locale.c

ucLinux is a very good embeded sytem. Most of company use this as their development OS.

	'\x00',    'F',    'e',    'b', '\x00',    'M',    'a',    'r', 
	'\x00',    'A',    'p',    'r', '\x00',    'M',    'a',    'y', 
	'\x00',    'J',    'u',    'n', '\x00',    'J',    'u',    'l', 
	'\x00',    'A',    'u',    'g', '\x00',    'S',    'e',    'p', 
	'\x00',    'O',    'c',    't', '\x00',    'N',    'o',    'v', 
	'\x00',    'D',    'e',    'c', '\x00',    'J',    'a',    'n', 
	   'u',    'a',    'r',    'y', '\x00',    'F',    'e',    'b', 
	   'r',    'u',    'a',    'r',    'y', '\x00',    'M',    'a', 
	   'r',    'c',    'h', '\x00',    'A',    'p',    'r',    'i', 
	   'l', '\x00',    'M',    'a',    'y', '\x00',    'J',    'u', 
	   'n',    'e', '\x00',    'J',    'u',    'l',    'y', '\x00', 
	   'A',    'u',    'g',    'u',    's',    't', '\x00',    'S', 
	   'e',    'p',    't',    'e',    'm',    'b',    'e',    'r', 
	'\x00',    'O',    'c',    't',    'o',    'b',    'e',    'r', 
	'\x00',    'N',    'o',    'v',    'e',    'm',    'b',    'e', 
	   'r', '\x00',    'D',    'e',    'c',    'e',    'm',    'b', 
	   'e',    'r', '\x00',    'A',    'M', '\x00',    'P',    'M', 
	'\x00',    '%',    'a',    ' ',    '%',    'b',    ' ',    '%', 
	   'e',    ' ',    '%',    'H',    ':',    '%',    'M',    ':', 
	   '%',    'S',    ' ',    '%',    'Y', '\x00',    '%',    'm', 
	   '/',    '%',    'd',    '/',    '%',    'y', '\x00',    '%', 
	   'H',    ':',    '%',    'M',    ':',    '%',    'S', '\x00', 
	   '%',    'I',    ':',    '%',    'M',    ':',    '%',    'S', 
	   ' ',    '%',    'p', '\x00',    '^',    '[',    'y',    'Y', 
	   ']', '\x00',    '^',    '[',    'n',    'N',    ']', '\x00', 
};

char *nl_langinfo(nl_item item)
{
	unsigned int c;
	unsigned int i;

	if ((c = _NL_ITEM_CATEGORY(item)) < C_LC_ALL) {
		if ((i = cat_start[c] + _NL_ITEM_INDEX(item)) < cat_start[c+1]) {
/*  			return (char *) C_locale_data + item_offset[i] + (i & 64); */
			return (char *) C_locale_data + nl_data[C_LC_ALL+1+i] + 2*(i & 64);
		}
	}
	return (char *) cat_start;	/* Conveniently, this is the empty string. */
}

#else  /* __LOCALE_C_ONLY */

#if defined(__UCLIBC_HAS_XLOCALE__) && !defined(__UCLIBC_DO_XLOCALE)

char *nl_langinfo(nl_item item)
{
	return __nl_langinfo_l(item, __UCLIBC_CURLOCALE);
}

#else  /* defined(__UCLIBC_HAS_XLOCALE__) && !defined(__UCLIBC_DO_XLOCALE) */

static const char empty[] = "";

char *__XL(nl_langinfo)(nl_item item   __LOCALE_PARAM )
{
	unsigned int c = _NL_ITEM_CATEGORY(item);
	unsigned int i = _NL_ITEM_INDEX(item);

	if ((c < LC_ALL) && (i < __LOCALE_PTR->category_item_count[c])) {
		return ((char **)(((char *) __LOCALE_PTR)
						  + __LOCALE_PTR->category_offsets[c]))[i];
	}

	return (char *) empty;
}

#endif /* defined(__UCLIBC_HAS_XLOCALE__) && !defined(__UCLIBC_DO_XLOCALE) */

#endif /* __LOCALE_C_ONLY */

#endif
/**********************************************************************/
#ifdef L_newlocale

#ifdef __UCLIBC_MJN3_ONLY__
#warning TODO: Move posix and utf8 strings.
#endif
static const char posix[] = "POSIX";
static const char utf8[] = "UTF-8";

static int find_locale(int category_mask, const char *p,
					   unsigned char *new_locale)
{
	int i;
	const unsigned char *s;
	uint16_t n;
	unsigned char lang_cult, codeset;

#if defined(__LOCALE_DATA_AT_MODIFIERS_LENGTH) && 1
	/* Support standard locale handling for @-modifiers. */

#ifdef __UCLIBC_MJN3_ONLY__
#warning REMINDER: Fix buf size in find_locale.
#endif
	char buf[18];	/* TODO: 7+{max codeset name length} */
	const char *q;

	if ((q = strchr(p,'@')) != NULL) {
		if ((((size_t)((q-p)-5)) > (sizeof(buf) - 5)) || (p[2] != '_')) {
			return 0;
		}
		/* locale name at least 5 chars long and 3rd char is '_' */
		s = LOCALE_AT_MODIFIERS;
		do {
			if (!strcmp(s+2, q+1)) {
				break;
			}
			s += 2 + *s;		/* TODO - fix this throughout */
		} while (*s);
		if (!*s) {
			return 0;
		}
		assert(q - p < sizeof(buf));
		memcpy(buf, p, q-p);
		buf[q-p] = 0;
		buf[2] = s[1];
		p = buf;
	}
#endif

	lang_cult = codeset = 0;	/* Assume C and default codeset.  */
	if (((*p == 'C') && !p[1]) || !strcmp(p, posix)) {
		goto FIND_LOCALE;
	}

	if ((strlen(p) > 5) && (p[5] == '.')) {	/* Codeset in locale name? */
		/* TODO: maybe CODESET_LIST + *s ??? */
		/* 7bit is 1, UTF-8 is 2, 8-bit is >= 3 */
		codeset = 2;
		if (strcmp(utf8,p+6) != 0) {/* TODO - fix! */
			s = CODESET_LIST;
			do {
				++codeset;		/* Increment codeset first. */
				if (!strcmp(CODESET_LIST+*s, p+6)) {
					goto FIND_LANG_CULT;
				}
			} while (*++s);
			return 0;			/* No matching codeset! */
		}
	}

 FIND_LANG_CULT:				/* Find language_culture number. */
	s = LOCALE_NAMES;
	do {						/* TODO -- do a binary search? */
		/* TODO -- fix gen_mmap!*/
		++lang_cult;			/* Increment first since C/POSIX is 0. */
		if (!strncmp(s,p,5)) { /* Found a matching locale name; */
			goto FIND_LOCALE;
		}
		s += 5;
	} while (lang_cult < __LOCALE_DATA_NUM_LOCALE_NAMES);
	return 0;					/* No matching language_culture! */

 FIND_LOCALE:					/* Find locale row matching name and codeset */
	s = LOCALES;
	n = 0;
	do {						/* TODO -- do a binary search? */
		if ((lang_cult == *s) && ((codeset == s[1]) || (codeset == s[2]))) {
			i = 1;
			s = new_locale + 1;
			do {
				if (category_mask & i) {
					/* Encode current locale row number. */
					((unsigned char *) s)[0] = (n >> 7) | 0x80;
					((unsigned char *) s)[1] = (n & 0x7f) | 0x80;
				}
				s += 2;
				i += i;
			} while (i < (1 << LC_ALL));

			return i;			/* Return non-zero */
		}
		s += __LOCALE_DATA_WIDTH_LOCALES;
		++n;
	} while (n <= __LOCALE_DATA_NUM_LOCALES); /* We started at 1!!! */

	return 0;					/* Unsupported locale. */
}

static unsigned char *composite_locale(int category_mask, const char *locale,
									   unsigned char *new_locale)
{
	char buf[MAX_LOCALE_STR];
	char *t;
	char *e;
	int c;
	int component_mask;

	if (!strchr(locale,'=')) {
		if (!find_locale(category_mask, locale, new_locale)) {
			return NULL;
		}
		return new_locale;
	}

	if (strlen(locale) >= sizeof(buf)) {
		return NULL;
	}
	stpcpy(buf, locale);

	component_mask = 0;
	t = strtok_r(buf, "=", &e);	/* This can't fail because of strchr test above. */
	do {
		c = 0;
		while (strcmp(CATEGORY_NAMES + (int) CATEGORY_NAMES[c], t)) {
			if (++c == LC_ALL) { /* Unknown category name! */
				return NULL;
			}
		}
		t = strtok_r(NULL, ";", &e);
		c = (1 << c);
		if (component_mask & c) { /* Multiple components for one category. */
			return NULL;
		}
		component_mask |= c;
		if ((category_mask & c) && (!t || !find_locale(c, t, new_locale))) {
			return NULL;
		}
	} while ((t = strtok_r(NULL, "=", &e)) != NULL);

	if (category_mask & ~component_mask) { /* Category component(s) missing. */
		return NULL;
	}

	return new_locale;
}

__locale_t __newlocale(int category_mask, const char *locale, __locale_t base)
{
	const unsigned char *p;
	int i, j, k;
	unsigned char new_selector[LOCALE_SELECTOR_SIZE];

	if (category_mask == (1 << LC_ALL)) {
		category_mask = LC_ALL_MASK;
	}

	if (!locale || (((unsigned int)(category_mask)) > LC_ALL_MASK)) {
	INVALID:
		__set_errno(EINVAL);
		return NULL;  /* No locale or illegal/unsupported category. */
	}

#ifdef __UCLIBC_MJN3_ONLY__
#warning TODO: Rename cur_locale to locale_selector.
#endif
	strcpy((char *) new_selector,
		   (base ? (char *) base->cur_locale : C_LOCALE_SELECTOR));

	if (!*locale) {			 /* locale == "", so check environment. */
#ifndef __UCLIBC_HAS_THREADS__
		static 				/* If no threads, then envstr can be static. */
#endif /*  __UCLIBC_HAS_THREADS__ */
			const char *envstr[4] = { "LC_ALL", NULL, "LANG", posix };

		i = 1;
		k = 0;
		do {
			if (category_mask & i) {
				/* Note: SUSv3 doesn't define a fallback mechanism here.
				 * So, if LC_ALL is invalid, we do _not_ continue trying
				 * the other environment vars. */
				envstr[1] = CATEGORY_NAMES + CATEGORY_NAMES[k];
				j = 0;
				do {
					p = envstr[j];
				} while ((++j < 4) && (!(p = getenv(p)) || !*p));


				/* The user set something... is it valid? */
				/* Note: Since we don't support user-supplied locales and
				 * alternate paths, we don't need to worry about special
				 * handling for suid/sgid apps. */
				if (!find_locale(i, p, new_selector)) {
					goto INVALID;
				}
			}
			i += i;
		} while (++k < LC_ALL);
	} else if (!composite_locale(category_mask, locale, new_selector)) {
		goto INVALID;
	}

#ifdef __UCLIBC_MJN3_ONLY__
#warning TODO: Do a compatible codeset check!
#endif

	/* If we get here, the new selector corresponds to a valid locale. */

#ifdef __UCLIBC_MJN3_ONLY__
#warning CONSIDER: Probably want a _locale_new func to allow for caching of locales.
#endif
#if 0
	if (base) {
		_locale_set_l(new_selector, base);
	} else {
		base = _locale_new(new_selector);
	}
#else
	if (!base) {
		if ((base = malloc(sizeof(__uclibc_locale_t))) == NULL) {
			return base;
		}
		_locale_init_l(base);
	}

	_locale_set_l(new_selector, base);
#endif

	return base;
}

weak_alias(__newlocale, newlocale)

#endif
/**********************************************************************/
#ifdef L_duplocale

#ifdef __UCLIBC_MJN3_ONLY__
#warning REMINDER: When we allocate ctype tables, remember to dup them.
#endif

__locale_t __duplocale(__locale_t dataset)
{
	__locale_t r;
	uint16_t * i2w;
	size_t n;

	assert(dataset != LC_GLOBAL_LOCALE);

	if ((r = malloc(sizeof(__uclibc_locale_t))) != NULL) {
		n = 2*dataset->collate.max_col_index+2;
		if ((i2w = calloc(n, sizeof(uint16_t)))
			!= NULL
			) {
			memcpy(r, dataset, sizeof(__uclibc_locale_t));
			r->collate.index2weight = i2w;
			memcpy(i2w, dataset->collate.index2weight, n * sizeof(uint16_t));
		} else {
			free(r);
			r = NULL;
		}
	}
	return r;
}

weak_alias(__duplocale, duplocale)

#endif
/**********************************************************************/
#ifdef L_freelocale

#ifdef __UCLIBC_MJN3_ONLY__
#warning REMINDER: When we allocate ctype tables, remember to free them.
#endif

void __freelocale(__locale_t dataset)
{
	assert(dataset != __global_locale);
	assert(dataset != LC_GLOBAL_LOCALE);

	free(dataset->collate.index2weight); /* Free collation data. */
	free(dataset);				/* Free locale */
}

weak_alias(__freelocale, freelocale)

#endif
/**********************************************************************/
#ifdef L_uselocale

__locale_t __uselocale(__locale_t dataset)
{
	__locale_t old;

	if (!dataset) {
		old = __UCLIBC_CURLOCALE;
	} else {
		if (dataset == LC_GLOBAL_LOCALE) {
			dataset = __global_locale;
		}
#ifdef __UCLIBC_HAS_THREADS__
		old = __curlocale_set(dataset);
#else
		old = __curlocale_var;
		__curlocale_var = dataset;
#endif
	}

	if (old == __global_locale) {
		return LC_GLOBAL_LOCALE;
	}
	return old;
}

weak_alias(__uselocale, uselocale)

#endif
/**********************************************************************/
#ifdef L___curlocale

#ifdef __UCLIBC_HAS_THREADS__

__locale_t weak_const_function __curlocale(void)
{
    return __curlocale_var; /* This is overriden by the thread version. */
}

__locale_t weak_function __curlocale_set(__locale_t newloc)
{
	__locale_t oldloc = __curlocale_var;
	assert(newloc != LC_GLOBAL_LOCALE);
	__curlocale_var = newloc;
	return oldloc;
}

#endif

#endif
/**********************************************************************/
#ifdef L___locale_mbrtowc_l

/* NOTE: This returns an int... not size_t.  Also, it is not a general
 * routine.  It is actually a very stripped-down version of mbrtowc
 * that takes a __locale_t arg.  This is used by strcoll and strxfrm.
 * It is also used above to generate wchar_t versions of the decimal point
 * and thousands seperator. */


#ifndef __CTYPE_HAS_UTF_8_LOCALES
#warning __CTYPE_HAS_UTF_8_LOCALES not set!
#endif
#ifndef __CTYPE_HAS_8_BIT_LOCALES
#warning __CTYPE_HAS_8_BIT_LOCALES not set!
#endif

#define Cc2wc_IDX_SHIFT		__LOCALE_DATA_Cc2wc_IDX_SHIFT
#define Cc2wc_ROW_LEN		__LOCALE_DATA_Cc2wc_ROW_LEN

extern size_t _wchar_utf8sntowcs(wchar_t *__restrict pwc, size_t wn,
								 const char **__restrict src, size_t n,
								 mbstate_t *ps, int allow_continuation);

int __locale_mbrtowc_l(wchar_t *__restrict dst,
					   const char *__restrict src,
					   __locale_t loc )
{
#ifdef __CTYPE_HAS_UTF_8_LOCALES
	if (loc->encoding == __ctype_encoding_utf8) {
		mbstate_t ps;
		const char *p = src;
		size_t r;
		ps.__mask = 0;
		r = _wchar_utf8sntowcs(dst, 1, &p, SIZE_MAX, &ps, 1);
		return (r == 1) ? (p-src) : r; /* Need to return 0 if nul char. */
	}
#endif

#ifdef __CTYPE_HAS_8_BIT_LOCALES
	assert((loc->encoding == __ctype_encoding_7_bit) || (loc->encoding == __ctype_encoding_8_bit));
#else
	assert(loc->encoding == __ctype_encoding_7_bit);
#endif

	if ((*dst = ((unsigned char)(*src))) < 0x80) {	/* ASCII... */
		return (*src != 0);
	}

#ifdef __CTYPE_HAS_8_BIT_LOCALES
	if (loc->encoding == __ctype_encoding_8_bit) {
		wchar_t wc = *dst - 0x80;
		*dst = loc->tbl8c2wc[
						(loc->idx8c2wc[wc >> Cc2wc_IDX_SHIFT]
						 << Cc2wc_IDX_SHIFT) + (wc & (Cc2wc_ROW_LEN - 1))];
		if (*dst) {
			return 1;
		}
	}
#endif

	return -1;
}

#endif
/**********************************************************************/