Home | History | Annotate | Line # | Download | only in modules
citrus_utf8.c revision 1.9 
      1 /*	$NetBSD: citrus_utf8.c,v 1.9 2003/06/26 12:09:58 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. 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_utf8.c,v 1.9 2003/06/26 12:09:58 tshiozak 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 
     81 #include "citrus_namespace.h"
     82 #include "citrus_types.h"
     83 #include "citrus_module.h"
     84 #include "citrus_ctype.h"
     85 #include "citrus_stdenc.h"
     86 #include "citrus_utf8.h"
     87 
     88 
     89 /* ----------------------------------------------------------------------
     90  * private stuffs used by templates
     91  */
     92 
     93 static int _UTF8_count_array[256];
     94 static int const *_UTF8_count = NULL;
     95 
     96 static u_int32_t _UTF8_range[] = {
     97 	0,	/*dummy*/
     98 	0x00000000, 0x00000080, 0x00000800, 0x00010000,
     99 	0x00200000, 0x04000000, 0x80000000,
    100 };
    101 
    102 typedef struct {
    103 	char ch[6];
    104 	int chlen;
    105 } _UTF8State;
    106 
    107 typedef struct {
    108 } _UTF8EncodingInfo;
    109 
    110 typedef struct {
    111 	_UTF8EncodingInfo	ei;
    112 	struct {
    113 		/* for future multi-locale facility */
    114 		_UTF8State	s_mblen;
    115 		_UTF8State	s_mbrlen;
    116 		_UTF8State	s_mbrtowc;
    117 		_UTF8State	s_mbtowc;
    118 		_UTF8State	s_mbsrtowcs;
    119 		_UTF8State	s_wcrtomb;
    120 		_UTF8State	s_wcsrtombs;
    121 		_UTF8State	s_wctomb;
    122 	} states;
    123 } _UTF8CTypeInfo;
    124 
    125 #define _CEI_TO_EI(_cei_)		(&(_cei_)->ei)
    126 #define _CEI_TO_STATE(_ei_, _func_)	(_ei_)->states.s_##_func_
    127 
    128 #define _FUNCNAME(m)			_citrus_UTF8_##m
    129 #define _ENCODING_INFO			_UTF8EncodingInfo
    130 #define _CTYPE_INFO			_UTF8CTypeInfo
    131 #define _ENCODING_STATE			_UTF8State
    132 #define _ENCODING_MB_CUR_MAX(_ei_)	6
    133 #define _ENCODING_IS_STATE_DEPENDENT	0
    134 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_)	0
    135 
    136 
    137 static __inline void
    138 _UTF8_init_count(void)
    139 {
    140 	int i;
    141 	if (!_UTF8_count) {
    142 		memset(_UTF8_count_array, 0, sizeof(_UTF8_count_array));
    143 		for (i = 0; i <= 0x7f; i++)
    144 			_UTF8_count_array[i] = 1;
    145 		for (i = 0xc0; i <= 0xdf; i++)
    146 			_UTF8_count_array[i] = 2;
    147 		for (i = 0xe0; i <= 0xef; i++)
    148 			_UTF8_count_array[i] = 3;
    149 		for (i = 0xf0; i <= 0xf7; i++)
    150 			_UTF8_count_array[i] = 4;
    151 		for (i = 0xf8; i <= 0xfb; i++)
    152 			_UTF8_count_array[i] = 5;
    153 		for (i = 0xfc; i <= 0xfd; i++)
    154 			_UTF8_count_array[i] = 6;
    155 		_UTF8_count = _UTF8_count_array;
    156 	}
    157 }
    158 
    159 static int
    160 _UTF8_findlen(wchar_t v)
    161 {
    162 	int i;
    163 	u_int32_t c;
    164 
    165 	c = (u_int32_t)v;	/*XXX*/
    166 	for (i = 1; i < sizeof(_UTF8_range) / sizeof(_UTF8_range[0]); i++)
    167 		if (c >= _UTF8_range[i] && c < _UTF8_range[i + 1])
    168 			return i;
    169 
    170 	return -1;	/*out of range*/
    171 }
    172 
    173 static __inline void
    174 /*ARGSUSED*/
    175 _citrus_UTF8_init_state(_UTF8EncodingInfo *ei, _UTF8State *s)
    176 {
    177 	memset(s, 0, sizeof(*s));
    178 }
    179 
    180 static __inline void
    181 /*ARGSUSED*/
    182 _citrus_UTF8_pack_state(_UTF8EncodingInfo *ei, void *pspriv,
    183 			const _UTF8State *s)
    184 {
    185 	memcpy(pspriv, (const void *)s, sizeof(*s));
    186 }
    187 
    188 static __inline void
    189 /*ARGSUSED*/
    190 _citrus_UTF8_unpack_state(_UTF8EncodingInfo *ei, _UTF8State *s,
    191 			  const void *pspriv)
    192 {
    193 	memcpy((void *)s, pspriv, sizeof(*s));
    194 }
    195 
    196 static int
    197 _citrus_UTF8_mbrtowc_priv(_UTF8EncodingInfo *ei, wchar_t *pwc, const char **s,
    198 			  size_t n, _UTF8State *psenc, size_t *nresult)
    199 {
    200 	wchar_t wchar;
    201 	const char *s0;
    202 	int c;
    203 	int i;
    204 	int chlenbak;
    205 
    206 	_DIAGASSERT(nresult != 0);
    207 	_DIAGASSERT(ei != NULL);
    208 	_DIAGASSERT(s != NULL);
    209 	_DIAGASSERT(psenc != NULL);
    210 
    211 	s0 = *s;
    212 
    213 	if (s0 == NULL) {
    214 		_citrus_UTF8_init_state(ei, psenc);
    215 		*nresult = 0; /* state independent */
    216 		return (0);
    217 	}
    218 
    219 	chlenbak = psenc->chlen;
    220 
    221 	/* make sure we have the first byte in the buffer */
    222 	switch (psenc->chlen) {
    223 	case 0:
    224 		if (n < 1) {
    225 			goto restart;
    226 		}
    227 		psenc->ch[0] = *s0++;
    228 		psenc->chlen = 1;
    229 		n--;
    230 		break;
    231 	case 1: case 2: case 3: case 4: case 5:
    232 		break;
    233 	default:
    234 		/* illegal state */
    235 		goto ilseq;
    236 	}
    237 
    238 	c = _UTF8_count[psenc->ch[0] & 0xff];
    239 	if (c == 0)
    240 		goto ilseq;
    241 	while (psenc->chlen < c) {
    242 		if (n < 1) {
    243 			goto restart;
    244 		}
    245 		psenc->ch[psenc->chlen] = *s0++;
    246 		psenc->chlen++;
    247 		n--;
    248 	}
    249 
    250 	switch (c) {
    251 	case 1:
    252 		wchar = psenc->ch[0] & 0xff;
    253 		break;
    254 	case 2: case 3: case 4: case 5: case 6:
    255 		wchar = psenc->ch[0] & (0x7f >> c);
    256 		for (i = 1; i < c; i++) {
    257 			if ((psenc->ch[i] & 0xc0) != 0x80)
    258 				goto ilseq;
    259 			wchar <<= 6;
    260 			wchar |= (psenc->ch[i] & 0x3f);
    261 		}
    262 
    263 		_DIAGASSERT(findlen(wchar) == c);
    264 
    265 		break;
    266 	}
    267 
    268 	*s = s0;
    269 
    270 	psenc->chlen = 0;
    271 
    272 	if (pwc)
    273 		*pwc = wchar;
    274 
    275 	if (!wchar)
    276 		*nresult = 0;
    277 	else
    278 		*nresult = c - chlenbak;
    279 
    280 	return (0);
    281 
    282 ilseq:
    283 	psenc->chlen = 0;
    284 	*nresult = (size_t)-1;
    285 	return (EILSEQ);
    286 
    287 restart:
    288 	*s = s0;
    289 	*nresult = (size_t)-2;
    290 	return (0);
    291 }
    292 
    293 static int
    294 _citrus_UTF8_wcrtomb_priv(_UTF8EncodingInfo *ei, char *s, size_t n, wchar_t wc,
    295 			  _UTF8State *psenc, size_t *nresult)
    296 {
    297 	int cnt, i, ret;
    298 	wchar_t c;
    299 
    300 	_DIAGASSERT(ei != NULL);
    301 	_DIAGASSERT(nresult != 0);
    302 	_DIAGASSERT(s != NULL);
    303 
    304 	cnt = _UTF8_findlen(wc);
    305 	if (cnt <= 0 || cnt > 6) {
    306 		/* invalid UCS4 value */
    307 		ret = EILSEQ;
    308 		goto err;
    309 	}
    310 	if (n < cnt) {
    311 		/* bound check failure */
    312 		ret = E2BIG;
    313 		goto err;
    314 	}
    315 
    316 	c = wc;
    317 	if (s) {
    318 		for (i = cnt - 1; i > 0; i--) {
    319 			s[i] = 0x80 | (c & 0x3f);
    320 			c >>= 6;
    321 		}
    322 		s[0] = c;
    323 		if (cnt == 1)
    324 			s[0] &= 0x7f;
    325 		else {
    326 			s[0] &= (0x7f >> cnt);
    327 			s[0] |= ((0xff00 >> cnt) & 0xff);
    328 		}
    329 	}
    330 
    331 	*nresult = (size_t)cnt;
    332 	return 0;
    333 
    334 err:
    335 	*nresult = (size_t)-1;
    336 	return ret;
    337 }
    338 
    339 static __inline int
    340 /*ARGSUSED*/
    341 _citrus_UTF8_stdenc_wctocs(_UTF8EncodingInfo * __restrict ei,
    342 			   _csid_t * __restrict csid,
    343 			   _index_t * __restrict idx,
    344 			   wchar_t wc)
    345 {
    346 
    347 	_DIAGASSERT(csid != NULL && idx != NULL);
    348 
    349 	*csid = 0;
    350 	*idx = (_citrus_index_t)wc;
    351 
    352 	return (0);
    353 }
    354 
    355 static __inline int
    356 /*ARGSUSED*/
    357 _citrus_UTF8_stdenc_cstowc(_UTF8EncodingInfo * __restrict ei,
    358 			   wchar_t * __restrict wc,
    359 			   _csid_t csid, _index_t idx)
    360 {
    361 
    362 	_DIAGASSERT(wc != NULL);
    363 
    364 	if (csid != 0)
    365 		return (EILSEQ);
    366 
    367 	*wc = (wchar_t)idx;
    368 
    369 	return (0);
    370 }
    371 
    372 static int
    373 /*ARGSUSED*/
    374 _citrus_UTF8_encoding_module_init(_UTF8EncodingInfo * __restrict ei,
    375 				  const void * __restrict var, size_t lenvar)
    376 {
    377 	_DIAGASSERT(ei != NULL);
    378 
    379 	_UTF8_init_count();
    380 	memset((void *)ei, 0, sizeof(*ei));
    381 
    382 	return 0;
    383 }
    384 
    385 static void
    386 /*ARGSUSED*/
    387 _citrus_UTF8_encoding_module_uninit(_UTF8EncodingInfo *ei)
    388 {
    389 }
    390 
    391 
    392 /* ----------------------------------------------------------------------
    393  * public interface for ctype
    394  */
    395 
    396 _CITRUS_CTYPE_DECLS(UTF8);
    397 _CITRUS_CTYPE_DEF_OPS(UTF8);
    398 
    399 #include "citrus_ctype_template.h"
    400 
    401 /* ----------------------------------------------------------------------
    402  * public interface for stdenc
    403  */
    404 
    405 _CITRUS_STDENC_DECLS(UTF8);
    406 _CITRUS_STDENC_DEF_OPS(UTF8);
    407 
    408 #include "citrus_stdenc_template.h"
    409