gl/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
   6    it under the terms of the GNU General Public License as published by
   7    the Free Software Foundation; either version 3, 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
  13    GNU General Public License for more details.
  14
  15    You should have received a copy of the GNU General Public License along
  16    with this program; if not, write to the Free Software Foundation,
  17    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
  18
  19 /* Written by Bruno Haible <bruno@clisp.org>.  */
  20
  21 #include <config.h>
  22
  23 /* Specification.  */
  24 #include "localcharset.h"
  25
  26 #include <stddef.h>
  27 #include <stdio.h>
  28 #include <string.h>
  29 #include <stdlib.h>
  30
  31 #if defined _WIN32 || defined __WIN32__
  32 # define WIN32_NATIVE
  33 #endif
  34
  35 #if defined __EMX__
  36 /* Assume EMX program runs on OS/2, even if compiled under DOS.  */
  37 # define OS2
  38 #endif
  39
  40 #if !defined WIN32_NATIVE
  41 # if HAVE_LANGINFO_CODESET
  42 #  include <langinfo.h>
  43 # else
  44 #  if 0 /* see comment below */
  45 #   include <locale.h>
  46 #  endif
  47 # endif
  48 # ifdef __CYGWIN__
  49 #  define WIN32_LEAN_AND_MEAN
  50 #  include <windows.h>
  51 # endif
  52 #elif defined WIN32_NATIVE
  53 # define WIN32_LEAN_AND_MEAN
  54 # include <windows.h>
  55 #endif
  56 #if defined OS2
  57 # define INCL_DOS
  58 # include <os2.h>
  59 #endif
  60
  61 #if ENABLE_RELOCATABLE
  62 # include "relocatable.h"
  63 #else
  64 # define relocate(pathname) (pathname)
  65 #endif
  66
  67 /* Get LIBDIR.  */
  68 #ifndef LIBDIR
  69 # include "configmake.h"
  70 #endif
  71
  72 #if defined _WIN32 || defined __WIN32__ || defined __CYGWIN__ || defined __EMX__ || defined __DJGPP__
  73   /* Win32, Cygwin, OS/2, DOS */
  74 # define ISSLASH(C) ((C) == '/' || (C) == '\\')
  75 #endif
  76
  77 #ifndef DIRECTORY_SEPARATOR
  78 # define DIRECTORY_SEPARATOR '/'
  79 #endif
  80
  81 #ifndef ISSLASH
  82 # define ISSLASH(C) ((C) == DIRECTORY_SEPARATOR)
  83 #endif
  84
  85 #if HAVE_DECL_GETC_UNLOCKED
  86 # undef getc
  87 # define getc getc_unlocked
  88 #endif
  89
  90 /* The following static variable is declared 'volatile' to avoid a
  91    possible multithread problem in the function get_charset_aliases. If we
  92    are running in a threaded environment, and if two threads initialize
  93    'charset_aliases' simultaneously, both will produce the same value,
  94    and everything will be ok if the two assignments to 'charset_aliases'
  95    are atomic. But I don't know what will happen if the two assignments mix.  */
  96 #if __STDC__ != 1
  97 # define volatile /* empty */
  98 #endif
  99 /* Pointer to the contents of the charset.alias file, if it has already been
 100    read, else NULL.  Its format is:
 101    ALIAS_1 '\0' CANONICAL_1 '\0' ... ALIAS_n '\0' CANONICAL_n '\0' '\0'  */
 102 static const char * volatile charset_aliases;
 103
 104 /* Return a pointer to the contents of the charset.alias file.  */
 105 static const char *
 106 get_charset_aliases (void)
 107 {
 108   const char *cp;
 109
 110   cp = charset_aliases;
 111   if (cp == NULL)
 112     {
 113 #if !(defined VMS || defined WIN32_NATIVE || defined __CYGWIN__)
 114       FILE *fp;
 115       const char *dir;
 116       const char *base = "charset.alias";
 117       char *file_name;
 118
 119       /* Make it possible to override the charset.alias location.  This is
 120          necessary for running the testsuite before "make install".  */
 121       dir = getenv ("CHARSETALIASDIR");
 122       if (dir == NULL || dir[0] == '\0')
 123         dir = relocate (LIBDIR);
 124
 125       /* Concatenate dir and base into freshly allocated file_name.  */
 126       {
 127         size_t dir_len = strlen (dir);
 128         size_t base_len = strlen (base);
 129         int add_slash = (dir_len > 0 && !ISSLASH (dir[dir_len - 1]));
 130         file_name = (char *) malloc (dir_len + add_slash + base_len + 1);
 131         if (file_name != NULL)
 132           {
 133             memcpy (file_name, dir, dir_len);
 134             if (add_slash)
 135               file_name[dir_len] = DIRECTORY_SEPARATOR;
 136             memcpy (file_name + dir_len + add_slash, base, base_len + 1);
 137           }
 138       }
 139
 140       if (file_name == NULL || (fp = fopen (file_name, "r")) == NULL)
 141         /* Out of memory or file not found, treat it as empty.  */
 142         cp = "";
 143       else
 144         {
 145           /* Parse the file's contents.  */
 146           char *res_ptr = NULL;
 147           size_t res_size = 0;
 148
 149           for (;;)
 150             {
 151               int c;
 152               char buf1[50+1];
 153               char buf2[50+1];
 154               size_t l1, l2;
 155               char *old_res_ptr;
 156
 157               c = getc (fp);
 158               if (c == EOF)
 159                 break;
 160               if (c == '\n' || c == ' ' || c == '\t')
 161                 continue;
 162               if (c == '#')
 163                 {
 164                   /* Skip comment, to end of line.  */
 165                   do
 166                     c = getc (fp);
 167                   while (!(c == EOF || c == '\n'));
 168                   if (c == EOF)
 169                     break;
 170                   continue;
 171                 }
 172               ungetc (c, fp);
 173               if (fscanf (fp, "%50s %50s", buf1, buf2) < 2)
 174                 break;
 175               l1 = strlen (buf1);
 176               l2 = strlen (buf2);
 177               old_res_ptr = res_ptr;
 178               if (res_size == 0)
 179                 {
 180                   res_size = l1 + 1 + l2 + 1;
 181                   res_ptr = (char *) malloc (res_size + 1);
 182                 }
 183               else
 184                 {
 185                   res_size += l1 + 1 + l2 + 1;
 186                   res_ptr = (char *) realloc (res_ptr, res_size + 1);
 187                 }
 188               if (res_ptr == NULL)
 189                 {
 190                   /* Out of memory. */
 191                   res_size = 0;
 192                   if (old_res_ptr != NULL)
 193                     free (old_res_ptr);
 194                   break;
 195                 }
 196               strcpy (res_ptr + res_size - (l2 + 1) - (l1 + 1), buf1);
 197               strcpy (res_ptr + res_size - (l2 + 1), buf2);
 198             }
 199           fclose (fp);
 200           if (res_size == 0)
 201             cp = "";
 202           else
 203             {
 204               *(res_ptr + res_size) = '\0';
 205               cp = res_ptr;
 206             }
 207         }
 208
 209       if (file_name != NULL)
 210         free (file_name);
 211
 212 #else
 213
 214 # if defined VMS
 215       /* To avoid the troubles of an extra file charset.alias_vms in the
 216          sources of many GNU packages, simply inline the aliases here.  */
 217       /* The list of encodings is taken from the OpenVMS 7.3-1 documentation
 218          "Compaq C Run-Time Library Reference Manual for OpenVMS systems"
 219          section 10.7 "Handling Different Character Sets".  */
 220       cp = "ISO8859-1" "\0" "ISO-8859-1" "\0"
 221            "ISO8859-2" "\0" "ISO-8859-2" "\0"
 222            "ISO8859-5" "\0" "ISO-8859-5" "\0"
 223            "ISO8859-7" "\0" "ISO-8859-7" "\0"
 224            "ISO8859-8" "\0" "ISO-8859-8" "\0"
 225            "ISO8859-9" "\0" "ISO-8859-9" "\0"
 226            /* Japanese */
 227            "eucJP" "\0" "EUC-JP" "\0"
 228            "SJIS" "\0" "SHIFT_JIS" "\0"
 229            "DECKANJI" "\0" "DEC-KANJI" "\0"
 230            "SDECKANJI" "\0" "EUC-JP" "\0"
 231            /* Chinese */
 232            "eucTW" "\0" "EUC-TW" "\0"
 233            "DECHANYU" "\0" "DEC-HANYU" "\0"
 234            "DECHANZI" "\0" "GB2312" "\0"
 235            /* Korean */
 236            "DECKOREAN" "\0" "EUC-KR" "\0";
 237 # endif
 238
 239 # if defined WIN32_NATIVE || defined __CYGWIN__
 240       /* To avoid the troubles of installing a separate file in the same
 241          directory as the DLL and of retrieving the DLL's directory at
 242          runtime, simply inline the aliases here.  */
 243
 244       cp = "CP936" "\0" "GBK" "\0"
 245            "CP1361" "\0" "JOHAB" "\0"
 246            "CP20127" "\0" "ASCII" "\0"
 247            "CP20866" "\0" "KOI8-R" "\0"
 248            "CP20936" "\0" "GB2312" "\0"
 249            "CP21866" "\0" "KOI8-RU" "\0"
 250            "CP28591" "\0" "ISO-8859-1" "\0"
 251            "CP28592" "\0" "ISO-8859-2" "\0"
 252            "CP28593" "\0" "ISO-8859-3" "\0"
 253            "CP28594" "\0" "ISO-8859-4" "\0"
 254            "CP28595" "\0" "ISO-8859-5" "\0"
 255            "CP28596" "\0" "ISO-8859-6" "\0"
 256            "CP28597" "\0" "ISO-8859-7" "\0"
 257            "CP28598" "\0" "ISO-8859-8" "\0"
 258            "CP28599" "\0" "ISO-8859-9" "\0"
 259            "CP28605" "\0" "ISO-8859-15" "\0"
 260            "CP38598" "\0" "ISO-8859-8" "\0"
 261            "CP51932" "\0" "EUC-JP" "\0"
 262            "CP51936" "\0" "GB2312" "\0"
 263            "CP51949" "\0" "EUC-KR" "\0"
 264            "CP51950" "\0" "EUC-TW" "\0"
 265            "CP54936" "\0" "GB18030" "\0"
 266            "CP65001" "\0" "UTF-8" "\0";
 267 # endif
 268 #endif
 269
 270       charset_aliases = cp;
 271     }
 272
 273   return cp;
 274 }
 275
 276 /* Determine the current locale's character encoding, and canonicalize it
 277    into one of the canonical names listed in config.charset.
 278    The result must not be freed; it is statically allocated.
 279    If the canonical name cannot be determined, the result is a non-canonical
 280    name.  */
 281
 282 #ifdef STATIC
 283 STATIC
 284 #endif
 285 const char *
 286 locale_charset (void)
 287 {
 288   const char *codeset;
 289   const char *aliases;
 290
 291 #if !(defined WIN32_NATIVE || defined OS2)
 292
 293 # if HAVE_LANGINFO_CODESET
 294
 295   /* Most systems support nl_langinfo (CODESET) nowadays.  */
 296   codeset = nl_langinfo (CODESET);
 297
 298 #  ifdef __CYGWIN__
 299   /* Cygwin 2006 does not have locales.  nl_langinfo (CODESET) always
 300      returns "US-ASCII".  As long as this is not fixed, return the suffix
 301      of the locale name from the environment variables (if present) or
 302      the codepage as a number.  */
 303   if (codeset != NULL && strcmp (codeset, "US-ASCII") == 0)
 304     {
 305       const char *locale;
 306       static char buf[2 + 10 + 1];
 307
 308       locale = getenv ("LC_ALL");
 309       if (locale == NULL || locale[0] == '\0')
 310         {
 311           locale = getenv ("LC_CTYPE");
 312           if (locale == NULL || locale[0] == '\0')
 313             locale = getenv ("LANG");
 314         }
 315       if (locale != NULL && locale[0] != '\0')
 316         {
 317           /* If the locale name contains an encoding after the dot, return
 318              it.  */
 319           const char *dot = strchr (locale, '.');
 320
 321           if (dot != NULL)
 322             {
 323               const char *modifier;
 324
 325               dot++;
 326               /* Look for the possible @... trailer and remove it, if any.  */
 327               modifier = strchr (dot, '@');
 328               if (modifier == NULL)
 329                 return dot;
 330               if (modifier - dot < sizeof (buf))
 331                 {
 332                   memcpy (buf, dot, modifier - dot);
 333                   buf [modifier - dot] = '\0';
 334                   return buf;
 335                 }
 336             }
 337         }
 338
 339       /* Woe32 has a function returning the locale's codepage as a number.  */
 340       sprintf (buf, "CP%u", GetACP ());
 341       codeset = buf;
 342     }
 343 #  endif
 344
 345 # else
 346
 347   /* On old systems which lack it, use setlocale or getenv.  */
 348   const char *locale = NULL;
 349
 350   /* But most old systems don't have a complete set of locales.  Some
 351      (like SunOS 4 or DJGPP) have only the C locale.  Therefore we don't
 352      use setlocale here; it would return "C" when it doesn't support the
 353      locale name the user has set.  */
 354 #  if 0
 355   locale = setlocale (LC_CTYPE, NULL);
 356 #  endif
 357   if (locale == NULL || locale[0] == '\0')
 358     {
 359       locale = getenv ("LC_ALL");
 360       if (locale == NULL || locale[0] == '\0')
 361         {
 362           locale = getenv ("LC_CTYPE");
 363           if (locale == NULL || locale[0] == '\0')
 364             locale = getenv ("LANG");
 365         }
 366     }
 367
 368   /* On some old systems, one used to set locale = "iso8859_1". On others,
 369      you set it to "language_COUNTRY.charset". In any case, we resolve it
 370      through the charset.alias file.  */
 371   codeset = locale;
 372
 373 # endif
 374
 375 #elif defined WIN32_NATIVE
 376
 377   static char buf[2 + 10 + 1];
 378
 379   /* Woe32 has a function returning the locale's codepage as a number.  */
 380   sprintf (buf, "CP%u", GetACP ());
 381   codeset = buf;
 382
 383 #elif defined OS2
 384
 385   const char *locale;
 386   static char buf[2 + 10 + 1];
 387   ULONG cp[3];
 388   ULONG cplen;
 389
 390   /* Allow user to override the codeset, as set in the operating system,
 391      with standard language environment variables.  */
 392   locale = getenv ("LC_ALL");
 393   if (locale == NULL || locale[0] == '\0')
 394     {
 395       locale = getenv ("LC_CTYPE");
 396       if (locale == NULL || locale[0] == '\0')
 397         locale = getenv ("LANG");
 398     }
 399   if (locale != NULL && locale[0] != '\0')
 400     {
 401       /* If the locale name contains an encoding after the dot, return it.  */
 402       const char *dot = strchr (locale, '.');
 403
 404       if (dot != NULL)
 405         {
 406           const char *modifier;
 407
 408           dot++;
 409           /* Look for the possible @... trailer and remove it, if any.  */
 410           modifier = strchr (dot, '@');
 411           if (modifier == NULL)
 412             return dot;
 413           if (modifier - dot < sizeof (buf))
 414             {
 415               memcpy (buf, dot, modifier - dot);
 416               buf [modifier - dot] = '\0';
 417               return buf;
 418             }
 419         }
 420
 421       /* Resolve through the charset.alias file.  */
 422       codeset = locale;
 423     }
 424   else
 425     {
 426       /* OS/2 has a function returning the locale's codepage as a number.  */
 427       if (DosQueryCp (sizeof (cp), cp, &cplen))
 428         codeset = "";
 429       else
 430         {
 431           sprintf (buf, "CP%u", cp[0]);
 432           codeset = buf;
 433         }
 434     }
 435
 436 #endif
 437
 438   if (codeset == NULL)
 439     /* The canonical name cannot be determined.  */
 440     codeset = "";
 441
 442   /* Resolve alias. */
 443   for (aliases = get_charset_aliases ();
 444        *aliases != '\0';
 445        aliases += strlen (aliases) + 1, aliases += strlen (aliases) + 1)
 446     if (strcmp (codeset, aliases) == 0
 447         || (aliases[0] == '*' && aliases[1] == '\0'))
 448       {
 449         codeset = aliases + strlen (aliases) + 1;
 450         break;
 451       }
 452
 453   /* Don't return an empty string.  GNU libc and GNU libiconv interpret
 454      the empty string as denoting "the locale's character encoding",
 455      thus GNU libiconv would call this function a second time.  */
 456   if (codeset[0] == '\0')
 457     codeset = "ASCII";
 458
 459   return codeset;
 460 }