intl/localename.c

   1 /* Determine the current selected locale.
   2    Copyright (C) 1995-1999, 2000-2002 Free Software Foundation, Inc.
   3
   4    This program is free software; you can redistribute it and/or modify it
   5    under the terms of the GNU Library General Public License as published
   6    by the Free Software Foundation; either version 2, or (at your option)
   7    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 GNU
  12    Library General Public License for more details.
  13
  14    You should have received a copy of the GNU Library General Public
  15    License along with this program; if not, write to the Free Software
  16    Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA 02110-1301,
  17    USA.  */
  18
  19 /* Written by Ulrich Drepper <drepper@gnu.org>, 1995.  */
  20 /* Win32 code written by Tor Lillqvist <tml@iki.fi>.  */
  21
  22 #ifdef HAVE_CONFIG_H
  23 # include <config.h>
  24 #endif
  25
  26 #include <stdlib.h>
  27 #include <locale.h>
  28
  29 #if defined _WIN32 || defined __WIN32__
  30 # undef WIN32   /* avoid warning on mingw32 */
  31 # define WIN32
  32 #endif
  33
  34 #ifdef WIN32
  35 # define WIN32_LEAN_AND_MEAN
  36 # include <windows.h>
  37 /* Mingw headers don't have latest language and sublanguage codes.  */
  38 # ifndef LANG_AFRIKAANS
  39 # define LANG_AFRIKAANS 0x36
  40 # endif
  41 # ifndef LANG_ALBANIAN
  42 # define LANG_ALBANIAN 0x1c
  43 # endif
  44 # ifndef LANG_ARABIC
  45 # define LANG_ARABIC 0x01
  46 # endif
  47 # ifndef LANG_ARMENIAN
  48 # define LANG_ARMENIAN 0x2b
  49 # endif
  50 # ifndef LANG_ASSAMESE
  51 # define LANG_ASSAMESE 0x4d
  52 # endif
  53 # ifndef LANG_AZERI
  54 # define LANG_AZERI 0x2c
  55 # endif
  56 # ifndef LANG_BASQUE
  57 # define LANG_BASQUE 0x2d
  58 # endif
  59 # ifndef LANG_BELARUSIAN
  60 # define LANG_BELARUSIAN 0x23
  61 # endif
  62 # ifndef LANG_BENGALI
  63 # define LANG_BENGALI 0x45
  64 # endif
  65 # ifndef LANG_CATALAN
  66 # define LANG_CATALAN 0x03
  67 # endif
  68 # ifndef LANG_DIVEHI
  69 # define LANG_DIVEHI 0x65
  70 # endif
  71 # ifndef LANG_ESTONIAN
  72 # define LANG_ESTONIAN 0x25
  73 # endif
  74 # ifndef LANG_FAEROESE
  75 # define LANG_FAEROESE 0x38
  76 # endif
  77 # ifndef LANG_FARSI
  78 # define LANG_FARSI 0x29
  79 # endif
  80 # ifndef LANG_GALICIAN
  81 # define LANG_GALICIAN 0x56
  82 # endif
  83 # ifndef LANG_GEORGIAN
  84 # define LANG_GEORGIAN 0x37
  85 # endif
  86 # ifndef LANG_GUJARATI
  87 # define LANG_GUJARATI 0x47
  88 # endif
  89 # ifndef LANG_HEBREW
  90 # define LANG_HEBREW 0x0d
  91 # endif
  92 # ifndef LANG_HINDI
  93 # define LANG_HINDI 0x39
  94 # endif
  95 # ifndef LANG_INDONESIAN
  96 # define LANG_INDONESIAN 0x21
  97 # endif
  98 # ifndef LANG_KANNADA
  99 # define LANG_KANNADA 0x4b
 100 # endif
 101 # ifndef LANG_KASHMIRI
 102 # define LANG_KASHMIRI 0x60
 103 # endif
 104 # ifndef LANG_KAZAK
 105 # define LANG_KAZAK 0x3f
 106 # endif
 107 # ifndef LANG_KONKANI
 108 # define LANG_KONKANI 0x57
 109 # endif
 110 # ifndef LANG_KYRGYZ
 111 # define LANG_KYRGYZ 0x40
 112 # endif
 113 # ifndef LANG_LATVIAN
 114 # define LANG_LATVIAN 0x26
 115 # endif
 116 # ifndef LANG_LITHUANIAN
 117 # define LANG_LITHUANIAN 0x27
 118 # endif
 119 # ifndef LANG_MACEDONIAN
 120 # define LANG_MACEDONIAN 0x2f
 121 # endif
 122 # ifndef LANG_MALAY
 123 # define LANG_MALAY 0x3e
 124 # endif
 125 # ifndef LANG_MALAYALAM
 126 # define LANG_MALAYALAM 0x4c
 127 # endif
 128 # ifndef LANG_MANIPURI
 129 # define LANG_MANIPURI 0x58
 130 # endif
 131 # ifndef LANG_MARATHI
 132 # define LANG_MARATHI 0x4e
 133 # endif
 134 # ifndef LANG_MONGOLIAN
 135 # define LANG_MONGOLIAN 0x50
 136 # endif
 137 # ifndef LANG_NEPALI
 138 # define LANG_NEPALI 0x61
 139 # endif
 140 # ifndef LANG_ORIYA
 141 # define LANG_ORIYA 0x48
 142 # endif
 143 # ifndef LANG_PUNJABI
 144 # define LANG_PUNJABI 0x46
 145 # endif
 146 # ifndef LANG_SANSKRIT
 147 # define LANG_SANSKRIT 0x4f
 148 # endif
 149 # ifndef LANG_SERBIAN
 150 # define LANG_SERBIAN 0x1a
 151 # endif
 152 # ifndef LANG_SINDHI
 153 # define LANG_SINDHI 0x59
 154 # endif
 155 # ifndef LANG_SLOVAK
 156 # define LANG_SLOVAK 0x1b
 157 # endif
 158 # ifndef LANG_SORBIAN
 159 # define LANG_SORBIAN 0x2e
 160 # endif
 161 # ifndef LANG_SWAHILI
 162 # define LANG_SWAHILI 0x41
 163 # endif
 164 # ifndef LANG_SYRIAC
 165 # define LANG_SYRIAC 0x5a
 166 # endif
 167 # ifndef LANG_TAMIL
 168 # define LANG_TAMIL 0x49
 169 # endif
 170 # ifndef LANG_TATAR
 171 # define LANG_TATAR 0x44
 172 # endif
 173 # ifndef LANG_TELUGU
 174 # define LANG_TELUGU 0x4a
 175 # endif
 176 # ifndef LANG_THAI
 177 # define LANG_THAI 0x1e
 178 # endif
 179 # ifndef LANG_UKRAINIAN
 180 # define LANG_UKRAINIAN 0x22
 181 # endif
 182 # ifndef LANG_URDU
 183 # define LANG_URDU 0x20
 184 # endif
 185 # ifndef LANG_UZBEK
 186 # define LANG_UZBEK 0x43
 187 # endif
 188 # ifndef LANG_VIETNAMESE
 189 # define LANG_VIETNAMESE 0x2a
 190 # endif
 191 # ifndef SUBLANG_ARABIC_SAUDI_ARABIA
 192 # define SUBLANG_ARABIC_SAUDI_ARABIA 0x01
 193 # endif
 194 # ifndef SUBLANG_ARABIC_IRAQ
 195 # define SUBLANG_ARABIC_IRAQ 0x02
 196 # endif
 197 # ifndef SUBLANG_ARABIC_EGYPT
 198 # define SUBLANG_ARABIC_EGYPT 0x03
 199 # endif
 200 # ifndef SUBLANG_ARABIC_LIBYA
 201 # define SUBLANG_ARABIC_LIBYA 0x04
 202 # endif
 203 # ifndef SUBLANG_ARABIC_ALGERIA
 204 # define SUBLANG_ARABIC_ALGERIA 0x05
 205 # endif
 206 # ifndef SUBLANG_ARABIC_MOROCCO
 207 # define SUBLANG_ARABIC_MOROCCO 0x06
 208 # endif
 209 # ifndef SUBLANG_ARABIC_TUNISIA
 210 # define SUBLANG_ARABIC_TUNISIA 0x07
 211 # endif
 212 # ifndef SUBLANG_ARABIC_OMAN
 213 # define SUBLANG_ARABIC_OMAN 0x08
 214 # endif
 215 # ifndef SUBLANG_ARABIC_YEMEN
 216 # define SUBLANG_ARABIC_YEMEN 0x09
 217 # endif
 218 # ifndef SUBLANG_ARABIC_SYRIA
 219 # define SUBLANG_ARABIC_SYRIA 0x0a
 220 # endif
 221 # ifndef SUBLANG_ARABIC_JORDAN
 222 # define SUBLANG_ARABIC_JORDAN 0x0b
 223 # endif
 224 # ifndef SUBLANG_ARABIC_LEBANON
 225 # define SUBLANG_ARABIC_LEBANON 0x0c
 226 # endif
 227 # ifndef SUBLANG_ARABIC_KUWAIT
 228 # define SUBLANG_ARABIC_KUWAIT 0x0d
 229 # endif
 230 # ifndef SUBLANG_ARABIC_UAE
 231 # define SUBLANG_ARABIC_UAE 0x0e
 232 # endif
 233 # ifndef SUBLANG_ARABIC_BAHRAIN
 234 # define SUBLANG_ARABIC_BAHRAIN 0x0f
 235 # endif
 236 # ifndef SUBLANG_ARABIC_QATAR
 237 # define SUBLANG_ARABIC_QATAR 0x10
 238 # endif
 239 # ifndef SUBLANG_AZERI_LATIN
 240 # define SUBLANG_AZERI_LATIN 0x01
 241 # endif
 242 # ifndef SUBLANG_AZERI_CYRILLIC
 243 # define SUBLANG_AZERI_CYRILLIC 0x02
 244 # endif
 245 # ifndef SUBLANG_CHINESE_MACAU
 246 # define SUBLANG_CHINESE_MACAU 0x05
 247 # endif
 248 # ifndef SUBLANG_ENGLISH_SOUTH_AFRICA
 249 # define SUBLANG_ENGLISH_SOUTH_AFRICA 0x07
 250 # endif
 251 # ifndef SUBLANG_ENGLISH_JAMAICA
 252 # define SUBLANG_ENGLISH_JAMAICA 0x08
 253 # endif
 254 # ifndef SUBLANG_ENGLISH_CARIBBEAN
 255 # define SUBLANG_ENGLISH_CARIBBEAN 0x09
 256 # endif
 257 # ifndef SUBLANG_ENGLISH_BELIZE
 258 # define SUBLANG_ENGLISH_BELIZE 0x0a
 259 # endif
 260 # ifndef SUBLANG_ENGLISH_TRINIDAD
 261 # define SUBLANG_ENGLISH_TRINIDAD 0x0b
 262 # endif
 263 # ifndef SUBLANG_ENGLISH_ZIMBABWE
 264 # define SUBLANG_ENGLISH_ZIMBABWE 0x0c
 265 # endif
 266 # ifndef SUBLANG_ENGLISH_PHILIPPINES
 267 # define SUBLANG_ENGLISH_PHILIPPINES 0x0d
 268 # endif
 269 # ifndef SUBLANG_FRENCH_LUXEMBOURG
 270 # define SUBLANG_FRENCH_LUXEMBOURG 0x05
 271 # endif
 272 # ifndef SUBLANG_FRENCH_MONACO
 273 # define SUBLANG_FRENCH_MONACO 0x06
 274 # endif
 275 # ifndef SUBLANG_GERMAN_LUXEMBOURG
 276 # define SUBLANG_GERMAN_LUXEMBOURG 0x04
 277 # endif
 278 # ifndef SUBLANG_GERMAN_LIECHTENSTEIN
 279 # define SUBLANG_GERMAN_LIECHTENSTEIN 0x05
 280 # endif
 281 # ifndef SUBLANG_KASHMIRI_INDIA
 282 # define SUBLANG_KASHMIRI_INDIA 0x02
 283 # endif
 284 # ifndef SUBLANG_MALAY_MALAYSIA
 285 # define SUBLANG_MALAY_MALAYSIA 0x01
 286 # endif
 287 # ifndef SUBLANG_MALAY_BRUNEI_DARUSSALAM
 288 # define SUBLANG_MALAY_BRUNEI_DARUSSALAM 0x02
 289 # endif
 290 # ifndef SUBLANG_NEPALI_INDIA
 291 # define SUBLANG_NEPALI_INDIA 0x02
 292 # endif
 293 # ifndef SUBLANG_SERBIAN_LATIN
 294 # define SUBLANG_SERBIAN_LATIN 0x02
 295 # endif
 296 # ifndef SUBLANG_SERBIAN_CYRILLIC
 297 # define SUBLANG_SERBIAN_CYRILLIC 0x03
 298 # endif
 299 # ifndef SUBLANG_SPANISH_GUATEMALA
 300 # define SUBLANG_SPANISH_GUATEMALA 0x04
 301 # endif
 302 # ifndef SUBLANG_SPANISH_COSTA_RICA
 303 # define SUBLANG_SPANISH_COSTA_RICA 0x05
 304 # endif
 305 # ifndef SUBLANG_SPANISH_PANAMA
 306 # define SUBLANG_SPANISH_PANAMA 0x06
 307 # endif
 308 # ifndef SUBLANG_SPANISH_DOMINICAN_REPUBLIC
 309 # define SUBLANG_SPANISH_DOMINICAN_REPUBLIC 0x07
 310 # endif
 311 # ifndef SUBLANG_SPANISH_VENEZUELA
 312 # define SUBLANG_SPANISH_VENEZUELA 0x08
 313 # endif
 314 # ifndef SUBLANG_SPANISH_COLOMBIA
 315 # define SUBLANG_SPANISH_COLOMBIA 0x09
 316 # endif
 317 # ifndef SUBLANG_SPANISH_PERU
 318 # define SUBLANG_SPANISH_PERU 0x0a
 319 # endif
 320 # ifndef SUBLANG_SPANISH_ARGENTINA
 321 # define SUBLANG_SPANISH_ARGENTINA 0x0b
 322 # endif
 323 # ifndef SUBLANG_SPANISH_ECUADOR
 324 # define SUBLANG_SPANISH_ECUADOR 0x0c
 325 # endif
 326 # ifndef SUBLANG_SPANISH_CHILE
 327 # define SUBLANG_SPANISH_CHILE 0x0d
 328 # endif
 329 # ifndef SUBLANG_SPANISH_URUGUAY
 330 # define SUBLANG_SPANISH_URUGUAY 0x0e
 331 # endif
 332 # ifndef SUBLANG_SPANISH_PARAGUAY
 333 # define SUBLANG_SPANISH_PARAGUAY 0x0f
 334 # endif
 335 # ifndef SUBLANG_SPANISH_BOLIVIA
 336 # define SUBLANG_SPANISH_BOLIVIA 0x10
 337 # endif
 338 # ifndef SUBLANG_SPANISH_EL_SALVADOR
 339 # define SUBLANG_SPANISH_EL_SALVADOR 0x11
 340 # endif
 341 # ifndef SUBLANG_SPANISH_HONDURAS
 342 # define SUBLANG_SPANISH_HONDURAS 0x12
 343 # endif
 344 # ifndef SUBLANG_SPANISH_NICARAGUA
 345 # define SUBLANG_SPANISH_NICARAGUA 0x13
 346 # endif
 347 # ifndef SUBLANG_SPANISH_PUERTO_RICO
 348 # define SUBLANG_SPANISH_PUERTO_RICO 0x14
 349 # endif
 350 # ifndef SUBLANG_SWEDISH_FINLAND
 351 # define SUBLANG_SWEDISH_FINLAND 0x02
 352 # endif
 353 # ifndef SUBLANG_URDU_PAKISTAN
 354 # define SUBLANG_URDU_PAKISTAN 0x01
 355 # endif
 356 # ifndef SUBLANG_URDU_INDIA
 357 # define SUBLANG_URDU_INDIA 0x02
 358 # endif
 359 # ifndef SUBLANG_UZBEK_LATIN
 360 # define SUBLANG_UZBEK_LATIN 0x01
 361 # endif
 362 # ifndef SUBLANG_UZBEK_CYRILLIC
 363 # define SUBLANG_UZBEK_CYRILLIC 0x02
 364 # endif
 365 #endif
 366
 367 /* XPG3 defines the result of 'setlocale (category, NULL)' as:
 368    "Directs 'setlocale()' to query 'category' and return the current
 369     setting of 'local'."
 370    However it does not specify the exact format.  Neither do SUSV2 and
 371    ISO C 99.  So we can use this feature only on selected systems (e.g.
 372    those using GNU C Library).  */
 373 #if defined _LIBC || (defined __GNU_LIBRARY__ && __GNU_LIBRARY__ >= 2)
 374 # define HAVE_LOCALE_NULL
 375 #endif
 376
 377 /* Determine the current locale's name, and canonicalize it into XPG syntax
 378      language[_territory[.codeset]][@modifier]
 379    The codeset part in the result is not reliable; the locale_charset()
 380    should be used for codeset information instead.
 381    The result must not be freed; it is statically allocated.  */
 382
 383 const char *
 384 _nl_locale_name (category, categoryname)
 385      int category;
 386      const char *categoryname;
 387 {
 388   const char *retval;
 389
 390 #ifndef WIN32
 391
 392   /* Use the POSIX methods of looking to 'LC_ALL', 'LC_xxx', and 'LANG'.
 393      On some systems this can be done by the 'setlocale' function itself.  */
 394 # if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 395   retval = setlocale (category, NULL);
 396 # else
 397   /* Setting of LC_ALL overwrites all other.  */
 398   retval = getenv ("LC_ALL");
 399   if (retval == NULL || retval[0] == '\0')
 400     {
 401       /* Next comes the name of the desired category.  */
 402       retval = getenv (categoryname);
 403       if (retval == NULL || retval[0] == '\0')
 404         {
 405           /* Last possibility is the LANG environment variable.  */
 406           retval = getenv ("LANG");
 407           if (retval == NULL || retval[0] == '\0')
 408             /* We use C as the default domain.  POSIX says this is
 409                implementation defined.  */
 410             retval = "C";
 411         }
 412     }
 413 # endif
 414
 415   return retval;
 416
 417 #else /* WIN32 */
 418
 419   /* Return an XPG style locale name language[_territory][@modifier].
 420      Don't even bother determining the codeset; it's not useful in this
 421      context, because message catalogs are not specific to a single
 422      codeset.  */
 423
 424   LCID lcid;
 425   LANGID langid;
 426   int primary, sub;
 427
 428   /* Let the user override the system settings through environment
 429      variables, as on POSIX systems.  */
 430   retval = getenv ("LC_ALL");
 431   if (retval != NULL && retval[0] != '\0')
 432     return retval;
 433   retval = getenv (categoryname);
 434   if (retval != NULL && retval[0] != '\0')
 435     return retval;
 436   retval = getenv ("LANG");
 437   if (retval != NULL && retval[0] != '\0')
 438     return retval;
 439
 440   /* Use native Win32 API locale ID.  */
 441   lcid = GetThreadLocale ();
 442
 443   /* Strip off the sorting rules, keep only the language part.  */
 444   langid = LANGIDFROMLCID (lcid);
 445
 446   /* Split into language and territory part.  */
 447   primary = PRIMARYLANGID (langid);
 448   sub = SUBLANGID (langid);
 449
 450   /* Dispatch on language.
 451      See also http://www.unicode.org/unicode/onlinedat/languages.html .
 452      For details about languages, see http://www.ethnologue.com/ .  */
 453   switch (primary)
 454     {
 455     case LANG_AFRIKAANS: return "af_ZA";
 456     case LANG_ALBANIAN: return "sq_AL";
 457     case 0x5e: /* AMHARIC */ return "am_ET";
 458     case LANG_ARABIC:
 459       switch (sub)
 460         {
 461         case SUBLANG_ARABIC_SAUDI_ARABIA: return "ar_SA";
 462         case SUBLANG_ARABIC_IRAQ: return "ar_IQ";
 463         case SUBLANG_ARABIC_EGYPT: return "ar_EG";
 464         case SUBLANG_ARABIC_LIBYA: return "ar_LY";
 465         case SUBLANG_ARABIC_ALGERIA: return "ar_DZ";
 466         case SUBLANG_ARABIC_MOROCCO: return "ar_MA";
 467         case SUBLANG_ARABIC_TUNISIA: return "ar_TN";
 468         case SUBLANG_ARABIC_OMAN: return "ar_OM";
 469         case SUBLANG_ARABIC_YEMEN: return "ar_YE";
 470         case SUBLANG_ARABIC_SYRIA: return "ar_SY";
 471         case SUBLANG_ARABIC_JORDAN: return "ar_JO";
 472         case SUBLANG_ARABIC_LEBANON: return "ar_LB";
 473         case SUBLANG_ARABIC_KUWAIT: return "ar_KW";
 474         case SUBLANG_ARABIC_UAE: return "ar_AE";
 475         case SUBLANG_ARABIC_BAHRAIN: return "ar_BH";
 476         case SUBLANG_ARABIC_QATAR: return "ar_QA";
 477         }
 478       return "ar";
 479     case LANG_ARMENIAN: return "hy_AM";
 480     case LANG_ASSAMESE: return "as_IN";
 481     case LANG_AZERI:
 482       switch (sub)
 483         {
 484         /* FIXME: Adjust this when Azerbaijani locales appear on Unix.  */
 485         case SUBLANG_AZERI_LATIN: return "az_AZ@latin";
 486         case SUBLANG_AZERI_CYRILLIC: return "az_AZ@cyrillic";
 487         }
 488       return "az";
 489     case LANG_BASQUE:
 490       return "eu"; /* Ambiguous: could be "eu_ES" or "eu_FR".  */
 491     case LANG_BELARUSIAN: return "be_BY";
 492     case LANG_BENGALI: return "bn_IN";
 493     case LANG_BULGARIAN: return "bg_BG";
 494     case 0x55: /* BURMESE */ return "my_MM";
 495     case 0x53: /* CAMBODIAN */ return "km_KH";
 496     case LANG_CATALAN: return "ca_ES";
 497     case 0x5c: /* CHEROKEE */ return "chr_US";
 498     case LANG_CHINESE:
 499       switch (sub)
 500         {
 501         case SUBLANG_CHINESE_TRADITIONAL: return "zh_TW";
 502         case SUBLANG_CHINESE_SIMPLIFIED: return "zh_CN";
 503         case SUBLANG_CHINESE_HONGKONG: return "zh_HK";
 504         case SUBLANG_CHINESE_SINGAPORE: return "zh_SG";
 505         case SUBLANG_CHINESE_MACAU: return "zh_MO";
 506         }
 507       return "zh";
 508     case LANG_CROATIAN:         /* LANG_CROATIAN == LANG_SERBIAN
 509                                  * What used to be called Serbo-Croatian
 510                                  * should really now be two separate
 511                                  * languages because of political reasons.
 512                                  * (Says tml, who knows nothing about Serbian
 513                                  * or Croatian.)
 514                                  * (I can feel those flames coming already.)
 515                                  */
 516       switch (sub)
 517         {
 518         case SUBLANG_DEFAULT: return "hr_HR";
 519         case SUBLANG_SERBIAN_LATIN: return "sr_YU";
 520         case SUBLANG_SERBIAN_CYRILLIC: return "sr_YU@cyrillic";
 521         }
 522       return "hr";
 523     case LANG_CZECH: return "cs_CZ";
 524     case LANG_DANISH: return "da_DK";
 525     case LANG_DIVEHI: return "div_MV";
 526     case LANG_DUTCH:
 527       switch (sub)
 528         {
 529         case SUBLANG_DUTCH: return "nl_NL";
 530         case SUBLANG_DUTCH_BELGIAN: /* FLEMISH, VLAAMS */ return "nl_BE";
 531         }
 532       return "nl";
 533     case 0x66: /* EDO */ return "bin_NG";
 534     case LANG_ENGLISH:
 535       switch (sub)
 536         {
 537         /* SUBLANG_ENGLISH_US == SUBLANG_DEFAULT. Heh. I thought
 538          * English was the language spoken in England.
 539          * Oh well.
 540          */
 541         case SUBLANG_ENGLISH_US: return "en_US";
 542         case SUBLANG_ENGLISH_UK: return "en_GB";
 543         case SUBLANG_ENGLISH_AUS: return "en_AU";
 544         case SUBLANG_ENGLISH_CAN: return "en_CA";
 545         case SUBLANG_ENGLISH_NZ: return "en_NZ";
 546         case SUBLANG_ENGLISH_EIRE: return "en_IE";
 547         case SUBLANG_ENGLISH_SOUTH_AFRICA: return "en_ZA";
 548         case SUBLANG_ENGLISH_JAMAICA: return "en_JM";
 549         case SUBLANG_ENGLISH_CARIBBEAN: return "en_GD"; /* Grenada? */
 550         case SUBLANG_ENGLISH_BELIZE: return "en_BZ";
 551         case SUBLANG_ENGLISH_TRINIDAD: return "en_TT";
 552         case SUBLANG_ENGLISH_ZIMBABWE: return "en_ZW";
 553         case SUBLANG_ENGLISH_PHILIPPINES: return "en_PH";
 554         }
 555       return "en";
 556     case LANG_ESTONIAN: return "et_EE";
 557     case LANG_FAEROESE: return "fo_FO";
 558     case LANG_FARSI: return "fa_IR";
 559     case LANG_FINNISH: return "fi_FI";
 560     case LANG_FRENCH:
 561       switch (sub)
 562         {
 563         case SUBLANG_FRENCH: return "fr_FR";
 564         case SUBLANG_FRENCH_BELGIAN: /* WALLOON */ return "fr_BE";
 565         case SUBLANG_FRENCH_CANADIAN: return "fr_CA";
 566         case SUBLANG_FRENCH_SWISS: return "fr_CH";
 567         case SUBLANG_FRENCH_LUXEMBOURG: return "fr_LU";
 568         case SUBLANG_FRENCH_MONACO: return "fr_MC";
 569         }
 570       return "fr";
 571     case 0x62: /* FRISIAN */ return "fy_NL";
 572     case 0x67: /* FULFULDE */ return "ful_NG";
 573     case 0x3c: /* GAELIC */
 574       switch (sub)
 575         {
 576         case 0x01: /* SCOTTISH */ return "gd_GB";
 577         case 0x02: /* IRISH */ return "ga_IE";
 578         }
 579       return "C";
 580     case LANG_GALICIAN: return "gl_ES";
 581     case LANG_GEORGIAN: return "ka_GE";
 582     case LANG_GERMAN:
 583       switch (sub)
 584         {
 585         case SUBLANG_GERMAN: return "de_DE";
 586         case SUBLANG_GERMAN_SWISS: return "de_CH";
 587         case SUBLANG_GERMAN_AUSTRIAN: return "de_AT";
 588         case SUBLANG_GERMAN_LUXEMBOURG: return "de_LU";
 589         case SUBLANG_GERMAN_LIECHTENSTEIN: return "de_LI";
 590         }
 591       return "de";
 592     case LANG_GREEK: return "el_GR";
 593     case 0x74: /* GUARANI */ return "gn_PY";
 594     case LANG_GUJARATI: return "gu_IN";
 595     case 0x68: /* HAUSA */ return "ha_NG";
 596     case 0x75: /* HAWAIIAN */
 597       /* FIXME: Do they mean Hawaiian ("haw_US", 1000 speakers)
 598          or Hawaii Creole English ("cpe_US", 600000 speakers)?  */
 599       return "cpe_US";
 600     case LANG_HEBREW: return "he_IL";
 601     case LANG_HINDI: return "hi_IN";
 602     case LANG_HUNGARIAN: return "hu_HU";
 603     case 0x69: /* IBIBIO */ return "nic_NG";
 604     case LANG_ICELANDIC: return "is_IS";
 605     case 0x70: /* IGBO */ return "ibo_NG";
 606     case LANG_INDONESIAN: return "id_ID";
 607     case 0x5d: /* INUKTITUT */ return "iu_CA";
 608     case LANG_ITALIAN:
 609       switch (sub)
 610         {
 611         case SUBLANG_ITALIAN: return "it_IT";
 612         case SUBLANG_ITALIAN_SWISS: return "it_CH";
 613         }
 614       return "it";
 615     case LANG_JAPANESE: return "ja_JP";
 616     case LANG_KANNADA: return "kn_IN";
 617     case 0x71: /* KANURI */ return "kau_NG";
 618     case LANG_KASHMIRI:
 619       switch (sub)
 620         {
 621         case SUBLANG_DEFAULT: return "ks_PK";
 622         case SUBLANG_KASHMIRI_INDIA: return "ks_IN";
 623         }
 624       return "ks";
 625     case LANG_KAZAK: return "kk_KZ";
 626     case LANG_KONKANI:
 627       /* FIXME: Adjust this when such locales appear on Unix.  */
 628       return "kok_IN";
 629     case LANG_KOREAN: return "ko_KR";
 630     case LANG_KYRGYZ: return "ky_KG";
 631     case 0x54: /* LAO */ return "lo_LA";
 632     case 0x76: /* LATIN */ return "la_VA";
 633     case LANG_LATVIAN: return "lv_LV";
 634     case LANG_LITHUANIAN: return "lt_LT";
 635     case LANG_MACEDONIAN: return "mk_MK";
 636     case LANG_MALAY:
 637       switch (sub)
 638         {
 639         case SUBLANG_MALAY_MALAYSIA: return "ms_MY";
 640         case SUBLANG_MALAY_BRUNEI_DARUSSALAM: return "ms_BN";
 641         }
 642       return "ms";
 643     case LANG_MALAYALAM: return "ml_IN";
 644     case 0x3a: /* MALTESE */ return "mt_MT";
 645     case LANG_MANIPURI:
 646       /* FIXME: Adjust this when such locales appear on Unix.  */
 647       return "mni_IN";
 648     case LANG_MARATHI: return "mr_IN";
 649     case LANG_MONGOLIAN:
 650       return "mn"; /* Ambiguous: could be "mn_CN" or "mn_MN".  */
 651     case LANG_NEPALI:
 652       switch (sub)
 653         {
 654         case SUBLANG_DEFAULT: return "ne_NP";
 655         case SUBLANG_NEPALI_INDIA: return "ne_IN";
 656         }
 657       return "ne";
 658     case LANG_NORWEGIAN:
 659       switch (sub)
 660         {
 661         case SUBLANG_NORWEGIAN_BOKMAL: return "no_NO";
 662         case SUBLANG_NORWEGIAN_NYNORSK: return "nn_NO";
 663         }
 664       return "no";
 665     case LANG_ORIYA: return "or_IN";
 666     case 0x72: /* OROMO */ return "om_ET";
 667     case 0x79: /* PAPIAMENTU */ return "pap_AN";
 668     case 0x63: /* PASHTO */
 669       return "ps"; /* Ambiguous: could be "ps_PK" or "ps_AF".  */
 670     case LANG_POLISH: return "pl_PL";
 671     case LANG_PORTUGUESE:
 672       switch (sub)
 673         {
 674         case SUBLANG_PORTUGUESE: return "pt_PT";
 675         /* Hmm. SUBLANG_PORTUGUESE_BRAZILIAN == SUBLANG_DEFAULT.
 676            Same phenomenon as SUBLANG_ENGLISH_US == SUBLANG_DEFAULT. */
 677         case SUBLANG_PORTUGUESE_BRAZILIAN: return "pt_BR";
 678         }
 679       return "pt";
 680     case LANG_PUNJABI: return "pa_IN";
 681     case 0x17: /* RHAETO-ROMANCE */ return "rm_CH";
 682     case LANG_ROMANIAN: return "ro_RO";
 683     case LANG_RUSSIAN:
 684       return "ru"; /* Ambiguous: could be "ru_RU" or "ru_UA".  */
 685     case 0x3b: /* SAMI */ return "se_NO";
 686     case LANG_SANSKRIT: return "sa_IN";
 687     case LANG_SINDHI: return "sd";
 688     case 0x5b: /* SINHALESE */ return "si_LK";
 689     case LANG_SLOVAK: return "sk_SK";
 690     case LANG_SLOVENIAN: return "sl_SI";
 691     case 0x77: /* SOMALI */ return "so_SO";
 692     case LANG_SORBIAN:
 693       /* FIXME: Adjust this when such locales appear on Unix.  */
 694       return "wen_DE";
 695     case LANG_SPANISH:
 696       switch (sub)
 697         {
 698         case SUBLANG_SPANISH: return "es_ES";
 699         case SUBLANG_SPANISH_MEXICAN: return "es_MX";
 700         case SUBLANG_SPANISH_MODERN:
 701           return "es_ES@modern";        /* not seen on Unix */
 702         case SUBLANG_SPANISH_GUATEMALA: return "es_GT";
 703         case SUBLANG_SPANISH_COSTA_RICA: return "es_CR";
 704         case SUBLANG_SPANISH_PANAMA: return "es_PA";
 705         case SUBLANG_SPANISH_DOMINICAN_REPUBLIC: return "es_DO";
 706         case SUBLANG_SPANISH_VENEZUELA: return "es_VE";
 707         case SUBLANG_SPANISH_COLOMBIA: return "es_CO";
 708         case SUBLANG_SPANISH_PERU: return "es_PE";
 709         case SUBLANG_SPANISH_ARGENTINA: return "es_AR";
 710         case SUBLANG_SPANISH_ECUADOR: return "es_EC";
 711         case SUBLANG_SPANISH_CHILE: return "es_CL";
 712         case SUBLANG_SPANISH_URUGUAY: return "es_UY";
 713         case SUBLANG_SPANISH_PARAGUAY: return "es_PY";
 714         case SUBLANG_SPANISH_BOLIVIA: return "es_BO";
 715         case SUBLANG_SPANISH_EL_SALVADOR: return "es_SV";
 716         case SUBLANG_SPANISH_HONDURAS: return "es_HN";
 717         case SUBLANG_SPANISH_NICARAGUA: return "es_NI";
 718         case SUBLANG_SPANISH_PUERTO_RICO: return "es_PR";
 719         }
 720       return "es";
 721     case 0x30: /* SUTU */ return "bnt_TZ";
 722     case LANG_SWAHILI: return "sw_KE";
 723     case LANG_SWEDISH:
 724       switch (sub)
 725         {
 726         case SUBLANG_DEFAULT: return "sv_SE";
 727         case SUBLANG_SWEDISH_FINLAND: return "sv_FI";
 728         }
 729       return "sv";
 730     case LANG_SYRIAC: return "syr_TR"; /* An extinct language.  */
 731     case 0x64: /* TAGALOG */ return "tl_PH";
 732     case 0x28: /* TAJIK */ return "tg_TJ";
 733     case 0x5f: /* TAMAZIGHT */ return "ber_MA";
 734     case LANG_TAMIL:
 735       return "ta"; /* Ambiguous: could be "ta_IN" or "ta_LK" or "ta_SG".  */
 736     case LANG_TATAR: return "tt_RU";
 737     case LANG_TELUGU: return "te_IN";
 738     case LANG_THAI: return "th_TH";
 739     case 0x51: /* TIBETAN */ return "bo_CN";
 740     case 0x73: /* TIGRINYA */ return "ti_ET";
 741     case 0x31: /* TSONGA */ return "ts_ZA";
 742     case LANG_TURKISH: return "tr_TR";
 743     case 0x42: /* TURKMEN */ return "tk_TM";
 744     case LANG_UKRAINIAN: return "uk_UA";
 745     case LANG_URDU:
 746       switch (sub)
 747         {
 748         case SUBLANG_URDU_PAKISTAN: return "ur_PK";
 749         case SUBLANG_URDU_INDIA: return "ur_IN";
 750         }
 751       return "ur";
 752     case LANG_UZBEK:
 753       switch (sub)
 754         {
 755         /* FIXME: Adjust this when Uzbek locales appear on Unix.  */
 756         case SUBLANG_UZBEK_LATIN: return "uz_UZ@latin";
 757         case SUBLANG_UZBEK_CYRILLIC: return "uz_UZ@cyrillic";
 758         }
 759       return "uz";
 760     case 0x33: /* VENDA */ return "ven_ZA";
 761     case LANG_VIETNAMESE: return "vi_VN";
 762     case 0x52: /* WELSH */ return "cy_GB";
 763     case 0x34: /* XHOSA */ return "xh_ZA";
 764     case 0x78: /* YI */ return "sit_CN";
 765     case 0x3d: /* YIDDISH */ return "yi_IL";
 766     case 0x6a: /* YORUBA */ return "yo_NG";
 767     case 0x35: /* ZULU */ return "zu_ZA";
 768     default: return "C";
 769     }
 770
 771 #endif
 772 }