Home | History | Annotate | Line # | Download | only in modules
citrus_euc.c revision 1.4 
      1 /*	$NetBSD: citrus_euc.c,v 1.4 2002/03/28 10:29:11 yamt Exp $	*/
      2 
      3 /*-
      4  * Copyright (c)2002 Citrus Project,
      5  * All rights reserved.
      6  *
      7  * Redistribution and use in source and binary forms, with or without
      8  * modification, are permitted provided that the following conditions
      9  * are met:
     10  * 1. Redistributions of source code must retain the above copyright
     11  *    notice, this list of conditions and the following disclaimer.
     12  * 2. Redistributions in binary form must reproduce the above copyright
     13  *    notice, this list of conditions and the following disclaimer in the
     14  *    documentation and/or other materials provided with the distribution.
     15  *
     16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
     17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
     20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     26  * SUCH DAMAGE.
     27  */
     28 
     29 /*-
     30  * Copyright (c) 1993
     31  *	The Regents of the University of California.  All rights reserved.
     32  *
     33  * This code is derived from software contributed to Berkeley by
     34  * Paul Borman at Krystal Technologies.
     35  *
     36  * Redistribution and use in source and binary forms, with or without
     37  * modification, are permitted provided that the following conditions
     38  * are met:
     39  * 1. Redistributions of source code must retain the above copyright
     40  *    notice, this list of conditions and the following disclaimer.
     41  * 2. Redistributions in binary form must reproduce the above copyright
     42  *    notice, this list of conditions and the following disclaimer in the
     43  *    documentation and/or other materials provided with the distribution.
     44  * 3. All advertising materials mentioning features or use of this software
     45  *    must display the following acknowledgement:
     46  *	This product includes software developed by the University of
     47  *	California, Berkeley and its contributors.
     48  * 4. Neither the name of the University nor the names of its contributors
     49  *    may be used to endorse or promote products derived from this software
     50  *    without specific prior written permission.
     51  *
     52  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     53  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     54  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     55  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     56  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     57  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     58  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     59  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     60  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     61  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     62  * SUCH DAMAGE.
     63  */
     64 
     65 #include <sys/cdefs.h>
     66 #if defined(LIBC_SCCS) && !defined(lint)
     67 __RCSID("$NetBSD: citrus_euc.c,v 1.4 2002/03/28 10:29:11 yamt Exp $");
     68 #endif /* LIBC_SCCS and not lint */
     69 
     70 #include <assert.h>
     71 #include <errno.h>
     72 #include <string.h>
     73 #include <stdio.h>
     74 #include <stdlib.h>
     75 #include <stddef.h>
     76 #include <locale.h>
     77 #include <wchar.h>
     78 #include <sys/types.h>
     79 #include <limits.h>
     80 #include "citrus_module.h"
     81 #include "citrus_ctype.h"
     82 #include "citrus_euc.h"
     83 
     84 
     85 /* ----------------------------------------------------------------------
     86  * private stuffs used by templates
     87  */
     88 
     89 typedef struct {
     90 	char ch[3];
     91 	int chlen;
     92 } _EUCState;
     93 
     94 typedef struct {
     95 	unsigned	count[4];
     96 	wchar_t		bits[4];
     97 	wchar_t		mask;
     98 	unsigned	mb_cur_max;
     99 } _EUCEncodingInfo;
    100 
    101 typedef struct {
    102 	_EUCEncodingInfo	ei;
    103 	struct {
    104 		/* for future multi-locale facility */
    105 		_EUCState	s_mblen;
    106 		_EUCState	s_mbrlen;
    107 		_EUCState	s_mbrtowc;
    108 		_EUCState	s_mbtowc;
    109 		_EUCState	s_mbsrtowcs;
    110 		_EUCState	s_wcrtomb;
    111 		_EUCState	s_wcsrtombs;
    112 		_EUCState	s_wctomb;
    113 	} states;
    114 } _EUCCTypeInfo;
    115 
    116 #define	_SS2	0x008e
    117 #define	_SS3	0x008f
    118 
    119 #define _TO_EI(_cl_)			((_EUCEncodingInfo *)(_cl_))
    120 #define	_TO_CEI(_cl_)			((_EUCCTypeInfo *)(_cl_))
    121 #define _TO_STATE(_ps_)			((_EUCState *)(_ps_))
    122 #define _CEI_TO_EI(_cei_)		(&(_cei_)->ei)
    123 #define _CEI_TO_STATE(_cei_, _func_)	(_cei_)->states.s_##_func_
    124 
    125 #define _FUNCNAME(m)			_citrus_EUC_##m
    126 #define _ENCODING_INFO			_EUCEncodingInfo
    127 #define _CTYPE_INFO			_EUCCTypeInfo
    128 #define _ENCODING_STATE			_EUCState
    129 #define _ENCODING_MB_CUR_MAX(_ei_)	(_ei_)->mb_cur_max
    130 #define _ENCODING_IS_STATE_DEPENDENT	0
    131 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_)	0
    132 
    133 
    134 static __inline int
    135 _citrus_EUC_cs(unsigned int c)
    136 {
    137 	c &= 0xff;
    138 
    139 	return ((c & 0x80) ? c == _SS3 ? 3 : c == _SS2 ? 2 : 1 : 0);
    140 }
    141 
    142 static __inline int
    143 _citrus_EUC_parse_variable(_EUCEncodingInfo *ei,
    144 			   const void *var, size_t lenvar)
    145 {
    146 	const char *v, *e;
    147 	int x;
    148 
    149 	/* parse variable string */
    150 	if (!var)
    151 		return (EFTYPE);
    152 
    153 	v = (const char *) var;
    154 
    155 	while (*v == ' ' || *v == '\t')
    156 		++v;
    157 
    158 	ei->mb_cur_max = 1;
    159 	for (x = 0; x < 4; ++x) {
    160 		ei->count[x] = (int) strtol(v, (char **)&e, 0);
    161 		if (v == e || !(v = e) || ei->count[x]<1 || ei->count[x]>4) {
    162 			return (EFTYPE);
    163 		}
    164 		if (ei->mb_cur_max < ei->count[x])
    165 			ei->mb_cur_max = ei->count[x];
    166 		while (*v == ' ' || *v == '\t')
    167 			++v;
    168 		ei->bits[x] = (int) strtol(v, (char **)&e, 0);
    169 		if (v == e || !(v = e)) {
    170 			return (EFTYPE);
    171 		}
    172 		while (*v == ' ' || *v == '\t')
    173 			++v;
    174 	}
    175 	ei->mask = (int)strtol(v, (char **)&e, 0);
    176 	if (v == e || !(v = e)) {
    177 		return (EFTYPE);
    178 	}
    179 
    180 	return 0;
    181 }
    182 
    183 
    184 static __inline void
    185 /*ARGSUSED*/
    186 _citrus_EUC_init_state(_EUCEncodingInfo *ei, _EUCState *s)
    187 {
    188 	memset(s, 0, sizeof(*s));
    189 }
    190 
    191 static __inline void
    192 /*ARGSUSED*/
    193 _citrus_EUC_pack_state(_EUCEncodingInfo *ei, void *pspriv, const _EUCState *s)
    194 {
    195 	memcpy(pspriv, (const void *)s, sizeof(*s));
    196 }
    197 
    198 static __inline void
    199 /*ARGSUSED*/
    200 _citrus_EUC_unpack_state(_EUCEncodingInfo *ei, _EUCState *s,
    201 			 const void *pspriv)
    202 {
    203 	memcpy((void *)s, pspriv, sizeof(*s));
    204 }
    205 
    206 static int
    207 _citrus_EUC_mbrtowc_priv(_EUCEncodingInfo *ei, wchar_t *pwc, const char **s,
    208 			 size_t n, _EUCState *psenc, size_t *nresult)
    209 {
    210 	wchar_t wchar;
    211 	int c, cs, len;
    212 	int chlenbak;
    213 	const char *s0, *s1 = NULL;
    214 
    215 	_DIAGASSERT(nresult != 0);
    216 	_DIAGASSERT(ei != NULL);
    217 	_DIAGASSERT(psenc != NULL);
    218 	_DIAGASSERT(s != NULL);
    219 
    220 	s0 = *s;
    221 
    222 	if (s0 == NULL) {
    223 		_citrus_EUC_init_state(ei, psenc);
    224 		*nresult = 0; /* state independent */
    225 		return (0);
    226 	}
    227 
    228 	chlenbak = psenc->chlen;
    229 
    230 	/* make sure we have the first byte in the buffer */
    231 	switch (psenc->chlen) {
    232 	case 0:
    233 		if (n < 1)
    234 			goto restart;
    235 		psenc->ch[0] = *s0++;
    236 		psenc->chlen = 1;
    237 		n--;
    238 		break;
    239 	case 1:
    240 	case 2:
    241 		break;
    242 	default:
    243 		/* illgeal state */
    244 		goto encoding_error;
    245 	}
    246 
    247 	c = ei->count[cs = _citrus_EUC_cs(psenc->ch[0] & 0xff)];
    248 	if (c == 0)
    249 		goto encoding_error;
    250 	while (psenc->chlen < c) {
    251 		if (n < 1)
    252 			goto restart;
    253 		psenc->ch[psenc->chlen] = *s0++;
    254 		psenc->chlen++;
    255 		n--;
    256 	}
    257 	*s = s0;
    258 
    259 	switch (cs) {
    260 	case 3:
    261 	case 2:
    262 		/* skip SS2/SS3 */
    263 		len = c - 1;
    264 		s1 = &psenc->ch[1];
    265 		break;
    266 	case 1:
    267 	case 0:
    268 		len = c;
    269 		s1 = &psenc->ch[0];
    270 		break;
    271 	}
    272 	wchar = 0;
    273 	while (len-- > 0)
    274 		wchar = (wchar << 8) | (*s1++ & 0xff);
    275 	wchar = (wchar & ~ei->mask) | ei->bits[cs];
    276 
    277 	psenc->chlen = 0;
    278 	if (pwc)
    279 		*pwc = wchar;
    280 
    281 	if (!wchar) {
    282 		*nresult = 0;
    283 	} else {
    284 		*nresult = (size_t)(c - chlenbak);
    285 	}
    286 
    287 	return 0;
    288 
    289 encoding_error:
    290 	psenc->chlen = 0;
    291 	*nresult = (size_t)-1;
    292 	return (EILSEQ);
    293 
    294 restart:
    295 	*nresult = (size_t)-2;
    296 	*s = s0;
    297 	return (0);
    298 }
    299 
    300 static int
    301 _citrus_EUC_wcrtomb_priv(_EUCEncodingInfo *ei, char *s, size_t n, wchar_t wc,
    302 			 _EUCState *psenc, size_t *nresult)
    303 {
    304 	wchar_t m, nm;
    305 	int cs, i;
    306 
    307 	_DIAGASSERT(ei != NULL);
    308 	_DIAGASSERT(nresult != 0);
    309 	_DIAGASSERT(s != NULL);
    310 
    311 	m = wc & ei->mask;
    312 	nm = wc & ~m;
    313 
    314 	for (cs = 0;
    315 	     cs < sizeof(ei->count)/sizeof(ei->count[0]);
    316 	     cs++) {
    317 		if (m == ei->bits[cs])
    318 			break;
    319 	}
    320 	/* fallback case - not sure if it is necessary */
    321 	if (cs == sizeof(ei->count)/sizeof(ei->count[0]))
    322 		cs = 1;
    323 
    324 	i = ei->count[cs];
    325 	if (n < i)
    326 		goto ilseq;
    327 	m = (cs % 2) ? 0x80 : 0x00;
    328 	switch (cs) {
    329 	case 2:
    330 		*s++ = _SS2;
    331 		i--;
    332 		break;
    333 	case 3:
    334 		*s++ = _SS3;
    335 		i--;
    336 		break;
    337 	}
    338 
    339 	while (i-- > 0)
    340 		*s++ = ((nm >> (i << 3)) & 0xff) | m;
    341 
    342 	*nresult = (size_t)ei->count[cs];
    343 	return 0;
    344 
    345 ilseq:
    346 	*nresult = (size_t)-1;
    347 	return EILSEQ;	/*XXX*/
    348 }
    349 
    350 static int
    351 /*ARGSUSED*/
    352 _citrus_EUC_stdencoding_init(_EUCEncodingInfo * __restrict ei,
    353 			     const void * __restrict var, size_t lenvar)
    354 {
    355 
    356 	_DIAGASSERT(ei != NULL);
    357 
    358 	return (_citrus_EUC_parse_variable(ei, var, lenvar));
    359 }
    360 
    361 static void
    362 /*ARGSUSED*/
    363 _citrus_EUC_stdencoding_uninit(_EUCEncodingInfo * __restrict ei)
    364 {
    365 }
    366 
    367 /* ----------------------------------------------------------------------
    368  * public interface for ctype
    369  */
    370 
    371 _CITRUS_CTYPE_DECLS(EUC);
    372 _CITRUS_CTYPE_DEF_OPS(EUC);
    373 
    374 #include "citrus_ctype_template.h"
    375