2 Unix SMB/CIFS implementation.
3 Character set conversion Extensions
4 Copyright (C) Igor Vergeichik <iverg@mail.ru> 2001
5 Copyright (C) Andrew Tridgell 2001
6 Copyright (C) Simo Sorce 2001
7 Copyright (C) Martin Pool 2003
9 This program is free software; you can redistribute it and/or modify
10 it under the terms of the GNU General Public License as published by
11 the Free Software Foundation; either version 2 of the License, or
12 (at your option) any later version.
14 This program is distributed in the hope that it will be useful,
15 but WITHOUT ANY WARRANTY; without even the implied warranty of
16 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 GNU General Public License for more details.
19 You should have received a copy of the GNU General Public License
20 along with this program; if not, write to the Free Software
21 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
26 #endif /* HAVE_CONFIG_H */
35 #include <sys/param.h>
36 #ifdef HAVE_USABLE_ICONV
46 #include <netatalk/endian.h>
47 #include <atalk/logger.h>
48 #include <atalk/unicode.h>
49 #include <atalk/util.h>
50 #include "byteorder.h"
56 * @brief Character-set conversion routines built on our iconv.
58 * @note Samba's internal character set (at least in the 3.0 series)
59 * is always the same as the one for the Unix filesystem. It is
60 * <b>not</b> necessarily UTF-8 and may be different on machines that
61 * need i18n filenames to be compatible with Unix software. It does
62 * have to be a superset of ASCII. All multibyte sequences must start
63 * with a byte with the high bit set.
69 #define MAX_CHARSETS 10
71 #define CHECK_FLAGS(a,b) (((a)!=NULL) ? (*(a) & (b)) : 0 )
73 static atalk_iconv_t conv_handles[MAX_CHARSETS][MAX_CHARSETS];
74 static char* charset_names[MAX_CHARSETS];
75 static struct charset_functions* charsets[MAX_CHARSETS];
76 static char hexdig[] = "0123456789abcdef";
77 #define hextoint( c ) ( isdigit( c ) ? c - '0' : c + 10 - 'a' )
79 static char* read_charsets_from_env(charset_t ch)
85 if (( name = getenv( "ATALK_MAC_CHARSET" )) != NULL )
91 if (( name = getenv( "ATALK_UNIX_CHARSET" )) != NULL )
104 * Return the name of a charset to give to iconv().
106 static const char *charset_name(charset_t ch)
108 const char *ret = NULL;
109 static int first = 1;
110 static char macname[128];
111 static char unixname[128];
114 memset(macname, 0, sizeof(macname));
115 memset(unixname, 0, sizeof(unixname));
119 if (ch == CH_UCS2) ret = "UCS-2";
120 else if (ch == CH_UTF8) ret = "UTF8";
121 else if (ch == CH_UTF8_MAC) ret = "UTF8-MAC";
122 else if (ch == CH_UNIX) {
123 if (unixname[0] == '\0') {
124 ret = read_charsets_from_env(CH_UNIX);
125 strlcpy(unixname, ret, sizeof(unixname));
130 else if (ch == CH_MAC) {
131 if (macname[0] == '\0') {
132 ret = read_charsets_from_env(CH_MAC);
133 strlcpy(macname, ret, sizeof(macname));
140 ret = charset_names[ch];
142 #if defined(HAVE_NL_LANGINFO) && defined(CODESET)
143 if (ret && strcasecmp(ret, "LOCALE") == 0) {
144 const char *ln = NULL;
146 #ifdef HAVE_SETLOCALE
147 setlocale(LC_ALL, "");
149 ln = nl_langinfo(CODESET);
151 /* Check whether the charset name is supported
153 atalk_iconv_t handle = atalk_iconv_open(ln, "UCS-2");
154 if (handle == (atalk_iconv_t) -1) {
155 LOG(log_debug, logtype_default, "Locale charset '%s' unsupported, using ASCII instead", ln);
158 atalk_iconv_close(handle);
161 strlcpy(unixname, ln, sizeof(unixname));
165 #else /* system doesn't have LOCALE support */
166 if (ch == CH_UNIX) ret = NULL;
169 if (!ret || !*ret) ret = "ASCII";
173 struct charset_functions* get_charset_functions (charset_t ch)
175 if (charsets[ch] != NULL)
178 charsets[ch] = find_charset_functions(charset_name(ch));
184 void lazy_initialize_conv(void)
186 static int initialized = 0;
194 charset_t add_charset(char* name)
196 static charset_t max_charset_t = NUM_CHARSETS-1;
197 charset_t cur_charset_t = max_charset_t+1;
200 lazy_initialize_conv();
202 for (c1=0; c1<=max_charset_t;c1++) {
203 if ( strcasecmp(name, charset_name(c1)) == 0)
207 if ( cur_charset_t >= MAX_CHARSETS ) {
208 LOG (log_debug, logtype_default, "Adding charset %s failed, too many charsets (max. %u allowed)",
210 return (charset_t) -1;
213 /* First try to setup the required conversions */
215 conv_handles[cur_charset_t][CH_UCS2] = atalk_iconv_open( charset_name(CH_UCS2), name);
216 if (conv_handles[cur_charset_t][CH_UCS2] == (atalk_iconv_t)-1) {
217 LOG(log_error, logtype_default, "Required conversion from %s to %s not supported",
218 name, charset_name(CH_UCS2));
219 conv_handles[cur_charset_t][CH_UCS2] = NULL;
220 return (charset_t) -1;
223 conv_handles[CH_UCS2][cur_charset_t] = atalk_iconv_open( name, charset_name(CH_UCS2));
224 if (conv_handles[CH_UCS2][cur_charset_t] == (atalk_iconv_t)-1) {
225 LOG(log_error, logtype_default, "Required conversion from %s to %s not supported",
226 charset_name(CH_UCS2), name);
227 conv_handles[CH_UCS2][cur_charset_t] = NULL;
228 return (charset_t) -1;
231 /* register the new charset_t name */
232 charset_names[cur_charset_t] = strdup(name);
234 charsets[cur_charset_t] = get_charset_functions (cur_charset_t);
238 LOG(log_debug, logtype_default, "Added charset %s with handle %u", name, cur_charset_t);
240 return (cur_charset_t);
244 * Initialize iconv conversion descriptors.
246 * This is called the first time it is needed, and also called again
247 * every time the configuration is reloaded, because the charset or
248 * codepage might have changed.
250 void init_iconv(void)
254 for (c1=0;c1<NUM_CHARSETS;c1++) {
255 const char *name = charset_name((charset_t)c1);
257 conv_handles[c1][CH_UCS2] = atalk_iconv_open( charset_name(CH_UCS2), name);
258 if (conv_handles[c1][CH_UCS2] == (atalk_iconv_t)-1) {
259 LOG(log_error, logtype_default, "Required conversion from %s to %s not supported",
260 name, charset_name(CH_UCS2));
261 conv_handles[c1][CH_UCS2] = NULL;
264 if (c1 != CH_UCS2) { /* avoid lost memory, make valgrind happy */
265 conv_handles[CH_UCS2][c1] = atalk_iconv_open( name, charset_name(CH_UCS2));
266 if (conv_handles[CH_UCS2][c1] == (atalk_iconv_t)-1) {
267 LOG(log_error, logtype_default, "Required conversion from %s to %s not supported",
268 charset_name(CH_UCS2), name);
269 conv_handles[CH_UCS2][c1] = NULL;
273 charsets[c1] = get_charset_functions (c1);
278 * Convert string from one encoding to another, making error checking etc
280 * @param src pointer to source string (multibyte or singlebyte)
281 * @param srclen length of the source string in bytes
282 * @param dest pointer to destination string (multibyte or singlebyte)
283 * @param destlen maximal length allowed for string
284 * @returns the number of bytes occupied in the destination
286 static size_t convert_string_internal(charset_t from, charset_t to,
287 void const *src, size_t srclen,
288 void *dest, size_t destlen)
292 const char* inbuf = (const char*)src;
293 char* outbuf = (char*)dest;
294 char* o_save = outbuf;
295 atalk_iconv_t descriptor;
297 if (srclen == (size_t)-1)
298 srclen = strlen(src)+1;
300 lazy_initialize_conv();
302 descriptor = conv_handles[from][to];
304 if (descriptor == (atalk_iconv_t)-1 || descriptor == (atalk_iconv_t)0) {
310 retval = atalk_iconv(descriptor, &inbuf, &i_len, &outbuf, &o_len);
311 if(retval==(size_t)-1) {
312 const char *reason="unknown error";
315 reason="Incomplete multibyte sequence";
318 reason="No more room";
321 reason="Illegal multibyte sequence";
324 LOG(log_debug, logtype_default,"Conversion error: %s",reason);
328 /* Terminate the string */
329 if (to == CH_UCS2 && destlen-o_len >= 2) {
330 o_save[destlen-o_len] = 0;
331 o_save[destlen-o_len+1] = 0;
333 else if ( to != CH_UCS2 && destlen-o_len > 0 )
334 o_save[destlen-o_len] = 0;
336 /* FIXME: what should we do here, string *might* be unterminated. E2BIG? */
339 return destlen-o_len;
343 size_t convert_string(charset_t from, charset_t to,
344 void const *src, size_t srclen,
345 void *dest, size_t destlen)
349 ucs2_t buffer[MAXPATHLEN];
350 ucs2_t buffer2[MAXPATHLEN];
353 lazy_initialize_conv();
355 /* convert from_set to UCS2 */
356 if ((size_t)(-1) == ( o_len = convert_string_internal( from, CH_UCS2, src, srclen,
357 (char*) buffer, sizeof(buffer))) ) {
358 LOG(log_error, logtype_default, "Conversion failed ( %s to CH_UCS2 )", charset_name(from));
362 /* Do pre/decomposition */
363 if ( ((!(charsets[to]) || !(charsets[to]->flags & CHARSET_DECOMPOSED)) &&
364 (!(charsets[from]) || (charsets[from]->flags & CHARSET_DECOMPOSED))))
366 if ((charsets[to] && charsets[to]->flags & CHARSET_DECOMPOSED) )
369 i_len = sizeof(buffer2);
372 switch (composition) {
378 if ( (size_t)-1 == (i_len = precompose_w(buffer, o_len, u, &i_len)) )
382 if ( (size_t)-1 == (i_len = decompose_w(buffer, o_len, u, &i_len)) )
387 /* Convert UCS2 to to_set */
388 if ((size_t)(-1) == ( o_len = convert_string_internal( CH_UCS2, to, (char*) u, i_len, dest, destlen)) ) {
389 LOG(log_error, logtype_default, "Conversion failed (CH_UCS2 to %s):%s", charset_name(to), strerror(errno));
399 * Convert between character sets, allocating a new buffer for the result.
401 * @param srclen length of source buffer.
402 * @param dest always set at least to NULL
403 * @note -1 is not accepted for srclen.
405 * @returns Size in bytes of the converted string; or -1 in case of error.
408 static size_t convert_string_allocate_internal(charset_t from, charset_t to,
409 void const *src, size_t srclen, char **dest)
411 size_t i_len, o_len, destlen;
413 const char *inbuf = (const char *)src;
414 char *outbuf = NULL, *ob = NULL;
415 atalk_iconv_t descriptor;
419 if (src == NULL || srclen == (size_t)-1)
422 lazy_initialize_conv();
424 descriptor = conv_handles[from][to];
426 if (descriptor == (atalk_iconv_t)-1 || descriptor == (atalk_iconv_t)0) {
427 /* conversion not supported, return -1*/
428 LOG(log_debug, logtype_default, "convert_string_allocate: conversion not supported!");
432 destlen = MAX(srclen, 512);
434 destlen = destlen * 2;
435 outbuf = (char *)realloc(ob, destlen);
437 LOG(log_debug, logtype_default,"convert_string_allocate: realloc failed!");
443 inbuf = src; /* this restarts the whole conversion if buffer needed to be increased */
446 retval = atalk_iconv(descriptor,
449 if(retval == (size_t)-1) {
450 const char *reason="unknown error";
453 reason="Incomplete multibyte sequence";
458 reason="Illegal multibyte sequence";
461 LOG(log_debug, logtype_default,"Conversion error: %s(%s)",reason,inbuf);
467 destlen = destlen - o_len;
469 /* Terminate the string */
470 if (to == CH_UCS2 && o_len >= 2) {
473 *dest = (char *)realloc(ob,destlen+2);
475 else if ( to != CH_UCS2 && o_len > 0 ) {
477 *dest = (char *)realloc(ob,destlen+1);
480 goto convert; /* realloc */
483 if (destlen && !*dest) {
484 LOG(log_debug, logtype_default, "convert_string_allocate: out of memory!");
493 size_t convert_string_allocate(charset_t from, charset_t to,
494 void const *src, size_t srclen,
499 ucs2_t buffer[MAXPATHLEN];
500 ucs2_t buffer2[MAXPATHLEN];
503 lazy_initialize_conv();
507 /* convert from_set to UCS2 */
508 if ((size_t)(-1) == ( o_len = convert_string_internal( from, CH_UCS2, src, srclen,
509 buffer, sizeof(buffer))) ) {
510 LOG(log_error, logtype_default, "Conversion failed ( %s to CH_UCS2 )", charset_name(from));
514 /* Do pre/decomposition */
515 if ( ((!(charsets[to]) || !(charsets[to]->flags & CHARSET_DECOMPOSED)) &&
516 (!(charsets[from]) || (charsets[from]->flags & CHARSET_DECOMPOSED))))
518 if ((charsets[to] && charsets[to]->flags & CHARSET_DECOMPOSED) )
521 i_len = sizeof(buffer2);
524 switch (composition) {
530 if ( (size_t)-1 == (i_len = precompose_w(buffer, o_len, u, &i_len)) )
534 if ( (size_t)-1 == (i_len = decompose_w(buffer, o_len, u, &i_len)) )
539 /* Convert UCS2 to to_set */
540 if ((size_t)(-1) == ( o_len = convert_string_allocate_internal( CH_UCS2, to, (char*)u, i_len, dest)) )
541 LOG(log_error, logtype_default, "Conversion failed (CH_UCS2 to %s):%s", charset_name(to), strerror(errno));
547 size_t charset_strupper(charset_t ch, const char *src, size_t srclen, char *dest, size_t destlen)
552 size = convert_string_allocate_internal(ch, CH_UCS2, src, srclen,
554 if (size == (size_t)-1) {
558 if (!strupper_w((ucs2_t *)buffer) && (dest == src)) {
563 size = convert_string_internal(CH_UCS2, ch, buffer, size, dest, destlen);
568 size_t charset_strlower(charset_t ch, const char *src, size_t srclen, char *dest, size_t destlen)
573 size = convert_string_allocate_internal(ch, CH_UCS2, src, srclen,
575 if (size == (size_t)-1) {
579 if (!strlower_w((ucs2_t *)buffer) && (dest == src)) {
584 size = convert_string_internal(CH_UCS2, ch, buffer, size, dest, destlen);
590 size_t unix_strupper(const char *src, size_t srclen, char *dest, size_t destlen)
592 return charset_strupper( CH_UNIX, src, srclen, dest, destlen);
595 size_t unix_strlower(const char *src, size_t srclen, char *dest, size_t destlen)
597 return charset_strlower( CH_UNIX, src, srclen, dest, destlen);
600 size_t utf8_strupper(const char *src, size_t srclen, char *dest, size_t destlen)
602 return charset_strupper( CH_UTF8, src, srclen, dest, destlen);
605 size_t utf8_strlower(const char *src, size_t srclen, char *dest, size_t destlen)
607 return charset_strlower( CH_UTF8, src, srclen, dest, destlen);
611 * Copy a string from a charset_t char* src to a UCS2 destination, allocating a buffer
613 * @param dest always set at least to NULL
615 * @returns The number of bytes occupied by the string in the destination
616 * or -1 in case of error.
619 size_t charset_to_ucs2_allocate(charset_t ch, ucs2_t **dest, const char *src)
621 size_t src_len = strlen(src);
624 return convert_string_allocate(ch, CH_UCS2, src, src_len, (char**) dest);
628 * Copy a string from a charset_t char* src to a UTF-8 destination, allocating a buffer
630 * @param dest always set at least to NULL
632 * @returns The number of bytes occupied by the string in the destination
635 size_t charset_to_utf8_allocate(charset_t ch, char **dest, const char *src)
637 size_t src_len = strlen(src);
640 return convert_string_allocate(ch, CH_UTF8, src, src_len, dest);
644 * Copy a string from a UCS2 src to a unix char * destination, allocating a buffer
646 * @param dest always set at least to NULL
648 * @returns The number of bytes occupied by the string in the destination
651 size_t ucs2_to_charset(charset_t ch, const ucs2_t *src, char *dest, size_t destlen)
653 size_t src_len = (strlen_w(src)) * sizeof(ucs2_t);
654 return convert_string(CH_UCS2, ch, src, src_len, dest, destlen);
658 size_t ucs2_to_charset_allocate(charset_t ch, char **dest, const ucs2_t *src)
660 size_t src_len = (strlen_w(src)) * sizeof(ucs2_t);
662 return convert_string_allocate(CH_UCS2, ch, src, src_len, dest);
666 * Copy a string from a UTF-8 src to a unix char * destination, allocating a buffer
668 * @param dest always set at least to NULL
670 * @returns The number of bytes occupied by the string in the destination
673 size_t utf8_to_charset_allocate(charset_t ch, char **dest, const char *src)
675 size_t src_len = strlen(src);
677 return convert_string_allocate(CH_UTF8, ch, src, src_len, dest);
680 size_t charset_precompose ( charset_t ch, char * src, size_t inlen, char * dst, size_t outlen)
683 ucs2_t u[MAXPATHLEN];
687 if ((size_t)(-1) == (len = convert_string_allocate_internal(ch, CH_UCS2, src, inlen, &buffer)) )
692 if ( (size_t)-1 == (ilen = precompose_w((ucs2_t *)buffer, len, u, &ilen)) ) {
697 if ((size_t)(-1) == (len = convert_string_internal( CH_UCS2, ch, (char*)u, ilen, dst, outlen)) ) {
707 size_t charset_decompose ( charset_t ch, char * src, size_t inlen, char * dst, size_t outlen)
710 ucs2_t u[MAXPATHLEN];
714 if ((size_t)(-1) == (len = convert_string_allocate_internal(ch, CH_UCS2, src, inlen, &buffer)) )
719 if ( (size_t)-1 == (ilen = decompose_w((ucs2_t *)buffer, len, u, &ilen)) ) {
724 if ((size_t)(-1) == (len = convert_string_internal( CH_UCS2, ch, (char*)u, ilen, dst, outlen)) ) {
734 size_t utf8_precompose ( char * src, size_t inlen, char * dst, size_t outlen)
736 return charset_precompose ( CH_UTF8, src, inlen, dst, outlen);
739 size_t utf8_decompose ( char * src, size_t inlen, char * dst, size_t outlen)
741 return charset_decompose ( CH_UTF8, src, inlen, dst, outlen);
745 static char debugbuf[ MAXPATHLEN +1 ];
746 char * debug_out ( char * seq, size_t len)
752 p = (unsigned char*) seq;
755 for ( i = 0; i<=(len-1); i++)
757 sprintf(q, "%2.2x.", *p);
768 * Convert from MB to UCS2 charset
770 * CONV_UNESCAPEHEX: ':XX' will be converted to an UCS2 character
771 * CONV_IGNORE: return the first convertable characters.
773 * This will *not* work if the destination charset is not multibyte, i.e. UCS2->UCS2 will fail
774 * The (un)escape scheme is not compatible to the old cap style escape. This is bad, we need it
775 * for e.g. HFS cdroms.
778 static size_t pull_charset_flags (charset_t from_set, charset_t cap_charset, char* src, size_t srclen, char* dest, size_t destlen, u_int16_t *flags)
780 size_t i_len, o_len, hlen;
781 size_t retval, j = 0;
782 const char* inbuf = (const char*)src;
783 char* outbuf = (char*)dest;
784 atalk_iconv_t descriptor;
785 atalk_iconv_t descriptor_cap;
790 if (srclen == (size_t)-1)
791 srclen = strlen(src)+1;
793 lazy_initialize_conv();
795 descriptor = conv_handles[from_set][CH_UCS2];
796 descriptor_cap = conv_handles[cap_charset][CH_UCS2];
798 if (descriptor == (atalk_iconv_t)-1 || descriptor == (atalk_iconv_t)0) {
807 if ( flags && (*flags & CONV_UNESCAPEHEX)) {
808 if ( NULL != (s = strchr ( inbuf, ':'))) {
809 j = i_len - (s - inbuf);
810 if ( 0 == (i_len = (s - inbuf)))
815 retval = atalk_iconv(descriptor, &inbuf, &i_len, &outbuf, &o_len);
816 if(retval==(size_t)-1) {
817 if (errno == EILSEQ && flags && (*flags & CONV_IGNORE)) {
818 *flags |= CONV_REQMANGLE;
819 return destlen-o_len;
826 if (j && flags && (*flags & CONV_UNESCAPEHEX )) {
827 /* we're at the start on an hex encoded ucs2 char */
833 isxdigit( *(inbuf+1)) && isxdigit( *(inbuf+2)) ) {
835 while ( *inbuf == ':' && j >=3 &&
836 isxdigit( *(inbuf+1)) && isxdigit( *(inbuf+2)) ) {
838 h[hlen] = hextoint( *inbuf ) << 4;
840 h[hlen++] |= hextoint( *inbuf );
844 h_buf = (const char*) h;
845 if ((size_t) -1 == (retval = atalk_iconv(descriptor_cap, &h_buf, &hlen, &outbuf, &o_len)) ) {
846 if (errno == EILSEQ && CHECK_FLAGS(flags, CONV_IGNORE)) {
847 *flags |= CONV_REQMANGLE;
848 return destlen-o_len;
856 /* We have an invalid :xx sequence */
857 if (CHECK_FLAGS(flags, CONV_IGNORE)) {
858 *flags |= CONV_REQMANGLE;
859 return destlen-o_len;
869 goto conversion_loop;
874 return destlen-o_len;
878 * Convert from UCS2 to MB charset
880 * CONV_ESCAPEDOTS: escape leading dots
881 * CONV_ESCAPEHEX: unconvertable characters and '/' will be escaped to :XX
882 * CONV_IGNORE: unconvertable characters will be replaced with '_'
884 * CONV_IGNORE and CONV_ESCAPEHEX can't work together. Should we check this ?
885 * This will *not* work if the destination charset is not multibyte, i.e. UCS2->UCS2 will fail
886 * The escape scheme is not compatible to the old cap style escape. This is bad, we need it
887 * for e.g. HFS cdroms.
891 static size_t push_charset_flags (charset_t to_set, charset_t cap_set, char* src, size_t srclen, char* dest, size_t destlen, u_int16_t *flags)
893 size_t i_len, o_len, i;
894 size_t retval, j = 0;
895 const char* inbuf = (const char*)src;
896 char* outbuf = (char*)dest;
897 atalk_iconv_t descriptor;
899 char *buf, *buf_save;
902 lazy_initialize_conv();
904 descriptor = conv_handles[CH_UCS2][to_set];
906 if (descriptor == (atalk_iconv_t)-1 || descriptor == (atalk_iconv_t)0) {
914 if ( SVAL(inbuf,0) == 0x002e && flags && (*flags & CONV_ESCAPEDOTS)) { /* 0x002e = . */
926 if (flags) *flags |= CONV_REQESCAPE;
930 if ( flags && (*flags & CONV_ESCAPEHEX)) {
931 for ( i = 0; i < i_len; i+=2) {
932 if ( SVAL((inbuf+i),0) == 0x002f) { /* 0x002f = / */
934 if ( 0 == ( i_len = i))
937 } else if ( SVAL(inbuf+i,0) == 0x003a) { /* 0x003a = : */
944 retval = atalk_iconv(descriptor, &inbuf, &i_len, &outbuf, &o_len);
945 if (retval==(size_t)-1) {
946 if (errno == EILSEQ && CHECK_FLAGS(flags, CONV_IGNORE)) {
947 *flags |= CONV_REQMANGLE;
948 return destlen -o_len;
950 else if ( errno == EILSEQ && flags && (*flags & CONV_ESCAPEHEX)) {
955 if ((size_t) -1 == (buflen = convert_string_allocate_internal(CH_UCS2, cap_set, inbuf, 2, &buf)) )
964 *outbuf++ = hexdig[ ( *buf & 0xf0 ) >> 4 ];
965 *outbuf++ = hexdig[ *buf & 0x0f ];
974 if (flags) *flags |= CONV_REQESCAPE;
976 goto conversion_loop;
983 if (j && flags && (*flags & CONV_ESCAPEHEX)) {
988 o_save[destlen -o_len] = ':';
989 o_save[destlen -o_len+1] = '2';
990 o_save[destlen -o_len+2] = 'f';
997 goto conversion_loop;
999 return destlen -o_len;
1002 size_t convert_charset ( charset_t from_set, charset_t to_set, charset_t cap_charset, char* src, size_t src_len, char* dest, size_t dest_len, u_int16_t *flags)
1004 size_t i_len, o_len;
1006 ucs2_t buffer[MAXPATHLEN];
1007 ucs2_t buffer2[MAXPATHLEN];
1008 int composition = 0;
1010 lazy_initialize_conv();
1012 /* convert from_set to UCS2 */
1013 if ((size_t)(-1) == ( o_len = pull_charset_flags( from_set, cap_charset, src, src_len,
1014 (char *) buffer, sizeof(buffer), flags)) ) {
1015 LOG(log_error, logtype_default, "Conversion failed ( %s to CH_UCS2 )", charset_name(from_set));
1022 /* Do pre/decomposition */
1023 if (CHECK_FLAGS(flags, CONV_PRECOMPOSE) ||
1024 ((!(charsets[to_set]) || !(charsets[to_set]->flags & CHARSET_DECOMPOSED)) &&
1025 (!(charsets[from_set]) || (charsets[from_set]->flags & CHARSET_DECOMPOSED))))
1027 if (CHECK_FLAGS(flags, CONV_DECOMPOSE) || (charsets[to_set] && charsets[to_set]->flags & CHARSET_DECOMPOSED) )
1030 i_len = sizeof(buffer2);
1033 switch (composition) {
1039 if ( (size_t)-1 == (i_len = precompose_w(buffer, o_len, u, &i_len)) )
1040 return (size_t)(-1);
1043 if ( (size_t)-1 == (i_len = decompose_w(buffer, o_len, u, &i_len)) )
1044 return (size_t)(-1);
1048 /* Do case conversions */
1049 if (CHECK_FLAGS(flags, CONV_TOUPPER)) {
1052 if (CHECK_FLAGS(flags, CONV_TOLOWER)) {
1056 /* Convert UCS2 to to_set */
1057 if ((size_t)(-1) == ( o_len = push_charset_flags( to_set, cap_charset, (char *)u, i_len, dest, dest_len, flags )) ) {
1058 LOG(log_error, logtype_default,
1059 "Conversion failed (CH_UCS2 to %s):%s", charset_name(to_set), strerror(errno));