summaryrefslogtreecommitdiffstats
path: root/lib/libc/locale
diff options
context:
space:
mode:
authortjr <tjr@FreeBSD.org>2004-04-07 10:48:19 +0000
committertjr <tjr@FreeBSD.org>2004-04-07 10:48:19 +0000
commit54a18fa1d63375b790f3bf3157f6b64b294e5d16 (patch)
tree44e2a3f83138ea23f4b4f68cbea18e4f60c8745e /lib/libc/locale
parenta6980b04fc41a4c6dc314dc3aa00de7e7834ba7b (diff)
downloadFreeBSD-src-54a18fa1d63375b790f3bf3157f6b64b294e5d16.zip
FreeBSD-src-54a18fa1d63375b790f3bf3157f6b64b294e5d16.tar.gz
Allow partial multibyte characters to accumulate in conversion state
objects passed to mbrtowc(), mbsrtowcs(), and mbrlen(), as required by C99.
Diffstat (limited to 'lib/libc/locale')
-rw-r--r--lib/libc/locale/big5.c48
-rw-r--r--lib/libc/locale/euc.c50
-rw-r--r--lib/libc/locale/gb18030.c50
-rw-r--r--lib/libc/locale/gb2312.c47
-rw-r--r--lib/libc/locale/gbk.c48
-rw-r--r--lib/libc/locale/mbsinit.c13
-rw-r--r--lib/libc/locale/mskanji.c54
-rw-r--r--lib/libc/locale/none.c16
-rw-r--r--lib/libc/locale/setrunelocale.c7
-rw-r--r--lib/libc/locale/table.c2
-rw-r--r--lib/libc/locale/utf2.c51
-rw-r--r--lib/libc/locale/utf8.c53
12 files changed, 350 insertions, 89 deletions
diff --git a/lib/libc/locale/big5.c b/lib/libc/locale/big5.c
index 3b1dabe..ccf35f0 100644
--- a/lib/libc/locale/big5.c
+++ b/lib/libc/locale/big5.c
@@ -1,5 +1,5 @@
/*-
- * Copyright (c) 2002, 2003 Tim J. Robbins. All rights reserved.
+ * Copyright (c) 2002-2004 Tim J. Robbins. All rights reserved.
* Copyright (c) 1993
* The Regents of the University of California. All rights reserved.
*
@@ -38,34 +38,49 @@
#if defined(LIBC_SCCS) && !defined(lint)
static char sccsid[] = "@(#)big5.c 8.1 (Berkeley) 6/4/93";
#endif /* LIBC_SCCS and not lint */
-#include <sys/cdefs.h>
+#include <sys/param.h>
__FBSDID("$FreeBSD$");
-#include <sys/types.h>
#include <runetype.h>
#include <stdlib.h>
+#include <string.h>
#include <wchar.h>
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
int _BIG5_init(_RuneLocale *);
size_t _BIG5_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
+int _BIG5_mbsinit(const mbstate_t *);
size_t _BIG5_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+typedef struct {
+ int count;
+ u_char bytes[2];
+} _BIG5State;
+
int
_BIG5_init(_RuneLocale *rl)
{
__mbrtowc = _BIG5_mbrtowc;
__wcrtomb = _BIG5_wcrtomb;
+ __mbsinit = _BIG5_mbsinit;
_CurrentRuneLocale = rl;
__mb_cur_max = 2;
return (0);
}
+int
+_BIG5_mbsinit(const mbstate_t *ps)
+{
+
+ return (ps == NULL || ((_BIG5State *)ps)->count == 0);
+}
+
static __inline int
_big5_check(u_int c)
{
@@ -76,14 +91,28 @@ _big5_check(u_int c)
size_t
_BIG5_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
- mbstate_t * __restrict ps __unused)
+ mbstate_t * __restrict ps)
{
+ _BIG5State *bs;
wchar_t wc;
- int i, len;
+ int i, len, ocount;
+ size_t ncopy;
+
+ bs = (_BIG5State *)ps;
+
+ if (s == NULL) {
+ s = "";
+ n = 1;
+ pwc = NULL;
+ }
+
+ ncopy = MIN(MIN(n, MB_CUR_MAX), sizeof(bs->bytes) - bs->count);
+ memcpy(bs->bytes + bs->count, s, ncopy);
+ ocount = bs->count;
+ bs->count += ncopy;
+ s = (char *)bs->bytes;
+ n = bs->count;
- if (s == NULL)
- /* Reset to initial shift state (no-op) */
- return (0);
if (n == 0 || (size_t)(len = _big5_check(*s)) > n)
/* Incomplete multibyte sequence */
return ((size_t)-2);
@@ -93,7 +122,8 @@ _BIG5_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
wc = (wc << 8) | (unsigned char)*s++;
if (pwc != NULL)
*pwc = wc;
- return (wc == L'\0' ? 0 : len);
+ bs->count = 0;
+ return (wc == L'\0' ? 0 : len - ocount);
}
size_t
diff --git a/lib/libc/locale/euc.c b/lib/libc/locale/euc.c
index 74ed52e..490b1c8 100644
--- a/lib/libc/locale/euc.c
+++ b/lib/libc/locale/euc.c
@@ -1,5 +1,5 @@
/*-
- * Copyright (c) 2002, 2003 Tim J. Robbins. All rights reserved.
+ * Copyright (c) 2002-2004 Tim J. Robbins. All rights reserved.
* Copyright (c) 1993
* The Regents of the University of California. All rights reserved.
*
@@ -38,23 +38,25 @@
#if defined(LIBC_SCCS) && !defined(lint)
static char sccsid[] = "@(#)euc.c 8.1 (Berkeley) 6/4/93";
#endif /* LIBC_SCCS and not lint */
-#include <sys/cdefs.h>
+#include <sys/param.h>
__FBSDID("$FreeBSD$");
-#include <sys/types.h>
-
#include <errno.h>
+#include <limits.h>
#include <runetype.h>
#include <stdlib.h>
+#include <string.h>
#include <wchar.h>
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
int _EUC_init(_RuneLocale *);
size_t _EUC_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
+int _EUC_mbsinit(const mbstate_t *);
size_t _EUC_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
typedef struct {
@@ -63,6 +65,11 @@ typedef struct {
wchar_t mask;
} _EucInfo;
+typedef struct {
+ int count;
+ u_char bytes[MB_LEN_MAX];
+} _EucState;
+
int
_EUC_init(_RuneLocale *rl)
{
@@ -111,9 +118,17 @@ _EUC_init(_RuneLocale *rl)
__mb_cur_max = new__mb_cur_max;
__mbrtowc = _EUC_mbrtowc;
__wcrtomb = _EUC_wcrtomb;
+ __mbsinit = _EUC_mbsinit;
return (0);
}
+int
+_EUC_mbsinit(const mbstate_t *ps)
+{
+
+ return (ps == NULL || ((_EucState *)ps)->count == 0);
+}
+
#define CEI ((_EucInfo *)(_CurrentRuneLocale->variable))
#define _SS2 0x008e
@@ -130,14 +145,28 @@ _euc_set(u_int c)
size_t
_EUC_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
- mbstate_t * __restrict ps __unused)
+ mbstate_t * __restrict ps)
{
- int len, remain, set;
+ _EucState *es;
+ int len, ocount, remain, set;
wchar_t wc;
+ size_t ncopy;
+
+ es = (_EucState *)ps;
+
+ if (s == NULL) {
+ s = "";
+ n = 1;
+ pwc = NULL;
+ }
+
+ ncopy = MIN(MIN(n, MB_CUR_MAX), sizeof(es->bytes) - es->count);
+ memcpy(es->bytes + es->count, s, ncopy);
+ ocount = es->count;
+ es->count += ncopy;
+ s = (char *)es->bytes;
+ n = es->count;
- if (s == NULL)
- /* Reset to initial shift state (no-op) */
- return (0);
if (n == 0 || (size_t)(len = CEI->count[set = _euc_set(*s)]) > n)
/* Incomplete multibyte sequence */
return ((size_t)-2);
@@ -158,7 +187,8 @@ _EUC_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
wc = (wc & ~CEI->mask) | CEI->bits[set];
if (pwc != NULL)
*pwc = wc;
- return (wc == L'\0' ? 0 : len);
+ es->count = 0;
+ return (wc == L'\0' ? 0 : len - ocount);
}
size_t
diff --git a/lib/libc/locale/gb18030.c b/lib/libc/locale/gb18030.c
index dca718f..817602b 100644
--- a/lib/libc/locale/gb18030.c
+++ b/lib/libc/locale/gb18030.c
@@ -29,22 +29,30 @@
* See gb18030(5) for details.
*/
-#include <sys/cdefs.h>
+#include <sys/param.h>
__FBSDID("$FreeBSD$");
#include <errno.h>
#include <runetype.h>
#include <stdlib.h>
+#include <string.h>
#include <wchar.h>
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
int _GB18030_init(_RuneLocale *);
-size_t _GB18030_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
+size_t _GB18030_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
-size_t _GB18030_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+int _GB18030_mbsinit(const mbstate_t *);
+size_t _GB18030_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+
+typedef struct {
+ int count;
+ u_char bytes[4];
+} _GB18030State;
int
_GB18030_init(_RuneLocale *rl)
@@ -52,22 +60,44 @@ _GB18030_init(_RuneLocale *rl)
__mbrtowc = _GB18030_mbrtowc;
__wcrtomb = _GB18030_wcrtomb;
+ __mbsinit = _GB18030_mbsinit;
_CurrentRuneLocale = rl;
__mb_cur_max = 4;
return (0);
}
+int
+_GB18030_mbsinit(const mbstate_t *ps)
+{
+
+ return (ps == NULL || ((_GB18030State *)ps)->count == 0);
+}
+
size_t
_GB18030_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s,
- size_t n, mbstate_t * __restrict ps __unused)
+ size_t n, mbstate_t * __restrict ps)
{
+ _GB18030State *gs;
wchar_t wch;
- int ch, len;
+ int ch, len, ocount;
+ size_t ncopy;
+
+ gs = (_GB18030State *)ps;
+
+ if (s == NULL) {
+ s = "";
+ n = 1;
+ pwc = NULL;
+ }
+
+ ncopy = MIN(MIN(n, MB_CUR_MAX), sizeof(gs->bytes) - gs->count);
+ memcpy(gs->bytes + gs->count, s, ncopy);
+ ocount = gs->count;
+ gs->count += ncopy;
+ s = (char *)gs->bytes;
+ n = gs->count;
- if (s == NULL)
- /* Reset to initial shift state (no-op) */
- return (0);
if (n == 0)
/* Incomplete multibyte sequence */
return ((size_t)-2);
@@ -116,7 +146,8 @@ _GB18030_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s,
if (pwc != NULL)
*pwc = wch;
- return (wch == L'\0' ? 0 : len);
+ gs->count = 0;
+ return (wch == L'\0' ? 0 : len - ocount);
ilseq:
errno = EILSEQ;
return ((size_t)-1);
@@ -132,7 +163,6 @@ _GB18030_wcrtomb(char * __restrict s, wchar_t wc,
if (s == NULL)
/* Reset to initial shift state (no-op) */
return (1);
-
if ((wc & ~0x7fffffff) != 0)
goto ilseq;
if (wc & 0x7f000000) {
diff --git a/lib/libc/locale/gb2312.c b/lib/libc/locale/gb2312.c
index f1895ce..8509d04 100644
--- a/lib/libc/locale/gb2312.c
+++ b/lib/libc/locale/gb2312.c
@@ -1,4 +1,5 @@
/*-
+ * Copyright (c) 2004 Tim J. Robbins. All rights reserved.
* Copyright (c) 2003 David Xu <davidxu@freebsd.org>
* All rights reserved.
*
@@ -24,23 +25,30 @@
* SUCH DAMAGE.
*/
-#include <sys/cdefs.h>
+#include <sys/param.h>
__FBSDID("$FreeBSD$");
-#include <sys/types.h>
#include <runetype.h>
#include <stdlib.h>
+#include <string.h>
#include <wchar.h>
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
int _GB2312_init(_RuneLocale *);
size_t _GB2312_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
+int _GB2312_mbsinit(const mbstate_t *);
size_t _GB2312_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+typedef struct {
+ int count;
+ u_char bytes[2];
+} _GB2312State;
+
int
_GB2312_init(_RuneLocale *rl)
{
@@ -48,10 +56,18 @@ _GB2312_init(_RuneLocale *rl)
_CurrentRuneLocale = rl;
__mbrtowc = _GB2312_mbrtowc;
__wcrtomb = _GB2312_wcrtomb;
+ __mbsinit = _GB2312_mbsinit;
__mb_cur_max = 2;
return (0);
}
+int
+_GB2312_mbsinit(const mbstate_t *ps)
+{
+
+ return (ps == NULL || ((_GB2312State *)ps)->count == 0);
+}
+
static __inline int
_GB2312_check(const char *str, size_t n)
{
@@ -77,14 +93,28 @@ _GB2312_check(const char *str, size_t n)
size_t
_GB2312_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
- mbstate_t * __restrict ps __unused)
+ mbstate_t * __restrict ps)
{
+ _GB2312State *gs;
wchar_t wc;
- int i, len;
+ int i, len, ocount;
+ size_t ncopy;
+
+ gs = (_GB2312State *)gs;
+
+ if (s == NULL) {
+ s = "";
+ n = 1;
+ pwc = NULL;
+ }
+
+ ncopy = MIN(MIN(n, MB_CUR_MAX), sizeof(gs->bytes) - gs->count);
+ memcpy(gs->bytes + gs->count, s, ncopy);
+ ocount = gs->count;
+ gs->count += ncopy;
+ s = (char *)gs->bytes;
+ n = gs->count;
- if (s == NULL)
- /* Reset to initial shift state (no-op) */
- return (0);
if ((len = _GB2312_check(s, n)) < 0)
return ((size_t)len);
wc = 0;
@@ -93,7 +123,8 @@ _GB2312_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
wc = (wc << 8) | (unsigned char)*s++;
if (pwc != NULL)
*pwc = wc;
- return (wc == L'\0' ? 0 : len);
+ gs->count = 0;
+ return (wc == L'\0' ? 0 : len - ocount);
}
size_t
diff --git a/lib/libc/locale/gbk.c b/lib/libc/locale/gbk.c
index e5572bc..880884f 100644
--- a/lib/libc/locale/gbk.c
+++ b/lib/libc/locale/gbk.c
@@ -1,5 +1,5 @@
/*-
- * Copyright (c) 2002, 2003 Tim J. Robbins. All rights reserved.
+ * Copyright (c) 2002-2004 Tim J. Robbins. All rights reserved.
* Copyright (c) 1993
* The Regents of the University of California. All rights reserved.
*
@@ -35,34 +35,49 @@
* SUCH DAMAGE.
*/
-#include <sys/cdefs.h>
+#include <sys/param.h>
__FBSDID("$FreeBSD$");
-#include <sys/types.h>
#include <runetype.h>
#include <stdlib.h>
+#include <string.h>
#include <wchar.h>
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
int _GBK_init(_RuneLocale *);
size_t _GBK_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
+int _GBK_mbsinit(const mbstate_t *);
size_t _GBK_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+typedef struct {
+ int count;
+ u_char bytes[2];
+} _GBKState;
+
int
_GBK_init(_RuneLocale *rl)
{
__mbrtowc = _GBK_mbrtowc;
__wcrtomb = _GBK_wcrtomb;
+ __mbsinit = _GBK_mbsinit;
_CurrentRuneLocale = rl;
__mb_cur_max = 2;
return (0);
}
+int
+_GBK_mbsinit(const mbstate_t *ps)
+{
+
+ return (ps == NULL || ((_GBKState *)ps)->count == 0);
+}
+
static __inline int
_gbk_check(u_int c)
{
@@ -73,14 +88,28 @@ _gbk_check(u_int c)
size_t
_GBK_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
- mbstate_t * __restrict ps __unused)
+ mbstate_t * __restrict ps)
{
+ _GBKState *gs;
wchar_t wc;
- int i, len;
+ int i, len, ocount;
+ size_t ncopy;
+
+ gs = (_GBKState *)ps;
+
+ if (s == NULL) {
+ s = "";
+ n = 1;
+ pwc = NULL;
+ }
+
+ ncopy = MIN(MIN(n, MB_CUR_MAX), sizeof(gs->bytes) - gs->count);
+ memcpy(gs->bytes + gs->count, s, ncopy);
+ ocount = gs->count;
+ gs->count += ncopy;
+ s = (char *)gs->bytes;
+ n = gs->count;
- if (s == NULL)
- /* Reset to initial shift state (no-op) */
- return (0);
if (n == 0 || (size_t)(len = _gbk_check(*s)) > n)
/* Incomplete multibyte sequence */
return ((size_t)-2);
@@ -90,7 +119,8 @@ _GBK_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
wc = (wc << 8) | (unsigned char)*s++;
if (pwc != NULL)
*pwc = wc;
- return (wc == L'\0' ? 0 : len);
+ gs->count = 0;
+ return (wc == L'\0' ? 0 : len - ocount);
}
size_t
diff --git a/lib/libc/locale/mbsinit.c b/lib/libc/locale/mbsinit.c
index 923db2f..5470789 100644
--- a/lib/libc/locale/mbsinit.c
+++ b/lib/libc/locale/mbsinit.c
@@ -1,5 +1,5 @@
/*-
- * Copyright (c) 2002 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
@@ -29,14 +29,11 @@ __FBSDID("$FreeBSD$");
#include <wchar.h>
+extern int (*__mbsinit)(const mbstate_t *);
+
int
-mbsinit(const mbstate_t *ps __unused)
+mbsinit(const mbstate_t *ps)
{
- /*
- * Stateful multibyte conversion is not supported; there are no
- * states other than the initial state.
- */
-
- return (1);
+ return (__mbsinit(ps));
}
diff --git a/lib/libc/locale/mskanji.c b/lib/libc/locale/mskanji.c
index 653e3f8..43bae6c 100644
--- a/lib/libc/locale/mskanji.c
+++ b/lib/libc/locale/mskanji.c
@@ -1,5 +1,6 @@
/*
- * Copyright (c) 2002, 2003 Tim J. Robbins. All rights reserved.
+ * Copyright (c) 2002-2004 Tim J. Robbins. All rights reserved.
+ *
* ja_JP.SJIS locale table for BSD4.4/rune
* version 1.0
* (C) Sin'ichiro MIYATANI / Phase One, Inc
@@ -35,22 +36,29 @@
#if defined(LIBC_SCCS) && !defined(lint)
static char sccsid[] = "@(#)mskanji.c 1.0 (Phase One) 5/5/95";
#endif /* LIBC_SCCS and not lint */
-#include <sys/cdefs.h>
+#include <sys/param.h>
__FBSDID("$FreeBSD$");
-#include <sys/types.h>
#include <runetype.h>
#include <stdlib.h>
+#include <string.h>
#include <wchar.h>
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
int _MSKanji_init(_RuneLocale *);
-size_t _MSKanji_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
+size_t _MSKanji_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
-size_t _MSKanji_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+int _MSKanji_mbsinit(const mbstate_t *);
+size_t _MSKanji_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+
+typedef struct {
+ int count;
+ u_char bytes[2];
+} _MSKanjiState;
int
_MSKanji_init(_RuneLocale *rl)
@@ -58,21 +66,43 @@ _MSKanji_init(_RuneLocale *rl)
__mbrtowc = _MSKanji_mbrtowc;
__wcrtomb = _MSKanji_wcrtomb;
+ __mbsinit = _MSKanji_mbsinit;
_CurrentRuneLocale = rl;
__mb_cur_max = 2;
return (0);
}
+int
+_MSKanji_mbsinit(const mbstate_t *ps)
+{
+
+ return (ps == NULL || ((_MSKanjiState *)ps)->count == 0);
+}
+
size_t
_MSKanji_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
- mbstate_t * __restrict ps __unused)
+ mbstate_t * __restrict ps)
{
+ _MSKanjiState *ms;
wchar_t wc;
- int len;
+ int len, ocount;
+ size_t ncopy;
+
+ ms = (_MSKanjiState *)ps;
+
+ if (s == NULL) {
+ s = "";
+ n = 1;
+ pwc = NULL;
+ }
+
+ ncopy = MIN(MIN(n, MB_CUR_MAX), sizeof(ms->bytes) - ms->count);
+ memcpy(ms->bytes + ms->count, s, ncopy);
+ ocount = ms->count;
+ ms->count += ncopy;
+ s = (char *)ms->bytes;
+ n = ms->count;
- if (s == NULL)
- /* Reset to initial shift state (no-op) */
- return (0);
if (n == 0)
/* Incomplete multibyte sequence */
return ((size_t)-2);
@@ -87,7 +117,8 @@ _MSKanji_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
}
if (pwc != NULL)
*pwc = wc;
- return (wc == L'\0' ? 0 : len);
+ ms->count = 0;
+ return (wc == L'\0' ? 0 : len - ocount);
}
size_t
@@ -99,7 +130,6 @@ _MSKanji_wcrtomb(char * __restrict s, wchar_t wc,
if (s == NULL)
/* Reset to initial shift state (no-op) */
return (1);
-
len = (wc > 0x100) ? 2 : 1;
for (i = len; i-- > 0; )
*s++ = wc >> (i << 3);
diff --git a/lib/libc/locale/none.c b/lib/libc/locale/none.c
index 3c4d8c7..a94d040 100644
--- a/lib/libc/locale/none.c
+++ b/lib/libc/locale/none.c
@@ -1,5 +1,5 @@
/*-
- * Copyright (c) 2002, 2003 Tim J. Robbins. All rights reserved.
+ * Copyright (c) 2002-2004 Tim J. Robbins. All rights reserved.
* Copyright (c) 1993
* The Regents of the University of California. All rights reserved.
*
@@ -51,11 +51,13 @@ __FBSDID("$FreeBSD$");
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
int _none_init(_RuneLocale *);
size_t _none_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
+int _none_mbsinit(const mbstate_t *);
size_t _none_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
int
@@ -63,12 +65,24 @@ _none_init(_RuneLocale *rl)
{
__mbrtowc = _none_mbrtowc;
+ __mbsinit = _none_mbsinit;
__wcrtomb = _none_wcrtomb;
_CurrentRuneLocale = rl;
__mb_cur_max = 1;
return(0);
}
+int
+_none_mbsinit(const mbstate_t *ps __unused)
+{
+
+ /*
+ * Encoding is not state dependent - we are always in the
+ * initial state.
+ */
+ return (1);
+}
+
size_t
_none_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
mbstate_t * __restrict ps __unused)
diff --git a/lib/libc/locale/setrunelocale.c b/lib/libc/locale/setrunelocale.c
index 11378a0..0d11893 100644
--- a/lib/libc/locale/setrunelocale.c
+++ b/lib/libc/locale/setrunelocale.c
@@ -64,11 +64,13 @@ extern _RuneLocale *_Read_RuneMagi(FILE *);
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
extern rune_t __emulated_sgetrune(const char *, size_t, const char **);
extern int __emulated_sputrune(rune_t, char *, size_t, char **);
extern size_t _none_mbrtowc(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int _none_mbsinit(const mbstate_t *);
extern size_t _none_wcrtomb(char * __restrict, wchar_t,
mbstate_t * __restrict);
@@ -108,6 +110,7 @@ __setrunelocale(const char *encoding)
const char * __restrict, size_t, mbstate_t * __restrict);
static size_t (*Cached__wcrtomb)(char * __restrict, wchar_t,
mbstate_t * __restrict);
+ static int (*Cached__mbsinit)(const mbstate_t *);
/*
* The "C" and "POSIX" locale are always here.
@@ -117,6 +120,7 @@ __setrunelocale(const char *encoding)
__mb_cur_max = 1;
__mbrtowc = _none_mbrtowc;
__wcrtomb = _none_wcrtomb;
+ __mbsinit = _none_mbsinit;
return (0);
}
@@ -129,6 +133,7 @@ __setrunelocale(const char *encoding)
__mb_cur_max = Cached__mb_cur_max;
__mbrtowc = Cached__mbrtowc;
__wcrtomb = Cached__wcrtomb;
+ __mbsinit = Cached__mbsinit;
return (0);
}
@@ -154,6 +159,7 @@ __setrunelocale(const char *encoding)
__mbrtowc = NULL;
__wcrtomb = NULL;
+ __mbsinit = NULL;
rl->sputrune = __emulated_sputrune;
rl->sgetrune = __emulated_sgetrune;
if (strcmp(rl->encoding, "NONE") == 0)
@@ -186,6 +192,7 @@ __setrunelocale(const char *encoding)
CachedRuneLocale = _CurrentRuneLocale;
Cached__mb_cur_max = __mb_cur_max;
Cached__mbrtowc = __mbrtowc;
+ Cached__mbsinit = __mbsinit;
Cached__wcrtomb = __wcrtomb;
(void)strcpy(ctype_encoding, encoding);
} else
diff --git a/lib/libc/locale/table.c b/lib/libc/locale/table.c
index f93fc1a..7680b9c 100644
--- a/lib/libc/locale/table.c
+++ b/lib/libc/locale/table.c
@@ -46,6 +46,7 @@ __FBSDID("$FreeBSD$");
extern size_t _none_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
+extern int _none_mbsinit(const mbstate_t *);
extern size_t _none_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
extern rune_t __emulated_sgetrune(const char *, size_t, const char **);
extern int __emulated_sputrune(rune_t, char *, size_t, char **);
@@ -259,5 +260,6 @@ _RuneLocale *_CurrentRuneLocale = &_DefaultRuneLocale;
int __mb_cur_max = 1;
size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict) = _none_mbrtowc;
+int (*__mbsinit)(const mbstate_t *) = _none_mbsinit;
size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict) =
_none_wcrtomb;
diff --git a/lib/libc/locale/utf2.c b/lib/libc/locale/utf2.c
index 39befd6..af7cc2f 100644
--- a/lib/libc/locale/utf2.c
+++ b/lib/libc/locale/utf2.c
@@ -36,23 +36,29 @@
/* UTF2 is obsolete and will be removed in FreeBSD 6 -- use UTF-8 instead. */
#define OBSOLETE_IN_6
-#include <sys/cdefs.h>
+#include <sys/param.h>
__FBSDID("$FreeBSD$");
#include <errno.h>
#include <runetype.h>
-#include <stddef.h>
-#include <stdio.h>
#include <stdlib.h>
+#include <string.h>
#include <wchar.h>
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
-size_t _UTF2_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
+size_t _UTF2_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
-size_t _UTF2_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+int _UTF2_mbsinit(const mbstate_t *);
+size_t _UTF2_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+
+typedef struct {
+ int count;
+ u_char bytes[3];
+} _UTF2State;
int
_UTF2_init(_RuneLocale *rl)
@@ -60,21 +66,44 @@ _UTF2_init(_RuneLocale *rl)
__mbrtowc = _UTF2_mbrtowc;
__wcrtomb = _UTF2_wcrtomb;
+ __mbsinit = _UTF2_mbsinit;
_CurrentRuneLocale = rl;
__mb_cur_max = 3;
return (0);
}
+int
+_UTF2_mbsinit(const mbstate_t *ps)
+{
+
+ return (ps == NULL || ((_UTF2State *)ps)->count == 0);
+}
+
size_t
_UTF2_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
- mbstate_t * __restrict ps __unused)
+ mbstate_t * __restrict ps)
{
- int ch, i, len, mask;
+ _UTF2State *us;
+ int ch, i, len, mask, ocount;
wchar_t wch;
+ size_t ncopy;
+
+ us = (_UTF2State *)ps;
+
+ if (s == NULL) {
+ s = "";
+ n = 1;
+ pwc = NULL;
+ }
+
+ ncopy = MIN(MIN(n, MB_CUR_MAX), sizeof(us->bytes) - us->count);
+ memcpy(us->bytes + us->count, s, ncopy);
+ ocount = us->count;
+ us->count += ncopy;
+ s = (char *)us->bytes;
+ n = us->count;
- if (s == NULL)
- return (0);
if (n == 0)
return ((size_t)-2);
@@ -108,7 +137,8 @@ _UTF2_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
}
if (pwc != NULL)
*pwc = wch;
- return (wch == L'\0' ? 0 : len);
+ us->count = 0;
+ return (wch == L'\0' ? 0 : len - ocount);
}
size_t
@@ -119,6 +149,7 @@ _UTF2_wcrtomb(char * __restrict s, wchar_t wc,
int i, len;
if (s == NULL)
+ /* Reset to initial conversion state. */
return (1);
if ((wc & ~0x7f) == 0) {
diff --git a/lib/libc/locale/utf8.c b/lib/libc/locale/utf8.c
index e1cbdea..113cdaf 100644
--- a/lib/libc/locale/utf8.c
+++ b/lib/libc/locale/utf8.c
@@ -1,5 +1,5 @@
/*-
- * Copyright (c) 2002, 2003 Tim J. Robbins
+ * Copyright (c) 2002-2004 Tim J. Robbins
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
@@ -24,23 +24,29 @@
* SUCH DAMAGE.
*/
-#include <sys/cdefs.h>
+#include <sys/param.h>
__FBSDID("$FreeBSD$");
#include <errno.h>
#include <runetype.h>
-#include <stddef.h>
-#include <stdio.h>
#include <stdlib.h>
+#include <string.h>
#include <wchar.h>
extern size_t (*__mbrtowc)(wchar_t * __restrict, const char * __restrict,
size_t, mbstate_t * __restrict);
+extern int (*__mbsinit)(const mbstate_t *);
extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
-size_t _UTF8_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
+size_t _UTF8_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
mbstate_t * __restrict);
-size_t _UTF8_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+int _UTF8_mbsinit(const mbstate_t *);
+size_t _UTF8_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
+
+typedef struct {
+ int count;
+ u_char bytes[6];
+} _UTF8State;
int
_UTF8_init(_RuneLocale *rl)
@@ -48,22 +54,44 @@ _UTF8_init(_RuneLocale *rl)
__mbrtowc = _UTF8_mbrtowc;
__wcrtomb = _UTF8_wcrtomb;
+ __mbsinit = _UTF8_mbsinit;
_CurrentRuneLocale = rl;
__mb_cur_max = 6;
return (0);
}
+int
+_UTF8_mbsinit(const mbstate_t *ps)
+{
+
+ return (ps == NULL || ((_UTF8State *)ps)->count == 0);
+}
+
size_t
_UTF8_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
- mbstate_t * __restrict ps __unused)
+ mbstate_t * __restrict ps)
{
- int ch, i, len, mask;
+ _UTF8State *us;
+ int ch, i, len, mask, ocount;
wchar_t lbound, wch;
+ size_t ncopy;
+
+ us = (_UTF8State *)ps;
+
+ if (s == NULL) {
+ s = "";
+ n = 1;
+ pwc = NULL;
+ }
+
+ ncopy = MIN(MIN(n, MB_CUR_MAX), sizeof(us->bytes) - us->count);
+ memcpy(us->bytes + us->count, s, ncopy);
+ ocount = us->count;
+ us->count += ncopy;
+ s = (char *)us->bytes;
+ n = us->count;
- if (s == NULL)
- /* Reset to initial shift state (no-op) */
- return (0);
if (n == 0)
/* Incomplete multibyte sequence */
return ((size_t)-2);
@@ -143,7 +171,8 @@ _UTF8_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
}
if (pwc != NULL)
*pwc = wch;
- return (wch == L'\0' ? 0 : len);
+ us->count = 0;
+ return (wch == L'\0' ? 0 : len - ocount);
}
size_t
OpenPOWER on IntegriCloud