xref: /freebsd/lib/libiconv_modules/JOHAB/citrus_johab.c (revision fed1ca4b719c56c930f2259d80663cd34be812bb)
1 /* $FreeBSD$ */
2 /* $NetBSD: citrus_johab.c,v 1.4 2008/06/14 16:01:07 tnozaki Exp $ */
3 
4 /*-
5  * Copyright (c)2006 Citrus Project,
6  * All rights reserved.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  *
17  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27  * SUCH DAMAGE.
28  */
29 #include <sys/cdefs.h>
30 #include <sys/types.h>
31 
32 #include <assert.h>
33 #include <errno.h>
34 #include <limits.h>
35 #include <stdbool.h>
36 #include <stddef.h>
37 #include <stdint.h>
38 #include <stdio.h>
39 #include <stdlib.h>
40 #include <string.h>
41 #include <wchar.h>
42 
43 #include "citrus_namespace.h"
44 #include "citrus_types.h"
45 #include "citrus_bcs.h"
46 #include "citrus_module.h"
47 #include "citrus_stdenc.h"
48 #include "citrus_johab.h"
49 
50 /* ----------------------------------------------------------------------
51  * private stuffs used by templates
52  */
53 
54 typedef struct {
55 	int	 chlen;
56 	char	 ch[2];
57 } _JOHABState;
58 
59 typedef struct {
60 	int	 dummy;
61 } _JOHABEncodingInfo;
62 
63 #define _CEI_TO_EI(_cei_)		(&(_cei_)->ei)
64 #define _CEI_TO_STATE(_cei_, _func_)	(_cei_)->states.s_##_func_
65 
66 #define _FUNCNAME(m)			_citrus_JOHAB_##m
67 #define _ENCODING_INFO			_JOHABEncodingInfo
68 #define _ENCODING_STATE			_JOHABState
69 #define _ENCODING_MB_CUR_MAX(_ei_)		2
70 #define _ENCODING_IS_STATE_DEPENDENT		0
71 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_)	0
72 
73 
74 static __inline void
75 /*ARGSUSED*/
76 _citrus_JOHAB_init_state(_JOHABEncodingInfo * __restrict ei __unused,
77     _JOHABState * __restrict psenc)
78 {
79 
80 	psenc->chlen = 0;
81 }
82 
83 #if 0
84 static __inline void
85 /*ARGSUSED*/
86 _citrus_JOHAB_pack_state(_JOHABEncodingInfo * __restrict ei __unused,
87     void * __restrict pspriv, const _JOHABState * __restrict psenc)
88 {
89 
90 	memcpy(pspriv, (const void *)psenc, sizeof(*psenc));
91 }
92 
93 static __inline void
94 /*ARGSUSED*/
95 _citrus_JOHAB_unpack_state(_JOHABEncodingInfo * __restrict ei __unused,
96     _JOHABState * __restrict psenc, const void * __restrict pspriv)
97 {
98 
99 	memcpy((void *)psenc, pspriv, sizeof(*psenc));
100 }
101 #endif
102 
103 static void
104 /*ARGSUSED*/
105 _citrus_JOHAB_encoding_module_uninit(_JOHABEncodingInfo *ei __unused)
106 {
107 
108 	/* ei may be null */
109 }
110 
111 static int
112 /*ARGSUSED*/
113 _citrus_JOHAB_encoding_module_init(_JOHABEncodingInfo * __restrict ei __unused,
114     const void * __restrict var __unused, size_t lenvar __unused)
115 {
116 
117 	/* ei may be null */
118 	return (0);
119 }
120 
121 static __inline bool
122 ishangul(int l, int t)
123 {
124 
125 	return ((l >= 0x84 && l <= 0xD3) &&
126 	    ((t >= 0x41 && t <= 0x7E) || (t >= 0x81 && t <= 0xFE)));
127 }
128 
129 static __inline bool
130 isuda(int l, int t)
131 {
132 
133 	return ((l == 0xD8) &&
134 	    ((t >= 0x31 && t <= 0x7E) || (t >= 0x91 && t <= 0xFE)));
135 }
136 
137 static __inline bool
138 ishanja(int l, int t)
139 {
140 
141 	return (((l >= 0xD9 && l <= 0xDE) || (l >= 0xE0 && l <= 0xF9)) &&
142 	    ((t >= 0x31 && t <= 0x7E) || (t >= 0x91 && t <= 0xFE)));
143 }
144 
145 static int
146 /*ARGSUSED*/
147 _citrus_JOHAB_mbrtowc_priv(_JOHABEncodingInfo * __restrict ei,
148     wchar_t * __restrict pwc, char ** __restrict s, size_t n,
149     _JOHABState * __restrict psenc, size_t * __restrict nresult)
150 {
151 	char *s0;
152 	int l, t;
153 
154 	if (*s == NULL) {
155 		_citrus_JOHAB_init_state(ei, psenc);
156 		*nresult = _ENCODING_IS_STATE_DEPENDENT;
157 		return (0);
158 	}
159 	s0 = *s;
160 
161 	switch (psenc->chlen) {
162 	case 0:
163 		if (n-- < 1)
164 			goto restart;
165 		l = *s0++ & 0xFF;
166 		if (l <= 0x7F) {
167 			if (pwc != NULL)
168 				*pwc = (wchar_t)l;
169 			*nresult = (l == 0) ? 0 : 1;
170 			*s = s0;
171 			return (0);
172 		}
173 		psenc->ch[psenc->chlen++] = l;
174 		break;
175 	case 1:
176 		l = psenc->ch[0] & 0xFF;
177 		break;
178 	default:
179 		return (EINVAL);
180 	}
181 	if (n-- < 1) {
182 restart:
183 		*nresult = (size_t)-2;
184 		*s = s0;
185 		return (0);
186 	}
187 	t = *s0++ & 0xFF;
188 	if (!ishangul(l, t) && !isuda(l, t) && !ishanja(l, t)) {
189 		*nresult = (size_t)-1;
190 		return (EILSEQ);
191 	}
192 	if (pwc != NULL)
193 		*pwc = (wchar_t)(l << 8 | t);
194 	*nresult = s0 - *s;
195 	*s = s0;
196 	psenc->chlen = 0;
197 
198 	return (0);
199 }
200 
201 static int
202 /*ARGSUSED*/
203 _citrus_JOHAB_wcrtomb_priv(_JOHABEncodingInfo * __restrict ei __unused,
204     char * __restrict s, size_t n, wchar_t wc,
205     _JOHABState * __restrict psenc, size_t * __restrict nresult)
206 {
207 	int l, t;
208 
209 	if (psenc->chlen != 0)
210 		return (EINVAL);
211 
212 	/* XXX assume wchar_t as int */
213 	if ((uint32_t)wc <= 0x7F) {
214 		if (n < 1)
215 			goto e2big;
216 		*s = wc & 0xFF;
217 		*nresult = 1;
218 	} else if ((uint32_t)wc <= 0xFFFF) {
219 		if (n < 2) {
220 e2big:
221 			*nresult = (size_t)-1;
222 			return (E2BIG);
223 		}
224 		l = (wc >> 8) & 0xFF;
225 		t = wc & 0xFF;
226 		if (!ishangul(l, t) && !isuda(l, t) && !ishanja(l, t))
227 			goto ilseq;
228 		*s++ = l;
229 		*s = t;
230 		*nresult = 2;
231 	} else {
232 ilseq:
233 		*nresult = (size_t)-1;
234 		return (EILSEQ);
235 	}
236 	return (0);
237 
238 }
239 
240 static __inline int
241 /*ARGSUSED*/
242 _citrus_JOHAB_stdenc_wctocs(_JOHABEncodingInfo * __restrict ei __unused,
243     _csid_t * __restrict csid, _index_t * __restrict idx, wchar_t wc)
244 {
245 	int m, l, linear, t;
246 
247 	/* XXX assume wchar_t as int */
248 	if ((uint32_t)wc <= 0x7F) {
249 		*idx = (_index_t)wc;
250 		*csid = 0;
251 	} else if ((uint32_t)wc <= 0xFFFF) {
252 		l = (wc >> 8) & 0xFF;
253 		t = wc & 0xFF;
254 		if (ishangul(l, t) || isuda(l, t)) {
255 			*idx = (_index_t)wc;
256 			*csid = 1;
257 		} else {
258 			if (l >= 0xD9 && l <= 0xDE) {
259 				linear = l - 0xD9;
260 				m = 0x21;
261 			} else if (l >= 0xE0 && l <= 0xF9) {
262 				linear = l - 0xE0;
263 				m = 0x4A;
264 			} else
265 				return (EILSEQ);
266 			linear *= 188;
267 			if (t >= 0x31 && t <= 0x7E)
268 				linear += t - 0x31;
269 			else if (t >= 0x91 && t <= 0xFE)
270 				linear += t - 0x43;
271 			else
272 				return (EILSEQ);
273 			l = (linear / 94) + m;
274 			t = (linear % 94) + 0x21;
275 			*idx = (_index_t)((l << 8) | t);
276 			*csid = 2;
277 		}
278 	} else
279 		return (EILSEQ);
280 	return (0);
281 }
282 
283 static __inline int
284 /*ARGSUSED*/
285 _citrus_JOHAB_stdenc_cstowc(_JOHABEncodingInfo * __restrict ei __unused,
286     wchar_t * __restrict wc, _csid_t csid, _index_t idx)
287 {
288 	int m, n, l, linear, t;
289 
290 	switch (csid) {
291 	case 0:
292 	case 1:
293 		*wc = (wchar_t)idx;
294 		break;
295 	case 2:
296 		if (idx >= 0x2121 && idx <= 0x2C71) {
297 			m = 0xD9;
298 			n = 0x21;
299 		} else if (idx >= 0x4A21 && idx <= 0x7D7E) {
300 			m = 0xE0;
301 			n = 0x4A;
302 		} else
303 			return (EILSEQ);
304 		l = ((idx >> 8) & 0xFF) - n;
305 		t = (idx & 0xFF) - 0x21;
306 		linear = (l * 94) + t;
307 		l = (linear / 188) + m;
308 		t = linear % 188;
309 		t += (t <= 0x4D) ? 0x31 : 0x43;
310 		break;
311 	default:
312 		return (EILSEQ);
313 	}
314 	return (0);
315 }
316 
317 static __inline int
318 /*ARGSUSED*/
319 _citrus_JOHAB_stdenc_get_state_desc_generic(_JOHABEncodingInfo * __restrict ei __unused,
320     _JOHABState * __restrict psenc, int * __restrict rstate)
321 {
322 
323 	*rstate = (psenc->chlen == 0) ? _STDENC_SDGEN_INITIAL :
324 	    _STDENC_SDGEN_INCOMPLETE_CHAR;
325 	return (0);
326 }
327 
328 /* ----------------------------------------------------------------------
329  * public interface for stdenc
330  */
331 
332 _CITRUS_STDENC_DECLS(JOHAB);
333 _CITRUS_STDENC_DEF_OPS(JOHAB);
334 
335 #include "citrus_stdenc_template.h"
336