source: branches/samba-3.2.x/source/lib/iconv.c@ 341

Last change on this file since 341 was 341, checked in by Herwig Bauernfeind, 16 years ago

divers fix for Ticket #68 in 3.2
File size: 19.1 KB
Line 
1/*
2 Unix SMB/CIFS implementation.
3 minimal iconv implementation
4 Copyright (C) Andrew Tridgell 2001
5 Copyright (C) Jelmer Vernooij 2002,2003
6
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
11
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
16
17 You should have received a copy of the GNU General Public License
18 along with this program. If not, see <http://www.gnu.org/licenses/>.
19*/
20
21#include "includes.h"
22
23/*
24 * We have to use strcasecmp here as the character conversions
25 * haven't been initialised yet. JRA.
26 */
27
28#undef strcasecmp
29
30/**
31 * @file
32 *
33 * @brief Samba wrapper/stub for iconv character set conversion.
34 *
35 * iconv is the XPG2 interface for converting between character
36 * encodings. This file provides a Samba wrapper around it, and also
37 * a simple reimplementation that is used if the system does not
38 * implement iconv.
39 *
40 * Samba only works with encodings that are supersets of ASCII: ascii
41 * characters like whitespace can be tested for directly, multibyte
42 * sequences start with a byte with the high bit set, and strings are
43 * terminated by a nul byte.
44 *
45 * Note that the only function provided by iconv is conversion between
46 * characters. It doesn't directly support operations like
47 * uppercasing or comparison. We have to convert to UCS-2 and compare
48 * there.
49 *
50 * @sa Samba Developers Guide
51 **/
52
53static_decl_charset;
54
55static size_t ascii_pull(void *,const char **, size_t *, char **, size_t *);
56static size_t ascii_push(void *,const char **, size_t *, char **, size_t *);
57static size_t latin1_push(void *,const char **, size_t *, char **, size_t *);
58static size_t utf8_pull(void *,const char **, size_t *, char **, size_t *);
59static size_t utf8_push(void *,const char **, size_t *, char **, size_t *);
60static size_t ucs2hex_pull(void *,const char **, size_t *, char **, size_t *);
61static size_t ucs2hex_push(void *,const char **, size_t *, char **, size_t *);
62static size_t iconv_copy(void *,const char **, size_t *, char **, size_t *);
63static size_t iconv_swab (void *,const char **, size_t *, char **, size_t *);
64
65static struct charset_functions builtin_functions[] = {
66 /* windows is really neither UCS-2 not UTF-16 */
67 {"UCS-2LE", iconv_copy, iconv_copy},
68 {"UTF-16LE", iconv_copy, iconv_copy},
69 {"UCS-2BE", iconv_swab, iconv_swab},
70 {"UTF-16BE", iconv_swab, iconv_swab},
71
72 /* we include the UTF-8 alias to cope with differing locale settings */
73 {"UTF8", utf8_pull, utf8_push},
74 {"UTF-8", utf8_pull, utf8_push},
75 {"ASCII", ascii_pull, ascii_push},
76 {"646", ascii_pull, ascii_push},
77 {"ISO-8859-1", ascii_pull, latin1_push},
78 {"UCS2-HEX", ucs2hex_pull, ucs2hex_push},
79 {NULL, NULL, NULL}
80};
81
82static struct charset_functions *charsets = NULL;
83
84static struct charset_functions *find_charset_functions(const char *name)
85{
86 struct charset_functions *c = charsets;
87
88 while(c) {
89 if (strcasecmp(name, c->name) == 0) {
90 return c;
91 }
92 c = c->next;
93 }
94
95 return NULL;
96}
97
98NTSTATUS smb_register_charset(struct charset_functions *funcs)
99{
100 if (!funcs) {
101 return NT_STATUS_INVALID_PARAMETER;
102 }
103
104 DEBUG(5, ("Attempting to register new charset %s\n", funcs->name));
105 /* Check whether we already have this charset... */
106 if (find_charset_functions(funcs->name)) {
107 DEBUG(0, ("Duplicate charset %s, not registering\n", funcs->name));
108 return NT_STATUS_OBJECT_NAME_COLLISION;
109 }
110
111 funcs->next = funcs->prev = NULL;
112 DEBUG(5, ("Registered charset %s\n", funcs->name));
113 DLIST_ADD(charsets, funcs);
114 return NT_STATUS_OK;
115}
116
117static void lazy_initialize_iconv(void)
118{
119 static bool initialized;
120 int i;
121
122 if (!initialized) {
123 initialized = True;
124 for(i = 0; builtin_functions[i].name; i++)
125 smb_register_charset(&builtin_functions[i]);
126 static_init_charset;
127 }
128}
129
130#ifdef __OS2__
131// i could have done a static variable w/o this function. but i feel it's nicer this way. SCS
132// the purpose of this function is to save the to_name to get the korean and japanese code set working
133char * save_toname(char *toname, bool what)
134{
135 static char *to_name=NULL;
136
137 if ( what == 0 )
138 to_name = SMB_STRDUP(toname);
139
140 return to_name;
141}
142#endif
143
144#ifdef HAVE_NATIVE_ICONV
145/* if there was an error then reset the internal state,
146 this ensures that we don't have a shift state remaining for
147 character sets like SJIS */
148static size_t sys_iconv(void *cd,
149 const char **inbuf, size_t *inbytesleft,
150 char **outbuf, size_t *outbytesleft)
151{
152#ifdef __OS2__
153 uint16 *outbuf_uc = ( uint16 * )*outbuf;
154 char *to_name = save_toname(NULL, 1);
155#endif
156 size_t ret = iconv((iconv_t)cd,
157 (char **)inbuf, inbytesleft,
158 outbuf, outbytesleft);
159 if (ret == (size_t)-1) {
160 int saved_errno = errno;
161 iconv(cd, NULL, NULL, NULL, NULL);
162 errno = saved_errno;
163 }
164#ifdef __OS2__
165 /* Workaround for path separator on OS/2 */
166 else
167 {
168 if( (strstr(to_name, "949") != NULL) || /* Korean CP */
169 (strstr(to_name, "932") != NULL) || /* Japanese CP */
170 (strstr(to_name, "942") != NULL) || /* Japanese CP */
171 (strstr(to_name, "943") != NULL) ) /* Japanese CP */
172 {
173 while(( char * )outbuf_uc < *outbuf )
174 {
175 if( *outbuf_uc == 0x20a9 || /* Korean WON */
176 *outbuf_uc == 0x00a5 ) /* Japanese YEN */
177 *outbuf_uc = '\\';
178
179 outbuf_uc++;
180 }
181 }
182 }
183#endif
184 return ret;
185}
186#endif
187
188/**
189 * This is a simple portable iconv() implementaion.
190 *
191 * It only knows about a very small number of character sets - just
192 * enough that Samba works on systems that don't have iconv.
193 **/
194size_t smb_iconv(smb_iconv_t cd,
195 const char **inbuf, size_t *inbytesleft,
196 char **outbuf, size_t *outbytesleft)
197{
198 char cvtbuf[2048];
199 char *bufp = cvtbuf;
200 size_t bufsize;
201
202#ifdef __OS2__
203 save_toname(cd->to_name, 0);
204#endif
205
206 /* in many cases we can go direct */
207 if (cd->direct) {
208 return cd->direct(cd->cd_direct,
209 inbuf, inbytesleft, outbuf, outbytesleft);
210 }
211
212
213 /* otherwise we have to do it chunks at a time */
214 while (*inbytesleft > 0) {
215 bufp = cvtbuf;
216 bufsize = sizeof(cvtbuf);
217
218 if (cd->pull(cd->cd_pull,
219 inbuf, inbytesleft, &bufp, &bufsize) == -1
220 && errno != E2BIG) return -1;
221
222 bufp = cvtbuf;
223 bufsize = sizeof(cvtbuf) - bufsize;
224
225 if (cd->push(cd->cd_push,
226 (const char **)&bufp, &bufsize,
227 outbuf, outbytesleft) == -1) return -1;
228 }
229
230 return 0;
231}
232
233
234static bool is_utf16(const char *name)
235{
236 return strcasecmp(name, "UCS-2LE") == 0 ||
237 strcasecmp(name, "UTF-16LE") == 0;
238}
239
240/*
241 simple iconv_open() wrapper
242 */
243smb_iconv_t smb_iconv_open(const char *tocode, const char *fromcode)
244{
245 smb_iconv_t ret;
246 struct charset_functions *from, *to;
247
248 lazy_initialize_iconv();
249 from = charsets;
250 to = charsets;
251
252 ret = SMB_MALLOC_P(struct _smb_iconv_t);
253 if (!ret) {
254 errno = ENOMEM;
255 return (smb_iconv_t)-1;
256 }
257 memset(ret, 0, sizeof(struct _smb_iconv_t));
258
259 ret->from_name = SMB_STRDUP(fromcode);
260 ret->to_name = SMB_STRDUP(tocode);
261
262 /* check for the simplest null conversion */
263 if (strcasecmp(fromcode, tocode) == 0) {
264 ret->direct = iconv_copy;
265 return ret;
266 }
267
268 /* check if we have a builtin function for this conversion */
269 from = find_charset_functions(fromcode);
270 if(from)ret->pull = from->pull;
271
272 to = find_charset_functions(tocode);
273 if(to)ret->push = to->push;
274
275 /* check if we can use iconv for this conversion */
276#ifdef HAVE_NATIVE_ICONV
277 if (!ret->pull) {
278 ret->cd_pull = iconv_open("UTF-16LE", fromcode);
279 if (ret->cd_pull == (iconv_t)-1)
280 ret->cd_pull = iconv_open("UCS-2LE", fromcode);
281 if (ret->cd_pull != (iconv_t)-1)
282 ret->pull = sys_iconv;
283 }
284
285 if (!ret->push) {
286 ret->cd_push = iconv_open(tocode, "UTF-16LE");
287 if (ret->cd_push == (iconv_t)-1)
288 ret->cd_push = iconv_open(tocode, "UCS-2LE");
289 if (ret->cd_push != (iconv_t)-1)
290 ret->push = sys_iconv;
291 }
292#endif
293
294 /* check if there is a module available that can do this conversion */
295 if (!ret->pull && NT_STATUS_IS_OK(smb_probe_module("charset", fromcode))) {
296 if(!(from = find_charset_functions(fromcode)))
297 DEBUG(0, ("Module %s doesn't provide charset %s!\n", fromcode, fromcode));
298 else
299 ret->pull = from->pull;
300 }
301
302 if (!ret->push && NT_STATUS_IS_OK(smb_probe_module("charset", tocode))) {
303 if(!(to = find_charset_functions(tocode)))
304 DEBUG(0, ("Module %s doesn't provide charset %s!\n", tocode, tocode));
305 else
306 ret->push = to->push;
307 }
308
309 if (!ret->push || !ret->pull) {
310 SAFE_FREE(ret->from_name);
311 SAFE_FREE(ret->to_name);
312 SAFE_FREE(ret);
313 errno = EINVAL;
314 return (smb_iconv_t)-1;
315 }
316
317 /* check for conversion to/from ucs2 */
318 if (is_utf16(fromcode) && to) {
319 ret->direct = to->push;
320 ret->push = ret->pull = NULL;
321 return ret;
322 }
323
324 if (is_utf16(tocode) && from) {
325 ret->direct = from->pull;
326 ret->push = ret->pull = NULL;
327 return ret;
328 }
329
330 /* Check if we can do the conversion direct */
331#ifdef HAVE_NATIVE_ICONV
332 if (is_utf16(fromcode)) {
333 ret->direct = sys_iconv;
334 ret->cd_direct = ret->cd_push;
335 ret->cd_push = NULL;
336 return ret;
337 }
338 if (is_utf16(tocode)) {
339 ret->direct = sys_iconv;
340 ret->cd_direct = ret->cd_pull;
341 ret->cd_pull = NULL;
342 return ret;
343 }
344#endif
345
346 return ret;
347}
348
349/*
350 simple iconv_close() wrapper
351*/
352int smb_iconv_close (smb_iconv_t cd)
353{
354#ifdef HAVE_NATIVE_ICONV
355 if (cd->cd_direct) iconv_close((iconv_t)cd->cd_direct);
356 if (cd->cd_pull) iconv_close((iconv_t)cd->cd_pull);
357 if (cd->cd_push) iconv_close((iconv_t)cd->cd_push);
358#endif
359
360 SAFE_FREE(cd->from_name);
361 SAFE_FREE(cd->to_name);
362
363 memset(cd, 0, sizeof(*cd));
364 SAFE_FREE(cd);
365 return 0;
366}
367
368
369/**********************************************************************
370 the following functions implement the builtin character sets in Samba
371 and also the "test" character sets that are designed to test
372 multi-byte character set support for english users
373***********************************************************************/
374
375static size_t ascii_pull(void *cd, const char **inbuf, size_t *inbytesleft,
376 char **outbuf, size_t *outbytesleft)
377{
378 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
379 (*outbuf)[0] = (*inbuf)[0];
380 (*outbuf)[1] = 0;
381 (*inbytesleft) -= 1;
382 (*outbytesleft) -= 2;
383 (*inbuf) += 1;
384 (*outbuf) += 2;
385 }
386
387 if (*inbytesleft > 0) {
388 errno = E2BIG;
389 return -1;
390 }
391
392 return 0;
393}
394
395static size_t ascii_push(void *cd, const char **inbuf, size_t *inbytesleft,
396 char **outbuf, size_t *outbytesleft)
397{
398 int ir_count=0;
399
400 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
401 (*outbuf)[0] = (*inbuf)[0] & 0x7F;
402 if ((*inbuf)[1]) ir_count++;
403 (*inbytesleft) -= 2;
404 (*outbytesleft) -= 1;
405 (*inbuf) += 2;
406 (*outbuf) += 1;
407 }
408
409 if (*inbytesleft == 1) {
410 errno = EINVAL;
411 return -1;
412 }
413
414 if (*inbytesleft > 1) {
415 errno = E2BIG;
416 return -1;
417 }
418
419 return ir_count;
420}
421
422static size_t latin1_push(void *cd, const char **inbuf, size_t *inbytesleft,
423 char **outbuf, size_t *outbytesleft)
424{
425 int ir_count=0;
426
427 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
428 (*outbuf)[0] = (*inbuf)[0];
429 if ((*inbuf)[1]) ir_count++;
430 (*inbytesleft) -= 2;
431 (*outbytesleft) -= 1;
432 (*inbuf) += 2;
433 (*outbuf) += 1;
434 }
435
436 if (*inbytesleft == 1) {
437 errno = EINVAL;
438 return -1;
439 }
440
441 if (*inbytesleft > 1) {
442 errno = E2BIG;
443 return -1;
444 }
445
446 return ir_count;
447}
448
449static size_t ucs2hex_pull(void *cd, const char **inbuf, size_t *inbytesleft,
450 char **outbuf, size_t *outbytesleft)
451{
452 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
453 unsigned v;
454
455 if ((*inbuf)[0] != '@') {
456 /* seven bit ascii case */
457 (*outbuf)[0] = (*inbuf)[0];
458 (*outbuf)[1] = 0;
459 (*inbytesleft) -= 1;
460 (*outbytesleft) -= 2;
461 (*inbuf) += 1;
462 (*outbuf) += 2;
463 continue;
464 }
465 /* it's a hex character */
466 if (*inbytesleft < 5) {
467 errno = EINVAL;
468 return -1;
469 }
470
471 if (sscanf(&(*inbuf)[1], "%04x", &v) != 1) {
472 errno = EILSEQ;
473 return -1;
474 }
475
476 (*outbuf)[0] = v&0xff;
477 (*outbuf)[1] = v>>8;
478 (*inbytesleft) -= 5;
479 (*outbytesleft) -= 2;
480 (*inbuf) += 5;
481 (*outbuf) += 2;
482 }
483
484 if (*inbytesleft > 0) {
485 errno = E2BIG;
486 return -1;
487 }
488
489 return 0;
490}
491
492static size_t ucs2hex_push(void *cd, const char **inbuf, size_t *inbytesleft,
493 char **outbuf, size_t *outbytesleft)
494{
495 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
496 char buf[6];
497
498 if ((*inbuf)[1] == 0 &&
499 ((*inbuf)[0] & 0x80) == 0 &&
500 (*inbuf)[0] != '@') {
501 (*outbuf)[0] = (*inbuf)[0];
502 (*inbytesleft) -= 2;
503 (*outbytesleft) -= 1;
504 (*inbuf) += 2;
505 (*outbuf) += 1;
506 continue;
507 }
508 if (*outbytesleft < 5) {
509 errno = E2BIG;
510 return -1;
511 }
512 snprintf(buf, 6, "@%04x", SVAL(*inbuf, 0));
513 memcpy(*outbuf, buf, 5);
514 (*inbytesleft) -= 2;
515 (*outbytesleft) -= 5;
516 (*inbuf) += 2;
517 (*outbuf) += 5;
518 }
519
520 if (*inbytesleft == 1) {
521 errno = EINVAL;
522 return -1;
523 }
524
525 if (*inbytesleft > 1) {
526 errno = E2BIG;
527 return -1;
528 }
529
530 return 0;
531}
532
533static size_t iconv_swab(void *cd, const char **inbuf, size_t *inbytesleft,
534 char **outbuf, size_t *outbytesleft)
535{
536 int n;
537
538 n = MIN(*inbytesleft, *outbytesleft);
539
540 swab(*inbuf, *outbuf, (n&~1));
541 if (n&1) {
542 (*outbuf)[n-1] = 0;
543 }
544
545 (*inbytesleft) -= n;
546 (*outbytesleft) -= n;
547 (*inbuf) += n;
548 (*outbuf) += n;
549
550 if (*inbytesleft > 0) {
551 errno = E2BIG;
552 return -1;
553 }
554
555 return 0;
556}
557
558static size_t iconv_copy(void *cd, const char **inbuf, size_t *inbytesleft,
559 char **outbuf, size_t *outbytesleft)
560{
561 int n;
562
563 n = MIN(*inbytesleft, *outbytesleft);
564
565 memmove(*outbuf, *inbuf, n);
566
567 (*inbytesleft) -= n;
568 (*outbytesleft) -= n;
569 (*inbuf) += n;
570 (*outbuf) += n;
571
572 if (*inbytesleft > 0) {
573 errno = E2BIG;
574 return -1;
575 }
576
577 return 0;
578}
579
580static size_t utf8_pull(void *cd, const char **inbuf, size_t *inbytesleft,
581 char **outbuf, size_t *outbytesleft)
582{
583 size_t in_left=*inbytesleft, out_left=*outbytesleft;
584 const uint8 *c = (const uint8 *)*inbuf;
585 uint8 *uc = (uint8 *)*outbuf;
586
587 while (in_left >= 1 && out_left >= 2) {
588 unsigned int codepoint;
589
590 if ((c[0] & 0x80) == 0) {
591 uc[0] = c[0];
592 uc[1] = 0;
593 c += 1;
594 in_left -= 1;
595 out_left -= 2;
596 uc += 2;
597 continue;
598 }
599
600 if ((c[0] & 0xe0) == 0xc0) {
601 if (in_left < 2 ||
602 (c[1] & 0xc0) != 0x80) {
603 errno = EILSEQ;
604 goto error;
605 }
606 codepoint = (c[1]&0x3f) | ((c[0]&0x1f)<<6);
607 if (codepoint < 0x80) {
608 /* don't accept UTF-8 characters that are not minimally packed */
609 errno = EILSEQ;
610 goto error;
611 }
612 uc[1] = codepoint >> 8;
613 uc[0] = codepoint & 0xff;
614 c += 2;
615 in_left -= 2;
616 out_left -= 2;
617 uc += 2;
618 continue;
619 }
620
621 if ((c[0] & 0xf0) == 0xe0) {
622 if (in_left < 3 ||
623 (c[1] & 0xc0) != 0x80 ||
624 (c[2] & 0xc0) != 0x80) {
625 errno = EILSEQ;
626 goto error;
627 }
628 codepoint = (c[2]&0x3f) | ((c[1]&0x3f)<<6) | ((c[0]&0xf)<<12);
629 if (codepoint < 0x800) {
630 /* don't accept UTF-8 characters that are not minimally packed */
631 errno = EILSEQ;
632 goto error;
633 }
634 uc[1] = codepoint >> 8;
635 uc[0] = codepoint & 0xff;
636 c += 3;
637 in_left -= 3;
638 out_left -= 2;
639 uc += 2;
640 continue;
641 }
642
643 if ((c[0] & 0xf8) == 0xf0) {
644 if (in_left < 4 ||
645 (c[1] & 0xc0) != 0x80 ||
646 (c[2] & 0xc0) != 0x80 ||
647 (c[3] & 0xc0) != 0x80) {
648 errno = EILSEQ;
649 goto error;
650 }
651 codepoint =
652 (c[3]&0x3f) |
653 ((c[2]&0x3f)<<6) |
654 ((c[1]&0x3f)<<12) |
655 ((c[0]&0x7)<<18);
656 if (codepoint < 0x10000 || codepoint > 0x10ffff) {
657 /* don't accept UTF-8 characters that are not minimally packed */
658 errno = EILSEQ;
659 goto error;
660 }
661
662 codepoint -= 0x10000;
663
664 if (out_left < 4) {
665 errno = E2BIG;
666 goto error;
667 }
668
669 uc[0] = (codepoint>>10) & 0xFF;
670 uc[1] = (codepoint>>18) | 0xd8;
671 uc[2] = codepoint & 0xFF;
672 uc[3] = ((codepoint>>8) & 0x3) | 0xdc;
673 c += 4;
674 in_left -= 4;
675 out_left -= 4;
676 uc += 4;
677 continue;
678 }
679
680 /* we don't handle 5 byte sequences */
681 errno = EINVAL;
682 goto error;
683 }
684
685 if (in_left > 0) {
686 errno = E2BIG;
687 goto error;
688 }
689
690 *inbytesleft = in_left;
691 *outbytesleft = out_left;
692 *inbuf = (char *)c;
693 *outbuf = (char *)uc;
694 return 0;
695
696error:
697 *inbytesleft = in_left;
698 *outbytesleft = out_left;
699 *inbuf = (char *)c;
700 *outbuf = (char *)uc;
701 return -1;
702}
703
704static size_t utf8_push(void *cd, const char **inbuf, size_t *inbytesleft,
705 char **outbuf, size_t *outbytesleft)
706{
707 size_t in_left=*inbytesleft, out_left=*outbytesleft;
708 uint8 *c = (uint8 *)*outbuf;
709 const uint8 *uc = (const uint8 *)*inbuf;
710
711 while (in_left >= 2 && out_left >= 1) {
712 unsigned int codepoint;
713
714 if (uc[1] == 0 && !(uc[0] & 0x80)) {
715 /* simplest case */
716 c[0] = uc[0];
717 in_left -= 2;
718 out_left -= 1;
719 uc += 2;
720 c += 1;
721 continue;
722 }
723
724 if ((uc[1]&0xf8) == 0) {
725 /* next simplest case */
726 if (out_left < 2) {
727 errno = E2BIG;
728 goto error;
729 }
730 c[0] = 0xc0 | (uc[0]>>6) | (uc[1]<<2);
731 c[1] = 0x80 | (uc[0] & 0x3f);
732 in_left -= 2;
733 out_left -= 2;
734 uc += 2;
735 c += 2;
736 continue;
737 }
738
739 if ((uc[1] & 0xfc) == 0xdc) {
740 /* its the second part of a 4 byte sequence. Illegal */
741 if (in_left < 4) {
742 errno = EINVAL;
743 } else {
744 errno = EILSEQ;
745 }
746 goto error;
747 }
748
749 if ((uc[1] & 0xfc) != 0xd8) {
750 codepoint = uc[0] | (uc[1]<<8);
751 if (out_left < 3) {
752 errno = E2BIG;
753 goto error;
754 }
755 c[0] = 0xe0 | (codepoint >> 12);
756 c[1] = 0x80 | ((codepoint >> 6) & 0x3f);
757 c[2] = 0x80 | (codepoint & 0x3f);
758
759 in_left -= 2;
760 out_left -= 3;
761 uc += 2;
762 c += 3;
763 continue;
764 }
765
766 /* its the first part of a 4 byte sequence */
767 if (in_left < 4) {
768 errno = EINVAL;
769 goto error;
770 }
771 if ((uc[3] & 0xfc) != 0xdc) {
772 errno = EILSEQ;
773 goto error;
774 }
775 codepoint = 0x10000 + (uc[2] | ((uc[3] & 0x3)<<8) |
776 (uc[0]<<10) | ((uc[1] & 0x3)<<18));
777
778 if (out_left < 4) {
779 errno = E2BIG;
780 goto error;
781 }
782 c[0] = 0xf0 | (codepoint >> 18);
783 c[1] = 0x80 | ((codepoint >> 12) & 0x3f);
784 c[2] = 0x80 | ((codepoint >> 6) & 0x3f);
785 c[3] = 0x80 | (codepoint & 0x3f);
786
787 in_left -= 4;
788 out_left -= 4;
789 uc += 4;
790 c += 4;
791 }
792
793 if (in_left == 1) {
794 errno = EINVAL;
795 goto error;
796 }
797
798 if (in_left > 1) {
799 errno = E2BIG;
800 goto error;
801 }
802
803 *inbytesleft = in_left;
804 *outbytesleft = out_left;
805 *inbuf = (char *)uc;
806 *outbuf = (char *)c;
807
808 return 0;
809
810error:
811 *inbytesleft = in_left;
812 *outbytesleft = out_left;
813 *inbuf = (char *)uc;
814 *outbuf = (char *)c;
815 return -1;
816}
817
Note: See TracBrowser for help on using the repository browser.