xref: /illumos-gate/usr/src/lib/libc/port/locale/gbk.c (revision eda3ef2d)
14297a3b0SGarrett D'Amore /*
22d08521bSGarrett D'Amore  * Copyright 2013 Garrett D'Amore <garrett@damore.org>
36b5e5868SGarrett D'Amore  * Copyright 2010 Nexenta Systems, Inc.  All rights reserved.
44297a3b0SGarrett D'Amore  * Copyright (c) 2002-2004 Tim J. Robbins. All rights reserved.
54297a3b0SGarrett D'Amore  * Copyright (c) 1993
64297a3b0SGarrett D'Amore  *	The Regents of the University of California.  All rights reserved.
74297a3b0SGarrett D'Amore  *
84297a3b0SGarrett D'Amore  * This code is derived from software contributed to Berkeley by
94297a3b0SGarrett D'Amore  * Paul Borman at Krystal Technologies.
104297a3b0SGarrett D'Amore  *
114297a3b0SGarrett D'Amore  * Redistribution and use in source and binary forms, with or without
124297a3b0SGarrett D'Amore  * modification, are permitted provided that the following conditions
134297a3b0SGarrett D'Amore  * are met:
144297a3b0SGarrett D'Amore  * 1. Redistributions of source code must retain the above copyright
154297a3b0SGarrett D'Amore  *    notice, this list of conditions and the following disclaimer.
164297a3b0SGarrett D'Amore  * 2. Redistributions in binary form must reproduce the above copyright
174297a3b0SGarrett D'Amore  *    notice, this list of conditions and the following disclaimer in the
184297a3b0SGarrett D'Amore  *    documentation and/or other materials provided with the distribution.
194297a3b0SGarrett D'Amore  * 4. Neither the name of the University nor the names of its contributors
204297a3b0SGarrett D'Amore  *    may be used to endorse or promote products derived from this software
214297a3b0SGarrett D'Amore  *    without specific prior written permission.
224297a3b0SGarrett D'Amore  *
234297a3b0SGarrett D'Amore  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
244297a3b0SGarrett D'Amore  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
254297a3b0SGarrett D'Amore  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
264297a3b0SGarrett D'Amore  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
274297a3b0SGarrett D'Amore  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
284297a3b0SGarrett D'Amore  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
294297a3b0SGarrett D'Amore  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
304297a3b0SGarrett D'Amore  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
314297a3b0SGarrett D'Amore  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
324297a3b0SGarrett D'Amore  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
334297a3b0SGarrett D'Amore  * SUCH DAMAGE.
344297a3b0SGarrett D'Amore  */
354297a3b0SGarrett D'Amore 
364297a3b0SGarrett D'Amore #include "lint.h"
374297a3b0SGarrett D'Amore #include <sys/types.h>
384297a3b0SGarrett D'Amore #include <errno.h>
394297a3b0SGarrett D'Amore #include <stdlib.h>
404297a3b0SGarrett D'Amore #include <string.h>
414297a3b0SGarrett D'Amore #include <wchar.h>
424297a3b0SGarrett D'Amore #include "mblocal.h"
432d08521bSGarrett D'Amore #include "lctype.h"
444297a3b0SGarrett D'Amore 
454297a3b0SGarrett D'Amore static size_t	_GBK_mbrtowc(wchar_t *_RESTRICT_KYWD,
464297a3b0SGarrett D'Amore 		    const char *_RESTRICT_KYWD,
47*0ac311baSRobert Mustacchi 		    size_t, mbstate_t *_RESTRICT_KYWD, boolean_t);
484297a3b0SGarrett D'Amore static int	_GBK_mbsinit(const mbstate_t *);
494297a3b0SGarrett D'Amore static size_t	_GBK_wcrtomb(char *_RESTRICT_KYWD, wchar_t,
504297a3b0SGarrett D'Amore 		    mbstate_t *_RESTRICT_KYWD);
512d08521bSGarrett D'Amore static size_t	_GBK_mbsnrtowcs(wchar_t *_RESTRICT_KYWD,
522d08521bSGarrett D'Amore 		    const char **_RESTRICT_KYWD, size_t, size_t,
532d08521bSGarrett D'Amore 		    mbstate_t *_RESTRICT_KYWD);
542d08521bSGarrett D'Amore static size_t	_GBK_wcsnrtombs(char *_RESTRICT_KYWD,
552d08521bSGarrett D'Amore 		    const wchar_t **_RESTRICT_KYWD, size_t, size_t,
562d08521bSGarrett D'Amore 		    mbstate_t *_RESTRICT_KYWD);
574297a3b0SGarrett D'Amore 
582d08521bSGarrett D'Amore void
_GBK_init(struct lc_ctype * lct)592d08521bSGarrett D'Amore _GBK_init(struct lc_ctype *lct)
604297a3b0SGarrett D'Amore {
612d08521bSGarrett D'Amore 	lct->lc_mbrtowc = _GBK_mbrtowc;
622d08521bSGarrett D'Amore 	lct->lc_wcrtomb = _GBK_wcrtomb;
632d08521bSGarrett D'Amore 	lct->lc_mbsinit = _GBK_mbsinit;
642d08521bSGarrett D'Amore 	lct->lc_mbsnrtowcs = _GBK_mbsnrtowcs;
652d08521bSGarrett D'Amore 	lct->lc_wcsnrtombs = _GBK_wcsnrtombs;
662d08521bSGarrett D'Amore 	lct->lc_max_mblen = 2;
672d08521bSGarrett D'Amore 	lct->lc_is_ascii = 0;
684297a3b0SGarrett D'Amore }
694297a3b0SGarrett D'Amore 
704297a3b0SGarrett D'Amore static int
_GBK_mbsinit(const mbstate_t * ps)714297a3b0SGarrett D'Amore _GBK_mbsinit(const mbstate_t *ps)
724297a3b0SGarrett D'Amore {
734297a3b0SGarrett D'Amore 
744297a3b0SGarrett D'Amore 	return (ps == NULL || ((const _GBKState *)ps)->ch == 0);
754297a3b0SGarrett D'Amore }
764297a3b0SGarrett D'Amore 
774297a3b0SGarrett D'Amore static int
_gbk_check(uint_t c)784297a3b0SGarrett D'Amore _gbk_check(uint_t c)
794297a3b0SGarrett D'Amore {
804297a3b0SGarrett D'Amore 
814297a3b0SGarrett D'Amore 	c &= 0xff;
824297a3b0SGarrett D'Amore 	return ((c >= 0x81 && c <= 0xfe) ? 2 : 1);
834297a3b0SGarrett D'Amore }
844297a3b0SGarrett D'Amore 
854297a3b0SGarrett D'Amore static size_t
_GBK_mbrtowc(wchar_t * _RESTRICT_KYWD pwc,const char * _RESTRICT_KYWD s,size_t n,mbstate_t * _RESTRICT_KYWD ps,boolean_t zero)864297a3b0SGarrett D'Amore _GBK_mbrtowc(wchar_t *_RESTRICT_KYWD pwc, const char *_RESTRICT_KYWD s,
87*0ac311baSRobert Mustacchi     size_t n, mbstate_t *_RESTRICT_KYWD ps, boolean_t zero)
884297a3b0SGarrett D'Amore {
894297a3b0SGarrett D'Amore 	_GBKState *gs;
904297a3b0SGarrett D'Amore 	wchar_t wc;
914297a3b0SGarrett D'Amore 	size_t len;
924297a3b0SGarrett D'Amore 
934297a3b0SGarrett D'Amore 	gs = (_GBKState *)ps;
944297a3b0SGarrett D'Amore 
954297a3b0SGarrett D'Amore 	if ((gs->ch & ~0xFF) != 0) {
964297a3b0SGarrett D'Amore 		/* Bad conversion state. */
974297a3b0SGarrett D'Amore 		errno = EINVAL;
984297a3b0SGarrett D'Amore 		return ((size_t)-1);
994297a3b0SGarrett D'Amore 	}
1004297a3b0SGarrett D'Amore 
1014297a3b0SGarrett D'Amore 	if (s == NULL) {
1024297a3b0SGarrett D'Amore 		s = "";
1034297a3b0SGarrett D'Amore 		n = 1;
1044297a3b0SGarrett D'Amore 		pwc = NULL;
1054297a3b0SGarrett D'Amore 	}
1064297a3b0SGarrett D'Amore 
1074297a3b0SGarrett D'Amore 	if (n == 0)
1084297a3b0SGarrett D'Amore 		/* Incomplete multibyte sequence */
1094297a3b0SGarrett D'Amore 		return ((size_t)-2);
1104297a3b0SGarrett D'Amore 
1114297a3b0SGarrett D'Amore 	if (gs->ch != 0) {
1124297a3b0SGarrett D'Amore 		if (*s == '\0') {
1134297a3b0SGarrett D'Amore 			errno = EILSEQ;
1144297a3b0SGarrett D'Amore 			return ((size_t)-1);
1154297a3b0SGarrett D'Amore 		}
1164297a3b0SGarrett D'Amore 		wc = (gs->ch << 8) | (*s & 0xFF);
1174297a3b0SGarrett D'Amore 		if (pwc != NULL)
1184297a3b0SGarrett D'Amore 			*pwc = wc;
1194297a3b0SGarrett D'Amore 		gs->ch = 0;
1204297a3b0SGarrett D'Amore 		return (1);
1214297a3b0SGarrett D'Amore 	}
1224297a3b0SGarrett D'Amore 
1234297a3b0SGarrett D'Amore 	len = (size_t)_gbk_check(*s);
1244297a3b0SGarrett D'Amore 	wc = *s++ & 0xff;
1254297a3b0SGarrett D'Amore 	if (len == 2) {
1264297a3b0SGarrett D'Amore 		if (n < 2) {
1274297a3b0SGarrett D'Amore 			/* Incomplete multibyte sequence */
1284297a3b0SGarrett D'Amore 			gs->ch = wc;
1294297a3b0SGarrett D'Amore 			return ((size_t)-2);
1304297a3b0SGarrett D'Amore 		}
1314297a3b0SGarrett D'Amore 		if (*s == '\0') {
1324297a3b0SGarrett D'Amore 			errno = EILSEQ;
1334297a3b0SGarrett D'Amore 			return ((size_t)-1);
1344297a3b0SGarrett D'Amore 		}
1354297a3b0SGarrett D'Amore 		wc = (wc << 8) | (*s++ & 0xff);
1364297a3b0SGarrett D'Amore 		if (pwc != NULL)
1374297a3b0SGarrett D'Amore 			*pwc = wc;
1384297a3b0SGarrett D'Amore 		return (2);
1394297a3b0SGarrett D'Amore 	} else {
1404297a3b0SGarrett D'Amore 		if (pwc != NULL)
1414297a3b0SGarrett D'Amore 			*pwc = wc;
142*0ac311baSRobert Mustacchi 		if (zero || wc != L'\0') {
143*0ac311baSRobert Mustacchi 			return (1);
144*0ac311baSRobert Mustacchi 		} else {
145*0ac311baSRobert Mustacchi 			return (0);
146*0ac311baSRobert Mustacchi 		}
1474297a3b0SGarrett D'Amore 	}
1484297a3b0SGarrett D'Amore }
1494297a3b0SGarrett D'Amore 
1504297a3b0SGarrett D'Amore static size_t
_GBK_wcrtomb(char * _RESTRICT_KYWD s,wchar_t wc,mbstate_t * _RESTRICT_KYWD ps)1514297a3b0SGarrett D'Amore _GBK_wcrtomb(char *_RESTRICT_KYWD s, wchar_t wc, mbstate_t *_RESTRICT_KYWD ps)
1524297a3b0SGarrett D'Amore {
1534297a3b0SGarrett D'Amore 	_GBKState *gs;
1544297a3b0SGarrett D'Amore 
1554297a3b0SGarrett D'Amore 	gs = (_GBKState *)ps;
1564297a3b0SGarrett D'Amore 
1574297a3b0SGarrett D'Amore 	if (gs->ch != 0) {
1584297a3b0SGarrett D'Amore 		errno = EINVAL;
1594297a3b0SGarrett D'Amore 		return ((size_t)-1);
1604297a3b0SGarrett D'Amore 	}
1614297a3b0SGarrett D'Amore 
1624297a3b0SGarrett D'Amore 	if (s == NULL)
1634297a3b0SGarrett D'Amore 		/* Reset to initial shift state (no-op) */
1644297a3b0SGarrett D'Amore 		return (1);
1654297a3b0SGarrett D'Amore 	if (wc & 0x8000) {
1664297a3b0SGarrett D'Amore 		*s++ = (wc >> 8) & 0xff;
1674297a3b0SGarrett D'Amore 		*s = wc & 0xff;
1684297a3b0SGarrett D'Amore 		return (2);
1694297a3b0SGarrett D'Amore 	}
1704297a3b0SGarrett D'Amore 	*s = wc & 0xff;
1714297a3b0SGarrett D'Amore 	return (1);
1724297a3b0SGarrett D'Amore }
1732d08521bSGarrett D'Amore 
1742d08521bSGarrett D'Amore static size_t
_GBK_mbsnrtowcs(wchar_t * _RESTRICT_KYWD dst,const char ** _RESTRICT_KYWD src,size_t nms,size_t len,mbstate_t * _RESTRICT_KYWD ps)1752d08521bSGarrett D'Amore _GBK_mbsnrtowcs(wchar_t *_RESTRICT_KYWD dst, const char **_RESTRICT_KYWD src,
1762d08521bSGarrett D'Amore     size_t nms, size_t len, mbstate_t *_RESTRICT_KYWD ps)
1772d08521bSGarrett D'Amore {
1782d08521bSGarrett D'Amore 	return (__mbsnrtowcs_std(dst, src, nms, len, ps, _GBK_mbrtowc));
1792d08521bSGarrett D'Amore }
1802d08521bSGarrett D'Amore 
1812d08521bSGarrett D'Amore static size_t
_GBK_wcsnrtombs(char * _RESTRICT_KYWD dst,const wchar_t ** _RESTRICT_KYWD src,size_t nwc,size_t len,mbstate_t * _RESTRICT_KYWD ps)1822d08521bSGarrett D'Amore _GBK_wcsnrtombs(char *_RESTRICT_KYWD dst, const wchar_t **_RESTRICT_KYWD src,
1832d08521bSGarrett D'Amore     size_t nwc, size_t len, mbstate_t *_RESTRICT_KYWD ps)
1842d08521bSGarrett D'Amore {
1852d08521bSGarrett D'Amore 	return (__wcsnrtombs_std(dst, src, nwc, len, ps, _GBK_wcrtomb));
1862d08521bSGarrett D'Amore }
187