ceil = prefix_list; *colon; ceil = colon+1) {
 630                 for (colon = ceil; *colon && *colon != PATH_SEP; colon++);
 631                 len = colon - ceil;
 632                 if (len == 0 || len > PATH_MAX || !is_absolute_path(ceil))
 633                         continue;
 634                 strlcpy(buf, ceil, len+1);
 635                 if (normalize_path_copy(buf, buf) < 0)
 636                         continue;
 637                 len = strlen(buf);
 638                 if (len > 0 && buf[len-1] == '/')
 639                         buf[--len] = '\0';
 640
 641                 if (!strncmp(path, buf, len) &&
 642                     path[len] == '/' &&
 643                     len > max_len) {
 644                         max_len = len;
 645                 }
 646         }
 647
 648         return max_len;
 649 }
 650
 651 /* strip arbitrary amount of directory separators at end of path */
 652 static inline int chomp_trailing_dir_sep(const char *path, int len)
 653 {
 654         while (len && is_dir_sep(path[len - 1]))
 655                 len--;
 656         return len;
 657 }
 658
 659 /*
 660  * If path ends with suffix (complete path components), returns the
 661  * part before suffix (sans trailing directory separators).
 662  * Otherwise returns NULL.
 663  */
 664 char *strip_path_suffix(const char *path, const char *suffix)
 665 {
 666         int path_len = strlen(path), suffix_len = strlen(suffix);
 667
 668         while (suffix_len) {
 669                 if (!path_len)
 670                         return NULL;
 671
 672                 if (is_dir_sep(path[path_len - 1])) {
 673                         if (!is_dir_sep(suffix[suffix_len - 1]))
 674                                 return NULL;
 675                         path_len = chomp_trailing_dir_sep(path, path_len);
 676                         suffix_len = chomp_trailing_dir_sep(suffix, suffix_len);
 677                 }
 678                 else if (path[--path_len] != suffix[--suffix_len])
 679                         return NULL;
 680         }
 681
 682         if (path_len && !is_dir_sep(path[path_len - 1]))
 683                 return NULL;
 684         return xstrndup(path, chomp_trailing_dir_sep(path, path_len));
 685 }
 686
 687 int daemon_avoid_alias(const char *p)
 688 {
 689         int sl, ndot;
 690
 691         /*
 692          * This resurrects the belts and suspenders paranoia check by HPA
 693          * done in <435560F7.4080006@zytor.com> thread, now enter_repo()
 694          * does not do getcwd() based path canonicalization.
 695          *
 696          * sl becomes true immediately after seeing '/' and continues to
 697          * be true as long as dots continue after that without intervening
 698          * non-dot character.
 699          */
 700         if (!p || (*p != '/' && *p != '~'))
 701                 return -1;
 702         sl = 1; ndot = 0;
 703         p++;
 704
 705         while (1) {
 706                 char ch = *p++;
 707                 if (sl) {
 708                         if (ch == '.')
 709                                 ndot++;
 710                         else if (ch == '/') {
 711                                 if (ndot < 3)
 712                                         /* reject //, /./ and /../ */
 713                                         return -1;
 714                                 ndot = 0;
 715                         }
 716                         else if (ch == 0) {
 717                                 if (0 < ndot && ndot < 3)
 718                                         /* reject /.$ and /..$ */
 719                                         return -1;
 720                                 return 0;
 721                         }
 722                         else
 723                                 sl = ndot = 0;
 724                 }
 725                 else if (ch == 0)
 726                         return 0;
 727                 else if (ch == '/') {
 728                         sl = 1;
 729                         ndot = 0;
 730                 }
 731         }
 732 }
 733
 734 int offset_1st_component(const char *path)
 735 {
 736         if (has_dos_drive_prefix(path))
 737                 return 2 + is_dir_sep(path[2]);
 738         return is_dir_sep(path[0]);
 739 }