1 /* Character set conversion.
2 Copyright (C) 1999-2001, 2007, 2009-2023 Free Software Foundation, Inc.
3
4 This file is free software: you can redistribute it and/or modify
5 it under the terms of the GNU Lesser General Public License as
6 published by the Free Software Foundation; either version 2.1 of the
7 License, or (at your option) any later version.
8
9 This file is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU Lesser General Public License for more details.
13
14 You should have received a copy of the GNU Lesser General Public License
15 along with this program. If not, see <https://www.gnu.org/licenses/>. */
16
17 #include <config.h>
18
19 /* Specification. */
20 #include <iconv.h>
21
22 #include <stddef.h>
23
24 #if REPLACE_ICONV_UTF
25 # include <errno.h>
26 # include <stdint.h>
27 # include <stdlib.h>
28 # include "unistr.h"
29 #endif
30
31 #if REPLACE_ICONV_UTF
32
33 /* UTF-{16,32}{BE,LE} converters taken from GNU libiconv 1.11. */
34
35 /* Return code if invalid. (xxx_mbtowc) */
36 # define RET_ILSEQ -1
37 /* Return code if no bytes were read. (xxx_mbtowc) */
38 # define RET_TOOFEW -2
39
40 /* Return code if invalid. (xxx_wctomb) */
41 # define RET_ILUNI -1
42 /* Return code if output buffer is too small. (xxx_wctomb, xxx_reset) */
43 # define RET_TOOSMALL -2
44
45 /*
46 * UTF-16BE
47 */
48
49 /* Specification: RFC 2781 */
50
51 static int
52 utf16be_mbtowc (ucs4_t *pwc, const unsigned char *s, size_t n)
53 {
54 if (n >= 2)
55 {
56 ucs4_t wc = (s[0] << 8) + s[1];
57 if (wc >= 0xd800 && wc < 0xdc00)
58 {
59 if (n >= 4)
60 {
61 ucs4_t wc2 = (s[2] << 8) + s[3];
62 if (!(wc2 >= 0xdc00 && wc2 < 0xe000))
63 return RET_ILSEQ;
64 *pwc = 0x10000 + ((wc - 0xd800) << 10) + (wc2 - 0xdc00);
65 return 4;
66 }
67 }
68 else if (wc >= 0xdc00 && wc < 0xe000)
69 {
70 return RET_ILSEQ;
71 }
72 else
73 {
74 *pwc = wc;
75 return 2;
76 }
77 }
78 return RET_TOOFEW;
79 }
80
81 static int
82 utf16be_wctomb (unsigned char *r, ucs4_t wc, size_t n)
83 {
84 if (!(wc >= 0xd800 && wc < 0xe000))
85 {
86 if (wc < 0x10000)
87 {
88 if (n >= 2)
89 {
90 r[0] = (unsigned char) (wc >> 8);
91 r[1] = (unsigned char) wc;
92 return 2;
93 }
94 else
95 return RET_TOOSMALL;
96 }
97 else if (wc < 0x110000)
98 {
99 if (n >= 4)
100 {
101 ucs4_t wc1 = 0xd800 + ((wc - 0x10000) >> 10);
102 ucs4_t wc2 = 0xdc00 + ((wc - 0x10000) & 0x3ff);
103 r[0] = (unsigned char) (wc1 >> 8);
104 r[1] = (unsigned char) wc1;
105 r[2] = (unsigned char) (wc2 >> 8);
106 r[3] = (unsigned char) wc2;
107 return 4;
108 }
109 else
110 return RET_TOOSMALL;
111 }
112 }
113 return RET_ILUNI;
114 }
115
116 /*
117 * UTF-16LE
118 */
119
120 /* Specification: RFC 2781 */
121
122 static int
123 utf16le_mbtowc (ucs4_t *pwc, const unsigned char *s, size_t n)
124 {
125 if (n >= 2)
126 {
127 ucs4_t wc = s[0] + (s[1] << 8);
128 if (wc >= 0xd800 && wc < 0xdc00)
129 {
130 if (n >= 4)
131 {
132 ucs4_t wc2 = s[2] + (s[3] << 8);
133 if (!(wc2 >= 0xdc00 && wc2 < 0xe000))
134 return RET_ILSEQ;
135 *pwc = 0x10000 + ((wc - 0xd800) << 10) + (wc2 - 0xdc00);
136 return 4;
137 }
138 }
139 else if (wc >= 0xdc00 && wc < 0xe000)
140 {
141 return RET_ILSEQ;
142 }
143 else
144 {
145 *pwc = wc;
146 return 2;
147 }
148 }
149 return RET_TOOFEW;
150 }
151
152 static int
153 utf16le_wctomb (unsigned char *r, ucs4_t wc, size_t n)
154 {
155 if (!(wc >= 0xd800 && wc < 0xe000))
156 {
157 if (wc < 0x10000)
158 {
159 if (n >= 2)
160 {
161 r[0] = (unsigned char) wc;
162 r[1] = (unsigned char) (wc >> 8);
163 return 2;
164 }
165 else
166 return RET_TOOSMALL;
167 }
168 else if (wc < 0x110000)
169 {
170 if (n >= 4)
171 {
172 ucs4_t wc1 = 0xd800 + ((wc - 0x10000) >> 10);
173 ucs4_t wc2 = 0xdc00 + ((wc - 0x10000) & 0x3ff);
174 r[0] = (unsigned char) wc1;
175 r[1] = (unsigned char) (wc1 >> 8);
176 r[2] = (unsigned char) wc2;
177 r[3] = (unsigned char) (wc2 >> 8);
178 return 4;
179 }
180 else
181 return RET_TOOSMALL;
182 }
183 }
184 return RET_ILUNI;
185 }
186
187 /*
188 * UTF-32BE
189 */
190
191 /* Specification: Unicode 3.1 Standard Annex #19 */
192
193 static int
194 utf32be_mbtowc (ucs4_t *pwc, const unsigned char *s, size_t n)
195 {
196 if (n >= 4)
197 {
198 ucs4_t wc = (s[0] << 24) + (s[1] << 16) + (s[2] << 8) + s[3];
199 if (wc < 0x110000 && !(wc >= 0xd800 && wc < 0xe000))
200 {
201 *pwc = wc;
202 return 4;
203 }
204 else
205 return RET_ILSEQ;
206 }
207 return RET_TOOFEW;
208 }
209
210 static int
211 utf32be_wctomb (unsigned char *r, ucs4_t wc, size_t n)
212 {
213 if (wc < 0x110000 && !(wc >= 0xd800 && wc < 0xe000))
214 {
215 if (n >= 4)
216 {
217 r[0] = 0;
218 r[1] = (unsigned char) (wc >> 16);
219 r[2] = (unsigned char) (wc >> 8);
220 r[3] = (unsigned char) wc;
221 return 4;
222 }
223 else
224 return RET_TOOSMALL;
225 }
226 return RET_ILUNI;
227 }
228
229 /*
230 * UTF-32LE
231 */
232
233 /* Specification: Unicode 3.1 Standard Annex #19 */
234
235 static int
236 utf32le_mbtowc (ucs4_t *pwc, const unsigned char *s, size_t n)
237 {
238 if (n >= 4)
239 {
240 ucs4_t wc = s[0] + (s[1] << 8) + (s[2] << 16) + (s[3] << 24);
241 if (wc < 0x110000 && !(wc >= 0xd800 && wc < 0xe000))
242 {
243 *pwc = wc;
244 return 4;
245 }
246 else
247 return RET_ILSEQ;
248 }
249 return RET_TOOFEW;
250 }
251
252 static int
253 utf32le_wctomb (unsigned char *r, ucs4_t wc, size_t n)
254 {
255 if (wc < 0x110000 && !(wc >= 0xd800 && wc < 0xe000))
256 {
257 if (n >= 4)
258 {
259 r[0] = (unsigned char) wc;
260 r[1] = (unsigned char) (wc >> 8);
261 r[2] = (unsigned char) (wc >> 16);
262 r[3] = 0;
263 return 4;
264 }
265 else
266 return RET_TOOSMALL;
267 }
268 return RET_ILUNI;
269 }
270
271 #endif
272
273 size_t
274 rpl_iconv (iconv_t cd,
275 ICONV_CONST char **inbuf, size_t *inbytesleft,
276 char **outbuf, size_t *outbytesleft)
277 #undef iconv
278 {
279 #if REPLACE_ICONV_UTF
280 switch ((uintptr_t) cd)
281 {
282 {
283 int (*xxx_wctomb) (unsigned char *, ucs4_t, size_t);
284
285 case (uintptr_t) _ICONV_UTF8_UTF16BE:
286 xxx_wctomb = utf16be_wctomb;
287 goto loop_from_utf8;
288 case (uintptr_t) _ICONV_UTF8_UTF16LE:
289 xxx_wctomb = utf16le_wctomb;
290 goto loop_from_utf8;
291 case (uintptr_t) _ICONV_UTF8_UTF32BE:
292 xxx_wctomb = utf32be_wctomb;
293 goto loop_from_utf8;
294 case (uintptr_t) _ICONV_UTF8_UTF32LE:
295 xxx_wctomb = utf32le_wctomb;
296 goto loop_from_utf8;
297
298 loop_from_utf8:
299 if (inbuf == NULL || *inbuf == NULL)
300 return 0;
301 {
302 ICONV_CONST char *inptr = *inbuf;
303 size_t inleft = *inbytesleft;
304 char *outptr = *outbuf;
305 size_t outleft = *outbytesleft;
306 size_t res = 0;
307 while (inleft > 0)
308 {
309 ucs4_t uc;
310 int m = u8_mbtoucr (&uc, (const uint8_t *) inptr, inleft);
311 if (m <= 0)
312 {
313 if (m == -1)
314 {
315 errno = EILSEQ;
316 res = (size_t)(-1);
317 break;
318 }
319 if (m == -2)
320 {
321 errno = EINVAL;
322 res = (size_t)(-1);
323 break;
324 }
325 abort ();
326 }
327 else
328 {
329 int n = xxx_wctomb ((uint8_t *) outptr, uc, outleft);
330 if (n < 0)
331 {
332 if (n == RET_ILUNI)
333 {
334 errno = EILSEQ;
335 res = (size_t)(-1);
336 break;
337 }
338 if (n == RET_TOOSMALL)
339 {
340 errno = E2BIG;
341 res = (size_t)(-1);
342 break;
343 }
344 abort ();
345 }
346 else
347 {
348 inptr += m;
349 inleft -= m;
350 outptr += n;
351 outleft -= n;
352 }
353 }
354 }
355 *inbuf = inptr;
356 *inbytesleft = inleft;
357 *outbuf = outptr;
358 *outbytesleft = outleft;
359 return res;
360 }
361 }
362
363 {
364 int (*xxx_mbtowc) (ucs4_t *, const unsigned char *, size_t);
365
366 case (uintptr_t) _ICONV_UTF16BE_UTF8:
367 xxx_mbtowc = utf16be_mbtowc;
368 goto loop_to_utf8;
369 case (uintptr_t) _ICONV_UTF16LE_UTF8:
370 xxx_mbtowc = utf16le_mbtowc;
371 goto loop_to_utf8;
372 case (uintptr_t) _ICONV_UTF32BE_UTF8:
373 xxx_mbtowc = utf32be_mbtowc;
374 goto loop_to_utf8;
375 case (uintptr_t) _ICONV_UTF32LE_UTF8:
376 xxx_mbtowc = utf32le_mbtowc;
377 goto loop_to_utf8;
378
379 loop_to_utf8:
380 if (inbuf == NULL || *inbuf == NULL)
381 return 0;
382 {
383 ICONV_CONST char *inptr = *inbuf;
384 size_t inleft = *inbytesleft;
385 char *outptr = *outbuf;
386 size_t outleft = *outbytesleft;
387 size_t res = 0;
388 while (inleft > 0)
389 {
390 ucs4_t uc;
391 int m = xxx_mbtowc (&uc, (const uint8_t *) inptr, inleft);
392 if (m <= 0)
393 {
394 if (m == RET_ILSEQ)
395 {
396 errno = EILSEQ;
397 res = (size_t)(-1);
398 break;
399 }
400 if (m == RET_TOOFEW)
401 {
402 errno = EINVAL;
403 res = (size_t)(-1);
404 break;
405 }
406 abort ();
407 }
408 else
409 {
410 int n = u8_uctomb ((uint8_t *) outptr, uc, outleft);
411 if (n < 0)
412 {
413 if (n == -1)
414 {
415 errno = EILSEQ;
416 res = (size_t)(-1);
417 break;
418 }
419 if (n == -2)
420 {
421 errno = E2BIG;
422 res = (size_t)(-1);
423 break;
424 }
425 abort ();
426 }
427 else
428 {
429 inptr += m;
430 inleft -= m;
431 outptr += n;
432 outleft -= n;
433 }
434 }
435 }
436 *inbuf = inptr;
437 *inbytesleft = inleft;
438 *outbuf = outptr;
439 *outbytesleft = outleft;
440 return res;
441 }
442 }
443 }
444 #endif
445 return iconv (cd, inbuf, inbytesleft, outbuf, outbytesleft);
446 }