jnlib/w32-gettext.c

   1 /* w32-gettext.h - A simple gettext implementation for Windows targets.
   2    Copyright (C) 1995, 1996, 1997, 1999, 2005, 2007,
   3                  2008 Free Software Foundation, Inc.
   4
   5    This program is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public License
   7    as published by the Free Software Foundation; either version 2.1 of
   8    the License, or (at your option) any later version.
   9
  10    This program is distributed in the hope that it will be useful, but
  11    WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with this program; if not, see <http://www.gnu.org/licenses/>.
  17  */
  18
  19 #if HAVE_CONFIG_H
  20 #include <config.h>
  21 #endif
  22 #if !defined (_WIN32) && !defined (__CYGWIN32__)
  23 #  error This module may only be build for Windows or Cygwin32
  24 #endif
  25
  26 #include <stdlib.h>
  27 #include <stdio.h>
  28 #include <string.h>
  29 #include <errno.h>
  30 #include <ctype.h>
  31 #include <sys/types.h>
  32 #include <sys/stat.h>
  33 #include <stdint.h>
  34 #include <locale.h>
  35 #include <windows.h>
  36
  37 #ifdef JNLIB_IN_JNLIB
  38 #include "libjnlib-config.h"
  39 #endif
  40
  41 #ifndef jnlib_malloc
  42 # define jnlib_malloc(a)    malloc ((a))
  43 # define jnlib_calloc(a,b)  calloc ((a), (b))
  44 # define jnlib_free(a)      free ((a))
  45 # define jnlib_xstrdup(a)   my_xstrdup(a)
  46 #endif /*!jnlib_malloc*/
  47
  48
  49 \f
  50 /* localname.c from gettext BEGIN.  */
  51
  52 /* Determine the current selected locale.
  53    Copyright (C) 1995-1999, 2000-2003 Free Software Foundation, Inc.
  54
  55    This program is free software; you can redistribute it and/or modify it
  56    under the terms of the GNU Library General Public License as published
  57    by the Free Software Foundation; either version 2, or (at your option)
  58    any later version.
  59
  60    This program is distributed in the hope that it will be useful,
  61    but WITHOUT ANY WARRANTY; without even the implied warranty of
  62    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  63    Library General Public License for more details.
  64
  65    You should have received a copy of the GNU Library General Public
  66    License along with this program; if not, write to the Free Software
  67    Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
  68    USA.  */
  69
  70 /* Written by Ulrich Drepper <drepper@gnu.org>, 1995.  */
  71 /* Win32 code written by Tor Lillqvist <tml@iki.fi>.  */
  72 /* Renamed _nl_locale_name, removed unsed args, removed include files,
  73    non-W32 code and changed comments <wk@gnupg.org>.  */
  74
  75 /* Mingw headers don't have latest language and sublanguage codes.  */
  76 #ifndef LANG_AFRIKAANS
  77 #define LANG_AFRIKAANS 0x36
  78 #endif
  79 #ifndef LANG_ALBANIAN
  80 #define LANG_ALBANIAN 0x1c
  81 #endif
  82 #ifndef LANG_AMHARIC
  83 #define LANG_AMHARIC 0x5e
  84 #endif
  85 #ifndef LANG_ARABIC
  86 #define LANG_ARABIC 0x01
  87 #endif
  88 #ifndef LANG_ARMENIAN
  89 #define LANG_ARMENIAN 0x2b
  90 #endif
  91 #ifndef LANG_ASSAMESE
  92 #define LANG_ASSAMESE 0x4d
  93 #endif
  94 #ifndef LANG_AZERI
  95 #define LANG_AZERI 0x2c
  96 #endif
  97 #ifndef LANG_BASQUE
  98 #define LANG_BASQUE 0x2d
  99 #endif
 100 #ifndef LANG_BELARUSIAN
 101 #define LANG_BELARUSIAN 0x23
 102 #endif
 103 #ifndef LANG_BENGALI
 104 #define LANG_BENGALI 0x45
 105 #endif
 106 #ifndef LANG_BURMESE
 107 #define LANG_BURMESE 0x55
 108 #endif
 109 #ifndef LANG_CAMBODIAN
 110 #define LANG_CAMBODIAN 0x53
 111 #endif
 112 #ifndef LANG_CATALAN
 113 #define LANG_CATALAN 0x03
 114 #endif
 115 #ifndef LANG_CHEROKEE
 116 #define LANG_CHEROKEE 0x5c
 117 #endif
 118 #ifndef LANG_DIVEHI
 119 #define LANG_DIVEHI 0x65
 120 #endif
 121 #ifndef LANG_EDO
 122 #define LANG_EDO 0x66
 123 #endif
 124 #ifndef LANG_ESTONIAN
 125 #define LANG_ESTONIAN 0x25
 126 #endif
 127 #ifndef LANG_FAEROESE
 128 #define LANG_FAEROESE 0x38
 129 #endif
 130 #ifndef LANG_FARSI
 131 #define LANG_FARSI 0x29
 132 #endif
 133 #ifndef LANG_FRISIAN
 134 #define LANG_FRISIAN 0x62
 135 #endif
 136 #ifndef LANG_FULFULDE
 137 #define LANG_FULFULDE 0x67
 138 #endif
 139 #ifndef LANG_GAELIC
 140 #define LANG_GAELIC 0x3c
 141 #endif
 142 #ifndef LANG_GALICIAN
 143 #define LANG_GALICIAN 0x56
 144 #endif
 145 #ifndef LANG_GEORGIAN
 146 #define LANG_GEORGIAN 0x37
 147 #endif
 148 #ifndef LANG_GUARANI
 149 #define LANG_GUARANI 0x74
 150 #endif
 151 #ifndef LANG_GUJARATI
 152 #define LANG_GUJARATI 0x47
 153 #endif
 154 #ifndef LANG_HAUSA
 155 #define LANG_HAUSA 0x68
 156 #endif
 157 #ifndef LANG_HAWAIIAN
 158 #define LANG_HAWAIIAN 0x75
 159 #endif
 160 #ifndef LANG_HEBREW
 161 #define LANG_HEBREW 0x0d
 162 #endif
 163 #ifndef LANG_HINDI
 164 #define LANG_HINDI 0x39
 165 #endif
 166 #ifndef LANG_IBIBIO
 167 #define LANG_IBIBIO 0x69
 168 #endif
 169 #ifndef LANG_IGBO
 170 #define LANG_IGBO 0x70
 171 #endif
 172 #ifndef LANG_INDONESIAN
 173 #define LANG_INDONESIAN 0x21
 174 #endif
 175 #ifndef LANG_INUKTITUT
 176 #define LANG_INUKTITUT 0x5d
 177 #endif
 178 #ifndef LANG_KANNADA
 179 #define LANG_KANNADA 0x4b
 180 #endif
 181 #ifndef LANG_KANURI
 182 #define LANG_KANURI 0x71
 183 #endif
 184 #ifndef LANG_KASHMIRI
 185 #define LANG_KASHMIRI 0x60
 186 #endif
 187 #ifndef LANG_KAZAK
 188 #define LANG_KAZAK 0x3f
 189 #endif
 190 #ifndef LANG_KONKANI
 191 #define LANG_KONKANI 0x57
 192 #endif
 193 #ifndef LANG_KYRGYZ
 194 #define LANG_KYRGYZ 0x40
 195 #endif
 196 #ifndef LANG_LAO
 197 #define LANG_LAO 0x54
 198 #endif
 199 #ifndef LANG_LATIN
 200 #define LANG_LATIN 0x76
 201 #endif
 202 #ifndef LANG_LATVIAN
 203 #define LANG_LATVIAN 0x26
 204 #endif
 205 #ifndef LANG_LITHUANIAN
 206 #define LANG_LITHUANIAN 0x27
 207 #endif
 208 #ifndef LANG_MACEDONIAN
 209 #define LANG_MACEDONIAN 0x2f
 210 #endif
 211 #ifndef LANG_MALAY
 212 #define LANG_MALAY 0x3e
 213 #endif
 214 #ifndef LANG_MALAYALAM
 215 #define LANG_MALAYALAM 0x4c
 216 #endif
 217 #ifndef LANG_MALTESE
 218 #define LANG_MALTESE 0x3a
 219 #endif
 220 #ifndef LANG_MANIPURI
 221 #define LANG_MANIPURI 0x58
 222 #endif
 223 #ifndef LANG_MARATHI
 224 #define LANG_MARATHI 0x4e
 225 #endif
 226 #ifndef LANG_MONGOLIAN
 227 #define LANG_MONGOLIAN 0x50
 228 #endif
 229 #ifndef LANG_NEPALI
 230 #define LANG_NEPALI 0x61
 231 #endif
 232 #ifndef LANG_ORIYA
 233 #define LANG_ORIYA 0x48
 234 #endif
 235 #ifndef LANG_OROMO
 236 #define LANG_OROMO 0x72
 237 #endif
 238 #ifndef LANG_PAPIAMENTU
 239 #define LANG_PAPIAMENTU 0x79
 240 #endif
 241 #ifndef LANG_PASHTO
 242 #define LANG_PASHTO 0x63
 243 #endif
 244 #ifndef LANG_PUNJABI
 245 #define LANG_PUNJABI 0x46
 246 #endif
 247 #ifndef LANG_RHAETO_ROMANCE
 248 #define LANG_RHAETO_ROMANCE 0x17
 249 #endif
 250 #ifndef LANG_SAAMI
 251 #define LANG_SAAMI 0x3b
 252 #endif
 253 #ifndef LANG_SANSKRIT
 254 #define LANG_SANSKRIT 0x4f
 255 #endif
 256 #ifndef LANG_SERBIAN
 257 #define LANG_SERBIAN 0x1a
 258 #endif
 259 #ifndef LANG_SINDHI
 260 #define LANG_SINDHI 0x59
 261 #endif
 262 #ifndef LANG_SINHALESE
 263 #define LANG_SINHALESE 0x5b
 264 #endif
 265 #ifndef LANG_SLOVAK
 266 #define LANG_SLOVAK 0x1b
 267 #endif
 268 #ifndef LANG_SOMALI
 269 #define LANG_SOMALI 0x77
 270 #endif
 271 #ifndef LANG_SORBIAN
 272 #define LANG_SORBIAN 0x2e
 273 #endif
 274 #ifndef LANG_SUTU
 275 #define LANG_SUTU 0x30
 276 #endif
 277 #ifndef LANG_SWAHILI
 278 #define LANG_SWAHILI 0x41
 279 #endif
 280 #ifndef LANG_SYRIAC
 281 #define LANG_SYRIAC 0x5a
 282 #endif
 283 #ifndef LANG_TAGALOG
 284 #define LANG_TAGALOG 0x64
 285 #endif
 286 #ifndef LANG_TAJIK
 287 #define LANG_TAJIK 0x28
 288 #endif
 289 #ifndef LANG_TAMAZIGHT
 290 #define LANG_TAMAZIGHT 0x5f
 291 #endif
 292 #ifndef LANG_TAMIL
 293 #define LANG_TAMIL 0x49
 294 #endif
 295 #ifndef LANG_TATAR
 296 #define LANG_TATAR 0x44
 297 #endif
 298 #ifndef LANG_TELUGU
 299 #define LANG_TELUGU 0x4a
 300 #endif
 301 #ifndef LANG_THAI
 302 #define LANG_THAI 0x1e
 303 #endif
 304 #ifndef LANG_TIBETAN
 305 #define LANG_TIBETAN 0x51
 306 #endif
 307 #ifndef LANG_TIGRINYA
 308 #define LANG_TIGRINYA 0x73
 309 #endif
 310 #ifndef LANG_TSONGA
 311 #define LANG_TSONGA 0x31
 312 #endif
 313 #ifndef LANG_TSWANA
 314 #define LANG_TSWANA 0x32
 315 #endif
 316 #ifndef LANG_TURKMEN
 317 #define LANG_TURKMEN 0x42
 318 #endif
 319 #ifndef LANG_UKRAINIAN
 320 #define LANG_UKRAINIAN 0x22
 321 #endif
 322 #ifndef LANG_URDU
 323 #define LANG_URDU 0x20
 324 #endif
 325 #ifndef LANG_UZBEK
 326 #define LANG_UZBEK 0x43
 327 #endif
 328 #ifndef LANG_VENDA
 329 #define LANG_VENDA 0x33
 330 #endif
 331 #ifndef LANG_VIETNAMESE
 332 #define LANG_VIETNAMESE 0x2a
 333 #endif
 334 #ifndef LANG_WELSH
 335 #define LANG_WELSH 0x52
 336 #endif
 337 #ifndef LANG_XHOSA
 338 #define LANG_XHOSA 0x34
 339 #endif
 340 #ifndef LANG_YI
 341 #define LANG_YI 0x78
 342 #endif
 343 #ifndef LANG_YIDDISH
 344 #define LANG_YIDDISH 0x3d
 345 #endif
 346 #ifndef LANG_YORUBA
 347 #define LANG_YORUBA 0x6a
 348 #endif
 349 #ifndef LANG_ZULU
 350 #define LANG_ZULU 0x35
 351 #endif
 352 #ifndef SUBLANG_ARABIC_SAUDI_ARABIA
 353 #define SUBLANG_ARABIC_SAUDI_ARABIA 0x01
 354 #endif
 355 #ifndef SUBLANG_ARABIC_IRAQ
 356 #define SUBLANG_ARABIC_IRAQ 0x02
 357 #endif
 358 #ifndef SUBLANG_ARABIC_EGYPT
 359 #define SUBLANG_ARABIC_EGYPT 0x03
 360 #endif
 361 #ifndef SUBLANG_ARABIC_LIBYA
 362 #define SUBLANG_ARABIC_LIBYA 0x04
 363 #endif
 364 #ifndef SUBLANG_ARABIC_ALGERIA
 365 #define SUBLANG_ARABIC_ALGERIA 0x05
 366 #endif
 367 #ifndef SUBLANG_ARABIC_MOROCCO
 368 #define SUBLANG_ARABIC_MOROCCO 0x06
 369 #endif
 370 #ifndef SUBLANG_ARABIC_TUNISIA
 371 #define SUBLANG_ARABIC_TUNISIA 0x07
 372 #endif
 373 #ifndef SUBLANG_ARABIC_OMAN
 374 #define SUBLANG_ARABIC_OMAN 0x08
 375 #endif
 376 #ifndef SUBLANG_ARABIC_YEMEN
 377 #define SUBLANG_ARABIC_YEMEN 0x09
 378 #endif
 379 #ifndef SUBLANG_ARABIC_SYRIA
 380 #define SUBLANG_ARABIC_SYRIA 0x0a
 381 #endif
 382 #ifndef SUBLANG_ARABIC_JORDAN
 383 #define SUBLANG_ARABIC_JORDAN 0x0b
 384 #endif
 385 #ifndef SUBLANG_ARABIC_LEBANON
 386 #define SUBLANG_ARABIC_LEBANON 0x0c
 387 #endif
 388 #ifndef SUBLANG_ARABIC_KUWAIT
 389 #define SUBLANG_ARABIC_KUWAIT 0x0d
 390 #endif
 391 #ifndef SUBLANG_ARABIC_UAE
 392 #define SUBLANG_ARABIC_UAE 0x0e
 393 #endif
 394 #ifndef SUBLANG_ARABIC_BAHRAIN
 395 #define SUBLANG_ARABIC_BAHRAIN 0x0f
 396 #endif
 397 #ifndef SUBLANG_ARABIC_QATAR
 398 #define SUBLANG_ARABIC_QATAR 0x10
 399 #endif
 400 #ifndef SUBLANG_AZERI_LATIN
 401 #define SUBLANG_AZERI_LATIN 0x01
 402 #endif
 403 #ifndef SUBLANG_AZERI_CYRILLIC
 404 #define SUBLANG_AZERI_CYRILLIC 0x02
 405 #endif
 406 #ifndef SUBLANG_BENGALI_INDIA
 407 #define SUBLANG_BENGALI_INDIA 0x01
 408 #endif
 409 #ifndef SUBLANG_BENGALI_BANGLADESH
 410 #define SUBLANG_BENGALI_BANGLADESH 0x02
 411 #endif
 412 #ifndef SUBLANG_CHINESE_MACAU
 413 #define SUBLANG_CHINESE_MACAU 0x05
 414 #endif
 415 #ifndef SUBLANG_ENGLISH_SOUTH_AFRICA
 416 #define SUBLANG_ENGLISH_SOUTH_AFRICA 0x07
 417 #endif
 418 #ifndef SUBLANG_ENGLISH_JAMAICA
 419 #define SUBLANG_ENGLISH_JAMAICA 0x08
 420 #endif
 421 #ifndef SUBLANG_ENGLISH_CARIBBEAN
 422 #define SUBLANG_ENGLISH_CARIBBEAN 0x09
 423 #endif
 424 #ifndef SUBLANG_ENGLISH_BELIZE
 425 #define SUBLANG_ENGLISH_BELIZE 0x0a
 426 #endif
 427 #ifndef SUBLANG_ENGLISH_TRINIDAD
 428 #define SUBLANG_ENGLISH_TRINIDAD 0x0b
 429 #endif
 430 #ifndef SUBLANG_ENGLISH_ZIMBABWE
 431 #define SUBLANG_ENGLISH_ZIMBABWE 0x0c
 432 #endif
 433 #ifndef SUBLANG_ENGLISH_PHILIPPINES
 434 #define SUBLANG_ENGLISH_PHILIPPINES 0x0d
 435 #endif
 436 #ifndef SUBLANG_ENGLISH_INDONESIA
 437 #define SUBLANG_ENGLISH_INDONESIA 0x0e
 438 #endif
 439 #ifndef SUBLANG_ENGLISH_HONGKONG
 440 #define SUBLANG_ENGLISH_HONGKONG 0x0f
 441 #endif
 442 #ifndef SUBLANG_ENGLISH_INDIA
 443 #define SUBLANG_ENGLISH_INDIA 0x10
 444 #endif
 445 #ifndef SUBLANG_ENGLISH_MALAYSIA
 446 #define SUBLANG_ENGLISH_MALAYSIA 0x11
 447 #endif
 448 #ifndef SUBLANG_ENGLISH_SINGAPORE
 449 #define SUBLANG_ENGLISH_SINGAPORE 0x12
 450 #endif
 451 #ifndef SUBLANG_FRENCH_LUXEMBOURG
 452 #define SUBLANG_FRENCH_LUXEMBOURG 0x05
 453 #endif
 454 #ifndef SUBLANG_FRENCH_MONACO
 455 #define SUBLANG_FRENCH_MONACO 0x06
 456 #endif
 457 #ifndef SUBLANG_FRENCH_WESTINDIES
 458 #define SUBLANG_FRENCH_WESTINDIES 0x07
 459 #endif
 460 #ifndef SUBLANG_FRENCH_REUNION
 461 #define SUBLANG_FRENCH_REUNION 0x08
 462 #endif
 463 #ifndef SUBLANG_FRENCH_CONGO
 464 #define SUBLANG_FRENCH_CONGO 0x09
 465 #endif
 466 #ifndef SUBLANG_FRENCH_SENEGAL
 467 #define SUBLANG_FRENCH_SENEGAL 0x0a
 468 #endif
 469 #ifndef SUBLANG_FRENCH_CAMEROON
 470 #define SUBLANG_FRENCH_CAMEROON 0x0b
 471 #endif
 472 #ifndef SUBLANG_FRENCH_COTEDIVOIRE
 473 #define SUBLANG_FRENCH_COTEDIVOIRE 0x0c
 474 #endif
 475 #ifndef SUBLANG_FRENCH_MALI
 476 #define SUBLANG_FRENCH_MALI 0x0d
 477 #endif
 478 #ifndef SUBLANG_FRENCH_MOROCCO
 479 #define SUBLANG_FRENCH_MOROCCO 0x0e
 480 #endif
 481 #ifndef SUBLANG_FRENCH_HAITI
 482 #define SUBLANG_FRENCH_HAITI 0x0f
 483 #endif
 484 #ifndef SUBLANG_GERMAN_LUXEMBOURG
 485 #define SUBLANG_GERMAN_LUXEMBOURG 0x04
 486 #endif
 487 #ifndef SUBLANG_GERMAN_LIECHTENSTEIN
 488 #define SUBLANG_GERMAN_LIECHTENSTEIN 0x05
 489 #endif
 490 #ifndef SUBLANG_KASHMIRI_INDIA
 491 #define SUBLANG_KASHMIRI_INDIA 0x02
 492 #endif
 493 #ifndef SUBLANG_MALAY_MALAYSIA
 494 #define SUBLANG_MALAY_MALAYSIA 0x01
 495 #endif
 496 #ifndef SUBLANG_MALAY_BRUNEI_DARUSSALAM
 497 #define SUBLANG_MALAY_BRUNEI_DARUSSALAM 0x02
 498 #endif
 499 #ifndef SUBLANG_NEPALI_INDIA
 500 #define SUBLANG_NEPALI_INDIA 0x02
 501 #endif
 502 #ifndef SUBLANG_PUNJABI_INDIA
 503 #define SUBLANG_PUNJABI_INDIA 0x01
 504 #endif
 505 #ifndef SUBLANG_ROMANIAN_ROMANIA
 506 #define SUBLANG_ROMANIAN_ROMANIA 0x01
 507 #endif
 508 #ifndef SUBLANG_SERBIAN_LATIN
 509 #define SUBLANG_SERBIAN_LATIN 0x02
 510 #endif
 511 #ifndef SUBLANG_SERBIAN_CYRILLIC
 512 #define SUBLANG_SERBIAN_CYRILLIC 0x03
 513 #endif
 514 #ifndef SUBLANG_SINDHI_INDIA
 515 #define SUBLANG_SINDHI_INDIA 0x00
 516 #endif
 517 #ifndef SUBLANG_SINDHI_PAKISTAN
 518 #define SUBLANG_SINDHI_PAKISTAN 0x01
 519 #endif
 520 #ifndef SUBLANG_SPANISH_GUATEMALA
 521 #define SUBLANG_SPANISH_GUATEMALA 0x04
 522 #endif
 523 #ifndef SUBLANG_SPANISH_COSTA_RICA
 524 #define SUBLANG_SPANISH_COSTA_RICA 0x05
 525 #endif
 526 #ifndef SUBLANG_SPANISH_PANAMA
 527 #define SUBLANG_SPANISH_PANAMA 0x06
 528 #endif
 529 #ifndef SUBLANG_SPANISH_DOMINICAN_REPUBLIC
 530 #define SUBLANG_SPANISH_DOMINICAN_REPUBLIC 0x07
 531 #endif
 532 #ifndef SUBLANG_SPANISH_VENEZUELA
 533 #define SUBLANG_SPANISH_VENEZUELA 0x08
 534 #endif
 535 #ifndef SUBLANG_SPANISH_COLOMBIA
 536 #define SUBLANG_SPANISH_COLOMBIA 0x09
 537 #endif
 538 #ifndef SUBLANG_SPANISH_PERU
 539 #define SUBLANG_SPANISH_PERU 0x0a
 540 #endif
 541 #ifndef SUBLANG_SPANISH_ARGENTINA
 542 #define SUBLANG_SPANISH_ARGENTINA 0x0b
 543 #endif
 544 #ifndef SUBLANG_SPANISH_ECUADOR
 545 #define SUBLANG_SPANISH_ECUADOR 0x0c
 546 #endif
 547 #ifndef SUBLANG_SPANISH_CHILE
 548 #define SUBLANG_SPANISH_CHILE 0x0d
 549 #endif
 550 #ifndef SUBLANG_SPANISH_URUGUAY
 551 #define SUBLANG_SPANISH_URUGUAY 0x0e
 552 #endif
 553 #ifndef SUBLANG_SPANISH_PARAGUAY
 554 #define SUBLANG_SPANISH_PARAGUAY 0x0f
 555 #endif
 556 #ifndef SUBLANG_SPANISH_BOLIVIA
 557 #define SUBLANG_SPANISH_BOLIVIA 0x10
 558 #endif
 559 #ifndef SUBLANG_SPANISH_EL_SALVADOR
 560 #define SUBLANG_SPANISH_EL_SALVADOR 0x11
 561 #endif
 562 #ifndef SUBLANG_SPANISH_HONDURAS
 563 #define SUBLANG_SPANISH_HONDURAS 0x12
 564 #endif
 565 #ifndef SUBLANG_SPANISH_NICARAGUA
 566 #define SUBLANG_SPANISH_NICARAGUA 0x13
 567 #endif
 568 #ifndef SUBLANG_SPANISH_PUERTO_RICO
 569 #define SUBLANG_SPANISH_PUERTO_RICO 0x14
 570 #endif
 571 #ifndef SUBLANG_SWEDISH_FINLAND
 572 #define SUBLANG_SWEDISH_FINLAND 0x02
 573 #endif
 574 #ifndef SUBLANG_TAMAZIGHT_ARABIC
 575 #define SUBLANG_TAMAZIGHT_ARABIC 0x01
 576 #endif
 577 #ifndef SUBLANG_TAMAZIGHT_LATIN
 578 #define SUBLANG_TAMAZIGHT_LATIN 0x02
 579 #endif
 580 #ifndef SUBLANG_TIGRINYA_ETHIOPIA
 581 #define SUBLANG_TIGRINYA_ETHIOPIA 0x00
 582 #endif
 583 #ifndef SUBLANG_TIGRINYA_ERITREA
 584 #define SUBLANG_TIGRINYA_ERITREA 0x01
 585 #endif
 586 #ifndef SUBLANG_URDU_PAKISTAN
 587 #define SUBLANG_URDU_PAKISTAN 0x01
 588 #endif
 589 #ifndef SUBLANG_URDU_INDIA
 590 #define SUBLANG_URDU_INDIA 0x02
 591 #endif
 592 #ifndef SUBLANG_UZBEK_LATIN
 593 #define SUBLANG_UZBEK_LATIN 0x01
 594 #endif
 595 #ifndef SUBLANG_UZBEK_CYRILLIC
 596 #define SUBLANG_UZBEK_CYRILLIC 0x02
 597 #endif
 598
 599 /* Return an XPG style locale name
 600      language[_territory[.codeset]][@modifier].
 601    Don't even bother determining the codeset; it's not useful in this
 602    context, because message catalogs are not specific to a single
 603    codeset.  The result must not be freed; it is statically
 604    allocated.  */
 605 static const char *
 606 my_nl_locale_name (const char *categoryname)
 607 {
 608   const char *retval;
 609   LCID lcid;
 610   LANGID langid;
 611   int primary, sub;
 612
 613   /* Let the user override the system settings through environment
 614      variables, as on POSIX systems.  */
 615   retval = getenv ("LC_ALL");
 616   if (retval != NULL && retval[0] != '\0')
 617     return retval;
 618   retval = getenv (categoryname);
 619   if (retval != NULL && retval[0] != '\0')
 620     return retval;
 621   retval = getenv ("LANG");
 622   if (retval != NULL && retval[0] != '\0')
 623     return retval;
 624
 625   /* Use native Win32 API locale ID.  */
 626   lcid = GetThreadLocale ();
 627
 628   /* Strip off the sorting rules, keep only the language part.  */
 629   langid = LANGIDFROMLCID (lcid);
 630
 631   /* Split into language and territory part.  */
 632   primary = PRIMARYLANGID (langid);
 633   sub = SUBLANGID (langid);
 634
 635   /* Dispatch on language.
 636      See also http://www.unicode.org/unicode/onlinedat/languages.html .
 637      For details about languages, see http://www.ethnologue.com/ .  */
 638   switch (primary)
 639     {
 640     case LANG_AFRIKAANS: return "af_ZA";
 641     case LANG_ALBANIAN: return "sq_AL";
 642     case LANG_AMHARIC: return "am_ET";
 643     case LANG_ARABIC:
 644       switch (sub)
 645         {
 646         case SUBLANG_ARABIC_SAUDI_ARABIA: return "ar_SA";
 647         case SUBLANG_ARABIC_IRAQ: return "ar_IQ";
 648         case SUBLANG_ARABIC_EGYPT: return "ar_EG";
 649         case SUBLANG_ARABIC_LIBYA: return "ar_LY";
 650         case SUBLANG_ARABIC_ALGERIA: return "ar_DZ";
 651         case SUBLANG_ARABIC_MOROCCO: return "ar_MA";
 652         case SUBLANG_ARABIC_TUNISIA: return "ar_TN";
 653         case SUBLANG_ARABIC_OMAN: return "ar_OM";
 654         case SUBLANG_ARABIC_YEMEN: return "ar_YE";
 655         case SUBLANG_ARABIC_SYRIA: return "ar_SY";
 656         case SUBLANG_ARABIC_JORDAN: return "ar_JO";
 657         case SUBLANG_ARABIC_LEBANON: return "ar_LB";
 658         case SUBLANG_ARABIC_KUWAIT: return "ar_KW";
 659         case SUBLANG_ARABIC_UAE: return "ar_AE";
 660         case SUBLANG_ARABIC_BAHRAIN: return "ar_BH";
 661         case SUBLANG_ARABIC_QATAR: return "ar_QA";
 662         }
 663       return "ar";
 664     case LANG_ARMENIAN: return "hy_AM";
 665     case LANG_ASSAMESE: return "as_IN";
 666     case LANG_AZERI:
 667       switch (sub)
 668         {
 669         /* FIXME: Adjust this when Azerbaijani locales appear on Unix.  */
 670         case SUBLANG_AZERI_LATIN: return "az_AZ@latin";
 671         case SUBLANG_AZERI_CYRILLIC: return "az_AZ@cyrillic";
 672         }
 673       return "az";
 674     case LANG_BASQUE:
 675       return "eu"; /* Ambiguous: could be "eu_ES" or "eu_FR".  */
 676     case LANG_BELARUSIAN: return "be_BY";
 677     case LANG_BENGALI:
 678       switch (sub)
 679         {
 680         case SUBLANG_BENGALI_INDIA: return "bn_IN";
 681         case SUBLANG_BENGALI_BANGLADESH: return "bn_BD";
 682         }
 683       return "bn";
 684     case LANG_BULGARIAN: return "bg_BG";
 685     case LANG_BURMESE: return "my_MM";
 686     case LANG_CAMBODIAN: return "km_KH";
 687     case LANG_CATALAN: return "ca_ES";
 688     case LANG_CHEROKEE: return "chr_US";
 689     case LANG_CHINESE:
 690       switch (sub)
 691         {
 692         case SUBLANG_CHINESE_TRADITIONAL: return "zh_TW";
 693         case SUBLANG_CHINESE_SIMPLIFIED: return "zh_CN";
 694         case SUBLANG_CHINESE_HONGKONG: return "zh_HK";
 695         case SUBLANG_CHINESE_SINGAPORE: return "zh_SG";
 696         case SUBLANG_CHINESE_MACAU: return "zh_MO";
 697         }
 698       return "zh";
 699     case LANG_CROATIAN:         /* LANG_CROATIAN == LANG_SERBIAN
 700                                  * What used to be called Serbo-Croatian
 701                                  * should really now be two separate
 702                                  * languages because of political reasons.
 703                                  * (Says tml, who knows nothing about Serbian
 704                                  * or Croatian.)
 705                                  * (I can feel those flames coming already.)
 706                                  */
 707       switch (sub)
 708         {
 709         case SUBLANG_DEFAULT: return "hr_HR";
 710         case SUBLANG_SERBIAN_LATIN: return "sr_CS";
 711         case SUBLANG_SERBIAN_CYRILLIC: return "sr_CS@cyrillic";
 712         }
 713       return "hr";
 714     case LANG_CZECH: return "cs_CZ";
 715     case LANG_DANISH: return "da_DK";
 716     case LANG_DIVEHI: return "div_MV";
 717     case LANG_DUTCH:
 718       switch (sub)
 719         {
 720         case SUBLANG_DUTCH: return "nl_NL";
 721         case SUBLANG_DUTCH_BELGIAN: /* FLEMISH, VLAAMS */ return "nl_BE";
 722         }
 723       return "nl";
 724     case LANG_EDO: return "bin_NG";
 725     case LANG_ENGLISH:
 726       switch (sub)
 727         {
 728         /* SUBLANG_ENGLISH_US == SUBLANG_DEFAULT. Heh. I thought
 729          * English was the language spoken in England.
 730          * Oh well.
 731          */
 732         case SUBLANG_ENGLISH_US: return "en_US";
 733         case SUBLANG_ENGLISH_UK: return "en_GB";
 734         case SUBLANG_ENGLISH_AUS: return "en_AU";
 735         case SUBLANG_ENGLISH_CAN: return "en_CA";
 736         case SUBLANG_ENGLISH_NZ: return "en_NZ";
 737         case SUBLANG_ENGLISH_EIRE: return "en_IE";
 738         case SUBLANG_ENGLISH_SOUTH_AFRICA: return "en_ZA";
 739         case SUBLANG_ENGLISH_JAMAICA: return "en_JM";
 740         case SUBLANG_ENGLISH_CARIBBEAN: return "en_GD"; /* Grenada? */
 741         case SUBLANG_ENGLISH_BELIZE: return "en_BZ";
 742         case SUBLANG_ENGLISH_TRINIDAD: return "en_TT";
 743         case SUBLANG_ENGLISH_ZIMBABWE: return "en_ZW";
 744         case SUBLANG_ENGLISH_PHILIPPINES: return "en_PH";
 745         case SUBLANG_ENGLISH_INDONESIA: return "en_ID";
 746         case SUBLANG_ENGLISH_HONGKONG: return "en_HK";
 747         case SUBLANG_ENGLISH_INDIA: return "en_IN";
 748         case SUBLANG_ENGLISH_MALAYSIA: return "en_MY";
 749         case SUBLANG_ENGLISH_SINGAPORE: return "en_SG";
 750         }
 751       return "en";
 752     case LANG_ESTONIAN: return "et_EE";
 753     case LANG_FAEROESE: return "fo_FO";
 754     case LANG_FARSI: return "fa_IR";
 755     case LANG_FINNISH: return "fi_FI";
 756     case LANG_FRENCH:
 757       switch (sub)
 758         {
 759         case SUBLANG_FRENCH: return "fr_FR";
 760         case SUBLANG_FRENCH_BELGIAN: /* WALLOON */ return "fr_BE";
 761         case SUBLANG_FRENCH_CANADIAN: return "fr_CA";
 762         case SUBLANG_FRENCH_SWISS: return "fr_CH";
 763         case SUBLANG_FRENCH_LUXEMBOURG: return "fr_LU";
 764         case SUBLANG_FRENCH_MONACO: return "fr_MC";
 765         case SUBLANG_FRENCH_WESTINDIES: return "fr"; /* Caribbean? */
 766         case SUBLANG_FRENCH_REUNION: return "fr_RE";
 767         case SUBLANG_FRENCH_CONGO: return "fr_CG";
 768         case SUBLANG_FRENCH_SENEGAL: return "fr_SN";
 769         case SUBLANG_FRENCH_CAMEROON: return "fr_CM";
 770         case SUBLANG_FRENCH_COTEDIVOIRE: return "fr_CI";
 771         case SUBLANG_FRENCH_MALI: return "fr_ML";
 772         case SUBLANG_FRENCH_MOROCCO: return "fr_MA";
 773         case SUBLANG_FRENCH_HAITI: return "fr_HT";
 774         }
 775       return "fr";
 776     case LANG_FRISIAN: return "fy_NL";
 777     case LANG_FULFULDE: return "ful_NG";
 778     case LANG_GAELIC:
 779       switch (sub)
 780         {
 781         case 0x01: /* SCOTTISH */ return "gd_GB";
 782         case 0x02: /* IRISH */ return "ga_IE";
 783         }
 784       return "C";
 785     case LANG_GALICIAN: return "gl_ES";
 786     case LANG_GEORGIAN: return "ka_GE";
 787     case LANG_GERMAN:
 788       switch (sub)
 789         {
 790         case SUBLANG_GERMAN: return "de_DE";
 791         case SUBLANG_GERMAN_SWISS: return "de_CH";
 792         case SUBLANG_GERMAN_AUSTRIAN: return "de_AT";
 793         case SUBLANG_GERMAN_LUXEMBOURG: return "de_LU";
 794         case SUBLANG_GERMAN_LIECHTENSTEIN: return "de_LI";
 795         }
 796       return "de";
 797     case LANG_GREEK: return "el_GR";
 798     case LANG_GUARANI: return "gn_PY";
 799     case LANG_GUJARATI: return "gu_IN";
 800     case LANG_HAUSA: return "ha_NG";
 801     case LANG_HAWAIIAN:
 802       /* FIXME: Do they mean Hawaiian ("haw_US", 1000 speakers)
 803          or Hawaii Creole English ("cpe_US", 600000 speakers)?  */
 804       return "cpe_US";
 805     case LANG_HEBREW: return "he_IL";
 806     case LANG_HINDI: return "hi_IN";
 807     case LANG_HUNGARIAN: return "hu_HU";
 808     case LANG_IBIBIO: return "nic_NG";
 809     case LANG_ICELANDIC: return "is_IS";
 810     case LANG_IGBO: return "ibo_NG";
 811     case LANG_INDONESIAN: return "id_ID";
 812     case LANG_INUKTITUT: return "iu_CA";
 813     case LANG_ITALIAN:
 814       switch (sub)
 815         {
 816         case SUBLANG_ITALIAN: return "it_IT";
 817         case SUBLANG_ITALIAN_SWISS: return "it_CH";
 818         }
 819       return "it";
 820     case LANG_JAPANESE: return "ja_JP";
 821     case LANG_KANNADA: return "kn_IN";
 822     case LANG_KANURI: return "kau_NG";
 823     case LANG_KASHMIRI:
 824       switch (sub)
 825         {
 826         case SUBLANG_DEFAULT: return "ks_PK";
 827         case SUBLANG_KASHMIRI_INDIA: return "ks_IN";
 828         }
 829       return "ks";
 830     case LANG_KAZAK: return "kk_KZ";
 831     case LANG_KONKANI:
 832       /* FIXME: Adjust this when such locales appear on Unix.  */
 833       return "kok_IN";
 834     case LANG_KOREAN: return "ko_KR";
 835     case LANG_KYRGYZ: return "ky_KG";
 836     case LANG_LAO: return "lo_LA";
 837     case LANG_LATIN: return "la_VA";
 838     case LANG_LATVIAN: return "lv_LV";
 839     case LANG_LITHUANIAN: return "lt_LT";
 840     case LANG_MACEDONIAN: return "mk_MK";
 841     case LANG_MALAY:
 842       switch (sub)
 843         {
 844         case SUBLANG_MALAY_MALAYSIA: return "ms_MY";
 845         case SUBLANG_MALAY_BRUNEI_DARUSSALAM: return "ms_BN";
 846         }
 847       return "ms";
 848     case LANG_MALAYALAM: return "ml_IN";
 849     case LANG_MALTESE: return "mt_MT";
 850     case LANG_MANIPURI:
 851       /* FIXME: Adjust this when such locales appear on Unix.  */
 852       return "mni_IN";
 853     case LANG_MARATHI: return "mr_IN";
 854     case LANG_MONGOLIAN:
 855       return "mn"; /* Ambiguous: could be "mn_CN" or "mn_MN".  */
 856     case LANG_NEPALI:
 857       switch (sub)
 858         {
 859         case SUBLANG_DEFAULT: return "ne_NP";
 860         case SUBLANG_NEPALI_INDIA: return "ne_IN";
 861         }
 862       return "ne";
 863     case LANG_NORWEGIAN:
 864       switch (sub)
 865         {
 866         case SUBLANG_NORWEGIAN_BOKMAL: return "no_NO";
 867         case SUBLANG_NORWEGIAN_NYNORSK: return "nn_NO";
 868         }
 869       return "no";
 870     case LANG_ORIYA: return "or_IN";
 871     case LANG_OROMO: return "om_ET";
 872     case LANG_PAPIAMENTU: return "pap_AN";
 873     case LANG_PASHTO:
 874       return "ps"; /* Ambiguous: could be "ps_PK" or "ps_AF".  */
 875     case LANG_POLISH: return "pl_PL";
 876     case LANG_PORTUGUESE:
 877       switch (sub)
 878         {
 879         case SUBLANG_PORTUGUESE: return "pt_PT";
 880         /* Hmm. SUBLANG_PORTUGUESE_BRAZILIAN == SUBLANG_DEFAULT.
 881            Same phenomenon as SUBLANG_ENGLISH_US == SUBLANG_DEFAULT. */
 882         case SUBLANG_PORTUGUESE_BRAZILIAN: return "pt_BR";
 883         }
 884       return "pt";
 885     case LANG_PUNJABI:
 886       switch (sub)
 887         {
 888         case SUBLANG_PUNJABI_INDIA: return "pa_IN"; /* Gurmukhi script */
 889         }
 890       return "pa";
 891     case LANG_RHAETO_ROMANCE: return "rm_CH";
 892     case LANG_ROMANIAN:
 893       switch (sub)
 894         {
 895         case SUBLANG_ROMANIAN_ROMANIA: return "ro_RO";
 896         }
 897       return "ro";
 898     case LANG_RUSSIAN:
 899       return "ru"; /* Ambiguous: could be "ru_RU" or "ru_UA" or "ru_MD".  */
 900     case LANG_SAAMI: /* actually Northern Sami */ return "se_NO";
 901     case LANG_SANSKRIT: return "sa_IN";
 902     case LANG_SINDHI:
 903       switch (sub)
 904         {
 905         case SUBLANG_SINDHI_INDIA: return "sd_IN";
 906         case SUBLANG_SINDHI_PAKISTAN: return "sd_PK";
 907         }
 908       return "sd";
 909     case LANG_SINHALESE: return "si_LK";
 910     case LANG_SLOVAK: return "sk_SK";
 911     case LANG_SLOVENIAN: return "sl_SI";
 912     case LANG_SOMALI: return "so_SO";
 913     case LANG_SORBIAN:
 914       /* FIXME: Adjust this when such locales appear on Unix.  */
 915       return "wen_DE";
 916     case LANG_SPANISH:
 917       switch (sub)
 918         {
 919         case SUBLANG_SPANISH: return "es_ES";
 920         case SUBLANG_SPANISH_MEXICAN: return "es_MX";
 921         case SUBLANG_SPANISH_MODERN:
 922           return "es_ES@modern";        /* not seen on Unix */
 923         case SUBLANG_SPANISH_GUATEMALA: return "es_GT";
 924         case SUBLANG_SPANISH_COSTA_RICA: return "es_CR";
 925         case SUBLANG_SPANISH_PANAMA: return "es_PA";
 926         case SUBLANG_SPANISH_DOMINICAN_REPUBLIC: return "es_DO";
 927         case SUBLANG_SPANISH_VENEZUELA: return "es_VE";
 928         case SUBLANG_SPANISH_COLOMBIA: return "es_CO";
 929         case SUBLANG_SPANISH_PERU: return "es_PE";
 930         case SUBLANG_SPANISH_ARGENTINA: return "es_AR";
 931         case SUBLANG_SPANISH_ECUADOR: return "es_EC";
 932         case SUBLANG_SPANISH_CHILE: return "es_CL";
 933         case SUBLANG_SPANISH_URUGUAY: return "es_UY";
 934         case SUBLANG_SPANISH_PARAGUAY: return "es_PY";
 935         case SUBLANG_SPANISH_BOLIVIA: return "es_BO";
 936         case SUBLANG_SPANISH_EL_SALVADOR: return "es_SV";
 937         case SUBLANG_SPANISH_HONDURAS: return "es_HN";
 938         case SUBLANG_SPANISH_NICARAGUA: return "es_NI";
 939         case SUBLANG_SPANISH_PUERTO_RICO: return "es_PR";
 940         }
 941       return "es";
 942     case LANG_SUTU: return "bnt_TZ"; /* or "st_LS" or "nso_ZA"? */
 943     case LANG_SWAHILI: return "sw_KE";
 944     case LANG_SWEDISH:
 945       switch (sub)
 946         {
 947         case SUBLANG_DEFAULT: return "sv_SE";
 948         case SUBLANG_SWEDISH_FINLAND: return "sv_FI";
 949         }
 950       return "sv";
 951     case LANG_SYRIAC: return "syr_TR"; /* An extinct language.  */
 952     case LANG_TAGALOG: return "tl_PH";
 953     case LANG_TAJIK: return "tg_TJ";
 954     case LANG_TAMAZIGHT:
 955       switch (sub)
 956         {
 957         /* FIXME: Adjust this when Tamazight locales appear on Unix.  */
 958         case SUBLANG_TAMAZIGHT_ARABIC: return "ber_MA@arabic";
 959         case SUBLANG_TAMAZIGHT_LATIN: return "ber_MA@latin";
 960         }
 961       return "ber_MA";
 962     case LANG_TAMIL:
 963       return "ta"; /* Ambiguous: could be "ta_IN" or "ta_LK" or "ta_SG".  */
 964     case LANG_TATAR: return "tt_RU";
 965     case LANG_TELUGU: return "te_IN";
 966     case LANG_THAI: return "th_TH";
 967     case LANG_TIBETAN: return "bo_CN";
 968     case LANG_TIGRINYA:
 969       switch (sub)
 970         {
 971         case SUBLANG_TIGRINYA_ETHIOPIA: return "ti_ET";
 972         case SUBLANG_TIGRINYA_ERITREA: return "ti_ER";
 973         }
 974       return "ti";
 975     case LANG_TSONGA: return "ts_ZA";
 976     case LANG_TSWANA: return "tn_BW";
 977     case LANG_TURKISH: return "tr_TR";
 978     case LANG_TURKMEN: return "tk_TM";
 979     case LANG_UKRAINIAN: return "uk_UA";
 980     case LANG_URDU:
 981       switch (sub)
 982         {
 983         case SUBLANG_URDU_PAKISTAN: return "ur_PK";
 984         case SUBLANG_URDU_INDIA: return "ur_IN";
 985         }
 986       return "ur";
 987     case LANG_UZBEK:
 988       switch (sub)
 989         {
 990         case SUBLANG_UZBEK_LATIN: return "uz_UZ";
 991         case SUBLANG_UZBEK_CYRILLIC: return "uz_UZ@cyrillic";
 992         }
 993       return "uz";
 994     case LANG_VENDA:
 995       /* FIXME: It's not clear whether Venda has the ISO 639-2 two-letter code
 996          "ve" or not.
 997          http://www.loc.gov/standards/iso639-2/englangn.html has it, but
 998          http://lcweb.loc.gov/standards/iso639-2/codechanges.html doesn't,  */
 999       return "ven_ZA"; /* or "ve_ZA"? */
1000     case LANG_VIETNAMESE: return "vi_VN";
1001     case LANG_WELSH: return "cy_GB";
1002     case LANG_XHOSA: return "xh_ZA";
1003     case LANG_YI: return "sit_CN";
1004     case LANG_YIDDISH: return "yi_IL";
1005     case LANG_YORUBA: return "yo_NG";
1006     case LANG_ZULU: return "zu_ZA";
1007     default: return "C";
1008     }
1009 }
1010
1011 /* localname.c from gettext END.  */
1012
1013
1014 \f
1015 /* Support functions.  */
1016
1017 static __inline__ uint32_t
1018 do_swap_u32 (uint32_t i)
1019 {
1020   return (i << 24) | ((i & 0xff00) << 8) | ((i >> 8) & 0xff00) | (i >> 24);
1021 }
1022
1023 #define SWAPIT(flag, data) ((flag) ? do_swap_u32(data) : (data))
1024
1025
1026 /* We assume to have `unsigned long int' value with at least 32 bits.  */
1027 #define HASHWORDBITS 32
1028
1029 /* The so called `hashpjw' function by P.J. Weinberger
1030    [see Aho/Sethi/Ullman, COMPILERS: Principles, Techniques and Tools,
1031    1986, 1987 Bell Telephone Laboratories, Inc.]  */
1032 static __inline__ unsigned long
1033 hash_string( const char *str_param )
1034 {
1035   unsigned long int hval, g;
1036   const char *str = str_param;
1037
1038   hval = 0;
1039   while (*str != '\0')
1040     {
1041       hval <<= 4;
1042       hval += (unsigned long int) *str++;
1043       g = hval & ((unsigned long int) 0xf << (HASHWORDBITS - 4));
1044       if (g != 0)
1045         {
1046           hval ^= g >> (HASHWORDBITS - 8);
1047           hval ^= g;
1048         }
1049     }
1050   return hval;
1051 }
1052
1053 /* static char * */
1054 /* my_xstrdup (const char *s) */
1055 /* { */
1056 /*   size_t n = strlen (s) + 1; */
1057 /*   char *p = jnlib_malloc (n); */
1058 /*   if (!p) */
1059 /*     abort (); */
1060 /*   strcpy (p, s); */
1061 /*   return p; */
1062 /* } */
1063
1064
1065 \f
1066 /* Generic message catalog and gettext stuff.  */
1067
1068 /* The magic number of the GNU message catalog format.  */
1069 #define MAGIC         0x950412de
1070 #define MAGIC_SWAPPED 0xde120495
1071
1072 /* Revision number of the currently used .mo (binary) file format.  */
1073 #define MO_REVISION_NUMBER 0
1074
1075
1076 /* Header for binary .mo file format.  */
1077 struct mo_file_header
1078 {
1079   /* The magic number.  */
1080   uint32_t magic;
1081   /* The revision number of the file format.  */
1082   uint32_t revision;
1083   /* The number of strings pairs.  */
1084   uint32_t nstrings;
1085   /* Offset of table with start offsets of original strings.  */
1086   uint32_t orig_tab_offset;
1087   /* Offset of table with start offsets of translation strings.  */
1088   uint32_t trans_tab_offset;
1089   /* Size of hashing table.  */
1090   uint32_t hash_tab_size;
1091   /* Offset of first hashing entry.  */
1092   uint32_t hash_tab_offset;
1093 };
1094
1095
1096 struct string_desc
1097 {
1098   /* Length of addressed string.  */
1099   uint32_t length;
1100   /* Offset of string in file.  */
1101   uint32_t offset;
1102 };
1103
1104
1105 struct overflow_space_s
1106 {
1107   struct overflow_space_s *next;
1108   uint32_t idx;
1109   uint32_t length;
1110   char d[1];
1111 };
1112
1113 struct loaded_domain
1114 {
1115   char *data;
1116   char *data_native; /* Data mapped to the native version of the
1117                         string.  (Allocated along with DATA). */
1118   int must_swap;
1119   uint32_t nstrings;
1120   uint32_t *mapped;  /* 0   := Not mapped (original utf8).
1121                         1   := Mapped to native encoding in overflow space.
1122                         >=2 := Mapped to native encoding. The values
1123                                gives the length of the mapped string.
1124                                becuase the 0 is included and an empty
1125                                string is not allowed we will enver get
1126                                values 0 and 1.  */
1127   struct overflow_space_s *overflow_space;
1128   struct string_desc *orig_tab;
1129   struct string_desc *trans_tab;
1130   uint32_t hash_size;
1131   uint32_t *hash_tab;
1132 };
1133
1134
1135 /* The domain we use.  We only support one domain at this point.  This
1136    is why this implementation can not be shared.  Bindtextdomain and
1137    dgettext will simply cheat and always use this one domain.  */
1138 static struct loaded_domain *the_domain;
1139
1140 /* Global flag to switch gettext into an utf8 mode.  */
1141 static int want_utf8;
1142
1143
1144 \f
1145 /* Free the domain data.  */
1146 static void
1147 free_domain (struct loaded_domain *domain)
1148 {
1149   struct overflow_space_s *os, *os2;
1150
1151   jnlib_free (domain->data);
1152   jnlib_free (domain->mapped);
1153   for (os = domain->overflow_space; os; os = os2)
1154     {
1155       os2 = os->next;
1156       jnlib_free (os);
1157     }
1158   jnlib_free (domain);
1159 }
1160
1161
1162 static struct loaded_domain *
1163 load_domain (const char *filename)
1164 {
1165   FILE *fp;
1166   size_t size;
1167   struct stat st;
1168   struct mo_file_header *data = NULL;
1169   struct loaded_domain *domain = NULL;
1170   size_t to_read;
1171   char *read_ptr;
1172
1173   fp = fopen (filename, "rb");
1174   if (!fp)
1175     return NULL;
1176
1177   /* Determine the file size.  */
1178   if (fstat (fileno (fp), &st)
1179       || (size = (size_t) st.st_size) != st.st_size
1180       || size < sizeof (struct mo_file_header))
1181     {
1182       fclose (fp);
1183       return NULL;
1184     }
1185
1186   data = (2*size <= size)? NULL : jnlib_malloc (2*size);
1187   if (!data)
1188     {
1189       fclose (fp);
1190       return NULL;
1191     }
1192
1193   to_read = size;
1194   read_ptr = (char *) data;
1195   do
1196     {
1197       long int nb = fread (read_ptr, 1, to_read, fp);
1198       if (nb < to_read)
1199         {
1200           fclose (fp);
1201           jnlib_free (data);
1202           return NULL;
1203         }
1204       read_ptr += nb;
1205       to_read -= nb;
1206     }
1207   while (to_read > 0);
1208   fclose (fp);
1209
1210   /* Using the magic number we can test whether it really is a message
1211      catalog file.  */
1212   if (data->magic != MAGIC && data->magic != MAGIC_SWAPPED)
1213     {
1214       /* The magic number is wrong: not a message catalog file.  */
1215       jnlib_free (data);
1216       return NULL;
1217     }
1218
1219   domain = jnlib_calloc (1, sizeof *domain);
1220   if (!domain)
1221     {
1222       jnlib_free (data);
1223       return NULL;
1224     }
1225   domain->data = (char *) data;
1226   domain->data_native = (char *) data + size;
1227   domain->must_swap = data->magic != MAGIC;
1228
1229   /* Fill in the information about the available tables.  */
1230   switch (SWAPIT (domain->must_swap, data->revision))
1231     {
1232     case MO_REVISION_NUMBER:
1233       domain->nstrings = SWAPIT (domain->must_swap, data->nstrings);
1234       domain->orig_tab = (struct string_desc *)
1235         ((char *) data + SWAPIT (domain->must_swap, data->orig_tab_offset));
1236       domain->trans_tab = (struct string_desc *)
1237         ((char *) data + SWAPIT (domain->must_swap, data->trans_tab_offset));
1238       domain->hash_size = SWAPIT (domain->must_swap, data->hash_tab_size);
1239       domain->hash_tab = (uint32_t *)
1240         ((char *) data + SWAPIT (domain->must_swap, data->hash_tab_offset));
1241       break;
1242
1243     default:
1244       /* This is an invalid revision.   */
1245       jnlib_free (data);
1246       jnlib_free (domain);
1247       return NULL;
1248     }
1249
1250   /* Allocate an array to keep track of code page mappings.  */
1251   domain->mapped = jnlib_calloc (domain->nstrings, sizeof *domain->mapped);
1252   if (!domain->mapped)
1253     {
1254       jnlib_free (data);
1255       jnlib_free (domain);
1256       return NULL;
1257     }
1258
1259   return domain;
1260 }
1261
1262
1263 /* Return a malloced wide char string from an UTF-8 encoded input
1264    string STRING.  Caller must free this value. On failure returns
1265    NULL.  The result of calling this function with STRING set to NULL
1266    is not defined. */
1267 static wchar_t *
1268 utf8_to_wchar (const char *string, size_t length, size_t *retlen)
1269 {
1270   int n;
1271   wchar_t *result;
1272   size_t nbytes;
1273
1274   n = MultiByteToWideChar (CP_UTF8, 0, string, length, NULL, 0);
1275   if (n < 0 || (n+1) <= 0)
1276     return NULL;
1277
1278   nbytes = (size_t)(n+1) * sizeof(*result);
1279   if (nbytes / sizeof(*result) != (n+1))
1280     {
1281       errno = ENOMEM;
1282       return NULL;
1283     }
1284   result = jnlib_malloc (nbytes);
1285   if (!result)
1286     return NULL;
1287
1288   n = MultiByteToWideChar (CP_UTF8, 0, string, length, result, n);
1289   if (n < 0)
1290     {
1291       jnlib_free (result);
1292       return NULL;
1293     }
1294   *retlen = n;
1295   return result;
1296 }
1297
1298
1299 /* Return a malloced string encoded in UTF-8 from the wide char input
1300    string STRING.  Caller must free this value. On failure returns
1301    NULL.  The result of calling this function with STRING set to NULL
1302    is not defined. */
1303 static char *
1304 wchar_to_native (const wchar_t *string, size_t length, size_t *retlen)
1305 {
1306   int n;
1307   char *result;
1308
1309   n = WideCharToMultiByte (CP_ACP, 0, string, length, NULL, 0, NULL, NULL);
1310   if (n < 0 || (n+1) <= 0)
1311     return NULL;
1312
1313   result = jnlib_malloc (n+1);
1314   if (!result)
1315     return NULL;
1316
1317   n = WideCharToMultiByte (CP_ACP, 0, string, length, result, n, NULL, NULL);
1318   if (n < 0)
1319     {
1320       jnlib_free (result);
1321       return NULL;
1322     }
1323   *retlen = n;
1324   return result;
1325 }
1326
1327
1328 /* Convert UTF8 to the native codepage.  Caller must free the return value. */
1329 static char *
1330 utf8_to_native (const char *string, size_t length, size_t *retlen)
1331 {
1332   wchar_t *wstring;
1333   char *result;
1334   size_t newlen;
1335
1336   wstring = utf8_to_wchar (string, length, &newlen);
1337   if (wstring)
1338     {
1339       result = wchar_to_native (wstring, newlen, &newlen);
1340       jnlib_free (wstring);
1341     }
1342   else
1343     result = NULL;
1344   *retlen = result? newlen : 0;
1345   return result;
1346 }
1347
1348
1349
1350 \f
1351 /* Specify that the DOMAINNAME message catalog will be found
1352    in DIRNAME rather than in the system locale data base.  */
1353 char *
1354 bindtextdomain (const char *domainname, const char *dirname)
1355 {
1356   struct loaded_domain *domain = NULL;
1357   const char *catval_full;
1358   char *catval;
1359   char *fname;
1360
1361   /* DOMAINNAME is ignored.  We only support one domain.  */
1362
1363   /* DIRNAME is "$INSTALLDIR\share\locale".  */
1364
1365   /* First find out the category value.  */
1366   catval = NULL;
1367   catval_full = my_nl_locale_name ("LC_MESSAGES");
1368
1369   /* Normally, we would have to loop over all returned locales, and
1370      search for the right file.  See gettext intl/dcigettext.c for all
1371      the gory details.  Here, we only support the basic category, and
1372      ignore everything else.  */
1373   if (catval_full)
1374     {
1375       char *p;
1376
1377       catval = jnlib_malloc (strlen (catval_full) + 1);
1378       if (catval)
1379         {
1380           strcpy (catval, catval_full);
1381           p = strchr (catval, '_');
1382           if (p)
1383             *p = '\0';
1384         }
1385     }
1386   if (!catval)
1387     return NULL;
1388
1389   /* Now build the filename string.  The complete filename is this:
1390      DIRNAME + \ + CATVAL + \LC_MESSAGES\ + DOMAINNAME + .mo  */
1391   {
1392     int len = strlen (dirname) + 1 + strlen (catval) + 13
1393       + strlen (domainname) + 3 + 1;
1394     char *p;
1395
1396     fname = jnlib_malloc (len);
1397     if (!fname)
1398       {
1399         jnlib_free (catval);
1400         return NULL;
1401       }
1402
1403     p = fname;
1404     strcpy (p, dirname);
1405     p += strlen (dirname);
1406     *(p++) = '\\';
1407     strcpy (p, catval);
1408     p += strlen (catval);
1409     strcpy (p, "\\LC_MESSAGES\\");
1410     p += 13;
1411     strcpy (p, domainname);
1412     p += strlen (domainname);
1413     strcpy (p, ".mo");
1414   }
1415
1416   domain = load_domain (fname);
1417   jnlib_free (catval);
1418   jnlib_free (fname);
1419
1420   /* We should not be invoked twice, but this is how you would do
1421      it if it happened.  */
1422   if (the_domain)
1423     free_domain (the_domain);
1424   the_domain = domain;
1425
1426   /* For historic reasons we are not allowed to return a const char*. */
1427   return (char*)dirname;
1428 }
1429
1430
1431
1432 \f
1433 static const char *
1434 get_plural (const char *data, size_t datalen, unsigned long nplural)
1435 {
1436   const char *p;
1437   int idx;
1438
1439   /* We only support the Germanic rule.  */
1440   idx = (nplural == 1? 0 : 1);
1441
1442   for (; idx; idx--)
1443     {
1444       p = strchr (data, 0) + 1;
1445       if (p >= data+datalen)
1446         return "ERROR in GETTEXT (bad plural entry)";
1447       datalen -= (p-data);
1448       data = p;
1449     }
1450   return data;
1451 }
1452
1453
1454 static const char*
1455 get_string (struct loaded_domain *domain, uint32_t idx,
1456             int use_plural, unsigned long nplural)
1457 {
1458   struct overflow_space_s *os;
1459   const char *trans;  /* Pointer to the translated entry.  */
1460   size_t translen;    /* Length of that entry.  */
1461
1462   if (want_utf8)
1463     {
1464       trans = (domain->data
1465                + SWAPIT(domain->must_swap, domain->trans_tab[idx].offset));
1466       translen = SWAPIT(domain->must_swap, domain->trans_tab[idx].length);
1467     }
1468   else if (!domain->mapped[idx])
1469     {
1470       /* Not yet mapped.  Map from utf-8 to native encoding now.  */
1471       const char *p_utf8;
1472       size_t plen_utf8, buflen;
1473       char *buf;
1474
1475       p_utf8 = (domain->data
1476                 + SWAPIT(domain->must_swap, domain->trans_tab[idx].offset));
1477       plen_utf8 = SWAPIT(domain->must_swap, domain->trans_tab[idx].length);
1478
1479       buf = utf8_to_native (p_utf8, plen_utf8, &buflen);
1480       if (!buf)
1481         {
1482           trans = "ERROR in GETTEXT MALLOC";
1483           translen = 0;
1484         }
1485       else if (buflen <= plen_utf8 && buflen > 1)
1486         {
1487           /* Copy into the DATA_NATIVE area. */
1488           char *p_tmp;
1489
1490           p_tmp = (domain->data_native
1491                    + SWAPIT(domain->must_swap, domain->trans_tab[idx].offset));
1492           memcpy (p_tmp, buf, buflen);
1493           domain->mapped[idx] = buflen;
1494           trans = p_tmp;
1495           translen = buflen;
1496         }
1497       else
1498         {
1499           /* There is not enough space for the translation (or for
1500              whatever reason an empry string is used): Store it in the
1501              overflow_space and mark that in the mapped array.
1502              Because UTF-8 strings are in general longer than the
1503              Windows 2 byte encodings, we expect that this won't
1504              happen too often (if at all) and thus we use a linked
1505              list to manage this space. */
1506           os = jnlib_malloc (sizeof *os + buflen);
1507           if (os)
1508             {
1509               os->idx = idx;
1510               memcpy (os->d, buf, buflen);
1511               os->length = buflen;
1512               os->next = domain->overflow_space;
1513               domain->overflow_space = os;
1514               domain->mapped[idx] = 1;
1515               trans = os->d;
1516               translen = os->length;
1517             }
1518           else
1519             {
1520               trans = "ERROR in GETTEXT MALLOC";
1521               translen = 0;
1522             }
1523         }
1524       jnlib_free (buf);
1525     }
1526   else if (domain->mapped[idx] == 1)
1527     {
1528       /* The translated string is in the overflow_space. */
1529       for (os=domain->overflow_space; os; os = os->next)
1530         if (os->idx == idx)
1531           break;
1532       if (os)
1533         {
1534           trans = os->d;
1535           translen = os->length;
1536         }
1537       else
1538         {
1539           trans = "ERROR in GETTEXT (overflow space)\n";
1540           translen = 0;
1541         }
1542     }
1543   else
1544     {
1545       trans = (domain->data_native
1546                + SWAPIT(domain->must_swap, domain->trans_tab[idx].offset));
1547       translen = domain->mapped[idx];
1548     }
1549
1550   if (use_plural && translen)
1551     return get_plural (trans, translen, nplural);
1552   else
1553     return trans;
1554 }
1555
1556
1557 static const char *
1558 do_gettext (const char *msgid, const char *msgid2, unsigned long nplural)
1559 {
1560   struct loaded_domain *domain;
1561   uint32_t top, bottom, nstr;
1562
1563   if (!(domain = the_domain))
1564     goto not_found;
1565
1566   /* First try to use the hash table.  */
1567   if (domain->hash_size > 2 && domain->hash_tab)
1568     {
1569       /* Use the hashing table.  */
1570       uint32_t len = strlen (msgid);
1571       uint32_t hash_val = hash_string (msgid);
1572       uint32_t idx = hash_val % domain->hash_size;
1573       uint32_t incr = 1 + (hash_val % (domain->hash_size - 2));
1574
1575       while ( (nstr = SWAPIT (domain->must_swap, domain->hash_tab[idx])) )
1576         {
1577           nstr--;
1578           if (nstr < domain->nstrings
1579               && SWAPIT(domain->must_swap,
1580                         domain->orig_tab[nstr].length) >= len
1581               && !strcmp (msgid, (domain->data
1582                                   + SWAPIT(domain->must_swap,
1583                                            domain->orig_tab[nstr].offset))))
1584             {
1585               return get_string (domain, nstr, !!msgid2, nplural);
1586             }
1587
1588           if (idx >= domain->hash_size - incr)
1589             idx -= domain->hash_size - incr;
1590           else
1591             idx += incr;
1592         }
1593     }
1594
1595   /* Now we try the default method: binary search in the sorted array
1596      of messages.  */
1597   bottom = 0;
1598   top = domain->nstrings;
1599   while (bottom < top)
1600     {
1601       int cmp_val;
1602
1603       nstr = (bottom + top) / 2;
1604       cmp_val = strcmp (msgid, (domain->data
1605                                 + SWAPIT(domain->must_swap,
1606                                          domain->orig_tab[nstr].offset)));
1607       if (cmp_val < 0)
1608         top = nstr;
1609       else if (cmp_val > 0)
1610         bottom = nstr + 1;
1611       else
1612         return get_string (domain, nstr, !!msgid2, nplural);
1613     }
1614
1615  not_found:
1616   /* We use the standard Germanic rule if plural has been requested.  */
1617   return msgid2? (nplural == 1? msgid : msgid2) : msgid;
1618 }
1619
1620
1621 char *
1622 textdomain (const char *domainname)
1623 {
1624   /* For now, support only one domain.  */
1625   return (char*)domainname;
1626 }
1627
1628
1629 const char *
1630 gettext (const char *msgid)
1631 {
1632   return do_gettext (msgid, NULL, 0);
1633 }
1634
1635 char *
1636 dgettext (const char *domainname, const char *msgid)
1637 {
1638   (void)domainname;
1639
1640   /* For now, support only one domain.  */
1641   return (char*)do_gettext (msgid, NULL, 0);
1642 }
1643
1644 const char *
1645 ngettext (const char *msgid1, const char *msgid2, unsigned long int n)
1646 {
1647   /* We use the simple Germanic plural rule.  */
1648   return do_gettext (msgid1, msgid2, n);
1649 }
1650
1651
1652 /* Return the locale name as used by gettext.  The return value will
1653    never be NULL. */
1654 const char *
1655 gettext_localename (void)
1656 {
1657   const char *s;
1658
1659   s = my_nl_locale_name ("LC_MESSAGES");
1660   return s? s:"";
1661 }
1662
1663 void
1664 gettext_select_utf8 (int value)
1665 {
1666   want_utf8 = value;
1667 }
1668
1669
1670 #ifdef TEST
1671 int
1672 main (int argc, char **argv)
1673 {
1674   const char atext1[] =
1675     "Warning: You have entered an insecure passphrase.%%0A"
1676     "A passphrase should be at least %u character long.";
1677   const char atext2[] =
1678     "Warning: You have entered an insecure passphrase.%%0A"
1679     "A passphrase should be at least %u characters long.";
1680
1681   if (argc)
1682     {
1683       argc--;
1684       argv++;
1685     }
1686
1687   bindtextdomain ("gnupg2", "c:/programme/gnu/gnupg/share/locale");
1688
1689   printf ("locale is `%s'\n", gettext_localename ());
1690   fputs ("text with N=1:\n", stdout);
1691   fputs (ngettext (atext1, atext2, 1), stdout);
1692   fputs ("\n\ntext with N=2:\n", stdout);
1693   fputs (ngettext (atext1, atext2, 2), stdout);
1694   fputs ("\nready\n", stdout);
1695
1696   return 0;
1697 }
1698 /*
1699  * Local Variables:
1700  *  compile-command: "i586-mingw32msvc-gcc -DTEST -Wall -g w32-gettext.c"
1701  * End:
1702  */
1703 #endif /*TEST*/