1 /* $NetBSD: citrus_utf1632.c,v 1.13 2022/04/19 20:32:14 rillig Exp $ */
2
3 /*-
4 * Copyright (c)2003 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 #include <sys/cdefs.h>
30 #if defined(LIBC_SCCS) && !defined(lint)
31 __RCSID("$NetBSD: citrus_utf1632.c,v 1.13 2022/04/19 20:32:14 rillig Exp $");
32 #endif /* LIBC_SCCS and not lint */
33
34 #include <assert.h>
35 #include <errno.h>
36 #include <string.h>
37 #include <stdio.h>
38 #include <stdlib.h>
39 #include <stddef.h>
40 #include <limits.h>
41 #include <wchar.h>
42 #include <sys/types.h>
43 #include <machine/endian.h>
44
45 #include "citrus_namespace.h"
46 #include "citrus_types.h"
47 #include "citrus_module.h"
48 #include "citrus_stdenc.h"
49 #include "citrus_bcs.h"
50
51 #include "citrus_utf1632.h"
52
53
54 /* ----------------------------------------------------------------------
55 * private stuffs used by templates
56 */
57
58 typedef struct {
59 u_int8_t ch[4];
60 int chlen;
61 int current_endian;
62 } _UTF1632State;
63
64 typedef struct {
65 int preffered_endian;
66 unsigned int cur_max;
67 #define _ENDIAN_UNKNOWN 0
68 #define _ENDIAN_BIG 1
69 #define _ENDIAN_LITTLE 2
70 u_int32_t mode;
71 #define _MODE_UTF32 0x00000001U
72 #define _MODE_FORCE_ENDIAN 0x00000002U
73 } _UTF1632EncodingInfo;
74
75 #define _FUNCNAME(m) _citrus_UTF1632_##m
76 #define _ENCODING_INFO _UTF1632EncodingInfo
77 #define _ENCODING_STATE _UTF1632State
78 #define _ENCODING_MB_CUR_MAX(_ei_) ((_ei_)->cur_max)
79 #define _ENCODING_IS_STATE_DEPENDENT 0
80 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_) 0
81
82
83 static __inline void
84 /*ARGSUSED*/
_citrus_UTF1632_init_state(_UTF1632EncodingInfo * ei,_UTF1632State * s)85 _citrus_UTF1632_init_state(_UTF1632EncodingInfo *ei, _UTF1632State *s)
86 {
87 memset(s, 0, sizeof(*s));
88 }
89
90 static int
_citrus_UTF1632_mbrtowc_priv(_UTF1632EncodingInfo * ei,wchar_t * pwc,const char ** s,size_t n,_UTF1632State * psenc,size_t * nresult)91 _citrus_UTF1632_mbrtowc_priv(_UTF1632EncodingInfo *ei, wchar_t *pwc,
92 const char **s, size_t n, _UTF1632State *psenc,
93 size_t *nresult)
94 {
95 int chlenbak, endian, needlen;
96 wchar_t wc;
97 size_t result;
98 const char *s0;
99
100 _DIAGASSERT(nresult != 0);
101 _DIAGASSERT(ei != NULL);
102 _DIAGASSERT(s != NULL);
103 _DIAGASSERT(psenc != NULL);
104
105 s0 = *s;
106
107 if (s0 == NULL) {
108 _citrus_UTF1632_init_state(ei, psenc);
109 *nresult = 0; /* state independent */
110 return (0);
111 }
112
113 result = 0;
114 chlenbak = psenc->chlen;
115
116 refetch:
117 if ((ei->mode & _MODE_UTF32) != 0 || chlenbak>=2)
118 needlen = 4;
119 else
120 needlen = 2;
121
122 while (chlenbak < needlen) {
123 if (n==0)
124 goto restart;
125 psenc->ch[chlenbak++] = *s0++;
126 n--;
127 result++;
128 }
129
130 if (psenc->current_endian == _ENDIAN_UNKNOWN) {
131 if ((ei->mode & _MODE_FORCE_ENDIAN) == 0) {
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 }
159 psenc->current_endian = ei->preffered_endian;
160 }
161 endian = psenc->current_endian;
162
163 /* get wc */
164 if ((ei->mode & _MODE_UTF32) == 0) {
165 /* UTF16 */
166 if (needlen==2) {
167 switch (endian) {
168 case _ENDIAN_LITTLE:
169 wc = (psenc->ch[0] |
170 ((wchar_t)psenc->ch[1] << 8));
171 break;
172 case _ENDIAN_BIG:
173 wc = (psenc->ch[1] |
174 ((wchar_t)psenc->ch[0] << 8));
175 break;
176 default:
177 goto ilseq;
178 }
179 if (wc >= 0xD800 && wc <= 0xDBFF) {
180 /* surrogate high */
181 needlen=4;
182 goto refetch;
183 }
184 } else {
185 /* surrogate low */
186 wc -= 0xD800; /* wc : surrogate high (see above) */
187 wc <<= 10;
188 switch (endian) {
189 case _ENDIAN_LITTLE:
190 if (psenc->ch[3]<0xDC || psenc->ch[3]>0xDF)
191 goto ilseq;
192 wc |= psenc->ch[2];
193 wc |= (wchar_t)(psenc->ch[3] & 3) << 8;
194 break;
195 case _ENDIAN_BIG:
196 if (psenc->ch[2]<0xDC || psenc->ch[2]>0xDF)
197 goto ilseq;
198 wc |= psenc->ch[3];
199 wc |= (wchar_t)(psenc->ch[2] & 3) << 8;
200 break;
201 default:
202 goto ilseq;
203 }
204 wc += 0x10000;
205 }
206 } else {
207 /* UTF32 */
208 switch (endian) {
209 case _ENDIAN_LITTLE:
210 wc = (psenc->ch[0] |
211 ((wchar_t)psenc->ch[1] << 8) |
212 ((wchar_t)psenc->ch[2] << 16) |
213 ((wchar_t)psenc->ch[3] << 24));
214 break;
215 case _ENDIAN_BIG:
216 wc = (psenc->ch[3] |
217 ((wchar_t)psenc->ch[2] << 8) |
218 ((wchar_t)psenc->ch[1] << 16) |
219 ((wchar_t)psenc->ch[0] << 24));
220 break;
221 default:
222 goto ilseq;
223 }
224 if (wc >= 0xD800 && wc <= 0xDFFF)
225 goto ilseq;
226 }
227
228
229 *pwc = wc;
230 psenc->chlen = 0;
231 *nresult = result;
232 *s = s0;
233
234 return (0);
235
236 ilseq:
237 *nresult = (size_t)-1;
238 psenc->chlen = 0;
239 return (EILSEQ);
240
241 restart:
242 *nresult = (size_t)-2;
243 psenc->chlen = chlenbak;
244 *s = s0;
245 return (0);
246 }
247
248 static int
_citrus_UTF1632_wcrtomb_priv(_UTF1632EncodingInfo * ei,char * s,size_t n,wchar_t wc,_UTF1632State * psenc,size_t * nresult)249 _citrus_UTF1632_wcrtomb_priv(_UTF1632EncodingInfo *ei, char *s, size_t n,
250 wchar_t wc, _UTF1632State *psenc,
251 size_t *nresult)
252 {
253 wchar_t wc2;
254 static const char _bom[4] = {
255 #if BYTE_ORDER == BIG_ENDIAN
256 0x00, 0x00, 0xFE, 0xFF,
257 #else
258 0xFF, 0xFE, 0x00, 0x00,
259 #endif
260 };
261 const char *bom = &_bom[0];
262 size_t cnt;
263
264 _DIAGASSERT(ei != NULL);
265 _DIAGASSERT(nresult != 0);
266 _DIAGASSERT(s != NULL);
267
268 cnt = (size_t)0;
269 if (psenc->current_endian == _ENDIAN_UNKNOWN) {
270 if ((ei->mode & _MODE_FORCE_ENDIAN) == 0) {
271 if (ei->mode & _MODE_UTF32) {
272 cnt = 4;
273 } else {
274 cnt = 2;
275 #if BYTE_ORDER == BIG_ENDIAN
276 bom += 2;
277 #endif
278 }
279 if (n < cnt)
280 goto e2big;
281 memcpy(s, bom, cnt);
282 s += cnt, n -= cnt;
283 }
284 psenc->current_endian = ei->preffered_endian;
285 }
286
287 wc2 = 0;
288 if ((ei->mode & _MODE_UTF32)==0) {
289 /* UTF16 */
290 if (wc>0xFFFF) {
291 /* surrogate */
292 if (wc>0x10FFFF)
293 goto ilseq;
294 if (n < 4)
295 goto e2big;
296 cnt += 4;
297 wc -= 0x10000;
298 wc2 = (wc & 0x3FF) | 0xDC00;
299 wc = (wc>>10) | 0xD800;
300 } else {
301 if (n < 2)
302 goto e2big;
303 cnt += 2;
304 }
305
306 surrogate:
307 switch (psenc->current_endian) {
308 case _ENDIAN_BIG:
309 s[1] = wc;
310 s[0] = (wc >>= 8);
311 break;
312 case _ENDIAN_LITTLE:
313 s[0] = wc;
314 s[1] = (wc >>= 8);
315 break;
316 }
317 if (wc2!=0) {
318 wc = wc2;
319 wc2 = 0;
320 s += 2;
321 goto surrogate;
322 }
323 } else {
324 /* UTF32 */
325 if (wc >= 0xD800 && wc <= 0xDFFF)
326 goto ilseq;
327 if (n < 4)
328 goto e2big;
329 cnt += 4;
330 switch (psenc->current_endian) {
331 case _ENDIAN_BIG:
332 s[3] = wc;
333 s[2] = (wc >>= 8);
334 s[1] = (wc >>= 8);
335 s[0] = (wc >>= 8);
336 break;
337 case _ENDIAN_LITTLE:
338 s[0] = wc;
339 s[1] = (wc >>= 8);
340 s[2] = (wc >>= 8);
341 s[3] = (wc >>= 8);
342 break;
343 }
344 }
345 *nresult = cnt;
346
347 return 0;
348
349 ilseq:
350 *nresult = (size_t)-1;
351 return EILSEQ;
352 e2big:
353 *nresult = (size_t)-1;
354 return E2BIG;
355 }
356
357 static void
parse_variable(_UTF1632EncodingInfo * __restrict ei,const void * __restrict var,size_t lenvar)358 parse_variable(_UTF1632EncodingInfo * __restrict ei,
359 const void * __restrict var, size_t lenvar)
360 {
361 #define MATCH(x, act) \
362 do { \
363 if (lenvar >= (sizeof(#x)-1) && \
364 _bcs_strncasecmp(p, #x, sizeof(#x)-1) == 0) { \
365 act; \
366 lenvar -= sizeof(#x)-1; \
367 p += sizeof(#x)-1; \
368 } \
369 } while (0)
370 const char *p;
371 p = var;
372 while (lenvar>0) {
373 switch (*p) {
374 case 'B':
375 case 'b':
376 MATCH(big, ei->preffered_endian = _ENDIAN_BIG);
377 break;
378 case 'L':
379 case 'l':
380 MATCH(little, ei->preffered_endian = _ENDIAN_LITTLE);
381 break;
382 case 'F':
383 case 'f':
384 MATCH(force, ei->mode |= _MODE_FORCE_ENDIAN);
385 break;
386 case 'U':
387 case 'u':
388 MATCH(utf32, ei->mode |= _MODE_UTF32);
389 break;
390 }
391 p++;
392 lenvar--;
393 }
394 }
395
396 static int
397 /*ARGSUSED*/
_citrus_UTF1632_encoding_module_init(_UTF1632EncodingInfo * __restrict ei,const void * __restrict var,size_t lenvar)398 _citrus_UTF1632_encoding_module_init(_UTF1632EncodingInfo * __restrict ei,
399 const void * __restrict var,
400 size_t lenvar)
401 {
402 _DIAGASSERT(ei != NULL);
403
404 memset((void *)ei, 0, sizeof(*ei));
405
406 parse_variable(ei, var, lenvar);
407
408 if ((ei->mode&_MODE_UTF32)==0)
409 ei->cur_max = 6; /* endian + surrogate */
410 else
411 ei->cur_max = 8; /* endian + normal */
412
413 if (ei->preffered_endian == _ENDIAN_UNKNOWN) {
414 #if BYTE_ORDER == BIG_ENDIAN
415 ei->preffered_endian = _ENDIAN_BIG;
416 #else
417 ei->preffered_endian = _ENDIAN_LITTLE;
418 #endif
419 }
420
421 return (0);
422 }
423
424 static void
425 /*ARGSUSED*/
_citrus_UTF1632_encoding_module_uninit(_UTF1632EncodingInfo * ei)426 _citrus_UTF1632_encoding_module_uninit(_UTF1632EncodingInfo *ei)
427 {
428 }
429
430 static __inline int
431 /*ARGSUSED*/
_citrus_UTF1632_stdenc_wctocs(_UTF1632EncodingInfo * __restrict ei,_csid_t * __restrict csid,_index_t * __restrict idx,_wc_t wc)432 _citrus_UTF1632_stdenc_wctocs(_UTF1632EncodingInfo * __restrict ei,
433 _csid_t * __restrict csid,
434 _index_t * __restrict idx,
435 _wc_t wc)
436 {
437
438 _DIAGASSERT(csid != NULL && idx != NULL);
439
440 *csid = 0;
441 *idx = (_index_t)wc;
442
443 return (0);
444 }
445
446 static __inline int
447 /*ARGSUSED*/
_citrus_UTF1632_stdenc_cstowc(_UTF1632EncodingInfo * __restrict ei,_wc_t * __restrict wc,_csid_t csid,_index_t idx)448 _citrus_UTF1632_stdenc_cstowc(_UTF1632EncodingInfo * __restrict ei,
449 _wc_t * __restrict wc,
450 _csid_t csid, _index_t idx)
451 {
452
453 _DIAGASSERT(wc != NULL);
454
455 if (csid != 0)
456 return (EILSEQ);
457
458 *wc = (_wc_t)idx;
459
460 return (0);
461 }
462
463 static __inline int
464 /*ARGSUSED*/
_citrus_UTF1632_stdenc_get_state_desc_generic(_UTF1632EncodingInfo * __restrict ei,_UTF1632State * __restrict psenc,int * __restrict rstate)465 _citrus_UTF1632_stdenc_get_state_desc_generic(_UTF1632EncodingInfo * __restrict ei,
466 _UTF1632State * __restrict psenc,
467 int * __restrict rstate)
468 {
469
470 if (psenc->chlen == 0)
471 *rstate = _STDENC_SDGEN_INITIAL;
472 else
473 *rstate = _STDENC_SDGEN_INCOMPLETE_CHAR;
474
475 return 0;
476 }
477
478 /* ----------------------------------------------------------------------
479 * public interface for stdenc
480 */
481
482 _CITRUS_STDENC_DECLS(UTF1632);
483 _CITRUS_STDENC_DEF_OPS(UTF1632);
484
485 #include "citrus_stdenc_template.h"
486