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 *,const char **, size_t *, char **, size_t *);
55 static size_t ascii_push(void *,const char **, size_t *, char **, size_t *);
56 static size_t latin1_push(void *,const char **, size_t *, char **, size_t *);
57 static size_t utf8_pull(void *,const char **, size_t *, char **, size_t *);
58 static size_t utf8_push(void *,const char **, size_t *, char **, size_t *);
59 static size_t ucs2hex_pull(void *,const char **, size_t *, char **, size_t *);
60 static size_t ucs2hex_push(void *,const char **, size_t *, char **, size_t *);
61 static size_t iconv_copy(void *,const char **, size_t *, char **, size_t *);
62 static size_t iconv_swab (void *,const char **, size_t *, char **, size_t *);
64 static struct charset_functions builtin_functions[] = {
65 /* windows is really neither UCS-2 not UTF-16 */
66 {"UCS-2LE", iconv_copy, iconv_copy},
67 {"UTF-16LE", iconv_copy, iconv_copy},
68 {"UCS-2BE", iconv_swab, iconv_swab},
70 /* we include the UTF-8 alias to cope with differing locale settings */
71 {"UTF8", utf8_pull, utf8_push},
72 {"UTF-8", utf8_pull, utf8_push},
73 {"ASCII", ascii_pull, ascii_push},
74 {"646", ascii_pull, ascii_push},
75 {"ISO-8859-1", ascii_pull, latin1_push},
76 {"UCS2-HEX", ucs2hex_pull, ucs2hex_push},
80 static struct charset_functions *charsets = NULL;
82 static struct charset_functions *find_charset_functions(const char *name)
84 struct charset_functions *c = charsets;
87 if (strcasecmp(name, c->name) == 0) {
96 NTSTATUS smb_register_charset(struct charset_functions *funcs)
99 return NT_STATUS_INVALID_PARAMETER;
102 DEBUG(5, ("Attempting to register new charset %s\n", funcs->name));
103 /* Check whether we already have this charset... */
104 if (find_charset_functions(funcs->name)) {
105 DEBUG(0, ("Duplicate charset %s, not registering\n", funcs->name));
106 return NT_STATUS_OBJECT_NAME_COLLISION;
109 funcs->next = funcs->prev = NULL;
110 DEBUG(5, ("Registered charset %s\n", funcs->name));
111 DLIST_ADD(charsets, funcs);
115 static void lazy_initialize_iconv(void)
117 static BOOL initialized;
122 for(i = 0; builtin_functions[i].name; i++)
123 smb_register_charset(&builtin_functions[i]);
128 /* if there was an error then reset the internal state,
129 this ensures that we don't have a shift state remaining for
130 character sets like SJIS */
131 static size_t sys_iconv(void *cd,
132 const char **inbuf, size_t *inbytesleft,
133 char **outbuf, size_t *outbytesleft)
135 #ifdef HAVE_NATIVE_ICONV
136 size_t ret = iconv((iconv_t)cd,
137 (char **)inbuf, inbytesleft,
138 outbuf, outbytesleft);
139 if (ret == (size_t)-1) {
140 int saved_errno = errno;
141 iconv(cd, NULL, NULL, NULL, NULL);
152 * This is a simple portable iconv() implementaion.
154 * It only knows about a very small number of character sets - just
155 * enough that Samba works on systems that don't have iconv.
157 size_t smb_iconv(smb_iconv_t cd,
158 const char **inbuf, size_t *inbytesleft,
159 char **outbuf, size_t *outbytesleft)
165 /* in many cases we can go direct */
167 return cd->direct(cd->cd_direct,
168 inbuf, inbytesleft, outbuf, outbytesleft);
172 /* otherwise we have to do it chunks at a time */
173 while (*inbytesleft > 0) {
175 bufsize = sizeof(cvtbuf);
177 if (cd->pull(cd->cd_pull,
178 inbuf, inbytesleft, &bufp, &bufsize) == -1
179 && errno != E2BIG) return -1;
182 bufsize = sizeof(cvtbuf) - bufsize;
184 if (cd->push(cd->cd_push,
185 (const char **)&bufp, &bufsize,
186 outbuf, outbytesleft) == -1) return -1;
193 simple iconv_open() wrapper
195 smb_iconv_t smb_iconv_open(const char *tocode, const char *fromcode)
198 struct charset_functions *from, *to;
200 lazy_initialize_iconv();
204 ret = (smb_iconv_t)malloc(sizeof(*ret));
207 return (smb_iconv_t)-1;
209 memset(ret, 0, sizeof(*ret));
211 ret->from_name = strdup(fromcode);
212 ret->to_name = strdup(tocode);
214 /* check for the simplest null conversion */
215 if (strcasecmp(fromcode, tocode) == 0) {
216 ret->direct = iconv_copy;
220 /* check if we have a builtin function for this conversion */
221 from = find_charset_functions(fromcode);
222 if(from)ret->pull = from->pull;
224 to = find_charset_functions(tocode);
225 if(to)ret->push = to->push;
227 /* check if we can use iconv for this conversion */
228 #ifdef HAVE_NATIVE_ICONV
230 ret->cd_pull = iconv_open("UCS-2LE", fromcode);
231 if (ret->cd_pull != (iconv_t)-1)
232 ret->pull = sys_iconv;
236 ret->cd_push = iconv_open(tocode, "UCS-2LE");
237 if (ret->cd_push != (iconv_t)-1)
238 ret->push = sys_iconv;
242 /* check if there is a module available that can do this conversion */
243 if (!ret->pull && NT_STATUS_IS_OK(smb_probe_module("charset", fromcode))) {
244 if(!(from = find_charset_functions(fromcode)))
245 DEBUG(0, ("Module %s doesn't provide charset %s!\n", fromcode, fromcode));
247 ret->pull = from->pull;
250 if (!ret->push && NT_STATUS_IS_OK(smb_probe_module("charset", tocode))) {
251 if(!(to = find_charset_functions(tocode)))
252 DEBUG(0, ("Module %s doesn't provide charset %s!\n", tocode, tocode));
254 ret->push = to->push;
257 if (!ret->push || !ret->pull) {
258 SAFE_FREE(ret->from_name);
259 SAFE_FREE(ret->to_name);
262 return (smb_iconv_t)-1;
265 /* check for conversion to/from ucs2 */
266 if (strcasecmp(fromcode, "UCS-2LE") == 0 && to) {
267 ret->direct = to->push;
268 ret->push = ret->pull = NULL;
272 if (strcasecmp(tocode, "UCS-2LE") == 0 && from) {
273 ret->direct = from->pull;
274 ret->push = ret->pull = NULL;
278 /* Check if we can do the conversion direct */
279 #ifdef HAVE_NATIVE_ICONV
280 if (strcasecmp(fromcode, "UCS-2LE") == 0) {
281 ret->direct = sys_iconv;
282 ret->cd_direct = ret->cd_push;
286 if (strcasecmp(tocode, "UCS-2LE") == 0) {
287 ret->direct = sys_iconv;
288 ret->cd_direct = ret->cd_pull;
298 simple iconv_close() wrapper
300 int smb_iconv_close (smb_iconv_t cd)
302 #ifdef HAVE_NATIVE_ICONV
303 if (cd->cd_direct) iconv_close((iconv_t)cd->cd_direct);
304 if (cd->cd_pull) iconv_close((iconv_t)cd->cd_pull);
305 if (cd->cd_push) iconv_close((iconv_t)cd->cd_push);
308 SAFE_FREE(cd->from_name);
309 SAFE_FREE(cd->to_name);
311 memset(cd, 0, sizeof(*cd));
317 /**********************************************************************
318 the following functions implement the builtin character sets in Samba
319 and also the "test" character sets that are designed to test
320 multi-byte character set support for english users
321 ***********************************************************************/
323 static size_t ascii_pull(void *cd, const char **inbuf, size_t *inbytesleft,
324 char **outbuf, size_t *outbytesleft)
326 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
327 (*outbuf)[0] = (*inbuf)[0];
330 (*outbytesleft) -= 2;
335 if (*inbytesleft > 0) {
343 static size_t ascii_push(void *cd, const char **inbuf, size_t *inbytesleft,
344 char **outbuf, size_t *outbytesleft)
348 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
349 (*outbuf)[0] = (*inbuf)[0] & 0x7F;
350 if ((*inbuf)[1]) ir_count++;
352 (*outbytesleft) -= 1;
357 if (*inbytesleft == 1) {
362 if (*inbytesleft > 1) {
370 static size_t latin1_push(void *cd, const char **inbuf, size_t *inbytesleft,
371 char **outbuf, size_t *outbytesleft)
375 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
376 (*outbuf)[0] = (*inbuf)[0];
377 if ((*inbuf)[1]) ir_count++;
379 (*outbytesleft) -= 1;
384 if (*inbytesleft == 1) {
389 if (*inbytesleft > 1) {
397 static size_t ucs2hex_pull(void *cd, const char **inbuf, size_t *inbytesleft,
398 char **outbuf, size_t *outbytesleft)
400 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
403 if ((*inbuf)[0] != '@') {
404 /* seven bit ascii case */
405 (*outbuf)[0] = (*inbuf)[0];
408 (*outbytesleft) -= 2;
413 /* it's a hex character */
414 if (*inbytesleft < 5) {
419 if (sscanf(&(*inbuf)[1], "%04x", &v) != 1) {
424 (*outbuf)[0] = v&0xff;
427 (*outbytesleft) -= 2;
432 if (*inbytesleft > 0) {
440 static size_t ucs2hex_push(void *cd, const char **inbuf, size_t *inbytesleft,
441 char **outbuf, size_t *outbytesleft)
443 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
446 if ((*inbuf)[1] == 0 &&
447 ((*inbuf)[0] & 0x80) == 0 &&
448 (*inbuf)[0] != '@') {
449 (*outbuf)[0] = (*inbuf)[0];
451 (*outbytesleft) -= 1;
456 if (*outbytesleft < 5) {
460 snprintf(buf, 6, "@%04x", SVAL(*inbuf, 0));
461 memcpy(*outbuf, buf, 5);
463 (*outbytesleft) -= 5;
468 if (*inbytesleft == 1) {
473 if (*inbytesleft > 1) {
481 static size_t iconv_swab(void *cd, const char **inbuf, size_t *inbytesleft,
482 char **outbuf, size_t *outbytesleft)
486 n = MIN(*inbytesleft, *outbytesleft);
488 swab(*inbuf, *outbuf, (n&~1));
494 (*outbytesleft) -= n;
498 if (*inbytesleft > 0) {
506 static size_t iconv_copy(void *cd, const char **inbuf, size_t *inbytesleft,
507 char **outbuf, size_t *outbytesleft)
511 n = MIN(*inbytesleft, *outbytesleft);
513 memmove(*outbuf, *inbuf, n);
516 (*outbytesleft) -= n;
520 if (*inbytesleft > 0) {
528 static size_t utf8_pull(void *cd, const char **inbuf, size_t *inbytesleft,
529 char **outbuf, size_t *outbytesleft)
531 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
532 const unsigned char *c = (const unsigned char *)*inbuf;
533 unsigned char *uc = (unsigned char *)*outbuf;
536 if ((c[0] & 0x80) == 0) {
539 } else if ((c[0] & 0xf0) == 0xe0) {
540 if (*inbytesleft < 3) {
541 DEBUG(0,("short utf8 char\n"));
544 uc[1] = ((c[0]&0xF)<<4) | ((c[1]>>2)&0xF);
545 uc[0] = (c[1]<<6) | (c[2]&0x3f);
547 } else if ((c[0] & 0xe0) == 0xc0) {
548 if (*inbytesleft < 2) {
549 DEBUG(0,("short utf8 char\n"));
552 uc[1] = (c[0]>>2) & 0x7;
553 uc[0] = (c[0]<<6) | (c[1]&0x3f);
558 (*inbytesleft) -= len;
559 (*outbytesleft) -= 2;
563 if (*inbytesleft > 0) {
575 static size_t utf8_push(void *cd, const char **inbuf, size_t *inbytesleft,
576 char **outbuf, size_t *outbytesleft)
578 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
579 unsigned char *c = (unsigned char *)*outbuf;
580 const unsigned char *uc = (const unsigned char *)*inbuf;
584 if (*outbytesleft < 3) {
585 DEBUG(0,("short utf8 write\n"));
588 c[0] = 0xe0 | (uc[1]>>4);
589 c[1] = 0x80 | ((uc[1]&0xF)<<2) | (uc[0]>>6);
590 c[2] = 0x80 | (uc[0]&0x3f);
592 } else if (uc[1] | (uc[0] & 0x80)) {
593 if (*outbytesleft < 2) {
594 DEBUG(0,("short utf8 write\n"));
597 c[0] = 0xc0 | (uc[1]<<2) | (uc[0]>>6);
598 c[1] = 0x80 | (uc[0]&0x3f);
606 (*outbytesleft) -= len;
611 if (*inbytesleft == 1) {
616 if (*inbytesleft > 1) {