1 /* $NetBSD: citrus_utf1632.c,v 1.9 2008/06/14 16:01:08 tnozaki Exp $ */
2
3 /*-
4 * SPDX-License-Identifier: BSD-2-Clause
5 *
6 * Copyright (c)2003 Citrus Project,
7 * All rights reserved.
8 *
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
11 * are met:
12 * 1. Redistributions of source code must retain the above copyright
13 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 *
18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
22 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 * SUCH DAMAGE.
29 */
30
31 #include <sys/cdefs.h>
32 #include <sys/endian.h>
33 #include <sys/types.h>
34
35 #include <assert.h>
36 #include <errno.h>
37 #include <limits.h>
38 #include <stddef.h>
39 #include <stdio.h>
40 #include <stdlib.h>
41 #include <string.h>
42 #include <wchar.h>
43
44 #include "citrus_namespace.h"
45 #include "citrus_types.h"
46 #include "citrus_module.h"
47 #include "citrus_stdenc.h"
48 #include "citrus_bcs.h"
49
50 #include "citrus_utf1632.h"
51
52
53 /* ----------------------------------------------------------------------
54 * private stuffs used by templates
55 */
56
57 typedef struct {
58 int chlen;
59 int current_endian;
60 uint8_t ch[4];
61 } _UTF1632State;
62
63 #define _ENDIAN_UNKNOWN 0
64 #define _ENDIAN_BIG 1
65 #define _ENDIAN_LITTLE 2
66 #if BYTE_ORDER == BIG_ENDIAN
67 #define _ENDIAN_INTERNAL _ENDIAN_BIG
68 #define _ENDIAN_SWAPPED _ENDIAN_LITTLE
69 #else
70 #define _ENDIAN_INTERNAL _ENDIAN_LITTLE
71 #define _ENDIAN_SWAPPED _ENDIAN_BIG
72 #endif
73 #define _MODE_UTF32 0x00000001U
74 #define _MODE_FORCE_ENDIAN 0x00000002U
75
76 typedef struct {
77 int preffered_endian;
78 unsigned int cur_max;
79 unsigned int cur_min;
80 uint32_t mode;
81 } _UTF1632EncodingInfo;
82
83 #define _FUNCNAME(m) _citrus_UTF1632_##m
84 #define _ENCODING_INFO _UTF1632EncodingInfo
85 #define _ENCODING_STATE _UTF1632State
86 #define _ENCODING_MB_CUR_MAX(_ei_) ((_ei_)->cur_max)
87 #define _ENCODING_MB_CUR_MIN(_ei_) ((_ei_)->cur_min)
88 #define _ENCODING_IS_STATE_DEPENDENT 0
89 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_) 0
90
91
92 static __inline void
93 /*ARGSUSED*/
_citrus_UTF1632_init_state(_UTF1632EncodingInfo * ei __unused,_UTF1632State * s)94 _citrus_UTF1632_init_state(_UTF1632EncodingInfo *ei __unused,
95 _UTF1632State *s)
96 {
97
98 memset(s, 0, sizeof(*s));
99 }
100
101 static int
_citrus_UTF1632_mbrtowc_priv(_UTF1632EncodingInfo * ei,wchar_t * pwc,char ** s,size_t n,_UTF1632State * psenc,size_t * nresult)102 _citrus_UTF1632_mbrtowc_priv(_UTF1632EncodingInfo *ei, wchar_t *pwc,
103 char **s, size_t n, _UTF1632State *psenc, size_t *nresult)
104 {
105 char *s0;
106 size_t result;
107 wchar_t wc = L'\0';
108 int chlenbak, endian, needlen;
109
110 s0 = *s;
111
112 if (s0 == NULL) {
113 _citrus_UTF1632_init_state(ei, psenc);
114 *nresult = 0; /* state independent */
115 return (0);
116 }
117
118 result = 0;
119 chlenbak = psenc->chlen;
120
121 refetch:
122 needlen = ((ei->mode & _MODE_UTF32) != 0 || chlenbak >= 2) ? 4 : 2;
123
124 while (chlenbak < needlen) {
125 if (n == 0)
126 goto restart;
127 psenc->ch[chlenbak++] = *s0++;
128 n--;
129 result++;
130 }
131
132 /* judge endian marker */
133 if ((ei->mode & _MODE_UTF32) == 0) {
134 /* UTF16 */
135 if (psenc->ch[0] == 0xFE && psenc->ch[1] == 0xFF) {
136 psenc->current_endian = _ENDIAN_BIG;
137 chlenbak = 0;
138 goto refetch;
139 } else if (psenc->ch[0] == 0xFF && psenc->ch[1] == 0xFE) {
140 psenc->current_endian = _ENDIAN_LITTLE;
141 chlenbak = 0;
142 goto refetch;
143 }
144 } else {
145 /* UTF32 */
146 if (psenc->ch[0] == 0x00 && psenc->ch[1] == 0x00 &&
147 psenc->ch[2] == 0xFE && psenc->ch[3] == 0xFF) {
148 psenc->current_endian = _ENDIAN_BIG;
149 chlenbak = 0;
150 goto refetch;
151 } else if (psenc->ch[0] == 0xFF && psenc->ch[1] == 0xFE &&
152 psenc->ch[2] == 0x00 && psenc->ch[3] == 0x00) {
153 psenc->current_endian = _ENDIAN_LITTLE;
154 chlenbak = 0;
155 goto refetch;
156 }
157 }
158 endian = ((ei->mode & _MODE_FORCE_ENDIAN) != 0 ||
159 psenc->current_endian == _ENDIAN_UNKNOWN) ? ei->preffered_endian :
160 psenc->current_endian;
161
162 /* get wc */
163 if ((ei->mode & _MODE_UTF32) == 0) {
164 /* UTF16 */
165 if (needlen == 2) {
166 switch (endian) {
167 case _ENDIAN_LITTLE:
168 wc = (psenc->ch[0] |
169 ((wchar_t)psenc->ch[1] << 8));
170 break;
171 case _ENDIAN_BIG:
172 wc = (psenc->ch[1] |
173 ((wchar_t)psenc->ch[0] << 8));
174 break;
175 default:
176 goto ilseq;
177 }
178 if (wc >= 0xD800 && wc <= 0xDBFF) {
179 /* surrogate high */
180 needlen = 4;
181 goto refetch;
182 }
183 } else {
184 /* surrogate low */
185 wc -= 0xD800; /* wc : surrogate high (see above) */
186 wc <<= 10;
187 switch (endian) {
188 case _ENDIAN_LITTLE:
189 if (psenc->ch[3] < 0xDC || psenc->ch[3] > 0xDF)
190 goto ilseq;
191 wc |= psenc->ch[2];
192 wc |= (wchar_t)(psenc->ch[3] & 3) << 8;
193 break;
194 case _ENDIAN_BIG:
195 if (psenc->ch[2]<0xDC || psenc->ch[2]>0xDF)
196 goto ilseq;
197 wc |= psenc->ch[3];
198 wc |= (wchar_t)(psenc->ch[2] & 3) << 8;
199 break;
200 default:
201 goto ilseq;
202 }
203 wc += 0x10000;
204 }
205 } else {
206 /* UTF32 */
207 switch (endian) {
208 case _ENDIAN_LITTLE:
209 wc = (psenc->ch[0] |
210 ((wchar_t)psenc->ch[1] << 8) |
211 ((wchar_t)psenc->ch[2] << 16) |
212 ((wchar_t)psenc->ch[3] << 24));
213 break;
214 case _ENDIAN_BIG:
215 wc = (psenc->ch[3] |
216 ((wchar_t)psenc->ch[2] << 8) |
217 ((wchar_t)psenc->ch[1] << 16) |
218 ((wchar_t)psenc->ch[0] << 24));
219 break;
220 default:
221 goto ilseq;
222 }
223 if (wc >= 0xD800 && wc <= 0xDFFF)
224 goto ilseq;
225 }
226
227
228 *pwc = wc;
229 psenc->chlen = 0;
230 *nresult = result;
231 *s = s0;
232
233 return (0);
234
235 ilseq:
236 *nresult = (size_t)-1;
237 psenc->chlen = 0;
238 return (EILSEQ);
239
240 restart:
241 *nresult = (size_t)-2;
242 psenc->chlen = chlenbak;
243 *s = s0;
244 return (0);
245 }
246
247 static int
_citrus_UTF1632_wcrtomb_priv(_UTF1632EncodingInfo * ei,char * s,size_t n,wchar_t wc,_UTF1632State * psenc,size_t * nresult)248 _citrus_UTF1632_wcrtomb_priv(_UTF1632EncodingInfo *ei, char *s, size_t n,
249 wchar_t wc, _UTF1632State *psenc, size_t *nresult)
250 {
251 wchar_t wc2;
252 static const char _bom[4] = {
253 0x00, 0x00, 0xFE, 0xFF,
254 };
255 const char *bom = &_bom[0];
256 size_t cnt;
257
258 cnt = (size_t)0;
259 if (psenc->current_endian == _ENDIAN_UNKNOWN) {
260 if ((ei->mode & _MODE_FORCE_ENDIAN) == 0) {
261 if (ei->mode & _MODE_UTF32)
262 cnt = 4;
263 else {
264 cnt = 2;
265 bom += 2;
266 }
267 if (n < cnt)
268 goto e2big;
269 memcpy(s, bom, cnt);
270 s += cnt, n -= cnt;
271 }
272 psenc->current_endian = ei->preffered_endian;
273 }
274
275 wc2 = 0;
276 if ((ei->mode & _MODE_UTF32)==0) {
277 /* UTF16 */
278 if (wc > 0xFFFF) {
279 /* surrogate */
280 if (wc > 0x10FFFF)
281 goto ilseq;
282 if (n < 4)
283 goto e2big;
284 cnt += 4;
285 wc -= 0x10000;
286 wc2 = (wc & 0x3FF) | 0xDC00;
287 wc = (wc>>10) | 0xD800;
288 } else {
289 if (n < 2)
290 goto e2big;
291 cnt += 2;
292 }
293
294 surrogate:
295 switch (psenc->current_endian) {
296 case _ENDIAN_BIG:
297 s[1] = wc;
298 s[0] = (wc >>= 8);
299 break;
300 case _ENDIAN_LITTLE:
301 s[0] = wc;
302 s[1] = (wc >>= 8);
303 break;
304 }
305 if (wc2 != 0) {
306 wc = wc2;
307 wc2 = 0;
308 s += 2;
309 goto surrogate;
310 }
311 } else {
312 /* UTF32 */
313 if (wc >= 0xD800 && wc <= 0xDFFF)
314 goto ilseq;
315 if (n < 4)
316 goto e2big;
317 cnt += 4;
318 switch (psenc->current_endian) {
319 case _ENDIAN_BIG:
320 s[3] = wc;
321 s[2] = (wc >>= 8);
322 s[1] = (wc >>= 8);
323 s[0] = (wc >>= 8);
324 break;
325 case _ENDIAN_LITTLE:
326 s[0] = wc;
327 s[1] = (wc >>= 8);
328 s[2] = (wc >>= 8);
329 s[3] = (wc >>= 8);
330 break;
331 }
332 }
333 *nresult = cnt;
334
335 return (0);
336
337 ilseq:
338 *nresult = (size_t)-1;
339 return (EILSEQ);
340 e2big:
341 *nresult = (size_t)-1;
342 return (E2BIG);
343 }
344
345 static void
parse_variable(_UTF1632EncodingInfo * __restrict ei,const void * __restrict var,size_t lenvar)346 parse_variable(_UTF1632EncodingInfo * __restrict ei,
347 const void * __restrict var, size_t lenvar)
348 {
349 const char *p;
350
351 p = var;
352 while (lenvar > 0) {
353 switch (*p) {
354 case 'B':
355 case 'b':
356 MATCH(big, ei->preffered_endian = _ENDIAN_BIG);
357 break;
358 case 'L':
359 case 'l':
360 MATCH(little, ei->preffered_endian = _ENDIAN_LITTLE);
361 break;
362 case 'i':
363 case 'I':
364 MATCH(internal, ei->preffered_endian = _ENDIAN_INTERNAL);
365 break;
366 case 's':
367 case 'S':
368 MATCH(swapped, ei->preffered_endian = _ENDIAN_SWAPPED);
369 break;
370 case 'F':
371 case 'f':
372 MATCH(force, ei->mode |= _MODE_FORCE_ENDIAN);
373 break;
374 case 'U':
375 case 'u':
376 MATCH(utf32, ei->mode |= _MODE_UTF32);
377 break;
378 }
379 p++;
380 lenvar--;
381 }
382 }
383
384 static int
385 /*ARGSUSED*/
_citrus_UTF1632_encoding_module_init(_UTF1632EncodingInfo * __restrict ei,const void * __restrict var,size_t lenvar)386 _citrus_UTF1632_encoding_module_init(_UTF1632EncodingInfo * __restrict ei,
387 const void * __restrict var, size_t lenvar)
388 {
389
390 memset((void *)ei, 0, sizeof(*ei));
391
392 parse_variable(ei, var, lenvar);
393
394 ei->cur_min = ((ei->mode&_MODE_UTF32) == 0) ? 2 : 4;
395 ei->cur_max = ((ei->mode&_MODE_UTF32) == 0) ? 6 : 8;
396 /* 6: endian + surrogate */
397 /* 8: endian + normal */
398
399 if (ei->preffered_endian == _ENDIAN_UNKNOWN) {
400 ei->preffered_endian = _ENDIAN_BIG;
401 }
402
403 return (0);
404 }
405
406 static void
407 /*ARGSUSED*/
_citrus_UTF1632_encoding_module_uninit(_UTF1632EncodingInfo * ei __unused)408 _citrus_UTF1632_encoding_module_uninit(_UTF1632EncodingInfo *ei __unused)
409 {
410
411 }
412
413 static __inline int
414 /*ARGSUSED*/
_citrus_UTF1632_stdenc_wctocs(_UTF1632EncodingInfo * __restrict ei __unused,_csid_t * __restrict csid,_index_t * __restrict idx,_wc_t wc)415 _citrus_UTF1632_stdenc_wctocs(_UTF1632EncodingInfo * __restrict ei __unused,
416 _csid_t * __restrict csid, _index_t * __restrict idx, _wc_t wc)
417 {
418
419 *csid = 0;
420 *idx = (_index_t)wc;
421
422 return (0);
423 }
424
425 static __inline int
426 /*ARGSUSED*/
_citrus_UTF1632_stdenc_cstowc(_UTF1632EncodingInfo * __restrict ei __unused,_wc_t * __restrict wc,_csid_t csid,_index_t idx)427 _citrus_UTF1632_stdenc_cstowc(_UTF1632EncodingInfo * __restrict ei __unused,
428 _wc_t * __restrict wc, _csid_t csid, _index_t idx)
429 {
430
431 if (csid != 0)
432 return (EILSEQ);
433
434 *wc = (_wc_t)idx;
435
436 return (0);
437 }
438
439 static __inline int
440 /*ARGSUSED*/
_citrus_UTF1632_stdenc_get_state_desc_generic(_UTF1632EncodingInfo * __restrict ei __unused,_UTF1632State * __restrict psenc,int * __restrict rstate)441 _citrus_UTF1632_stdenc_get_state_desc_generic(_UTF1632EncodingInfo * __restrict ei __unused,
442 _UTF1632State * __restrict psenc, int * __restrict rstate)
443 {
444
445 *rstate = (psenc->chlen == 0) ? _STDENC_SDGEN_INITIAL :
446 _STDENC_SDGEN_INCOMPLETE_CHAR;
447 return (0);
448 }
449
450 /* ----------------------------------------------------------------------
451 * public interface for stdenc
452 */
453
454 _CITRUS_STDENC_DECLS(UTF1632);
455 _CITRUS_STDENC_DEF_OPS(UTF1632);
456
457 #include "citrus_stdenc_template.h"
458