mingw/share/gettext/intl/localcharset.c

   1 /* Determine a canonical name for the current locale's character encoding.
   2
   3    Copyright (C) 2000-2006, 2008-2010 Free Software Foundation, Inc.
   4
   5    This program is free software; you can redistribute it and/or modify it
   6    under the terms of the GNU Library General Public License as published
   7    by the Free Software Foundation; either version 2, or (at your option)
   8    any later version.
   9
  10    This program is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Library General Public License for more details.
  14
  15    You should have received a copy of the GNU Library General Public
  16    License along with this program; if not, write to the Free Software
  17    Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301,
  18    USA.  */
  19
  20 /* Written by Bruno Haible <bruno@clisp.org>.  */
  21
  22 #include <config.h>
  23
  24 /* Specification.  */
  25 #include "localcharset.h"
  26
  27 #include <fcntl.h>
  28 #include <stddef.h>
  29 #include <stdio.h>
  30 #include <string.h>
  31 #include <stdlib.h>
  32
  33 #if defined __APPLE__ && defined __MACH__ && HAVE_LANGINFO_CODESET
  34 # define DARWIN7 /* Darwin 7 or newer, i.e. MacOS X 10.3 or newer */
  35 #endif
  36
  37 #if defined _WIN32 || defined __WIN32__
  38 # define WIN32_NATIVE
  39 #endif
  40
  41 #if defined __EMX__
  42 /* Assume EMX program runs on OS/2, even if compiled under DOS.  */
  43 # ifndef OS2
  44 #  define OS2
  45 # endif
  46 #endif
  47
  48 #if !defined WIN32_NATIVE
  49 # include <unistd.h>
  50 # if HAVE_LANGINFO_CODESET
  51 #  include <langinfo.h>
  52 # else
  53 #  if 0 /* see comment below */
  54 #   include <locale.h>
  55 #  endif
  56 # endif
  57 # ifdef __CYGWIN__
  58 #  define WIN32_LEAN_AND_MEAN
  59 #  include <windows.h>
  60 # endif
  61 #elif defined WIN32_NATIVE
  62 # define WIN32_LEAN_AND_MEAN
  63 # include <windows.h>
  64 #endif
  65 #if defined OS2
  66 # define INCL_DOS
  67 # include <os2.h>
  68 #endif
  69
  70 #if ENABLE_RELOCATABLE
  71 # include "relocatable.h"
  72 #else
  73 # define relocate(pathname) (pathname)
  74 #endif
  75
  76 /* Get LIBDIR.  */
  77 #ifndef LIBDIR
  78 # include "configmake.h"
  79 #endif
  80
  81 /* Define O_NOFOLLOW to 0 on platforms where it does not exist.  */
  82 #ifndef O_NOFOLLOW
  83 # define O_NOFOLLOW 0
  84 #endif
  85
  86 #if defined _WIN32 || defined __WIN32__ || defined __CYGWIN__ || defined __EMX__ || defined __DJGPP__
  87   /* Win32, Cygwin, OS/2, DOS */
  88 # define ISSLASH(C) ((C) == '/' || (C) == '\\')
  89 #endif
  90
  91 #ifndef DIRECTORY_SEPARATOR
  92 # define DIRECTORY_SEPARATOR '/'
  93 #endif
  94
  95 #ifndef ISSLASH
  96 # define ISSLASH(C) ((C) == DIRECTORY_SEPARATOR)
  97 #endif
  98
  99 #if HAVE_DECL_GETC_UNLOCKED
 100 # undef getc
 101 # define getc getc_unlocked
 102 #endif
 103
 104 /* The following static variable is declared 'volatile' to avoid a
 105    possible multithread problem in the function get_charset_aliases. If we
 106    are running in a threaded environment, and if two threads initialize
 107    'charset_aliases' simultaneously, both will produce the same value,
 108    and everything will be ok if the two assignments to 'charset_aliases'
 109    are atomic. But I don't know what will happen if the two assignments mix.  */
 110 #if __STDC__ != 1
 111 # define volatile /* empty */
 112 #endif
 113 /* Pointer to the contents of the charset.alias file, if it has already been
 114    read, else NULL.  Its format is:
 115    ALIAS_1 '\0' CANONICAL_1 '\0' ... ALIAS_n '\0' CANONICAL_n '\0' '\0'  */
 116 static const char * volatile charset_aliases;
 117
 118 /* Return a pointer to the contents of the charset.alias file.  */
 119 static const char *
 120 get_charset_aliases (void)
 121 {
 122   const char *cp;
 123
 124   cp = charset_aliases;
 125   if (cp == NULL)
 126     {
 127 #if !(defined DARWIN7 || defined VMS || defined WIN32_NATIVE || defined __CYGWIN__)
 128       const char *dir;
 129       const char *base = "charset.alias";
 130       char *file_name;
 131
 132       /* Make it possible to override the charset.alias location.  This is
 133          necessary for running the testsuite before "make install".  */
 134       dir = getenv ("CHARSETALIASDIR");
 135       if (dir == NULL || dir[0] == '\0')
 136         dir = relocate (LIBDIR);
 137
 138       /* Concatenate dir and base into freshly allocated file_name.  */
 139       {
 140         size_t dir_len = strlen (dir);
 141         size_t base_len = strlen (base);
 142         int add_slash = (dir_len > 0 && !ISSLASH (dir[dir_len - 1]));
 143         file_name = (char *) malloc (dir_len + add_slash + base_len + 1);
 144         if (file_name != NULL)
 145           {
 146             memcpy (file_name, dir, dir_len);
 147             if (add_slash)
 148               file_name[dir_len] = DIRECTORY_SEPARATOR;
 149             memcpy (file_name + dir_len + add_slash, base, base_len + 1);
 150           }
 151       }
 152
 153       if (file_name == NULL)
 154         /* Out of memory.  Treat the file as empty.  */
 155         cp = "";
 156       else
 157         {
 158           int fd;
 159
 160           /* Open the file.  Reject symbolic links on platforms that support
 161              O_NOFOLLOW.  This is a security feature.  Without it, an attacker
 162              could retrieve parts of the contents (namely, the tail of the
 163              first line that starts with "* ") of an arbitrary file by placing
 164              a symbolic link to that file under the name "charset.alias" in
 165              some writable directory and defining the environment variable
 166              CHARSETALIASDIR to point to that directory.  */
 167           fd = open (file_name,
 168                      O_RDONLY | (HAVE_WORKING_O_NOFOLLOW ? O_NOFOLLOW : 0));
 169           if (fd < 0)
 170             /* File not found.  Treat it as empty.  */
 171             cp = "";
 172           else
 173             {
 174               FILE *fp;
 175
 176               fp = fdopen (fd, "r");
 177               if (fp == NULL)
 178                 {
 179                   /* Out of memory.  Treat the file as empty.  */
 180                   close (fd);
 181                   cp = "";
 182                 }
 183               else
 184                 {
 185                   /* Parse the file's contents.  */
 186                   char *res_ptr = NULL;
 187                   size_t res_size = 0;
 188
 189                   for (;;)
 190                     {
 191                       int c;
 192                       char buf1[50+1];
 193                       char buf2[50+1];
 194                       size_t l1, l2;
 195                       char *old_res_ptr;
 196
 197                       c = getc (fp);
 198                       if (c == EOF)
 199                         break;
 200                       if (c == '\n' || c == ' ' || c == '\t')
 201                         continue;
 202                       if (c == '#')
 203                         {
 204                           /* Skip comment, to end of line.  */
 205                           do
 206                             c = getc (fp);
 207                           while (!(c == EOF || c == '\n'));
 208                           if (c == EOF)
 209                             break;
 210                           continue;
 211                         }
 212                       ungetc (c, fp);
 213                       if (fscanf (fp, "%50s %50s", buf1, buf2) < 2)
 214                         break;
 215                       l1 = strlen (buf1);
 216                       l2 = strlen (buf2);
 217                       old_res_ptr = res_ptr;
 218                       if (res_size == 0)
 219                         {
 220                           res_size = l1 + 1 + l2 + 1;
 221                           res_ptr = (char *) malloc (res_size + 1);
 222                         }
 223                       else
 224                         {
 225                           res_size += l1 + 1 + l2 + 1;
 226                           res_ptr = (char *) realloc (res_ptr, res_size + 1);
 227                         }
 228                       if (res_ptr == NULL)
 229                         {
 230                           /* Out of memory. */
 231                           res_size = 0;
 232                           if (old_res_ptr != NULL)
 233                             free (old_res_ptr);
 234                           break;
 235                         }
 236                       strcpy (res_ptr + res_size - (l2 + 1) - (l1 + 1), buf1);
 237                       strcpy (res_ptr + res_size - (l2 + 1), buf2);
 238                     }
 239                   fclose (fp);
 240                   if (res_size == 0)
 241                     cp = "";
 242                   else
 243                     {
 244                       *(res_ptr + res_size) = '\0';
 245                       cp = res_ptr;
 246                     }
 247                 }
 248             }
 249
 250           free (file_name);
 251         }
 252
 253 #else
 254
 255 # if defined DARWIN7
 256       /* To avoid the trouble of installing a file that is shared by many
 257          GNU packages -- many packaging systems have problems with this --,
 258          simply inline the aliases here.  */
 259       cp = "ISO8859-1" "\0" "ISO-8859-1" "\0"
 260            "ISO8859-2" "\0" "ISO-8859-2" "\0"
 261            "ISO8859-4" "\0" "ISO-8859-4" "\0"
 262            "ISO8859-5" "\0" "ISO-8859-5" "\0"
 263            "ISO8859-7" "\0" "ISO-8859-7" "\0"
 264            "ISO8859-9" "\0" "ISO-8859-9" "\0"
 265            "ISO8859-13" "\0" "ISO-8859-13" "\0"
 266            "ISO8859-15" "\0" "ISO-8859-15" "\0"
 267            "KOI8-R" "\0" "KOI8-R" "\0"
 268            "KOI8-U" "\0" "KOI8-U" "\0"
 269            "CP866" "\0" "CP866" "\0"
 270            "CP949" "\0" "CP949" "\0"
 271            "CP1131" "\0" "CP1131" "\0"
 272            "CP1251" "\0" "CP1251" "\0"
 273            "eucCN" "\0" "GB2312" "\0"
 274            "GB2312" "\0" "GB2312" "\0"
 275            "eucJP" "\0" "EUC-JP" "\0"
 276            "eucKR" "\0" "EUC-KR" "\0"
 277            "Big5" "\0" "BIG5" "\0"
 278            "Big5HKSCS" "\0" "BIG5-HKSCS" "\0"
 279            "GBK" "\0" "GBK" "\0"
 280            "GB18030" "\0" "GB18030" "\0"
 281            "SJIS" "\0" "SHIFT_JIS" "\0"
 282            "ARMSCII-8" "\0" "ARMSCII-8" "\0"
 283            "PT154" "\0" "PT154" "\0"
 284          /*"ISCII-DEV" "\0" "?" "\0"*/
 285            "*" "\0" "UTF-8" "\0";
 286 # endif
 287
 288 # if defined VMS
 289       /* To avoid the troubles of an extra file charset.alias_vms in the
 290          sources of many GNU packages, simply inline the aliases here.  */
 291       /* The list of encodings is taken from the OpenVMS 7.3-1 documentation
 292          "Compaq C Run-Time Library Reference Manual for OpenVMS systems"
 293          section 10.7 "Handling Different Character Sets".  */
 294       cp = "ISO8859-1" "\0" "ISO-8859-1" "\0"
 295            "ISO8859-2" "\0" "ISO-8859-2" "\0"
 296            "ISO8859-5" "\0" "ISO-8859-5" "\0"
 297            "ISO8859-7" "\0" "ISO-8859-7" "\0"
 298            "ISO8859-8" "\0" "ISO-8859-8" "\0"
 299            "ISO8859-9" "\0" "ISO-8859-9" "\0"
 300            /* Japanese */
 301            "eucJP" "\0" "EUC-JP" "\0"
 302            "SJIS" "\0" "SHIFT_JIS" "\0"
 303            "DECKANJI" "\0" "DEC-KANJI" "\0"
 304            "SDECKANJI" "\0" "EUC-JP" "\0"
 305            /* Chinese */
 306            "eucTW" "\0" "EUC-TW" "\0"
 307            "DECHANYU" "\0" "DEC-HANYU" "\0"
 308            "DECHANZI" "\0" "GB2312" "\0"
 309            /* Korean */
 310            "DECKOREAN" "\0" "EUC-KR" "\0";
 311 # endif
 312
 313 # if defined WIN32_NATIVE || defined __CYGWIN__
 314       /* To avoid the troubles of installing a separate file in the same
 315          directory as the DLL and of retrieving the DLL's directory at
 316          runtime, simply inline the aliases here.  */
 317
 318       cp = "CP936" "\0" "GBK" "\0"
 319            "CP1361" "\0" "JOHAB" "\0"
 320            "CP20127" "\0" "ASCII" "\0"
 321            "CP20866" "\0" "KOI8-R" "\0"
 322            "CP20936" "\0" "GB2312" "\0"
 323            "CP21866" "\0" "KOI8-RU" "\0"
 324            "CP28591" "\0" "ISO-8859-1" "\0"
 325            "CP28592" "\0" "ISO-8859-2" "\0"
 326            "CP28593" "\0" "ISO-8859-3" "\0"
 327            "CP28594" "\0" "ISO-8859-4" "\0"
 328            "CP28595" "\0" "ISO-8859-5" "\0"
 329            "CP28596" "\0" "ISO-8859-6" "\0"
 330            "CP28597" "\0" "ISO-8859-7" "\0"
 331            "CP28598" "\0" "ISO-8859-8" "\0"
 332            "CP28599" "\0" "ISO-8859-9" "\0"
 333            "CP28605" "\0" "ISO-8859-15" "\0"
 334            "CP38598" "\0" "ISO-8859-8" "\0"
 335            "CP51932" "\0" "EUC-JP" "\0"
 336            "CP51936" "\0" "GB2312" "\0"
 337            "CP51949" "\0" "EUC-KR" "\0"
 338            "CP51950" "\0" "EUC-TW" "\0"
 339            "CP54936" "\0" "GB18030" "\0"
 340            "CP65001" "\0" "UTF-8" "\0";
 341 # endif
 342 #endif
 343
 344       charset_aliases = cp;
 345     }
 346
 347   return cp;
 348 }
 349
 350 /* Determine the current locale's character encoding, and canonicalize it
 351    into one of the canonical names listed in config.charset.
 352    The result must not be freed; it is statically allocated.
 353    If the canonical name cannot be determined, the result is a non-canonical
 354    name.  */
 355
 356 #ifdef STATIC
 357 STATIC
 358 #endif
 359 const char *
 360 locale_charset (void)
 361 {
 362   const char *codeset;
 363   const char *aliases;
 364
 365 #if !(defined WIN32_NATIVE || defined OS2)
 366
 367 # if HAVE_LANGINFO_CODESET
 368
 369   /* Most systems support nl_langinfo (CODESET) nowadays.  */
 370   codeset = nl_langinfo (CODESET);
 371
 372 #  ifdef __CYGWIN__
 373   /* Cygwin < 1.7 does not have locales.  nl_langinfo (CODESET) always
 374      returns "US-ASCII".  Return the suffix of the locale name from the
 375      environment variables (if present) or the codepage as a number.  */
 376   if (codeset != NULL && strcmp (codeset, "US-ASCII") == 0)
 377     {
 378       const char *locale;
 379       static char buf[2 + 10 + 1];
 380
 381       locale = getenv ("LC_ALL");
 382       if (locale == NULL || locale[0] == '\0')
 383         {
 384           locale = getenv ("LC_CTYPE");
 385           if (locale == NULL || locale[0] == '\0')
 386             locale = getenv ("LANG");
 387         }
 388       if (locale != NULL && locale[0] != '\0')
 389         {
 390           /* If the locale name contains an encoding after the dot, return
 391              it.  */
 392           const char *dot = strchr (locale, '.');
 393
 394           if (dot != NULL)
 395             {
 396               const char *modifier;
 397
 398               dot++;
 399               /* Look for the possible @... trailer and remove it, if any.  */
 400               modifier = strchr (dot, '@');
 401               if (modifier == NULL)
 402                 return dot;
 403               if (modifier - dot < sizeof (buf))
 404                 {
 405                   memcpy (buf, dot, modifier - dot);
 406                   buf [modifier - dot] = '\0';
 407                   return buf;
 408                 }
 409             }
 410         }
 411
 412       /* Woe32 has a function returning the locale's codepage as a number:
 413          GetACP().  This encoding is used by Cygwin, unless the user has set
 414          the environment variable CYGWIN=codepage:oem (which very few people
 415          do).
 416          Output directed to console windows needs to be converted (to
 417          GetOEMCP() if the console is using a raster font, or to
 418          GetConsoleOutputCP() if it is using a TrueType font).  Cygwin does
 419          this conversion transparently (see winsup/cygwin/fhandler_console.cc),
 420          converting to GetConsoleOutputCP().  This leads to correct results,
 421          except when SetConsoleOutputCP has been called and a raster font is
 422          in use.  */
 423       sprintf (buf, "CP%u", GetACP ());
 424       codeset = buf;
 425     }
 426 #  endif
 427
 428 # else
 429
 430   /* On old systems which lack it, use setlocale or getenv.  */
 431   const char *locale = NULL;
 432
 433   /* But most old systems don't have a complete set of locales.  Some
 434      (like SunOS 4 or DJGPP) have only the C locale.  Therefore we don't
 435      use setlocale here; it would return "C" when it doesn't support the
 436      locale name the user has set.  */
 437 #  if 0
 438   locale = setlocale (LC_CTYPE, NULL);
 439 #  endif
 440   if (locale == NULL || locale[0] == '\0')
 441     {
 442       locale = getenv ("LC_ALL");
 443       if (locale == NULL || locale[0] == '\0')
 444         {
 445           locale = getenv ("LC_CTYPE");
 446           if (locale == NULL || locale[0] == '\0')
 447             locale = getenv ("LANG");
 448         }
 449     }
 450
 451   /* On some old systems, one used to set locale = "iso8859_1". On others,
 452      you set it to "language_COUNTRY.charset". In any case, we resolve it
 453      through the charset.alias file.  */
 454   codeset = locale;
 455
 456 # endif
 457
 458 #elif defined WIN32_NATIVE
 459
 460   static char buf[2 + 10 + 1];
 461
 462   /* Woe32 has a function returning the locale's codepage as a number:
 463      GetACP().
 464      When the output goes to a console window, it needs to be provided in
 465      GetOEMCP() encoding if the console is using a raster font, or in
 466      GetConsoleOutputCP() encoding if it is using a TrueType font.
 467      But in GUI programs and for output sent to files and pipes, GetACP()
 468      encoding is the best bet.  */
 469   sprintf (buf, "CP%u", GetACP ());
 470   codeset = buf;
 471
 472 #elif defined OS2
 473
 474   const char *locale;
 475   static char buf[2 + 10 + 1];
 476   ULONG cp[3];
 477   ULONG cplen;
 478
 479   /* Allow user to override the codeset, as set in the operating system,
 480      with standard language environment variables.  */
 481   locale = getenv ("LC_ALL");
 482   if (locale == NULL || locale[0] == '\0')
 483     {
 484       locale = getenv ("LC_CTYPE");
 485       if (locale == NULL || locale[0] == '\0')
 486         locale = getenv ("LANG");
 487     }
 488   if (locale != NULL && locale[0] != '\0')
 489     {
 490       /* If the locale name contains an encoding after the dot, return it.  */
 491       const char *dot = strchr (locale, '.');
 492
 493       if (dot != NULL)
 494         {
 495           const char *modifier;
 496
 497           dot++;
 498           /* Look for the possible @... trailer and remove it, if any.  */
 499           modifier = strchr (dot, '@');
 500           if (modifier == NULL)
 501             return dot;
 502           if (modifier - dot < sizeof (buf))
 503             {
 504               memcpy (buf, dot, modifier - dot);
 505               buf [modifier - dot] = '\0';
 506               return buf;
 507             }
 508         }
 509
 510       /* Resolve through the charset.alias file.  */
 511       codeset = locale;
 512     }
 513   else
 514     {
 515       /* OS/2 has a function returning the locale's codepage as a number.  */
 516       if (DosQueryCp (sizeof (cp), cp, &cplen))
 517         codeset = "";
 518       else
 519         {
 520           sprintf (buf, "CP%u", cp[0]);
 521           codeset = buf;
 522         }
 523     }
 524
 525 #endif
 526
 527   if (codeset == NULL)
 528     /* The canonical name cannot be determined.  */
 529     codeset = "";
 530
 531   /* Resolve alias. */
 532   for (aliases = get_charset_aliases ();
 533        *aliases != '\0';
 534        aliases += strlen (aliases) + 1, aliases += strlen (aliases) + 1)
 535     if (strcmp (codeset, aliases) == 0
 536         || (aliases[0] == '*' && aliases[1] == '\0'))
 537       {
 538         codeset = aliases + strlen (aliases) + 1;
 539         break;
 540       }
 541
 542   /* Don't return an empty string.  GNU libc and GNU libiconv interpret
 543      the empty string as denoting "the locale's character encoding",
 544      thus GNU libiconv would call this function a second time.  */
 545   if (codeset[0] == '\0')
 546     codeset = "ASCII";
 547
 548   return codeset;
 549 }