X-Git-Url: http://erislabs.net/gitweb/?a=blobdiff_plain;ds=sidebyside;f=lib%2Flocalename.c;h=3ab88228a266426b2e16099669c22e0323da7556;hb=5bddd6312f3041adf1580f0c791dcb05c0d3ae92;hp=1310cb5d91b0757cbc978dd9884b8b6ddb434fa7;hpb=387283921ccb4f962e9e8eba2ef1d5e84a678b35;p=gnulib.git diff --git a/lib/localename.c b/lib/localename.c index 1310cb5d9..3ab88228a 100644 --- a/lib/localename.c +++ b/lib/localename.c @@ -1,5 +1,5 @@ /* Determine name of the currently selected locale. - Copyright (C) 1995-1999, 2000-2009 Free Software Foundation, Inc. + Copyright (C) 1995-2011 Free Software Foundation, Inc. This program is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as published @@ -29,11 +29,24 @@ # include "localename.h" #endif +#include +#include #include #include +#include + +#if HAVE_USELOCALE +/* MacOS X 10.5 defines the locale_t type in . */ +# if defined __APPLE__ && defined __MACH__ +# include +# endif +# include +# if !defined IN_LIBINTL +# include "glthread/lock.h" +# endif +#endif #if HAVE_CFLOCALECOPYCURRENT || HAVE_CFPREFERENCESCOPYAPPVALUE -# include # include # if HAVE_CFLOCALECOPYCURRENT # include @@ -46,7 +59,7 @@ # define WIN32_NATIVE #endif -#ifdef WIN32_NATIVE +#if defined WIN32_NATIVE || defined __CYGWIN__ /* WIN32 or Cygwin */ # define WIN32_LEAN_AND_MEAN # include /* List of language codes, sorted by value: @@ -1394,7 +1407,7 @@ gl_locale_name_canonicalize (char *name) #endif -#ifdef WIN32_NATIVE +#if defined WIN32_NATIVE || defined __CYGWIN__ /* WIN32 or Cygwin */ /* Canonicalize a Win32 native locale name to a Unix locale name. NAME is a sufficiently large buffer. @@ -2494,22 +2507,274 @@ gl_locale_name_from_win32_LCID (LCID lcid) #endif +#if HAVE_USELOCALE /* glibc or MacOS X */ + +/* Simple hash set of strings. We don't want to drag in lots of hash table + code here. */ + +# define SIZE_BITS (sizeof (size_t) * CHAR_BIT) + +/* A hash function for NUL-terminated char* strings using + the method described by Bruno Haible. + See http://www.haible.de/bruno/hashfunc.html. */ +static size_t +string_hash (const void *x) +{ + const char *s = (const char *) x; + size_t h = 0; + + for (; *s; s++) + h = *s + ((h << 9) | (h >> (SIZE_BITS - 9))); + + return h; +} + +/* A hash table of fixed size. Multiple threads can access it read-only + simultaneously, but only one thread can insert into it at the same time. */ + +/* A node in a hash bucket collision list. */ +struct hash_node + { + struct hash_node * volatile next; + char contents[100]; /* has variable size */ + }; + +# define HASH_TABLE_SIZE 257 +static struct hash_node * volatile struniq_hash_table[HASH_TABLE_SIZE] + /* = { NULL, ..., NULL } */; + +/* This lock protects the struniq_hash_table against multiple simultaneous + insertions. */ +gl_lock_define_initialized(static, struniq_lock) + +/* Store a copy of the given string in a string pool with indefinite extent. + Return a pointer to this copy. */ +static const char * +struniq (const char *string) +{ + size_t hashcode = string_hash (string); + size_t slot = hashcode % HASH_TABLE_SIZE; + size_t size; + struct hash_node *new_node; + struct hash_node *p; + for (p = struniq_hash_table[slot]; p != NULL; p = p->next) + if (strcmp (p->contents, string) == 0) + return p->contents; + size = strlen (string) + 1; + new_node = + (struct hash_node *) + malloc (offsetof (struct hash_node, contents[0]) + size); + if (new_node == NULL) + /* Out of memory. Return a statically allocated string. */ + return "C"; + memcpy (new_node->contents, string, size); + /* Lock while inserting new_node. */ + gl_lock_lock (struniq_lock); + /* Check whether another thread already added the string while we were + waiting on the lock. */ + for (p = struniq_hash_table[slot]; p != NULL; p = p->next) + if (strcmp (p->contents, string) == 0) + { + free (new_node); + new_node = p; + goto done; + } + /* Really insert new_node into the hash table. Fill new_node entirely first, + because other threads may be iterating over the linked list. */ + new_node->next = struniq_hash_table[slot]; + struniq_hash_table[slot] = new_node; + done: + /* Unlock after new_node is inserted. */ + gl_lock_unlock (struniq_lock); + return new_node->contents; +} + +#endif + + +#if defined IN_LIBINTL || HAVE_USELOCALE + +/* Like gl_locale_name_thread, except that the result is not in storage of + indefinite extent. */ +# if !defined IN_LIBINTL +static +# endif +const char * +gl_locale_name_thread_unsafe (int category, const char *categoryname) +{ +# if HAVE_USELOCALE + { + locale_t thread_locale = uselocale (NULL); + if (thread_locale != LC_GLOBAL_LOCALE) + { +# if __GLIBC__ >= 2 && !defined __UCLIBC__ + /* Work around an incorrect definition of the _NL_LOCALE_NAME macro in + glibc < 2.12. + See . */ + const char *name = + nl_langinfo (_NL_ITEM ((category), _NL_ITEM_INDEX (-1))); + if (name[0] == '\0') + /* Fallback code for glibc < 2.4, which did not implement + nl_langinfo (_NL_LOCALE_NAME (category)). */ + name = thread_locale->__names[category]; + return name; +# endif +# if defined __APPLE__ && defined __MACH__ /* MacOS X */ + /* The locale name is found deep in an undocumented data structure. + Since it's stored in a buffer of size 32 and newlocale() rejects + locale names of length > 31, we can assume that it is NUL terminated + in this buffer. But we need to make a copy of the locale name, of + indefinite extent. */ + struct _xlocale_part1_v0 /* used in MacOS X 10.5 */ + { + int32_t __refcount; + void (*__free_extra)(void *); + __darwin_mbstate_t __mbs[10]; + int64_t __magic; + }; + struct _xlocale_part1_v1 /* used in MacOS X >= 10.6.0 */ + { + int32_t __refcount; + void (*__free_extra)(void *); + __darwin_mbstate_t __mbs[10]; + /*pthread_lock_t*/ int __lock; + int64_t __magic; + }; + struct _xlocale_part2 + { + int64_t __magic; + unsigned char __collate_load_error; + unsigned char __collate_substitute_nontrivial; + unsigned char _messages_using_locale; + unsigned char _monetary_using_locale; + unsigned char _numeric_using_locale; + unsigned char _time_using_locale; + unsigned char __mlocale_changed; + unsigned char __nlocale_changed; + unsigned char __numeric_fp_cvt; + struct __xlocale_st_collate *__lc_collate; + struct __xlocale_st_runelocale *__lc_ctype; + struct __xlocale_st_messages *__lc_messages; + struct __xlocale_st_monetary *__lc_monetary; + struct __xlocale_st_numeric *__lc_numeric; + struct _xlocale *__lc_numeric_loc; + struct __xlocale_st_time *__lc_time; + /* more */ + }; + struct __xlocale_st_collate + { + int32_t __refcount; + void (*__free_extra)(void *); + char __encoding[32]; + /* more */ + }; + struct __xlocale_st_runelocale + { + int32_t __refcount; + void (*__free_extra)(void *); + char __ctype_encoding[32]; + /* more */ + }; + struct __xlocale_st_messages + { + int32_t __refcount; + void (*__free_extra)(void *); + char *_messages_locale_buf; + /* more */ + }; + struct __xlocale_st_monetary + { + int32_t __refcount; + void (*__free_extra)(void *); + char *_monetary_locale_buf; + /* more */ + }; + struct __xlocale_st_numeric { + int32_t __refcount; + void (*__free_extra)(void *); + char *_numeric_locale_buf; + /* more */ + }; + struct __xlocale_st_time { + int32_t __refcount; + void (*__free_extra)(void *); + char *_time_locale_buf; + /* more */ + }; + struct _xlocale_part2 *tlp; + if (((struct _xlocale_part1_v0 *) thread_locale)->__magic + == 0x786C6F63616C6530LL) + /* MacOS X 10.5 */ + tlp = + (struct _xlocale_part2 *) + &((struct _xlocale_part1_v0 *) thread_locale)->__magic; + else if (((struct _xlocale_part1_v1 *) thread_locale)->__magic + == 0x786C6F63616C6530LL) + /* MacOS X >= 10.6.0 */ + tlp = + (struct _xlocale_part2 *) + &((struct _xlocale_part1_v1 *) thread_locale)->__magic; + else + /* Unsupported version of MacOS X: The internals of 'struct _xlocale' + have changed again. */ + return ""; + switch (category) + { + case LC_CTYPE: + return tlp->__lc_ctype->__ctype_encoding; + case LC_NUMERIC: + return tlp->_numeric_using_locale + ? tlp->__lc_numeric->_numeric_locale_buf + : "C"; + case LC_TIME: + return tlp->_time_using_locale + ? tlp->__lc_time->_time_locale_buf + : "C"; + case LC_COLLATE: + return !tlp->__collate_load_error + ? tlp->__lc_collate->__encoding + : "C"; + case LC_MONETARY: + return tlp->_monetary_using_locale + ? tlp->__lc_monetary->_monetary_locale_buf + : "C"; + case LC_MESSAGES: + return tlp->_messages_using_locale + ? tlp->__lc_messages->_messages_locale_buf + : "C"; + default: /* We shouldn't get here. */ + return ""; + } +# endif + } + } +# endif + return NULL; +} + +#endif + +const char * +gl_locale_name_thread (int category, const char *categoryname) +{ +#if HAVE_USELOCALE + const char *name = gl_locale_name_thread_unsafe (category, categoryname); + if (name != NULL) + return struniq (name); +#endif + return NULL; +} + /* XPG3 defines the result of 'setlocale (category, NULL)' as: "Directs 'setlocale()' to query 'category' and return the current setting of 'local'." However it does not specify the exact format. Neither do SUSV2 and ISO C 99. So we can use this feature only on selected systems (e.g. those using GNU C Library). */ -#if defined _LIBC || (defined __GLIBC__ && __GLIBC__ >= 2) +#if defined _LIBC || ((defined __GLIBC__ && __GLIBC__ >= 2) && !defined __UCLIBC__) # define HAVE_LOCALE_NULL #endif -/* Determine the current locale's name, and canonicalize it into XPG syntax - language[_territory][.codeset][@modifier] - The codeset part in the result is not reliable; the locale_charset() - should be used for codeset information instead. - The result must not be freed; it is statically allocated. */ - const char * gl_locale_name_posix (int category, const char *categoryname) { @@ -2518,6 +2783,30 @@ gl_locale_name_posix (int category, const char *categoryname) #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL return setlocale (category, NULL); #else + /* On other systems we ignore what setlocale reports and instead look at the + environment variables directly. This is necessary + 1. on systems which have a facility for customizing the default locale + (MacOS X, native Windows, Cygwin) and where the system's setlocale() + function ignores this default locale (MacOS X, Cygwin), in two cases: + a. when the user missed to use the setlocale() override from libintl + (for example by not including ), + b. when setlocale supports only the "C" locale, such as on Cygwin + 1.5.x. In this case even the override from libintl cannot help. + 2. on all systems where setlocale supports only the "C" locale. */ + /* Strictly speaking, it is a POSIX violation to look at the environment + variables regardless whether setlocale has been called or not. POSIX + says: + "For C-language programs, the POSIX locale shall be the + default locale when the setlocale() function is not called." + But we assume that all programs that use internationalized APIs call + setlocale (LC_ALL, ""). */ + return gl_locale_name_environ (category, categoryname); +#endif +} + +const char * +gl_locale_name_environ (int category, const char *categoryname) +{ const char *retval; /* Setting of LC_ALL overrides all other. */ @@ -2531,10 +2820,21 @@ gl_locale_name_posix (int category, const char *categoryname) /* Last possibility is the LANG environment variable. */ retval = getenv ("LANG"); if (retval != NULL && retval[0] != '\0') - return retval; + { +#if HAVE_CFLOCALECOPYCURRENT || HAVE_CFPREFERENCESCOPYAPPVALUE + /* MacOS X 10.2 or newer. + Ignore invalid LANG value set by the Terminal application. */ + if (strcmp (retval, "UTF-8") != 0) +#endif +#if defined __CYGWIN__ + /* Cygwin. + Ignore dummy LANG value set by ~/.profile. */ + if (strcmp (retval, "C.UTF-8") != 0) +#endif + return retval; + } return NULL; -#endif } const char * @@ -2547,9 +2847,28 @@ gl_locale_name_default (void) implementation-defined locale. Some implementations may provide facilities for local installation administrators to set the default locale, customizing it for each location. POSIX:2001 does not require - such a facility. */ + such a facility. -#if !(HAVE_CFLOCALECOPYCURRENT || HAVE_CFPREFERENCESCOPYAPPVALUE || defined(WIN32_NATIVE)) + The systems with such a facility are MacOS X and Windows: They provide a + GUI that allows the user to choose a locale. + - On MacOS X, by default, none of LC_* or LANG are set. Starting with + MacOS X 10.4 or 10.5, LANG is set for processes launched by the + 'Terminal' application (but sometimes to an incorrect value "UTF-8"). + When no environment variable is set, setlocale (LC_ALL, "") uses the + "C" locale. + - On native Windows, by default, none of LC_* or LANG are set. + When no environment variable is set, setlocale (LC_ALL, "") uses the + locale chosen by the user. + - On Cygwin 1.5.x, by default, none of LC_* or LANG are set. + When no environment variable is set, setlocale (LC_ALL, "") uses the + "C" locale. + - On Cygwin 1.7, by default, LANG is set to "C.UTF-8" when the default + ~/.profile is executed. + When no environment variable is set, setlocale (LC_ALL, "") uses the + "C.UTF-8" locale, which operates in the same way as the "C" locale. + */ + +#if !(HAVE_CFLOCALECOPYCURRENT || HAVE_CFPREFERENCESCOPYAPPVALUE || defined WIN32_NATIVE || defined __CYGWIN__) /* The system does not have a way of setting the locale, other than the POSIX specified environment variables. We use C as default locale. */ @@ -2575,7 +2894,7 @@ gl_locale_name_default (void) CFLocaleRef locale = CFLocaleCopyCurrent (); CFStringRef name = CFLocaleGetIdentifier (locale); - if (CFStringGetCString (name, namebuf, sizeof(namebuf), + if (CFStringGetCString (name, namebuf, sizeof (namebuf), kCFStringEncodingASCII)) { gl_locale_name_canonicalize (namebuf); @@ -2588,7 +2907,8 @@ gl_locale_name_default (void) kCFPreferencesCurrentApplication); if (value != NULL && CFGetTypeID (value) == CFStringGetTypeID () - && CFStringGetCString ((CFStringRef)value, namebuf, sizeof(namebuf), + && CFStringGetCString ((CFStringRef)value, + namebuf, sizeof (namebuf), kCFStringEncodingASCII)) { gl_locale_name_canonicalize (namebuf); @@ -2603,7 +2923,7 @@ gl_locale_name_default (void) # endif -# if defined(WIN32_NATIVE) /* WIN32, not Cygwin */ +# if defined WIN32_NATIVE || defined __CYGWIN__ /* WIN32 or Cygwin */ { LCID lcid; @@ -2616,11 +2936,21 @@ gl_locale_name_default (void) #endif } +/* Determine the current locale's name, and canonicalize it into XPG syntax + language[_territory][.codeset][@modifier] + The codeset part in the result is not reliable; the locale_charset() + should be used for codeset information instead. + The result must not be freed; it is statically allocated. */ + const char * gl_locale_name (int category, const char *categoryname) { const char *retval; + retval = gl_locale_name_thread (category, categoryname); + if (retval != NULL) + return retval; + retval = gl_locale_name_posix (category, categoryname); if (retval != NULL) return retval;