xref: /netbsd/lib/libc/citrus/modules/citrus_euctw.c (revision bf9ec67e)
1 /*	$NetBSD: citrus_euctw.c,v 1.5 2002/03/28 10:53:48 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)1999 Citrus Project,
31  * All rights reserved.
32  *
33  * Redistribution and use in source and binary forms, with or without
34  * modification, are permitted provided that the following conditions
35  * are met:
36  * 1. Redistributions of source code must retain the above copyright
37  *    notice, this list of conditions and the following disclaimer.
38  * 2. Redistributions in binary form must reproduce the above copyright
39  *    notice, this list of conditions and the following disclaimer in the
40  *    documentation and/or other materials provided with the distribution.
41  *
42  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
43  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
44  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
45  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
46  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
47  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
48  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
49  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
50  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
51  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
52  * SUCH DAMAGE.
53  *
54  *	$Citrus: xpg4dl/FreeBSD/lib/libc/locale/euctw.c,v 1.13 2001/06/21 01:51:44 yamt Exp $
55  */
56 
57 #include <sys/cdefs.h>
58 #if defined(LIBC_SCCS) && !defined(lint)
59 __RCSID("$NetBSD: citrus_euctw.c,v 1.5 2002/03/28 10:53:48 yamt Exp $");
60 #endif /* LIBC_SCCS and not lint */
61 
62 #include <assert.h>
63 #include <errno.h>
64 #include <string.h>
65 #include <stdio.h>
66 #include <stdlib.h>
67 #include <stddef.h>
68 #include <locale.h>
69 #include <wchar.h>
70 #include <sys/types.h>
71 #include <limits.h>
72 #include "citrus_module.h"
73 #include "citrus_ctype.h"
74 #include "citrus_euctw.h"
75 
76 
77 /* ----------------------------------------------------------------------
78  * private stuffs used by templates
79  */
80 
81 typedef struct {
82 	char ch[4];
83 	int chlen;
84 } _EUCTWState;
85 
86 typedef struct {
87 	int dummy;
88 } _EUCTWEncodingInfo;
89 typedef struct {
90 	_EUCTWEncodingInfo	ei;
91 	struct {
92 		/* for future multi-locale facility */
93 		_EUCTWState	s_mblen;
94 		_EUCTWState	s_mbrlen;
95 		_EUCTWState	s_mbrtowc;
96 		_EUCTWState	s_mbtowc;
97 		_EUCTWState	s_mbsrtowcs;
98 		_EUCTWState	s_wcrtomb;
99 		_EUCTWState	s_wcsrtombs;
100 		_EUCTWState	s_wctomb;
101 	} states;
102 } _EUCTWCTypeInfo;
103 
104 #define	_SS2	0x008e
105 #define	_SS3	0x008f
106 
107 #define _CEI_TO_EI(_cei_)		(&(_cei_)->ei)
108 #define _CEI_TO_STATE(_cei_, _func_)	(_cei_)->states.s_##_func_
109 
110 #define _FUNCNAME(m)			_citrus_EUCTW_##m
111 #define _ENCODING_INFO			_EUCTWEncodingInfo
112 #define _CTYPE_INFO			_EUCTWCTypeInfo
113 #define _ENCODING_STATE			_EUCTWState
114 #define _ENCODING_MB_CUR_MAX(_ei_)	4
115 #define _ENCODING_IS_STATE_DEPENDENT	0
116 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_)	0
117 
118 static __inline int
119 _citrus_EUCTW_cs(u_int c)
120 {
121 	c &= 0xff;
122 
123 	return ((c & 0x80) ? (c == _SS2 ? 2 : 1) : 0);
124 }
125 
126 static __inline int
127 _citrus_EUCTW_count(int cs)
128 {
129 	switch (cs) {
130 	case 0:
131 		return 1;
132 	case 1:
133 		return 2;
134 	case 2:
135 		return 4;
136 	case 3:
137 		abort();
138 		/*NOTREACHED*/
139 	}
140 	return 0;
141 }
142 
143 static __inline void
144 /*ARGSUSED*/
145 _citrus_EUCTW_init_state(_EUCTWEncodingInfo * __restrict ei,
146 			 _EUCTWState * __restrict s)
147 {
148 	memset(s, 0, sizeof(*s));
149 }
150 
151 static __inline void
152 /*ARGSUSED*/
153 _citrus_EUCTW_pack_state(_EUCTWEncodingInfo * __restrict ei,
154 			 void * __restrict pspriv,
155 			 const _EUCTWState * __restrict s)
156 {
157 	memcpy(pspriv, (const void *)s, sizeof(*s));
158 }
159 
160 static __inline void
161 /*ARGSUSED*/
162 _citrus_EUCTW_unpack_state(_EUCTWEncodingInfo * __restrict ei,
163 			   _EUCTWState * __restrict s,
164 			   const void * __restrict pspriv)
165 {
166 	memcpy((void *)s, pspriv, sizeof(*s));
167 }
168 
169 static int
170 /*ARGSUSED*/
171 _citrus_EUCTW_stdencoding_init(_EUCTWEncodingInfo * __restrict ei,
172 			       const void * __restrict var, size_t lenvar)
173 {
174 
175 	_DIAGASSERT(ei != NULL);
176 
177 	memset((void *)ei, 0, sizeof(*ei));
178 
179 	return 0;
180 }
181 
182 static void
183 /*ARGSUSED*/
184 _citrus_EUCTW_stdencoding_uninit(_EUCTWEncodingInfo *ei)
185 {
186 }
187 
188 static int
189 _citrus_EUCTW_mbrtowc_priv(_EUCTWEncodingInfo * __restrict ei,
190 			   wchar_t * __restrict pwc,
191 			   const char ** __restrict s,
192 			   size_t n, _EUCTWState * __restrict psenc,
193 			   size_t * __restrict nresult)
194 {
195 	wchar_t wchar;
196 	int c, cs;
197 	int chlenbak;
198 	const char *s0;
199 
200 	_DIAGASSERT(nresult != 0);
201 	_DIAGASSERT(ei != NULL);
202 	_DIAGASSERT(psenc != NULL);
203 	_DIAGASSERT(s != NULL);
204 
205 	s0 = *s;
206 
207 	if (s0 == NULL) {
208 		_citrus_EUCTW_init_state(ei, psenc);
209 		*nresult = 0; /* state independent */
210 		return (0);
211 	}
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 	case 2:
224 		break;
225 	default:
226 		/* illgeal state */
227 		goto ilseq;
228 	}
229 
230 	c = _citrus_EUCTW_count(cs = _citrus_EUCTW_cs(psenc->ch[0] & 0xff));
231 	if (c == 0)
232 		goto ilseq;
233 	while (psenc->chlen < c) {
234 		if (n < 1)
235 			goto ilseq;
236 		psenc->ch[psenc->chlen] = *s0++;
237 		psenc->chlen++;
238 		n--;
239 	}
240 
241 	wchar = 0;
242 	switch (cs) {
243 	case 0:
244 		if (psenc->ch[0] & 0x80)
245 			goto ilseq;
246 		wchar = psenc->ch[0] & 0xff;
247 		break;
248 	case 1:
249 		if (!(psenc->ch[0] & 0x80) || !(psenc->ch[1] & 0x80))
250 			goto ilseq;
251 		wchar = ((psenc->ch[0] & 0xff) << 8) | (psenc->ch[1] & 0xff);
252 		wchar |= 'G' << 24;
253 		break;
254 	case 2:
255 		if ((u_char)psenc->ch[1] < 0xa1 || 0xa7 < (u_char)psenc->ch[1])
256 			goto ilseq;
257 		if (!(psenc->ch[2] & 0x80) || !(psenc->ch[3] & 0x80))
258 			goto ilseq;
259 		wchar = ((psenc->ch[2] & 0xff) << 8) | (psenc->ch[3] & 0xff);
260 		wchar |= ('G' + psenc->ch[1] - 0xa1) << 24;
261 		break;
262 	default:
263 		goto ilseq;
264 	}
265 
266 	*s = s0;
267 	psenc->chlen = 0;
268 
269 	if (pwc)
270 		*pwc = wchar;
271 
272 	if (!wchar)
273 		*nresult = 0;
274 	else
275 		*nresult = c - chlenbak;
276 
277 	return (0);
278 
279 ilseq:
280 	psenc->chlen = 0;
281 	*nresult = (size_t)-1;
282 	return (EILSEQ);
283 
284 restart:
285 	*s = s0;
286 	*nresult = (size_t)-1;
287 	return (0);
288 }
289 
290 static int
291 _citrus_EUCTW_wcrtomb_priv(_EUCTWEncodingInfo * __restrict ei,
292 			   char * __restrict s, size_t n, wchar_t wc,
293 			   _EUCTWState * __restrict psenc,
294 			   size_t * __restrict nresult)
295 {
296 	wchar_t cs = wc & 0x7f000080;
297 	wchar_t v;
298 	int i, len, clen;
299 
300 	_DIAGASSERT(ei != NULL);
301 	_DIAGASSERT(nresult != 0);
302 	_DIAGASSERT(s != NULL);
303 
304 	clen = 1;
305 	if (wc & 0x00007f00)
306 		clen = 2;
307 	if ((wc & 0x007f0000) && !(wc & 0x00800000))
308 		clen = 3;
309 
310 	if (clen == 1 && cs == 0x00000000) {
311 		/* ASCII */
312 		len = 1;
313 		if (n < len)
314 			goto ilseq;
315 		v = wc & 0x0000007f;
316 	} else if (clen == 2 && cs == ('G' << 24)) {
317 		/* CNS-11643-1 */
318 		len = 2;
319 		if (n < len)
320 			goto ilseq;
321 		v = wc & 0x00007f7f;
322 		v |= 0x00008080;
323 	} else if (clen == 2 && 'H' <= (cs >> 24) && (cs >> 24) <= 'M') {
324 		/* CNS-11643-[2-7] */
325 		len = 4;
326 		if (n < len)
327 			goto ilseq;
328 		*s++ = _SS2;
329 		*s++ = (cs >> 24) - 'H' + 0xa2;
330 		v = wc & 0x00007f7f;
331 		v |= 0x00008080;
332 	} else
333 		goto ilseq;
334 
335 	i = clen;
336 	while (i-- > 0)
337 		*s++ = (v >> (i << 3)) & 0xff;
338 
339 	*nresult = len;
340 	return (0);
341 
342 ilseq:
343 	*nresult = (size_t)-1;
344 	return (EILSEQ);
345 }
346 
347 /* ----------------------------------------------------------------------
348  * public interface for ctype
349  */
350 
351 _CITRUS_CTYPE_DECLS(EUCTW);
352 _CITRUS_CTYPE_DEF_OPS(EUCTW);
353 
354 #include "citrus_ctype_template.h"
355