1 /* locale.c -- locale module.
2 Copyright (C) 2003, 2004
3 National Institute of Advanced Industrial Science and Technology (AIST)
4 Registration Number H15PRO112
6 This file is part of the m17n library.
8 The m17n library is free software; you can redistribute it and/or
9 modify it under the terms of the GNU Lesser General Public License
10 as published by the Free Software Foundation; either version 2.1 of
11 the License, or (at your option) any later version.
13 The m17n library is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 Lesser General Public License for more details.
18 You should have received a copy of the GNU Lesser General Public
19 License along with the m17n library; if not, write to the Free
20 Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
24 @addtogroup m17nLocale
25 @brief Locale objects and API for them.
27 The m17n library represents locale related information as objects
31 @addtogroup m17nLocale
32 @brief ¥í¥±¡¼¥ë¥ª¥Ö¥¸¥§¥¯¥È¤È¤½¤ì¤Ë´Ø¤¹¤ë API.
34 m17n ¥é¥¤¥Ö¥é¥ê¤Ï¥í¥±¡¼¥ë´ØÏ¢¾ðÊó¤ò #MLocale ·¿¤Î¥ª¥Ö¥¸¥§¥¯¥È¤Çɽ¸½¤¹¤ë¡£ */
38 #if !defined (FOR_DOXYGEN) || defined (DOXYGEN_INTERNAL_MODULE)
39 /*** @addtogroup m17nInternal
49 #ifdef HAVE_LANGINFO_H
55 #include "m17n-misc.h"
62 static MSymbol M_locale;
63 static MSymbol M_xfrm;
66 /** Structure of locales. */
80 /** The current locales of each category. */
81 MLocale *mlocale__collate, *mlocale__ctype;
82 MLocale *mlocale__messages, *mlocale__time;
84 /* These are currently not used. */
86 MLocale *mlocale_monetary, *mlocale_numeric, ;
89 /** Parse locale name NAME and return a newly created MLocale object.
90 If the locale is not supported by the system, return NULL. */
93 make_locale (const char *name)
95 char *current, *new, *str;
100 str = setlocale (LC_CTYPE, NULL);
101 len = strlen (str) + 1;
102 current = alloca (len);
103 memcpy (current, str, len);
105 if (! (new = setlocale (LC_CTYPE, name)))
109 M17N_OBJECT (locale, NULL, MERROR_LOCALE);
110 locale->name = msymbol (new);
111 msymbol_put (locale->name, M_locale, (void *) locale);
112 M17N_OBJECT_UNREF (locale);
114 len = strlen (new) + 1;
116 memcpy (str, new, len);
124 for (i = 0; str[i]; i++)
125 if (str[i] == '_' || str[i] == '.' || str[i] == '@')
130 /* The first field is for language. */
131 locale->language = msymbol (name);
133 /* The field following '_' is for territory. */
134 locale->territory = msymbol (name);
136 /* The field following '.' is for codeset. */
137 locale->codeset = msymbol (name);
139 /* The other field is for modifier. */
140 locale->modifier = msymbol (name);
147 #ifdef HAVE_NL_LANGINFO
149 /* If we can use nl_langinfo () to retrieve a codeset name, respect
150 it over the codeset name extracted from the locale name. */
151 locale->codeset = msymbol (nl_langinfo (CODESET));
155 /* If the locale name specifies a codeset, get the corresponding
157 if (locale->codeset != Mnil)
159 locale->coding = mconv_resolve_coding (locale->codeset);
160 if (locale->coding == Mnil)
161 locale->coding = Mcoding_us_ascii;
164 locale->coding = Mcoding_us_ascii;
166 setlocale (LC_CTYPE, current);
171 /** Decode the byte sequence at BUF of length SIZE bytes by the coding
172 system associated with LOCALE, and return a generated M-text. */
175 decode_locale (unsigned char *buf, int size, MLocale *locale)
177 return mconv_decode_buffer (locale->coding, buf, size);
181 /** Encode the M-text MT by the coding system associated with LOCALE,
182 and store the resulting bytes in the memory area at BUF of *SIZE
183 bytes. If the area is too short, allocate a new and wider area.
184 Store the length of the generated bytes in the place pointed by
185 SIZE, and return the address of those bytes. */
187 static unsigned char *
188 encode_locale (MText *mt, unsigned char *buf, int *size, MLocale *locale)
190 int nbytes = mconv_encode_buffer (locale->coding, mt, buf, *size - 1);
197 MTABLE_REALLOC (buf, *size, MERROR_LOCALE);
198 nbytes = mconv_encode_buffer (mlocale__ctype->coding, mt, buf,
200 } while (nbytes < 0);
208 /** Structure of transformed strings. The function mtext_coll ()
209 caches this object in an M-text as a text property. */
212 /* Common header for a managed object. */
215 /* Locale corresponding to <str>. */
218 /** Result of strxfrm. */
224 free_xfrm (void *object)
226 MXfrm *xfrm = (MXfrm *) object;
228 M17N_OBJECT_UNREF (xfrm->locale);
235 MTextProperty *prop = mtext_get_property (mt, 0, M_xfrm);
238 unsigned char *buf, *newbuf;
243 if (prop->end == mt->nchars)
245 xfrm = (MXfrm *) prop->val;
246 if (xfrm->locale == mlocale__ctype)
249 mtext_detach_property (prop);
254 newbuf = encode_locale (mt, buf, &size, mlocale__ctype);
255 M17N_OBJECT (xfrm, free_xfrm, MERROR_MTEXT);
256 xfrm->str = malloc (size);
257 request = strxfrm (xfrm->str, (char *) newbuf, size);
260 xfrm->str = realloc (xfrm->str, request);
261 strxfrm (xfrm->str, (char *) newbuf, size);
265 prop = mtext_property (M_xfrm, xfrm, MTEXTPROP_VOLATILE_WEAK);
266 mtext_attach_property (mt, 0, mt->nchars, prop);
267 M17N_OBJECT_UNREF (prop);
277 M_locale = msymbol_as_managing_key (" locale");
279 Mterritory = msymbol ("territory");
280 Mcodeset = msymbol ("codeset");
282 mlocale__collate = mlocale_set (LC_COLLATE, NULL);
283 M17N_OBJECT_REF (mlocale__collate);
284 mlocale__ctype = mlocale_set (LC_CTYPE, NULL);
285 M17N_OBJECT_REF (mlocale__ctype);
286 mlocale__messages = mlocale_set (LC_MESSAGES, NULL);
287 M17N_OBJECT_REF (mlocale__messages);
288 mlocale__time = mlocale_set (LC_TIME, NULL);
289 M17N_OBJECT_REF (mlocale__time);
291 M_xfrm = msymbol_as_managing_key (" xfrm");
298 M17N_OBJECT_UNREF (mlocale__collate);
299 M17N_OBJECT_UNREF (mlocale__ctype);
300 M17N_OBJECT_UNREF (mlocale__messages);
301 M17N_OBJECT_UNREF (mlocale__time);
305 #endif /* !FOR_DOXYGEN || DOXYGEN_INTERNAL_MODULE */
309 /*** @addtogroup m17nLocale */
313 /***en The symbol whose name is "territory". */
314 /***ja "territory" ¤È¤¤¤¦Ì¾Á°¤ò»ý¤Ä¥·¥ó¥Ü¥ë. */
318 /***en The symbol whose name is "modifier". */
319 /***ja "modifier" ¤È¤¤¤¦Ì¾Á°¤ò»ý¤Ä¥·¥ó¥Ü¥ë. */
323 /***en The symbol whose name is "codeset". */
324 /***ja "codeset" ¤È¤¤¤¦Ì¾Á°¤ò»ý¤Ä¥·¥ó¥Ü¥ë. */
330 @brief Set the current locale.
332 The mlocale_set () function sets or query a part of the current
333 locale. The part is specified by $CATEGORY which must be a valid
334 first argument to <tt>setlocale ()</tt>.
336 If $LOCALE is not NULL, the locale of the specified part is set to
337 $LOCALE. If $LOCALE is not supported by the system, the current
338 locale is not changed.
340 If $LOCALE is NULL, the current locale of the specified part is
344 If the call is successful, mlocale_set () returns an opaque locale
345 object that corresponds to the locale. The name of the locale can
346 be acquired by the function mlocale_get_prop ().
348 Otherwise, it returns NULL. */
351 @brief ¸½ºß¤Î¥í¥±¡¼¥ë¤òÀßÄꤹ¤ë.
353 ´Ø¿ô mlocale_set () ¤Ï¸½ºß¤Î¥í¥±¡¼¥ë¤Î°ìÉô¤òÀßÄꤷ¤¿¤êÌ䤤¹ç¤ï¤»¤¿¤ê¤¹¤ë¡£¤³¤³¤Ç°ìÉô¤È¤Ï
354 $CATEGORY ¤Ç»ØÄꤵ¤ì¡¢<tt>setlocale ()</tt> ¤Î͸ú¤ÊÂè°ì°ú¿ô¤È¤Ê¤ë¤â¤Î¤Ç¤Ê¤¯¤Æ¤Ï¤Ê¤é¤Ê¤¤¡£
356 $LOCALE ¤¬ NULL ¤Ç¤Ê¤±¤ì¤Ð¡¢»ØÄꤷ¤¿Éôʬ¤Î¥í¥±¡¼¥ë¤¬$LOCALE ¤ËÀßÄꤵ¤ì¤ë¡£
357 $LOCALE ¤¬¥·¥¹¥Æ¥à¤Ë¥µ¥Ý¡¼¥È¤µ¤ì¤Æ¤¤¤Ê¤±¤ì¤Ð¡¢ÀßÄê¤Ï¹Ô¤ï¤ì¤º¡¢¸½ºß¤Î¥í¥±¡¼¥ë¤ÏÊѤï¤é¤Ê¤¤¡£
359 $LOCALE ¤¬ NULL ¤Ê¤é¤Ð¡¢¸½ºß¤Î¥í¥±¡¼¥ë¤Î»ØÄꤷ¤¿Éôʬ¤òÌ䤤¹ç¤ï¤»¤ë¡£
363 ¸Æ¤Ó½Ð¤·¤ËÀ®¸ù¤¹¤ì¤Ð¡¢mlocale_set () ¤Ï¥í¥±¡¼¥ë¤ËÂбþ¤¹¤ë opaque
364 ¥í¥±¡¼¥ë¥ª¥Ö¥¸¥§¥¯¥È¤òÊÖ¤¹¡£¥í¥±¡¼¥ë¤Î̾Á°¤Ï´Ø¿ô
365 mlocale_get_prop () ¤Ë¤è¤Ã¤ÆÆÀ¤ë¤³¤È¤¬¤Ç¤¤ë¡£
367 ¤½¤¦¤Ç¤Ê¤±¤ì¤Ð NULL ¤òÊÖ¤¹¡£
375 mlocale_set (int category, const char *name)
380 new = setlocale (category, name);
384 locale = (MLocale *) msymbol_get (msymbol (new), M_locale);
386 locale = make_locale (new);
389 if (name && (category == LC_ALL || category == LC_COLLATE))
391 M17N_OBJECT_REF (locale);
392 M17N_OBJECT_UNREF (mlocale__collate);
393 mlocale__collate = locale;
395 else if (name && (category == LC_ALL || category == LC_CTYPE))
397 M17N_OBJECT_REF (locale);
398 M17N_OBJECT_UNREF (mlocale__ctype);
399 mlocale__ctype = locale;
401 if (name && (category == LC_ALL || category == LC_MESSAGES))
403 M17N_OBJECT_REF (locale);
404 M17N_OBJECT_UNREF (mlocale__messages);
405 mlocale__messages = locale;
407 if (name && (category == LC_ALL || category == LC_TIME))
409 M17N_OBJECT_REF (locale);
410 M17N_OBJECT_UNREF (mlocale__time);
411 mlocale__time = locale;
419 @brief Get the value of a locale property.
421 The mlocale_get_prop () function returns the value of a property
422 $KEY of local $LOCALE. $KEY must be #Mname, #Mlanguage,
423 #Mterritory, #Mcodeset, #Mmodifier, or #Mcoding. */
426 @brief ¥í¥±¡¼¥ë¥×¥í¥Ñ¥Æ¥£¤ÎÃͤòÆÀ¤ë.
428 ´Ø¿ô mlocale_get_prop () ¤Ï¡¢¥í¥±¡¼¥ë $LOCALE ¤Î $KEY ¥×¥í¥Ñ¥Æ¥£¤ÎÃͤòÊÖ¤¹¡£
429 $KEY ¤Ï #Mname, #Mlanguage, #Mterritory, #Mcodeset, #Mmodifier,
430 #Mcoding ¤Î¤¤¤º¤ì¤«¤Ç¤¢¤ë¡£ */
433 mlocale_get_prop (MLocale *locale, MSymbol key)
436 return locale->coding;
439 if (key == Mlanguage)
440 return locale->language;
441 if (key == Mterritory)
442 return locale->territory;
444 return locale->codeset;
445 if (key == Mmodifier)
446 return locale->modifier;
452 @brief Format date and time.
454 The mtext_ftime () function formats the broken-down time $TM
455 according to the format specification $FORMAT and append the
456 result to the M-text $MT. The formating is done according to the
457 locale $LOCALE (if not NULL) or the current locale (LC_TIME).
459 The meaning of the arguments $TM and $FORMAT are the same as those
466 @brief ÆüÉդȻþ´Ö¤ò¥Õ¥©¡¼¥Þ¥Ã¥È¤¹¤ë.
468 ´Ø¿ô mtext_ftime () ¤Ï»þ¹ï¥Ç¡¼¥¿ (broken-down time) $TM ¤ò$FORMAT
469 ¤Ç»ØÄꤵ¤ì¤¿·Á¼°¤ËÀ¶½ñ¤·¡¢·ë²Ì¤òM-text $MT ¤ËÉղ乤롣¥Õ¥©¡¼¥Þ¥Ã¥È¤Ï
470 NULL ¤Ç¤Ê¤±¤ì¤Ð ¥í¥±¡¼¥ë $LOCALE ¤Ë¡¢¤Þ¤¿¤Ï¸½ºß¤Î¥í¥±¡¼¥ë(LC_TIME) ¤Ë½¾¤¦¡£
472 °ú¿ô $TM ¤È $FORMAT ¤Î°ÕÌ£¤Ï strftime () ¤Î¾ì¹ç¤ÈƱ¤¸¡£
479 mtext_ftime (MText *mt, const char *format, const struct tm *tm,
484 size_t nbytes, nchars;
485 char *current_locale = NULL;
489 char *str = setlocale (LC_TIME, NULL);
490 int len = strlen (str) + 1;
492 current_locale = alloca (len);
493 memcpy (current_locale, str, len);
494 mlocale_set (LC_TIME, msymbol_name (locale->name));
500 MTABLE_ALLOCA (buf, bufsize, MERROR_MTEXT);
502 nbytes = strftime ((char *) buf, bufsize, format, tm);
511 MText *work = decode_locale (buf, nbytes, mlocale__time);
515 nchars = work->nchars;
516 mtext_cat (mt, work);
517 M17N_OBJECT_UNREF (work);
526 mlocale_set (LC_TIME, current_locale);
534 @brief Get an environment variable.
536 The mtext_getenv () function searches the environment variable
537 list for a string that matches the string pointed to by $NAME.
539 If there is a match, the function decodes the value according to
540 the current locale (LC_CTYPE) into an M-text, and return that
543 If there is no match, the function returns NULL. */
545 @brief ´Ä¶ÊÑ¿ô¤òÆÀ¤ë.
547 ´Ø¿ô mtext_getenv () ¤Ï $NAME
548 ¤Ç»Ø¤µ¤ì¤ëʸ»úÎó¤È¹çÃפ¹¤ëʸ»úÎó¤ò´Ä¶ÊÑ¿ô¤Î¥ê¥¹¥ÈÃ椫¤éõ¤¹¡£
550 ¸«¤Ä¤«¤Ã¤¿¾ì¹ç¤Ë¤Ï¡¢¤½¤ÎÃͤò¸½ºß¤Î¥í¥±¡¼¥ë(LC_CTYPE) ¤Ë½¾¤Ã¤Æ
551 M-text ¤Ë¥Ç¥³¡¼¥É¤·¡¢¤½¤ÎM-text ¤òÊÖ¤¹¡£
553 ¸«¤Ä¤«¤é¤Ê¤±¤ì¤Ð¡¢NULL ¤òÊÖ¤¹¡£ */
556 mtext_getenv (const char *name)
558 char *p = getenv (name);
562 return decode_locale ((unsigned char *) p, strlen (p), mlocale__ctype);
568 @brief Change or add an environment variable.
570 The mtext_putenv () function changes or adds the value of
571 environment variables according to M-text $MT. It calls the
572 function <tt>putenv</tt> with an argument generated by encoding
573 $MT according to the current locale (LC_CTYPE).
576 This function returns zero on success, or -1 if an error
579 @brief ´Ä¶ÊÑ¿ô¤òÊѹ¹¡¿Äɲ乤ë.
581 ´Ø¿ô mtext_putenv () ¤Ï M-text $MT
582 ¤Ë½¾¤Ã¤Æ¡¢´Ä¶ÊÑ¿ô¤ÎÃͤòÊѹ¹¤·¤¿¤êÄɲä·¤¿¤ê¤¹¤ë¡£¤³¤Î´Ø¿ô¤Ï¡¢¸½ºß¤Î¥í¥±¡¼¥ë
583 (LC_CTYPE) ¤Ë½¾¤Ã¤Æ$MT ¤ò¥¨¥ó¥³¡¼¥É¤·¡¢¤½¤ì¤ò°ú¿ô¤È¤·¤Æ´Ø¿ô <tt>putenv</tt> ¤ò¸Æ¤Ö¡£
586 ¤³¤Î´Ø¿ô¤Ï¡¢À®¸ù¤·¤¿¾ì¹ç¤Ë¤Ï 0 ¤ò¡¢¥¨¥é¡¼¤¬µ¯¤³¤ì¤Ð -1 ¤òÊÖ¤¹¡£
591 mtext_putenv (MText *mt)
593 unsigned char buf[1024];
595 unsigned char *newbuf;
598 newbuf = encode_locale (mt, buf, &size, mlocale__ctype);
599 result = putenv ((char *) newbuf);
608 @brief Compare two M-texts using the current locale.
610 The mtext_coll () function compares the two M-texts $MT1 and $MT2.
611 It returns an integer less than, equal to, or greater than zero if
612 $MT1 is found, respectively, to be less than, to match, or to be
613 greater than $MT2. The comparison is based on texts as
614 appropriate for the current locale (LC_COLLATE).
616 This function makes use of information that is automatically
617 cached in the M-texts as a text property. So, the second call of
618 this function with $MT1 or $MT2 finishes faster than the first
621 @brief ¸½ºß¤Î¥í¥±¡¼¥ë¤òÍѤ¤¤Æ£²¤Ä¤Î M-text ¤òÈæ³Ó¤¹¤ë.
623 ´Ø¿ô mtext_coll () ¤Ï£²¤Ä¤Î M-text $MT1 ¤È $MT2
624 ¤òÈæ³Ó¤¹¤ë¡£Ìá¤êÃͤÏÉé¤ÎÀ°¿ôÃÍ, 0, Àµ¤ÎÀ°¿ôÃͤΤ¤¤º¤ì¤«¤Ç¤¢¤ê¡¢¤½¤ì¤¾¤ì
625 $MT1 ¤¬ $MT2 ¤è¤ê¾®¤µ¤¤¡¢Æ±¤¸¡¢Â礤¤¾ì¹ç¤ËÁêÅö¤¹¤ë¡£Èæ³Ó¤Ï¸½ºß¤Î¥í¥±¡¼¥ë
626 (LC_COLLATE) ¤Ë´ð¤Å¤¤¤Æ¹Ô¤ï¤ì¤ë¡£
629 ¤Î¥Æ¥¥¹¥È¥×¥í¥Ñ¥Æ¥£¤È¤·¤Æ¼«Æ°Åª¤Ë¥¥ã¥Ã¥·¥å¤µ¤ì¤ë¾ðÊó¤òÍøÍѤ¹¤ë¤Î¤Ç¡¢£²ÅÙÌܰʹߤÎƱ¤¸Èæ³Ó¤Ï£±ÅÙÌܤè¤ê®¤¯¼Â¹Ô¤µ¤ì¤ë¡£ */
632 mtext_coll (MText *mt1, MText *mt2)
636 if (mt1->nchars == 0)
637 return (mt2->nchars == 0 ? 0 : -1);
638 else if (mt2->nchars == 0)
641 str1 = get_xfrm (mt1);
642 str2 = get_xfrm (mt2);
643 return strcoll (str1, str2);