3 <<setlocale>>, <<localeconv>>---select or query locale
16 <<setlocale>> is the facility defined by ANSI C to condition the
22 and <<"C">> values for <[locale]>; strings representing other locales are not
31 for a given language, a three character string per ISO 639-3.
37 <<"C-charset">>.
38 Instead of <<"C-">>, you can also specify <<"C.">>.  Both variations allow
40 for instance <<"C.UTF-8">>, which keeps all settings as in the C locale,
41 but uses the UTF-8 charset.
44 <<"UTF-8">>, <<"JIS">>, <<"EUCJP">>, <<"SJIS">>, <<"KOI8-R">>, <<"KOI8-U">>,
45 <<"KOI8-T">>, <<"GEORGIAN-PS">>, <<"PT154">>, <<"TIS-620">>, <<"ISO-8859-x">>
53 <<"EUCKR">> are also recognized with dash, <<"EUC-JP">> and <<"EUC-KR">>.
73 These characters have a width of 1 for singlebyte charsets and UTF-8,
74 and a width of 2 for multibyte charsets other than UTF-8. Specifying
78 to enforce single-width character properties.
86 `<<locale.h>>') describing the locale-specific conventions currently
100 this implementation, always those of the C locale).
103 ANSI C requires <<setlocale>>, but the only locale required across all
104 implementations is the C locale.
107 There is no ISO-8859-12 codepage.  It's also refused by this implementation.
114  * Copyright (c) 1996 - 2002 FreeBSD Project
115  * Copyright (c) 1991, 1993
179  * Default locale per POSIX.  Can be overridden on a per-target base.
182 #define DEFAULT_LOCALE	"C"
194   { "C", "C", "C", "C", "C", "C", "C", },
226   { "C", "C", DEFAULT_LOCALE, "C", "C", "C", "C", },
260 	    = "C";
274       if (strcmp (locale, "POSIX") && strcmp (locale, "C")  in setlocale()
278   return "C";  in setlocale()
293     return category != LC_ALL ? __get_global_locale ()->categories[category]  in setlocale()
300     strcpy (new_categories[i], __get_global_locale ()->categories[i]);  in setlocale()
365 	      if ((len = r - locale) > ENCODING_LEN)  in setlocale()
383 #pragma GCC diagnostic ignored "-Wpragmas"  in setlocale()
384 #pragma GCC diagnostic ignored "-Wunknown-warning-option"  in setlocale()
389 #pragma GCC diagnostic ignored "-Wanalyzer-overlapping-buffers"  in setlocale()
391 	      strcpy (new_categories[i], new_categories[i-1]);  in setlocale()
410       strcpy (saved_categories[i], __get_global_locale ()->categories[i]);  in setlocale()
420 		  strcpy (new_categories[j], "C");  in setlocale()
438   strcpy (global_locale_string, __get_global_locale ()->categories[1]);  in currentlocale()
441     if (strcmp (__get_global_locale ()->categories[1],  in currentlocale()
442 		__get_global_locale ()->categories[i]))  in currentlocale()
448 			 __get_global_locale ()->categories[i]);  in currentlocale()
458   /* At this point a full-featured system would just load the locale  in __loadlocale()
462      one in POSIX-style, or one in the old newlib style to maintain  in __loadlocale()
469   char *end, *c = NULL;  in __loadlocale()  local
482   if (!strcmp (new_locale, loc->categories[category]))  in __loadlocale()
483     return loc->categories[category];  in __loadlocale()
488   /* "POSIX" is translated to "C", as on Linux. */  in __loadlocale()
490     strcpy (locale, "C");  in __loadlocale()
491   if (!strcmp (locale, "C"))				/* Default "C" locale */  in __loadlocale()
493   else if (locale[0] == 'C'  in __loadlocale()
494 	   && (locale[1] == '-'		/* Old newlib style */  in __loadlocale()
495 	       || locale[1] == '.'))	/* Extension for the C locale to allow  in __loadlocale()
497 					   sticking to the C locale in terms  in __loadlocale()
503       c = locale + 2;  in __loadlocale()
504       strcpy (charset, c);  in __loadlocale()
508       c += strlen (charset);  in __loadlocale()
512       c = locale;  in __loadlocale()
516       if (c[0] < 'a' || c[0] > 'z'  in __loadlocale()
517 	  || c[1] < 'a' || c[1] > 'z')  in __loadlocale()
519       c += 2;  in __loadlocale()
520       /* Allow three character Language per ISO 639-3 */  in __loadlocale()
521       if (c[0] >= 'a' && c[0] <= 'z')  in __loadlocale()
522       	++c;  in __loadlocale()
523       if (c[0] == '_')  in __loadlocale()
526 	  ++c;  in __loadlocale()
527 	  if (c[0] < 'A' || c[0] > 'Z'  in __loadlocale()
528 	      || c[1] < 'A' || c[1] > 'Z')  in __loadlocale()
530 	  c += 2;  in __loadlocale()
532       if (c[0] == '.')  in __loadlocale()
537 	  ++c;  in __loadlocale()
538 	  strcpy (charset, c);  in __loadlocale()
542 	  c += strlen (charset);  in __loadlocale()
544       else if (c[0] == '\0' || c[0] == '@')  in __loadlocale()
546 	strcpy (charset, "ISO-8859-1");  in __loadlocale()
551   if (c && c[0] == '@')  in __loadlocale()
553       /* Modifier "cjksingle" is recognized to enforce single-width mode. */  in __loadlocale()
557       if (!strcmp (c + 1, "cjksingle"))  in __loadlocale()
559       else if (!strcmp (c + 1, "cjknarrow"))  in __loadlocale()
561       else if (!strcmp (c + 1, "cjkwide"))  in __loadlocale()
569       if (strcasecmp (charset, "UTF-8") && strcasecmp (charset, "UTF8"))  in __loadlocale()
571       strcpy (charset, "UTF-8");  in __loadlocale()
590       c = charset + 3;  in __loadlocale()
591       if (*c == '-')  in __loadlocale()
592 	++c;  in __loadlocale()
593       if (!strcasecmp (c, "JP"))  in __loadlocale()
614       /* Must be exactly one of ISO-8859-1, [...] ISO-8859-16, except for  in __loadlocale()
615          ISO-8859-12.  This code also recognizes the aliases without dashes. */  in __loadlocale()
618       c = charset + 3;  in __loadlocale()
619       if (*c == '-')  in __loadlocale()
620 	++c;  in __loadlocale()
621       if (strncasecmp (c, "8859", 4))  in __loadlocale()
623       c += 4;  in __loadlocale()
624       if (*c == '-')  in __loadlocale()
625 	++c;  in __loadlocale()
626       val = strtol (c, &end, 10);  in __loadlocale()
629       strcpy (charset, "ISO-8859-");  in __loadlocale()
630       c = charset + 9;  in __loadlocale()
632       	*c++ = '1';  in __loadlocale()
633       *c++ = val % 10 + '0';  in __loadlocale()
634       *c = '\0';  in __loadlocale()
644     case 'C':  in __loadlocale()
645     case 'c':  in __loadlocale()
696       /* KOI8-R, KOI8-U, KOI8-T and the aliases without dash */  in __loadlocale()
699       c = charset + 4;  in __loadlocale()
700       if (*c == '-')  in __loadlocale()
701 	++c;  in __loadlocale()
702       if (*c == 'R' || *c == 'r')  in __loadlocale()
707       else if (*c == 'U' || *c == 'u')  in __loadlocale()
712       else if (*c == 'T' || *c == 't')  in __loadlocale()
739       /* GEORGIAN-PS and the alias without dash */  in __loadlocale()
742 	  c = charset + 8;  in __loadlocale()
743 	  if (*c == '-')  in __loadlocale()
744 	    ++c;  in __loadlocale()
745 	  if (strcasecmp (c, "PS"))  in __loadlocale()
781       c = charset + 3;  in __loadlocale()
782       if (*c == '-')  in __loadlocale()
783 	++c;  in __loadlocale()
784       if (strcmp (c, "620"))  in __loadlocale()
804       strcpy (loc->ctype_codeset, charset);  in __loadlocale()
805       loc->mb_cur_max[0] = mbc_max;  in __loadlocale()
807       loc->wctomb = l_wctomb;  in __loadlocale()
808       loc->mbtowc = l_mbtowc;  in __loadlocale()
810       /* Set CJK width mode (1: ambiguous-wide, 0: normal, -1: disabled). */  in __loadlocale()
813          single-byte charsets, and double width for multi-byte charsets  in __loadlocale()
814          other than UTF-8. For UTF-8, use single width.  in __loadlocale()
818       loc->cjk_lang = cjkwide ||  in __loadlocale()
821 	loc->cjk_lang = -1;	/* Disable CJK dual-width */  in __loadlocale()
830   return strcpy(loc->categories[category], new_locale);  in __loadlocale()
861   return __get_current_ctype_locale ()->mb_cur_max[0];  in __locale_mb_cur_max()
863   return __get_current_locale ()->mb_cur_max[0];  in __locale_mb_cur_max()
871   return locale->ctype_ptr;  in __locale_ctype_ptr_l()
877   return __get_current_locale ()->ctype_ptr;  in __locale_ctype_ptr()