size_t charset_decompose ( charset_t ch, char * src, size_t inlen, char * dst, size_t outlen)
{
char *buffer;
ucs2_t u[MAXPATHLEN];
size_t len;
size_t ilen;
if ((size_t)(-1) == (len = convert_string_allocate_internal(ch, CH_UCS2, src, inlen, &buffer)) )
return len;
ilen=sizeof(u);
if ( (size_t)-1 == (ilen = decompose_w((ucs2_t *)buffer, len, u, &ilen)) ) {
free (buffer);
return (size_t)(-1);
}
if ((size_t)(-1) == (len = convert_string_internal( CH_UCS2, ch, (char*)u, ilen, dst, outlen)) ) {
free (buffer);
return (size_t)(-1);
}
free(buffer);
return (len);
}
size_t convert_string(charset_t from, charset_t to,
void const *src, size_t srclen,
void *dest, size_t destlen)
{
size_t i_len, o_len;
ucs2_t *u;
ucs2_t buffer[MAXPATHLEN];
ucs2_t buffer2[MAXPATHLEN];
/* convert from_set to UCS2 */
if ((size_t)-1 == ( o_len = convert_string_internal( from, CH_UCS2, src, srclen,
(char*) buffer, sizeof(buffer))) ) {
LOG(log_error, logtype_default, "Conversion failed ( %s to CH_UCS2 )", charset_name(from));
return (size_t) -1;
}
/* Do pre/decomposition */
i_len = sizeof(buffer2);
u = buffer2;
if (charsets[to] && (charsets[to]->flags & CHARSET_DECOMPOSED) ) {
if ( (size_t)-1 == (i_len = decompose_w(buffer, o_len, u, &i_len)) )
return (size_t)-1;
}
else if (!charsets[from] || (charsets[from]->flags & CHARSET_DECOMPOSED)) {
if ( (size_t)-1 == (i_len = precompose_w(buffer, o_len, u, &i_len)) )
return (size_t)-1;
}
else {
u = buffer;
i_len = o_len;
}
/* Convert UCS2 to to_set */
if ((size_t)(-1) == ( o_len = convert_string_internal( CH_UCS2, to, (char*) u, i_len, dest, destlen)) ) {
LOG(log_error, logtype_default, "Conversion failed (CH_UCS2 to %s):%s", charset_name(to), strerror(errno));
return (size_t) -1;
}
return o_len;
}
size_t charset_strlower(charset_t ch, const char *src, size_t srclen, char *dest, size_t destlen)
{
size_t size;
char *buffer;
size = convert_string_allocate_internal(ch, CH_UCS2, src, srclen,
(char **) &buffer);
if (size == (size_t)-1) {
SAFE_FREE(buffer);
return size;
}
if (!strlower_w((ucs2_t *)buffer) && (dest == src)) {
free(buffer);
return srclen;
}
size = convert_string_internal(CH_UCS2, ch, buffer, size, dest, destlen);
free(buffer);
return size;
}
size_t convert_string(charset_t from, charset_t to,
void const *src, size_t srclen,
void *dest, size_t destlen, BOOL allow_bad_conv)
{
/*
* NB. We deliberately don't do a strlen here if srclen == -1.
* This is very expensive over millions of calls and is taken
* care of in the slow path in convert_string_internal. JRA.
*/
#ifdef DEVELOPER
SMB_ASSERT(destlen != (size_t)-1);
#endif
if (srclen == 0)
return 0;
if (from != CH_UTF16LE && from != CH_UTF16BE && to != CH_UTF16LE && to != CH_UTF16BE) {
const unsigned char *p = (const unsigned char *)src;
unsigned char *q = (unsigned char *)dest;
size_t slen = srclen;
size_t dlen = destlen;
unsigned char lastp = '\0';
size_t retval = 0;
/* If all characters are ascii, fast path here. */
while (slen && dlen) {
if ((lastp = *p) <= 0x7f) {
*q++ = *p++;
if (slen != (size_t)-1) {
slen--;
}
dlen--;
retval++;
if (!lastp)
break;
} else {
#ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
goto general_case;
#else
return retval + convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
#endif
}
}
if (!dlen) {
/* Even if we fast path we should note if we ran out of room. */
if (((slen != (size_t)-1) && slen) ||
((slen == (size_t)-1) && lastp)) {
errno = E2BIG;
}
}
return retval;
} else if (from == CH_UTF16LE && to != CH_UTF16LE) {
const unsigned char *p = (const unsigned char *)src;
unsigned char *q = (unsigned char *)dest;
size_t retval = 0;
size_t slen = srclen;
size_t dlen = destlen;
unsigned char lastp = '\0';
/* If all characters are ascii, fast path here. */
while (((slen == (size_t)-1) || (slen >= 2)) && dlen) {
if (((lastp = *p) <= 0x7f) && (p[1] == 0)) {
*q++ = *p;
if (slen != (size_t)-1) {
slen -= 2;
}
p += 2;
dlen--;
retval++;
if (!lastp)
break;
} else {
#ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
goto general_case;
#else
return retval + convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
#endif
}
}
if (!dlen) {
/* Even if we fast path we should note if we ran out of room. */
if (((slen != (size_t)-1) && slen) ||
((slen == (size_t)-1) && lastp)) {
errno = E2BIG;
}
}
return retval;
} else if (from != CH_UTF16LE && from != CH_UTF16BE && to == CH_UTF16LE) {
const unsigned char *p = (const unsigned char *)src;
unsigned char *q = (unsigned char *)dest;
size_t retval = 0;
size_t slen = srclen;
size_t dlen = destlen;
unsigned char lastp = '\0';
/* If all characters are ascii, fast path here. */
while (slen && (dlen >= 2)) {
if ((lastp = *p) <= 0x7F) {
*q++ = *p++;
*q++ = '\0';
if (slen != (size_t)-1) {
slen--;
}
dlen -= 2;
retval += 2;
if (!lastp)
break;
} else {
#ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
goto general_case;
#else
return retval + convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
#endif
}
}
if (!dlen) {
/* Even if we fast path we should note if we ran out of room. */
if (((slen != (size_t)-1) && slen) ||
((slen == (size_t)-1) && lastp)) {
errno = E2BIG;
}
}
return retval;
}
#ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
general_case:
#endif
return convert_string_internal(from, to, src, srclen, dest, destlen, allow_bad_conv);
}
