xref: /freebsd/lib/libkiconv/xlat16_iconv.c (revision 5b20c5e1f8d59f14901f0c0e4b6b577eaa3f9569)
1 /*-
2  * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
3  *
4  * Copyright (c) 2003, 2005 Ryuichiro Imura
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  * $FreeBSD$
29  */
30 
31 /*
32  * kiconv(3) requires shared linked, and reduce module size
33  * when statically linked.
34  */
35 
36 #ifdef PIC
37 
38 #include <sys/types.h>
39 #include <sys/iconv.h>
40 #include <sys/sysctl.h>
41 
42 #include <ctype.h>
43 #include <dlfcn.h>
44 #include <err.h>
45 #include <errno.h>
46 #include <locale.h>
47 #include <stdio.h>
48 #include <stdlib.h>
49 #include <string.h>
50 #include <wctype.h>
51 
52 #include "quirks.h"
53 
54 struct xlat16_table {
55 	uint32_t *	idx[0x200];
56 	void *		data;
57 	size_t		size;
58 };
59 
60 static struct xlat16_table kiconv_xlat16_open(const char *, const char *, int);
61 static int chklocale(int, const char *);
62 
63 #ifdef ICONV_DLOPEN
64 typedef void *iconv_t;
65 static int my_iconv_init(void);
66 static iconv_t (*my_iconv_open)(const char *, const char *);
67 static size_t (*my_iconv)(iconv_t, char **, size_t *, char **, size_t *);
68 static int (*my_iconv_close)(iconv_t);
69 #else
70 #include <iconv.h>
71 #define my_iconv_init() 0
72 #define my_iconv_open iconv_open
73 #define my_iconv iconv
74 #define my_iconv_close iconv_close
75 #endif
76 static size_t my_iconv_char(iconv_t, u_char **, size_t *, u_char **, size_t *);
77 
78 int
79 kiconv_add_xlat16_cspair(const char *tocode, const char *fromcode, int flag)
80 {
81 	int error;
82 	size_t idxsize;
83 	struct xlat16_table xt;
84 	void *data;
85 	char *p;
86 	const char unicode[] = ENCODING_UNICODE;
87 
88 	if ((flag & KICONV_WCTYPE) == 0 &&
89 	    strcmp(unicode, tocode) != 0 &&
90 	    strcmp(unicode, fromcode) != 0 &&
91 	    kiconv_lookupconv(unicode) == 0) {
92 		error = kiconv_add_xlat16_cspair(unicode, fromcode, flag);
93 		if (error)
94 			return (-1);
95 		error = kiconv_add_xlat16_cspair(tocode, unicode, flag);
96 		return (error);
97 	}
98 
99 	if (kiconv_lookupcs(tocode, fromcode) == 0)
100 		return (0);
101 
102 	if (flag & KICONV_WCTYPE)
103 		xt = kiconv_xlat16_open(fromcode, fromcode, flag);
104 	else
105 		xt = kiconv_xlat16_open(tocode, fromcode, flag);
106 	if (xt.size == 0)
107 		return (-1);
108 
109 	idxsize = sizeof(xt.idx);
110 
111 	if ((idxsize + xt.size) > ICONV_CSMAXDATALEN) {
112 		errno = E2BIG;
113 		return (-1);
114 	}
115 
116 	if ((data = malloc(idxsize + xt.size)) != NULL) {
117 		p = data;
118 		memcpy(p, xt.idx, idxsize);
119 		p += idxsize;
120 		memcpy(p, xt.data, xt.size);
121 		error = kiconv_add_xlat16_table(tocode, fromcode, data,
122 		    (int)(idxsize + xt.size));
123 		return (error);
124 	}
125 
126 	return (-1);
127 }
128 
129 int
130 kiconv_add_xlat16_cspairs(const char *foreigncode, const char *localcode)
131 {
132 	int error, locale;
133 
134 	error = kiconv_add_xlat16_cspair(foreigncode, localcode,
135 	    KICONV_FROM_LOWER | KICONV_FROM_UPPER);
136 	if (error)
137 		return (error);
138 	error = kiconv_add_xlat16_cspair(localcode, foreigncode,
139 	    KICONV_LOWER | KICONV_UPPER);
140 	if (error)
141 		return (error);
142 	locale = chklocale(LC_CTYPE, localcode);
143 	if (locale == 0) {
144 		error = kiconv_add_xlat16_cspair(KICONV_WCTYPE_NAME, localcode,
145 		    KICONV_WCTYPE);
146 		if (error)
147 			return (error);
148 	}
149 
150 	return (0);
151 }
152 
153 static struct xlat16_table
154 kiconv_xlat16_open(const char *tocode, const char *fromcode, int lcase)
155 {
156 	u_char src[3], dst[4], *srcp, *dstp, ud, ld;
157 	int us, ls, ret;
158 	uint16_t c;
159 	uint32_t table[0x80];
160 	size_t inbytesleft, outbytesleft, pre_q_size, post_q_size;
161 	struct xlat16_table xt;
162 	struct quirk_replace_list *pre_q_list, *post_q_list;
163 	iconv_t cd;
164 	char *p;
165 
166 	xt.data = NULL;
167 	xt.size = 0;
168 
169 	src[2] = '\0';
170 	dst[3] = '\0';
171 
172 	ret = my_iconv_init();
173 	if (ret)
174 		return (xt);
175 
176 	cd = my_iconv_open(search_quirk(tocode, fromcode, &pre_q_list, &pre_q_size),
177 	    search_quirk(fromcode, tocode, &post_q_list, &post_q_size));
178 	if (cd == (iconv_t) (-1))
179 		return (xt);
180 
181 	if ((xt.data = malloc(0x200 * 0x80 * sizeof(uint32_t))) == NULL)
182 		return (xt);
183 
184 	p = xt.data;
185 
186 	for (ls = 0 ; ls < 0x200 ; ls++) {
187 		xt.idx[ls] = NULL;
188 		for (us = 0 ; us < 0x80 ; us++) {
189 			srcp = src;
190 			dstp = dst;
191 
192 			inbytesleft = 2;
193 			outbytesleft = 3;
194 			bzero(dst, outbytesleft);
195 
196 			c = ((ls & 0x100 ? us | 0x80 : us) << 8) | (u_char)ls;
197 
198 			if (lcase & KICONV_WCTYPE) {
199 				if ((c & 0xff) == 0)
200 					c >>= 8;
201 				if (iswupper(c)) {
202 					c = towlower(c);
203 					if ((c & 0xff00) == 0)
204 						c <<= 8;
205 					table[us] = c | XLAT16_HAS_LOWER_CASE;
206 				} else if (iswlower(c)) {
207 					c = towupper(c);
208 					if ((c & 0xff00) == 0)
209 						c <<= 8;
210 					table[us] = c | XLAT16_HAS_UPPER_CASE;
211 				} else
212 					table[us] = 0;
213 				/*
214 				 * store not NULL
215 				 */
216 				if (table[us])
217 					xt.idx[ls] = table;
218 
219 				continue;
220 			}
221 
222 			c = quirk_vendor2unix(c, pre_q_list, pre_q_size);
223 			src[0] = (u_char)(c >> 8);
224 			src[1] = (u_char)c;
225 
226 			ret = my_iconv_char(cd, &srcp, &inbytesleft,
227 				&dstp, &outbytesleft);
228 			if (ret == -1) {
229 				table[us] = 0;
230 				continue;
231 			}
232 
233 			ud = (u_char)dst[0];
234 			ld = (u_char)dst[1];
235 
236 			switch(outbytesleft) {
237 			case 0:
238 #ifdef XLAT16_ACCEPT_3BYTE_CHR
239 				table[us] = (ud << 8) | ld;
240 				table[us] |= (u_char)dst[2] << 16;
241 				table[us] |= XLAT16_IS_3BYTE_CHR;
242 #else
243 				table[us] = 0;
244 				continue;
245 #endif
246 				break;
247 			case 1:
248 				table[us] = quirk_unix2vendor((ud << 8) | ld,
249 				    post_q_list, post_q_size);
250 				if ((table[us] >> 8) == 0)
251 					table[us] |= XLAT16_ACCEPT_NULL_OUT;
252 				break;
253 			case 2:
254 				table[us] = ud;
255 				if (lcase & KICONV_LOWER && ud != tolower(ud)) {
256 					table[us] |= (u_char)tolower(ud) << 16;
257 					table[us] |= XLAT16_HAS_LOWER_CASE;
258 				}
259 				if (lcase & KICONV_UPPER && ud != toupper(ud)) {
260 					table[us] |= (u_char)toupper(ud) << 16;
261 					table[us] |= XLAT16_HAS_UPPER_CASE;
262 				}
263 				break;
264 			}
265 
266 			switch(inbytesleft) {
267 			case 0:
268 				if ((ls & 0xff) == 0)
269 					table[us] |= XLAT16_ACCEPT_NULL_IN;
270 				break;
271 			case 1:
272 				c = ls > 0xff ? us | 0x80 : us;
273 				if (lcase & KICONV_FROM_LOWER && c != tolower(c)) {
274 					table[us] |= (u_char)tolower(c) << 16;
275 					table[us] |= XLAT16_HAS_FROM_LOWER_CASE;
276 				}
277 				if (lcase & KICONV_FROM_UPPER && c != toupper(c)) {
278 					table[us] |= (u_char)toupper(c) << 16;
279 					table[us] |= XLAT16_HAS_FROM_UPPER_CASE;
280 				}
281 				break;
282 			}
283 
284 			if (table[us] == 0)
285 				continue;
286 
287 			/*
288 			 * store not NULL
289 			 */
290 			xt.idx[ls] = table;
291 		}
292 		if (xt.idx[ls]) {
293 			memcpy(p, table, sizeof(table));
294 			p += sizeof(table);
295 		}
296 	}
297 	my_iconv_close(cd);
298 
299 	xt.size = p - (char *)xt.data;
300 	xt.data = realloc(xt.data, xt.size);
301 	return (xt);
302 }
303 
304 static int
305 chklocale(int category, const char *code)
306 {
307 	char *p;
308 	int error = -1;
309 
310 	p = strchr(setlocale(category, NULL), '.');
311 	if (p++) {
312 		error = strcasecmp(code, p);
313 		if (error) {
314 			/* XXX - can't avoid calling quirk here... */
315 			error = strcasecmp(code, kiconv_quirkcs(p,
316 			    KICONV_VENDOR_MICSFT));
317 		}
318 	}
319 	return (error);
320 }
321 
322 #ifdef ICONV_DLOPEN
323 static int
324 my_iconv_init(void)
325 {
326 	void *iconv_lib;
327 
328 	iconv_lib = dlopen("libiconv.so", RTLD_LAZY | RTLD_GLOBAL);
329 	if (iconv_lib == NULL) {
330 		warn("Unable to load iconv library: %s\n", dlerror());
331 		errno = ENOENT;
332 		return (-1);
333 	}
334 	my_iconv_open = dlsym(iconv_lib, "iconv_open");
335 	my_iconv = dlsym(iconv_lib, "iconv");
336 	my_iconv_close = dlsym(iconv_lib, "iconv_close");
337 
338 	return (0);
339 }
340 #endif
341 
342 static size_t
343 my_iconv_char(iconv_t cd, u_char **ibuf, size_t * ilen, u_char **obuf,
344 	size_t * olen)
345 {
346 	u_char *sp, *dp, ilocal[3], olocal[3];
347 	u_char c1, c2;
348 	int ret;
349 	size_t ir, or;
350 
351 	sp = *ibuf;
352 	dp = *obuf;
353 	ir = *ilen;
354 
355 	bzero(*obuf, *olen);
356 	ret = my_iconv(cd, (char **)&sp, ilen, (char **)&dp, olen);
357 	c1 = (*obuf)[0];
358 	c2 = (*obuf)[1];
359 
360 	if (ret == -1) {
361 		if (*ilen == ir - 1 && (*ibuf)[1] == '\0' && (c1 || c2))
362 			return (0);
363 		else
364 			return (-1);
365 	}
366 
367 	/*
368 	 * We must judge if inbuf is a single byte char or double byte char.
369 	 * Here, to judge, try first byte(*sp) conversion and compare.
370 	 */
371 	ir = 1;
372 	or = 3;
373 
374 	bzero(olocal, or);
375 	memcpy(ilocal, *ibuf, sizeof(ilocal));
376 	sp = ilocal;
377 	dp = olocal;
378 
379 	if ((my_iconv(cd,(char **)&sp, &ir, (char **)&dp, &or)) != -1) {
380 		if (olocal[0] != c1)
381 			return (ret);
382 
383 		if (olocal[1] == c2 && (*ibuf)[1] == '\0') {
384 			/*
385 			 * inbuf is a single byte char
386 			 */
387 			*ilen = 1;
388 			*olen = or;
389 			return (ret);
390 		}
391 
392 		switch(or) {
393 		case 0:
394 		case 1:
395 			if (olocal[1] == c2) {
396 				/*
397 				 * inbuf is a single byte char,
398 				 * so return false here.
399 				 */
400 				return (-1);
401 			} else {
402 				/*
403 				 * inbuf is a double byte char
404 				 */
405 				return (ret);
406 			}
407 			break;
408 		case 2:
409 			/*
410 			 * should compare second byte of inbuf
411 			 */
412 			break;
413 		}
414 	} else {
415 		/*
416 		 * inbuf clould not be splitted, so inbuf is
417 		 * a double byte char.
418 		 */
419 		return (ret);
420 	}
421 
422 	/*
423 	 * try second byte(*(sp+1)) conversion, and compare
424 	 */
425 	ir = 1;
426 	or = 3;
427 
428 	bzero(olocal, or);
429 
430 	sp = ilocal + 1;
431 	dp = olocal;
432 
433 	if ((my_iconv(cd,(char **)&sp, &ir, (char **)&dp, &or)) != -1) {
434 		if (olocal[0] == c2)
435 			/*
436 			 * inbuf is a single byte char
437 			 */
438 			return (-1);
439 	}
440 
441 	return (ret);
442 }
443 
444 #else /* statically linked */
445 
446 #include <sys/types.h>
447 #include <sys/iconv.h>
448 #include <errno.h>
449 
450 int
451 kiconv_add_xlat16_cspair(const char *tocode __unused, const char *fromcode __unused,
452     int flag __unused)
453 {
454 
455 	errno = EINVAL;
456 	return (-1);
457 }
458 
459 int
460 kiconv_add_xlat16_cspairs(const char *tocode __unused, const char *fromcode __unused)
461 {
462 	errno = EINVAL;
463 	return (-1);
464 }
465 
466 #endif /* PIC */
467