| 1 |
22 |
khays |
/* Determine a canonical name for the current locale's character encoding.
|
| 2 |
|
|
|
| 3 |
|
|
Copyright (C) 2000-2003 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 |
|
|
#ifdef HAVE_CONFIG_H
|
| 23 |
|
|
# include <config.h>
|
| 24 |
|
|
#endif
|
| 25 |
|
|
|
| 26 |
|
|
/* Specification. */
|
| 27 |
|
|
#include "localcharset.h"
|
| 28 |
|
|
|
| 29 |
|
|
#if HAVE_STDDEF_H
|
| 30 |
|
|
# include <stddef.h>
|
| 31 |
|
|
#endif
|
| 32 |
|
|
|
| 33 |
|
|
#include <stdio.h>
|
| 34 |
|
|
#if HAVE_STRING_H
|
| 35 |
|
|
# include <string.h>
|
| 36 |
|
|
#else
|
| 37 |
|
|
# include <strings.h>
|
| 38 |
|
|
#endif
|
| 39 |
|
|
#if HAVE_STDLIB_H
|
| 40 |
|
|
# include <stdlib.h>
|
| 41 |
|
|
#endif
|
| 42 |
|
|
|
| 43 |
|
|
#if defined _WIN32 || defined __WIN32__
|
| 44 |
|
|
# undef WIN32 /* avoid warning on mingw32 */
|
| 45 |
|
|
# define WIN32
|
| 46 |
|
|
#endif
|
| 47 |
|
|
|
| 48 |
|
|
#if defined __EMX__
|
| 49 |
|
|
/* Assume EMX program runs on OS/2, even if compiled under DOS. */
|
| 50 |
|
|
# define OS2
|
| 51 |
|
|
#endif
|
| 52 |
|
|
|
| 53 |
|
|
#if !defined WIN32
|
| 54 |
|
|
# if HAVE_LANGINFO_CODESET
|
| 55 |
|
|
# include <langinfo.h>
|
| 56 |
|
|
# else
|
| 57 |
|
|
# if HAVE_SETLOCALE
|
| 58 |
|
|
# include <locale.h>
|
| 59 |
|
|
# endif
|
| 60 |
|
|
# endif
|
| 61 |
|
|
#elif defined WIN32
|
| 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 |
|
|
#if defined _WIN32 || defined __WIN32__ || defined __EMX__ || defined __DJGPP__
|
| 77 |
|
|
/* Win32, OS/2, DOS */
|
| 78 |
|
|
# define ISSLASH(C) ((C) == '/' || (C) == '\\')
|
| 79 |
|
|
#endif
|
| 80 |
|
|
|
| 81 |
|
|
#ifndef DIRECTORY_SEPARATOR
|
| 82 |
|
|
# define DIRECTORY_SEPARATOR '/'
|
| 83 |
|
|
#endif
|
| 84 |
|
|
|
| 85 |
|
|
#ifndef ISSLASH
|
| 86 |
|
|
# define ISSLASH(C) ((C) == DIRECTORY_SEPARATOR)
|
| 87 |
|
|
#endif
|
| 88 |
|
|
|
| 89 |
|
|
#ifdef HAVE_GETC_UNLOCKED
|
| 90 |
|
|
# undef getc
|
| 91 |
|
|
# define getc getc_unlocked
|
| 92 |
|
|
#endif
|
| 93 |
|
|
|
| 94 |
|
|
/* The following static variable is declared 'volatile' to avoid a
|
| 95 |
|
|
possible multithread problem in the function get_charset_aliases. If we
|
| 96 |
|
|
are running in a threaded environment, and if two threads initialize
|
| 97 |
|
|
'charset_aliases' simultaneously, both will produce the same value,
|
| 98 |
|
|
and everything will be ok if the two assignments to 'charset_aliases'
|
| 99 |
|
|
are atomic. But I don't know what will happen if the two assignments mix. */
|
| 100 |
|
|
#if __STDC__ != 1
|
| 101 |
|
|
# define volatile /* empty */
|
| 102 |
|
|
#endif
|
| 103 |
|
|
/* Pointer to the contents of the charset.alias file, if it has already been
|
| 104 |
|
|
read, else NULL. Its format is:
|
| 105 |
|
|
ALIAS_1 '\0' CANONICAL_1 '\0' ... ALIAS_n '\0' CANONICAL_n '\0' '\0' */
|
| 106 |
|
|
static const char * volatile charset_aliases;
|
| 107 |
|
|
|
| 108 |
|
|
/* Return a pointer to the contents of the charset.alias file. */
|
| 109 |
|
|
static const char *
|
| 110 |
|
|
get_charset_aliases ()
|
| 111 |
|
|
{
|
| 112 |
|
|
const char *cp;
|
| 113 |
|
|
|
| 114 |
|
|
cp = charset_aliases;
|
| 115 |
|
|
if (cp == NULL)
|
| 116 |
|
|
{
|
| 117 |
|
|
#if !(defined VMS || defined WIN32)
|
| 118 |
|
|
FILE *fp;
|
| 119 |
|
|
const char *dir = relocate (LIBDIR);
|
| 120 |
|
|
const char *base = "charset.alias";
|
| 121 |
|
|
char *file_name;
|
| 122 |
|
|
|
| 123 |
|
|
/* Concatenate dir and base into freshly allocated file_name. */
|
| 124 |
|
|
{
|
| 125 |
|
|
size_t dir_len = strlen (dir);
|
| 126 |
|
|
size_t base_len = strlen (base);
|
| 127 |
|
|
int add_slash = (dir_len > 0 && !ISSLASH (dir[dir_len - 1]));
|
| 128 |
|
|
file_name = (char *) malloc (dir_len + add_slash + base_len + 1);
|
| 129 |
|
|
if (file_name != NULL)
|
| 130 |
|
|
{
|
| 131 |
|
|
memcpy (file_name, dir, dir_len);
|
| 132 |
|
|
if (add_slash)
|
| 133 |
|
|
file_name[dir_len] = DIRECTORY_SEPARATOR;
|
| 134 |
|
|
memcpy (file_name + dir_len + add_slash, base, base_len + 1);
|
| 135 |
|
|
}
|
| 136 |
|
|
}
|
| 137 |
|
|
|
| 138 |
|
|
if (file_name == NULL || (fp = fopen (file_name, "r")) == NULL)
|
| 139 |
|
|
/* Out of memory or file not found, treat it as empty. */
|
| 140 |
|
|
cp = "";
|
| 141 |
|
|
else
|
| 142 |
|
|
{
|
| 143 |
|
|
/* Parse the file's contents. */
|
| 144 |
|
|
int c;
|
| 145 |
|
|
char buf1[50+1];
|
| 146 |
|
|
char buf2[50+1];
|
| 147 |
|
|
char *res_ptr = NULL;
|
| 148 |
|
|
size_t res_size = 0;
|
| 149 |
|
|
size_t l1, l2;
|
| 150 |
|
|
|
| 151 |
|
|
for (;;)
|
| 152 |
|
|
{
|
| 153 |
|
|
c = getc (fp);
|
| 154 |
|
|
if (c == EOF)
|
| 155 |
|
|
break;
|
| 156 |
|
|
if (c == '\n' || c == ' ' || c == '\t')
|
| 157 |
|
|
continue;
|
| 158 |
|
|
if (c == '#')
|
| 159 |
|
|
{
|
| 160 |
|
|
/* Skip comment, to end of line. */
|
| 161 |
|
|
do
|
| 162 |
|
|
c = getc (fp);
|
| 163 |
|
|
while (!(c == EOF || c == '\n'));
|
| 164 |
|
|
if (c == EOF)
|
| 165 |
|
|
break;
|
| 166 |
|
|
continue;
|
| 167 |
|
|
}
|
| 168 |
|
|
ungetc (c, fp);
|
| 169 |
|
|
if (fscanf (fp, "%50s %50s", buf1, buf2) < 2)
|
| 170 |
|
|
break;
|
| 171 |
|
|
l1 = strlen (buf1);
|
| 172 |
|
|
l2 = strlen (buf2);
|
| 173 |
|
|
if (res_size == 0)
|
| 174 |
|
|
{
|
| 175 |
|
|
res_size = l1 + 1 + l2 + 1;
|
| 176 |
|
|
res_ptr = (char *) malloc (res_size + 1);
|
| 177 |
|
|
}
|
| 178 |
|
|
else
|
| 179 |
|
|
{
|
| 180 |
|
|
res_size += l1 + 1 + l2 + 1;
|
| 181 |
|
|
res_ptr = (char *) realloc (res_ptr, res_size + 1);
|
| 182 |
|
|
}
|
| 183 |
|
|
if (res_ptr == NULL)
|
| 184 |
|
|
{
|
| 185 |
|
|
/* Out of memory. */
|
| 186 |
|
|
res_size = 0;
|
| 187 |
|
|
break;
|
| 188 |
|
|
}
|
| 189 |
|
|
strcpy (res_ptr + res_size - (l2 + 1) - (l1 + 1), buf1);
|
| 190 |
|
|
strcpy (res_ptr + res_size - (l2 + 1), buf2);
|
| 191 |
|
|
}
|
| 192 |
|
|
fclose (fp);
|
| 193 |
|
|
if (res_size == 0)
|
| 194 |
|
|
cp = "";
|
| 195 |
|
|
else
|
| 196 |
|
|
{
|
| 197 |
|
|
*(res_ptr + res_size) = '\0';
|
| 198 |
|
|
cp = res_ptr;
|
| 199 |
|
|
}
|
| 200 |
|
|
}
|
| 201 |
|
|
|
| 202 |
|
|
if (file_name != NULL)
|
| 203 |
|
|
free (file_name);
|
| 204 |
|
|
|
| 205 |
|
|
#else
|
| 206 |
|
|
|
| 207 |
|
|
# if defined VMS
|
| 208 |
|
|
/* To avoid the troubles of an extra file charset.alias_vms in the
|
| 209 |
|
|
sources of many GNU packages, simply inline the aliases here. */
|
| 210 |
|
|
/* The list of encodings is taken from the OpenVMS 7.3-1 documentation
|
| 211 |
|
|
"Compaq C Run-Time Library Reference Manual for OpenVMS systems"
|
| 212 |
|
|
section 10.7 "Handling Different Character Sets". */
|
| 213 |
|
|
cp = "ISO8859-1" "\0" "ISO-8859-1" "\0"
|
| 214 |
|
|
"ISO8859-2" "\0" "ISO-8859-2" "\0"
|
| 215 |
|
|
"ISO8859-5" "\0" "ISO-8859-5" "\0"
|
| 216 |
|
|
"ISO8859-7" "\0" "ISO-8859-7" "\0"
|
| 217 |
|
|
"ISO8859-8" "\0" "ISO-8859-8" "\0"
|
| 218 |
|
|
"ISO8859-9" "\0" "ISO-8859-9" "\0"
|
| 219 |
|
|
/* Japanese */
|
| 220 |
|
|
"eucJP" "\0" "EUC-JP" "\0"
|
| 221 |
|
|
"SJIS" "\0" "SHIFT_JIS" "\0"
|
| 222 |
|
|
"DECKANJI" "\0" "DEC-KANJI" "\0"
|
| 223 |
|
|
"SDECKANJI" "\0" "EUC-JP" "\0"
|
| 224 |
|
|
/* Chinese */
|
| 225 |
|
|
"eucTW" "\0" "EUC-TW" "\0"
|
| 226 |
|
|
"DECHANYU" "\0" "DEC-HANYU" "\0"
|
| 227 |
|
|
"DECHANZI" "\0" "GB2312" "\0"
|
| 228 |
|
|
/* Korean */
|
| 229 |
|
|
"DECKOREAN" "\0" "EUC-KR" "\0";
|
| 230 |
|
|
# endif
|
| 231 |
|
|
|
| 232 |
|
|
# if defined WIN32
|
| 233 |
|
|
/* To avoid the troubles of installing a separate file in the same
|
| 234 |
|
|
directory as the DLL and of retrieving the DLL's directory at
|
| 235 |
|
|
runtime, simply inline the aliases here. */
|
| 236 |
|
|
|
| 237 |
|
|
cp = "CP936" "\0" "GBK" "\0"
|
| 238 |
|
|
"CP1361" "\0" "JOHAB" "\0"
|
| 239 |
|
|
"CP20127" "\0" "ASCII" "\0"
|
| 240 |
|
|
"CP20866" "\0" "KOI8-R" "\0"
|
| 241 |
|
|
"CP21866" "\0" "KOI8-RU" "\0"
|
| 242 |
|
|
"CP28591" "\0" "ISO-8859-1" "\0"
|
| 243 |
|
|
"CP28592" "\0" "ISO-8859-2" "\0"
|
| 244 |
|
|
"CP28593" "\0" "ISO-8859-3" "\0"
|
| 245 |
|
|
"CP28594" "\0" "ISO-8859-4" "\0"
|
| 246 |
|
|
"CP28595" "\0" "ISO-8859-5" "\0"
|
| 247 |
|
|
"CP28596" "\0" "ISO-8859-6" "\0"
|
| 248 |
|
|
"CP28597" "\0" "ISO-8859-7" "\0"
|
| 249 |
|
|
"CP28598" "\0" "ISO-8859-8" "\0"
|
| 250 |
|
|
"CP28599" "\0" "ISO-8859-9" "\0"
|
| 251 |
|
|
"CP28605" "\0" "ISO-8859-15" "\0";
|
| 252 |
|
|
# endif
|
| 253 |
|
|
#endif
|
| 254 |
|
|
|
| 255 |
|
|
charset_aliases = cp;
|
| 256 |
|
|
}
|
| 257 |
|
|
|
| 258 |
|
|
return cp;
|
| 259 |
|
|
}
|
| 260 |
|
|
|
| 261 |
|
|
/* Determine the current locale's character encoding, and canonicalize it
|
| 262 |
|
|
into one of the canonical names listed in config.charset.
|
| 263 |
|
|
The result must not be freed; it is statically allocated.
|
| 264 |
|
|
If the canonical name cannot be determined, the result is a non-canonical
|
| 265 |
|
|
name. */
|
| 266 |
|
|
|
| 267 |
|
|
#ifdef STATIC
|
| 268 |
|
|
STATIC
|
| 269 |
|
|
#endif
|
| 270 |
|
|
const char *
|
| 271 |
|
|
locale_charset ()
|
| 272 |
|
|
{
|
| 273 |
|
|
const char *codeset;
|
| 274 |
|
|
const char *aliases;
|
| 275 |
|
|
|
| 276 |
|
|
#if !(defined WIN32 || defined OS2)
|
| 277 |
|
|
|
| 278 |
|
|
# if HAVE_LANGINFO_CODESET
|
| 279 |
|
|
|
| 280 |
|
|
/* Most systems support nl_langinfo (CODESET) nowadays. */
|
| 281 |
|
|
codeset = nl_langinfo (CODESET);
|
| 282 |
|
|
|
| 283 |
|
|
# else
|
| 284 |
|
|
|
| 285 |
|
|
/* On old systems which lack it, use setlocale or getenv. */
|
| 286 |
|
|
const char *locale = NULL;
|
| 287 |
|
|
|
| 288 |
|
|
/* But most old systems don't have a complete set of locales. Some
|
| 289 |
|
|
(like SunOS 4 or DJGPP) have only the C locale. Therefore we don't
|
| 290 |
|
|
use setlocale here; it would return "C" when it doesn't support the
|
| 291 |
|
|
locale name the user has set. */
|
| 292 |
|
|
# if HAVE_SETLOCALE && 0
|
| 293 |
|
|
locale = setlocale (LC_CTYPE, NULL);
|
| 294 |
|
|
# endif
|
| 295 |
|
|
if (locale == NULL || locale[0] == '\0')
|
| 296 |
|
|
{
|
| 297 |
|
|
locale = getenv ("LC_ALL");
|
| 298 |
|
|
if (locale == NULL || locale[0] == '\0')
|
| 299 |
|
|
{
|
| 300 |
|
|
locale = getenv ("LC_CTYPE");
|
| 301 |
|
|
if (locale == NULL || locale[0] == '\0')
|
| 302 |
|
|
locale = getenv ("LANG");
|
| 303 |
|
|
}
|
| 304 |
|
|
}
|
| 305 |
|
|
|
| 306 |
|
|
/* On some old systems, one used to set locale = "iso8859_1". On others,
|
| 307 |
|
|
you set it to "language_COUNTRY.charset". In any case, we resolve it
|
| 308 |
|
|
through the charset.alias file. */
|
| 309 |
|
|
codeset = locale;
|
| 310 |
|
|
|
| 311 |
|
|
# endif
|
| 312 |
|
|
|
| 313 |
|
|
#elif defined WIN32
|
| 314 |
|
|
|
| 315 |
|
|
static char buf[2 + 10 + 1];
|
| 316 |
|
|
|
| 317 |
|
|
/* Woe32 has a function returning the locale's codepage as a number. */
|
| 318 |
|
|
sprintf (buf, "CP%u", GetACP ());
|
| 319 |
|
|
codeset = buf;
|
| 320 |
|
|
|
| 321 |
|
|
#elif defined OS2
|
| 322 |
|
|
|
| 323 |
|
|
const char *locale;
|
| 324 |
|
|
static char buf[2 + 10 + 1];
|
| 325 |
|
|
ULONG cp[3];
|
| 326 |
|
|
ULONG cplen;
|
| 327 |
|
|
|
| 328 |
|
|
/* Allow user to override the codeset, as set in the operating system,
|
| 329 |
|
|
with standard language environment variables. */
|
| 330 |
|
|
locale = getenv ("LC_ALL");
|
| 331 |
|
|
if (locale == NULL || locale[0] == '\0')
|
| 332 |
|
|
{
|
| 333 |
|
|
locale = getenv ("LC_CTYPE");
|
| 334 |
|
|
if (locale == NULL || locale[0] == '\0')
|
| 335 |
|
|
locale = getenv ("LANG");
|
| 336 |
|
|
}
|
| 337 |
|
|
if (locale != NULL && locale[0] != '\0')
|
| 338 |
|
|
{
|
| 339 |
|
|
/* If the locale name contains an encoding after the dot, return it. */
|
| 340 |
|
|
const char *dot = strchr (locale, '.');
|
| 341 |
|
|
|
| 342 |
|
|
if (dot != NULL)
|
| 343 |
|
|
{
|
| 344 |
|
|
const char *modifier;
|
| 345 |
|
|
|
| 346 |
|
|
dot++;
|
| 347 |
|
|
/* Look for the possible @... trailer and remove it, if any. */
|
| 348 |
|
|
modifier = strchr (dot, '@');
|
| 349 |
|
|
if (modifier == NULL)
|
| 350 |
|
|
return dot;
|
| 351 |
|
|
if (modifier - dot < sizeof (buf))
|
| 352 |
|
|
{
|
| 353 |
|
|
memcpy (buf, dot, modifier - dot);
|
| 354 |
|
|
buf [modifier - dot] = '\0';
|
| 355 |
|
|
return buf;
|
| 356 |
|
|
}
|
| 357 |
|
|
}
|
| 358 |
|
|
|
| 359 |
|
|
/* Resolve through the charset.alias file. */
|
| 360 |
|
|
codeset = locale;
|
| 361 |
|
|
}
|
| 362 |
|
|
else
|
| 363 |
|
|
{
|
| 364 |
|
|
/* OS/2 has a function returning the locale's codepage as a number. */
|
| 365 |
|
|
if (DosQueryCp (sizeof (cp), cp, &cplen))
|
| 366 |
|
|
codeset = "";
|
| 367 |
|
|
else
|
| 368 |
|
|
{
|
| 369 |
|
|
sprintf (buf, "CP%u", cp[0]);
|
| 370 |
|
|
codeset = buf;
|
| 371 |
|
|
}
|
| 372 |
|
|
}
|
| 373 |
|
|
|
| 374 |
|
|
#endif
|
| 375 |
|
|
|
| 376 |
|
|
if (codeset == NULL)
|
| 377 |
|
|
/* The canonical name cannot be determined. */
|
| 378 |
|
|
codeset = "";
|
| 379 |
|
|
|
| 380 |
|
|
/* Resolve alias. */
|
| 381 |
|
|
for (aliases = get_charset_aliases ();
|
| 382 |
|
|
*aliases != '\0';
|
| 383 |
|
|
aliases += strlen (aliases) + 1, aliases += strlen (aliases) + 1)
|
| 384 |
|
|
if (strcmp (codeset, aliases) == 0
|
| 385 |
|
|
|| (aliases[0] == '*' && aliases[1] == '\0'))
|
| 386 |
|
|
{
|
| 387 |
|
|
codeset = aliases + strlen (aliases) + 1;
|
| 388 |
|
|
break;
|
| 389 |
|
|
}
|
| 390 |
|
|
|
| 391 |
|
|
/* Don't return an empty string. GNU libc and GNU libiconv interpret
|
| 392 |
|
|
the empty string as denoting "the locale's character encoding",
|
| 393 |
|
|
thus GNU libiconv would call this function a second time. */
|
| 394 |
|
|
if (codeset[0] == '\0')
|
| 395 |
|
|
codeset = "ASCII";
|
| 396 |
|
|
|
| 397 |
|
|
return codeset;
|
| 398 |
|
|
}
|