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                         strbuf_add(&user_path, home, strlen(home));
 320                 } else {
 321                         struct passwd *pw = getpw_str(username, username_len);
 322                         if (!pw)
 323                                 goto return_null;
 324                         strbuf_add(&user_path, pw->pw_dir, strlen(pw->pw_dir));
 325                 }
 326                 to_copy = first_slash;
 327         }
 328         strbuf_add(&user_path, to_copy, strlen(to_copy));
 329         return strbuf_detach(&user_path, NULL);
 330 return_null:
 331         strbuf_release(&user_path);
 332         return NULL;
 333 }
 334
 335 /*
 336  * First, one directory to try is determined by the following algorithm.
 337  *
 338  * (0) If "strict" is given, the path is used as given and no DWIM is
 339  *     done. Otherwise:
 340  * (1) "~/path" to mean path under the running user's home directory;
 341  * (2) "~user/path" to mean path under named user's home directory;
 342  * (3) "relative/path" to mean cwd relative directory; or
 343  * (4) "/absolute/path" to mean absolute directory.
 344  *
 345  * Unless "strict" is given, we try access() for existence of "%s.git/.git",
 346  * "%s/.git", "%s.git", "%s" in this order.  The first one that exists is
 347  * what we try.
 348  *
 349  * Second, we try chdir() to that.  Upon failure, we return NULL.
 350  *
 351  * Then, we try if the current directory is a valid git repository.
 352  * Upon failure, we return NULL.
 353  *
 354  * If all goes well, we return the directory we used to chdir() (but
 355  * before ~user is expanded), avoiding getcwd() resolving symbolic
 356  * links.  User relative paths are also returned as they are given,
 357  * except DWIM suffixing.
 358  */
 359 char *enter_repo(char *path, int strict)
 360 {
 361         static char used_path[PATH_MAX];
 362         static char validated_path[PATH_MAX];
 363
 364         if (!path)
 365                 return NULL;
 366
 367         if (!strict) {
 368                 static const char *suffix[] = {
 369                         ".git/.git", "/.git", ".git", "", NULL,
 370                 };
 371                 int len = strlen(path);
 372                 int i;
 373                 while ((1 < len) && (path[len-1] == '/')) {
 374                         path[len-1] = 0;
 375                         len--;
 376                 }
 377                 if (PATH_MAX <= len)
 378                         return NULL;
 379                 if (path[0] == '~') {
 380                         char *newpath = expand_user_path(path);
 381                         if (!newpath || (PATH_MAX - 10 < strlen(newpath))) {
 382                                 free(newpath);
 383                                 return NULL;
 384                         }
 385                         /*
 386                          * Copy back into the static buffer. A pity
 387                          * since newpath was not bounded, but other
 388                          * branches of the if are limited by PATH_MAX
 389                          * anyway.
 390                          */
 391                         strcpy(used_path, newpath); free(newpath);
 392                         strcpy(validated_path, path);
 393                         path = used_path;
 394                 }
 395                 else if (PATH_MAX - 10 < len)
 396                         return NULL;
 397                 else {
 398                         path = strcpy(used_path, path);
 399                         strcpy(validated_path, path);
 400                 }
 401                 len = strlen(path);
 402                 for (i = 0; suffix[i]; i++) {
 403                         strcpy(path + len, suffix[i]);
 404                         if (!access(path, F_OK)) {
 405                                 strcat(validated_path, suffix[i]);
 406                                 break;
 407                         }
 408                 }
 409                 if (!suffix[i] || chdir(path))
 410                         return NULL;
 411                 path = validated_path;
 412         }
 413         else if (chdir(path))
 414                 return NULL;
 415
 416         if (access("objects", X_OK) == 0 && access("refs", X_OK) == 0 &&
 417             validate_headref("HEAD") == 0) {
 418                 set_git_dir(".");
 419                 check_repository_format();
 420                 return path;
 421         }
 422
 423         return NULL;
 424 }
 425
 426 int set_shared_perm(const char *path, int mode)
 427 {
 428         struct stat st;
 429         int tweak, shared, orig_mode;
 430
 431         if (!shared_repository) {
 432                 if (mode)
 433                         return chmod(path, mode & ~S_IFMT);
 434                 return 0;
 435         }
 436         if (!mode) {
 437                 if (lstat(path, &st) < 0)
 438                         return -1;
 439                 mode = st.st_mode;
 440                 orig_mode = mode;
 441         } else
 442                 orig_mode = 0;
 443         if (shared_repository < 0)
 444                 shared = -shared_repository;
 445         else
 446                 shared = shared_repository;
 447         tweak = shared;
 448
 449         if (!(mode & S_IWUSR))
 450                 tweak &= ~0222;
 451         if (mode & S_IXUSR)
 452                 /* Copy read bits to execute bits */
 453                 tweak |= (tweak & 0444) >> 2;
 454         if (shared_repository < 0)
 455                 mode = (mode & ~0777) | tweak;
 456         else
 457                 mode |= tweak;
 458
 459         if (S_ISDIR(mode)) {
 460                 /* Copy read bits to execute bits */
 461                 mode |= (shared & 0444) >> 2;
 462                 mode |= FORCE_DIR_SET_GID;
 463         }
 464
 465         if (((shared_repository < 0
 466               ? (orig_mode & (FORCE_DIR_SET_GID | 0777))
 467               : (orig_mode & mode)) != mode) &&
 468             chmod(path, (mode & ~S_IFMT)) < 0)
 469                 return -2;
 470         return 0;
 471 }
 472
 473 const char *make_relative_path(const char *abs, const char *base)
 474 {
 475         static char buf[PATH_MAX + 1];
 476         int i = 0, j = 0;
 477
 478         if (!base || !base[0])
 479                 return abs;
 480         while (base[i]) {
 481                 if (is_dir_sep(base[i])) {
 482                         if (!is_dir_sep(abs[j]))
 483                                 return abs;
 484                         while (is_dir_sep(base[i]))
 485                                 i++;
 486                         while (is_dir_sep(abs[j]))
 487                                 j++;
 488                         continue;
 489                 } else if (abs[j] != base[i]) {
 490                         return abs;
 491                 }
 492                 i++;
 493                 j++;
 494         }
 495         if (
 496             /* "/foo" is a prefix of "/foo" */
 497             abs[j] &&
 498             /* "/foo" is not a prefix of "/foobar" */
 499             !is_dir_sep(base[i-1]) && !is_dir_sep(abs[j])
 500            )
 501                 return abs;
 502         while (is_dir_sep(abs[j]))
 503                 j++;
 504         if (!abs[j])
 505                 strcpy(buf, ".");
 506         else
 507                 strcpy(buf, abs + j);
 508         return buf;
 509 }
 510
 511 /*
 512  * It is okay if dst == src, but they should not overlap otherwise.
 513  *
 514  * Performs the following normalizations on src, storing the result in dst:
 515  * - Ensures that components are separated by '/' (Windows only)
 516  * - Squashes sequences of '/'.
 517  * - Removes "." components.
 518  * - Removes ".." components, and the components the precede them.
 519  * Returns failure (non-zero) if a ".." component appears as first path
 520  * component anytime during the normalization. Otherwise, returns success (0).
 521  *
 522  * Note that this function is purely textual.  It does not follow symlinks,
 523  * verify the existence of the path, or make any system calls.
 524  */
 525 int normalize_path_copy(char *dst, const char *src)
 526 {
 527         char *dst0;
 528
 529         if (has_dos_drive_prefix(src)) {
 530                 *dst++ = *src++;
 531                 *dst++ = *src++;
 532         }
 533         dst0 = dst;
 534
 535         if (is_dir_sep(*src)) {
 536                 *dst++ = '/';
 537                 while (is_dir_sep(*src))
 538                         src++;
 539         }
 540
 541         for (;;) {
 542                 char c = *src;
 543
 544                 /*
 545                  * A path component that begins with . could be
 546                  * special:
 547                  * (1) "." and ends   -- ignore and terminate.
 548                  * (2) "./"           -- ignore them, eat slash and continue.
 549                  * (3) ".." and ends  -- strip one and terminate.
 550                  * (4) "../"          -- strip one, eat slash and continue.
 551                  */
 552                 if (c == '.') {
 553                         if (!src[1]) {
 554                                 /* (1) */
 555                                 src++;
 556                         } else if (is_dir_sep(src[1])) {
 557                                 /* (2) */
 558                                 src += 2;
 559                                 while (is_dir_sep(*src))
 560                                         src++;
 561                                 continue;
 562                         } else if (src[1] == '.') {
 563                                 if (!src[2]) {
 564                                         /* (3) */
 565                                         src += 2;
 566                                         goto up_one;
 567                                 } else if (is_dir_sep(src[2])) {
 568                                         /* (4) */
 569                                         src += 3;
 570                                         while (is_dir_sep(*src))
 571                                                 src++;
 572                                         goto up_one;
 573                                 }
 574                         }
 575                 }
 576
 577                 /* copy up to the next '/', and eat all '/' */
 578                 while ((c = *src++) != '\0' && !is_dir_sep(c))
 579                         *dst++ = c;
 580                 if (is_dir_sep(c)) {
 581                         *dst++ = '/';
 582                         while (is_dir_sep(c))
 583                                 c = *src++;
 584                         src--;
 585                 } else if (!c)
 586                         break;
 587                 continue;
 588
 589         up_one:
 590                 /*
 591                  * dst0..dst is prefix portion, and dst[-1] is '/';
 592                  * go up one level.
 593                  */
 594                 dst--;  /* go to trailing '/' */
 595                 if (dst <= dst0)
 596                         return -1;
 597                 /* Windows: dst[-1] cannot be backslash anymore */
 598                 while (dst0 < dst && dst[-1] != '/')
 599                         dst--;
 600         }
 601         *dst = '\0';
 602         return 0;
 603 }
 604
 605 /*
 606  * path = Canonical absolute path
 607  * prefix_list = Colon-separated list of absolute paths
 608  *
 609  * Determines, for each path in prefix_list, whether the "prefix" really
 610  * is an ancestor directory of path.  Returns the length of the longest
 611  * ancestor directory, excluding any trailing slashes, or -1 if no prefix
 612  * is an ancestor.  (Note that this means 0 is returned if prefix_list is
 613  * "/".) "/foo" is not considered an ancestor of "/foobar".  Directories
 614  * are not considered to be their own ancestors.  path must be in a
 615  * canonical form: empty components, or "." or ".." components are not
 616  * allowed.  prefix_list may be null, which is like "".
 617  */
 618 int longest_ancestor_length(const char *path, const char *prefix_list)
 619 {
 620         char buf[PATH_MAX+1];
 621         const char *ceil, *colon;
 622         int len, max_len = -1;
 623
 624         if (prefix_list == NULL || !strcmp(path, "/"))
 625                 return -1;
 626
 627         for (colon = ceil = prefix_list; *colon; ceil = colon+1) {
 628                 for (colon = ceil; *colon && *colon != PATH_SEP; colon++);
 629                 len = colon - ceil;
 630                 if (len == 0 || len > PATH_MAX || !is_absolute_path(ceil))
 631                         continue;
 632                 strlcpy(buf, ceil, len+1);
 633                 if (normalize_path_copy(buf, buf) < 0)
 634                         continue;
 635                 len = strlen(buf);
 636                 if (len > 0 && buf[len-1] == '/')
 637                         buf[--len] = '\0';
 638
 639                 if (!strncmp(path, buf, len) &&
 640                     path[len] == '/' &&
 641                     len > max_len) {
 642                         max_len = len;
 643                 }
 644         }
 645
 646         return max_len;
 647 }
 648
 649 /* strip arbitrary amount of directory separators at end of path */
 650 static inline int chomp_trailing_dir_sep(const char *path, int len)
 651 {
 652         while (len && is_dir_sep(path[len - 1]))
 653                 len--;
 654         return len;
 655 }
 656
 657 /*
 658  * If path ends with suffix (complete path components), returns the
 659  * part before suffix (sans trailing directory separators).
 660  * Otherwise returns NULL.
 661  */
 662 char *strip_path_suffix(const char *path, const char *suffix)
 663 {
 664         int path_len = strlen(path), suffix_len = strlen(suffix);
 665
 666         while (suffix_len) {
 667                 if (!path_len)
 668                         return NULL;
 669
 670                 if (is_dir_sep(path[path_len - 1])) {
 671                         if (!is_dir_sep(suffix[suffix_len - 1]))
 672                                 return NULL;
 673                         path_len = chomp_trailing_dir_sep(path, path_len);
 674                         suffix_len = chomp_trailing_dir_sep(suffix, suffix_len);
 675                 }
 676                 else if (path[--path_len] != suffix[--suffix_len])
 677                         return NULL;
 678         }
 679
 680         if (path_len && !is_dir_sep(path[path_len - 1]))
 681                 return NULL;
 682         return xstrndup(path, chomp_trailing_dir_sep(path, path_len));
 683 }
 684
 685 int daemon_avoid_alias(const char *p)
 686 {
 687         int sl, ndot;
 688
 689         /*
 690          * This resurrects the belts and suspenders paranoia check by HPA
 691          * done in <435560F7.4080006@zytor.com> thread, now enter_repo()
 692          * does not do getcwd() based path canonicalization.
 693          *
 694          * sl becomes true immediately after seeing '/' and continues to
 695          * be true as long as dots continue after that without intervening
 696          * non-dot character.
 697          */
 698         if (!p || (*p != '/' && *p != '~'))
 699                 return -1;
 700         sl = 1; ndot = 0;
 701         p++;
 702
 703         while (1) {
 704                 char ch = *p++;
 705                 if (sl) {
 706                         if (ch == '.')
 707                                 ndot++;
 708                         else if (ch == '/') {
 709                                 if (ndot < 3)
 710                                         /* reject //, /./ and /../ */
 711                                         return -1;
 712                                 ndot = 0;
 713                         }
 714                         else if (ch == 0) {
 715                                 if (0 < ndot && ndot < 3)
 716                                         /* reject /.$ and /..$ */
 717                                         return -1;
 718                                 return 0;
 719                         }
 720                         else
 721                                 sl = ndot = 0;
 722                 }
 723                 else if (ch == 0)
 724                         return 0;
 725                 else if (ch == '/') {
 726                         sl = 1;
 727                         ndot = 0;
 728                 }
 729         }
 730 }
 731
 732 int offset_1st_component(const char *path)
 733 {
 734         if (has_dos_drive_prefix(path))
 735                 return 2 + is_dir_sep(path[2]);
 736         return is_dir_sep(path[0]);
 737 }