Home | History | Annotate | Line # | Download | only in modules
citrus_big5.c revision 1.9 
      1 /*	$NetBSD: citrus_big5.c,v 1.9 2005/10/29 18:02:04 tshiozak 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. Neither the name of the University nor the names of its contributors
     45  *    may be used to endorse or promote products derived from this software
     46  *    without specific prior written permission.
     47  *
     48  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     49  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     50  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     51  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     52  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     53  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     54  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     55  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     56  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     57  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     58  * SUCH DAMAGE.
     59  */
     60 
     61 #include <sys/cdefs.h>
     62 #if defined(LIBC_SCCS) && !defined(lint)
     63 __RCSID("$NetBSD: citrus_big5.c,v 1.9 2005/10/29 18:02:04 tshiozak Exp $");
     64 #endif /* LIBC_SCCS and not lint */
     65 
     66 #include <assert.h>
     67 #include <errno.h>
     68 #include <string.h>
     69 #include <stdio.h>
     70 #include <stdlib.h>
     71 #include <stddef.h>
     72 #include <locale.h>
     73 #include <wchar.h>
     74 #include <sys/types.h>
     75 #include <limits.h>
     76 
     77 #include "citrus_namespace.h"
     78 #include "citrus_types.h"
     79 #include "citrus_module.h"
     80 #include "citrus_ctype.h"
     81 #include "citrus_stdenc.h"
     82 #include "citrus_big5.h"
     83 
     84 /* ----------------------------------------------------------------------
     85  * private stuffs used by templates
     86  */
     87 
     88 typedef struct {
     89 	char ch[2];
     90 	int chlen;
     91 } _BIG5State;
     92 
     93 typedef struct {
     94 	int dummy;
     95 } _BIG5EncodingInfo;
     96 
     97 typedef struct {
     98 	_BIG5EncodingInfo	ei;
     99 	struct {
    100 		/* for future multi-locale facility */
    101 		_BIG5State	s_mblen;
    102 		_BIG5State	s_mbrlen;
    103 		_BIG5State	s_mbrtowc;
    104 		_BIG5State	s_mbtowc;
    105 		_BIG5State	s_mbsrtowcs;
    106 		_BIG5State	s_wcrtomb;
    107 		_BIG5State	s_wcsrtombs;
    108 		_BIG5State	s_wctomb;
    109 	} states;
    110 } _BIG5CTypeInfo;
    111 
    112 #define _CEI_TO_EI(_cei_)		(&(_cei_)->ei)
    113 #define _CEI_TO_STATE(_cei_, _func_)	(_cei_)->states.s_##_func_
    114 
    115 #define _FUNCNAME(m)			_citrus_BIG5_##m
    116 #define _ENCODING_INFO			_BIG5EncodingInfo
    117 #define _CTYPE_INFO			_BIG5CTypeInfo
    118 #define _ENCODING_STATE			_BIG5State
    119 #define _ENCODING_MB_CUR_MAX(_ei_)	2
    120 #define _ENCODING_IS_STATE_DEPENDENT	0
    121 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_)	0
    122 
    123 
    124 static __inline void
    125 /*ARGSUSED*/
    126 _citrus_BIG5_init_state(_BIG5EncodingInfo * __restrict ei,
    127 			_BIG5State * __restrict s)
    128 {
    129 	memset(s, 0, sizeof(*s));
    130 }
    131 
    132 static __inline void
    133 /*ARGSUSED*/
    134 _citrus_BIG5_pack_state(_BIG5EncodingInfo * __restrict ei,
    135 			void * __restrict pspriv,
    136 			const _BIG5State * __restrict s)
    137 {
    138 	memcpy(pspriv, (const void *)s, sizeof(*s));
    139 }
    140 
    141 static __inline void
    142 /*ARGSUSED*/
    143 _citrus_BIG5_unpack_state(_BIG5EncodingInfo * __restrict ei,
    144 			  _BIG5State * __restrict s,
    145 			  const void * __restrict pspriv)
    146 {
    147 	memcpy((void *)s, pspriv, sizeof(*s));
    148 }
    149 
    150 static __inline int
    151 _citrus_BIG5_check(u_int c)
    152 {
    153 	c &= 0xff;
    154 	return ((c >= 0xa1 && c <= 0xfe) ? 2 : 1);
    155 }
    156 
    157 static __inline int
    158 _citrus_BIG5_check2(u_int c)
    159 {
    160 	c &= 0xff;
    161 	if ((c >= 0x40 && c <= 0x7f) || (c >= 0xa1 && c <= 0xfe))
    162 		return 1;
    163 	else
    164 		return 0;
    165 }
    166 
    167 static int
    168 /*ARGSUSED*/
    169 _citrus_BIG5_encoding_module_init(_BIG5EncodingInfo * __restrict ei,
    170 				  const void * __restrict var, size_t lenvar)
    171 {
    172 	_DIAGASSERT(ei != NULL);
    173 
    174 	memset((void *)ei, 0, sizeof(*ei));
    175 
    176 	return (0);
    177 }
    178 
    179 static void
    180 /*ARGSUSED*/
    181 _citrus_BIG5_encoding_module_uninit(_BIG5EncodingInfo *ei)
    182 {
    183 }
    184 
    185 static int
    186 /*ARGSUSED*/
    187 _citrus_BIG5_mbrtowc_priv(_BIG5EncodingInfo * __restrict ei,
    188 			  wchar_t * __restrict pwc,
    189 			  const char ** __restrict s, size_t n,
    190 			  _BIG5State * __restrict psenc,
    191 			  size_t * __restrict nresult)
    192 {
    193 	wchar_t wchar;
    194 	int c;
    195 	int chlenbak;
    196 	const char *s0;
    197 
    198 	_DIAGASSERT(nresult != 0);
    199 	_DIAGASSERT(ei != NULL);
    200 	_DIAGASSERT(psenc != NULL);
    201 	_DIAGASSERT(s != NULL && *s != NULL);
    202 
    203 	s0 = *s;
    204 
    205 	if (s0 == NULL) {
    206 		_citrus_BIG5_init_state(ei, psenc);
    207 		*nresult = 0;
    208 		return (0);
    209 	}
    210 
    211 	chlenbak = psenc->chlen;
    212 
    213 	/* make sure we have the first byte in the buffer */
    214 	switch (psenc->chlen) {
    215 	case 0:
    216 		if (n < 1)
    217 			goto restart;
    218 		psenc->ch[0] = *s0++;
    219 		psenc->chlen = 1;
    220 		n--;
    221 		break;
    222 	case 1:
    223 		break;
    224 	default:
    225 		/* illegal state */
    226 		goto ilseq;
    227 	}
    228 
    229 	c = _citrus_BIG5_check(psenc->ch[0] & 0xff);
    230 	if (c == 0)
    231 		goto ilseq;
    232 	while (psenc->chlen < c) {
    233 		if (n < 1) {
    234 			goto restart;
    235 		}
    236 		psenc->ch[psenc->chlen] = *s0++;
    237 		psenc->chlen++;
    238 		n--;
    239 	}
    240 
    241 	switch (c) {
    242 	case 1:
    243 		wchar = psenc->ch[0] & 0xff;
    244 		break;
    245 	case 2:
    246 		if (!_citrus_BIG5_check2(psenc->ch[1] & 0xff))
    247 			goto ilseq;
    248 		wchar = ((psenc->ch[0] & 0xff) << 8) | (psenc->ch[1] & 0xff);
    249 		break;
    250 	default:
    251 		/* illegal state */
    252 		goto ilseq;
    253 	}
    254 
    255 	*s = s0;
    256 	psenc->chlen = 0;
    257 	if (pwc)
    258 		*pwc = wchar;
    259 	if (!wchar)
    260 		*nresult = 0;
    261 	else
    262 		*nresult = c - chlenbak;
    263 
    264 	return (0);
    265 
    266 ilseq:
    267 	psenc->chlen = 0;
    268 	*nresult = (size_t)-1;
    269 	return (EILSEQ);
    270 
    271 restart:
    272 	*s = s0;
    273 	*nresult = (size_t)-2;
    274 	return (0);
    275 }
    276 
    277 static int
    278 /*ARGSUSED*/
    279 _citrus_BIG5_wcrtomb_priv(_BIG5EncodingInfo * __restrict ei,
    280 			  char * __restrict s,
    281 			  size_t n, wchar_t wc, _BIG5State * __restrict psenc,
    282 			  size_t * __restrict nresult)
    283 {
    284 	int l, ret;
    285 
    286 	_DIAGASSERT(ei != NULL);
    287 	_DIAGASSERT(nresult != 0);
    288 	_DIAGASSERT(s != NULL);
    289 
    290 	/* check invalid sequence */
    291 	if (wc & ~0xffff) {
    292 		ret = EILSEQ;
    293 		goto err;
    294 	}
    295 
    296 	if (wc & 0x8000) {
    297 		if (_citrus_BIG5_check((wc >> 8) & 0xff) != 2 ||
    298 		    !_citrus_BIG5_check2(wc & 0xff)) {
    299 			ret = EILSEQ;
    300 			goto err;
    301 		}
    302 		l = 2;
    303 	} else {
    304 		if (wc & ~0xff || !_citrus_BIG5_check(wc & 0xff)) {
    305 			ret = EILSEQ;
    306 			goto err;
    307 		}
    308 		l = 1;
    309 	}
    310 
    311 	if (n < l) {
    312 		/* bound check failure */
    313 		ret = E2BIG;
    314 		goto err;
    315 	}
    316 
    317 	if (l == 2) {
    318 		s[0] = (wc >> 8) & 0xff;
    319 		s[1] = wc & 0xff;
    320 	} else
    321 		s[0] = wc & 0xff;
    322 
    323 	*nresult = l;
    324 
    325 	return 0;
    326 
    327 err:
    328 	*nresult = (size_t)-1;
    329 	return ret;
    330 }
    331 
    332 static __inline int
    333 /*ARGSUSED*/
    334 _citrus_BIG5_stdenc_wctocs(_BIG5EncodingInfo * __restrict ei,
    335 			   _csid_t * __restrict csid,
    336 			   _index_t * __restrict idx, wchar_t wc)
    337 {
    338 
    339 	_DIAGASSERT(csid != NULL && idx != NULL);
    340 
    341 	if (wc<0x100)
    342 		*csid = 0;
    343 	else
    344 		*csid = 1;
    345 	*idx = (_index_t)wc;
    346 
    347 	return 0;
    348 }
    349 
    350 static __inline int
    351 /*ARGSUSED*/
    352 _citrus_BIG5_stdenc_cstowc(_BIG5EncodingInfo * __restrict ei,
    353 			   wchar_t * __restrict wc,
    354 			   _csid_t csid, _index_t idx)
    355 {
    356 	u_int8_t h, l;
    357 
    358 	_DIAGASSERT(wc != NULL);
    359 
    360 	switch (csid) {
    361 	case 0:
    362 		if (idx>=0x80U)
    363 			return EILSEQ;
    364 		*wc = (wchar_t)idx;
    365 		break;
    366 	case 1:
    367 		if (idx>=0x10000U)
    368 			return EILSEQ;
    369 		h = idx >> 8;
    370 		l = idx;
    371 		if (h<0xA1 || h>0xF9 || l<0x40 || l>0xFE)
    372 			return EILSEQ;
    373 		*wc = (wchar_t)idx;
    374 		break;
    375 	default:
    376 		return EILSEQ;
    377 	}
    378 
    379 	return 0;
    380 }
    381 
    382 static __inline int
    383 /*ARGSUSED*/
    384 _citrus_BIG5_stdenc_get_state_desc_generic(_BIG5EncodingInfo * __restrict ei,
    385 					   _BIG5State * __restrict psenc,
    386 					   int * __restrict rstate)
    387 {
    388 
    389 	if (psenc->chlen == 0)
    390 		*rstate = _STDENC_SDGEN_INITIAL;
    391 	else
    392 		*rstate = _STDENC_SDGEN_INCOMPLETE_CHAR;
    393 
    394 	return 0;
    395 }
    396 
    397 /* ----------------------------------------------------------------------
    398  * public interface for ctype
    399  */
    400 
    401 _CITRUS_CTYPE_DECLS(BIG5);
    402 _CITRUS_CTYPE_DEF_OPS(BIG5);
    403 
    404 #include "citrus_ctype_template.h"
    405 
    406 
    407 /* ----------------------------------------------------------------------
    408  * public interface for stdenc
    409  */
    410 
    411 _CITRUS_STDENC_DECLS(BIG5);
    412 _CITRUS_STDENC_DEF_OPS(BIG5);
    413 
    414 #include "citrus_stdenc_template.h"
    415