xref: /freebsd/lib/libc/locale/mskanji.c (revision dc36d6f9bb1753f3808552f3afd30eda9a7b206a)
1d915a14eSPedro F. Giffuni /*-
2d915a14eSPedro F. Giffuni  * SPDX-License-Identifier: BSD-4-Clause
3d915a14eSPedro F. Giffuni  *
47b247341SBaptiste Daroussin  * Copyright 2013 Garrett D'Amore <garrett@damore.org>
57b247341SBaptiste Daroussin  * Copyright 2010 Nexenta Systems, Inc.  All rights reserved.
6ca2dae42STim J. Robbins  * Copyright (c) 2002-2004 Tim J. Robbins. All rights reserved.
7ca2dae42STim J. Robbins  *
816f76e6fSJulian Elischer  *    ja_JP.SJIS locale table for BSD4.4/rune
916f76e6fSJulian Elischer  *    version 1.0
1016f76e6fSJulian Elischer  *    (C) Sin'ichiro MIYATANI / Phase One, Inc
1116f76e6fSJulian Elischer  *    May 12, 1995
1216f76e6fSJulian Elischer  *
133c87aa1dSDavid Chisnall  * Copyright (c) 2011 The FreeBSD Foundation
14*5b5fa75aSEd Maste  *
153c87aa1dSDavid Chisnall  * Portions of this software were developed by David Chisnall
163c87aa1dSDavid Chisnall  * under sponsorship from the FreeBSD Foundation.
173c87aa1dSDavid Chisnall  *
1816f76e6fSJulian Elischer  * Redistribution and use in source and binary forms, with or without
1916f76e6fSJulian Elischer  * modification, are permitted provided that the following conditions
2016f76e6fSJulian Elischer  * are met:
2116f76e6fSJulian Elischer  * 1. Redistributions of source code must retain the above copyright
2216f76e6fSJulian Elischer  *    notice, this list of conditions and the following disclaimer.
2316f76e6fSJulian Elischer  * 2. Redistributions in binary form must reproduce the above copyright
2416f76e6fSJulian Elischer  *    notice, this list of conditions and the following disclaimer in the
2516f76e6fSJulian Elischer  *    documentation and/or other materials provided with the distribution.
2616f76e6fSJulian Elischer  * 3. All advertising materials mentioning features or use of this software
2716f76e6fSJulian Elischer  *    must display the following acknowledgement:
2816f76e6fSJulian Elischer  *      This product includes software developed by Phase One, Inc.
2916f76e6fSJulian Elischer  * 4. The name of Phase One, Inc. may be used to endorse or promote products
3016f76e6fSJulian Elischer  *    derived from this software without specific prior written permission.
3116f76e6fSJulian Elischer  *
3216f76e6fSJulian Elischer  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
3316f76e6fSJulian Elischer  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
3416f76e6fSJulian Elischer  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
3516f76e6fSJulian Elischer  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
3616f76e6fSJulian Elischer  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
3716f76e6fSJulian Elischer  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
3816f76e6fSJulian Elischer  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
3916f76e6fSJulian Elischer  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
4016f76e6fSJulian Elischer  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
4116f76e6fSJulian Elischer  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
4216f76e6fSJulian Elischer  * SUCH DAMAGE.
4316f76e6fSJulian Elischer  */
4416f76e6fSJulian Elischer 
45e94c6cb4SAlexey Zelkin #include <sys/types.h>
46fc813796STim J. Robbins #include <errno.h>
4702f4f60aSTim J. Robbins #include <runetype.h>
4816f76e6fSJulian Elischer #include <stdlib.h>
49ca2dae42STim J. Robbins #include <string.h>
5002f4f60aSTim J. Robbins #include <wchar.h>
512051a8f2STim J. Robbins #include "mblocal.h"
5202f4f60aSTim J. Robbins 
53367ed4e1SAndrey A. Chernov extern int __mb_sb_limit;
54367ed4e1SAndrey A. Chernov 
55e94c6cb4SAlexey Zelkin static size_t	_MSKanji_mbrtowc(wchar_t * __restrict, const char * __restrict,
56e94c6cb4SAlexey Zelkin 		    size_t, mbstate_t * __restrict);
57e94c6cb4SAlexey Zelkin static int	_MSKanji_mbsinit(const mbstate_t *);
58e94c6cb4SAlexey Zelkin static size_t	_MSKanji_wcrtomb(char * __restrict, wchar_t,
5902f4f60aSTim J. Robbins 		    mbstate_t * __restrict);
607b247341SBaptiste Daroussin static size_t	_MSKanji_mbsnrtowcs(wchar_t * __restrict,
617b247341SBaptiste Daroussin 		    const char ** __restrict, size_t, size_t,
627b247341SBaptiste Daroussin 		    mbstate_t * __restrict);
637b247341SBaptiste Daroussin static size_t	_MSKanji_wcsnrtombs(char * __restrict,
647b247341SBaptiste Daroussin 		    const wchar_t ** __restrict, size_t, size_t,
657b247341SBaptiste Daroussin 		    mbstate_t * __restrict);
6616f76e6fSJulian Elischer 
67ca2dae42STim J. Robbins typedef struct {
68b666b593STim J. Robbins 	wchar_t	ch;
69ca2dae42STim J. Robbins } _MSKanjiState;
70ca2dae42STim J. Robbins 
7116f76e6fSJulian Elischer int
_MSKanji_init(struct xlocale_ctype * l,_RuneLocale * rl)723c87aa1dSDavid Chisnall _MSKanji_init(struct xlocale_ctype *l, _RuneLocale *rl)
7316f76e6fSJulian Elischer {
7416f76e6fSJulian Elischer 
753c87aa1dSDavid Chisnall 	l->__mbrtowc = _MSKanji_mbrtowc;
763c87aa1dSDavid Chisnall 	l->__wcrtomb = _MSKanji_wcrtomb;
777b247341SBaptiste Daroussin 	l->__mbsnrtowcs = _MSKanji_mbsnrtowcs;
787b247341SBaptiste Daroussin 	l->__wcsnrtombs = _MSKanji_wcsnrtombs;
793c87aa1dSDavid Chisnall 	l->__mbsinit = _MSKanji_mbsinit;
803c87aa1dSDavid Chisnall 	l->runes = rl;
813c87aa1dSDavid Chisnall 	l->__mb_cur_max = 2;
82ae7abb26SAndrey A. Chernov 	l->__mb_sb_limit = 224;
8316f76e6fSJulian Elischer 	return (0);
8416f76e6fSJulian Elischer }
8516f76e6fSJulian Elischer 
86e94c6cb4SAlexey Zelkin static int
_MSKanji_mbsinit(const mbstate_t * ps)87ca2dae42STim J. Robbins _MSKanji_mbsinit(const mbstate_t *ps)
88ca2dae42STim J. Robbins {
89ca2dae42STim J. Robbins 
90b666b593STim J. Robbins 	return (ps == NULL || ((const _MSKanjiState *)ps)->ch == 0);
91ca2dae42STim J. Robbins }
92ca2dae42STim J. Robbins 
93e94c6cb4SAlexey Zelkin static size_t
_MSKanji_mbrtowc(wchar_t * __restrict pwc,const char * __restrict s,size_t n,mbstate_t * __restrict ps)9402f4f60aSTim J. Robbins _MSKanji_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
95ca2dae42STim J. Robbins     mbstate_t * __restrict ps)
9616f76e6fSJulian Elischer {
97ca2dae42STim J. Robbins 	_MSKanjiState *ms;
9802f4f60aSTim J. Robbins 	wchar_t wc;
9916f76e6fSJulian Elischer 
100ca2dae42STim J. Robbins 	ms = (_MSKanjiState *)ps;
101ca2dae42STim J. Robbins 
102b666b593STim J. Robbins 	if ((ms->ch & ~0xFF) != 0) {
103b666b593STim J. Robbins 		/* Bad conversion state. */
104fc813796STim J. Robbins 		errno = EINVAL;
105fc813796STim J. Robbins 		return ((size_t)-1);
106fc813796STim J. Robbins 	}
107fc813796STim J. Robbins 
108ca2dae42STim J. Robbins 	if (s == NULL) {
109ca2dae42STim J. Robbins 		s = "";
110ca2dae42STim J. Robbins 		n = 1;
111ca2dae42STim J. Robbins 		pwc = NULL;
112ca2dae42STim J. Robbins 	}
113ca2dae42STim J. Robbins 
11402f4f60aSTim J. Robbins 	if (n == 0)
11502f4f60aSTim J. Robbins 		/* Incomplete multibyte sequence */
11602f4f60aSTim J. Robbins 		return ((size_t)-2);
117b666b593STim J. Robbins 
118b666b593STim J. Robbins 	if (ms->ch != 0) {
119b666b593STim J. Robbins 		if (*s == '\0') {
120b666b593STim J. Robbins 			errno = EILSEQ;
121b666b593STim J. Robbins 			return ((size_t)-1);
122b666b593STim J. Robbins 		}
123b666b593STim J. Robbins 		wc = (ms->ch << 8) | (*s & 0xFF);
124b666b593STim J. Robbins 		if (pwc != NULL)
125b666b593STim J. Robbins 			*pwc = wc;
126b666b593STim J. Robbins 		ms->ch = 0;
127b666b593STim J. Robbins 		return (1);
128b666b593STim J. Robbins 	}
12902f4f60aSTim J. Robbins 	wc = *s++ & 0xff;
13002f4f60aSTim J. Robbins 	if ((wc > 0x80 && wc < 0xa0) || (wc >= 0xe0 && wc < 0xfd)) {
131b666b593STim J. Robbins 		if (n < 2) {
13202f4f60aSTim J. Robbins 			/* Incomplete multibyte sequence */
133b666b593STim J. Robbins 			ms->ch = wc;
13402f4f60aSTim J. Robbins 			return ((size_t)-2);
135b666b593STim J. Robbins 		}
13688af941aSTim J. Robbins 		if (*s == '\0') {
13788af941aSTim J. Robbins 			errno = EILSEQ;
13888af941aSTim J. Robbins 			return ((size_t)-1);
13988af941aSTim J. Robbins 		}
14002f4f60aSTim J. Robbins 		wc = (wc << 8) | (*s++ & 0xff);
14102f4f60aSTim J. Robbins 		if (pwc != NULL)
14202f4f60aSTim J. Robbins 			*pwc = wc;
143b666b593STim J. Robbins 		return (2);
144b666b593STim J. Robbins 	} else {
145b666b593STim J. Robbins 		if (pwc != NULL)
146b666b593STim J. Robbins 			*pwc = wc;
147b666b593STim J. Robbins 		return (wc == L'\0' ? 0 : 1);
148b666b593STim J. Robbins 	}
149d891f268STim J. Robbins }
150d891f268STim J. Robbins 
151e94c6cb4SAlexey Zelkin static size_t
_MSKanji_wcrtomb(char * __restrict s,wchar_t wc,mbstate_t * __restrict ps)152fc813796STim J. Robbins _MSKanji_wcrtomb(char * __restrict s, wchar_t wc, mbstate_t * __restrict ps)
15316f76e6fSJulian Elischer {
154fc813796STim J. Robbins 	_MSKanjiState *ms;
15516f76e6fSJulian Elischer 	int len, i;
15616f76e6fSJulian Elischer 
157fc813796STim J. Robbins 	ms = (_MSKanjiState *)ps;
158fc813796STim J. Robbins 
159b666b593STim J. Robbins 	if (ms->ch != 0) {
160fc813796STim J. Robbins 		errno = EINVAL;
161fc813796STim J. Robbins 		return ((size_t)-1);
162fc813796STim J. Robbins 	}
163fc813796STim J. Robbins 
16402f4f60aSTim J. Robbins 	if (s == NULL)
16502f4f60aSTim J. Robbins 		/* Reset to initial shift state (no-op) */
16602f4f60aSTim J. Robbins 		return (1);
16702f4f60aSTim J. Robbins 	len = (wc > 0x100) ? 2 : 1;
16802f4f60aSTim J. Robbins 	for (i = len; i-- > 0; )
16902f4f60aSTim J. Robbins 		*s++ = wc >> (i << 3);
170d891f268STim J. Robbins 	return (len);
17116f76e6fSJulian Elischer }
1727b247341SBaptiste Daroussin 
1737b247341SBaptiste Daroussin static size_t
_MSKanji_mbsnrtowcs(wchar_t * __restrict dst,const char ** __restrict src,size_t nms,size_t len,mbstate_t * __restrict ps)1747b247341SBaptiste Daroussin _MSKanji_mbsnrtowcs(wchar_t * __restrict dst,
1757b247341SBaptiste Daroussin     const char ** __restrict src, size_t nms,
1767b247341SBaptiste Daroussin     size_t len, mbstate_t * __restrict ps)
1777b247341SBaptiste Daroussin {
1787b247341SBaptiste Daroussin 	return (__mbsnrtowcs_std(dst, src, nms, len, ps, _MSKanji_mbrtowc));
1797b247341SBaptiste Daroussin }
1807b247341SBaptiste Daroussin 
1817b247341SBaptiste Daroussin static size_t
_MSKanji_wcsnrtombs(char * __restrict dst,const wchar_t ** __restrict src,size_t nwc,size_t len,mbstate_t * __restrict ps)1827b247341SBaptiste Daroussin _MSKanji_wcsnrtombs(char * __restrict dst,
1837b247341SBaptiste Daroussin     const wchar_t ** __restrict src, size_t nwc,
1847b247341SBaptiste Daroussin     size_t len, mbstate_t * __restrict ps)
1857b247341SBaptiste Daroussin {
1867b247341SBaptiste Daroussin 	return (__wcsnrtombs_std(dst, src, nwc, len, ps, _MSKanji_wcrtomb));
1877b247341SBaptiste Daroussin }
188