2 Unix SMB/CIFS implementation.
3 minimal iconv implementation
4 Copyright (C) Andrew Tridgell 2001
5 Copyright (C) Jelmer Vernooij 2002,2003
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 2 of the License, or
10 (at your option) any later version.
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.
17 You should have received a copy of the GNU General Public License
18 along with this program; if not, write to the Free Software
19 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
25 * We have to use strcasecmp here as the character conversions
26 * haven't been initialised yet. JRA.
34 * @brief Samba wrapper/stub for iconv character set conversion.
36 * iconv is the XPG2 interface for converting between character
37 * encodings. This file provides a Samba wrapper around it, and also
38 * a simple reimplementation that is used if the system does not
41 * Samba only works with encodings that are supersets of ASCII: ascii
42 * characters like whitespace can be tested for directly, multibyte
43 * sequences start with a byte with the high bit set, and strings are
44 * terminated by a nul byte.
46 * Note that the only function provided by iconv is conversion between
47 * characters. It doesn't directly support operations like
48 * uppercasing or comparison. We have to convert to UCS-2 and compare
51 * @sa Samba Developers Guide
54 static size_t ascii_pull(void *,char **, size_t *, char **, size_t *);
55 static size_t ascii_push(void *,char **, size_t *, char **, size_t *);
56 static size_t latin1_push(void *,char **, size_t *, char **, size_t *);
57 static size_t utf8_pull(void *,char **, size_t *, char **, size_t *);
58 static size_t utf8_push(void *,char **, size_t *, char **, size_t *);
59 static size_t ucs2hex_pull(void *,char **, size_t *, char **, size_t *);
60 static size_t ucs2hex_push(void *,char **, size_t *, char **, size_t *);
61 static size_t iconv_copy(void *,char **, size_t *, char **, size_t *);
63 static struct charset_functions builtin_functions[] = {
64 {"UCS-2LE", iconv_copy, iconv_copy},
65 {"UTF8", utf8_pull, utf8_push},
66 {"ASCII", ascii_pull, ascii_push},
67 {"646", ascii_pull, ascii_push},
68 {"ISO-8859-1", ascii_pull, latin1_push},
69 {"UCS2-HEX", ucs2hex_pull, ucs2hex_push},
73 static struct charset_functions *charsets = NULL;
75 static struct charset_functions *find_charset_functions(const char *name)
77 struct charset_functions *c = charsets;
80 if (strcasecmp(name, c->name) == 0) {
89 NTSTATUS smb_register_charset(struct charset_functions *funcs)
92 return NT_STATUS_INVALID_PARAMETER;
95 DEBUG(5, ("Attempting to register new charset %s\n", funcs->name));
96 /* Check whether we already have this charset... */
97 if (find_charset_functions(funcs->name)) {
98 DEBUG(0, ("Duplicate charset %s, not registering\n", funcs->name));
99 return NT_STATUS_OBJECT_NAME_COLLISION;
102 funcs->next = funcs->prev = NULL;
103 DEBUG(5, ("Registered charset %s\n", funcs->name));
104 DLIST_ADD(charsets, funcs);
108 void lazy_initialize_iconv(void)
110 static BOOL initialized;
115 for(i = 0; builtin_functions[i].name; i++)
116 smb_register_charset(&builtin_functions[i]);
121 /* if there was an error then reset the internal state,
122 this ensures that we don't have a shift state remaining for
123 character sets like SJIS */
124 static size_t sys_iconv(void *cd,
125 char **inbuf, size_t *inbytesleft,
126 char **outbuf, size_t *outbytesleft)
128 #ifdef HAVE_NATIVE_ICONV
129 size_t ret = iconv((iconv_t)cd,
131 outbuf, outbytesleft);
132 if (ret == (size_t)-1) iconv(cd, NULL, NULL, NULL, NULL);
141 * This is a simple portable iconv() implementaion.
143 * It only knows about a very small number of character sets - just
144 * enough that Samba works on systems that don't have iconv.
146 size_t smb_iconv(smb_iconv_t cd,
147 char **inbuf, size_t *inbytesleft,
148 char **outbuf, size_t *outbytesleft)
154 /* in many cases we can go direct */
156 return cd->direct(cd->cd_direct,
157 (char **)inbuf, inbytesleft, outbuf, outbytesleft);
161 /* otherwise we have to do it chunks at a time */
162 while (*inbytesleft > 0) {
164 bufsize = sizeof(cvtbuf);
166 if (cd->pull(cd->cd_pull,
167 (char **)inbuf, inbytesleft, &bufp, &bufsize) == -1
168 && errno != E2BIG) return -1;
171 bufsize = sizeof(cvtbuf) - bufsize;
173 if (cd->push(cd->cd_push,
175 outbuf, outbytesleft) == -1) return -1;
182 simple iconv_open() wrapper
184 smb_iconv_t smb_iconv_open(const char *tocode, const char *fromcode)
187 struct charset_functions *from, *to;
189 lazy_initialize_iconv();
193 ret = (smb_iconv_t)malloc(sizeof(*ret));
196 return (smb_iconv_t)-1;
198 memset(ret, 0, sizeof(*ret));
200 ret->from_name = strdup(fromcode);
201 ret->to_name = strdup(tocode);
203 /* check for the simplest null conversion */
204 if (strcasecmp(fromcode, tocode) == 0) {
205 ret->direct = iconv_copy;
209 /* check if we have a builtin function for this conversion */
210 from = find_charset_functions(fromcode);
211 if(from)ret->pull = from->pull;
213 to = find_charset_functions(tocode);
214 if(to)ret->push = to->push;
216 /* check if we can use iconv for this conversion */
217 #ifdef HAVE_NATIVE_ICONV
219 ret->cd_pull = iconv_open("UCS-2LE", fromcode);
220 if (ret->cd_pull != (iconv_t)-1)
221 ret->pull = sys_iconv;
225 ret->cd_push = iconv_open(tocode, "UCS-2LE");
226 if (ret->cd_push != (iconv_t)-1)
227 ret->push = sys_iconv;
231 /* check if there is a module available that can do this conversion */
232 if (!ret->pull && NT_STATUS_IS_OK(smb_probe_module("charset", fromcode))) {
233 if(!(from = find_charset_functions(fromcode)))
234 DEBUG(0, ("Module %s doesn't provide charset %s!\n", fromcode, fromcode));
236 ret->pull = from->pull;
239 if (!ret->push && NT_STATUS_IS_OK(smb_probe_module("charset", tocode))) {
240 if(!(to = find_charset_functions(tocode)))
241 DEBUG(0, ("Module %s doesn't provide charset %s!\n", tocode, tocode));
243 ret->push = to->push;
246 if (!ret->push || !ret->pull) {
247 SAFE_FREE(ret->from_name);
248 SAFE_FREE(ret->to_name);
251 return (smb_iconv_t)-1;
254 /* check for conversion to/from ucs2 */
255 if (strcasecmp(fromcode, "UCS-2LE") == 0 && to) {
256 ret->direct = to->push;
257 ret->push = ret->pull = NULL;
261 if (strcasecmp(tocode, "UCS-2LE") == 0 && from) {
262 ret->direct = from->pull;
263 ret->push = ret->pull = NULL;
267 /* Check if we can do the conversion direct */
268 #ifdef HAVE_NATIVE_ICONV
269 if (strcasecmp(fromcode, "UCS-2LE") == 0) {
270 ret->direct = sys_iconv;
271 ret->cd_direct = ret->cd_push;
275 if (strcasecmp(tocode, "UCS-2LE") == 0) {
276 ret->direct = sys_iconv;
277 ret->cd_direct = ret->cd_pull;
287 simple iconv_close() wrapper
289 int smb_iconv_close (smb_iconv_t cd)
291 #ifdef HAVE_NATIVE_ICONV
292 if (cd->cd_direct) iconv_close((iconv_t)cd->cd_direct);
293 if (cd->cd_pull) iconv_close((iconv_t)cd->cd_pull);
294 if (cd->cd_push) iconv_close((iconv_t)cd->cd_push);
297 SAFE_FREE(cd->from_name);
298 SAFE_FREE(cd->to_name);
300 memset(cd, 0, sizeof(*cd));
306 /**********************************************************************
307 the following functions implement the builtin character sets in Samba
308 and also the "test" character sets that are designed to test
309 multi-byte character set support for english users
310 ***********************************************************************/
312 static size_t ascii_pull(void *cd, char **inbuf, size_t *inbytesleft,
313 char **outbuf, size_t *outbytesleft)
315 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
316 (*outbuf)[0] = (*inbuf)[0];
319 (*outbytesleft) -= 2;
324 if (*inbytesleft > 0) {
332 static size_t ascii_push(void *cd, char **inbuf, size_t *inbytesleft,
333 char **outbuf, size_t *outbytesleft)
337 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
338 (*outbuf)[0] = (*inbuf)[0] & 0x7F;
339 if ((*inbuf)[1]) ir_count++;
341 (*outbytesleft) -= 1;
346 if (*inbytesleft == 1) {
351 if (*inbytesleft > 1) {
359 static size_t latin1_push(void *cd, char **inbuf, size_t *inbytesleft,
360 char **outbuf, size_t *outbytesleft)
364 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
365 (*outbuf)[0] = (*inbuf)[0];
366 if ((*inbuf)[1]) ir_count++;
368 (*outbytesleft) -= 1;
373 if (*inbytesleft == 1) {
378 if (*inbytesleft > 1) {
386 static size_t ucs2hex_pull(void *cd, char **inbuf, size_t *inbytesleft,
387 char **outbuf, size_t *outbytesleft)
389 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
392 if ((*inbuf)[0] != '@') {
393 /* seven bit ascii case */
394 (*outbuf)[0] = (*inbuf)[0];
397 (*outbytesleft) -= 2;
402 /* it's a hex character */
403 if (*inbytesleft < 5) {
408 if (sscanf(&(*inbuf)[1], "%04x", &v) != 1) {
413 (*outbuf)[0] = v&0xff;
416 (*outbytesleft) -= 2;
421 if (*inbytesleft > 0) {
429 static size_t ucs2hex_push(void *cd, char **inbuf, size_t *inbytesleft,
430 char **outbuf, size_t *outbytesleft)
432 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
435 if ((*inbuf)[1] == 0 &&
436 ((*inbuf)[0] & 0x80) == 0 &&
437 (*inbuf)[0] != '@') {
438 (*outbuf)[0] = (*inbuf)[0];
440 (*outbytesleft) -= 1;
445 if (*outbytesleft < 5) {
449 snprintf(buf, 6, "@%04x", SVAL(*inbuf, 0));
450 memcpy(*outbuf, buf, 5);
452 (*outbytesleft) -= 5;
457 if (*inbytesleft == 1) {
462 if (*inbytesleft > 1) {
471 static size_t iconv_copy(void *cd, char **inbuf, size_t *inbytesleft,
472 char **outbuf, size_t *outbytesleft)
476 n = MIN(*inbytesleft, *outbytesleft);
478 memmove(*outbuf, *inbuf, n);
481 (*outbytesleft) -= n;
485 if (*inbytesleft > 0) {
493 static size_t utf8_pull(void *cd, char **inbuf, size_t *inbytesleft,
494 char **outbuf, size_t *outbytesleft)
496 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
497 unsigned char *c = (unsigned char *)*inbuf;
498 unsigned char *uc = (unsigned char *)*outbuf;
501 if ((c[0] & 0x80) == 0) {
504 } else if ((c[0] & 0xf0) == 0xe0) {
505 if (*inbytesleft < 3) {
506 DEBUG(0,("short utf8 char\n"));
509 uc[1] = ((c[0]&0xF)<<4) | ((c[1]>>2)&0xF);
510 uc[0] = (c[1]<<6) | (c[2]&0x3f);
512 } else if ((c[0] & 0xe0) == 0xc0) {
513 if (*inbytesleft < 2) {
514 DEBUG(0,("short utf8 char\n"));
517 uc[1] = (c[0]>>2) & 0x7;
518 uc[0] = (c[0]<<6) | (c[1]&0x3f);
523 (*inbytesleft) -= len;
524 (*outbytesleft) -= 2;
528 if (*inbytesleft > 0) {
540 static size_t utf8_push(void *cd, char **inbuf, size_t *inbytesleft,
541 char **outbuf, size_t *outbytesleft)
543 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
544 unsigned char *c = (unsigned char *)*outbuf;
545 unsigned char *uc = (unsigned char *)*inbuf;
549 if (*outbytesleft < 3) {
550 DEBUG(0,("short utf8 write\n"));
553 c[0] = 0xe0 | (uc[1]>>4);
554 c[1] = 0x80 | ((uc[1]&0xF)<<2) | (uc[0]>>6);
555 c[2] = 0x80 | (uc[0]&0x3f);
557 } else if (uc[1] | (uc[0] & 0x80)) {
558 if (*outbytesleft < 2) {
559 DEBUG(0,("short utf8 write\n"));
562 c[0] = 0xc0 | (uc[1]<<2) | (uc[0]>>6);
563 c[1] = 0x80 | (uc[0]&0x3f);
571 (*outbytesleft) -= len;
576 if (*inbytesleft == 1) {
581 if (*inbytesleft > 1) {