2 Unix SMB/CIFS implementation.
3 minimal iconv implementation
4 Copyright (C) Andrew Tridgell 2001
5 Copyright (C) Jelmer Vernooij 2002
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.
24 static size_t ascii_pull(void *,char **, size_t *, char **, size_t *);
25 static size_t ascii_push(void *,char **, size_t *, char **, size_t *);
26 static size_t utf8_pull(void *,char **, size_t *, char **, size_t *);
27 static size_t utf8_push(void *,char **, size_t *, char **, size_t *);
28 static size_t weird_pull(void *,char **, size_t *, char **, size_t *);
29 static size_t weird_push(void *,char **, size_t *, char **, size_t *);
30 static size_t ucs2hex_pull(void *,char **, size_t *, char **, size_t *);
31 static size_t ucs2hex_push(void *,char **, size_t *, char **, size_t *);
32 static size_t iconv_copy(void *,char **, size_t *, char **, size_t *);
34 struct charset_functions builtin_functions[] = {
35 {"UCS-2LE", iconv_copy, iconv_copy},
36 {"UTF8", utf8_pull, utf8_push},
37 {"ASCII", ascii_pull, ascii_push},
38 {"WEIRD", weird_pull, weird_push},
39 {"UCS2-HEX", ucs2hex_pull, ucs2hex_push},
43 static struct charset_functions *charsets = NULL;
45 BOOL smb_register_charset(struct charset_functions *funcs)
47 struct charset_functions *c = charsets;
49 DEBUG(5, ("Attempting to register new charset %s\n", funcs->name));
50 /* Check whether we already have this charset... */
52 if(!strcasecmp(c->name, funcs->name)){
53 DEBUG(2, ("Duplicate charset %s, not registering\n", funcs->name));
59 funcs->next = funcs->prev = NULL;
60 DEBUG(5, ("Registered charset %s\n", c->name));
61 DLIST_ADD(charsets, funcs);
65 void lazy_initialize_iconv(void)
67 static BOOL initialized = False;
72 for(i = 0; builtin_functions[i].name; i++)
73 smb_register_charset(&builtin_functions[i]);
77 /* if there was an error then reset the internal state,
78 this ensures that we don't have a shift state remaining for
79 character sets like SJIS */
80 static size_t sys_iconv(void *cd,
81 char **inbuf, size_t *inbytesleft,
82 char **outbuf, size_t *outbytesleft)
84 #ifdef HAVE_NATIVE_ICONV
85 size_t ret = iconv((iconv_t)cd,
87 outbuf, outbytesleft);
88 if (ret == (size_t)-1) iconv(cd, NULL, NULL, NULL, NULL);
97 this is a simple portable iconv() implementaion. It only knows about
98 a very small number of character sets - just enough that Samba works
99 on systems that don't have iconv
101 size_t smb_iconv(smb_iconv_t cd,
102 const char **inbuf, size_t *inbytesleft,
103 char **outbuf, size_t *outbytesleft)
109 /* in many cases we can go direct */
111 return cd->direct(cd->cd_direct,
112 (char **)inbuf, inbytesleft, outbuf, outbytesleft);
116 /* otherwise we have to do it chunks at a time */
117 while (*inbytesleft > 0) {
119 bufsize = sizeof(cvtbuf);
121 if (cd->pull(cd->cd_pull,
122 (char **)inbuf, inbytesleft, &bufp, &bufsize) == -1
123 && errno != E2BIG) return -1;
126 bufsize = sizeof(cvtbuf) - bufsize;
128 if (cd->push(cd->cd_push,
130 outbuf, outbytesleft) == -1) return -1;
137 simple iconv_open() wrapper
139 smb_iconv_t smb_iconv_open(const char *tocode, const char *fromcode)
142 struct charset_functions *from, *to;
144 lazy_initialize_iconv();
148 ret = (smb_iconv_t)malloc(sizeof(*ret));
151 return (smb_iconv_t)-1;
153 memset(ret, 0, sizeof(*ret));
155 ret->from_name = strdup(fromcode);
156 ret->to_name = strdup(tocode);
158 /* check for the simplest null conversion */
159 if (strcmp(fromcode, tocode) == 0) {
160 ret->direct = iconv_copy;
165 if (strcasecmp(from->name, fromcode) == 0) break;
170 if (strcasecmp(to->name, tocode) == 0) break;
174 #ifdef HAVE_NATIVE_ICONV
176 ret->pull = sys_iconv;
177 ret->cd_pull = iconv_open("UCS-2LE", fromcode);
178 if (ret->cd_pull == (iconv_t)-1) goto failed;
182 ret->push = sys_iconv;
183 ret->cd_push = iconv_open(tocode, "UCS-2LE");
184 if (ret->cd_push == (iconv_t)-1) goto failed;
192 /* check for conversion to/from ucs2 */
193 if (strcasecmp(fromcode, "UCS-2LE") == 0 && to) {
194 ret->direct = to->push;
197 if (strcasecmp(tocode, "UCS-2LE") == 0 && from) {
198 ret->direct = from->pull;
202 #ifdef HAVE_NATIVE_ICONV
203 if (strcasecmp(fromcode, "UCS-2LE") == 0) {
204 ret->direct = sys_iconv;
205 ret->cd_direct = ret->cd_push;
209 if (strcasecmp(tocode, "UCS-2LE") == 0) {
210 ret->direct = sys_iconv;
211 ret->cd_direct = ret->cd_pull;
217 /* the general case has to go via a buffer */
218 if (!ret->pull) ret->pull = from->pull;
219 if (!ret->push) ret->push = to->push;
225 return (smb_iconv_t)-1;
229 simple iconv_close() wrapper
231 int smb_iconv_close (smb_iconv_t cd)
233 #ifdef HAVE_NATIVE_ICONV
234 if (cd->cd_direct) iconv_close((iconv_t)cd->cd_direct);
235 if (cd->cd_pull) iconv_close((iconv_t)cd->cd_pull);
236 if (cd->cd_push) iconv_close((iconv_t)cd->cd_push);
239 SAFE_FREE(cd->from_name);
240 SAFE_FREE(cd->to_name);
242 memset(cd, 0, sizeof(*cd));
248 /**********************************************************************
249 the following functions implement the builtin character sets in Samba
250 and also the "test" character sets that are designed to test
251 multi-byte character set support for english users
252 ***********************************************************************/
254 static size_t ascii_pull(void *cd, char **inbuf, size_t *inbytesleft,
255 char **outbuf, size_t *outbytesleft)
257 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
258 (*outbuf)[0] = (*inbuf)[0];
261 (*outbytesleft) -= 2;
266 if (*inbytesleft > 0) {
274 static size_t ascii_push(void *cd, char **inbuf, size_t *inbytesleft,
275 char **outbuf, size_t *outbytesleft)
279 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
280 (*outbuf)[0] = (*inbuf)[0] & 0x7F;
281 if ((*inbuf)[1]) ir_count++;
283 (*outbytesleft) -= 1;
288 if (*inbytesleft == 1) {
293 if (*inbytesleft > 1) {
302 static size_t ucs2hex_pull(void *cd, char **inbuf, size_t *inbytesleft,
303 char **outbuf, size_t *outbytesleft)
305 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
308 if ((*inbuf)[0] != '@') {
309 /* seven bit ascii case */
310 (*outbuf)[0] = (*inbuf)[0];
313 (*outbytesleft) -= 2;
318 /* it's a hex character */
319 if (*inbytesleft < 5) {
324 if (sscanf(&(*inbuf)[1], "%04x", &v) != 1) {
329 (*outbuf)[0] = v&0xff;
332 (*outbytesleft) -= 2;
337 if (*inbytesleft > 0) {
345 static size_t ucs2hex_push(void *cd, char **inbuf, size_t *inbytesleft,
346 char **outbuf, size_t *outbytesleft)
348 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
351 if ((*inbuf)[1] == 0 &&
352 ((*inbuf)[0] & 0x80) == 0 &&
353 (*inbuf)[0] != '@') {
354 (*outbuf)[0] = (*inbuf)[0];
356 (*outbytesleft) -= 1;
361 if (*outbytesleft < 5) {
365 snprintf(buf, 6, "@%04x", SVAL(*inbuf, 0));
366 memcpy(*outbuf, buf, 5);
368 (*outbytesleft) -= 5;
373 if (*inbytesleft == 1) {
378 if (*inbytesleft > 1) {
387 /* the "weird" character set is very useful for testing multi-byte
388 support and finding bugs. Don't use on a production system!
400 static size_t weird_pull(void *cd, char **inbuf, size_t *inbytesleft,
401 char **outbuf, size_t *outbytesleft)
403 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
406 for (i=0;weird_table[i].from;i++) {
407 if (strncmp((*inbuf),
409 weird_table[i].len) == 0) {
410 if (*inbytesleft < weird_table[i].len) {
411 DEBUG(0,("ERROR: truncated weird string\n"));
412 /* smb_panic("weird_pull"); */
415 (*outbuf)[0] = weird_table[i].from;
417 (*inbytesleft) -= weird_table[i].len;
418 (*outbytesleft) -= 2;
419 (*inbuf) += weird_table[i].len;
427 (*outbuf)[0] = (*inbuf)[0];
430 (*outbytesleft) -= 2;
435 if (*inbytesleft > 0) {
443 static size_t weird_push(void *cd, char **inbuf, size_t *inbytesleft,
444 char **outbuf, size_t *outbytesleft)
448 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
451 for (i=0;weird_table[i].from;i++) {
452 if ((*inbuf)[0] == weird_table[i].from &&
454 if (*outbytesleft < weird_table[i].len) {
455 DEBUG(0,("No room for weird character\n"));
456 /* smb_panic("weird_push"); */
458 memcpy(*outbuf, weird_table[i].to,
461 (*outbytesleft) -= weird_table[i].len;
463 (*outbuf) += weird_table[i].len;
471 (*outbuf)[0] = (*inbuf)[0];
472 if ((*inbuf)[1]) ir_count++;
474 (*outbytesleft) -= 1;
479 if (*inbytesleft == 1) {
484 if (*inbytesleft > 1) {
492 static size_t iconv_copy(void *cd, char **inbuf, size_t *inbytesleft,
493 char **outbuf, size_t *outbytesleft)
497 n = MIN(*inbytesleft, *outbytesleft);
499 memmove(*outbuf, *inbuf, n);
502 (*outbytesleft) -= n;
506 if (*inbytesleft > 0) {
514 static size_t utf8_pull(void *cd, char **inbuf, size_t *inbytesleft,
515 char **outbuf, size_t *outbytesleft)
517 while (*inbytesleft >= 1 && *outbytesleft >= 2) {
518 unsigned char *c = (unsigned char *)*inbuf;
519 unsigned char *uc = (unsigned char *)*outbuf;
522 if ((c[0] & 0x80) == 0) {
525 } else if ((c[0] & 0xf0) == 0xe0) {
526 if (*inbytesleft < 3) {
527 DEBUG(0,("short utf8 char\n"));
530 uc[1] = ((c[0]&0xF)<<4) | ((c[1]>>2)&0xF);
531 uc[0] = (c[1]<<6) | (c[2]&0x3f);
533 } else if ((c[0] & 0xe0) == 0xc0) {
534 if (*inbytesleft < 2) {
535 DEBUG(0,("short utf8 char\n"));
538 uc[1] = (c[0]>>2) & 0x7;
539 uc[0] = (c[0]<<6) | (c[1]&0x3f);
544 (*inbytesleft) -= len;
545 (*outbytesleft) -= 2;
549 if (*inbytesleft > 0) {
561 static size_t utf8_push(void *cd, char **inbuf, size_t *inbytesleft,
562 char **outbuf, size_t *outbytesleft)
564 while (*inbytesleft >= 2 && *outbytesleft >= 1) {
565 unsigned char *c = (unsigned char *)*outbuf;
566 unsigned char *uc = (unsigned char *)*inbuf;
570 if (*outbytesleft < 3) {
571 DEBUG(0,("short utf8 write\n"));
574 c[0] = 0xe0 | (uc[1]>>4);
575 c[1] = 0x80 | ((uc[1]&0xF)<<2) | (uc[0]>>6);
576 c[2] = 0x80 | (uc[0]&0x3f);
578 } else if (uc[1] | (uc[0] & 0x80)) {
579 if (*outbytesleft < 2) {
580 DEBUG(0,("short utf8 write\n"));
583 c[0] = 0xc0 | (uc[1]<<2) | (uc[0]>>6);
584 c[1] = 0x80 | (uc[0]&0x3f);
592 (*outbytesleft) -= len;
597 if (*inbytesleft == 1) {
602 if (*inbytesleft > 1) {