This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
- the Free Software Foundation; either version 2 of the License, or
+ the Free Software Foundation; either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
*/
#include "includes.h"
+/* We can parameterize this if someone complains.... JRA. */
+
+char lp_failed_convert_char(void)
+{
+ return '_';
+}
+
/**
* @file
*
{
const char *ret = NULL;
- if (ch == CH_UCS2) ret = "UCS-2LE";
+ if (ch == CH_UTF16LE) ret = "UTF-16LE";
+ else if (ch == CH_UTF16BE) ret = "UTF-16BE";
else if (ch == CH_UNIX) ret = lp_unix_charset();
else if (ch == CH_DOS) ret = lp_dos_charset();
else if (ch == CH_DISPLAY) ret = lp_display_charset();
}
}
+/**
+ * Destroy global objects allocated by init_iconv()
+ **/
+void gfree_charcnv(void)
+{
+ int c1, c2;
+
+ for (c1=0;c1<NUM_CHARSETS;c1++) {
+ for (c2=0;c2<NUM_CHARSETS;c2++) {
+ if ( conv_handles[c1][c2] ) {
+ smb_iconv_close( conv_handles[c1][c2] );
+ conv_handles[c1][c2] = 0;
+ }
+ }
+ }
+}
+
/**
* Initialize iconv conversion descriptors.
*
/* so that charset_name() works we need to get the UNIX<->UCS2 going
first */
- if (!conv_handles[CH_UNIX][CH_UCS2])
- conv_handles[CH_UNIX][CH_UCS2] = smb_iconv_open("UCS-2LE", "ASCII");
+ if (!conv_handles[CH_UNIX][CH_UTF16LE])
+ conv_handles[CH_UNIX][CH_UTF16LE] = smb_iconv_open(charset_name(CH_UTF16LE), "ASCII");
- if (!conv_handles[CH_UCS2][CH_UNIX])
- conv_handles[CH_UCS2][CH_UNIX] = smb_iconv_open("ASCII", "UCS-2LE");
+ if (!conv_handles[CH_UTF16LE][CH_UNIX])
+ conv_handles[CH_UTF16LE][CH_UNIX] = smb_iconv_open("ASCII", charset_name(CH_UTF16LE));
for (c1=0;c1<NUM_CHARSETS;c1++) {
for (c2=0;c2<NUM_CHARSETS;c2++) {
conv_handles[c1][c2] = smb_iconv_open(n2,n1);
if (conv_handles[c1][c2] == (smb_iconv_t)-1) {
- DEBUG(0,("Conversion from %s to %s not supported\n",
+ DEBUG(0,("init_iconv: Conversion from %s to %s not supported\n",
charset_name((charset_t)c1), charset_name((charset_t)c2)));
- conv_handles[c1][c2] = NULL;
+ if (c1 != CH_UTF16LE && c1 != CH_UTF16BE) {
+ n1 = "ASCII";
+ }
+ if (c2 != CH_UTF16LE && c2 != CH_UTF16BE) {
+ n2 = "ASCII";
+ }
+ DEBUG(0,("init_iconv: Attempting to replace with conversion from %s to %s\n",
+ n1, n2 ));
+ conv_handles[c1][c2] = smb_iconv_open(n2,n1);
+ if (!conv_handles[c1][c2]) {
+ DEBUG(0,("init_iconv: Conversion from %s to %s failed", n1, n2));
+ smb_panic("init_iconv: conv_handle initialization failed");
+ }
}
}
}
descriptor = conv_handles[from][to];
if (srclen == (size_t)-1) {
- if (from == CH_UCS2) {
+ if (from == CH_UTF16LE || from == CH_UTF16BE) {
srclen = (strlen_w((const smb_ucs2_t *)src)+1) * 2;
} else {
srclen = strlen((const char *)src)+1;
again:
- retval = smb_iconv(descriptor, (char **)&inbuf, &i_len, &outbuf, &o_len);
+ retval = smb_iconv(descriptor, &inbuf, &i_len, &outbuf, &o_len);
if(retval==(size_t)-1) {
const char *reason="unknown error";
switch(errno) {
break;
case E2BIG:
reason="No more room";
- if (!conv_silent)
- DEBUG(3, ("convert_string_internal: Required %lu, available %lu\n",
- (unsigned long)srclen, (unsigned long)destlen));
- /* we are not sure we need srclen bytes,
- may be more, may be less.
- We only know we need more than destlen
- bytes ---simo */
- break;
+ if (!conv_silent) {
+ if (from == CH_UNIX) {
+ DEBUG(3,("E2BIG: convert_string(%s,%s): srclen=%u destlen=%u - '%s'\n",
+ charset_name(from), charset_name(to),
+ (unsigned int)srclen, (unsigned int)destlen, (const char *)src));
+ } else {
+ DEBUG(3,("E2BIG: convert_string(%s,%s): srclen=%u destlen=%u\n",
+ charset_name(from), charset_name(to),
+ (unsigned int)srclen, (unsigned int)destlen));
+ }
+ }
+ break;
case EILSEQ:
reason="Illegal multibyte sequence";
if (!conv_silent)
if (o_len == 0 || i_len == 0)
return destlen - o_len;
- if (from == CH_UCS2 && to != CH_UCS2) {
- /* Can't convert from ucs2 to multibyte. Just truncate this char to ascii. */
+ if (((from == CH_UTF16LE)||(from == CH_UTF16BE)) &&
+ ((to != CH_UTF16LE)||(to != CH_UTF16BE))) {
+ /* Can't convert from utf16 any endian to multibyte.
+ Replace with the default fail char.
+ */
if (i_len < 2)
return destlen - o_len;
if (i_len >= 2) {
- *outbuf = inbuf[0];
+ *outbuf = lp_failed_convert_char();
outbuf++;
o_len--;
/* Keep trying with the next char... */
goto again;
- } else if (from != CH_UCS2 && to == CH_UCS2) {
- /* Can't convert to ucs2 - just widen by adding zero. */
+ } else if (from != CH_UTF16LE && from != CH_UTF16BE && to == CH_UTF16LE) {
+ /* Can't convert to UTF16LE - just widen by adding the
+ default fail char then zero.
+ */
if (o_len < 2)
return destlen - o_len;
- outbuf[0] = inbuf[0];
+ outbuf[0] = lp_failed_convert_char();
outbuf[1] = '\0';
inbuf++;
/* Keep trying with the next char... */
goto again;
- } else if (from != CH_UCS2 && to != CH_UCS2) {
- /* Failed multibyte to multibyte. Just copy 1 char and
+ } else if (from != CH_UTF16LE && from != CH_UTF16BE &&
+ to != CH_UTF16LE && to != CH_UTF16BE) {
+ /* Failed multibyte to multibyte. Just copy the default fail char and
try again. */
- outbuf[0] = inbuf[0];
+ outbuf[0] = lp_failed_convert_char();
inbuf++;
i_len--;
if (srclen == 0)
return 0;
- if (from != CH_UCS2 && to != CH_UCS2) {
+ 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;
+ unsigned char lastp = '\0';
size_t retval = 0;
/* If all characters are ascii, fast path here. */
#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_UCS2 && to != CH_UCS2) {
+ } 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;
+ unsigned char lastp = '\0';
/* If all characters are ascii, fast path here. */
while (((slen == (size_t)-1) || (slen >= 2)) && dlen) {
#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_UCS2 && to == CH_UCS2) {
+ } 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;
+ unsigned char lastp = '\0';
/* If all characters are ascii, fast path here. */
while (slen && (dlen >= 2)) {
#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;
}
**/
size_t convert_string_allocate(TALLOC_CTX *ctx, charset_t from, charset_t to,
- void const *src, size_t srclen, void **dest, BOOL allow_bad_conv)
+ void const *src, size_t srclen, void *dst, BOOL allow_bad_conv)
{
size_t i_len, o_len, destlen = MAX(srclen, 512);
size_t retval;
const char *inbuf = (const char *)src;
char *outbuf = NULL, *ob = NULL;
smb_iconv_t descriptor;
+ void **dest = (void **)dst;
*dest = NULL;
if (descriptor == (smb_iconv_t)-1 || descriptor == (smb_iconv_t)0) {
if (!conv_silent)
DEBUG(0,("convert_string_allocate: Conversion not supported.\n"));
- if (allow_bad_conv)
- goto use_as_is;
return (size_t)-1;
}
destlen = destlen * 2;
}
- if (ctx)
- ob = (char *)talloc_realloc(ctx, ob, destlen);
- else
- ob = (char *)Realloc(ob, destlen);
+ if (ctx) {
+ ob = (char *)TALLOC_REALLOC(ctx, ob, destlen);
+ } else {
+ ob = (char *)SMB_REALLOC(ob, destlen);
+ }
if (!ob) {
DEBUG(0, ("convert_string_allocate: realloc failed!\n"));
- if (!ctx)
- SAFE_FREE(outbuf);
return (size_t)-1;
- } else {
- outbuf = ob;
}
+ outbuf = ob;
i_len = srclen;
o_len = destlen;
again:
retval = smb_iconv(descriptor,
- (char **)&inbuf, &i_len,
+ &inbuf, &i_len,
&outbuf, &o_len);
if(retval == (size_t)-1) {
const char *reason="unknown error";
out:
destlen = destlen - o_len;
- if (ctx)
- *dest = (char *)talloc_realloc(ctx,ob,destlen);
- else
- *dest = (char *)Realloc(ob,destlen);
- if (destlen && !*dest) {
+ if (ctx) {
+ ob = (char *)TALLOC_REALLOC(ctx,ob,destlen);
+ } else {
+ ob = (char *)SMB_REALLOC(ob,destlen);
+ }
+
+ if (destlen && !ob) {
DEBUG(0, ("convert_string_allocate: out of memory!\n"));
- if (!ctx)
- SAFE_FREE(ob);
return (size_t)-1;
}
+ *dest = ob;
return destlen;
use_as_is:
if (o_len == 0 || i_len == 0)
goto out;
- if (from == CH_UCS2 && to != CH_UCS2) {
- /* Can't convert from ucs2 to multibyte. Just truncate this char to ascii. */
+ if (((from == CH_UTF16LE)||(from == CH_UTF16BE)) &&
+ ((to != CH_UTF16LE)||(to != CH_UTF16BE))) {
+ /* Can't convert from utf16 any endian to multibyte.
+ Replace with the default fail char.
+ */
+
if (i_len < 2)
goto out;
if (i_len >= 2) {
- *outbuf = inbuf[0];
+ *outbuf = lp_failed_convert_char();
outbuf++;
o_len--;
/* Keep trying with the next char... */
goto again;
- } else if (from != CH_UCS2 && to == CH_UCS2) {
- /* Can't convert to ucs2 - just widen by adding zero. */
+ } else if (from != CH_UTF16LE && from != CH_UTF16BE && to == CH_UTF16LE) {
+ /* Can't convert to UTF16LE - just widen by adding the
+ default fail char then zero.
+ */
if (o_len < 2)
goto out;
- outbuf[0] = inbuf[0];
+ outbuf[0] = lp_failed_convert_char();
outbuf[1] = '\0';
inbuf++;
/* Keep trying with the next char... */
goto again;
- } else if (from != CH_UCS2 && to != CH_UCS2) {
- /* Failed multibyte to multibyte. Just copy 1 char and
- try again. */
- outbuf[0] = inbuf[0];
+ } else if (from != CH_UTF16LE && from != CH_UTF16BE &&
+ to != CH_UTF16LE && to != CH_UTF16BE) {
+ /* Failed multibyte to multibyte. Just copy the default fail char and
+ try again. */
+ outbuf[0] = lp_failed_convert_char();
inbuf++;
i_len--;
*
* @returns Size in bytes of the converted string; or -1 in case of error.
**/
-static size_t convert_string_talloc(TALLOC_CTX *ctx, charset_t from, charset_t to,
- void const *src, size_t srclen, void **dest, BOOL allow_bad_conv)
+size_t convert_string_talloc(TALLOC_CTX *ctx, charset_t from, charset_t to,
+ void const *src, size_t srclen, void *dst,
+ BOOL allow_bad_conv)
{
+ void **dest = (void **)dst;
size_t dest_len;
*dest = NULL;
return srclen;
}
- size = convert_string(CH_UCS2, CH_UNIX, buffer, size, dest, destlen, True);
+ size = convert_string(CH_UTF16LE, CH_UNIX, buffer, size, dest, destlen, True);
free(buffer);
return size;
}
while (1) {
if (*p & 0x80)
break;
- *q++ = toupper(*p);
+ *q++ = toupper_ascii(*p);
if (!*p)
break;
p++;
/* MB case. */
size_t size;
wpstring buffer;
- size = convert_string(CH_UNIX, CH_UCS2, s, -1, buffer, sizeof(buffer), True);
+ size = convert_string(CH_UNIX, CH_UTF16LE, s, -1, buffer, sizeof(buffer), True);
if (size == (size_t)-1) {
return NULL;
}
strupper_w(buffer);
- size = convert_string(CH_UCS2, CH_UNIX, buffer, -1, out_buffer, sizeof(out_buffer), True);
+ size = convert_string(CH_UTF16LE, CH_UNIX, buffer, -1, out_buffer, sizeof(out_buffer), True);
if (size == (size_t)-1) {
return NULL;
}
}
- return strdup(out_buffer);
+ return SMB_STRDUP(out_buffer);
}
size_t unix_strlower(const char *src, size_t srclen, char *dest, size_t destlen)
size_t size;
smb_ucs2_t *buffer = NULL;
- size = convert_string_allocate(NULL, CH_UNIX, CH_UCS2, src, srclen,
- (void **) &buffer, True);
+ size = convert_string_allocate(NULL, CH_UNIX, CH_UTF16LE, src, srclen,
+ (void **)(void *)&buffer, True);
if (size == (size_t)-1 || !buffer) {
smb_panic("failed to create UCS2 buffer");
}
SAFE_FREE(buffer);
return srclen;
}
- size = convert_string(CH_UCS2, CH_UNIX, buffer, size, dest, destlen, True);
+ size = convert_string(CH_UTF16LE, CH_UNIX, buffer, size, dest, destlen, True);
SAFE_FREE(buffer);
return size;
}
for (i = 0; buffer[i] != 0 && (i < buffer_len); i++) {
unsigned char mb[10];
/* Convert one smb_ucs2_t character at a time. */
- size_t mb_len = convert_string(CH_UCS2, CH_DOS, buffer+i, sizeof(smb_ucs2_t), mb, sizeof(mb), False);
+ size_t mb_len = convert_string(CH_UTF16LE, CH_DOS, buffer+i, sizeof(smb_ucs2_t), mb, sizeof(mb), False);
if ((mb_len != (size_t)-1) && (dest_len + mb_len <= MAX_NETBIOSNAME_LEN - 1)) {
memcpy((char *)dest + dest_len, mb, mb_len);
dest_len += mb_len;
if (flags & STR_TERMINATE) {
if (src_len == (size_t)-1) {
- src_len = strlen(src) + 1;
+ src_len = strlen((const char *)src) + 1;
} else {
- size_t len = strnlen(src, src_len);
+ size_t len = strnlen((const char *)src, src_len);
if (len < src_len)
len++;
src_len = len;
ret = convert_string(CH_DOS, CH_UNIX, src, src_len, dest, dest_len, True);
if (ret == (size_t)-1) {
+ ret = 0;
dest_len = 0;
}
- if (dest_len)
- dest[MIN(ret, dest_len-1)] = 0;
- else
+ if (dest_len && ret) {
+ /* Did we already process the terminating zero ? */
+ if (dest[MIN(ret-1, dest_len-1)] != 0) {
+ dest[MIN(ret, dest_len-1)] = 0;
+ }
+ } else {
dest[0] = 0;
+ }
return src_len;
}
size_t pull_ascii_nstring(char *dest, size_t dest_len, const void *src)
{
- return pull_ascii(dest, src, dest_len, sizeof(nstring), STR_TERMINATE);
+ return pull_ascii(dest, src, dest_len, sizeof(nstring)-1, STR_TERMINATE);
}
/**
/* ucs2 is always a multiple of 2 bytes */
dest_len &= ~1;
- ret = convert_string(CH_UNIX, CH_UCS2, src, src_len, dest, dest_len, True);
+ ret = convert_string(CH_UNIX, CH_UTF16LE, src, src_len, dest, dest_len, True);
if (ret == (size_t)-1) {
return 0;
}
len += ret;
if (flags & STR_UPPER) {
- smb_ucs2_t *dest_ucs2 = dest;
+ smb_ucs2_t *dest_ucs2 = (smb_ucs2_t *)dest;
size_t i;
- for (i = 0; i < (dest_len / 2) && dest_ucs2[i]; i++) {
+
+ /* We check for i < (ret / 2) below as the dest string isn't null
+ terminated if STR_TERMINATE isn't set. */
+
+ for (i = 0; i < (ret / 2) && i < (dest_len / 2) && dest_ucs2[i]; i++) {
smb_ucs2_t v = toupper_w(dest_ucs2[i]);
if (v != dest_ucs2[i]) {
dest_ucs2[i] = v;
size_t src_len = strlen(src)+1;
*dest = NULL;
- return convert_string_talloc(ctx, CH_UNIX, CH_UCS2, src, src_len, (void **)dest, True);
+ return convert_string_talloc(ctx, CH_UNIX, CH_UTF16LE, src, src_len, (void **)dest, True);
}
size_t src_len = strlen(src)+1;
*dest = NULL;
- return convert_string_allocate(NULL, CH_UNIX, CH_UCS2, src, src_len, (void **)dest, True);
+ return convert_string_allocate(NULL, CH_UNIX, CH_UTF16LE, src, src_len, (void **)dest, True);
}
/**
if (flags & STR_TERMINATE) {
/* src_len -1 is the default for null terminated strings. */
if (src_len != (size_t)-1) {
- size_t len = strnlen_w(src, src_len/2);
+ size_t len = strnlen_w((const smb_ucs2_t *)src,
+ src_len/2);
if (len < src_len/2)
len++;
src_len = len*2;
if (src_len != (size_t)-1)
src_len &= ~1;
- ret = convert_string(CH_UCS2, CH_UNIX, src, src_len, dest, dest_len, True);
+ ret = convert_string(CH_UTF16LE, CH_UNIX, src, src_len, dest, dest_len, True);
if (ret == (size_t)-1) {
return 0;
}
if (src_len == (size_t)-1)
src_len = ret*2;
- if (dest_len)
- dest[MIN(ret, dest_len-1)] = 0;
- else
+ if (dest_len && ret) {
+ /* Did we already process the terminating zero ? */
+ if (dest[MIN(ret-1, dest_len-1)] != 0) {
+ dest[MIN(ret, dest_len-1)] = 0;
+ }
+ } else {
dest[0] = 0;
+ }
return src_len;
}
{
size_t src_len = (strlen_w(src)+1) * sizeof(smb_ucs2_t);
*dest = NULL;
- return convert_string_talloc(ctx, CH_UCS2, CH_UNIX, src, src_len, (void **)dest, True);
+ return convert_string_talloc(ctx, CH_UTF16LE, CH_UNIX, src, src_len, (void **)dest, True);
}
/**
{
size_t src_len = (strlen_w(src)+1) * sizeof(smb_ucs2_t);
*dest = NULL;
- return convert_string_allocate(NULL, CH_UCS2, CH_UNIX, src, src_len, (void **)dest, True);
+ return convert_string_allocate(NULL, CH_UTF16LE, CH_UNIX, src, src_len, (void **)dest, True);
}
/**
return convert_string_allocate(NULL, CH_UTF8, CH_UNIX, src, src_len, (void **)dest, True);
}
+/**
+ * Copy a string from a DOS src to a unix char * destination, allocating a buffer using talloc
+ *
+ * @param dest always set at least to NULL
+ *
+ * @returns The number of bytes occupied by the string in the destination
+ **/
+
+size_t pull_ascii_talloc(TALLOC_CTX *ctx, char **dest, const char *src)
+{
+ size_t src_len = strlen(src)+1;
+ *dest = NULL;
+ return convert_string_talloc(ctx, CH_DOS, CH_UNIX, src, src_len, (void **)dest, True);
+}
+
/**
Copy a string from a char* src to a unicode or ascii
dos codepage destination choosing unicode or ascii based on the
The resulting string in "dest" is always null terminated.
**/
-size_t pull_string_fn(const char *function, unsigned int line, const void *base_ptr, char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
+size_t pull_string_fn(const char *function, unsigned int line,
+ const void *base_ptr, uint16 smb_flags2, char *dest,
+ const void *src, size_t dest_len, size_t src_len,
+ int flags)
{
#ifdef DEVELOPER
if (dest_len != (size_t)-1)
clobber_region(function, line, dest, dest_len);
#endif
+ if ((base_ptr == NULL) && ((flags & (STR_ASCII|STR_UNICODE)) == 0)) {
+ smb_panic("No base ptr to get flg2 and neither ASCII nor "
+ "UNICODE defined");
+ }
+
if (!(flags & STR_ASCII) && \
((flags & STR_UNICODE || \
- (SVAL(base_ptr, smb_flg2) & FLAGS2_UNICODE_STRINGS)))) {
+ (smb_flags2 & FLAGS2_UNICODE_STRINGS)))) {
return pull_ucs2(base_ptr, dest, src, dest_len, src_len, flags);
}
return pull_ascii(dest, src, dest_len, src_len, flags);
return 0;
}
-/****************************************************************
- Calculate the size (in bytes) of the next multibyte character in
- our internal character set. Note that p must be pointing to a
- valid mb char, not within one.
-****************************************************************/
+/*
+ Return the unicode codepoint for the next multi-byte CH_UNIX character
+ in the string. The unicode codepoint (codepoint_t) is an unsinged 32 bit value.
+
+ Also return the number of bytes consumed (which tells the caller
+ how many bytes to skip to get to the next CH_UNIX character).
-size_t next_mb_char_size(const char *s)
+ Return INVALID_CODEPOINT if the next character cannot be converted.
+*/
+
+codepoint_t next_codepoint(const char *str, size_t *size)
{
- size_t i;
+ /* It cannot occupy more than 4 bytes in UTF16 format */
+ uint8_t buf[4];
+ smb_iconv_t descriptor;
+ size_t ilen_orig;
+ size_t ilen;
+ size_t olen;
+ char *outbuf;
+
+ if ((str[0] & 0x80) == 0) {
+ *size = 1;
+ return (codepoint_t)str[0];
+ }
- if (!(*s & 0x80))
- return 1; /* ascii. */
+ /* We assume that no multi-byte character can take
+ more than 5 bytes. This is OK as we only
+ support codepoints up to 1M */
- conv_silent = True;
- for ( i = 1; i <=4; i++ ) {
- smb_ucs2_t uc;
- if (convert_string(CH_UNIX, CH_UCS2, s, i, &uc, 2, False) == 2) {
-#if 0 /* JRATEST */
- DEBUG(10,("next_mb_char_size: size %u at string %s\n",
- (unsigned int)i, s));
-#endif
- conv_silent = False;
- return i;
+ ilen_orig = strnlen(str, 5);
+ ilen = ilen_orig;
+
+ lazy_initialize_conv();
+
+ descriptor = conv_handles[CH_UNIX][CH_UTF16LE];
+ if (descriptor == (smb_iconv_t)-1 || descriptor == (smb_iconv_t)0) {
+ *size = 1;
+ return INVALID_CODEPOINT;
+ }
+
+ /* This looks a little strange, but it is needed to cope
+ with codepoints above 64k which are encoded as per RFC2781. */
+ olen = 2;
+ outbuf = (char *)buf;
+ smb_iconv(descriptor, &str, &ilen, &outbuf, &olen);
+ if (olen == 2) {
+ /* We failed to convert to a 2 byte character.
+ See if we can convert to a 4 UTF16-LE byte char encoding.
+ */
+ olen = 4;
+ outbuf = (char *)buf;
+ smb_iconv(descriptor, &str, &ilen, &outbuf, &olen);
+ if (olen == 4) {
+ /* We didn't convert any bytes */
+ *size = 1;
+ return INVALID_CODEPOINT;
}
+ olen = 4 - olen;
+ } else {
+ olen = 2 - olen;
}
- /* We're hosed - we don't know how big this is... */
- DEBUG(10,("next_mb_char_size: unknown size at string %s\n", s));
- conv_silent = False;
- return 1;
+
+ *size = ilen_orig - ilen;
+
+ if (olen == 2) {
+ /* 2 byte, UTF16-LE encoded value. */
+ return (codepoint_t)SVAL(buf, 0);
+ }
+ if (olen == 4) {
+ /* Decode a 4 byte UTF16-LE character manually.
+ See RFC2871 for the encoding machanism.
+ */
+ codepoint_t w1 = SVAL(buf,0) & ~0xD800;
+ codepoint_t w2 = SVAL(buf,2) & ~0xDC00;
+
+ return (codepoint_t)0x10000 +
+ (w1 << 10) + w2;
+ }
+
+ /* no other length is valid */
+ return INVALID_CODEPOINT;
}