tests/test-canonicalize.c

   1 /* Test of execution of file name canonicalization.
   2    Copyright (C) 2007-2025 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 /* Written by Bruno Haible <bruno@clisp.org>, 2007.  */
  18
  19 /* Don't use __attribute__ __nonnull__ in this compilation unit.  Otherwise gcc
  20    may "optimize" the null_ptr function, when its result gets passed to a
  21    function that has an argument declared as _GL_ARG_NONNULL.  */
  22 #define _GL_ARG_NONNULL(params)
  23
  24 #include <config.h>
  25
  26 #include "canonicalize.h"
  27
  28 #include <errno.h>
  29 #include <fcntl.h>
  30 #include <stdio.h>
  31 #include <stdlib.h>
  32 #include <string.h>
  33 #include <sys/stat.h>
  34 #include <unistd.h>
  35
  36 #include "same-inode.h"
  37 #include "ignore-value.h"
  38
  39 #if GNULIB_defined_canonicalize_file_name
  40 # include "null-ptr.h"
  41 #endif
  42
  43 #include "macros.h"
  44
  45 #define BASE "t-can.tmp"
  46
  47 int
  48 main (void)
  49 {
  50   /* Setup some hierarchy to be used by this test.  Start by removing
  51      any leftovers from a previous partial run.  */
  52   {
  53     int fd;
  54     ignore_value (system ("rm -rf " BASE " ise"));
  55     ASSERT (mkdir (BASE, 0700) == 0);
  56     fd = creat (BASE "/tra", 0600);
  57     ASSERT (0 <= fd);
  58     ASSERT (close (fd) == 0);
  59   }
  60
  61   /* Check // handling (the easy cases, without symlinks).
  62      This // handling is not mandated by POSIX.  However, many applications
  63      expect that canonicalize_filename_mode "canonicalizes" the file name,
  64      that is, that different results of canonicalize_filename_mode correspond
  65      to different files (except for hard links).  */
  66   {
  67     char *result0 = canonicalize_file_name ("/etc/passwd");
  68     if (result0 != NULL) /* This file does not exist on native Windows.  */
  69       {
  70         char *result;
  71
  72         result = canonicalize_filename_mode ("/etc/passwd", CAN_MISSING);
  73         ASSERT (result != NULL && strcmp (result, result0) == 0);
  74
  75         result = canonicalize_filename_mode ("/etc//passwd", CAN_MISSING);
  76         ASSERT (result != NULL && strcmp (result, result0) == 0);
  77
  78         result = canonicalize_filename_mode ("/etc///passwd", CAN_MISSING);
  79         ASSERT (result != NULL && strcmp (result, result0) == 0);
  80
  81         /* On Windows, the syntax //host/share/filename denotes a file
  82            in a directory named 'share', exported from host 'host'.
  83            See also m4/double-slash-root.m4.  */
  84 #if !(defined _WIN32 || defined __CYGWIN__)
  85         result = canonicalize_filename_mode ("//etc/passwd", CAN_MISSING);
  86         ASSERT (result != NULL && strcmp (result, result0) == 0);
  87
  88         result = canonicalize_filename_mode ("//etc//passwd", CAN_MISSING);
  89         ASSERT (result != NULL && strcmp (result, result0) == 0);
  90
  91         result = canonicalize_filename_mode ("//etc///passwd", CAN_MISSING);
  92         ASSERT (result != NULL && strcmp (result, result0) == 0);
  93 #endif
  94
  95         result = canonicalize_filename_mode ("///etc/passwd", CAN_MISSING);
  96         ASSERT (result != NULL && strcmp (result, result0) == 0);
  97
  98         result = canonicalize_filename_mode ("///etc//passwd", CAN_MISSING);
  99         ASSERT (result != NULL && strcmp (result, result0) == 0);
 100
 101         result = canonicalize_filename_mode ("///etc///passwd", CAN_MISSING);
 102         ASSERT (result != NULL && strcmp (result, result0) == 0);
 103       }
 104   }
 105
 106   /* Check for ., .., intermediate // handling, and for error cases.  */
 107   {
 108     char *result1 = canonicalize_file_name (BASE "//./..//" BASE "/tra");
 109     char *result2 = canonicalize_filename_mode (BASE "//./..//" BASE "/tra",
 110                                                 CAN_EXISTING);
 111     ASSERT (result1 != NULL);
 112     ASSERT (result2 != NULL);
 113     ASSERT (strcmp (result1, result2) == 0);
 114     ASSERT (strstr (result1, "/" BASE "/tra")
 115             == result1 + strlen (result1) - strlen ("/" BASE "/tra"));
 116     free (result1);
 117     free (result2);
 118
 119     errno = 0;
 120     result1 = canonicalize_file_name ("");
 121     ASSERT (result1 == NULL);
 122     ASSERT (errno == ENOENT);
 123
 124     errno = 0;
 125     result2 = canonicalize_filename_mode ("", CAN_EXISTING);
 126     ASSERT (result2 == NULL);
 127     ASSERT (errno == ENOENT);
 128
 129     /* This test works only if the canonicalize_file_name implementation
 130        comes from gnulib.  If it comes from libc, we have no way to prevent
 131        gcc from "optimizing" the null_ptr function in invalid ways.  See
 132        <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93156>.  */
 133 #if GNULIB_defined_canonicalize_file_name
 134     errno = 0;
 135     result1 = canonicalize_file_name (null_ptr ());
 136     ASSERT (result1 == NULL);
 137     ASSERT (errno == EINVAL);
 138 #endif
 139
 140     errno = 0;
 141     result2 = canonicalize_filename_mode (NULL, CAN_EXISTING);
 142     ASSERT (result2 == NULL);
 143     ASSERT (errno == EINVAL);
 144
 145     errno = 0;
 146     result2 = canonicalize_filename_mode (".", CAN_MISSING | CAN_ALL_BUT_LAST);
 147     ASSERT (result2 == NULL);
 148     ASSERT (errno == EINVAL);
 149   }
 150
 151   /* Check that a non-directory with trailing slash yields NULL.  */
 152   {
 153     char *result1;
 154     char *result2;
 155     errno = 0;
 156     result1 = canonicalize_file_name (BASE "/tra/");
 157     ASSERT (result1 == NULL);
 158     ASSERT (errno == ENOTDIR);
 159     errno = 0;
 160     result2 = canonicalize_filename_mode (BASE "/tra/", CAN_EXISTING);
 161     ASSERT (result2 == NULL);
 162     ASSERT (errno == ENOTDIR);
 163   }
 164
 165   /* Check that a missing directory yields NULL.  */
 166   {
 167     char *result1;
 168     char *result2;
 169     errno = 0;
 170     result1 = canonicalize_file_name (BASE "/zzz/..");
 171     ASSERT (result1 == NULL);
 172     ASSERT (errno == ENOENT);
 173     errno = 0;
 174     result2 = canonicalize_filename_mode (BASE "/zzz/..", CAN_EXISTING);
 175     ASSERT (result2 == NULL);
 176     ASSERT (errno == ENOENT);
 177   }
 178
 179   /* From here on out, tests involve symlinks.  */
 180   if (symlink (BASE "/ket", "ise") != 0)
 181     {
 182       ASSERT (remove (BASE "/tra") == 0);
 183       ASSERT (rmdir (BASE) == 0);
 184       if (test_exit_status != EXIT_SUCCESS)
 185         return test_exit_status;
 186       fputs ("skipping test: symlinks not supported on this file system\n",
 187              stderr);
 188       return 77;
 189     }
 190   ASSERT (symlink ("bef", BASE "/plo") == 0);
 191   ASSERT (symlink ("tra", BASE "/huk") == 0);
 192   ASSERT (symlink ("lum", BASE "/bef") == 0);
 193   ASSERT (symlink ("wum", BASE "/ouk") == 0);
 194   ASSERT (symlink ("../ise", BASE "/ket") == 0);
 195   ASSERT (mkdir (BASE "/lum", 0700) == 0);
 196   ASSERT (symlink ("s", BASE "/p") == 0);
 197   ASSERT (symlink ("d", BASE "/s") == 0);
 198   ASSERT (mkdir (BASE "/d", 0700) == 0);
 199   ASSERT (close (creat (BASE "/d/2", 0600)) == 0);
 200   ASSERT (symlink ("../s/2", BASE "/d/1") == 0);
 201   ASSERT (symlink ("//.//../..", BASE "/droot") == 0);
 202
 203   /* Check that symbolic links are not resolved, with CAN_NOLINKS.  */
 204   {
 205     char *result1 = canonicalize_filename_mode (BASE "/huk", CAN_NOLINKS);
 206     ASSERT (result1 != NULL);
 207     ASSERT (str_endswith (result1, "/" BASE "/huk"));
 208     free (result1);
 209   }
 210
 211   /* Check that the symbolic link to a file can be resolved.  */
 212   {
 213     char *result1 = canonicalize_file_name (BASE "/huk");
 214     char *result2 = canonicalize_file_name (BASE "/tra");
 215     char *result3 = canonicalize_filename_mode (BASE "/huk", CAN_EXISTING);
 216     ASSERT (result1 != NULL);
 217     ASSERT (result2 != NULL);
 218     ASSERT (result3 != NULL);
 219     ASSERT (strcmp (result1, result2) == 0);
 220     ASSERT (strcmp (result2, result3) == 0);
 221     ASSERT (str_endswith (result1, "/" BASE "/tra"));
 222     free (result1);
 223     free (result2);
 224     free (result3);
 225   }
 226
 227   /* Check that the symbolic link to a directory can be resolved.  */
 228   {
 229     char *result1 = canonicalize_file_name (BASE "/plo");
 230     char *result2 = canonicalize_file_name (BASE "/bef");
 231     char *result3 = canonicalize_file_name (BASE "/lum");
 232     char *result4 = canonicalize_filename_mode (BASE "/plo", CAN_EXISTING);
 233     ASSERT (result1 != NULL);
 234     ASSERT (result2 != NULL);
 235     ASSERT (result3 != NULL);
 236     ASSERT (result4 != NULL);
 237     ASSERT (strcmp (result1, result2) == 0);
 238     ASSERT (strcmp (result2, result3) == 0);
 239     ASSERT (strcmp (result3, result4) == 0);
 240     ASSERT (str_endswith (result1, "/" BASE "/lum"));
 241     free (result1);
 242     free (result2);
 243     free (result3);
 244     free (result4);
 245   }
 246
 247   /* Check that a symbolic link to a nonexistent file yields NULL.  */
 248   {
 249     char *result1;
 250     char *result2;
 251     errno = 0;
 252     result1 = canonicalize_file_name (BASE "/ouk");
 253     ASSERT (result1 == NULL);
 254     ASSERT (errno == ENOENT);
 255     errno = 0;
 256     result2 = canonicalize_filename_mode (BASE "/ouk", CAN_EXISTING);
 257     ASSERT (result2 == NULL);
 258     ASSERT (errno == ENOENT);
 259   }
 260
 261   /* Check that a non-directory symlink with trailing slash yields NULL,
 262      and likewise for other troublesome suffixes.  */
 263   {
 264     char const *const file_name[]
 265       = {
 266          BASE "/huk/",
 267          BASE "/huk/.",
 268          BASE "/huk/./",
 269          BASE "/huk/./.",
 270          BASE "/huk/x",
 271          BASE "/huk/..",
 272          BASE "/huk/../",
 273          BASE "/huk/../.",
 274          BASE "/huk/../x",
 275          BASE "/huk/./..",
 276          BASE "/huk/././../x",
 277         };
 278     for (int i = 0; i < sizeof file_name / sizeof *file_name; i++)
 279       {
 280         errno = 0;
 281         ASSERT (!canonicalize_file_name (file_name[i]));
 282         ASSERT (errno == ENOTDIR);
 283         errno = 0;
 284         ASSERT (!canonicalize_filename_mode (file_name[i], CAN_EXISTING));
 285         ASSERT (errno == ENOTDIR);
 286       }
 287   }
 288
 289   /* Check that a missing directory via symlink yields NULL.  */
 290   {
 291     char *result1;
 292     char *result2;
 293     errno = 0;
 294     result1 = canonicalize_file_name (BASE "/ouk/..");
 295     ASSERT (result1 == NULL);
 296     ASSERT (errno == ENOENT);
 297     errno = 0;
 298     result2 = canonicalize_filename_mode (BASE "/ouk/..", CAN_EXISTING);
 299     ASSERT (result2 == NULL);
 300     ASSERT (errno == ENOENT);
 301   }
 302
 303   /* Check that a loop of symbolic links is detected.  */
 304   {
 305     char *result1;
 306     char *result2;
 307     errno = 0;
 308     result1 = canonicalize_file_name ("ise");
 309     ASSERT (result1 == NULL);
 310     ASSERT (errno == ELOOP);
 311     errno = 0;
 312     result2 = canonicalize_filename_mode ("ise", CAN_EXISTING);
 313     ASSERT (result2 == NULL);
 314     ASSERT (errno == ELOOP);
 315   }
 316
 317   /* Check that alternate modes can resolve missing basenames.  */
 318   {
 319     char *result1 = canonicalize_filename_mode (BASE "/zzz", CAN_ALL_BUT_LAST);
 320     char *result2 = canonicalize_filename_mode (BASE "/zzz", CAN_MISSING);
 321     char *result3 = canonicalize_filename_mode (BASE "/zzz/", CAN_ALL_BUT_LAST);
 322     char *result4 = canonicalize_filename_mode (BASE "/zzz/", CAN_MISSING);
 323     ASSERT (result1 != NULL);
 324     ASSERT (result2 != NULL);
 325     ASSERT (result3 != NULL);
 326     ASSERT (result4 != NULL);
 327     ASSERT (strcmp (result1, result2) == 0);
 328     ASSERT (strcmp (result2, result3) == 0);
 329     ASSERT (strcmp (result3, result4) == 0);
 330     ASSERT (str_endswith (result1, "/" BASE "/zzz"));
 331     free (result1);
 332     free (result2);
 333     free (result3);
 334     free (result4);
 335   }
 336
 337   /* Check that alternate modes can resolve broken symlink basenames.  */
 338   {
 339     char *result1 = canonicalize_filename_mode (BASE "/ouk", CAN_ALL_BUT_LAST);
 340     char *result2 = canonicalize_filename_mode (BASE "/ouk", CAN_MISSING);
 341     char *result3 = canonicalize_filename_mode (BASE "/ouk/", CAN_ALL_BUT_LAST);
 342     char *result4 = canonicalize_filename_mode (BASE "/ouk/", CAN_MISSING);
 343     ASSERT (result1 != NULL);
 344     ASSERT (result2 != NULL);
 345     ASSERT (result3 != NULL);
 346     ASSERT (result4 != NULL);
 347     ASSERT (strcmp (result1, result2) == 0);
 348     ASSERT (strcmp (result2, result3) == 0);
 349     ASSERT (strcmp (result3, result4) == 0);
 350     ASSERT (str_endswith (result1, "/" BASE "/wum"));
 351     free (result1);
 352     free (result2);
 353     free (result3);
 354     free (result4);
 355   }
 356
 357   /* Check that alternate modes can handle missing dirnames.  */
 358   {
 359     char *result1 = canonicalize_filename_mode ("t-can.zzz/zzz", CAN_ALL_BUT_LAST);
 360     char *result2 = canonicalize_filename_mode ("t-can.zzz/zzz", CAN_MISSING);
 361     ASSERT (result1 == NULL);
 362     ASSERT (result2 != NULL);
 363     ASSERT (str_endswith (result2, "/t-can.zzz/zzz"));
 364     free (result2);
 365   }
 366
 367   /* Ensure that the following is resolved properly.
 368      Before 2007-09-27, it would mistakenly report a loop.  */
 369   {
 370     char *result1 = canonicalize_filename_mode (BASE, CAN_EXISTING);
 371     char *result2 = canonicalize_filename_mode (BASE "/p/1", CAN_EXISTING);
 372     ASSERT (result1 != NULL);
 373     ASSERT (result2 != NULL);
 374     ASSERT (strcmp (result2 + strlen (result1), "/d/2") == 0);
 375     free (result1);
 376     free (result2);
 377   }
 378
 379   /* Check that leading // within symlinks is honored correctly.  */
 380   {
 381     struct stat st1;
 382     struct stat st2;
 383     char *result1 = canonicalize_file_name ("//.");
 384     char *result2 = canonicalize_filename_mode ("//.", CAN_EXISTING);
 385     char *result3 = canonicalize_file_name (BASE "/droot");
 386     char *result4 = canonicalize_filename_mode (BASE "/droot", CAN_EXISTING);
 387     ASSERT (result1);
 388     ASSERT (result2);
 389     ASSERT (result3);
 390     ASSERT (result4);
 391     ASSERT (stat ("/", &st1) == 0);
 392     ASSERT (stat ("//", &st2) == 0);
 393     bool same = psame_inode (&st1, &st2);
 394 #if defined __MVS__
 395     /* On IBM z/OS, "/" and "//" both canonicalize to themselves, yet they both
 396        have st_dev == st_ino == 1.  */
 397     same = false;
 398 #endif
 399     if (same)
 400       {
 401         ASSERT (strcmp (result1, "/") == 0);
 402         ASSERT (strcmp (result2, "/") == 0);
 403         ASSERT (strcmp (result3, "/") == 0);
 404         ASSERT (strcmp (result4, "/") == 0);
 405       }
 406     else
 407       {
 408         ASSERT (strcmp (result1, "//") == 0);
 409         ASSERT (strcmp (result2, "//") == 0);
 410         ASSERT (strcmp (result3, "//") == 0);
 411         ASSERT (strcmp (result4, "//") == 0);
 412       }
 413     free (result1);
 414     free (result2);
 415     free (result3);
 416     free (result4);
 417   }
 418
 419 #if !(defined _WIN32 && !defined __CYGWIN__)
 420   /* Check a device file.  */
 421   {
 422     char *result1 = canonicalize_file_name ("/dev/null");
 423     char *result2 = canonicalize_filename_mode ("/dev/null", CAN_ALL_BUT_LAST);
 424     ASSERT (result1 != NULL);
 425     ASSERT (result2 != NULL);
 426     ASSERT (strcmp (result1, result2) == 0);
 427   }
 428 #endif
 429
 430   /* Cleanup.  */
 431   ASSERT (remove (BASE "/droot") == 0);
 432   ASSERT (remove (BASE "/d/1") == 0);
 433   ASSERT (remove (BASE "/d/2") == 0);
 434   ASSERT (remove (BASE "/d") == 0);
 435   ASSERT (remove (BASE "/s") == 0);
 436   ASSERT (remove (BASE "/p") == 0);
 437   ASSERT (remove (BASE "/plo") == 0);
 438   ASSERT (remove (BASE "/huk") == 0);
 439   ASSERT (remove (BASE "/bef") == 0);
 440   ASSERT (remove (BASE "/ouk") == 0);
 441   ASSERT (remove (BASE "/ket") == 0);
 442   ASSERT (remove (BASE "/lum") == 0);
 443   ASSERT (remove (BASE "/tra") == 0);
 444   ASSERT (remove (BASE) == 0);
 445   ASSERT (remove ("ise") == 0);
 446
 447   return test_exit_status;
 448 }