lib/canonicalize.c

   1 /* Return the canonical absolute name of a given file.
   2    Copyright (C) 1996-2020 Free Software Foundation, Inc.
   3
   4    This program is free software: you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation; either version 3 of the License, or
   7    (at your option) any later version.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
  16
  17 #include <config.h>
  18
  19 #include "canonicalize.h"
  20
  21 #include <errno.h>
  22 #include <stdlib.h>
  23 #include <string.h>
  24 #include <sys/stat.h>
  25 #include <unistd.h>
  26
  27 #include "areadlink.h"
  28 #include "file-set.h"
  29 #include "hash-triple.h"
  30 #include "pathmax.h"
  31 #include "xalloc.h"
  32 #include "xgetcwd.h"
  33 #include "filename.h"
  34
  35 #define MULTIPLE_BITS_SET(i) (((i) & ((i) - 1)) != 0)
  36
  37 /* In this file, we cannot handle file names longer than PATH_MAX.
  38    On systems with no file name length limit, use a fallback.  */
  39 #ifndef PATH_MAX
  40 # define PATH_MAX 8192
  41 #endif
  42
  43 #ifndef DOUBLE_SLASH_IS_DISTINCT_ROOT
  44 # define DOUBLE_SLASH_IS_DISTINCT_ROOT 0
  45 #endif
  46
  47 #if ISSLASH ('\\')
  48 # define SLASHES "/\\"
  49 #else
  50 # define SLASHES "/"
  51 #endif
  52
  53 #if !((HAVE_CANONICALIZE_FILE_NAME && FUNC_REALPATH_WORKS)      \
  54       || GNULIB_CANONICALIZE_LGPL)
  55 /* Return the canonical absolute name of file NAME.  A canonical name
  56    does not contain any ".", ".." components nor any repeated file name
  57    separators ('/') or symlinks.  All components must exist.
  58    The result is malloc'd.  */
  59
  60 char *
  61 canonicalize_file_name (const char *name)
  62 {
  63   return canonicalize_filename_mode (name, CAN_EXISTING);
  64 }
  65 #endif /* !HAVE_CANONICALIZE_FILE_NAME */
  66
  67 /* Return true if we've already seen the triple, <FILENAME, dev, ino>.
  68    If *HT is not initialized, initialize it.  */
  69 static bool
  70 seen_triple (Hash_table **ht, char const *filename, struct stat const *st)
  71 {
  72   if (*ht == NULL)
  73     {
  74       size_t initial_capacity = 7;
  75       *ht = hash_initialize (initial_capacity,
  76                             NULL,
  77                             triple_hash,
  78                             triple_compare_ino_str,
  79                             triple_free);
  80       if (*ht == NULL)
  81         xalloc_die ();
  82     }
  83
  84   if (seen_file (*ht, filename, st))
  85     return true;
  86
  87   record_file (*ht, filename, st);
  88   return false;
  89 }
  90
  91 /* Return the canonical absolute name of file NAME, while treating
  92    missing elements according to CAN_MODE.  A canonical name
  93    does not contain any ".", ".." components nor any repeated file name
  94    separators ('/') or, depending on other CAN_MODE flags, symlinks.
  95    Whether components must exist or not depends on canonicalize mode.
  96    The result is malloc'd.  */
  97
  98 char *
  99 canonicalize_filename_mode (const char *name, canonicalize_mode_t can_mode)
 100 {
 101   char *rname, *dest, *extra_buf = NULL;
 102   char const *start;
 103   char const *end;
 104   char const *rname_limit;
 105   size_t extra_len = 0;
 106   Hash_table *ht = NULL;
 107   int saved_errno;
 108   int can_flags = can_mode & ~CAN_MODE_MASK;
 109   bool logical = can_flags & CAN_NOLINKS;
 110   size_t prefix_len;
 111
 112   can_mode &= CAN_MODE_MASK;
 113
 114   if (MULTIPLE_BITS_SET (can_mode))
 115     {
 116       errno = EINVAL;
 117       return NULL;
 118     }
 119
 120   if (name == NULL)
 121     {
 122       errno = EINVAL;
 123       return NULL;
 124     }
 125
 126   if (name[0] == '\0')
 127     {
 128       errno = ENOENT;
 129       return NULL;
 130     }
 131
 132   /* This is always zero for Posix hosts, but can be 2 for MS-Windows
 133      and MS-DOS X:/foo/bar file names.  */
 134   prefix_len = FILE_SYSTEM_PREFIX_LEN (name);
 135
 136   if (!IS_ABSOLUTE_FILE_NAME (name))
 137     {
 138       rname = xgetcwd ();
 139       if (!rname)
 140         return NULL;
 141       dest = strchr (rname, '\0');
 142       if (dest - rname < PATH_MAX)
 143         {
 144           char *p = xrealloc (rname, PATH_MAX);
 145           dest = p + (dest - rname);
 146           rname = p;
 147           rname_limit = rname + PATH_MAX;
 148         }
 149       else
 150         {
 151           rname_limit = dest;
 152         }
 153       start = name;
 154       prefix_len = FILE_SYSTEM_PREFIX_LEN (rname);
 155     }
 156   else
 157     {
 158       rname = xmalloc (PATH_MAX);
 159       rname_limit = rname + PATH_MAX;
 160       dest = rname;
 161       if (prefix_len)
 162         {
 163           memcpy (rname, name, prefix_len);
 164           dest += prefix_len;
 165         }
 166       *dest++ = '/';
 167       if (DOUBLE_SLASH_IS_DISTINCT_ROOT)
 168         {
 169           if (ISSLASH (name[1]) && !ISSLASH (name[2]) && !prefix_len)
 170             *dest++ = '/';
 171           *dest = '\0';
 172         }
 173       start = name + prefix_len;
 174     }
 175
 176   for ( ; *start; start = end)
 177     {
 178       /* Skip sequence of multiple file name separators.  */
 179       while (ISSLASH (*start))
 180         ++start;
 181
 182       /* Find end of component.  */
 183       for (end = start; *end && !ISSLASH (*end); ++end)
 184         /* Nothing.  */;
 185
 186       if (end - start == 0)
 187         break;
 188       else if (end - start == 1 && start[0] == '.')
 189         /* nothing */;
 190       else if (end - start == 2 && start[0] == '.' && start[1] == '.')
 191         {
 192           /* Back up to previous component, ignore if at root already.  */
 193           if (dest > rname + prefix_len + 1)
 194             for (--dest; dest > rname && !ISSLASH (dest[-1]); --dest)
 195               continue;
 196           if (DOUBLE_SLASH_IS_DISTINCT_ROOT && dest == rname + 1
 197               && !prefix_len && ISSLASH (*dest) && !ISSLASH (dest[1]))
 198             dest++;
 199         }
 200       else
 201         {
 202           struct stat st;
 203
 204           if (!ISSLASH (dest[-1]))
 205             *dest++ = '/';
 206
 207           if (dest + (end - start) >= rname_limit)
 208             {
 209               ptrdiff_t dest_offset = dest - rname;
 210               size_t new_size = rname_limit - rname;
 211
 212               if (end - start + 1 > PATH_MAX)
 213                 new_size += end - start + 1;
 214               else
 215                 new_size += PATH_MAX;
 216               rname = xrealloc (rname, new_size);
 217               rname_limit = rname + new_size;
 218
 219               dest = rname + dest_offset;
 220             }
 221
 222           dest = memcpy (dest, start, end - start);
 223           dest += end - start;
 224           *dest = '\0';
 225
 226           if (logical && (can_mode == CAN_MISSING))
 227             {
 228               /* Avoid the stat in this case as it's inconsequential.
 229                  i.e. we're neither resolving symlinks or testing
 230                  component existence.  */
 231               st.st_mode = 0;
 232             }
 233           else if ((logical ? stat (rname, &st) : lstat (rname, &st)) != 0)
 234             {
 235               /* FIXME: If errno == EOVERFLOW here, the entry exists.  */
 236               saved_errno = errno;
 237               if (can_mode == CAN_EXISTING)
 238                 goto error;
 239               if (can_mode == CAN_ALL_BUT_LAST)
 240                 {
 241                   if (end[strspn (end, SLASHES)] || saved_errno != ENOENT)
 242                     goto error;
 243                   continue;
 244                 }
 245               st.st_mode = 0;
 246             }
 247
 248           if (S_ISLNK (st.st_mode))
 249             {
 250               char *buf;
 251               size_t n, len;
 252
 253               /* Detect loops.  We cannot use the cycle-check module here,
 254                  since it's actually possible to encounter the same symlink
 255                  more than once in a given traversal.  However, encountering
 256                  the same symlink,NAME pair twice does indicate a loop.  */
 257               if (seen_triple (&ht, name, &st))
 258                 {
 259                   if (can_mode == CAN_MISSING)
 260                     continue;
 261                   saved_errno = ELOOP;
 262                   goto error;
 263                 }
 264
 265               buf = areadlink_with_size (rname, st.st_size);
 266               if (!buf)
 267                 {
 268                   if (can_mode == CAN_MISSING && errno != ENOMEM)
 269                     continue;
 270                   saved_errno = errno;
 271                   goto error;
 272                 }
 273
 274               n = strlen (buf);
 275               len = strlen (end);
 276
 277               if (!extra_len)
 278                 {
 279                   extra_len =
 280                     ((n + len + 1) > PATH_MAX) ? (n + len + 1) : PATH_MAX;
 281                   extra_buf = xmalloc (extra_len);
 282                 }
 283               else if ((n + len + 1) > extra_len)
 284                 {
 285                   extra_len = n + len + 1;
 286                   extra_buf = xrealloc (extra_buf, extra_len);
 287                 }
 288
 289               /* Careful here, end may be a pointer into extra_buf... */
 290               memmove (&extra_buf[n], end, len + 1);
 291               name = end = memcpy (extra_buf, buf, n);
 292
 293               if (IS_ABSOLUTE_FILE_NAME (buf))
 294                 {
 295                   size_t pfxlen = FILE_SYSTEM_PREFIX_LEN (buf);
 296
 297                   if (pfxlen)
 298                     memcpy (rname, buf, pfxlen);
 299                   dest = rname + pfxlen;
 300                   *dest++ = '/'; /* It's an absolute symlink */
 301                   if (DOUBLE_SLASH_IS_DISTINCT_ROOT)
 302                     {
 303                       if (ISSLASH (buf[1]) && !ISSLASH (buf[2]) && !pfxlen)
 304                         *dest++ = '/';
 305                       *dest = '\0';
 306                     }
 307                   /* Install the new prefix to be in effect hereafter.  */
 308                   prefix_len = pfxlen;
 309                 }
 310               else
 311                 {
 312                   /* Back up to previous component, ignore if at root
 313                      already: */
 314                   if (dest > rname + prefix_len + 1)
 315                     for (--dest; dest > rname && !ISSLASH (dest[-1]); --dest)
 316                       continue;
 317                   if (DOUBLE_SLASH_IS_DISTINCT_ROOT && dest == rname + 1
 318                       && ISSLASH (*dest) && !ISSLASH (dest[1]) && !prefix_len)
 319                     dest++;
 320                 }
 321
 322               free (buf);
 323             }
 324           else
 325             {
 326               if (!S_ISDIR (st.st_mode) && *end && (can_mode != CAN_MISSING))
 327                 {
 328                   saved_errno = ENOTDIR;
 329                   goto error;
 330                 }
 331             }
 332         }
 333     }
 334   if (dest > rname + prefix_len + 1 && ISSLASH (dest[-1]))
 335     --dest;
 336   if (DOUBLE_SLASH_IS_DISTINCT_ROOT && dest == rname + 1 && !prefix_len
 337       && ISSLASH (*dest) && !ISSLASH (dest[1]))
 338     dest++;
 339   *dest = '\0';
 340   if (rname_limit != dest + 1)
 341     rname = xrealloc (rname, dest - rname + 1);
 342
 343   free (extra_buf);
 344   if (ht)
 345     hash_free (ht);
 346   return rname;
 347
 348 error:
 349   free (extra_buf);
 350   free (rname);
 351   if (ht)
 352     hash_free (ht);
 353   errno = saved_errno;
 354   return NULL;
 355 }