gnulib/import/stat.c

   1 /* Work around platform bugs in stat.
   2    Copyright (C) 2009-2019 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 Eric Blake and Bruno Haible.  */
  18
  19 /* If the user's config.h happens to include <sys/stat.h>, let it include only
  20    the system's <sys/stat.h> here, so that orig_stat doesn't recurse to
  21    rpl_stat.  */
  22 #define __need_system_sys_stat_h
  23 #include <config.h>
  24
  25 /* Get the original definition of stat.  It might be defined as a macro.  */
  26 #include <sys/types.h>
  27 #include <sys/stat.h>
  28 #undef __need_system_sys_stat_h
  29
  30 #if defined _WIN32 && ! defined __CYGWIN__
  31 # define WINDOWS_NATIVE
  32 #endif
  33
  34 #if !defined WINDOWS_NATIVE
  35
  36 static int
  37 orig_stat (const char *filename, struct stat *buf)
  38 {
  39   return stat (filename, buf);
  40 }
  41
  42 #endif
  43
  44 /* Specification.  */
  45 #ifdef __osf__
  46 /* Write "sys/stat.h" here, not <sys/stat.h>, otherwise OSF/1 5.1 DTK cc
  47    eliminates this include because of the preliminary #include <sys/stat.h>
  48    above.  */
  49 # include "sys/stat.h"
  50 #else
  51 # include <sys/stat.h>
  52 #endif
  53
  54 #include "stat-time.h"
  55
  56 #include <errno.h>
  57 #include <limits.h>
  58 #include <stdbool.h>
  59 #include <string.h>
  60 #include "filename.h"
  61 #include "malloca.h"
  62 #include "verify.h"
  63
  64 #ifdef WINDOWS_NATIVE
  65 # define WIN32_LEAN_AND_MEAN
  66 # include <windows.h>
  67 # include "stat-w32.h"
  68 #endif
  69
  70 #ifdef WINDOWS_NATIVE
  71 /* Return TRUE if the given file name denotes an UNC root.  */
  72 static BOOL
  73 is_unc_root (const char *rname)
  74 {
  75   /* Test whether it has the syntax '\\server\share'.  */
  76   if (ISSLASH (rname[0]) && ISSLASH (rname[1]))
  77     {
  78       /* It starts with two slashes.  Find the next slash.  */
  79       const char *p = rname + 2;
  80       const char *q = p;
  81       while (*q != '\0' && !ISSLASH (*q))
  82         q++;
  83       if (q > p && *q != '\0')
  84         {
  85           /* Found the next slash at q.  */
  86           q++;
  87           const char *r = q;
  88           while (*r != '\0' && !ISSLASH (*r))
  89             r++;
  90           if (r > q && *r == '\0')
  91             return TRUE;
  92         }
  93     }
  94   return FALSE;
  95 }
  96 #endif
  97
  98 /* Store information about NAME into ST.  Work around bugs with
  99    trailing slashes.  Mingw has other bugs (such as st_ino always
 100    being 0 on success) which this wrapper does not work around.  But
 101    at least this implementation provides the ability to emulate fchdir
 102    correctly.  */
 103
 104 int
 105 rpl_stat (char const *name, struct stat *buf)
 106 {
 107 #ifdef WINDOWS_NATIVE
 108   /* Fill the fields ourselves, because the original stat function returns
 109      values for st_atime, st_mtime, st_ctime that depend on the current time
 110      zone.  See
 111      <https://lists.gnu.org/r/bug-gnulib/2017-04/msg00134.html>  */
 112   /* XXX Should we convert to wchar_t* and prepend '\\?\', in order to work
 113      around length limitations
 114      <https://docs.microsoft.com/en-us/windows/desktop/FileIO/naming-a-file> ?  */
 115
 116   /* POSIX <https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_13>
 117      specifies: "More than two leading <slash> characters shall be treated as
 118      a single <slash> character."  */
 119   if (ISSLASH (name[0]) && ISSLASH (name[1]) && ISSLASH (name[2]))
 120     {
 121       name += 2;
 122       while (ISSLASH (name[1]))
 123         name++;
 124     }
 125
 126   size_t len = strlen (name);
 127   size_t drive_prefix_len = (HAS_DEVICE (name) ? 2 : 0);
 128
 129   /* Remove trailing slashes (except the very first one, at position
 130      drive_prefix_len), but remember their presence.  */
 131   size_t rlen;
 132   bool check_dir = false;
 133
 134   rlen = len;
 135   while (rlen > drive_prefix_len && ISSLASH (name[rlen-1]))
 136     {
 137       check_dir = true;
 138       if (rlen == drive_prefix_len + 1)
 139         break;
 140       rlen--;
 141     }
 142
 143   /* Handle '' and 'C:'.  */
 144   if (!check_dir && rlen == drive_prefix_len)
 145     {
 146       errno = ENOENT;
 147       return -1;
 148     }
 149
 150   /* Handle '\\'.  */
 151   if (rlen == 1 && ISSLASH (name[0]) && len >= 2)
 152     {
 153       errno = ENOENT;
 154       return -1;
 155     }
 156
 157   const char *rname;
 158   char *malloca_rname;
 159   if (rlen == len)
 160     {
 161       rname = name;
 162       malloca_rname = NULL;
 163     }
 164   else
 165     {
 166       malloca_rname = malloca (rlen + 1);
 167       if (malloca_rname == NULL)
 168         {
 169           errno = ENOMEM;
 170           return -1;
 171         }
 172       memcpy (malloca_rname, name, rlen);
 173       malloca_rname[rlen] = '\0';
 174       rname = malloca_rname;
 175     }
 176
 177   /* There are two ways to get at the requested information:
 178        - by scanning the parent directory and examining the relevant
 179          directory entry,
 180        - by opening the file directly.
 181      The first approach fails for root directories (e.g. 'C:\') and
 182      UNC root directories (e.g. '\\server\share').
 183      The second approach fails for some system files (e.g. 'C:\pagefile.sys'
 184      and 'C:\hiberfil.sys'): ERROR_SHARING_VIOLATION.
 185      The second approach gives more information (in particular, correct
 186      st_dev, st_ino, st_nlink fields).
 187      So we use the second approach and, as a fallback except for root and
 188      UNC root directories, also the first approach.  */
 189   {
 190     int ret;
 191
 192     {
 193       /* Approach based on the file.  */
 194
 195       /* Open a handle to the file.
 196          CreateFile
 197          <https://docs.microsoft.com/en-us/windows/desktop/api/fileapi/nf-fileapi-createfilea>
 198          <https://docs.microsoft.com/en-us/windows/desktop/FileIO/creating-and-opening-files>  */
 199       HANDLE h =
 200         CreateFile (rname,
 201                     FILE_READ_ATTRIBUTES,
 202                     FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE,
 203                     NULL,
 204                     OPEN_EXISTING,
 205                     /* FILE_FLAG_POSIX_SEMANTICS (treat file names that differ only
 206                        in case as different) makes sense only when applied to *all*
 207                        filesystem operations.  */
 208                     FILE_FLAG_BACKUP_SEMANTICS /* | FILE_FLAG_POSIX_SEMANTICS */,
 209                     NULL);
 210       if (h != INVALID_HANDLE_VALUE)
 211         {
 212           ret = _gl_fstat_by_handle (h, rname, buf);
 213           CloseHandle (h);
 214           goto done;
 215         }
 216     }
 217
 218     /* Test for root and UNC root directories.  */
 219     if ((rlen == drive_prefix_len + 1 && ISSLASH (rname[drive_prefix_len]))
 220         || is_unc_root (rname))
 221       goto failed;
 222
 223     /* Fallback.  */
 224     {
 225       /* Approach based on the directory entry.  */
 226
 227       if (strchr (rname, '?') != NULL || strchr (rname, '*') != NULL)
 228         {
 229           /* Other Windows API functions would fail with error
 230              ERROR_INVALID_NAME.  */
 231           if (malloca_rname != NULL)
 232             freea (malloca_rname);
 233           errno = ENOENT;
 234           return -1;
 235         }
 236
 237       /* Get the details about the directory entry.  This can be done through
 238          FindFirstFile
 239          <https://docs.microsoft.com/en-us/windows/desktop/api/fileapi/nf-fileapi-findfirstfilea>
 240          <https://docs.microsoft.com/en-us/windows/desktop/api/minwinbase/ns-minwinbase-_win32_find_dataa>
 241          or through
 242          FindFirstFileEx with argument FindExInfoBasic
 243          <https://docs.microsoft.com/en-us/windows/desktop/api/fileapi/nf-fileapi-findfirstfileexa>
 244          <https://docs.microsoft.com/en-us/windows/desktop/api/minwinbase/ne-minwinbase-findex_info_levels>
 245          <https://docs.microsoft.com/en-us/windows/desktop/api/minwinbase/ns-minwinbase-_win32_find_dataa>  */
 246       WIN32_FIND_DATA info;
 247       HANDLE h = FindFirstFile (rname, &info);
 248       if (h == INVALID_HANDLE_VALUE)
 249         goto failed;
 250
 251       /* Test for error conditions before starting to fill *buf.  */
 252       if (sizeof (buf->st_size) <= 4 && info.nFileSizeHigh > 0)
 253         {
 254           FindClose (h);
 255           if (malloca_rname != NULL)
 256             freea (malloca_rname);
 257           errno = EOVERFLOW;
 258           return -1;
 259         }
 260
 261 # if _GL_WINDOWS_STAT_INODES
 262       buf->st_dev = 0;
 263 #  if _GL_WINDOWS_STAT_INODES == 2
 264       buf->st_ino._gl_ino[0] = buf->st_ino._gl_ino[1] = 0;
 265 #  else /* _GL_WINDOWS_STAT_INODES == 1 */
 266       buf->st_ino = 0;
 267 #  endif
 268 # else
 269       /* st_ino is not wide enough for identifying a file on a device.
 270          Without st_ino, st_dev is pointless.  */
 271       buf->st_dev = 0;
 272       buf->st_ino = 0;
 273 # endif
 274
 275       /* st_mode.  */
 276       unsigned int mode =
 277         /* XXX How to handle FILE_ATTRIBUTE_REPARSE_POINT ?  */
 278         ((info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) ? _S_IFDIR | S_IEXEC_UGO : _S_IFREG)
 279         | S_IREAD_UGO
 280         | ((info.dwFileAttributes & FILE_ATTRIBUTE_READONLY) ? 0 : S_IWRITE_UGO);
 281       if (!(info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY))
 282         {
 283           /* Determine whether the file is executable by looking at the file
 284              name suffix.  */
 285           if (info.nFileSizeHigh > 0 || info.nFileSizeLow > 0)
 286             {
 287               const char *last_dot = NULL;
 288               const char *p;
 289               for (p = info.cFileName; *p != '\0'; p++)
 290                 if (*p == '.')
 291                   last_dot = p;
 292               if (last_dot != NULL)
 293                 {
 294                   const char *suffix = last_dot + 1;
 295                   if (_stricmp (suffix, "exe") == 0
 296                       || _stricmp (suffix, "bat") == 0
 297                       || _stricmp (suffix, "cmd") == 0
 298                       || _stricmp (suffix, "com") == 0)
 299                     mode |= S_IEXEC_UGO;
 300                 }
 301             }
 302         }
 303       buf->st_mode = mode;
 304
 305       /* st_nlink.  Ignore hard links here.  */
 306       buf->st_nlink = 1;
 307
 308       /* There's no easy way to map the Windows SID concept to an integer.  */
 309       buf->st_uid = 0;
 310       buf->st_gid = 0;
 311
 312       /* st_rdev is irrelevant for normal files and directories.  */
 313       buf->st_rdev = 0;
 314
 315       /* st_size.  */
 316       if (sizeof (buf->st_size) <= 4)
 317         /* Range check already done above.  */
 318         buf->st_size = info.nFileSizeLow;
 319       else
 320         buf->st_size = ((long long) info.nFileSizeHigh << 32) | (long long) info.nFileSizeLow;
 321
 322       /* st_atime, st_mtime, st_ctime.  */
 323 # if _GL_WINDOWS_STAT_TIMESPEC
 324       buf->st_atim = _gl_convert_FILETIME_to_timespec (&info.ftLastAccessTime);
 325       buf->st_mtim = _gl_convert_FILETIME_to_timespec (&info.ftLastWriteTime);
 326       buf->st_ctim = _gl_convert_FILETIME_to_timespec (&info.ftCreationTime);
 327 # else
 328       buf->st_atime = _gl_convert_FILETIME_to_POSIX (&info.ftLastAccessTime);
 329       buf->st_mtime = _gl_convert_FILETIME_to_POSIX (&info.ftLastWriteTime);
 330       buf->st_ctime = _gl_convert_FILETIME_to_POSIX (&info.ftCreationTime);
 331 # endif
 332
 333       FindClose (h);
 334
 335       ret = 0;
 336     }
 337
 338    done:
 339     if (ret >= 0 && check_dir && !S_ISDIR (buf->st_mode))
 340       {
 341         errno = ENOTDIR;
 342         ret = -1;
 343       }
 344     if (malloca_rname != NULL)
 345       {
 346         int saved_errno = errno;
 347         freea (malloca_rname);
 348         errno = saved_errno;
 349       }
 350     return ret;
 351   }
 352
 353  failed:
 354   {
 355     DWORD error = GetLastError ();
 356     #if 0
 357     fprintf (stderr, "rpl_stat error 0x%x\n", (unsigned int) error);
 358     #endif
 359
 360     if (malloca_rname != NULL)
 361       freea (malloca_rname);
 362
 363     switch (error)
 364       {
 365       /* Some of these errors probably cannot happen with the specific flags
 366          that we pass to CreateFile.  But who knows...  */
 367       case ERROR_FILE_NOT_FOUND: /* The last component of rname does not exist.  */
 368       case ERROR_PATH_NOT_FOUND: /* Some directory component in rname does not exist.  */
 369       case ERROR_BAD_PATHNAME:   /* rname is such as '\\server'.  */
 370       case ERROR_BAD_NET_NAME:   /* rname is such as '\\server\nonexistentshare'.  */
 371       case ERROR_INVALID_NAME:   /* rname contains wildcards, misplaced colon, etc.  */
 372       case ERROR_DIRECTORY:
 373         errno = ENOENT;
 374         break;
 375
 376       case ERROR_ACCESS_DENIED:  /* rname is such as 'C:\System Volume Information\foo'.  */
 377       case ERROR_SHARING_VIOLATION: /* rname is such as 'C:\pagefile.sys' (second approach only).  */
 378                                     /* XXX map to EACCES or EPERM? */
 379         errno = EACCES;
 380         break;
 381
 382       case ERROR_OUTOFMEMORY:
 383         errno = ENOMEM;
 384         break;
 385
 386       case ERROR_WRITE_PROTECT:
 387         errno = EROFS;
 388         break;
 389
 390       case ERROR_WRITE_FAULT:
 391       case ERROR_READ_FAULT:
 392       case ERROR_GEN_FAILURE:
 393         errno = EIO;
 394         break;
 395
 396       case ERROR_BUFFER_OVERFLOW:
 397       case ERROR_FILENAME_EXCED_RANGE:
 398         errno = ENAMETOOLONG;
 399         break;
 400
 401       case ERROR_DELETE_PENDING: /* XXX map to EACCES or EPERM? */
 402         errno = EPERM;
 403         break;
 404
 405       default:
 406         errno = EINVAL;
 407         break;
 408       }
 409
 410     return -1;
 411   }
 412 #else
 413   int result = orig_stat (name, buf);
 414   if (result == 0)
 415     {
 416 # if REPLACE_FUNC_STAT_FILE
 417       /* Solaris 9 mistakenly succeeds when given a non-directory with a
 418          trailing slash.  */
 419       if (!S_ISDIR (buf->st_mode))
 420         {
 421           size_t len = strlen (name);
 422           if (ISSLASH (name[len - 1]))
 423             {
 424               errno = ENOTDIR;
 425               return -1;
 426             }
 427         }
 428 # endif /* REPLACE_FUNC_STAT_FILE */
 429       result = stat_time_normalize (result, buf);
 430     }
 431   return result;
 432 #endif
 433 }