intl/localcharset.c

   1 /* Determine a canonical name for the current locale's character encoding.
   2
   3    Copyright (C) 2000-2006 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 <stddef.h>
  28 #include <stdio.h>
  29 #include <string.h>
  30 #include <stdlib.h>
  31
  32 #if defined _WIN32 || defined __WIN32__
  33 # define WIN32_NATIVE
  34 #endif
  35
  36 #if defined __EMX__
  37 /* Assume EMX program runs on OS/2, even if compiled under DOS.  */
  38 # define OS2
  39 #endif
  40
  41 #if !defined WIN32_NATIVE
  42 # if HAVE_LANGINFO_CODESET
  43 #  include <langinfo.h>
  44 # else
  45 #  if 0 /* see comment below */
  46 #   include <locale.h>
  47 #  endif
  48 # endif
  49 # ifdef __CYGWIN__
  50 #  define WIN32_LEAN_AND_MEAN
  51 #  include <windows.h>
  52 # endif
  53 #elif defined WIN32_NATIVE
  54 # define WIN32_LEAN_AND_MEAN
  55 # include <windows.h>
  56 #endif
  57 #if defined OS2
  58 # define INCL_DOS
  59 # include <os2.h>
  60 #endif
  61
  62 #if ENABLE_RELOCATABLE
  63 # include "relocatable.h"
  64 #else
  65 # define relocate(pathname) (pathname)
  66 #endif
  67
  68 /* Get LIBDIR.  */
  69 #ifndef LIBDIR
  70 # include "configmake.h"
  71 #endif
  72
  73 #if defined _WIN32 || defined __WIN32__ || defined __CYGWIN__ || defined __EMX__ || defined __DJGPP__
  74   /* Win32, Cygwin, OS/2, DOS */
  75 # define ISSLASH(C) ((C) == '/' || (C) == '\\')
  76 #endif
  77
  78 #ifndef DIRECTORY_SEPARATOR
  79 # define DIRECTORY_SEPARATOR '/'
  80 #endif
  81
  82 #ifndef ISSLASH
  83 # define ISSLASH(C) ((C) == DIRECTORY_SEPARATOR)
  84 #endif
  85
  86 #if HAVE_DECL_GETC_UNLOCKED
  87 # undef getc
  88 # define getc getc_unlocked
  89 #endif
  90
  91 /* The following static variable is declared 'volatile' to avoid a
  92    possible multithread problem in the function get_charset_aliases. If we
  93    are running in a threaded environment, and if two threads initialize
  94    'charset_aliases' simultaneously, both will produce the same value,
  95    and everything will be ok if the two assignments to 'charset_aliases'
  96    are atomic. But I don't know what will happen if the two assignments mix.  */
  97 #if __STDC__ != 1
  98 # define volatile /* empty */
  99 #endif
 100 /* Pointer to the contents of the charset.alias file, if it has already been
 101    read, else NULL.  Its format is:
 102    ALIAS_1 '\0' CANONICAL_1 '\0' ... ALIAS_n '\0' CANONICAL_n '\0' '\0'  */
 103 static const char * volatile charset_aliases;
 104
 105 /* Return a pointer to the contents of the charset.alias file.  */
 106 static const char *
 107 get_charset_aliases (void)
 108 {
 109   const char *cp;
 110
 111   cp = charset_aliases;
 112   if (cp == NULL)
 113     {
 114 #if !(defined VMS || defined WIN32_NATIVE || defined __CYGWIN__)
 115       FILE *fp;
 116       const char *dir;
 117       const char *base = "charset.alias";
 118       char *file_name;
 119
 120       /* Make it possible to override the charset.alias location.  This is
 121          necessary for running the testsuite before "make install".  */
 122       dir = getenv ("CHARSETALIASDIR");
 123       if (dir == NULL || dir[0] == '\0')
 124         dir = relocate (LIBDIR);
 125
 126       /* Concatenate dir and base into freshly allocated file_name.  */
 127       {
 128         size_t dir_len = strlen (dir);
 129         size_t base_len = strlen (base);
 130         int add_slash = (dir_len > 0 && !ISSLASH (dir[dir_len - 1]));
 131         file_name = (char *) malloc (dir_len + add_slash + base_len + 1);
 132         if (file_name != NULL)
 133           {
 134             memcpy (file_name, dir, dir_len);
 135             if (add_slash)
 136               file_name[dir_len] = DIRECTORY_SEPARATOR;
 137             memcpy (file_name + dir_len + add_slash, base, base_len + 1);
 138           }
 139       }
 140
 141       if (file_name == NULL || (fp = fopen (file_name, "r")) == NULL)
 142         /* Out of memory or file not found, treat it as empty.  */
 143         cp = "";
 144       else
 145         {
 146           /* Parse the file's contents.  */
 147           char *res_ptr = NULL;
 148           size_t res_size = 0;
 149
 150           for (;;)
 151             {
 152               int c;
 153               char buf1[50+1];
 154               char buf2[50+1];
 155               size_t l1, l2;
 156               char *old_res_ptr;
 157
 158               c = getc (fp);
 159               if (c == EOF)
 160                 break;
 161               if (c == '\n' || c == ' ' || c == '\t')
 162                 continue;
 163               if (c == '#')
 164                 {
 165                   /* Skip comment, to end of line.  */
 166                   do
 167                     c = getc (fp);
 168                   while (!(c == EOF || c == '\n'));
 169                   if (c == EOF)
 170                     break;
 171                   continue;
 172                 }
 173               ungetc (c, fp);
 174               if (fscanf (fp, "%50s %50s", buf1, buf2) < 2)
 175                 break;
 176               l1 = strlen (buf1);
 177               l2 = strlen (buf2);
 178               old_res_ptr = res_ptr;
 179               if (res_size == 0)
 180                 {
 181                   res_size = l1 + 1 + l2 + 1;
 182                   res_ptr = (char *) malloc (res_size + 1);
 183                 }
 184               else
 185                 {
 186                   res_size += l1 + 1 + l2 + 1;
 187                   res_ptr = (char *) realloc (res_ptr, res_size + 1);
 188                 }
 189               if (res_ptr == NULL)
 190                 {
 191                   /* Out of memory. */
 192                   res_size = 0;
 193                   if (old_res_ptr != NULL)
 194                     free (old_res_ptr);
 195                   break;
 196                 }
 197               strcpy (res_ptr + res_size - (l2 + 1) - (l1 + 1), buf1);
 198               strcpy (res_ptr + res_size - (l2 + 1), buf2);
 199             }
 200           fclose (fp);
 201           if (res_size == 0)
 202             cp = "";
 203           else
 204             {
 205               *(res_ptr + res_size) = '\0';
 206               cp = res_ptr;
 207             }
 208         }
 209
 210       if (file_name != NULL)
 211         free (file_name);
 212
 213 #else
 214
 215 # if defined VMS
 216       /* To avoid the troubles of an extra file charset.alias_vms in the
 217          sources of many GNU packages, simply inline the aliases here.  */
 218       /* The list of encodings is taken from the OpenVMS 7.3-1 documentation
 219          "Compaq C Run-Time Library Reference Manual for OpenVMS systems"
 220          section 10.7 "Handling Different Character Sets".  */
 221       cp = "ISO8859-1" "\0" "ISO-8859-1" "\0"
 222            "ISO8859-2" "\0" "ISO-8859-2" "\0"
 223            "ISO8859-5" "\0" "ISO-8859-5" "\0"
 224            "ISO8859-7" "\0" "ISO-8859-7" "\0"
 225            "ISO8859-8" "\0" "ISO-8859-8" "\0"
 226            "ISO8859-9" "\0" "ISO-8859-9" "\0"
 227            /* Japanese */
 228            "eucJP" "\0" "EUC-JP" "\0"
 229            "SJIS" "\0" "SHIFT_JIS" "\0"
 230            "DECKANJI" "\0" "DEC-KANJI" "\0"
 231            "SDECKANJI" "\0" "EUC-JP" "\0"
 232            /* Chinese */
 233            "eucTW" "\0" "EUC-TW" "\0"
 234            "DECHANYU" "\0" "DEC-HANYU" "\0"
 235            "DECHANZI" "\0" "GB2312" "\0"
 236            /* Korean */
 237            "DECKOREAN" "\0" "EUC-KR" "\0";
 238 # endif
 239
 240 # if defined WIN32_NATIVE || defined __CYGWIN__
 241       /* To avoid the troubles of installing a separate file in the same
 242          directory as the DLL and of retrieving the DLL's directory at
 243          runtime, simply inline the aliases here.  */
 244
 245       cp = "CP936" "\0" "GBK" "\0"
 246            "CP1361" "\0" "JOHAB" "\0"
 247            "CP20127" "\0" "ASCII" "\0"
 248            "CP20866" "\0" "KOI8-R" "\0"
 249            "CP20936" "\0" "GB2312" "\0"
 250            "CP21866" "\0" "KOI8-RU" "\0"
 251            "CP28591" "\0" "ISO-8859-1" "\0"
 252            "CP28592" "\0" "ISO-8859-2" "\0"
 253            "CP28593" "\0" "ISO-8859-3" "\0"
 254            "CP28594" "\0" "ISO-8859-4" "\0"
 255            "CP28595" "\0" "ISO-8859-5" "\0"
 256            "CP28596" "\0" "ISO-8859-6" "\0"
 257            "CP28597" "\0" "ISO-8859-7" "\0"
 258            "CP28598" "\0" "ISO-8859-8" "\0"
 259            "CP28599" "\0" "ISO-8859-9" "\0"
 260            "CP28605" "\0" "ISO-8859-15" "\0"
 261            "CP38598" "\0" "ISO-8859-8" "\0"
 262            "CP51932" "\0" "EUC-JP" "\0"
 263            "CP51936" "\0" "GB2312" "\0"
 264            "CP51949" "\0" "EUC-KR" "\0"
 265            "CP51950" "\0" "EUC-TW" "\0"
 266            "CP54936" "\0" "GB18030" "\0"
 267            "CP65001" "\0" "UTF-8" "\0";
 268 # endif
 269 #endif
 270
 271       charset_aliases = cp;
 272     }
 273
 274   return cp;
 275 }
 276
 277 /* Determine the current locale's character encoding, and canonicalize it
 278    into one of the canonical names listed in config.charset.
 279    The result must not be freed; it is statically allocated.
 280    If the canonical name cannot be determined, the result is a non-canonical
 281    name.  */
 282
 283 #ifdef STATIC
 284 STATIC
 285 #endif
 286 const char *
 287 locale_charset (void)
 288 {
 289   const char *codeset;
 290   const char *aliases;
 291
 292 #if !(defined WIN32_NATIVE || defined OS2)
 293
 294 # if HAVE_LANGINFO_CODESET
 295
 296   /* Most systems support nl_langinfo (CODESET) nowadays.  */
 297   codeset = nl_langinfo (CODESET);
 298
 299 #  ifdef __CYGWIN__
 300   /* Cygwin 2006 does not have locales.  nl_langinfo (CODESET) always
 301      returns "US-ASCII".  As long as this is not fixed, return the suffix
 302      of the locale name from the environment variables (if present) or
 303      the codepage as a number.  */
 304   if (codeset != NULL && strcmp (codeset, "US-ASCII") == 0)
 305     {
 306       const char *locale;
 307       static char buf[2 + 10 + 1];
 308
 309       locale = getenv ("LC_ALL");
 310       if (locale == NULL || locale[0] == '\0')
 311         {
 312           locale = getenv ("LC_CTYPE");
 313           if (locale == NULL || locale[0] == '\0')
 314             locale = getenv ("LANG");
 315         }
 316       if (locale != NULL && locale[0] != '\0')
 317         {
 318           /* If the locale name contains an encoding after the dot, return
 319              it.  */
 320           const char *dot = strchr (locale, '.');
 321
 322           if (dot != NULL)
 323             {
 324               const char *modifier;
 325
 326               dot++;
 327               /* Look for the possible @... trailer and remove it, if any.  */
 328               modifier = strchr (dot, '@');
 329               if (modifier == NULL)
 330                 return dot;
 331               if (modifier - dot < sizeof (buf))
 332                 {
 333                   memcpy (buf, dot, modifier - dot);
 334                   buf [modifier - dot] = '\0';
 335                   return buf;
 336                 }
 337             }
 338         }
 339
 340       /* Woe32 has a function returning the locale's codepage as a number.  */
 341       sprintf (buf, "CP%u", GetACP ());
 342       codeset = buf;
 343     }
 344 #  endif
 345
 346 # else
 347
 348   /* On old systems which lack it, use setlocale or getenv.  */
 349   const char *locale = NULL;
 350
 351   /* But most old systems don't have a complete set of locales.  Some
 352      (like SunOS 4 or DJGPP) have only the C locale.  Therefore we don't
 353      use setlocale here; it would return "C" when it doesn't support the
 354      locale name the user has set.  */
 355 #  if 0
 356   locale = setlocale (LC_CTYPE, NULL);
 357 #  endif
 358   if (locale == NULL || locale[0] == '\0')
 359     {
 360       locale = getenv ("LC_ALL");
 361       if (locale == NULL || locale[0] == '\0')
 362         {
 363           locale = getenv ("LC_CTYPE");
 364           if (locale == NULL || locale[0] == '\0')
 365             locale = getenv ("LANG");
 366         }
 367     }
 368
 369   /* On some old systems, one used to set locale = "iso8859_1". On others,
 370      you set it to "language_COUNTRY.charset". In any case, we resolve it
 371      through the charset.alias file.  */
 372   codeset = locale;
 373
 374 # endif
 375
 376 #elif defined WIN32_NATIVE
 377
 378   static char buf[2 + 10 + 1];
 379
 380   /* Woe32 has a function returning the locale's codepage as a number.  */
 381   sprintf (buf, "CP%u", GetACP ());
 382   codeset = buf;
 383
 384 #elif defined OS2
 385
 386   const char *locale;
 387   static char buf[2 + 10 + 1];
 388   ULONG cp[3];
 389   ULONG cplen;
 390
 391   /* Allow user to override the codeset, as set in the operating system,
 392      with standard language environment variables.  */
 393   locale = getenv ("LC_ALL");
 394   if (locale == NULL || locale[0] == '\0')
 395     {
 396       locale = getenv ("LC_CTYPE");
 397       if (locale == NULL || locale[0] == '\0')
 398         locale = getenv ("LANG");
 399     }
 400   if (locale != NULL && locale[0] != '\0')
 401     {
 402       /* If the locale name contains an encoding after the dot, return it.  */
 403       const char *dot = strchr (locale, '.');
 404
 405       if (dot != NULL)
 406         {
 407           const char *modifier;
 408
 409           dot++;
 410           /* Look for the possible @... trailer and remove it, if any.  */
 411           modifier = strchr (dot, '@');
 412           if (modifier == NULL)
 413             return dot;
 414           if (modifier - dot < sizeof (buf))
 415             {
 416               memcpy (buf, dot, modifier - dot);
 417               buf [modifier - dot] = '\0';
 418               return buf;
 419             }
 420         }
 421
 422       /* Resolve through the charset.alias file.  */
 423       codeset = locale;
 424     }
 425   else
 426     {
 427       /* OS/2 has a function returning the locale's codepage as a number.  */
 428       if (DosQueryCp (sizeof (cp), cp, &cplen))
 429         codeset = "";
 430       else
 431         {
 432           sprintf (buf, "CP%u", cp[0]);
 433           codeset = buf;
 434         }
 435     }
 436
 437 #endif
 438
 439   if (codeset == NULL)
 440     /* The canonical name cannot be determined.  */
 441     codeset = "";
 442
 443   /* Resolve alias. */
 444   for (aliases = get_charset_aliases ();
 445        *aliases != '\0';
 446        aliases += strlen (aliases) + 1, aliases += strlen (aliases) + 1)
 447     if (strcmp (codeset, aliases) == 0
 448         || (aliases[0] == '*' && aliases[1] == '\0'))
 449       {
 450         codeset = aliases + strlen (aliases) + 1;
 451         break;
 452       }
 453
 454   /* Don't return an empty string.  GNU libc and GNU libiconv interpret
 455      the empty string as denoting "the locale's character encoding",
 456      thus GNU libiconv would call this function a second time.  */
 457   if (codeset[0] == '\0')
 458     codeset = "ASCII";
 459
 460   return codeset;
 461 }