1 /* setlocale() function that respects the locale chosen by the user.
2    Copyright (C) 2009, 2015-2016 Free Software Foundation, Inc.
3    Written by Bruno Haible <bruno@clisp.org>, 2009.
4 
5    This program is free software: you can redistribute it and/or modify
6    it under the terms of the GNU Lesser General Public License as published by
7    the Free Software Foundation; either version 2.1 of the License, or
8    (at your option) any later version.
9 
10    This program is distributed in the hope that it will be useful,
11    but WITHOUT ANY WARRANTY; without even the implied warranty of
12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13    GNU Lesser General Public License for more details.
14 
15    You should have received a copy of the GNU Lesser General Public License
16    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
17 
18 #ifdef HAVE_CONFIG_H
19 # include <config.h>
20 #endif
21 
22 /* Override setlocale() and newlocale() so that when the default locale is
23    requested (locale = "") and no relevant environment variable is set, the
24    locale chosen by the user is used.
25    This matters on MacOS X 10 and Windows.
26    See the comments in localename.c, function gl_locale_name_default.  */
27 
28 #include <locale.h>
29 #include <stdlib.h>
30 #include <string.h>
31 
32 /* When building a DLL, we must export some functions.  Note that because
33    the functions are only defined for binary backward compatibility, we
34    don't need to use __declspec(dllimport) in any case.  */
35 #if HAVE_VISIBILITY && BUILDING_DLL
36 # define DLL_EXPORTED __attribute__((__visibility__("default")))
37 #elif defined _MSC_VER && BUILDING_DLL
38 # define DLL_EXPORTED __declspec(dllexport)
39 #else
40 # define DLL_EXPORTED
41 #endif
42 
43 #include "gettextP.h"
44 
45 #if (defined __APPLE__ && defined __MACH__) || defined _WIN32 || defined __WIN32__ || defined __CYGWIN__
46 
47 # undef setlocale
48 # undef newlocale
49 
50 /* Return string representation of locale category CATEGORY.  */
51 static const char *
category_to_name(int category)52 category_to_name (int category)
53 {
54   const char *retval;
55 
56   switch (category)
57   {
58   case LC_COLLATE:
59     retval = "LC_COLLATE";
60     break;
61   case LC_CTYPE:
62     retval = "LC_CTYPE";
63     break;
64   case LC_MONETARY:
65     retval = "LC_MONETARY";
66     break;
67   case LC_NUMERIC:
68     retval = "LC_NUMERIC";
69     break;
70   case LC_TIME:
71     retval = "LC_TIME";
72     break;
73   case LC_MESSAGES:
74     retval = "LC_MESSAGES";
75     break;
76   default:
77     /* If you have a better idea for a default value let me know.  */
78     retval = "LC_XXX";
79   }
80 
81   return retval;
82 }
83 
84 # if (defined _WIN32 || defined __WIN32__) && ! defined __CYGWIN__
85 
86 /* The native Win32 setlocale() function expects locale names of the form
87    "German" or "German_Germany" or "DEU", but not "de" or "de_DE".  We need
88    to convert the names from the form with ISO 639 language code and ISO 3166
89    country code to the form with English names or with three-letter identifier.
90    The three-letter identifiers known by a Windows XP SP2 or SP3 are:
91      AFK  Afrikaans_South Africa.1252
92      ARA  Arabic_Saudi Arabia.1256
93      ARB  Arabic_Lebanon.1256
94      ARE  Arabic_Egypt.1256
95      ARG  Arabic_Algeria.1256
96      ARH  Arabic_Bahrain.1256
97      ARI  Arabic_Iraq.1256
98      ARJ  Arabic_Jordan.1256
99      ARK  Arabic_Kuwait.1256
100      ARL  Arabic_Libya.1256
101      ARM  Arabic_Morocco.1256
102      ARO  Arabic_Oman.1256
103      ARQ  Arabic_Qatar.1256
104      ARS  Arabic_Syria.1256
105      ART  Arabic_Tunisia.1256
106      ARU  Arabic_U.A.E..1256
107      ARY  Arabic_Yemen.1256
108      AZE  Azeri (Latin)_Azerbaijan.1254
109      BEL  Belarusian_Belarus.1251
110      BGR  Bulgarian_Bulgaria.1251
111      BSB  Bosnian_Bosnia and Herzegovina.1250
112      BSC  Bosnian (Cyrillic)_Bosnia and Herzegovina.1250  (wrong encoding!)
113      CAT  Catalan_Spain.1252
114      CHH  Chinese_Hong Kong S.A.R..950
115      CHI  Chinese_Singapore.936
116      CHS  Chinese_People's Republic of China.936
117      CHT  Chinese_Taiwan.950
118      CSY  Czech_Czech Republic.1250
119      CYM  Welsh_United Kingdom.1252
120      DAN  Danish_Denmark.1252
121      DEA  German_Austria.1252
122      DEC  German_Liechtenstein.1252
123      DEL  German_Luxembourg.1252
124      DES  German_Switzerland.1252
125      DEU  German_Germany.1252
126      ELL  Greek_Greece.1253
127      ENA  English_Australia.1252
128      ENB  English_Caribbean.1252
129      ENC  English_Canada.1252
130      ENG  English_United Kingdom.1252
131      ENI  English_Ireland.1252
132      ENJ  English_Jamaica.1252
133      ENL  English_Belize.1252
134      ENP  English_Republic of the Philippines.1252
135      ENS  English_South Africa.1252
136      ENT  English_Trinidad and Tobago.1252
137      ENU  English_United States.1252
138      ENW  English_Zimbabwe.1252
139      ENZ  English_New Zealand.1252
140      ESA  Spanish_Panama.1252
141      ESB  Spanish_Bolivia.1252
142      ESC  Spanish_Costa Rica.1252
143      ESD  Spanish_Dominican Republic.1252
144      ESE  Spanish_El Salvador.1252
145      ESF  Spanish_Ecuador.1252
146      ESG  Spanish_Guatemala.1252
147      ESH  Spanish_Honduras.1252
148      ESI  Spanish_Nicaragua.1252
149      ESL  Spanish_Chile.1252
150      ESM  Spanish_Mexico.1252
151      ESN  Spanish_Spain.1252
152      ESO  Spanish_Colombia.1252
153      ESP  Spanish_Spain.1252
154      ESR  Spanish_Peru.1252
155      ESS  Spanish_Argentina.1252
156      ESU  Spanish_Puerto Rico.1252
157      ESV  Spanish_Venezuela.1252
158      ESY  Spanish_Uruguay.1252
159      ESZ  Spanish_Paraguay.1252
160      ETI  Estonian_Estonia.1257
161      EUQ  Basque_Spain.1252
162      FAR  Farsi_Iran.1256
163      FIN  Finnish_Finland.1252
164      FOS  Faroese_Faroe Islands.1252
165      FPO  Filipino_Philippines.1252
166      FRA  French_France.1252
167      FRB  French_Belgium.1252
168      FRC  French_Canada.1252
169      FRL  French_Luxembourg.1252
170      FRM  French_Principality of Monaco.1252
171      FRS  French_Switzerland.1252
172      FYN  Frisian_Netherlands.1252
173      GLC  Galician_Spain.1252
174      HEB  Hebrew_Israel.1255
175      HRB  Croatian_Bosnia and Herzegovina.1250
176      HRV  Croatian_Croatia.1250
177      HUN  Hungarian_Hungary.1250
178      IND  Indonesian_Indonesia.1252
179      IRE  Irish_Ireland.1252
180      ISL  Icelandic_Iceland.1252
181      ITA  Italian_Italy.1252
182      ITS  Italian_Switzerland.1252
183      IUK  Inuktitut (Latin)_Canada.1252
184      JPN  Japanese_Japan.932
185      KKZ  Kazakh_Kazakhstan.1251
186      KOR  Korean_Korea.949
187      KYR  Kyrgyz_Kyrgyzstan.1251
188      LBX  Luxembourgish_Luxembourg.1252
189      LTH  Lithuanian_Lithuania.1257
190      LVI  Latvian_Latvia.1257
191      MKI  FYRO Macedonian_Former Yugoslav Republic of Macedonia.1251
192      MON  Mongolian_Mongolia.1251
193      MPD  Mapudungun_Chile.1252
194      MSB  Malay_Brunei Darussalam.1252
195      MSL  Malay_Malaysia.1252
196      MWK  Mohawk_Canada.1252
197      NLB  Dutch_Belgium.1252
198      NLD  Dutch_Netherlands.1252
199      NON  Norwegian-Nynorsk_Norway.1252
200      NOR  Norwegian (Bokmål)_Norway.1252
201      NSO  Northern Sotho_South Africa.1252
202      PLK  Polish_Poland.1250
203      PTB  Portuguese_Brazil.1252
204      PTG  Portuguese_Portugal.1252
205      QUB  Quechua_Bolivia.1252
206      QUE  Quechua_Ecuador.1252
207      QUP  Quechua_Peru.1252
208      RMC  Romansh_Switzerland.1252
209      ROM  Romanian_Romania.1250
210      RUS  Russian_Russia.1251
211      SKY  Slovak_Slovakia.1250
212      SLV  Slovenian_Slovenia.1250
213      SMA  Sami (Southern)_Norway.1252
214      SMB  Sami (Southern)_Sweden.1252
215      SME  Sami (Northern)_Norway.1252
216      SMF  Sami (Northern)_Sweden.1252
217      SMG  Sami (Northern)_Finland.1252
218      SMJ  Sami (Lule)_Norway.1252
219      SMK  Sami (Lule)_Sweden.1252
220      SMN  Sami (Inari)_Finland.1252
221      SMS  Sami (Skolt)_Finland.1252
222      SQI  Albanian_Albania.1250
223      SRB  Serbian (Cyrillic)_Serbia and Montenegro.1251
224      SRL  Serbian (Latin)_Serbia and Montenegro.1250
225      SRN  Serbian (Cyrillic)_Bosnia and Herzegovina.1251
226      SRS  Serbian (Latin)_Bosnia and Herzegovina.1250
227      SVE  Swedish_Sweden.1252
228      SVF  Swedish_Finland.1252
229      SWK  Swahili_Kenya.1252
230      THA  Thai_Thailand.874
231      TRK  Turkish_Turkey.1254
232      TSN  Tswana_South Africa.1252
233      TTT  Tatar_Russia.1251
234      UKR  Ukrainian_Ukraine.1251
235      URD  Urdu_Islamic Republic of Pakistan.1256
236      USA  English_United States.1252
237      UZB  Uzbek (Latin)_Uzbekistan.1254
238      VIT  Vietnamese_Viet Nam.1258
239      XHO  Xhosa_South Africa.1252
240      ZHH  Chinese_Hong Kong S.A.R..950
241      ZHI  Chinese_Singapore.936
242      ZHM  Chinese_Macau S.A.R..950
243      ZUL  Zulu_South Africa.1252
244  */
245 
246 /* Table from ISO 639 language code, optionally with country or script suffix,
247    to English name.
248    Keep in sync with the gl_locale_name_from_win32_LANGID function in
249    localename.c!  */
250 struct table_entry
251 {
252   const char *code;
253   const char *english;
254 };
255 static const struct table_entry language_table[] =
256   {
257     { "af", "Afrikaans" },
258     { "am", "Amharic" },
259     { "ar", "Arabic" },
260     { "arn", "Mapudungun" },
261     { "as", "Assamese" },
262     { "az@cyrillic", "Azeri (Cyrillic)" },
263     { "az@latin", "Azeri (Latin)" },
264     { "ba", "Bashkir" },
265     { "be", "Belarusian" },
266     { "ber", "Tamazight" },
267     { "ber@arabic", "Tamazight (Arabic)" },
268     { "ber@latin", "Tamazight (Latin)" },
269     { "bg", "Bulgarian" },
270     { "bin", "Edo" },
271     { "bn", "Bengali" },
272     { "bn_BD", "Bengali (Bangladesh)" },
273     { "bn_IN", "Bengali (India)" },
274     { "bnt", "Sutu" },
275     { "bo", "Tibetan" },
276     { "br", "Breton" },
277     { "bs", "BSB" }, /* "Bosnian (Latin)" */
278     { "bs@cyrillic", "BSC" }, /* Bosnian (Cyrillic) */
279     { "ca", "Catalan" },
280     { "chr", "Cherokee" },
281     { "co", "Corsican" },
282     { "cpe", "Hawaiian" },
283     { "cs", "Czech" },
284     { "cy", "Welsh" },
285     { "da", "Danish" },
286     { "de", "German" },
287     { "dsb", "Lower Sorbian" },
288     { "dv", "Divehi" },
289     { "el", "Greek" },
290     { "en", "English" },
291     { "es", "Spanish" },
292     { "et", "Estonian" },
293     { "eu", "Basque" },
294     { "fa", "Farsi" },
295     { "ff", "Fulfulde" },
296     { "fi", "Finnish" },
297     { "fo", "Faroese" }, /* "Faeroese" does not work */
298     { "fr", "French" },
299     { "fy", "Frisian" },
300     { "ga", "IRE" }, /* Gaelic (Ireland) */
301     { "gd", "Gaelic (Scotland)" },
302     { "gd", "Scottish Gaelic" },
303     { "gl", "Galician" },
304     { "gn", "Guarani" },
305     { "gsw", "Alsatian" },
306     { "gu", "Gujarati" },
307     { "ha", "Hausa" },
308     { "he", "Hebrew" },
309     { "hi", "Hindi" },
310     { "hr", "Croatian" },
311     { "hsb", "Upper Sorbian" },
312     { "hu", "Hungarian" },
313     { "hy", "Armenian" },
314     { "id", "Indonesian" },
315     { "ig", "Igbo" },
316     { "ii", "Yi" },
317     { "is", "Icelandic" },
318     { "it", "Italian" },
319     { "iu", "IUK" }, /* Inuktitut */
320     { "ja", "Japanese" },
321     { "ka", "Georgian" },
322     { "kk", "Kazakh" },
323     { "kl", "Greenlandic" },
324     { "km", "Cambodian" },
325     { "km", "Khmer" },
326     { "kn", "Kannada" },
327     { "ko", "Korean" },
328     { "kok", "Konkani" },
329     { "kr", "Kanuri" },
330     { "ks", "Kashmiri" },
331     { "ks_IN", "Kashmiri_India" },
332     { "ks_PK", "Kashmiri (Arabic)_Pakistan" },
333     { "ky", "Kyrgyz" },
334     { "la", "Latin" },
335     { "lb", "Luxembourgish" },
336     { "lo", "Lao" },
337     { "lt", "Lithuanian" },
338     { "lv", "Latvian" },
339     { "mi", "Maori" },
340     { "mk", "FYRO Macedonian" },
341     { "mk", "Macedonian" },
342     { "ml", "Malayalam" },
343     { "mn", "Mongolian" },
344     { "mni", "Manipuri" },
345     { "moh", "Mohawk" },
346     { "mr", "Marathi" },
347     { "ms", "Malay" },
348     { "mt", "Maltese" },
349     { "my", "Burmese" },
350     { "nb", "NOR" }, /* Norwegian Bokmål */
351     { "ne", "Nepali" },
352     { "nic", "Ibibio" },
353     { "nl", "Dutch" },
354     { "nn", "NON" }, /* Norwegian Nynorsk */
355     { "no", "Norwegian" },
356     { "nso", "Northern Sotho" },
357     { "nso", "Sepedi" },
358     { "oc", "Occitan" },
359     { "om", "Oromo" },
360     { "or", "Oriya" },
361     { "pa", "Punjabi" },
362     { "pap", "Papiamentu" },
363     { "pl", "Polish" },
364     { "prs", "Dari" },
365     { "ps", "Pashto" },
366     { "pt", "Portuguese" },
367     { "qu", "Quechua" },
368     { "qut", "K'iche'" },
369     { "rm", "Romansh" },
370     { "ro", "Romanian" },
371     { "ru", "Russian" },
372     { "rw", "Kinyarwanda" },
373     { "sa", "Sanskrit" },
374     { "sah", "Yakut" },
375     { "sd", "Sindhi" },
376     { "se", "Sami (Northern)" },
377     { "se", "Northern Sami" },
378     { "si", "Sinhalese" },
379     { "sk", "Slovak" },
380     { "sl", "Slovenian" },
381     { "sma", "Sami (Southern)" },
382     { "sma", "Southern Sami" },
383     { "smj", "Sami (Lule)" },
384     { "smj", "Lule Sami" },
385     { "smn", "Sami (Inari)" },
386     { "smn", "Inari Sami" },
387     { "sms", "Sami (Skolt)" },
388     { "sms", "Skolt Sami" },
389     { "so", "Somali" },
390     { "sq", "Albanian" },
391     { "sr", "Serbian (Latin)" },
392     { "sr@cyrillic", "SRB" }, /* Serbian (Cyrillic) */
393     { "sw", "Swahili" },
394     { "syr", "Syriac" },
395     { "ta", "Tamil" },
396     { "te", "Telugu" },
397     { "tg", "Tajik" },
398     { "th", "Thai" },
399     { "ti", "Tigrinya" },
400     { "tk", "Turkmen" },
401     { "tl", "Filipino" },
402     { "tn", "Tswana" },
403     { "tr", "Turkish" },
404     { "ts", "Tsonga" },
405     { "tt", "Tatar" },
406     { "ug", "Uighur" },
407     { "uk", "Ukrainian" },
408     { "ur", "Urdu" },
409     { "uz", "Uzbek" },
410     { "uz", "Uzbek (Latin)" },
411     { "uz@cyrillic", "Uzbek (Cyrillic)" },
412     { "ve", "Venda" },
413     { "vi", "Vietnamese" },
414     { "wen", "Sorbian" },
415     { "wo", "Wolof" },
416     { "xh", "Xhosa" },
417     { "yi", "Yiddish" },
418     { "yo", "Yoruba" },
419     { "zh", "Chinese" },
420     { "zu", "Zulu" }
421   };
422 
423 /* Table from ISO 3166 country code to English name.
424    Keep in sync with the gl_locale_name_from_win32_LANGID function in
425    localename.c!  */
426 static const struct table_entry country_table[] =
427   {
428     { "AE", "U.A.E." },
429     { "AF", "Afghanistan" },
430     { "AL", "Albania" },
431     { "AM", "Armenia" },
432     { "AN", "Netherlands Antilles" },
433     { "AR", "Argentina" },
434     { "AT", "Austria" },
435     { "AU", "Australia" },
436     { "AZ", "Azerbaijan" },
437     { "BA", "Bosnia and Herzegovina" },
438     { "BD", "Bangladesh" },
439     { "BE", "Belgium" },
440     { "BG", "Bulgaria" },
441     { "BH", "Bahrain" },
442     { "BN", "Brunei Darussalam" },
443     { "BO", "Bolivia" },
444     { "BR", "Brazil" },
445     { "BT", "Bhutan" },
446     { "BY", "Belarus" },
447     { "BZ", "Belize" },
448     { "CA", "Canada" },
449     { "CG", "Congo" },
450     { "CH", "Switzerland" },
451     { "CI", "Cote d'Ivoire" },
452     { "CL", "Chile" },
453     { "CM", "Cameroon" },
454     { "CN", "People's Republic of China" },
455     { "CO", "Colombia" },
456     { "CR", "Costa Rica" },
457     { "CS", "Serbia and Montenegro" },
458     { "CZ", "Czech Republic" },
459     { "DE", "Germany" },
460     { "DK", "Denmark" },
461     { "DO", "Dominican Republic" },
462     { "DZ", "Algeria" },
463     { "EC", "Ecuador" },
464     { "EE", "Estonia" },
465     { "EG", "Egypt" },
466     { "ER", "Eritrea" },
467     { "ES", "Spain" },
468     { "ET", "Ethiopia" },
469     { "FI", "Finland" },
470     { "FO", "Faroe Islands" },
471     { "FR", "France" },
472     { "GB", "United Kingdom" },
473     { "GD", "Caribbean" },
474     { "GE", "Georgia" },
475     { "GL", "Greenland" },
476     { "GR", "Greece" },
477     { "GT", "Guatemala" },
478     { "HK", "Hong Kong" },
479     { "HK", "Hong Kong S.A.R." },
480     { "HN", "Honduras" },
481     { "HR", "Croatia" },
482     { "HT", "Haiti" },
483     { "HU", "Hungary" },
484     { "ID", "Indonesia" },
485     { "IE", "Ireland" },
486     { "IL", "Israel" },
487     { "IN", "India" },
488     { "IQ", "Iraq" },
489     { "IR", "Iran" },
490     { "IS", "Iceland" },
491     { "IT", "Italy" },
492     { "JM", "Jamaica" },
493     { "JO", "Jordan" },
494     { "JP", "Japan" },
495     { "KE", "Kenya" },
496     { "KG", "Kyrgyzstan" },
497     { "KH", "Cambodia" },
498     { "KR", "South Korea" },
499     { "KW", "Kuwait" },
500     { "KZ", "Kazakhstan" },
501     { "LA", "Laos" },
502     { "LB", "Lebanon" },
503     { "LI", "Liechtenstein" },
504     { "LK", "Sri Lanka" },
505     { "LT", "Lithuania" },
506     { "LU", "Luxembourg" },
507     { "LV", "Latvia" },
508     { "LY", "Libya" },
509     { "MA", "Morocco" },
510     { "MC", "Principality of Monaco" },
511     { "MD", "Moldava" },
512     { "MD", "Moldova" },
513     { "ME", "Montenegro" },
514     { "MK", "Former Yugoslav Republic of Macedonia" },
515     { "ML", "Mali" },
516     { "MM", "Myanmar" },
517     { "MN", "Mongolia" },
518     { "MO", "Macau S.A.R." },
519     { "MT", "Malta" },
520     { "MV", "Maldives" },
521     { "MX", "Mexico" },
522     { "MY", "Malaysia" },
523     { "NG", "Nigeria" },
524     { "NI", "Nicaragua" },
525     { "NL", "Netherlands" },
526     { "NO", "Norway" },
527     { "NP", "Nepal" },
528     { "NZ", "New Zealand" },
529     { "OM", "Oman" },
530     { "PA", "Panama" },
531     { "PE", "Peru" },
532     { "PH", "Philippines" },
533     { "PK", "Islamic Republic of Pakistan" },
534     { "PL", "Poland" },
535     { "PR", "Puerto Rico" },
536     { "PT", "Portugal" },
537     { "PY", "Paraguay" },
538     { "QA", "Qatar" },
539     { "RE", "Reunion" },
540     { "RO", "Romania" },
541     { "RS", "Serbia" },
542     { "RU", "Russia" },
543     { "RW", "Rwanda" },
544     { "SA", "Saudi Arabia" },
545     { "SE", "Sweden" },
546     { "SG", "Singapore" },
547     { "SI", "Slovenia" },
548     { "SK", "Slovak" },
549     { "SN", "Senegal" },
550     { "SO", "Somalia" },
551     { "SR", "Suriname" },
552     { "SV", "El Salvador" },
553     { "SY", "Syria" },
554     { "TH", "Thailand" },
555     { "TJ", "Tajikistan" },
556     { "TM", "Turkmenistan" },
557     { "TN", "Tunisia" },
558     { "TR", "Turkey" },
559     { "TT", "Trinidad and Tobago" },
560     { "TW", "Taiwan" },
561     { "TZ", "Tanzania" },
562     { "UA", "Ukraine" },
563     { "US", "United States" },
564     { "UY", "Uruguay" },
565     { "VA", "Vatican" },
566     { "VE", "Venezuela" },
567     { "VN", "Viet Nam" },
568     { "YE", "Yemen" },
569     { "ZA", "South Africa" },
570     { "ZW", "Zimbabwe" }
571   };
572 
573 /* Given a string STRING, find the set of indices i such that TABLE[i].code is
574    the given STRING.  It is a range [lo,hi-1].  */
575 typedef struct { size_t lo; size_t hi; } range_t;
576 static void
search(const struct table_entry * table,size_t table_size,const char * string,range_t * result)577 search (const struct table_entry *table, size_t table_size, const char *string,
578         range_t *result)
579 {
580   /* The table is sorted.  Perform a binary search.  */
581   size_t hi = table_size;
582   size_t lo = 0;
583   while (lo < hi)
584     {
585       /* Invariant:
586          for i < lo, strcmp (table[i].code, string) < 0,
587          for i >= hi, strcmp (table[i].code, string) > 0.  */
588       size_t mid = (hi + lo) >> 1; /* >= lo, < hi */
589       int cmp = strcmp (table[mid].code, string);
590       if (cmp < 0)
591         lo = mid + 1;
592       else if (cmp > 0)
593         hi = mid;
594       else
595         {
596           /* Found an i with
597                strcmp (language_table[i].code, string) == 0.
598              Find the entire interval of such i.  */
599           {
600             size_t i;
601 
602             for (i = mid; i > lo; )
603               {
604                 i--;
605                 if (strcmp (table[i].code, string) < 0)
606                   {
607                     lo = i + 1;
608                     break;
609                   }
610               }
611           }
612           {
613             size_t i;
614 
615             for (i = mid; i < hi; i++)
616               {
617                 if (strcmp (table[i].code, string) > 0)
618                   {
619                     hi = i;
620                     break;
621                   }
622               }
623           }
624           /* The set of i with
625                strcmp (language_table[i].code, string) == 0
626              is the interval [lo, hi-1].  */
627           break;
628         }
629     }
630   result->lo = lo;
631   result->hi = hi;
632 }
633 
634 /* Like setlocale, but accept also locale names in the form ll or ll_CC,
635    where ll is an ISO 639 language code and CC is an ISO 3166 country code.  */
636 static char *
setlocale_unixlike(int category,const char * locale)637 setlocale_unixlike (int category, const char *locale)
638 {
639   char *result;
640   char llCC_buf[64];
641   char ll_buf[64];
642   char CC_buf[64];
643 
644   /* First, try setlocale with the original argument unchanged.  */
645   result = setlocale (category, locale);
646   if (result != NULL)
647     return result;
648 
649   /* Otherwise, assume the argument is in the form
650        language[_territory][.codeset][@modifier]
651      and try to map it using the tables.  */
652   if (strlen (locale) < sizeof (llCC_buf))
653     {
654       /* Second try: Remove the codeset part.  */
655       {
656         const char *p = locale;
657         char *q = llCC_buf;
658 
659         /* Copy the part before the dot.  */
660         for (; *p != '\0' && *p != '.'; p++, q++)
661           *q = *p;
662         if (*p == '.')
663           /* Skip the part up to the '@', if any.  */
664           for (; *p != '\0' && *p != '@'; p++)
665             ;
666         /* Copy the part starting with '@', if any.  */
667         for (; *p != '\0'; p++, q++)
668           *q = *p;
669         *q = '\0';
670       }
671       /* llCC_buf now contains
672            language[_territory][@modifier]
673        */
674       if (strcmp (llCC_buf, locale) != 0)
675         {
676           result = setlocale (category, llCC_buf);
677           if (result != NULL)
678             return result;
679         }
680       /* Look it up in language_table.  */
681       {
682         range_t range;
683         size_t i;
684 
685         search (language_table,
686                 sizeof (language_table) / sizeof (language_table[0]),
687                 llCC_buf,
688                 &range);
689 
690         for (i = range.lo; i < range.hi; i++)
691           {
692             /* Try the replacement in language_table[i].  */
693             result = setlocale (category, language_table[i].english);
694             if (result != NULL)
695               return result;
696           }
697       }
698       /* Split language[_territory][@modifier]
699          into  ll_buf = language[@modifier]
700          and   CC_buf = territory
701        */
702       {
703         const char *underscore = strchr (llCC_buf, '_');
704         if (underscore != NULL)
705           {
706             const char *territory_start = underscore + 1;
707             const char *territory_end = strchr (territory_start, '@');
708             if (territory_end == NULL)
709               territory_end = territory_start + strlen (territory_start);
710 
711             memcpy (ll_buf, llCC_buf, underscore - llCC_buf);
712             strcpy (ll_buf + (underscore - llCC_buf), territory_end);
713 
714             memcpy (CC_buf, territory_start, territory_end - territory_start);
715             CC_buf[territory_end - territory_start] = '\0';
716 
717             {
718               /* Look up ll_buf in language_table
719                  and CC_buf in country_table.  */
720               range_t language_range;
721 
722               search (language_table,
723                       sizeof (language_table) / sizeof (language_table[0]),
724                       ll_buf,
725                       &language_range);
726               if (language_range.lo < language_range.hi)
727                 {
728                   range_t country_range;
729 
730                   search (country_table,
731                           sizeof (country_table) / sizeof (country_table[0]),
732                           CC_buf,
733                           &country_range);
734                   if (country_range.lo < country_range.hi)
735                     {
736                       size_t i;
737                       size_t j;
738 
739                       for (i = language_range.lo; i < language_range.hi; i++)
740                         for (j = country_range.lo; j < country_range.hi; j++)
741                           {
742                             /* Concatenate the replacements.  */
743                             const char *part1 = language_table[i].english;
744                             size_t part1_len = strlen (part1);
745                             const char *part2 = country_table[j].english;
746                             size_t part2_len = strlen (part2) + 1;
747                             char buf[64+64];
748 
749                             if (!(part1_len + 1 + part2_len <= sizeof (buf)))
750                               abort ();
751                             memcpy (buf, part1, part1_len);
752                             buf[part1_len] = '_';
753                             memcpy (buf + part1_len + 1, part2, part2_len);
754 
755                             /* Try the concatenated replacements.  */
756                             result = setlocale (category, buf);
757                             if (result != NULL)
758                               return result;
759                           }
760                     }
761 
762                   /* Try omitting the country entirely.  This may set a locale
763                      corresponding to the wrong country, but is better than
764                      failing entirely.  */
765                   {
766                     size_t i;
767 
768                     for (i = language_range.lo; i < language_range.hi; i++)
769                       {
770                         /* Try only the language replacement.  */
771                         result =
772                           setlocale (category, language_table[i].english);
773                         if (result != NULL)
774                           return result;
775                       }
776                   }
777                 }
778             }
779           }
780       }
781     }
782 
783   /* Failed.  */
784   return NULL;
785 }
786 
787 # else
788 #  define setlocale_unixlike setlocale
789 # endif
790 
791 # if LC_MESSAGES == 1729
792 
793 /* The system does not store an LC_MESSAGES locale category.  Do it here.  */
794 static char lc_messages_name[64] = "C";
795 
796 /* Like setlocale, but support also LC_MESSAGES.  */
797 static char *
setlocale_single(int category,const char * locale)798 setlocale_single (int category, const char *locale)
799 {
800   if (category == LC_MESSAGES)
801     {
802       if (locale != NULL)
803         {
804           lc_messages_name[sizeof (lc_messages_name) - 1] = '\0';
805           strncpy (lc_messages_name, locale, sizeof (lc_messages_name) - 1);
806         }
807       return lc_messages_name;
808     }
809   else
810     return setlocale_unixlike (category, locale);
811 }
812 
813 # else
814 #  define setlocale_single setlocale_unixlike
815 # endif
816 
817 DLL_EXPORTED
818 char *
libintl_setlocale(int category,const char * locale)819 libintl_setlocale (int category, const char *locale)
820 {
821   if (locale != NULL && locale[0] == '\0')
822     {
823       /* A request to the set the current locale to the default locale.  */
824       if (category == LC_ALL)
825         {
826           /* Set LC_CTYPE first.  Then the other categories.  */
827           static int const categories[] =
828             {
829               LC_NUMERIC,
830               LC_TIME,
831               LC_COLLATE,
832               LC_MONETARY,
833               LC_MESSAGES
834             };
835           char *saved_locale;
836           const char *base_name;
837           unsigned int i;
838 
839           /* Back up the old locale, in case one of the steps fails.  */
840           saved_locale = setlocale (LC_ALL, NULL);
841           if (saved_locale == NULL)
842             return NULL;
843           saved_locale = strdup (saved_locale);
844           if (saved_locale == NULL)
845             return NULL;
846 
847           /* Set LC_CTYPE category.  Set all other categories (except possibly
848              LC_MESSAGES) to the same value in the same call; this is likely to
849              save calls.  */
850           base_name =
851             gl_locale_name_environ (LC_CTYPE, category_to_name (LC_CTYPE));
852           if (base_name == NULL)
853             base_name = gl_locale_name_default ();
854 
855           if (setlocale_unixlike (LC_ALL, base_name) == NULL)
856             goto fail;
857 # if (defined _WIN32 || defined __WIN32__) && ! defined __CYGWIN__
858           /* On native Windows, setlocale(LC_ALL,...) may succeed but set the
859              LC_CTYPE category to an invalid value ("C") when it does not
860              support the specified encoding.  Report a failure instead.  */
861           if (strchr (base_name, '.') != NULL
862               && strcmp (setlocale (LC_CTYPE, NULL), "C") == 0)
863             goto fail;
864 # endif
865 
866           for (i = 0; i < sizeof (categories) / sizeof (categories[0]); i++)
867             {
868               int cat = categories[i];
869               const char *name;
870 
871               name = gl_locale_name_environ (cat, category_to_name (cat));
872               if (name == NULL)
873                 name = gl_locale_name_default ();
874 
875               /* If name is the same as base_name, it has already been set
876                  through the setlocale call before the loop.  */
877               if (strcmp (name, base_name) != 0
878 # if LC_MESSAGES == 1729
879                   || cat == LC_MESSAGES
880 # endif
881                  )
882                 if (setlocale_single (cat, name) == NULL)
883                   goto fail;
884             }
885 
886           /* All steps were successful.  */
887           ++_nl_msg_cat_cntr;
888           free (saved_locale);
889           return setlocale (LC_ALL, NULL);
890 
891         fail:
892           if (saved_locale[0] != '\0') /* don't risk an endless recursion */
893             setlocale (LC_ALL, saved_locale);
894           free (saved_locale);
895           return NULL;
896         }
897       else
898         {
899           char *result;
900           const char *name =
901             gl_locale_name_environ (category, category_to_name (category));
902           if (name == NULL)
903             name = gl_locale_name_default ();
904 
905           result = setlocale_single (category, name);
906           if (result != NULL)
907             ++_nl_msg_cat_cntr;
908           return result;
909         }
910     }
911   else
912     {
913 # if (defined _WIN32 || defined __WIN32__) && ! defined __CYGWIN__
914       if (category == LC_ALL && locale != NULL && strchr (locale, '.') != NULL)
915         {
916           char *saved_locale;
917 
918           /* Back up the old locale.  */
919           saved_locale = setlocale (LC_ALL, NULL);
920           if (saved_locale == NULL)
921             return NULL;
922           saved_locale = strdup (saved_locale);
923           if (saved_locale == NULL)
924             return NULL;
925 
926           if (setlocale_unixlike (LC_ALL, locale) == NULL)
927             {
928               free (saved_locale);
929               return NULL;
930             }
931 
932           /* On native Windows, setlocale(LC_ALL,...) may succeed but set the
933              LC_CTYPE category to an invalid value ("C") when it does not
934              support the specified encoding.  Report a failure instead.  */
935           if (strcmp (setlocale (LC_CTYPE, NULL), "C") == 0)
936             {
937               if (saved_locale[0] != '\0') /* don't risk an endless recursion */
938                 setlocale (LC_ALL, saved_locale);
939               free (saved_locale);
940               return NULL;
941             }
942 
943           /* It was really successful.  */
944           ++_nl_msg_cat_cntr;
945           free (saved_locale);
946           return setlocale (LC_ALL, NULL);
947         }
948       else
949 # endif
950         {
951           char *result = setlocale_single (category, locale);
952           if (result != NULL)
953             ++_nl_msg_cat_cntr;
954           return result;
955         }
956     }
957 }
958 
959 # if HAVE_NEWLOCALE
960 
961 DLL_EXPORTED
962 locale_t
libintl_newlocale(int category_mask,const char * locale,locale_t base)963 libintl_newlocale (int category_mask, const char *locale, locale_t base)
964 {
965   if (category_mask != 0 && locale != NULL && locale[0] == '\0')
966     {
967       /* A request to construct a locale_t object that refers to the default
968          locale.  */
969 
970       /* Set LC_CTYPE first.  Then the other categories.  */
971       static struct { int cat; int mask; } const categories[] =
972         {
973           { LC_CTYPE,    LC_CTYPE_MASK },
974           { LC_NUMERIC,  LC_NUMERIC_MASK },
975           { LC_TIME,     LC_TIME_MASK },
976           { LC_COLLATE,  LC_COLLATE_MASK },
977           { LC_MONETARY, LC_MONETARY_MASK },
978           { LC_MESSAGES, LC_MESSAGES_MASK }
979         };
980 
981       locale_t orig_base = base;
982 
983       if ((LC_ALL_MASK & ~category_mask) == 0)
984         {
985           const char *base_name;
986           unsigned int i;
987 
988           /* Set LC_CTYPE category.  Set all other categories (except possibly
989              LC_MESSAGES) to the same value in the same call; this is likely to
990              save calls.  */
991           base_name =
992             gl_locale_name_environ (LC_CTYPE, category_to_name (LC_CTYPE));
993           if (base_name == NULL)
994             base_name = gl_locale_name_default ();
995 
996           base = newlocale (LC_ALL_MASK, base_name, base);
997           if (base == NULL)
998             return NULL;
999 
1000           for (i = 1; i < sizeof (categories) / sizeof (categories[0]); i++)
1001             {
1002               int category = categories[i].cat;
1003               int category_mask = categories[i].mask;
1004               const char *name;
1005 
1006               name =
1007                 gl_locale_name_environ (category, category_to_name (category));
1008               if (name == NULL)
1009                 name = gl_locale_name_default ();
1010 
1011               /* If name is the same as base_name, it has already been set
1012                  through the setlocale call before the loop.  */
1013               if (strcmp (name, base_name) != 0)
1014                 {
1015                   locale_t copy = newlocale (category_mask, name, base);
1016                   if (copy == NULL)
1017                     goto fail;
1018                   /* No need to call freelocale (base) if copy != base; the
1019                      newlocale function already takes care of doing it.  */
1020                   base = copy;
1021                 }
1022             }
1023         }
1024       else
1025         {
1026           unsigned int i;
1027 
1028           for (i = 0; i < sizeof (categories) / sizeof (categories[0]); i++)
1029             {
1030               int cat_mask = categories[i].mask;
1031 
1032               if ((category_mask & cat_mask) != 0)
1033                 {
1034                   int cat = categories[i].cat;
1035                   const char *name;
1036                   locale_t copy;
1037 
1038                   name = gl_locale_name_environ (cat, category_to_name (cat));
1039                   if (name == NULL)
1040                     name = gl_locale_name_default ();
1041 
1042                   copy = newlocale (cat_mask, name, base);
1043                   if (copy == NULL)
1044                     goto fail;
1045                   /* No need to call freelocale (base) if copy != base; the
1046                      newlocale function already takes care of doing it.  */
1047                   base = copy;
1048                 }
1049             }
1050         }
1051 
1052       /* All steps were successful.  */
1053       return base;
1054 
1055     fail:
1056       if (base != NULL && orig_base == NULL)
1057         {
1058           int saved_errno = errno;
1059           freelocale (base);
1060           errno = saved_errno;
1061         }
1062       return NULL;
1063     }
1064   else
1065     return newlocale (category_mask, locale, base);
1066 }
1067 
1068 # endif
1069 
1070 #endif
1071