Fix a few bugs in the new extended locale functions.

Move stub gettext functions to a stub libintl to make switching in
  gnu gettext easier.  Also add a few gnu-isms.
Change to using hidden names with global weak aliases for the extended
  locale functions, as expected by libstd++.
Slightly rework the locale data generation stuff to allow pregenerated
  locale data to be used with buildroot.

9 files changed, 82 insertions, 154 deletions

diff --git a/libc/misc/ctype/ctype.c b/libc/misc/ctype/ctype.c
index 0eb97140d..a89e1e75c 100644
--- a/libc/misc/ctype/ctype.c
+++ b/libc/misc/ctype/ctype.c
@@ -256,8 +256,9 @@ int tolower(int c)
 #ifdef L_tolower_l
 
 #undef tolower_l
+#undef __tolower_l
 
-int tolower_l(int c, __locale_t l)
+int __tolower_l(int c, __locale_t l)
 {
 #if defined(__UCLIBC_HAS_CTYPE_ENFORCED__)
 	assert(CTYPE_DOMAIN_CHECK(c));
@@ -265,6 +266,8 @@ int tolower_l(int c, __locale_t l)
 	return __UCLIBC_CTYPE_IN_TO_DOMAIN(c) ? l->__ctype_tolower[c] : c;
 }
 
+weak_alias(__tolower_l, tolower_l)
+
 #endif
 /**********************************************************************/
 #ifdef L_toupper
@@ -282,8 +285,9 @@ int toupper(int c)
 #ifdef L_toupper_l
 
 #undef toupper_l
+#undef __toupper_l
 
-int toupper_l(int c, __locale_t l)
+int __toupper_l(int c, __locale_t l)
 {
 #if defined(__UCLIBC_HAS_CTYPE_ENFORCED__)
 	assert(CTYPE_DOMAIN_CHECK(c));
@@ -291,6 +295,8 @@ int toupper_l(int c, __locale_t l)
 	return __UCLIBC_CTYPE_IN_TO_DOMAIN(c) ? l->__ctype_toupper[c] : c;
 }
 
+weak_alias(__toupper_l, toupper_l)
+
 #endif
 /**********************************************************************/
 #if defined(L_isascii) || defined(L_isascii_l)
@@ -300,6 +306,8 @@ int __XL(isascii)(int c)
 	return __isascii(c);		/* locale-independent */
 }
 
+__XL_ALIAS(isascii)
+
 #endif
 /**********************************************************************/
 #if defined(L_toascii) || defined(L_toascii_l)
@@ -309,6 +317,8 @@ int __XL(toascii)(int c)
 	return __toascii(c);		/* locale-independent */
 }
 
+__XL_ALIAS(toascii)
+
 #endif
 /**********************************************************************/
 /* old uClibc extensions */
diff --git a/libc/misc/gnu/obstack.c b/libc/misc/gnu/obstack.c
index 87b294c25..a3f7f3f45 100644
--- a/libc/misc/gnu/obstack.c
+++ b/libc/misc/gnu/obstack.c
@@ -466,7 +466,7 @@ _obstack_memory_used (h)
 #  if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
 #   include <libintl.h>
 #   ifndef _
-#    define _(Str) gettext (Str)
+#    define _(Str) __dcgettext (NULL, Str, LC_MESSAGES)
 #   endif
 #  else
 #   define _(Str) (Str)
diff --git a/libc/misc/intl/Makefile b/libc/misc/intl/Makefile
index 75fbb321f..9975e5eb1 100644
--- a/libc/misc/intl/Makefile
+++ b/libc/misc/intl/Makefile
@@ -25,8 +25,7 @@ TOPDIR=../../../
 include $(TOPDIR)Rules.mak
 
 MSRC= intl.c
-MOBJ= gettext.o ngettext.o  dgettext.o dcgettext.o dngettext.o dcngettext.o \
-	textdomain.o bindtextdomain.o bind_textdomain_codeset.o
+MOBJ= __uClibc_dgettext.o __uClibc_dcgettext.o
 
 OBJS=$(MOBJ)
 
diff --git a/libc/misc/intl/intl.c b/libc/misc/intl/intl.c
index 183ffcc92..eca6fa96a 100644
--- a/libc/misc/intl/intl.c
+++ b/libc/misc/intl/intl.c
@@ -16,7 +16,7 @@
  */
 
 /*
- *  Supply some weaks for gettext and friends.  Used by strerror*().
+ *  Supply some weaks for use by strerror*(), etc.
  */
 
 #include <stdlib.h>
@@ -27,123 +27,27 @@
 #include <libintl.h>
 
 /**********************************************************************/
-#ifdef L_gettext
-
-char *weak_function gettext(const char *msgid)
-{
-	return (char *) msgid;
-}
-
-#endif
-/**********************************************************************/
-#ifdef L_dgettext
+#ifdef L___uClibc_dgettext
 
 char *__uClibc_dgettext(const char *domainname,
-									  const char *msgid)
+						const char *msgid)
 {
 	return (char *) msgid;
 }
 
 weak_alias (__uClibc_dgettext, __dgettext)
-weak_alias (__uClibc_dgettext, dgettext)
 
 #endif
 /**********************************************************************/
-#ifdef L_dcgettext
+#ifdef L___uClibc_dcgettext
 
 char * __uClibc_dcgettext(const char *domainname,
-									   const char *msgid, int category)
+						  const char *msgid, int category)
 {
 	return (char *) msgid;
 }
 
 weak_alias (__uClibc_dcgettext, __dcgettext)
-weak_alias (__uClibc_dcgettext, dcgettext)
-
-#endif
-/**********************************************************************/
-#ifdef L_ngettext
-
-char *weak_function ngettext(const char *msgid1, const char *msgid2,
-							 unsigned long int n)
-{
-	return (char *) ((n == 1) ? msgid1 : msgid2);
-}
-
-#endif
-/**********************************************************************/
-#ifdef L_dngettext
-
-char *weak_function dngettext(const char *domainname, const char *msgid1,
-							  const char *msgid2, unsigned long int n)
-{
-	return (char *) ((n == 1) ? msgid1 : msgid2);
-}
-
-#endif
-/**********************************************************************/
-#ifdef L_dcngettext
-
-char *weak_function dcngettext(const char *domainname, const char *msgid1,
-							   const char *msgid2, unsigned long int n,
-							   int category)
-{
-	return (char *) ((n == 1) ? msgid1 : msgid2);
-}
-
-#endif
-/**********************************************************************/
-#ifdef L_textdomain
-
-char *weak_function textdomain(const char *domainname)
-{
-	static const char default_str[] = "messages";
-
-	if (domainname && *domainname && strcmp(domainname, default_str)) {
-		__set_errno(EINVAL);
-		return NULL;
-	}
-	return (char *) default_str;
-}
-
-#endif
-/**********************************************************************/
-#ifdef L_bindtextdomain
-
-char *weak_function bindtextdomain(const char *domainname, const char *dirname)
-{
-	static const char dir[] = "/";
-
-	if (!domainname || !*domainname
-		|| (dirname
-#if 1
-			&& ((dirname[0] != '/') || dirname[1])
-#else
-			&& strcmp(dirname, dir)
-#endif
-			)
-		) {
-		__set_errno(EINVAL);
-		return NULL;
-	}
-
-	return (char *) dir;
-}
-
-#endif
-/**********************************************************************/
-#ifdef L_bind_textdomain_codeset
-
-/* Specify the character encoding in which the messages from the
-   DOMAINNAME message catalog will be returned.  */
-char *weak_function bind_textdomain_codeset(const char *domainname,
-											const char *codeset)
-{
-	if (!domainname || !*domainname || codeset) {
-		__set_errno(EINVAL);
-	}
-	return NULL;
-}
 
 #endif
 /**********************************************************************/
diff --git a/libc/misc/locale/.cvsignore b/libc/misc/locale/.cvsignore
index 11e4e51bc..3209b4ee4 100644
--- a/libc/misc/locale/.cvsignore
+++ b/libc/misc/locale/.cvsignore
@@ -1 +1,2 @@
+locale_data.c
 locale_data.o
diff --git a/libc/misc/locale/Makefile b/libc/misc/locale/Makefile
index 8ad041e69..17b208b2e 100644
--- a/libc/misc/locale/Makefile
+++ b/libc/misc/locale/Makefile
@@ -36,18 +36,14 @@ ifeq ($(UCLIBC_HAS_XLOCALE),y)
 	MOBJx += nl_langinfo_l.o duplocale.o freelocale.o uselocale.o __curlocale.o
 endif
 
-OBJS= $(MOBJ) $(MOBJx)
-
-ifeq ($(UCLIBC_HAS_LOCALE),y)
-	OBJS += $(COBJS) locale_data.o
-endif
-
 DATA=
 ifeq ($(UCLIBC_HAS_LOCALE),y)
-	DATA += locale_data.o
+	DATA = locale_data.o
 endif
 
-all: $(DATA) $(OBJS) $(LIBC)
+OBJS= $(MOBJ) $(MOBJx) $(DATA)
+
+all: $(OBJS) $(LIBC)
 
 
 $(LIBC): ar-target
@@ -63,12 +59,14 @@ $(MOBJx): $(MSRC)
 	$(CC) $(CFLAGS) -DL_$* -D__UCLIBC_DO_XLOCALE $< -c -o $*.o
 	$(STRIPTOOL) -x -R .note -R .comment $*.o
 
-$(OBJS): Makefile
+locale_data.o:
+	ln -sf ../../../extra/locale/locale_data.c
+	$(CC) $(CFLAGS) -c -D__WCHAR_ENABLED -I../../../extra/locale locale_data.c -o $@
+	$(STRIPTOOL) -x -R .note -R .comment $*.o
 
-# locale_data.o:
-# 	make -C $(TOPDIR)/extra/locale
+$(OBJS): Makefile
 
 clean:
-	rm -f *.[oa] *~ core
+	rm -f *.[oa] *~ core locale_data.c
 
 .PHONY: data
diff --git a/libc/misc/locale/locale.c b/libc/misc/locale/locale.c
index b101f6dad..8f0cbd415 100644
--- a/libc/misc/locale/locale.c
+++ b/libc/misc/locale/locale.c
@@ -16,7 +16,6 @@
  */
 
 /* Nov. 1, 2002
- *
  * Reworked setlocale() return values and locale arg processing to
  *   be more like glibc.  Applications expecting to be able to
  *   query locale settings should now work... at the cost of almost
@@ -24,8 +23,15 @@
  * Fixed a bug in the internal fixed-size-string locale specifier code.
  *
  * Dec 20, 2002
- *
  * Added in collation support and updated stub nl_langinfo.
+ *
+ * Aug 1, 2003
+ * Added glibc-like extended locale stuff (newlocale, duplocale, etc).
+ *
+ * Aug 18, 2003
+ * Bug in duplocale... collation data wasn't copied.
+ * Bug in newlocale... translate 1<<LC_ALL to LC_ALL_MASK.
+ * Bug in _wchar_utf8sntowcs... fix cut-n-paste error.
  */
 
 
@@ -48,6 +54,8 @@
 #include <assert.h>
 #include <errno.h>
 #include <ctype.h>
+#warning devel code
+#include <stdio.h>
 
 #undef __LOCALE_C_ONLY
 #ifndef __UCLIBC_HAS_LOCALE__
@@ -249,10 +257,8 @@ char *setlocale(int category, const char *locale)
 	}
 
 	if (locale != NULL) {		/* Not just a query... */
-		if (!__newlocale((category == LC_ALL) ? LC_ALL_MASK : (1 << category),
-						 locale, __global_locale)
-			) { /* Failed! */
-			return NULL;
+		if (!__newlocale((1 << category), locale, __global_locale)) {
+			return NULL;		/* Failed! */
 		}
 		update_hr_locale(__global_locale->cur_locale);
 	}
@@ -436,35 +442,36 @@ static int init_cur_collate(int der_num, __collate_t *cur_collate)
 	cur_collate->ti_mask = (1 << cur_collate->ti_shift)-1;
 	cur_collate->ii_mask = (1 << cur_collate->ii_shift)-1;
 
-/*	 printf("base=%d  num_col_base: %d  %d\n", cdd->base_idx ,cur_collate->num_col_base, cdb->num_col_base); */
+/* 	fflush(stdout); */
+/* 	fprintf(stderr,"base=%d  num_col_base: %d  %d\n", cdd->base_idx ,cur_collate->num_col_base, cdb->num_col_base); */
 
 	n = (sizeof(coldata_header_t) + cdh->num_base * sizeof(coldata_base_t)
 		 + cdh->num_der * sizeof(coldata_der_t))/2;
 
-/*	 printf("n   = %d\n", n); */
+/* 	fprintf(stderr,"n   = %d\n", n); */
 	cur_collate->index2weight_tbl = __locale_collate_tbl + n + cdb->index2weight_offset;
-/*	 printf("i2w = %d\n", n + cdb->index2weight_offset); */
+/* 	fprintf(stderr,"i2w = %d\n", n + cdb->index2weight_offset); */
 	n += cdh->num_index2weight;
 	cur_collate->index2ruleidx_tbl = __locale_collate_tbl + n + cdb->index2ruleidx_offset;
-/*	 printf("i2r = %d\n", n + cdb->index2ruleidx_offset); */
+/* 	fprintf(stderr,"i2r = %d\n", n + cdb->index2ruleidx_offset); */
 	n += cdh->num_index2ruleidx;
 	cur_collate->multistart_tbl = __locale_collate_tbl + n + cdd->multistart_offset;
-/*	 printf("mts = %d\n", n + cdb->multistart_offset); */
+/* 	fprintf(stderr,"mts = %d\n", n + cdb->multistart_offset); */
 	n += cdh->num_multistart;
 	cur_collate->overrides_tbl = __locale_collate_tbl + n + cdd->overrides_offset;
-/*	 printf("ovr = %d\n", n + cdd->overrides_offset); */
+/* 	fprintf(stderr,"ovr = %d\n", n + cdd->overrides_offset); */
 	n += cdh->num_override;
 	cur_collate->ruletable = __locale_collate_tbl + n;
-/*	 printf("rtb = %d\n", n); */
+/* 	fprintf(stderr, "rtb = %d\n", n); */
 	n += cdh->num_ruletable;
 	cur_collate->weightstr = __locale_collate_tbl + n;
-/*	 printf("wts = %d\n", n); */
+/* 	fprintf(stderr,"wts = %d\n", n); */
 	n += cdh->num_weightstr;
 	cur_collate->wcs2colidt_tbl = __locale_collate_tbl + n
 		+ (((unsigned long)(cdb->wcs2colidt_offset_hi)) << 16)
 		+ cdb->wcs2colidt_offset_low;
-/*	 printf("wcs = %lu\n", n	+ (((unsigned long)(cdb->wcs2colidt_offset_hi)) << 16) */
-/* 		   + cdb->wcs2colidt_offset_low); */
+/* 	fprintf(stderr,"wcs = %lu\n", n	+ (((unsigned long)(cdb->wcs2colidt_offset_hi)) << 16) */
+/* 			+ cdb->wcs2colidt_offset_low); */
 
 	cur_collate->MAX_WEIGHTS = cdh->MAX_WEIGHTS;
 
@@ -494,13 +501,15 @@ static int init_cur_collate(int der_num, __collate_t *cur_collate)
 		w = *p++;
 		do {
 			i = *p++;
-/* 			fprintf(stderr, "	i=%d w=%d *p=%d\n", i, w, *p); */
+/* 			fprintf(stderr, "	i=%d (%#x) w=%d *p=%d\n", i, i, w, *p); */
 			cur_collate->index2weight[i-1] = w++;
 			cur_collate->index2ruleidx[i-1] = *p++;
 		} while (--n);
 	}
+	assert(*p == 1);
 	while (*++p) {
 		i = *p;
+/* 		fprintf(stderr, "	i=%d (%#x) w=%d *p=%d\n", i, i, p[1], p[2]); */
 		cur_collate->index2weight[i-1] = *++p;
 		cur_collate->index2ruleidx[i-1] = *++p;
 	}
@@ -869,7 +878,7 @@ char *nl_langinfo(nl_item item)
 
 char *nl_langinfo(nl_item item)
 {
-	return nl_langinfo_l(item, __UCLIBC_CURLOCALE);
+	return __nl_langinfo_l(item, __UCLIBC_CURLOCALE);
 }
 
 #else  /* defined(__UCLIBC_HAS_XLOCALE__) && !defined(__UCLIBC_DO_XLOCALE) */
@@ -1056,6 +1065,10 @@ __locale_t __newlocale(int category_mask, const char *locale, __locale_t base)
 	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);
@@ -1141,20 +1154,22 @@ weak_alias(__newlocale, newlocale)
 #warning REMINDER: When we allocate ctype tables, remember to dup them.
 #endif
 
-__locale_t duplocale(__locale_t dataset)
+__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) {
-		if ((i2w = calloc(2*dataset->collate.max_col_index+2,
-						  sizeof(uint16_t)))
+		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;
@@ -1163,6 +1178,8 @@ __locale_t duplocale(__locale_t dataset)
 	return r;
 }
 
+weak_alias(__duplocale, duplocale)
+
 #endif
 /**********************************************************************/
 #ifdef L_freelocale
@@ -1171,7 +1188,7 @@ __locale_t duplocale(__locale_t dataset)
 #warning REMINDER: When we allocate ctype tables, remember to free them.
 #endif
 
-void freelocale(__locale_t dataset)
+void __freelocale(__locale_t dataset)
 {
 	assert(dataset != __global_locale);
 	assert(dataset != LC_GLOBAL_LOCALE);
@@ -1180,11 +1197,13 @@ void freelocale(__locale_t dataset)
 	free(dataset);				/* Free locale */
 }
 
+weak_alias(__freelocale, freelocale)
+
 #endif
 /**********************************************************************/
 #ifdef L_uselocale
 
-__locale_t uselocale(__locale_t dataset)
+__locale_t __uselocale(__locale_t dataset)
 {
 	__locale_t old;
 
@@ -1208,6 +1227,8 @@ __locale_t uselocale(__locale_t dataset)
 	return old;
 }
 
+weak_alias(__uselocale, uselocale)
+
 #endif
 /**********************************************************************/
 #ifdef L___curlocale
@@ -1283,8 +1304,8 @@ int __locale_mbrtowc_l(wchar_t *__restrict dst,
 #ifdef __CTYPE_HAS_8_BIT_LOCALES
 	if (loc->encoding == __ctype_encoding_8_bit) {
 		wchar_t wc = *dst - 0x80;
-		*dst = __LOCALE_PTR->tbl8c2wc[
-						(__LOCALE_PTR->idx8c2wc[wc >> Cc2wc_IDX_SHIFT]
+		*dst = loc->tbl8c2wc[
+						(loc->idx8c2wc[wc >> Cc2wc_IDX_SHIFT]
 						 << Cc2wc_IDX_SHIFT) + (wc & (Cc2wc_ROW_LEN - 1))];
 		if (*dst) {
 			return 1;
diff --git a/libc/misc/time/time.c b/libc/misc/time/time.c
index 0ae91f69a..9bc8bc4cf 100644
--- a/libc/misc/time/time.c
+++ b/libc/misc/time/time.c
@@ -652,7 +652,7 @@ size_t strftime(char *__restrict s, size_t maxsize,
 				const char *__restrict format,
 				const struct tm *__restrict timeptr)
 {
-	return strftime_l(s, maxsize, format, timeptr, __UCLIBC_CURLOCALE);
+	return __strftime_l(s, maxsize, format, timeptr, __UCLIBC_CURLOCALE);
 }
 
 #else  /* defined(__UCLIBC_HAS_XLOCALE__) && !defined(__UCLIBC_DO_XLOCALE) */
@@ -1106,6 +1106,8 @@ size_t __XL(strftime)(char *__restrict s, size_t maxsize,
 	goto LOOP;
 }
 
+__XL_ALIAS(strftime)
+
 #endif /* defined(__UCLIBC_HAS_XLOCALE__) && !defined(__UCLIBC_DO_XLOCALE) */
 
 #endif
@@ -1127,7 +1129,7 @@ size_t __XL(strftime)(char *__restrict s, size_t maxsize,
 char *strptime(const char *__restrict buf, const char *__restrict format,
 			   struct tm *__restrict tm)
 {
-	return strptime_l(buf, format, tm, __UCLIBC_CURLOCALE);
+	return __strptime_l(buf, format, tm, __UCLIBC_CURLOCALE);
 }
 
 #else  /* defined(__UCLIBC_HAS_XLOCALE__) && !defined(__UCLIBC_DO_XLOCALE) */
@@ -1482,6 +1484,8 @@ char *__XL(strptime)(const char *__restrict buf, const char *__restrict format,
 	return NULL;
 }
 
+__XL_ALIAS(strptime)
+
 #endif /* defined(__UCLIBC_HAS_XLOCALE__) && !defined(__UCLIBC_DO_XLOCALE) */
 
 #endif
diff --git a/libc/misc/wctype/wctype.c b/libc/misc/wctype/wctype.c
index e0dde3a1f..851d2625b 100644
--- a/libc/misc/wctype/wctype.c
+++ b/libc/misc/wctype/wctype.c
@@ -44,15 +44,6 @@
 
 #ifdef __UCLIBC_HAS_XLOCALE__
 #include <xlocale.h>
-
-extern int __iswctype_l (wint_t __wc, wctype_t __desc, __locale_t __locale);
-     __THROW;
-extern wint_t __towlower_l (wint_t __wc, __locale_t __locale) __THROW;
-
-extern wint_t __towupper_l (wint_t __wc, __locale_t __locale) __THROW;
-extern wint_t __towctrans_l (wint_t __wc, wctrans_t __desc,
-			   __locale_t __locale) __THROW;
-
 #endif /* __UCLIBC_HAS_XLOCALE__ */
 
 /* We know wide char support is enabled.  We wouldn't be here otherwise. */