• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* setlocale() function that respects the locale chosen by the user.
2    Copyright (C) 2009, 2011, 2013, 2018-2019 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 <https://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 <stdio.h>
30 #include <stdlib.h>
31 #include <string.h>
32 
33 /* When building a DLL, we must export some functions.  Note that because
34    the functions are only defined for binary backward compatibility, we
35    don't need to use __declspec(dllimport) in any case.  */
36 #if HAVE_VISIBILITY && BUILDING_DLL
37 # define DLL_EXPORTED __attribute__((__visibility__("default")))
38 #elif defined _MSC_VER && BUILDING_DLL
39 # define DLL_EXPORTED __declspec(dllexport)
40 #else
41 # define DLL_EXPORTED
42 #endif
43 
44 #include "gettextP.h"
45 
46 #if HAVE_CFLOCALECOPYPREFERREDLANGUAGES || HAVE_CFPREFERENCESCOPYAPPVALUE
47 # if HAVE_CFLOCALECOPYPREFERREDLANGUAGES
48 #  include <CoreFoundation/CFLocale.h>
49 # elif HAVE_CFPREFERENCESCOPYAPPVALUE
50 #  include <CoreFoundation/CFPreferences.h>
51 # endif
52 # include <CoreFoundation/CFPropertyList.h>
53 # include <CoreFoundation/CFArray.h>
54 # include <CoreFoundation/CFString.h>
55 #endif
56 
57 #if (defined __APPLE__ && defined __MACH__) || defined _WIN32 || defined __CYGWIN__
58 
59 # undef setlocale
60 # undef newlocale
61 
62 /* Return string representation of locale category CATEGORY.  */
63 static const char *
category_to_name(int category)64 category_to_name (int category)
65 {
66   const char *retval;
67 
68   switch (category)
69   {
70   case LC_COLLATE:
71     retval = "LC_COLLATE";
72     break;
73   case LC_CTYPE:
74     retval = "LC_CTYPE";
75     break;
76   case LC_MONETARY:
77     retval = "LC_MONETARY";
78     break;
79   case LC_NUMERIC:
80     retval = "LC_NUMERIC";
81     break;
82   case LC_TIME:
83     retval = "LC_TIME";
84     break;
85   case LC_MESSAGES:
86     retval = "LC_MESSAGES";
87     break;
88   default:
89     /* If you have a better idea for a default value let me know.  */
90     retval = "LC_XXX";
91   }
92 
93   return retval;
94 }
95 
96 # if defined _WIN32 && ! defined __CYGWIN__
97 
98 /* The native Windows setlocale() function expects locale names of the form
99    "German" or "German_Germany" or "DEU", but not "de" or "de_DE".  We need
100    to convert the names from the form with ISO 639 language code and ISO 3166
101    country code to the form with English names or with three-letter identifier.
102    The three-letter identifiers known by a Windows XP SP2 or SP3 are:
103      AFK  Afrikaans_South Africa.1252
104      ARA  Arabic_Saudi Arabia.1256
105      ARB  Arabic_Lebanon.1256
106      ARE  Arabic_Egypt.1256
107      ARG  Arabic_Algeria.1256
108      ARH  Arabic_Bahrain.1256
109      ARI  Arabic_Iraq.1256
110      ARJ  Arabic_Jordan.1256
111      ARK  Arabic_Kuwait.1256
112      ARL  Arabic_Libya.1256
113      ARM  Arabic_Morocco.1256
114      ARO  Arabic_Oman.1256
115      ARQ  Arabic_Qatar.1256
116      ARS  Arabic_Syria.1256
117      ART  Arabic_Tunisia.1256
118      ARU  Arabic_U.A.E..1256
119      ARY  Arabic_Yemen.1256
120      AZE  Azeri (Latin)_Azerbaijan.1254
121      BEL  Belarusian_Belarus.1251
122      BGR  Bulgarian_Bulgaria.1251
123      BSB  Bosnian_Bosnia and Herzegovina.1250
124      BSC  Bosnian (Cyrillic)_Bosnia and Herzegovina.1250  (wrong encoding!)
125      CAT  Catalan_Spain.1252
126      CHH  Chinese_Hong Kong S.A.R..950
127      CHI  Chinese_Singapore.936
128      CHS  Chinese_People's Republic of China.936
129      CHT  Chinese_Taiwan.950
130      CSY  Czech_Czech Republic.1250
131      CYM  Welsh_United Kingdom.1252
132      DAN  Danish_Denmark.1252
133      DEA  German_Austria.1252
134      DEC  German_Liechtenstein.1252
135      DEL  German_Luxembourg.1252
136      DES  German_Switzerland.1252
137      DEU  German_Germany.1252
138      ELL  Greek_Greece.1253
139      ENA  English_Australia.1252
140      ENB  English_Caribbean.1252
141      ENC  English_Canada.1252
142      ENG  English_United Kingdom.1252
143      ENI  English_Ireland.1252
144      ENJ  English_Jamaica.1252
145      ENL  English_Belize.1252
146      ENP  English_Republic of the Philippines.1252
147      ENS  English_South Africa.1252
148      ENT  English_Trinidad and Tobago.1252
149      ENU  English_United States.1252
150      ENW  English_Zimbabwe.1252
151      ENZ  English_New Zealand.1252
152      ESA  Spanish_Panama.1252
153      ESB  Spanish_Bolivia.1252
154      ESC  Spanish_Costa Rica.1252
155      ESD  Spanish_Dominican Republic.1252
156      ESE  Spanish_El Salvador.1252
157      ESF  Spanish_Ecuador.1252
158      ESG  Spanish_Guatemala.1252
159      ESH  Spanish_Honduras.1252
160      ESI  Spanish_Nicaragua.1252
161      ESL  Spanish_Chile.1252
162      ESM  Spanish_Mexico.1252
163      ESN  Spanish_Spain.1252
164      ESO  Spanish_Colombia.1252
165      ESP  Spanish_Spain.1252
166      ESR  Spanish_Peru.1252
167      ESS  Spanish_Argentina.1252
168      ESU  Spanish_Puerto Rico.1252
169      ESV  Spanish_Venezuela.1252
170      ESY  Spanish_Uruguay.1252
171      ESZ  Spanish_Paraguay.1252
172      ETI  Estonian_Estonia.1257
173      EUQ  Basque_Spain.1252
174      FAR  Farsi_Iran.1256
175      FIN  Finnish_Finland.1252
176      FOS  Faroese_Faroe Islands.1252
177      FPO  Filipino_Philippines.1252
178      FRA  French_France.1252
179      FRB  French_Belgium.1252
180      FRC  French_Canada.1252
181      FRL  French_Luxembourg.1252
182      FRM  French_Principality of Monaco.1252
183      FRS  French_Switzerland.1252
184      FYN  Frisian_Netherlands.1252
185      GLC  Galician_Spain.1252
186      HEB  Hebrew_Israel.1255
187      HRB  Croatian_Bosnia and Herzegovina.1250
188      HRV  Croatian_Croatia.1250
189      HUN  Hungarian_Hungary.1250
190      IND  Indonesian_Indonesia.1252
191      IRE  Irish_Ireland.1252
192      ISL  Icelandic_Iceland.1252
193      ITA  Italian_Italy.1252
194      ITS  Italian_Switzerland.1252
195      IUK  Inuktitut (Latin)_Canada.1252
196      JPN  Japanese_Japan.932
197      KKZ  Kazakh_Kazakhstan.1251
198      KOR  Korean_Korea.949
199      KYR  Kyrgyz_Kyrgyzstan.1251
200      LBX  Luxembourgish_Luxembourg.1252
201      LTH  Lithuanian_Lithuania.1257
202      LVI  Latvian_Latvia.1257
203      MKI  FYRO Macedonian_Former Yugoslav Republic of Macedonia.1251
204      MON  Mongolian_Mongolia.1251
205      MPD  Mapudungun_Chile.1252
206      MSB  Malay_Brunei Darussalam.1252
207      MSL  Malay_Malaysia.1252
208      MWK  Mohawk_Canada.1252
209      NLB  Dutch_Belgium.1252
210      NLD  Dutch_Netherlands.1252
211      NON  Norwegian-Nynorsk_Norway.1252
212      NOR  Norwegian (Bokmål)_Norway.1252
213      NSO  Northern Sotho_South Africa.1252
214      PLK  Polish_Poland.1250
215      PTB  Portuguese_Brazil.1252
216      PTG  Portuguese_Portugal.1252
217      QUB  Quechua_Bolivia.1252
218      QUE  Quechua_Ecuador.1252
219      QUP  Quechua_Peru.1252
220      RMC  Romansh_Switzerland.1252
221      ROM  Romanian_Romania.1250
222      RUS  Russian_Russia.1251
223      SKY  Slovak_Slovakia.1250
224      SLV  Slovenian_Slovenia.1250
225      SMA  Sami (Southern)_Norway.1252
226      SMB  Sami (Southern)_Sweden.1252
227      SME  Sami (Northern)_Norway.1252
228      SMF  Sami (Northern)_Sweden.1252
229      SMG  Sami (Northern)_Finland.1252
230      SMJ  Sami (Lule)_Norway.1252
231      SMK  Sami (Lule)_Sweden.1252
232      SMN  Sami (Inari)_Finland.1252
233      SMS  Sami (Skolt)_Finland.1252
234      SQI  Albanian_Albania.1250
235      SRB  Serbian (Cyrillic)_Serbia and Montenegro.1251
236      SRL  Serbian (Latin)_Serbia and Montenegro.1250
237      SRN  Serbian (Cyrillic)_Bosnia and Herzegovina.1251
238      SRS  Serbian (Latin)_Bosnia and Herzegovina.1250
239      SVE  Swedish_Sweden.1252
240      SVF  Swedish_Finland.1252
241      SWK  Swahili_Kenya.1252
242      THA  Thai_Thailand.874
243      TRK  Turkish_Turkey.1254
244      TSN  Tswana_South Africa.1252
245      TTT  Tatar_Russia.1251
246      UKR  Ukrainian_Ukraine.1251
247      URD  Urdu_Islamic Republic of Pakistan.1256
248      USA  English_United States.1252
249      UZB  Uzbek (Latin)_Uzbekistan.1254
250      VIT  Vietnamese_Viet Nam.1258
251      XHO  Xhosa_South Africa.1252
252      ZHH  Chinese_Hong Kong S.A.R..950
253      ZHI  Chinese_Singapore.936
254      ZHM  Chinese_Macau S.A.R..950
255      ZUL  Zulu_South Africa.1252
256  */
257 
258 /* Table from ISO 639 language code, optionally with country or script suffix,
259    to English name.
260    Keep in sync with the gl_locale_name_from_win32_LANGID function in
261    localename.c!  */
262 struct table_entry
263 {
264   const char *code;
265   const char *english;
266 };
267 static const struct table_entry language_table[] =
268   {
269     { "af", "Afrikaans" },
270     { "am", "Amharic" },
271     { "ar", "Arabic" },
272     { "arn", "Mapudungun" },
273     { "as", "Assamese" },
274     { "az@cyrillic", "Azeri (Cyrillic)" },
275     { "az@latin", "Azeri (Latin)" },
276     { "ba", "Bashkir" },
277     { "be", "Belarusian" },
278     { "ber", "Tamazight" },
279     { "ber@arabic", "Tamazight (Arabic)" },
280     { "ber@latin", "Tamazight (Latin)" },
281     { "bg", "Bulgarian" },
282     { "bin", "Edo" },
283     { "bn", "Bengali" },
284     { "bn_BD", "Bengali (Bangladesh)" },
285     { "bn_IN", "Bengali (India)" },
286     { "bnt", "Sutu" },
287     { "bo", "Tibetan" },
288     { "br", "Breton" },
289     { "bs", "BSB" }, /* "Bosnian (Latin)" */
290     { "bs@cyrillic", "BSC" }, /* Bosnian (Cyrillic) */
291     { "ca", "Catalan" },
292     { "chr", "Cherokee" },
293     { "co", "Corsican" },
294     { "cpe", "Hawaiian" },
295     { "cs", "Czech" },
296     { "cy", "Welsh" },
297     { "da", "Danish" },
298     { "de", "German" },
299     { "dsb", "Lower Sorbian" },
300     { "dv", "Divehi" },
301     { "el", "Greek" },
302     { "en", "English" },
303     { "es", "Spanish" },
304     { "et", "Estonian" },
305     { "eu", "Basque" },
306     { "fa", "Farsi" },
307     { "ff", "Fulfulde" },
308     { "fi", "Finnish" },
309     { "fo", "Faroese" }, /* "Faeroese" does not work */
310     { "fr", "French" },
311     { "fy", "Frisian" },
312     { "ga", "IRE" }, /* Gaelic (Ireland) */
313     { "gd", "Gaelic (Scotland)" },
314     { "gd", "Scottish Gaelic" },
315     { "gl", "Galician" },
316     { "gn", "Guarani" },
317     { "gsw", "Alsatian" },
318     { "gu", "Gujarati" },
319     { "ha", "Hausa" },
320     { "he", "Hebrew" },
321     { "hi", "Hindi" },
322     { "hr", "Croatian" },
323     { "hsb", "Upper Sorbian" },
324     { "hu", "Hungarian" },
325     { "hy", "Armenian" },
326     { "id", "Indonesian" },
327     { "ig", "Igbo" },
328     { "ii", "Yi" },
329     { "is", "Icelandic" },
330     { "it", "Italian" },
331     { "iu", "IUK" }, /* Inuktitut */
332     { "ja", "Japanese" },
333     { "ka", "Georgian" },
334     { "kk", "Kazakh" },
335     { "kl", "Greenlandic" },
336     { "km", "Cambodian" },
337     { "km", "Khmer" },
338     { "kn", "Kannada" },
339     { "ko", "Korean" },
340     { "kok", "Konkani" },
341     { "kr", "Kanuri" },
342     { "ks", "Kashmiri" },
343     { "ks_IN", "Kashmiri_India" },
344     { "ks_PK", "Kashmiri (Arabic)_Pakistan" },
345     { "ky", "Kyrgyz" },
346     { "la", "Latin" },
347     { "lb", "Luxembourgish" },
348     { "lo", "Lao" },
349     { "lt", "Lithuanian" },
350     { "lv", "Latvian" },
351     { "mi", "Maori" },
352     { "mk", "FYRO Macedonian" },
353     { "mk", "Macedonian" },
354     { "ml", "Malayalam" },
355     { "mn", "Mongolian" },
356     { "mni", "Manipuri" },
357     { "moh", "Mohawk" },
358     { "mr", "Marathi" },
359     { "ms", "Malay" },
360     { "mt", "Maltese" },
361     { "my", "Burmese" },
362     { "nb", "NOR" }, /* Norwegian Bokmål */
363     { "ne", "Nepali" },
364     { "nic", "Ibibio" },
365     { "nl", "Dutch" },
366     { "nn", "NON" }, /* Norwegian Nynorsk */
367     { "no", "Norwegian" },
368     { "nso", "Northern Sotho" },
369     { "nso", "Sepedi" },
370     { "oc", "Occitan" },
371     { "om", "Oromo" },
372     { "or", "Oriya" },
373     { "pa", "Punjabi" },
374     { "pap", "Papiamentu" },
375     { "pl", "Polish" },
376     { "prs", "Dari" },
377     { "ps", "Pashto" },
378     { "pt", "Portuguese" },
379     { "qu", "Quechua" },
380     { "qut", "K'iche'" },
381     { "rm", "Romansh" },
382     { "ro", "Romanian" },
383     { "ru", "Russian" },
384     { "rw", "Kinyarwanda" },
385     { "sa", "Sanskrit" },
386     { "sah", "Yakut" },
387     { "sd", "Sindhi" },
388     { "se", "Sami (Northern)" },
389     { "se", "Northern Sami" },
390     { "si", "Sinhalese" },
391     { "sk", "Slovak" },
392     { "sl", "Slovenian" },
393     { "sma", "Sami (Southern)" },
394     { "sma", "Southern Sami" },
395     { "smj", "Sami (Lule)" },
396     { "smj", "Lule Sami" },
397     { "smn", "Sami (Inari)" },
398     { "smn", "Inari Sami" },
399     { "sms", "Sami (Skolt)" },
400     { "sms", "Skolt Sami" },
401     { "so", "Somali" },
402     { "sq", "Albanian" },
403     { "sr", "Serbian (Latin)" },
404     { "sr@cyrillic", "SRB" }, /* Serbian (Cyrillic) */
405     { "sv", "Swedish" },
406     { "sw", "Swahili" },
407     { "syr", "Syriac" },
408     { "ta", "Tamil" },
409     { "te", "Telugu" },
410     { "tg", "Tajik" },
411     { "th", "Thai" },
412     { "ti", "Tigrinya" },
413     { "tk", "Turkmen" },
414     { "tl", "Filipino" },
415     { "tn", "Tswana" },
416     { "tr", "Turkish" },
417     { "ts", "Tsonga" },
418     { "tt", "Tatar" },
419     { "ug", "Uighur" },
420     { "uk", "Ukrainian" },
421     { "ur", "Urdu" },
422     { "uz", "Uzbek" },
423     { "uz", "Uzbek (Latin)" },
424     { "uz@cyrillic", "Uzbek (Cyrillic)" },
425     { "ve", "Venda" },
426     { "vi", "Vietnamese" },
427     { "wen", "Sorbian" },
428     { "wo", "Wolof" },
429     { "xh", "Xhosa" },
430     { "yi", "Yiddish" },
431     { "yo", "Yoruba" },
432     { "zh", "Chinese" },
433     { "zu", "Zulu" }
434   };
435 
436 /* Table from ISO 3166 country code to English name.
437    Keep in sync with the gl_locale_name_from_win32_LANGID function in
438    localename.c!  */
439 static const struct table_entry country_table[] =
440   {
441     { "AE", "U.A.E." },
442     { "AF", "Afghanistan" },
443     { "AL", "Albania" },
444     { "AM", "Armenia" },
445     { "AN", "Netherlands Antilles" },
446     { "AR", "Argentina" },
447     { "AT", "Austria" },
448     { "AU", "Australia" },
449     { "AZ", "Azerbaijan" },
450     { "BA", "Bosnia and Herzegovina" },
451     { "BD", "Bangladesh" },
452     { "BE", "Belgium" },
453     { "BG", "Bulgaria" },
454     { "BH", "Bahrain" },
455     { "BN", "Brunei Darussalam" },
456     { "BO", "Bolivia" },
457     { "BR", "Brazil" },
458     { "BT", "Bhutan" },
459     { "BY", "Belarus" },
460     { "BZ", "Belize" },
461     { "CA", "Canada" },
462     { "CG", "Congo" },
463     { "CH", "Switzerland" },
464     { "CI", "Cote d'Ivoire" },
465     { "CL", "Chile" },
466     { "CM", "Cameroon" },
467     { "CN", "People's Republic of China" },
468     { "CO", "Colombia" },
469     { "CR", "Costa Rica" },
470     { "CS", "Serbia and Montenegro" },
471     { "CZ", "Czech Republic" },
472     { "DE", "Germany" },
473     { "DK", "Denmark" },
474     { "DO", "Dominican Republic" },
475     { "DZ", "Algeria" },
476     { "EC", "Ecuador" },
477     { "EE", "Estonia" },
478     { "EG", "Egypt" },
479     { "ER", "Eritrea" },
480     { "ES", "Spain" },
481     { "ET", "Ethiopia" },
482     { "FI", "Finland" },
483     { "FO", "Faroe Islands" },
484     { "FR", "France" },
485     { "GB", "United Kingdom" },
486     { "GD", "Caribbean" },
487     { "GE", "Georgia" },
488     { "GL", "Greenland" },
489     { "GR", "Greece" },
490     { "GT", "Guatemala" },
491     { "HK", "Hong Kong" },
492     { "HK", "Hong Kong S.A.R." },
493     { "HN", "Honduras" },
494     { "HR", "Croatia" },
495     { "HT", "Haiti" },
496     { "HU", "Hungary" },
497     { "ID", "Indonesia" },
498     { "IE", "Ireland" },
499     { "IL", "Israel" },
500     { "IN", "India" },
501     { "IQ", "Iraq" },
502     { "IR", "Iran" },
503     { "IS", "Iceland" },
504     { "IT", "Italy" },
505     { "JM", "Jamaica" },
506     { "JO", "Jordan" },
507     { "JP", "Japan" },
508     { "KE", "Kenya" },
509     { "KG", "Kyrgyzstan" },
510     { "KH", "Cambodia" },
511     { "KR", "South Korea" },
512     { "KW", "Kuwait" },
513     { "KZ", "Kazakhstan" },
514     { "LA", "Laos" },
515     { "LB", "Lebanon" },
516     { "LI", "Liechtenstein" },
517     { "LK", "Sri Lanka" },
518     { "LT", "Lithuania" },
519     { "LU", "Luxembourg" },
520     { "LV", "Latvia" },
521     { "LY", "Libya" },
522     { "MA", "Morocco" },
523     { "MC", "Principality of Monaco" },
524     { "MD", "Moldava" },
525     { "MD", "Moldova" },
526     { "ME", "Montenegro" },
527     { "MK", "Former Yugoslav Republic of Macedonia" },
528     { "ML", "Mali" },
529     { "MM", "Myanmar" },
530     { "MN", "Mongolia" },
531     { "MO", "Macau S.A.R." },
532     { "MT", "Malta" },
533     { "MV", "Maldives" },
534     { "MX", "Mexico" },
535     { "MY", "Malaysia" },
536     { "NG", "Nigeria" },
537     { "NI", "Nicaragua" },
538     { "NL", "Netherlands" },
539     { "NO", "Norway" },
540     { "NP", "Nepal" },
541     { "NZ", "New Zealand" },
542     { "OM", "Oman" },
543     { "PA", "Panama" },
544     { "PE", "Peru" },
545     { "PH", "Philippines" },
546     { "PK", "Islamic Republic of Pakistan" },
547     { "PL", "Poland" },
548     { "PR", "Puerto Rico" },
549     { "PT", "Portugal" },
550     { "PY", "Paraguay" },
551     { "QA", "Qatar" },
552     { "RE", "Reunion" },
553     { "RO", "Romania" },
554     { "RS", "Serbia" },
555     { "RU", "Russia" },
556     { "RW", "Rwanda" },
557     { "SA", "Saudi Arabia" },
558     { "SE", "Sweden" },
559     { "SG", "Singapore" },
560     { "SI", "Slovenia" },
561     { "SK", "Slovak" },
562     { "SN", "Senegal" },
563     { "SO", "Somalia" },
564     { "SR", "Suriname" },
565     { "SV", "El Salvador" },
566     { "SY", "Syria" },
567     { "TH", "Thailand" },
568     { "TJ", "Tajikistan" },
569     { "TM", "Turkmenistan" },
570     { "TN", "Tunisia" },
571     { "TR", "Turkey" },
572     { "TT", "Trinidad and Tobago" },
573     { "TW", "Taiwan" },
574     { "TZ", "Tanzania" },
575     { "UA", "Ukraine" },
576     { "US", "United States" },
577     { "UY", "Uruguay" },
578     { "VA", "Vatican" },
579     { "VE", "Venezuela" },
580     { "VN", "Viet Nam" },
581     { "YE", "Yemen" },
582     { "ZA", "South Africa" },
583     { "ZW", "Zimbabwe" }
584   };
585 
586 /* Given a string STRING, find the set of indices i such that TABLE[i].code is
587    the given STRING.  It is a range [lo,hi-1].  */
588 typedef struct { size_t lo; size_t hi; } range_t;
589 static void
search(const struct table_entry * table,size_t table_size,const char * string,range_t * result)590 search (const struct table_entry *table, size_t table_size, const char *string,
591         range_t *result)
592 {
593   /* The table is sorted.  Perform a binary search.  */
594   size_t hi = table_size;
595   size_t lo = 0;
596   while (lo < hi)
597     {
598       /* Invariant:
599          for i < lo, strcmp (table[i].code, string) < 0,
600          for i >= hi, strcmp (table[i].code, string) > 0.  */
601       size_t mid = (hi + lo) >> 1; /* >= lo, < hi */
602       int cmp = strcmp (table[mid].code, string);
603       if (cmp < 0)
604         lo = mid + 1;
605       else if (cmp > 0)
606         hi = mid;
607       else
608         {
609           /* Found an i with
610                strcmp (language_table[i].code, string) == 0.
611              Find the entire interval of such i.  */
612           {
613             size_t i;
614 
615             for (i = mid; i > lo; )
616               {
617                 i--;
618                 if (strcmp (table[i].code, string) < 0)
619                   {
620                     lo = i + 1;
621                     break;
622                   }
623               }
624           }
625           {
626             size_t i;
627 
628             for (i = mid + 1; i < hi; i++)
629               {
630                 if (strcmp (table[i].code, string) > 0)
631                   {
632                     hi = i;
633                     break;
634                   }
635               }
636           }
637           /* The set of i with
638                strcmp (language_table[i].code, string) == 0
639              is the interval [lo, hi-1].  */
640           break;
641         }
642     }
643   result->lo = lo;
644   result->hi = hi;
645 }
646 
647 /* Like setlocale, but accept also locale names in the form ll or ll_CC,
648    where ll is an ISO 639 language code and CC is an ISO 3166 country code.  */
649 static char *
setlocale_unixlike(int category,const char * locale)650 setlocale_unixlike (int category, const char *locale)
651 {
652   char *result;
653   char llCC_buf[64];
654   char ll_buf[64];
655   char CC_buf[64];
656 
657   /* The native Windows implementation of setlocale understands the special
658      locale name "C", but not "POSIX".  Therefore map "POSIX" to "C".  */
659   if (locale != NULL && strcmp (locale, "POSIX") == 0)
660     locale = "C";
661 
662   /* First, try setlocale with the original argument unchanged.  */
663   result = setlocale (category, locale);
664   if (result != NULL)
665     return result;
666 
667   /* Otherwise, assume the argument is in the form
668        language[_territory][.codeset][@modifier]
669      and try to map it using the tables.  */
670   if (strlen (locale) < sizeof (llCC_buf))
671     {
672       /* Second try: Remove the codeset part.  */
673       {
674         const char *p = locale;
675         char *q = llCC_buf;
676 
677         /* Copy the part before the dot.  */
678         for (; *p != '\0' && *p != '.'; p++, q++)
679           *q = *p;
680         if (*p == '.')
681           /* Skip the part up to the '@', if any.  */
682           for (; *p != '\0' && *p != '@'; p++)
683             ;
684         /* Copy the part starting with '@', if any.  */
685         for (; *p != '\0'; p++, q++)
686           *q = *p;
687         *q = '\0';
688       }
689       /* llCC_buf now contains
690            language[_territory][@modifier]
691        */
692       if (strcmp (llCC_buf, locale) != 0)
693         {
694           result = setlocale (category, llCC_buf);
695           if (result != NULL)
696             return result;
697         }
698       /* Look it up in language_table.  */
699       {
700         range_t range;
701         size_t i;
702 
703         search (language_table,
704                 sizeof (language_table) / sizeof (language_table[0]),
705                 llCC_buf,
706                 &range);
707 
708         for (i = range.lo; i < range.hi; i++)
709           {
710             /* Try the replacement in language_table[i].  */
711             result = setlocale (category, language_table[i].english);
712             if (result != NULL)
713               return result;
714           }
715       }
716       /* Split language[_territory][@modifier]
717          into  ll_buf = language[@modifier]
718          and   CC_buf = territory
719        */
720       {
721         const char *underscore = strchr (llCC_buf, '_');
722         if (underscore != NULL)
723           {
724             const char *territory_start = underscore + 1;
725             const char *territory_end = strchr (territory_start, '@');
726             if (territory_end == NULL)
727               territory_end = territory_start + strlen (territory_start);
728 
729             memcpy (ll_buf, llCC_buf, underscore - llCC_buf);
730             strcpy (ll_buf + (underscore - llCC_buf), territory_end);
731 
732             memcpy (CC_buf, territory_start, territory_end - territory_start);
733             CC_buf[territory_end - territory_start] = '\0';
734 
735             {
736               /* Look up ll_buf in language_table
737                  and CC_buf in country_table.  */
738               range_t language_range;
739 
740               search (language_table,
741                       sizeof (language_table) / sizeof (language_table[0]),
742                       ll_buf,
743                       &language_range);
744               if (language_range.lo < language_range.hi)
745                 {
746                   range_t country_range;
747 
748                   search (country_table,
749                           sizeof (country_table) / sizeof (country_table[0]),
750                           CC_buf,
751                           &country_range);
752                   if (country_range.lo < country_range.hi)
753                     {
754                       size_t i;
755                       size_t j;
756 
757                       for (i = language_range.lo; i < language_range.hi; i++)
758                         for (j = country_range.lo; j < country_range.hi; j++)
759                           {
760                             /* Concatenate the replacements.  */
761                             const char *part1 = language_table[i].english;
762                             size_t part1_len = strlen (part1);
763                             const char *part2 = country_table[j].english;
764                             size_t part2_len = strlen (part2) + 1;
765                             char buf[64+64];
766 
767                             if (!(part1_len + 1 + part2_len <= sizeof (buf)))
768                               abort ();
769                             memcpy (buf, part1, part1_len);
770                             buf[part1_len] = '_';
771                             memcpy (buf + part1_len + 1, part2, part2_len);
772 
773                             /* Try the concatenated replacements.  */
774                             result = setlocale (category, buf);
775                             if (result != NULL)
776                               return result;
777                           }
778                     }
779 
780                   /* Try omitting the country entirely.  This may set a locale
781                      corresponding to the wrong country, but is better than
782                      failing entirely.  */
783                   {
784                     size_t i;
785 
786                     for (i = language_range.lo; i < language_range.hi; i++)
787                       {
788                         /* Try only the language replacement.  */
789                         result =
790                           setlocale (category, language_table[i].english);
791                         if (result != NULL)
792                           return result;
793                       }
794                   }
795                 }
796             }
797           }
798       }
799     }
800 
801   /* Failed.  */
802   return NULL;
803 }
804 
805 # elif defined __ANDROID__
806 
807 /* Like setlocale, but accept also the locale names "C" and "POSIX".  */
808 static char *
setlocale_unixlike(int category,const char * locale)809 setlocale_unixlike (int category, const char *locale)
810 {
811   char *result = setlocale (category, locale);
812   if (result == NULL)
813     switch (category)
814       {
815       case LC_CTYPE:
816       case LC_NUMERIC:
817       case LC_TIME:
818       case LC_COLLATE:
819       case LC_MONETARY:
820       case LC_MESSAGES:
821       case LC_ALL:
822       case LC_PAPER:
823       case LC_NAME:
824       case LC_ADDRESS:
825       case LC_TELEPHONE:
826       case LC_MEASUREMENT:
827         if (locale == NULL
828             || strcmp (locale, "C") == 0 || strcmp (locale, "POSIX") == 0)
829           result = (char *) "C";
830         break;
831       default:
832         break;
833       }
834   return result;
835 }
836 #  define setlocale setlocale_unixlike
837 
838 # else
839 #  define setlocale_unixlike setlocale
840 # endif
841 
842 # if LC_MESSAGES == 1729
843 
844 /* The system does not store an LC_MESSAGES locale category.  Do it here.  */
845 static char lc_messages_name[64] = "C";
846 
847 /* Like setlocale, but support also LC_MESSAGES.  */
848 static char *
setlocale_single(int category,const char * locale)849 setlocale_single (int category, const char *locale)
850 {
851   if (category == LC_MESSAGES)
852     {
853       if (locale != NULL)
854         {
855           lc_messages_name[sizeof (lc_messages_name) - 1] = '\0';
856           strncpy (lc_messages_name, locale, sizeof (lc_messages_name) - 1);
857         }
858       return lc_messages_name;
859     }
860   else
861     return setlocale_unixlike (category, locale);
862 }
863 
864 # else
865 #  define setlocale_single setlocale_unixlike
866 # endif
867 
868 # if defined __APPLE__ && defined __MACH__
869 
870 /* Mapping from language to main territory where that language is spoken.  */
871 static char const locales_with_principal_territory[][6 + 1] =
872   {
873                 /* Language     Main territory */
874     "ace_ID",   /* Achinese     Indonesia */
875     "af_ZA",    /* Afrikaans    South Africa */
876     "ak_GH",    /* Akan         Ghana */
877     "am_ET",    /* Amharic      Ethiopia */
878     "an_ES",    /* Aragonese    Spain */
879     "ang_GB",   /* Old English  Britain */
880     "arn_CL",   /* Mapudungun   Chile */
881     "as_IN",    /* Assamese     India */
882     "ast_ES",   /* Asturian     Spain */
883     "av_RU",    /* Avaric       Russia */
884     "awa_IN",   /* Awadhi       India */
885     "az_AZ",    /* Azerbaijani  Azerbaijan */
886     "ban_ID",   /* Balinese     Indonesia */
887     "be_BY",    /* Belarusian   Belarus */
888     "bej_SD",   /* Beja         Sudan */
889     "bem_ZM",   /* Bemba        Zambia */
890     "bg_BG",    /* Bulgarian    Bulgaria */
891     "bho_IN",   /* Bhojpuri     India */
892     "bi_VU",    /* Bislama      Vanuatu */
893     "bik_PH",   /* Bikol        Philippines */
894     "bin_NG",   /* Bini         Nigeria */
895     "bm_ML",    /* Bambara      Mali */
896     "bn_IN",    /* Bengali      India */
897     "bo_CN",    /* Tibetan      China */
898     "br_FR",    /* Breton       France */
899     "bs_BA",    /* Bosnian      Bosnia */
900     "bug_ID",   /* Buginese     Indonesia */
901     "ca_ES",    /* Catalan      Spain */
902     "ce_RU",    /* Chechen      Russia */
903     "ceb_PH",   /* Cebuano      Philippines */
904     "co_FR",    /* Corsican     France */
905     "cr_CA",    /* Cree         Canada */
906     /* Don't put "crh_UZ" or "crh_UA" here.  That would be asking for fruitless
907        political discussion.  */
908     "cs_CZ",    /* Czech        Czech Republic */
909     "csb_PL",   /* Kashubian    Poland */
910     "cy_GB",    /* Welsh        Britain */
911     "da_DK",    /* Danish       Denmark */
912     "de_DE",    /* German       Germany */
913     "din_SD",   /* Dinka        Sudan */
914     "doi_IN",   /* Dogri        India */
915     "dsb_DE",   /* Lower Sorbian        Germany */
916     "dv_MV",    /* Divehi       Maldives */
917     "dz_BT",    /* Dzongkha     Bhutan */
918     "ee_GH",    /* Éwé          Ghana */
919     "el_GR",    /* Greek        Greece */
920     /* Don't put "en_GB" or "en_US" here.  That would be asking for fruitless
921        political discussion.  */
922     "es_ES",    /* Spanish      Spain */
923     "et_EE",    /* Estonian     Estonia */
924     "fa_IR",    /* Persian      Iran */
925     "fi_FI",    /* Finnish      Finland */
926     "fil_PH",   /* Filipino     Philippines */
927     "fj_FJ",    /* Fijian       Fiji */
928     "fo_FO",    /* Faroese      Faeroe Islands */
929     "fon_BJ",   /* Fon          Benin */
930     "fr_FR",    /* French       France */
931     "fur_IT",   /* Friulian     Italy */
932     "fy_NL",    /* Western Frisian      Netherlands */
933     "ga_IE",    /* Irish        Ireland */
934     "gd_GB",    /* Scottish Gaelic      Britain */
935     "gon_IN",   /* Gondi        India */
936     "gsw_CH",   /* Swiss German Switzerland */
937     "gu_IN",    /* Gujarati     India */
938     "he_IL",    /* Hebrew       Israel */
939     "hi_IN",    /* Hindi        India */
940     "hil_PH",   /* Hiligaynon   Philippines */
941     "hr_HR",    /* Croatian     Croatia */
942     "hsb_DE",   /* Upper Sorbian        Germany */
943     "ht_HT",    /* Haitian      Haiti */
944     "hu_HU",    /* Hungarian    Hungary */
945     "hy_AM",    /* Armenian     Armenia */
946     "id_ID",    /* Indonesian   Indonesia */
947     "ig_NG",    /* Igbo         Nigeria */
948     "ii_CN",    /* Sichuan Yi   China */
949     "ilo_PH",   /* Iloko        Philippines */
950     "is_IS",    /* Icelandic    Iceland */
951     "it_IT",    /* Italian      Italy */
952     "ja_JP",    /* Japanese     Japan */
953     "jab_NG",   /* Hyam         Nigeria */
954     "jv_ID",    /* Javanese     Indonesia */
955     "ka_GE",    /* Georgian     Georgia */
956     "kab_DZ",   /* Kabyle       Algeria */
957     "kaj_NG",   /* Jju          Nigeria */
958     "kam_KE",   /* Kamba        Kenya */
959     "kmb_AO",   /* Kimbundu     Angola */
960     "kcg_NG",   /* Tyap         Nigeria */
961     "kdm_NG",   /* Kagoma       Nigeria */
962     "kg_CD",    /* Kongo        Democratic Republic of Congo */
963     "kk_KZ",    /* Kazakh       Kazakhstan */
964     "kl_GL",    /* Kalaallisut  Greenland */
965     "km_KH",    /* Central Khmer        Cambodia */
966     "kn_IN",    /* Kannada      India */
967     "ko_KR",    /* Korean       Korea (South) */
968     "kok_IN",   /* Konkani      India */
969     "kr_NG",    /* Kanuri       Nigeria */
970     "kru_IN",   /* Kurukh       India */
971     "ky_KG",    /* Kyrgyz       Kyrgyzstan */
972     "lg_UG",    /* Ganda        Uganda */
973     "li_BE",    /* Limburgish   Belgium */
974     "lo_LA",    /* Laotian      Laos */
975     "lt_LT",    /* Lithuanian   Lithuania */
976     "lu_CD",    /* Luba-Katanga Democratic Republic of Congo */
977     "lua_CD",   /* Luba-Lulua   Democratic Republic of Congo */
978     "luo_KE",   /* Luo          Kenya */
979     "lv_LV",    /* Latvian      Latvia */
980     "mad_ID",   /* Madurese     Indonesia */
981     "mag_IN",   /* Magahi       India */
982     "mai_IN",   /* Maithili     India */
983     "mak_ID",   /* Makasar      Indonesia */
984     "man_ML",   /* Mandingo     Mali */
985     "men_SL",   /* Mende        Sierra Leone */
986     "mfe_MU",   /* Mauritian Creole     Mauritius */
987     "mg_MG",    /* Malagasy     Madagascar */
988     "mi_NZ",    /* Maori        New Zealand */
989     "min_ID",   /* Minangkabau  Indonesia */
990     "mk_MK",    /* Macedonian   North Macedonia */
991     "ml_IN",    /* Malayalam    India */
992     "mn_MN",    /* Mongolian    Mongolia */
993     "mni_IN",   /* Manipuri     India */
994     "mos_BF",   /* Mossi        Burkina Faso */
995     "mr_IN",    /* Marathi      India */
996     "ms_MY",    /* Malay        Malaysia */
997     "mt_MT",    /* Maltese      Malta */
998     "mwr_IN",   /* Marwari      India */
999     "my_MM",    /* Burmese      Myanmar */
1000     "na_NR",    /* Nauru        Nauru */
1001     "nah_MX",   /* Nahuatl      Mexico */
1002     "nap_IT",   /* Neapolitan   Italy */
1003     "nb_NO",    /* Norwegian Bokmål    Norway */
1004     "nds_DE",   /* Low Saxon    Germany */
1005     "ne_NP",    /* Nepali       Nepal */
1006     "nl_NL",    /* Dutch        Netherlands */
1007     "nn_NO",    /* Norwegian Nynorsk    Norway */
1008     "no_NO",    /* Norwegian    Norway */
1009     "nr_ZA",    /* South Ndebele        South Africa */
1010     "nso_ZA",   /* Northern Sotho       South Africa */
1011     "ny_MW",    /* Chichewa     Malawi */
1012     "nym_TZ",   /* Nyamwezi     Tanzania */
1013     "nyn_UG",   /* Nyankole     Uganda */
1014     "oc_FR",    /* Occitan      France */
1015     "oj_CA",    /* Ojibwa       Canada */
1016     "or_IN",    /* Oriya        India */
1017     "pa_IN",    /* Punjabi      India */
1018     "pag_PH",   /* Pangasinan   Philippines */
1019     "pam_PH",   /* Pampanga     Philippines */
1020     "pap_AN",   /* Papiamento   Netherlands Antilles - this line can be removed in 2018 */
1021     "pbb_CO",   /* Páez         Colombia */
1022     "pl_PL",    /* Polish       Poland */
1023     "ps_AF",    /* Pashto       Afghanistan */
1024     "pt_PT",    /* Portuguese   Portugal */
1025     "raj_IN",   /* Rajasthani   India */
1026     "rm_CH",    /* Romansh      Switzerland */
1027     "rn_BI",    /* Kirundi      Burundi */
1028     "ro_RO",    /* Romanian     Romania */
1029     "ru_RU",    /* Russian      Russia */
1030     "rw_RW",    /* Kinyarwanda  Rwanda */
1031     "sa_IN",    /* Sanskrit     India */
1032     "sah_RU",   /* Yakut        Russia */
1033     "sas_ID",   /* Sasak        Indonesia */
1034     "sat_IN",   /* Santali      India */
1035     "sc_IT",    /* Sardinian    Italy */
1036     "scn_IT",   /* Sicilian     Italy */
1037     "sg_CF",    /* Sango        Central African Republic */
1038     "shn_MM",   /* Shan         Myanmar */
1039     "si_LK",    /* Sinhala      Sri Lanka */
1040     "sid_ET",   /* Sidamo       Ethiopia */
1041     "sk_SK",    /* Slovak       Slovakia */
1042     "sl_SI",    /* Slovenian    Slovenia */
1043     "sm_WS",    /* Samoan       Samoa */
1044     "smn_FI",   /* Inari Sami   Finland */
1045     "sms_FI",   /* Skolt Sami   Finland */
1046     "so_SO",    /* Somali       Somalia */
1047     "sq_AL",    /* Albanian     Albania */
1048     "sr_RS",    /* Serbian      Serbia */
1049     "srr_SN",   /* Serer        Senegal */
1050     "suk_TZ",   /* Sukuma       Tanzania */
1051     "sus_GN",   /* Susu         Guinea */
1052     "sv_SE",    /* Swedish      Sweden */
1053     "te_IN",    /* Telugu       India */
1054     "tem_SL",   /* Timne        Sierra Leone */
1055     "tet_ID",   /* Tetum        Indonesia */
1056     "tg_TJ",    /* Tajik        Tajikistan */
1057     "th_TH",    /* Thai         Thailand */
1058     "ti_ER",    /* Tigrinya     Eritrea */
1059     "tiv_NG",   /* Tiv          Nigeria */
1060     "tk_TM",    /* Turkmen      Turkmenistan */
1061     "tl_PH",    /* Tagalog      Philippines */
1062     "to_TO",    /* Tonga        Tonga */
1063     "tpi_PG",   /* Tok Pisin    Papua New Guinea */
1064     "tr_TR",    /* Turkish      Turkey */
1065     "tum_MW",   /* Tumbuka      Malawi */
1066     "ug_CN",    /* Uighur       China */
1067     "uk_UA",    /* Ukrainian    Ukraine */
1068     "umb_AO",   /* Umbundu      Angola */
1069     "ur_PK",    /* Urdu         Pakistan */
1070     "uz_UZ",    /* Uzbek        Uzbekistan */
1071     "ve_ZA",    /* Venda        South Africa */
1072     "vi_VN",    /* Vietnamese   Vietnam */
1073     "wa_BE",    /* Walloon      Belgium */
1074     "wal_ET",   /* Walamo       Ethiopia */
1075     "war_PH",   /* Waray        Philippines */
1076     "wen_DE",   /* Sorbian      Germany */
1077     "yao_MW",   /* Yao          Malawi */
1078     "zap_MX"    /* Zapotec      Mexico */
1079   };
1080 
1081 /* Compare just the language part of two locale names.  */
1082 static int
langcmp(const char * locale1,const char * locale2)1083 langcmp (const char *locale1, const char *locale2)
1084 {
1085   size_t locale1_len;
1086   size_t locale2_len;
1087   int cmp;
1088 
1089   {
1090     const char *locale1_end = strchr (locale1, '_');
1091     if (locale1_end != NULL)
1092       locale1_len = locale1_end - locale1;
1093     else
1094       locale1_len = strlen (locale1);
1095   }
1096   {
1097     const char *locale2_end = strchr (locale2, '_');
1098     if (locale2_end != NULL)
1099       locale2_len = locale2_end - locale2;
1100     else
1101       locale2_len = strlen (locale2);
1102   }
1103 
1104   if (locale1_len < locale2_len)
1105     {
1106       cmp = memcmp (locale1, locale2, locale1_len);
1107       if (cmp == 0)
1108         cmp = -1;
1109     }
1110   else
1111     {
1112       cmp = memcmp (locale1, locale2, locale2_len);
1113       if (locale1_len > locale2_len && cmp == 0)
1114         cmp = 1;
1115     }
1116 
1117   return cmp;
1118 }
1119 
1120 /* Given a locale name, return the main locale with the same language,
1121    or NULL if not found.
1122    For example: "fr_DE" -> "fr_FR".  */
1123 static const char *
get_main_locale_with_same_language(const char * locale)1124 get_main_locale_with_same_language (const char *locale)
1125 {
1126 #  define table locales_with_principal_territory
1127   /* The table is sorted.  Perform a binary search.  */
1128   size_t hi = sizeof (table) / sizeof (table[0]);
1129   size_t lo = 0;
1130   while (lo < hi)
1131     {
1132       /* Invariant:
1133          for i < lo, langcmp (table[i], locale) < 0,
1134          for i >= hi, langcmp (table[i], locale) > 0.  */
1135       size_t mid = (hi + lo) >> 1; /* >= lo, < hi */
1136       int cmp = langcmp (table[mid], locale);
1137       if (cmp < 0)
1138         lo = mid + 1;
1139       else if (cmp > 0)
1140         hi = mid;
1141       else
1142         {
1143           /* Found an i with
1144                langcmp (language_table[i], locale) == 0.
1145              Verify that it is the only such i.  */
1146           if (mid > lo && langcmp (table[mid - 1], locale) >= 0)
1147             abort ();
1148           if (mid + 1 < hi && langcmp (table[mid + 1], locale) <= 0)
1149             abort ();
1150           return table[mid];
1151         }
1152     }
1153 #  undef table
1154   return NULL;
1155 }
1156 
1157 /* Mapping from territory to main language that is spoken in that territory.  */
1158 static char const locales_with_principal_language[][6 + 1] =
1159   {
1160     /* This is based on the set of existing locales in glibc, with duplicates
1161        removed, and on the Wikipedia pages named "Languages of <territory>".
1162        If in doubt, use the locale that exists in macOS.  For example, the only
1163        "*_IN" locale in macOS 10.13 is "hi_IN", so use that.  */
1164     /* A useful shell function for producing a line of this table is:
1165          func_line ()
1166          {
1167            # Usage: func_line ll_CC
1168            ll=`echo "$1" | sed -e 's|_.*||'`
1169            cc=`echo "$1" | sed -e 's|^.*_||'`
1170            llx=`sed -n -e "s|^${ll} ||p" < gettext-tools/doc/ISO_639`
1171            ccx=`expand gettext-tools/doc/ISO_3166 | sed -n -e "s|^${cc}  *||p"`
1172            echo "    \"$1\",    /$X* ${llx} ${ccx} *$X/"
1173          }
1174      */
1175               /* Main language  Territory */
1176     "ca_AD",    /* Catalan      Andorra */
1177     "ar_AE",    /* Arabic       United Arab Emirates */
1178     "ps_AF",    /* Pashto       Afghanistan */
1179     "en_AG",    /* English      Antigua and Barbuda */
1180     "sq_AL",    /* Albanian     Albania */
1181     "hy_AM",    /* Armenian     Armenia */
1182     "pap_AN",   /* Papiamento   Netherlands Antilles - this line can be removed in 2018 */
1183     "pt_AO",    /* Portuguese   Angola */
1184     "es_AR",    /* Spanish      Argentina */
1185     "de_AT",    /* German       Austria */
1186     "en_AU",    /* English      Australia */
1187     /* Aruba has two official languages: "nl_AW", "pap_AW".  */
1188     "az_AZ",    /* Azerbaijani  Azerbaijan */
1189     "bs_BA",    /* Bosnian      Bosnia */
1190     "bn_BD",    /* Bengali      Bangladesh */
1191     "nl_BE",    /* Dutch        Belgium */
1192     "fr_BF",    /* French       Burkina Faso */
1193     "bg_BG",    /* Bulgarian    Bulgaria */
1194     "ar_BH",    /* Arabic       Bahrain */
1195     "rn_BI",    /* Kirundi      Burundi */
1196     "fr_BJ",    /* French       Benin */
1197     "es_BO",    /* Spanish      Bolivia */
1198     "pt_BR",    /* Portuguese   Brazil */
1199     "dz_BT",    /* Dzongkha     Bhutan */
1200     "en_BW",    /* English      Botswana */
1201     "be_BY",    /* Belarusian   Belarus */
1202     "en_CA",    /* English      Canada */
1203     "fr_CD",    /* French       Democratic Republic of Congo */
1204     "sg_CF",    /* Sango        Central African Republic */
1205     "de_CH",    /* German       Switzerland */
1206     "es_CL",    /* Spanish      Chile */
1207     "zh_CN",    /* Chinese      China */
1208     "es_CO",    /* Spanish      Colombia */
1209     "es_CR",    /* Spanish      Costa Rica */
1210     "es_CU",    /* Spanish      Cuba */
1211     /* Curaçao has three official languages: "nl_CW", "pap_CW", "en_CW".  */
1212     "el_CY",    /* Greek        Cyprus */
1213     "cs_CZ",    /* Czech        Czech Republic */
1214     "de_DE",    /* German       Germany */
1215     /* Djibouti has two official languages: "ar_DJ" and "fr_DJ".  */
1216     "da_DK",    /* Danish       Denmark */
1217     "es_DO",    /* Spanish      Dominican Republic */
1218     "ar_DZ",    /* Arabic       Algeria */
1219     "es_EC",    /* Spanish      Ecuador */
1220     "et_EE",    /* Estonian     Estonia */
1221     "ar_EG",    /* Arabic       Egypt */
1222     "ti_ER",    /* Tigrinya     Eritrea */
1223     "es_ES",    /* Spanish      Spain */
1224     "am_ET",    /* Amharic      Ethiopia */
1225     "fi_FI",    /* Finnish      Finland */
1226     /* Fiji has three official languages: "en_FJ", "fj_FJ", "hif_FJ".  */
1227     "fo_FO",    /* Faroese      Faeroe Islands */
1228     "fr_FR",    /* French       France */
1229     "en_GB",    /* English      Britain */
1230     "ka_GE",    /* Georgian     Georgia */
1231     "en_GH",    /* English      Ghana */
1232     "kl_GL",    /* Kalaallisut  Greenland */
1233     "fr_GN",    /* French       Guinea */
1234     "el_GR",    /* Greek        Greece */
1235     "es_GT",    /* Spanish      Guatemala */
1236     "zh_HK",    /* Chinese      Hong Kong */
1237     "es_HN",    /* Spanish      Honduras */
1238     "hr_HR",    /* Croatian     Croatia */
1239     "ht_HT",    /* Haitian      Haiti */
1240     "hu_HU",    /* Hungarian    Hungary */
1241     "id_ID",    /* Indonesian   Indonesia */
1242     "en_IE",    /* English      Ireland */
1243     "he_IL",    /* Hebrew       Israel */
1244     "hi_IN",    /* Hindi        India */
1245     "ar_IQ",    /* Arabic       Iraq */
1246     "fa_IR",    /* Persian      Iran */
1247     "is_IS",    /* Icelandic    Iceland */
1248     "it_IT",    /* Italian      Italy */
1249     "ar_JO",    /* Arabic       Jordan */
1250     "ja_JP",    /* Japanese     Japan */
1251     "sw_KE",    /* Swahili      Kenya */
1252     "ky_KG",    /* Kyrgyz       Kyrgyzstan */
1253     "km_KH",    /* Central Khmer        Cambodia */
1254     "ko_KR",    /* Korean       Korea (South) */
1255     "ar_KW",    /* Arabic       Kuwait */
1256     "kk_KZ",    /* Kazakh       Kazakhstan */
1257     "lo_LA",    /* Laotian      Laos */
1258     "ar_LB",    /* Arabic       Lebanon */
1259     "de_LI",    /* German       Liechtenstein */
1260     "si_LK",    /* Sinhala      Sri Lanka */
1261     "lt_LT",    /* Lithuanian   Lithuania */
1262     /* Luxembourg has three official languages: "lb_LU", "fr_LU", "de_LU".  */
1263     "lv_LV",    /* Latvian      Latvia */
1264     "ar_LY",    /* Arabic       Libya */
1265     "ar_MA",    /* Arabic       Morocco */
1266     "sr_ME",    /* Serbian      Montenegro */
1267     "mg_MG",    /* Malagasy     Madagascar */
1268     "mk_MK",    /* Macedonian   North Macedonia */
1269     "fr_ML",    /* French       Mali */
1270     "my_MM",    /* Burmese      Myanmar */
1271     "mn_MN",    /* Mongolian    Mongolia */
1272     "mt_MT",    /* Maltese      Malta */
1273     "mfe_MU",   /* Mauritian Creole     Mauritius */
1274     "dv_MV",    /* Divehi       Maldives */
1275     "ny_MW",    /* Chichewa     Malawi */
1276     "es_MX",    /* Spanish      Mexico */
1277     "ms_MY",    /* Malay        Malaysia */
1278     "en_NG",    /* English      Nigeria */
1279     "es_NI",    /* Spanish      Nicaragua */
1280     "nl_NL",    /* Dutch        Netherlands */
1281     "no_NO",    /* Norwegian    Norway */
1282     "ne_NP",    /* Nepali       Nepal */
1283     "na_NR",    /* Nauru        Nauru */
1284     "niu_NU",   /* Niuean       Niue */
1285     "en_NZ",    /* English      New Zealand */
1286     "ar_OM",    /* Arabic       Oman */
1287     "es_PA",    /* Spanish      Panama */
1288     "es_PE",    /* Spanish      Peru */
1289     "tpi_PG",   /* Tok Pisin    Papua New Guinea */
1290     "fil_PH",   /* Filipino     Philippines */
1291     "pa_PK",    /* Punjabi      Pakistan */
1292     "pl_PL",    /* Polish       Poland */
1293     "es_PR",    /* Spanish      Puerto Rico */
1294     "pt_PT",    /* Portuguese   Portugal */
1295     "es_PY",    /* Spanish      Paraguay */
1296     "ar_QA",    /* Arabic       Qatar */
1297     "ro_RO",    /* Romanian     Romania */
1298     "sr_RS",    /* Serbian      Serbia */
1299     "ru_RU",    /* Russian      Russia */
1300     "rw_RW",    /* Kinyarwanda  Rwanda */
1301     "ar_SA",    /* Arabic       Saudi Arabia */
1302     "en_SC",    /* English      Seychelles */
1303     "ar_SD",    /* Arabic       Sudan */
1304     "sv_SE",    /* Swedish      Sweden */
1305     "en_SG",    /* English      Singapore */
1306     "sl_SI",    /* Slovenian    Slovenia */
1307     "sk_SK",    /* Slovak       Slovakia */
1308     "en_SL",    /* English      Sierra Leone */
1309     "fr_SN",    /* French       Senegal */
1310     "so_SO",    /* Somali       Somalia */
1311     "ar_SS",    /* Arabic       South Sudan */
1312     "es_SV",    /* Spanish      El Salvador */
1313     "ar_SY",    /* Arabic       Syria */
1314     "th_TH",    /* Thai         Thailand */
1315     "tg_TJ",    /* Tajik        Tajikistan */
1316     "tk_TM",    /* Turkmen      Turkmenistan */
1317     "ar_TN",    /* Arabic       Tunisia */
1318     "to_TO",    /* Tonga        Tonga */
1319     "tr_TR",    /* Turkish      Turkey */
1320     "zh_TW",    /* Chinese      Taiwan */
1321     "sw_TZ",    /* Swahili      Tanzania */
1322     "uk_UA",    /* Ukrainian    Ukraine */
1323     "lg_UG",    /* Ganda        Uganda */
1324     "en_US",    /* English      United States of America */
1325     "es_UY",    /* Spanish      Uruguay */
1326     "uz_UZ",    /* Uzbek        Uzbekistan */
1327     "es_VE",    /* Spanish      Venezuela */
1328     "vi_VN",    /* Vietnamese   Vietnam */
1329     "bi_VU",    /* Bislama      Vanuatu */
1330     "sm_WS",    /* Samoan       Samoa */
1331     "ar_YE",    /* Arabic       Yemen */
1332     "en_ZA",    /* English      South Africa */
1333     "en_ZM",    /* English      Zambia */
1334     "en_ZW"     /* English      Zimbabwe */
1335   };
1336 
1337 /* Compare just the territory part of two locale names.  */
1338 static int
terrcmp(const char * locale1,const char * locale2)1339 terrcmp (const char *locale1, const char *locale2)
1340 {
1341   const char *territory1 = strrchr (locale1, '_') + 1;
1342   const char *territory2 = strrchr (locale2, '_') + 1;
1343 
1344   return strcmp (territory1, territory2);
1345 }
1346 
1347 /* Given a locale name, return the locale corresponding to the main language
1348    with the same territory, or NULL if not found.
1349    For example: "fr_DE" -> "de_DE".  */
1350 static const char *
get_main_locale_with_same_territory(const char * locale)1351 get_main_locale_with_same_territory (const char *locale)
1352 {
1353   if (strrchr (locale, '_') != NULL)
1354     {
1355 #  define table locales_with_principal_language
1356       /* The table is sorted.  Perform a binary search.  */
1357       size_t hi = sizeof (table) / sizeof (table[0]);
1358       size_t lo = 0;
1359       while (lo < hi)
1360         {
1361           /* Invariant:
1362              for i < lo, terrcmp (table[i], locale) < 0,
1363              for i >= hi, terrcmp (table[i], locale) > 0.  */
1364           size_t mid = (hi + lo) >> 1; /* >= lo, < hi */
1365           int cmp = terrcmp (table[mid], locale);
1366           if (cmp < 0)
1367             lo = mid + 1;
1368           else if (cmp > 0)
1369             hi = mid;
1370           else
1371             {
1372               /* Found an i with
1373                    terrcmp (language_table[i], locale) == 0.
1374                  Verify that it is the only such i.  */
1375               if (mid > lo && terrcmp (table[mid - 1], locale) >= 0)
1376                 abort ();
1377               if (mid + 1 < hi && terrcmp (table[mid + 1], locale) <= 0)
1378                 abort ();
1379               return table[mid];
1380             }
1381         }
1382 #  undef table
1383     }
1384   return NULL;
1385 }
1386 
1387 # endif
1388 
1389 DLL_EXPORTED
1390 char *
libintl_setlocale(int category,const char * locale)1391 libintl_setlocale (int category, const char *locale)
1392 {
1393   if (locale != NULL && locale[0] == '\0')
1394     {
1395       /* A request to the set the current locale to the default locale.  */
1396       if (category == LC_ALL)
1397         {
1398           /* Set LC_CTYPE first.  Then the other categories.  */
1399           static int const categories[] =
1400             {
1401               LC_CTYPE,
1402               LC_NUMERIC,
1403               LC_TIME,
1404               LC_COLLATE,
1405               LC_MONETARY,
1406               LC_MESSAGES
1407             };
1408           char *saved_locale;
1409           const char *base_name;
1410           unsigned int i;
1411 
1412           /* Back up the old locale, in case one of the steps fails.  */
1413           saved_locale = setlocale (LC_ALL, NULL);
1414           if (saved_locale == NULL)
1415             return NULL;
1416           saved_locale = strdup (saved_locale);
1417           if (saved_locale == NULL)
1418             return NULL;
1419 
1420           /* Set LC_CTYPE category.  Set all other categories (except possibly
1421              LC_MESSAGES) to the same value in the same call; this is likely to
1422              save calls.  */
1423           base_name =
1424             gl_locale_name_environ (LC_CTYPE, category_to_name (LC_CTYPE));
1425           if (base_name == NULL)
1426             base_name = gl_locale_name_default ();
1427 
1428           if (setlocale_unixlike (LC_ALL, base_name) != NULL)
1429             {
1430               /* LC_CTYPE category already set.  */
1431               i = 1;
1432             }
1433           else
1434             {
1435               /* On Mac OS X, "UTF-8" is a valid locale name for LC_CTYPE but
1436                  not for LC_ALL.  Therefore this call may fail.  So, try
1437                  another base_name.  */
1438               base_name = "C";
1439               if (setlocale_unixlike (LC_ALL, base_name) == NULL)
1440                 goto fail;
1441               i = 0;
1442             }
1443 # if defined _WIN32 && ! defined __CYGWIN__
1444           /* On native Windows, setlocale(LC_ALL,...) may succeed but set the
1445              LC_CTYPE category to an invalid value ("C") when it does not
1446              support the specified encoding.  Report a failure instead.  */
1447           if (strchr (base_name, '.') != NULL
1448               && strcmp (setlocale (LC_CTYPE, NULL), "C") == 0)
1449             goto fail;
1450 # endif
1451 
1452           for (; i < sizeof (categories) / sizeof (categories[0]); i++)
1453             {
1454               int cat = categories[i];
1455               const char *name;
1456 
1457               name = gl_locale_name_environ (cat, category_to_name (cat));
1458               if (name == NULL)
1459                 name = gl_locale_name_default ();
1460 
1461               /* If name is the same as base_name, it has already been set
1462                  through the setlocale call before the loop.  */
1463               if (strcmp (name, base_name) != 0
1464 # if LC_MESSAGES == 1729
1465                   || cat == LC_MESSAGES
1466 # endif
1467                  )
1468                 if (setlocale_single (cat, name) == NULL)
1469 # if defined __APPLE__ && defined __MACH__
1470                   {
1471                     /* On Mac OS X 10.13, some locales can be set through
1472                        System Preferences > Language & Region, that are not
1473                        supported by libc.  The system's setlocale() falls
1474                        back to "C" for these locale categories.  We can do
1475                        better, by trying an existing locale with the same
1476                        language or an existing locale with the same territory.
1477                        If we can't, print a warning, to limit user
1478                        expectations.  */
1479                     int warn = 0;
1480 
1481                     if (cat == LC_CTYPE)
1482                       warn = (setlocale_single (cat, "UTF-8") == NULL);
1483                     else if (cat == LC_MESSAGES)
1484                       {
1485 #  if HAVE_CFLOCALECOPYPREFERREDLANGUAGES || HAVE_CFPREFERENCESCOPYAPPVALUE /* MacOS X 10.4 or newer */
1486                         /* Take the primary language preference.  */
1487 #   if HAVE_CFLOCALECOPYPREFERREDLANGUAGES /* MacOS X 10.5 or newer */
1488                         CFArrayRef prefArray = CFLocaleCopyPreferredLanguages ();
1489 #   elif HAVE_CFPREFERENCESCOPYAPPVALUE /* MacOS X 10.4 or newer */
1490                         CFTypeRef preferences =
1491                           CFPreferencesCopyAppValue (CFSTR ("AppleLanguages"),
1492                                                      kCFPreferencesCurrentApplication);
1493                         if (preferences != NULL
1494                             && CFGetTypeID (preferences) == CFArrayGetTypeID ())
1495                           {
1496                             CFArrayRef prefArray = (CFArrayRef)preferences;
1497 #   endif
1498                             int n = CFArrayGetCount (prefArray);
1499                             if (n > 0)
1500                               {
1501                                 char buf[256];
1502                                 CFTypeRef element = CFArrayGetValueAtIndex (prefArray, 0);
1503                                 if (element != NULL
1504                                     && CFGetTypeID (element) == CFStringGetTypeID ()
1505                                     && CFStringGetCString ((CFStringRef)element,
1506                                                            buf, sizeof (buf),
1507                                                            kCFStringEncodingASCII))
1508                                   {
1509                                     /* Remove the country.
1510                                        E.g. "zh-Hans-DE" -> "zh-Hans".  */
1511                                     char *last_minus = strrchr (buf, '-');
1512                                     if (last_minus != NULL)
1513                                       *last_minus = '\0';
1514 
1515                                     /* Convert to Unix locale name.
1516                                        E.g. "zh-Hans" -> "zh_CN".  */
1517                                     gl_locale_name_canonicalize (buf);
1518 
1519                                     /* Try setlocale with this value.  */
1520                                     if (setlocale_single (cat, buf) == NULL)
1521                                       {
1522                                         const char *last_try =
1523                                           get_main_locale_with_same_language (buf);
1524 
1525                                         if (last_try == NULL
1526                                             || setlocale_single (cat, last_try) == NULL)
1527                                           warn = 1;
1528                                       }
1529                                   }
1530                               }
1531 #   if HAVE_CFLOCALECOPYPREFERREDLANGUAGES /* MacOS X 10.5 or newer */
1532                         CFRelease (prefArray);
1533 #   elif HAVE_CFPREFERENCESCOPYAPPVALUE /* MacOS X 10.4 or newer */
1534                           }
1535 #   endif
1536 #  else
1537                         const char *last_try =
1538                           get_main_locale_with_same_language (name);
1539 
1540                         if (last_try == NULL
1541                             || setlocale_single (cat, last_try) == NULL)
1542                           warn = 1;
1543 #  endif
1544                       }
1545                     else
1546                       {
1547                         /* For LC_NUMERIC, the application should use the locale
1548                            properties kCFLocaleDecimalSeparator,
1549                            kCFLocaleGroupingSeparator.
1550                            For LC_TIME, the application should use the locale
1551                            property kCFLocaleCalendarIdentifier.
1552                            For LC_COLLATE, the application should use the locale
1553                            properties kCFLocaleCollationIdentifier,
1554                            kCFLocaleCollatorIdentifier.
1555                            For LC_MONETARY, the applicationshould use the locale
1556                            properties kCFLocaleCurrencySymbol,
1557                            kCFLocaleCurrencyCode.
1558                            But since most applications don't have macOS specific
1559                            code like this, try an existing locale with the same
1560                            territory.  */
1561                         const char *last_try =
1562                           get_main_locale_with_same_territory (name);
1563 
1564                         if (last_try == NULL
1565                             || setlocale_single (cat, last_try) == NULL)
1566                           warn = 1;
1567                       }
1568 
1569                     if (warn)
1570                       {
1571                         /* Warn only if the environment variable
1572                            SETLOCALE_VERBOSE is set.  Otherwise these warnings
1573                            are just annoyances, since normal users won't invoke
1574                            'localedef'.  */
1575                         const char *verbose = getenv ("SETLOCALE_VERBOSE");
1576                         if (verbose != NULL && verbose[0] != '\0')
1577                           fprintf (stderr,
1578                                    "Warning: Failed to set locale category %s to %s.\n",
1579                                    category_to_name (cat), name);
1580                       }
1581                   }
1582 # else
1583                   goto fail;
1584 # endif
1585             }
1586 
1587           /* All steps were successful.  */
1588           ++_nl_msg_cat_cntr;
1589           free (saved_locale);
1590           return setlocale (LC_ALL, NULL);
1591 
1592         fail:
1593           if (saved_locale[0] != '\0') /* don't risk an endless recursion */
1594             setlocale (LC_ALL, saved_locale);
1595           free (saved_locale);
1596           return NULL;
1597         }
1598       else
1599         {
1600           char *result;
1601           const char *name =
1602             gl_locale_name_environ (category, category_to_name (category));
1603           if (name == NULL)
1604             name = gl_locale_name_default ();
1605 
1606           result = setlocale_single (category, name);
1607           if (result != NULL)
1608             ++_nl_msg_cat_cntr;
1609           return result;
1610         }
1611     }
1612   else
1613     {
1614 # if defined _WIN32 && ! defined __CYGWIN__
1615       if (category == LC_ALL && locale != NULL && strchr (locale, '.') != NULL)
1616         {
1617           char *saved_locale;
1618 
1619           /* Back up the old locale.  */
1620           saved_locale = setlocale (LC_ALL, NULL);
1621           if (saved_locale == NULL)
1622             return NULL;
1623           saved_locale = strdup (saved_locale);
1624           if (saved_locale == NULL)
1625             return NULL;
1626 
1627           if (setlocale_unixlike (LC_ALL, locale) == NULL)
1628             {
1629               free (saved_locale);
1630               return NULL;
1631             }
1632 
1633           /* On native Windows, setlocale(LC_ALL,...) may succeed but set the
1634              LC_CTYPE category to an invalid value ("C") when it does not
1635              support the specified encoding.  Report a failure instead.  */
1636           if (strcmp (setlocale (LC_CTYPE, NULL), "C") == 0)
1637             {
1638               if (saved_locale[0] != '\0') /* don't risk an endless recursion */
1639                 setlocale (LC_ALL, saved_locale);
1640               free (saved_locale);
1641               return NULL;
1642             }
1643 
1644           /* It was really successful.  */
1645           ++_nl_msg_cat_cntr;
1646           free (saved_locale);
1647           return setlocale (LC_ALL, NULL);
1648         }
1649       else
1650 # endif
1651         {
1652           char *result = setlocale_single (category, locale);
1653           if (result != NULL)
1654             ++_nl_msg_cat_cntr;
1655           return result;
1656         }
1657     }
1658 }
1659 
1660 # if HAVE_NEWLOCALE
1661 
1662 DLL_EXPORTED
1663 locale_t
1664 libintl_newlocale (int category_mask, const char *locale, locale_t base)
1665 {
1666   if (category_mask != 0 && locale != NULL && locale[0] == '\0')
1667     {
1668       /* A request to construct a locale_t object that refers to the default
1669          locale.  */
1670 
1671       /* Set LC_CTYPE first.  Then the other categories.  */
1672       static struct { int cat; int mask; } const categories[] =
1673         {
1674           { LC_CTYPE,    LC_CTYPE_MASK },
1675           { LC_NUMERIC,  LC_NUMERIC_MASK },
1676           { LC_TIME,     LC_TIME_MASK },
1677           { LC_COLLATE,  LC_COLLATE_MASK },
1678           { LC_MONETARY, LC_MONETARY_MASK },
1679           { LC_MESSAGES, LC_MESSAGES_MASK }
1680         };
1681 
1682       locale_t orig_base = base;
1683 
1684       if ((LC_ALL_MASK & ~category_mask) == 0)
1685         {
1686           const char *base_name;
1687           unsigned int i;
1688 
1689           /* Set LC_CTYPE category.  Set all other categories (except possibly
1690              LC_MESSAGES) to the same value in the same call; this is likely to
1691              save calls.  */
1692           base_name =
1693             gl_locale_name_environ (LC_CTYPE, category_to_name (LC_CTYPE));
1694           if (base_name == NULL)
1695             base_name = gl_locale_name_default ();
1696 
1697           base = newlocale (LC_ALL_MASK, base_name, base);
1698           if (base == NULL)
1699             return NULL;
1700 
1701           for (i = 1; i < sizeof (categories) / sizeof (categories[0]); i++)
1702             {
1703               int category = categories[i].cat;
1704               int category_mask = categories[i].mask;
1705               const char *name;
1706 
1707               name =
1708                 gl_locale_name_environ (category, category_to_name (category));
1709               if (name == NULL)
1710                 name = gl_locale_name_default ();
1711 
1712               /* If name is the same as base_name, it has already been set
1713                  through the setlocale call before the loop.  */
1714               if (strcmp (name, base_name) != 0)
1715                 {
1716                   locale_t copy = newlocale (category_mask, name, base);
1717                   if (copy == NULL)
1718                     goto fail;
1719                   /* No need to call freelocale (base) if copy != base; the
1720                      newlocale function already takes care of doing it.  */
1721                   base = copy;
1722                 }
1723             }
1724         }
1725       else
1726         {
1727           unsigned int i;
1728 
1729           for (i = 0; i < sizeof (categories) / sizeof (categories[0]); i++)
1730             {
1731               int cat_mask = categories[i].mask;
1732 
1733               if ((category_mask & cat_mask) != 0)
1734                 {
1735                   int cat = categories[i].cat;
1736                   const char *name;
1737                   locale_t copy;
1738 
1739                   name = gl_locale_name_environ (cat, category_to_name (cat));
1740                   if (name == NULL)
1741                     name = gl_locale_name_default ();
1742 
1743                   copy = newlocale (cat_mask, name, base);
1744                   if (copy == NULL)
1745                     goto fail;
1746                   /* No need to call freelocale (base) if copy != base; the
1747                      newlocale function already takes care of doing it.  */
1748                   base = copy;
1749                 }
1750             }
1751         }
1752 
1753       /* All steps were successful.  */
1754       return base;
1755 
1756     fail:
1757       if (base != NULL && orig_base == NULL)
1758         {
1759           int saved_errno = errno;
1760           freelocale (base);
1761           errno = saved_errno;
1762         }
1763       return NULL;
1764     }
1765   else
1766     return newlocale (category_mask, locale, base);
1767 }
1768 
1769 # endif
1770 
1771 #endif
1772