1 | /* |
---|
2 | FUNCTION |
---|
3 | <<setlocale>>, <<localeconv>>---select or query locale |
---|
4 | |
---|
5 | INDEX |
---|
6 | setlocale |
---|
7 | INDEX |
---|
8 | localeconv |
---|
9 | INDEX |
---|
10 | _setlocale_r |
---|
11 | INDEX |
---|
12 | _localeconv_r |
---|
13 | |
---|
14 | SYNOPSIS |
---|
15 | #include <locale.h> |
---|
16 | char *setlocale(int <[category]>, const char *<[locale]>); |
---|
17 | lconv *localeconv(void); |
---|
18 | |
---|
19 | char *_setlocale_r(void *<[reent]>, |
---|
20 | int <[category]>, const char *<[locale]>); |
---|
21 | lconv *_localeconv_r(void *<[reent]>); |
---|
22 | |
---|
23 | DESCRIPTION |
---|
24 | <<setlocale>> is the facility defined by ANSI C to condition the |
---|
25 | execution environment for international collating and formatting |
---|
26 | information; <<localeconv>> reports on the settings of the current |
---|
27 | locale. |
---|
28 | |
---|
29 | This is a minimal implementation, supporting only the required <<"POSIX">> |
---|
30 | and <<"C">> values for <[locale]>; strings representing other locales are not |
---|
31 | honored unless _MB_CAPABLE is defined. |
---|
32 | |
---|
33 | If _MB_CAPABLE is defined, POSIX locale strings are allowed, following |
---|
34 | the form |
---|
35 | |
---|
36 | language[_TERRITORY][.charset][@@modifier] |
---|
37 | |
---|
38 | <<"language">> is a two character string per ISO 639, or, if not available |
---|
39 | for a given language, a three character string per ISO 639-3. |
---|
40 | <<"TERRITORY">> is a country code per ISO 3166. For <<"charset">> and |
---|
41 | <<"modifier">> see below. |
---|
42 | |
---|
43 | Additionally to the POSIX specifier, the following extension is supported |
---|
44 | for backward compatibility with older implementations using newlib: |
---|
45 | <<"C-charset">>. |
---|
46 | Instead of <<"C-">>, you can also specify <<"C.">>. Both variations allow |
---|
47 | to specify language neutral locales while using other charsets than ASCII, |
---|
48 | for instance <<"C.UTF-8">>, which keeps all settings as in the C locale, |
---|
49 | but uses the UTF-8 charset. |
---|
50 | |
---|
51 | The following charsets are recognized: |
---|
52 | <<"UTF-8">>, <<"JIS">>, <<"EUCJP">>, <<"SJIS">>, <<"KOI8-R">>, <<"KOI8-U">>, |
---|
53 | <<"GEORGIAN-PS">>, <<"PT154">>, <<"TIS-620">>, <<"ISO-8859-x">> with |
---|
54 | 1 <= x <= 16, or <<"CPxxx">> with xxx in [437, 720, 737, 775, 850, 852, 855, |
---|
55 | 857, 858, 862, 866, 874, 932, 1125, 1250, 1251, 1252, 1253, 1254, 1255, 1256, |
---|
56 | 1257, 1258]. |
---|
57 | |
---|
58 | Charsets are case insensitive. For instance, <<"EUCJP">> and <<"eucJP">> |
---|
59 | are equivalent. Charset names with dashes can also be written without |
---|
60 | dashes, as in <<"UTF8">>, <<"iso88591">> or <<"koi8r">>. <<"EUCJP">> and |
---|
61 | <<"EUCKR">> are also recognized with dash, <<"EUC-JP">> and <<"EUC-KR">>. |
---|
62 | |
---|
63 | Full support for all of the above charsets requires that newlib has been |
---|
64 | build with multibyte support and support for all ISO and Windows Codepage. |
---|
65 | Otherwise all singlebyte charsets are simply mapped to ASCII. Right now, |
---|
66 | only newlib for Cygwin is built with full charset support by default. |
---|
67 | Under Cygwin, this implementation additionally supports the charsets |
---|
68 | <<"GBK">>, <<"GB2312">>, <<"eucCN">>, <<"eucKR">>, and <<"Big5">>. Cygwin |
---|
69 | does not support <<"JIS">>. |
---|
70 | |
---|
71 | Cygwin additionally supports locales from the file |
---|
72 | /usr/share/locale/locale.alias. |
---|
73 | |
---|
74 | (<<"">> is also accepted; if given, the settings are read from the |
---|
75 | corresponding LC_* environment variables and $LANG according to POSIX rules.) |
---|
76 | |
---|
77 | This implementation also supports the modifier <<"cjknarrow">>, which |
---|
78 | affects how the functions <<wcwidth>> and <<wcswidth>> handle characters |
---|
79 | from the "CJK Ambiguous Width" category of characters described at |
---|
80 | http://www.unicode.org/reports/tr11/#Ambiguous. These characters have a width |
---|
81 | of 1 for singlebyte charsets and a width of 2 for multibyte charsets |
---|
82 | other than UTF-8. For UTF-8, their width depends on the language specifier: |
---|
83 | it is 2 for <<"zh">> (Chinese), <<"ja">> (Japanese), and <<"ko">> (Korean), |
---|
84 | and 1 for everything else. Specifying <<"cjknarrow">> forces a width of 1, |
---|
85 | independent of charset and language. |
---|
86 | |
---|
87 | If you use <<NULL>> as the <[locale]> argument, <<setlocale>> returns a |
---|
88 | pointer to the string representing the current locale. The acceptable |
---|
89 | values for <[category]> are defined in `<<locale.h>>' as macros |
---|
90 | beginning with <<"LC_">>. |
---|
91 | |
---|
92 | <<localeconv>> returns a pointer to a structure (also defined in |
---|
93 | `<<locale.h>>') describing the locale-specific conventions currently |
---|
94 | in effect. |
---|
95 | |
---|
96 | <<_localeconv_r>> and <<_setlocale_r>> are reentrant versions of |
---|
97 | <<localeconv>> and <<setlocale>> respectively. The extra argument |
---|
98 | <[reent]> is a pointer to a reentrancy structure. |
---|
99 | |
---|
100 | RETURNS |
---|
101 | A successful call to <<setlocale>> returns a pointer to a string |
---|
102 | associated with the specified category for the new locale. The string |
---|
103 | returned by <<setlocale>> is such that a subsequent call using that |
---|
104 | string will restore that category (or all categories in case of LC_ALL), |
---|
105 | to that state. The application shall not modify the string returned |
---|
106 | which may be overwritten by a subsequent call to <<setlocale>>. |
---|
107 | On error, <<setlocale>> returns <<NULL>>. |
---|
108 | |
---|
109 | <<localeconv>> returns a pointer to a structure of type <<lconv>>, |
---|
110 | which describes the formatting and collating conventions in effect (in |
---|
111 | this implementation, always those of the C locale). |
---|
112 | |
---|
113 | PORTABILITY |
---|
114 | ANSI C requires <<setlocale>>, but the only locale required across all |
---|
115 | implementations is the C locale. |
---|
116 | |
---|
117 | NOTES |
---|
118 | There is no ISO-8859-12 codepage. It's also refused by this implementation. |
---|
119 | |
---|
120 | No supporting OS subroutines are required. |
---|
121 | */ |
---|
122 | |
---|
123 | /* Parts of this code are originally taken from FreeBSD. */ |
---|
124 | /* |
---|
125 | * Copyright (c) 1996 - 2002 FreeBSD Project |
---|
126 | * Copyright (c) 1991, 1993 |
---|
127 | * The Regents of the University of California. All rights reserved. |
---|
128 | * |
---|
129 | * This code is derived from software contributed to Berkeley by |
---|
130 | * Paul Borman at Krystal Technologies. |
---|
131 | * |
---|
132 | * Redistribution and use in source and binary forms, with or without |
---|
133 | * modification, are permitted provided that the following conditions |
---|
134 | * are met: |
---|
135 | * 1. Redistributions of source code must retain the above copyright |
---|
136 | * notice, this list of conditions and the following disclaimer. |
---|
137 | * 2. Redistributions in binary form must reproduce the above copyright |
---|
138 | * notice, this list of conditions and the following disclaimer in the |
---|
139 | * documentation and/or other materials provided with the distribution. |
---|
140 | * 4. Neither the name of the University nor the names of its contributors |
---|
141 | * may be used to endorse or promote products derived from this software |
---|
142 | * without specific prior written permission. |
---|
143 | * |
---|
144 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND |
---|
145 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
---|
146 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
---|
147 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE |
---|
148 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
---|
149 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
---|
150 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
---|
151 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
---|
152 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
---|
153 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
---|
154 | * SUCH DAMAGE. |
---|
155 | */ |
---|
156 | |
---|
157 | #include <newlib.h> |
---|
158 | #include <errno.h> |
---|
159 | #include <string.h> |
---|
160 | #include <limits.h> |
---|
161 | #include <reent.h> |
---|
162 | #include <stdlib.h> |
---|
163 | #include <wchar.h> |
---|
164 | #include "setlocale.h" |
---|
165 | #include "../ctype/ctype_.h" |
---|
166 | #include "../stdlib/local.h" |
---|
167 | |
---|
168 | #ifdef __CYGWIN__ /* Has to be kept available as exported symbol for |
---|
169 | backward compatibility. Set it in setlocale, but |
---|
170 | otherwise ignore it. Applications compiled after |
---|
171 | 2010 don't use it anymore. */ |
---|
172 | int __EXPORT __mb_cur_max = 6; |
---|
173 | #endif |
---|
174 | |
---|
175 | char *_PathLocale = NULL; |
---|
176 | |
---|
177 | #ifdef _MB_CAPABLE |
---|
178 | /* |
---|
179 | * Category names for getenv() |
---|
180 | */ |
---|
181 | static char *categories[_LC_LAST] = { |
---|
182 | "LC_ALL", |
---|
183 | "LC_COLLATE", |
---|
184 | "LC_CTYPE", |
---|
185 | "LC_MONETARY", |
---|
186 | "LC_NUMERIC", |
---|
187 | "LC_TIME", |
---|
188 | "LC_MESSAGES", |
---|
189 | }; |
---|
190 | #endif /* _MB_CAPABLE */ |
---|
191 | |
---|
192 | /* |
---|
193 | * Default locale per POSIX. Can be overridden on a per-target base. |
---|
194 | */ |
---|
195 | #ifndef DEFAULT_LOCALE |
---|
196 | #define DEFAULT_LOCALE "C" |
---|
197 | #endif |
---|
198 | |
---|
199 | #ifdef _MB_CAPABLE |
---|
200 | /* |
---|
201 | * This variable can be changed by any outside mechanism. This allows, |
---|
202 | * for instance, to load the default locale from a file. |
---|
203 | */ |
---|
204 | char __default_locale[ENCODING_LEN + 1] = DEFAULT_LOCALE; |
---|
205 | |
---|
206 | const struct __locale_t __C_locale = |
---|
207 | { |
---|
208 | { "C", "C", "C", "C", "C", "C", "C", }, |
---|
209 | __ascii_wctomb, |
---|
210 | __ascii_mbtowc, |
---|
211 | 0, |
---|
212 | DEFAULT_CTYPE_PTR, |
---|
213 | { |
---|
214 | ".", "", "", "", "", "", "", "", "", "", |
---|
215 | CHAR_MAX, CHAR_MAX, CHAR_MAX, CHAR_MAX, |
---|
216 | CHAR_MAX, CHAR_MAX, CHAR_MAX, CHAR_MAX, |
---|
217 | CHAR_MAX, CHAR_MAX, CHAR_MAX, CHAR_MAX, |
---|
218 | CHAR_MAX, CHAR_MAX |
---|
219 | }, |
---|
220 | #ifndef __HAVE_LOCALE_INFO__ |
---|
221 | "\1", |
---|
222 | "ASCII", |
---|
223 | "ASCII", |
---|
224 | #else /* __HAVE_LOCALE_INFO__ */ |
---|
225 | { |
---|
226 | { NULL, NULL }, /* LC_ALL */ |
---|
227 | #ifdef __CYGWIN__ |
---|
228 | { &_C_collate_locale, NULL }, /* LC_COLLATE */ |
---|
229 | #else |
---|
230 | { NULL, NULL }, /* LC_COLLATE */ |
---|
231 | #endif |
---|
232 | { &_C_ctype_locale, NULL }, /* LC_CTYPE */ |
---|
233 | { &_C_monetary_locale, NULL }, /* LC_MONETARY */ |
---|
234 | { &_C_numeric_locale, NULL }, /* LC_NUMERIC */ |
---|
235 | { &_C_time_locale, NULL }, /* LC_TIME */ |
---|
236 | { &_C_messages_locale, NULL }, /* LC_MESSAGES */ |
---|
237 | }, |
---|
238 | #endif /* __HAVE_LOCALE_INFO__ */ |
---|
239 | }; |
---|
240 | #endif /* _MB_CAPABLE */ |
---|
241 | |
---|
242 | struct __locale_t __global_locale = |
---|
243 | { |
---|
244 | { "C", "C", DEFAULT_LOCALE, "C", "C", "C", "C", }, |
---|
245 | #ifdef __CYGWIN__ |
---|
246 | __utf8_wctomb, |
---|
247 | __utf8_mbtowc, |
---|
248 | #else |
---|
249 | __ascii_wctomb, |
---|
250 | __ascii_mbtowc, |
---|
251 | #endif |
---|
252 | 0, |
---|
253 | DEFAULT_CTYPE_PTR, |
---|
254 | { |
---|
255 | ".", "", "", "", "", "", "", "", "", "", |
---|
256 | CHAR_MAX, CHAR_MAX, CHAR_MAX, CHAR_MAX, |
---|
257 | CHAR_MAX, CHAR_MAX, CHAR_MAX, CHAR_MAX, |
---|
258 | CHAR_MAX, CHAR_MAX, CHAR_MAX, CHAR_MAX, |
---|
259 | CHAR_MAX, CHAR_MAX |
---|
260 | }, |
---|
261 | #ifndef __HAVE_LOCALE_INFO__ |
---|
262 | "\1", |
---|
263 | "ASCII", |
---|
264 | "ASCII", |
---|
265 | #else /* __HAVE_LOCALE_INFO__ */ |
---|
266 | { |
---|
267 | { NULL, NULL }, /* LC_ALL */ |
---|
268 | #ifdef __CYGWIN__ |
---|
269 | { &_C_collate_locale, NULL }, /* LC_COLLATE */ |
---|
270 | #else |
---|
271 | { NULL, NULL }, /* LC_COLLATE */ |
---|
272 | #endif |
---|
273 | { &_C_ctype_locale, NULL }, /* LC_CTYPE */ |
---|
274 | { &_C_monetary_locale, NULL }, /* LC_MONETARY */ |
---|
275 | { &_C_numeric_locale, NULL }, /* LC_NUMERIC */ |
---|
276 | { &_C_time_locale, NULL }, /* LC_TIME */ |
---|
277 | { &_C_messages_locale, NULL }, /* LC_MESSAGES */ |
---|
278 | }, |
---|
279 | #endif /* __HAVE_LOCALE_INFO__ */ |
---|
280 | }; |
---|
281 | |
---|
282 | #ifdef _MB_CAPABLE |
---|
283 | /* Renamed from current_locale_string to make clear this is only the |
---|
284 | *global* string for setlocale (LC_ALL, NULL). There's no equivalent |
---|
285 | functionality for uselocale. */ |
---|
286 | static char global_locale_string[_LC_LAST * (ENCODING_LEN + 1/*"/"*/ + 1)]; |
---|
287 | static char *currentlocale (void); |
---|
288 | |
---|
289 | #endif /* _MB_CAPABLE */ |
---|
290 | |
---|
291 | char * |
---|
292 | _setlocale_r (struct _reent *p, |
---|
293 | int category, |
---|
294 | const char *locale) |
---|
295 | { |
---|
296 | #ifndef _MB_CAPABLE |
---|
297 | if (locale) |
---|
298 | { |
---|
299 | if (strcmp (locale, "POSIX") && strcmp (locale, "C") |
---|
300 | && strcmp (locale, "")) |
---|
301 | return NULL; |
---|
302 | } |
---|
303 | return "C"; |
---|
304 | #else /* _MB_CAPABLE */ |
---|
305 | static char new_categories[_LC_LAST][ENCODING_LEN + 1]; |
---|
306 | static char saved_categories[_LC_LAST][ENCODING_LEN + 1]; |
---|
307 | int i, j, len, saverr; |
---|
308 | const char *env, *r; |
---|
309 | |
---|
310 | if (category < LC_ALL || category >= _LC_LAST) |
---|
311 | { |
---|
312 | p->_errno = EINVAL; |
---|
313 | return NULL; |
---|
314 | } |
---|
315 | |
---|
316 | if (locale == NULL) |
---|
317 | return category != LC_ALL ? __get_global_locale ()->categories[category] |
---|
318 | : currentlocale(); |
---|
319 | |
---|
320 | /* |
---|
321 | * Default to the current locale for everything. |
---|
322 | */ |
---|
323 | for (i = 1; i < _LC_LAST; ++i) |
---|
324 | strcpy (new_categories[i], __get_global_locale ()->categories[i]); |
---|
325 | |
---|
326 | /* |
---|
327 | * Now go fill up new_categories from the locale argument |
---|
328 | */ |
---|
329 | if (!*locale) |
---|
330 | { |
---|
331 | if (category == LC_ALL) |
---|
332 | { |
---|
333 | for (i = 1; i < _LC_LAST; ++i) |
---|
334 | { |
---|
335 | env = __get_locale_env (p, i); |
---|
336 | if (strlen (env) > ENCODING_LEN) |
---|
337 | { |
---|
338 | p->_errno = EINVAL; |
---|
339 | return NULL; |
---|
340 | } |
---|
341 | strcpy (new_categories[i], env); |
---|
342 | } |
---|
343 | } |
---|
344 | else |
---|
345 | { |
---|
346 | env = __get_locale_env (p, category); |
---|
347 | if (strlen (env) > ENCODING_LEN) |
---|
348 | { |
---|
349 | p->_errno = EINVAL; |
---|
350 | return NULL; |
---|
351 | } |
---|
352 | strcpy (new_categories[category], env); |
---|
353 | } |
---|
354 | } |
---|
355 | else if (category != LC_ALL) |
---|
356 | { |
---|
357 | if (strlen (locale) > ENCODING_LEN) |
---|
358 | { |
---|
359 | p->_errno = EINVAL; |
---|
360 | return NULL; |
---|
361 | } |
---|
362 | strcpy (new_categories[category], locale); |
---|
363 | } |
---|
364 | else |
---|
365 | { |
---|
366 | if ((r = strchr (locale, '/')) == NULL) |
---|
367 | { |
---|
368 | if (strlen (locale) > ENCODING_LEN) |
---|
369 | { |
---|
370 | p->_errno = EINVAL; |
---|
371 | return NULL; |
---|
372 | } |
---|
373 | for (i = 1; i < _LC_LAST; ++i) |
---|
374 | strcpy (new_categories[i], locale); |
---|
375 | } |
---|
376 | else |
---|
377 | { |
---|
378 | for (i = 1; r[1] == '/'; ++r) |
---|
379 | ; |
---|
380 | if (!r[1]) |
---|
381 | { |
---|
382 | p->_errno = EINVAL; |
---|
383 | return NULL; /* Hmm, just slashes... */ |
---|
384 | } |
---|
385 | do |
---|
386 | { |
---|
387 | if (i == _LC_LAST) |
---|
388 | break; /* Too many slashes... */ |
---|
389 | if ((len = r - locale) > ENCODING_LEN) |
---|
390 | { |
---|
391 | p->_errno = EINVAL; |
---|
392 | return NULL; |
---|
393 | } |
---|
394 | strlcpy (new_categories[i], locale, len + 1); |
---|
395 | i++; |
---|
396 | while (*r == '/') |
---|
397 | r++; |
---|
398 | locale = r; |
---|
399 | while (*r && *r != '/') |
---|
400 | r++; |
---|
401 | } |
---|
402 | while (*locale); |
---|
403 | while (i < _LC_LAST) |
---|
404 | { |
---|
405 | strcpy (new_categories[i], new_categories[i-1]); |
---|
406 | i++; |
---|
407 | } |
---|
408 | } |
---|
409 | } |
---|
410 | |
---|
411 | if (category != LC_ALL) |
---|
412 | return __loadlocale (__get_global_locale (), category, |
---|
413 | new_categories[category]); |
---|
414 | |
---|
415 | for (i = 1; i < _LC_LAST; ++i) |
---|
416 | { |
---|
417 | strcpy (saved_categories[i], __get_global_locale ()->categories[i]); |
---|
418 | if (__loadlocale (__get_global_locale (), i, new_categories[i]) == NULL) |
---|
419 | { |
---|
420 | saverr = p->_errno; |
---|
421 | for (j = 1; j < i; j++) |
---|
422 | { |
---|
423 | strcpy (new_categories[j], saved_categories[j]); |
---|
424 | if (__loadlocale (__get_global_locale (), j, new_categories[j]) |
---|
425 | == NULL) |
---|
426 | { |
---|
427 | strcpy (new_categories[j], "C"); |
---|
428 | __loadlocale (__get_global_locale (), j, new_categories[j]); |
---|
429 | } |
---|
430 | } |
---|
431 | p->_errno = saverr; |
---|
432 | return NULL; |
---|
433 | } |
---|
434 | } |
---|
435 | return currentlocale (); |
---|
436 | #endif /* _MB_CAPABLE */ |
---|
437 | } |
---|
438 | |
---|
439 | #ifdef _MB_CAPABLE |
---|
440 | static char * |
---|
441 | currentlocale () |
---|
442 | { |
---|
443 | int i; |
---|
444 | |
---|
445 | strcpy (global_locale_string, __get_global_locale ()->categories[1]); |
---|
446 | |
---|
447 | for (i = 2; i < _LC_LAST; ++i) |
---|
448 | if (strcmp (__get_global_locale ()->categories[1], |
---|
449 | __get_global_locale ()->categories[i])) |
---|
450 | { |
---|
451 | for (i = 2; i < _LC_LAST; ++i) |
---|
452 | { |
---|
453 | (void)strcat(global_locale_string, "/"); |
---|
454 | (void)strcat(global_locale_string, |
---|
455 | __get_global_locale ()->categories[i]); |
---|
456 | } |
---|
457 | break; |
---|
458 | } |
---|
459 | return global_locale_string; |
---|
460 | } |
---|
461 | |
---|
462 | extern void __set_ctype (struct __locale_t *, const char *charset); |
---|
463 | |
---|
464 | char * |
---|
465 | __loadlocale (struct __locale_t *loc, int category, const char *new_locale) |
---|
466 | { |
---|
467 | /* At this point a full-featured system would just load the locale |
---|
468 | specific data from the locale files. |
---|
469 | What we do here for now is to check the incoming string for correctness. |
---|
470 | The string must be in one of the allowed locale strings, either |
---|
471 | one in POSIX-style, or one in the old newlib style to maintain |
---|
472 | backward compatibility. If the local string is correct, the charset |
---|
473 | is extracted and stored in ctype_codeset or message_charset |
---|
474 | dependent on the cateogry. */ |
---|
475 | char *locale = NULL; |
---|
476 | char charset[ENCODING_LEN + 1]; |
---|
477 | long val = 0; |
---|
478 | char *end, *c = NULL; |
---|
479 | int mbc_max; |
---|
480 | wctomb_p l_wctomb; |
---|
481 | mbtowc_p l_mbtowc; |
---|
482 | int cjknarrow = 0; |
---|
483 | |
---|
484 | /* Avoid doing everything twice if nothing has changed. |
---|
485 | |
---|
486 | duplocale relies on this test to go wrong so the locale is actually |
---|
487 | duplicated when required. Any change here has to be synced with a |
---|
488 | matching change in duplocale. */ |
---|
489 | if (!strcmp (new_locale, loc->categories[category])) |
---|
490 | return loc->categories[category]; |
---|
491 | |
---|
492 | #ifdef __CYGWIN__ |
---|
493 | /* This additional code handles the case that the incoming locale string |
---|
494 | is not valid. If so, it calls the function __set_locale_from_locale_alias, |
---|
495 | which is only available on Cygwin right now. The function reads the |
---|
496 | file /usr/share/locale/locale.alias. The file contains locale aliases |
---|
497 | and their replacement locale. For instance, the alias "french" is |
---|
498 | translated to "fr_FR.ISO-8859-1", the alias "thai" is translated to |
---|
499 | "th_TH.TIS-620". If successful, the function returns with a pointer |
---|
500 | to the second argument, which is a buffer in which the replacement locale |
---|
501 | gets stored. Otherwise the function returns NULL. */ |
---|
502 | char tmp_locale[ENCODING_LEN + 1]; |
---|
503 | int ret = 0; |
---|
504 | |
---|
505 | restart: |
---|
506 | if (!locale) |
---|
507 | locale = (char *) new_locale; |
---|
508 | else if (locale != tmp_locale) |
---|
509 | { |
---|
510 | locale = __set_locale_from_locale_alias (locale, tmp_locale); |
---|
511 | if (!locale) |
---|
512 | return NULL; |
---|
513 | } |
---|
514 | # define FAIL goto restart |
---|
515 | #else |
---|
516 | locale = new_locale; |
---|
517 | # define FAIL return NULL |
---|
518 | #endif |
---|
519 | |
---|
520 | /* "POSIX" is translated to "C", as on Linux. */ |
---|
521 | if (!strcmp (locale, "POSIX")) |
---|
522 | strcpy (locale, "C"); |
---|
523 | if (!strcmp (locale, "C")) /* Default "C" locale */ |
---|
524 | strcpy (charset, "ASCII"); |
---|
525 | else if (locale[0] == 'C' |
---|
526 | && (locale[1] == '-' /* Old newlib style */ |
---|
527 | || locale[1] == '.')) /* Extension for the C locale to allow |
---|
528 | specifying different charsets while |
---|
529 | sticking to the C locale in terms |
---|
530 | of sort order, etc. Proposed in |
---|
531 | the Debian project. */ |
---|
532 | { |
---|
533 | char *chp; |
---|
534 | |
---|
535 | c = locale + 2; |
---|
536 | strcpy (charset, c); |
---|
537 | if ((chp = strchr (charset, '@'))) |
---|
538 | /* Strip off modifier */ |
---|
539 | *chp = '\0'; |
---|
540 | c += strlen (charset); |
---|
541 | } |
---|
542 | else /* POSIX style */ |
---|
543 | { |
---|
544 | c = locale; |
---|
545 | |
---|
546 | /* Don't use ctype macros here, they might be localized. */ |
---|
547 | /* Language */ |
---|
548 | if (c[0] < 'a' || c[0] > 'z' |
---|
549 | || c[1] < 'a' || c[1] > 'z') |
---|
550 | FAIL; |
---|
551 | c += 2; |
---|
552 | /* Allow three character Language per ISO 639-3 */ |
---|
553 | if (c[0] >= 'a' && c[0] <= 'z') |
---|
554 | ++c; |
---|
555 | if (c[0] == '_') |
---|
556 | { |
---|
557 | /* Territory */ |
---|
558 | ++c; |
---|
559 | if (c[0] < 'A' || c[0] > 'Z' |
---|
560 | || c[1] < 'A' || c[1] > 'Z') |
---|
561 | FAIL; |
---|
562 | c += 2; |
---|
563 | } |
---|
564 | if (c[0] == '.') |
---|
565 | { |
---|
566 | /* Charset */ |
---|
567 | char *chp; |
---|
568 | |
---|
569 | ++c; |
---|
570 | strcpy (charset, c); |
---|
571 | if ((chp = strchr (charset, '@'))) |
---|
572 | /* Strip off modifier */ |
---|
573 | *chp = '\0'; |
---|
574 | c += strlen (charset); |
---|
575 | } |
---|
576 | else if (c[0] == '\0' || c[0] == '@') |
---|
577 | /* End of string or just a modifier */ |
---|
578 | #ifdef __CYGWIN__ |
---|
579 | /* The Cygwin-only function __set_charset_from_locale checks |
---|
580 | for the default charset which is connected to the given locale. |
---|
581 | The function uses Windows functions in turn so it can't be easily |
---|
582 | adapted to other targets. However, if any other target provides |
---|
583 | equivalent functionality, preferrably using the same function name |
---|
584 | it would be sufficient to change the guarding #ifdef. */ |
---|
585 | __set_charset_from_locale (locale, charset); |
---|
586 | #else |
---|
587 | strcpy (charset, "ISO-8859-1"); |
---|
588 | #endif |
---|
589 | else |
---|
590 | /* Invalid string */ |
---|
591 | FAIL; |
---|
592 | } |
---|
593 | if (c && c[0] == '@') |
---|
594 | { |
---|
595 | /* Modifier */ |
---|
596 | /* Only one modifier is recognized right now. "cjknarrow" is used |
---|
597 | to modify the behaviour of wcwidth() for East Asian languages. |
---|
598 | For details see the comment at the end of this function. */ |
---|
599 | if (!strcmp (c + 1, "cjknarrow")) |
---|
600 | cjknarrow = 1; |
---|
601 | } |
---|
602 | /* We only support this subset of charsets. */ |
---|
603 | switch (charset[0]) |
---|
604 | { |
---|
605 | case 'U': |
---|
606 | case 'u': |
---|
607 | if (strcasecmp (charset, "UTF-8") && strcasecmp (charset, "UTF8")) |
---|
608 | FAIL; |
---|
609 | strcpy (charset, "UTF-8"); |
---|
610 | mbc_max = 6; |
---|
611 | l_wctomb = __utf8_wctomb; |
---|
612 | l_mbtowc = __utf8_mbtowc; |
---|
613 | break; |
---|
614 | #ifndef __CYGWIN__ |
---|
615 | /* Cygwin does not support JIS at all. */ |
---|
616 | case 'J': |
---|
617 | case 'j': |
---|
618 | if (strcasecmp (charset, "JIS")) |
---|
619 | FAIL; |
---|
620 | strcpy (charset, "JIS"); |
---|
621 | mbc_max = 8; |
---|
622 | l_wctomb = __jis_wctomb; |
---|
623 | l_mbtowc = __jis_mbtowc; |
---|
624 | break; |
---|
625 | #endif /* !__CYGWIN__ */ |
---|
626 | case 'E': |
---|
627 | case 'e': |
---|
628 | if (strncasecmp (charset, "EUC", 3)) |
---|
629 | FAIL; |
---|
630 | c = charset + 3; |
---|
631 | if (*c == '-') |
---|
632 | ++c; |
---|
633 | if (!strcasecmp (c, "JP")) |
---|
634 | { |
---|
635 | strcpy (charset, "EUCJP"); |
---|
636 | mbc_max = 3; |
---|
637 | l_wctomb = __eucjp_wctomb; |
---|
638 | l_mbtowc = __eucjp_mbtowc; |
---|
639 | } |
---|
640 | #ifdef __CYGWIN__ |
---|
641 | /* Newlib does neither provide EUC-KR nor EUC-CN, and Cygwin's |
---|
642 | implementation requires Windows support. */ |
---|
643 | else if (!strcasecmp (c, "KR")) |
---|
644 | { |
---|
645 | strcpy (charset, "EUCKR"); |
---|
646 | mbc_max = 2; |
---|
647 | l_wctomb = __kr_wctomb; |
---|
648 | l_mbtowc = __kr_mbtowc; |
---|
649 | } |
---|
650 | else if (!strcasecmp (c, "CN")) |
---|
651 | { |
---|
652 | strcpy (charset, "EUCCN"); |
---|
653 | mbc_max = 2; |
---|
654 | l_wctomb = __gbk_wctomb; |
---|
655 | l_mbtowc = __gbk_mbtowc; |
---|
656 | } |
---|
657 | #endif /* __CYGWIN__ */ |
---|
658 | else |
---|
659 | FAIL; |
---|
660 | break; |
---|
661 | case 'S': |
---|
662 | case 's': |
---|
663 | if (strcasecmp (charset, "SJIS")) |
---|
664 | FAIL; |
---|
665 | strcpy (charset, "SJIS"); |
---|
666 | mbc_max = 2; |
---|
667 | l_wctomb = __sjis_wctomb; |
---|
668 | l_mbtowc = __sjis_mbtowc; |
---|
669 | break; |
---|
670 | case 'I': |
---|
671 | case 'i': |
---|
672 | /* Must be exactly one of ISO-8859-1, [...] ISO-8859-16, except for |
---|
673 | ISO-8859-12. This code also recognizes the aliases without dashes. */ |
---|
674 | if (strncasecmp (charset, "ISO", 3)) |
---|
675 | FAIL; |
---|
676 | c = charset + 3; |
---|
677 | if (*c == '-') |
---|
678 | ++c; |
---|
679 | if (strncasecmp (c, "8859", 4)) |
---|
680 | FAIL; |
---|
681 | c += 4; |
---|
682 | if (*c == '-') |
---|
683 | ++c; |
---|
684 | val = strtol (c, &end, 10); |
---|
685 | if (val < 1 || val > 16 || val == 12 || *end) |
---|
686 | FAIL; |
---|
687 | strcpy (charset, "ISO-8859-"); |
---|
688 | c = charset + 9; |
---|
689 | if (val > 10) |
---|
690 | *c++ = '1'; |
---|
691 | *c++ = val % 10 + '0'; |
---|
692 | *c = '\0'; |
---|
693 | mbc_max = 1; |
---|
694 | #ifdef _MB_EXTENDED_CHARSETS_ISO |
---|
695 | l_wctomb = __iso_wctomb (val); |
---|
696 | l_mbtowc = __iso_mbtowc (val); |
---|
697 | #else /* !_MB_EXTENDED_CHARSETS_ISO */ |
---|
698 | l_wctomb = __ascii_wctomb; |
---|
699 | l_mbtowc = __ascii_mbtowc; |
---|
700 | #endif /* _MB_EXTENDED_CHARSETS_ISO */ |
---|
701 | break; |
---|
702 | case 'C': |
---|
703 | case 'c': |
---|
704 | if (charset[1] != 'P' && charset[1] != 'p') |
---|
705 | FAIL; |
---|
706 | strncpy (charset, "CP", 2); |
---|
707 | val = strtol (charset + 2, &end, 10); |
---|
708 | if (*end) |
---|
709 | FAIL; |
---|
710 | switch (val) |
---|
711 | { |
---|
712 | case 437: |
---|
713 | case 720: |
---|
714 | case 737: |
---|
715 | case 775: |
---|
716 | case 850: |
---|
717 | case 852: |
---|
718 | case 855: |
---|
719 | case 857: |
---|
720 | case 858: |
---|
721 | case 862: |
---|
722 | case 866: |
---|
723 | case 874: |
---|
724 | case 1125: |
---|
725 | case 1250: |
---|
726 | case 1251: |
---|
727 | case 1252: |
---|
728 | case 1253: |
---|
729 | case 1254: |
---|
730 | case 1255: |
---|
731 | case 1256: |
---|
732 | case 1257: |
---|
733 | case 1258: |
---|
734 | mbc_max = 1; |
---|
735 | #ifdef _MB_EXTENDED_CHARSETS_WINDOWS |
---|
736 | l_wctomb = __cp_wctomb (val); |
---|
737 | l_mbtowc = __cp_mbtowc (val); |
---|
738 | #else /* !_MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
739 | l_wctomb = __ascii_wctomb; |
---|
740 | l_mbtowc = __ascii_mbtowc; |
---|
741 | #endif /* _MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
742 | break; |
---|
743 | case 932: |
---|
744 | mbc_max = 2; |
---|
745 | l_wctomb = __sjis_wctomb; |
---|
746 | l_mbtowc = __sjis_mbtowc; |
---|
747 | break; |
---|
748 | default: |
---|
749 | FAIL; |
---|
750 | } |
---|
751 | break; |
---|
752 | case 'K': |
---|
753 | case 'k': |
---|
754 | /* KOI8-R, KOI8-U and the aliases without dash */ |
---|
755 | if (strncasecmp (charset, "KOI8", 4)) |
---|
756 | FAIL; |
---|
757 | c = charset + 4; |
---|
758 | if (*c == '-') |
---|
759 | ++c; |
---|
760 | if (*c == 'R' || *c == 'r') |
---|
761 | { |
---|
762 | val = 20866; |
---|
763 | strcpy (charset, "CP20866"); |
---|
764 | } |
---|
765 | else if (*c == 'U' || *c == 'u') |
---|
766 | { |
---|
767 | val = 21866; |
---|
768 | strcpy (charset, "CP21866"); |
---|
769 | } |
---|
770 | else |
---|
771 | FAIL; |
---|
772 | mbc_max = 1; |
---|
773 | #ifdef _MB_EXTENDED_CHARSETS_WINDOWS |
---|
774 | l_wctomb = __cp_wctomb (val); |
---|
775 | l_mbtowc = __cp_mbtowc (val); |
---|
776 | #else /* !_MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
777 | l_wctomb = __ascii_wctomb; |
---|
778 | l_mbtowc = __ascii_mbtowc; |
---|
779 | #endif /* _MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
780 | break; |
---|
781 | case 'A': |
---|
782 | case 'a': |
---|
783 | if (strcasecmp (charset, "ASCII")) |
---|
784 | FAIL; |
---|
785 | strcpy (charset, "ASCII"); |
---|
786 | mbc_max = 1; |
---|
787 | l_wctomb = __ascii_wctomb; |
---|
788 | l_mbtowc = __ascii_mbtowc; |
---|
789 | break; |
---|
790 | case 'G': |
---|
791 | case 'g': |
---|
792 | #ifdef __CYGWIN__ |
---|
793 | /* Newlib does not provide GBK/GB2312 and Cygwin's implementation |
---|
794 | requires Windows support. */ |
---|
795 | if (!strcasecmp (charset, "GBK") |
---|
796 | || !strcasecmp (charset, "GB2312")) |
---|
797 | { |
---|
798 | strcpy (charset, charset[2] == '2' ? "GB2312" : "GBK"); |
---|
799 | mbc_max = 2; |
---|
800 | l_wctomb = __gbk_wctomb; |
---|
801 | l_mbtowc = __gbk_mbtowc; |
---|
802 | } |
---|
803 | else |
---|
804 | #endif /* __CYGWIN__ */ |
---|
805 | /* GEORGIAN-PS and the alias without dash */ |
---|
806 | if (!strncasecmp (charset, "GEORGIAN", 8)) |
---|
807 | { |
---|
808 | c = charset + 8; |
---|
809 | if (*c == '-') |
---|
810 | ++c; |
---|
811 | if (strcasecmp (c, "PS")) |
---|
812 | FAIL; |
---|
813 | val = 101; |
---|
814 | strcpy (charset, "CP101"); |
---|
815 | mbc_max = 1; |
---|
816 | #ifdef _MB_EXTENDED_CHARSETS_WINDOWS |
---|
817 | l_wctomb = __cp_wctomb (val); |
---|
818 | l_mbtowc = __cp_mbtowc (val); |
---|
819 | #else /* !_MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
820 | l_wctomb = __ascii_wctomb; |
---|
821 | l_mbtowc = __ascii_mbtowc; |
---|
822 | #endif /* _MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
823 | } |
---|
824 | else |
---|
825 | FAIL; |
---|
826 | break; |
---|
827 | case 'P': |
---|
828 | case 'p': |
---|
829 | /* PT154 */ |
---|
830 | if (strcasecmp (charset, "PT154")) |
---|
831 | FAIL; |
---|
832 | val = 102; |
---|
833 | strcpy (charset, "CP102"); |
---|
834 | mbc_max = 1; |
---|
835 | #ifdef _MB_EXTENDED_CHARSETS_WINDOWS |
---|
836 | l_wctomb = __cp_wctomb (val); |
---|
837 | l_mbtowc = __cp_mbtowc (val); |
---|
838 | #else /* !_MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
839 | l_wctomb = __ascii_wctomb; |
---|
840 | l_mbtowc = __ascii_mbtowc; |
---|
841 | #endif /* _MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
842 | break; |
---|
843 | case 'T': |
---|
844 | case 't': |
---|
845 | if (strncasecmp (charset, "TIS", 3)) |
---|
846 | FAIL; |
---|
847 | c = charset + 3; |
---|
848 | if (*c == '-') |
---|
849 | ++c; |
---|
850 | if (strcasecmp (c, "620")) |
---|
851 | FAIL; |
---|
852 | val = 874; |
---|
853 | strcpy (charset, "CP874"); |
---|
854 | mbc_max = 1; |
---|
855 | #ifdef _MB_EXTENDED_CHARSETS_WINDOWS |
---|
856 | l_wctomb = __cp_wctomb (val); |
---|
857 | l_mbtowc = __cp_mbtowc (val); |
---|
858 | #else /* !_MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
859 | l_wctomb = __ascii_wctomb; |
---|
860 | l_mbtowc = __ascii_mbtowc; |
---|
861 | #endif /* _MB_EXTENDED_CHARSETS_WINDOWS */ |
---|
862 | break; |
---|
863 | #ifdef __CYGWIN__ |
---|
864 | /* Newlib does not provide Big5 and Cygwin's implementation |
---|
865 | requires Windows support. */ |
---|
866 | case 'B': |
---|
867 | case 'b': |
---|
868 | if (strcasecmp (charset, "BIG5")) |
---|
869 | FAIL; |
---|
870 | strcpy (charset, "BIG5"); |
---|
871 | mbc_max = 2; |
---|
872 | l_wctomb = __big5_wctomb; |
---|
873 | l_mbtowc = __big5_mbtowc; |
---|
874 | break; |
---|
875 | #endif /* __CYGWIN__ */ |
---|
876 | default: |
---|
877 | FAIL; |
---|
878 | } |
---|
879 | switch (category) |
---|
880 | { |
---|
881 | case LC_CTYPE: |
---|
882 | #ifndef __HAVE_LOCALE_INFO__ |
---|
883 | strcpy (loc->ctype_codeset, charset); |
---|
884 | loc->mb_cur_max[0] = mbc_max; |
---|
885 | #endif |
---|
886 | #ifdef __CYGWIN__ |
---|
887 | __mb_cur_max = mbc_max; /* Only for backward compat */ |
---|
888 | #endif |
---|
889 | loc->wctomb = l_wctomb; |
---|
890 | loc->mbtowc = l_mbtowc; |
---|
891 | __set_ctype (loc, charset); |
---|
892 | /* Determine the width for the "CJK Ambiguous Width" category of |
---|
893 | characters. This is used in wcwidth(). Assume single width for |
---|
894 | single-byte charsets, and double width for multi-byte charsets |
---|
895 | other than UTF-8. For UTF-8, use double width for the East Asian |
---|
896 | languages ("ja", "ko", "zh"), and single width for everything else. |
---|
897 | Single width can also be forced with the "@cjknarrow" modifier. */ |
---|
898 | loc->cjk_lang = !cjknarrow && mbc_max > 1 |
---|
899 | && (charset[0] != 'U' |
---|
900 | || strncmp (locale, "ja", 2) == 0 |
---|
901 | || strncmp (locale, "ko", 2) == 0 |
---|
902 | || strncmp (locale, "zh", 2) == 0); |
---|
903 | #ifdef __HAVE_LOCALE_INFO__ |
---|
904 | ret = __ctype_load_locale (loc, locale, (void *) l_wctomb, charset, |
---|
905 | mbc_max); |
---|
906 | #endif /* __HAVE_LOCALE_INFO__ */ |
---|
907 | break; |
---|
908 | case LC_MESSAGES: |
---|
909 | #ifdef __HAVE_LOCALE_INFO__ |
---|
910 | ret = __messages_load_locale (loc, locale, (void *) l_wctomb, charset); |
---|
911 | if (!ret) |
---|
912 | #else |
---|
913 | strcpy (loc->message_codeset, charset); |
---|
914 | #endif /* __HAVE_LOCALE_INFO__ */ |
---|
915 | break; |
---|
916 | #ifdef __HAVE_LOCALE_INFO__ |
---|
917 | #ifdef __CYGWIN__ |
---|
918 | /* Right now only Cygwin supports a __collate_load_locale function at all. */ |
---|
919 | case LC_COLLATE: |
---|
920 | ret = __collate_load_locale (loc, locale, (void *) l_mbtowc, charset); |
---|
921 | break; |
---|
922 | #endif |
---|
923 | case LC_MONETARY: |
---|
924 | ret = __monetary_load_locale (loc, locale, (void *) l_wctomb, charset); |
---|
925 | break; |
---|
926 | case LC_NUMERIC: |
---|
927 | ret = __numeric_load_locale (loc, locale, (void *) l_wctomb, charset); |
---|
928 | break; |
---|
929 | case LC_TIME: |
---|
930 | ret = __time_load_locale (loc, locale, (void *) l_wctomb, charset); |
---|
931 | break; |
---|
932 | #endif /* __HAVE_LOCALE_INFO__ */ |
---|
933 | default: |
---|
934 | break; |
---|
935 | } |
---|
936 | #ifdef __HAVE_LOCALE_INFO__ |
---|
937 | if (ret) |
---|
938 | FAIL; |
---|
939 | #endif /* __HAVE_LOCALE_INFO__ */ |
---|
940 | return strcpy(loc->categories[category], new_locale); |
---|
941 | } |
---|
942 | |
---|
943 | const char * |
---|
944 | __get_locale_env (struct _reent *p, int category) |
---|
945 | { |
---|
946 | const char *env; |
---|
947 | |
---|
948 | /* 1. check LC_ALL. */ |
---|
949 | env = _getenv_r (p, categories[0]); |
---|
950 | |
---|
951 | /* 2. check LC_* */ |
---|
952 | if (env == NULL || !*env) |
---|
953 | env = _getenv_r (p, categories[category]); |
---|
954 | |
---|
955 | /* 3. check LANG */ |
---|
956 | if (env == NULL || !*env) |
---|
957 | env = _getenv_r (p, "LANG"); |
---|
958 | |
---|
959 | /* 4. if none is set, fall to default locale */ |
---|
960 | if (env == NULL || !*env) |
---|
961 | env = __default_locale; |
---|
962 | |
---|
963 | return env; |
---|
964 | } |
---|
965 | #endif /* _MB_CAPABLE */ |
---|
966 | |
---|
967 | int |
---|
968 | __locale_mb_cur_max (void) |
---|
969 | { |
---|
970 | #ifdef __HAVE_LOCALE_INFO__ |
---|
971 | return __get_current_ctype_locale ()->mb_cur_max[0]; |
---|
972 | #else |
---|
973 | return __get_current_locale ()->mb_cur_max[0]; |
---|
974 | #endif |
---|
975 | } |
---|
976 | |
---|
977 | const char * |
---|
978 | __locale_ctype_ptr_l (struct __locale_t *locale) |
---|
979 | { |
---|
980 | return locale->ctype_ptr; |
---|
981 | } |
---|
982 | |
---|
983 | const char * |
---|
984 | __locale_ctype_ptr (void) |
---|
985 | { |
---|
986 | return __get_current_locale ()->ctype_ptr; |
---|
987 | } |
---|
988 | |
---|
989 | #ifndef _REENT_ONLY |
---|
990 | |
---|
991 | char * |
---|
992 | setlocale (int category, |
---|
993 | const char *locale) |
---|
994 | { |
---|
995 | return _setlocale_r (_REENT, category, locale); |
---|
996 | } |
---|
997 | |
---|
998 | #endif |
---|