-/*
+/*
Unix SMB/CIFS implementation.
Character set conversion Extensions
Copyright (C) Igor Vergeichik <iverg@mail.ru> 2001
Copyright (C) Andrew Tridgell 2001
Copyright (C) Simo Sorce 2001
Copyright (C) Martin Pool 2003
-
+
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,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
-
+
You should have received a copy of the GNU General Public License
- along with this program; if not, write to the Free Software
- Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "includes.h"
* @file
*
* @brief Character-set conversion routines built on our iconv.
- *
+ *
* @note Samba's internal character set (at least in the 3.0 series)
* is always the same as the one for the Unix filesystem. It is
* <b>not</b> necessarily UTF-8 and may be different on machines that
static smb_iconv_t conv_handles[NUM_CHARSETS][NUM_CHARSETS];
-static BOOL conv_silent; /* Should we do a debug if the conversion fails ? */
+static bool conv_silent; /* Should we do a debug if the conversion fails ? */
+static bool initialized;
/**
* Return the name of a charset to give to iconv().
{
const char *ret = NULL;
- if (ch == CH_UCS2) ret = "UTF-16LE";
+ 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();
}
ret = ln;
}
-#ifdef HAVE_SETLOCALE
- /* We set back the locale to C to get ASCII-compatible toupper/lower functions.
- For now we do not need any other POSIX localisations anyway. When we should
- really need localized string functions one day we need to write our own
- ascii_tolower etc.
- */
- setlocale(LC_ALL, "C");
- #endif
-
#endif
if (!ret || !*ret) ret = "ASCII";
void lazy_initialize_conv(void)
{
- static int initialized = False;
-
if (!initialized) {
- initialized = True;
load_case_tables();
init_iconv();
+ initialized = true;
}
}
+/**
+ * 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;
+ }
+ }
+ }
+ initialized = false;
+}
+
/**
* Initialize iconv conversion descriptors.
*
void init_iconv(void)
{
int c1, c2;
- BOOL did_reload = False;
+ bool did_reload = False;
/* 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(charset_name(CH_UCS2), "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", charset_name(CH_UCS2));
+ 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++) {
if (conv_handles[c1][c2] == (smb_iconv_t)-1) {
DEBUG(0,("init_iconv: Conversion from %s to %s not supported\n",
charset_name((charset_t)c1), charset_name((charset_t)c2)));
- if (c1 != CH_UCS2) {
+ if (c1 != CH_UTF16LE && c1 != CH_UTF16BE) {
n1 = "ASCII";
}
- if (c2 != CH_UCS2) {
+ if (c2 != CH_UTF16LE && c2 != CH_UTF16BE) {
n2 = "ASCII";
}
DEBUG(0,("init_iconv: Attempting to replace with conversion from %s to %s\n",
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.");
+ smb_panic("init_iconv: conv_handle initialization failed");
}
}
}
* codepage changes? */
/* XXX: Is the did_reload test too strict? */
conv_silent = True;
- init_doschar_table();
init_valid_table();
conv_silent = False;
}
static size_t convert_string_internal(charset_t from, charset_t to,
void const *src, size_t srclen,
- void *dest, size_t destlen, BOOL allow_bad_conv)
+ void *dest, size_t destlen, bool allow_bad_conv)
{
size_t i_len, o_len;
size_t retval;
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;
DEBUG(3,("convert_string_internal: Conversion error: %s(%s)\n",reason,inbuf));
if (allow_bad_conv)
goto use_as_is;
- break;
+ return (size_t)-1;
case E2BIG:
reason="No more room";
if (!conv_silent) {
DEBUG(3,("convert_string_internal: Conversion error: %s(%s)\n",reason,inbuf));
if (allow_bad_conv)
goto use_as_is;
- break;
+
+ return (size_t)-1;
default:
if (!conv_silent)
DEBUG(0,("convert_string_internal: Conversion error: %s(%s)\n",reason,inbuf));
- break;
+ return (size_t)-1;
}
/* smb_panic(reason); */
}
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. Replace with the default fail char. */
+ 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) {
/* 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 the default fail char then 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;
/* Keep trying with the next char... */
goto again;
- } else if (from != CH_UCS2 && to != CH_UCS2) {
+ } 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();
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)
+ void *dest, size_t destlen, bool allow_bad_conv)
{
/*
* NB. We deliberately don't do a strlen here if srclen == -1.
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;
#ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
goto general_case;
#else
- return retval + convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
+ size_t ret = convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
+ if (ret == (size_t)-1) {
+ return ret;
+ }
+ return retval + ret;
#endif
}
}
}
}
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;
#ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
goto general_case;
#else
- return retval + convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
+ size_t ret = convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
+ if (ret == (size_t)-1) {
+ return ret;
+ }
+ return retval + ret;
#endif
}
}
}
}
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;
#ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
goto general_case;
#else
- return retval + convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
+ size_t ret = convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
+ if (ret == (size_t)-1) {
+ return ret;
+ }
+ return retval + ret;
#endif
}
}
}
/**
- * Convert between character sets, allocating a new buffer for the result.
+ * Convert between character sets, allocating a new buffer using talloc for the result.
*
- * @param ctx TALLOC_CTX to use to allocate with. If NULL use malloc.
* @param srclen length of source buffer.
* @param dest always set at least to NULL
+ * @parm converted_size set to the number of bytes occupied by the string in
+ * the destination on success.
* @note -1 is not accepted for srclen.
*
- * @returns Size in bytes of the converted string; or -1 in case of error.
+ * @return true if new buffer was correctly allocated, and string was
+ * converted.
*
* Ensure the srclen contains the terminating zero.
- *
+ *
* I hate the goto's in this function. It's embarressing.....
* There has to be a cleaner way to do this. JRA.
- **/
+ */
+bool convert_string_talloc(TALLOC_CTX *ctx, charset_t from, charset_t to,
+ void const *src, size_t srclen, void *dst,
+ size_t *converted_size, bool allow_bad_conv)
-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)
{
- size_t i_len, o_len, destlen = MAX(srclen, 512);
+ size_t i_len, o_len, destlen = (srclen * 3) / 2;
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 (src == NULL || srclen == (size_t)-1)
- return (size_t)-1;
- if (srclen == 0)
- return 0;
+ if (!converted_size) {
+ errno = EINVAL;
+ return false;
+ }
+
+ if (src == NULL || srclen == (size_t)-1) {
+ errno = EINVAL;
+ return false;
+ }
+ if (srclen == 0) {
+ ob = ((ctx != NULL) ? talloc_strdup(ctx, "") : SMB_STRDUP(""));
+ if (ob == NULL) {
+ errno = ENOMEM;
+ return false;
+ }
+ *dest = ob;
+ *converted_size = 0;
+ return true;
+ }
lazy_initialize_conv();
if (descriptor == (smb_iconv_t)-1 || descriptor == (smb_iconv_t)0) {
if (!conv_silent)
- DEBUG(0,("convert_string_allocate: Conversion not supported.\n"));
- return (size_t)-1;
+ DEBUG(0,("convert_string_talloc: Conversion not supported.\n"));
+ errno = EOPNOTSUPP;
+ return false;
}
convert:
- if ((destlen*2) < destlen) {
+ /* +2 is for ucs2 null termination. */
+ if ((destlen*2)+2 < destlen) {
/* wrapped ! abort. */
if (!conv_silent)
- DEBUG(0, ("convert_string_allocate: destlen wrapped !\n"));
+ DEBUG(0, ("convert_string_talloc: destlen wrapped !\n"));
if (!ctx)
SAFE_FREE(outbuf);
- return (size_t)-1;
+ errno = EOPNOTSUPP;
+ return false;
} else {
destlen = destlen * 2;
}
- if (ctx)
- ob = (char *)TALLOC_REALLOC(ctx, ob, destlen);
- else
- ob = (char *)SMB_REALLOC(ob, destlen);
+ /* +2 is for ucs2 null termination. */
+ ob = (char *)TALLOC_REALLOC(ctx, ob, destlen + 2);
if (!ob) {
- DEBUG(0, ("convert_string_allocate: realloc failed!\n"));
- if (!ctx)
- SAFE_FREE(outbuf);
- return (size_t)-1;
- } else {
- outbuf = ob;
+ DEBUG(0, ("convert_string_talloc: realloc failed!\n"));
+ errno = ENOMEM;
+ return false;
}
+ outbuf = ob;
i_len = srclen;
o_len = destlen;
case EINVAL:
reason="Incomplete multibyte sequence";
if (!conv_silent)
- DEBUG(3,("convert_string_allocate: Conversion error: %s(%s)\n",reason,inbuf));
+ DEBUG(3,("convert_string_talloc: Conversion error: %s(%s)\n",reason,inbuf));
if (allow_bad_conv)
goto use_as_is;
break;
case E2BIG:
- goto convert;
+ goto convert;
case EILSEQ:
reason="Illegal multibyte sequence";
if (!conv_silent)
- DEBUG(3,("convert_string_allocate: Conversion error: %s(%s)\n",reason,inbuf));
+ DEBUG(3,("convert_string_talloc: Conversion error: %s(%s)\n",reason,inbuf));
if (allow_bad_conv)
goto use_as_is;
break;
if (!conv_silent)
DEBUG(0,("Conversion error: %s(%s)\n",reason,inbuf));
/* smb_panic(reason); */
- return (size_t)-1;
+ TALLOC_FREE(ob);
+ return false;
}
out:
destlen = destlen - o_len;
- if (ctx)
- *dest = (char *)TALLOC_REALLOC(ctx,ob,destlen);
- else
- *dest = (char *)SMB_REALLOC(ob,destlen);
- if (destlen && !*dest) {
- DEBUG(0, ("convert_string_allocate: out of memory!\n"));
- if (!ctx)
- SAFE_FREE(ob);
- return (size_t)-1;
+ /* Don't shrink unless we're reclaiming a lot of
+ * space. This is in the hot codepath and these
+ * reallocs *cost*. JRA.
+ */
+ if (o_len > 1024) {
+ /* We're shrinking here so we know the +2 is safe from wrap. */
+ ob = (char *)TALLOC_REALLOC(ctx,ob,destlen + 2);
+ }
+
+ if (destlen && !ob) {
+ DEBUG(0, ("convert_string_talloc: out of memory!\n"));
+ errno = ENOMEM;
+ return false;
}
- return destlen;
+ *dest = ob;
+
+ /* Must ucs2 null terminate in the extra space we allocated. */
+ ob[destlen] = '\0';
+ ob[destlen+1] = '\0';
+
+ *converted_size = destlen;
+ return true;
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 use the default fail char. */
+ 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;
/* 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 the default fail char then 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;
/* Keep trying with the next char... */
goto again;
- } else if (from != CH_UCS2 && to != CH_UCS2) {
+ } 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. */
+ try again. */
outbuf[0] = lp_failed_convert_char();
inbuf++;
}
}
-/**
- * Convert between character sets, allocating a new buffer using talloc for the result.
- *
- * @param srclen length of source buffer.
- * @param dest always set at least to NULL
- * @note -1 is not accepted for srclen.
- *
- * @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 dest_len;
-
- *dest = NULL;
- dest_len=convert_string_allocate(ctx, from, to, src, srclen, dest, allow_bad_conv);
- if (dest_len == (size_t)-1)
- return (size_t)-1;
- if (*dest == NULL)
- return (size_t)-1;
- return dest_len;
-}
-
size_t unix_strupper(const char *src, size_t srclen, char *dest, size_t destlen)
{
size_t size;
smb_ucs2_t *buffer;
-
- size = push_ucs2_allocate(&buffer, src);
- if (size == (size_t)-1) {
- smb_panic("failed to create UCS2 buffer");
+
+ if (!push_ucs2_talloc(NULL, &buffer, src, &size)) {
+ return (size_t)-1;
}
+
if (!strupper_w(buffer) && (dest == src)) {
- free(buffer);
+ TALLOC_FREE(buffer);
return srclen;
}
-
- size = convert_string(CH_UCS2, CH_UNIX, buffer, size, dest, destlen, True);
- free(buffer);
+
+ size = convert_string(CH_UTF16LE, CH_UNIX, buffer, size, dest, destlen, True);
+ TALLOC_FREE(buffer);
return size;
}
/**
- strdup() a unix string to upper case.
- Max size is pstring.
+ talloc_strdup() a unix string to upper case.
**/
-char *strdup_upper(const char *s)
+char *talloc_strdup_upper(TALLOC_CTX *ctx, const char *s)
{
- pstring out_buffer;
+ char *out_buffer = talloc_strdup(ctx,s);
const unsigned char *p = (const unsigned char *)s;
unsigned char *q = (unsigned char *)out_buffer;
+ if (!q) {
+ return NULL;
+ }
+
/* this is quite a common operation, so we want it to be
fast. We optimise for the ascii case, knowing that all our
supported multi-byte character sets are ascii-compatible
(ie. they match for the first 128 chars) */
- while (1) {
+ while (*p) {
if (*p & 0x80)
break;
- *q++ = toupper(*p);
- if (!*p)
- break;
+ *q++ = toupper_ascii_fast(*p);
p++;
- if (p - ( const unsigned char *)s >= sizeof(pstring))
- break;
}
if (*p) {
/* MB case. */
- size_t size;
- wpstring buffer;
- size = convert_string(CH_UNIX, CH_UCS2, s, -1, buffer, sizeof(buffer), True);
- if (size == (size_t)-1) {
+ size_t converted_size, converted_size2;
+ smb_ucs2_t *ubuf = NULL;
+
+ /* We're not using the ascii buffer above. */
+ TALLOC_FREE(out_buffer);
+
+ if (!convert_string_talloc(ctx, CH_UNIX, CH_UTF16LE, s,
+ strlen(s)+1, (void *)&ubuf,
+ &converted_size, True))
+ {
return NULL;
}
- strupper_w(buffer);
-
- size = convert_string(CH_UCS2, CH_UNIX, buffer, -1, out_buffer, sizeof(out_buffer), True);
- if (size == (size_t)-1) {
+ strupper_w(ubuf);
+
+ if (!convert_string_talloc(ctx, CH_UTF16LE, CH_UNIX, ubuf,
+ converted_size, (void *)&out_buffer,
+ &converted_size2, True))
+ {
+ TALLOC_FREE(ubuf);
return NULL;
}
+
+ /* Don't need the intermediate buffer
+ * anymore.
+ */
+ TALLOC_FREE(ubuf);
}
- return SMB_STRDUP(out_buffer);
+ return out_buffer;
+}
+
+char *strupper_talloc(TALLOC_CTX *ctx, const char *s) {
+ return talloc_strdup_upper(ctx, s);
}
+
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);
- if (size == (size_t)-1 || !buffer) {
+
+ if (!convert_string_talloc(NULL, CH_UNIX, CH_UTF16LE, src, srclen,
+ (void **)(void *)&buffer, &size,
+ True))
+ {
smb_panic("failed to create UCS2 buffer");
}
if (!strlower_w(buffer) && (dest == src)) {
- SAFE_FREE(buffer);
+ TALLOC_FREE(buffer);
return srclen;
}
- size = convert_string(CH_UCS2, CH_UNIX, buffer, size, dest, destlen, True);
- SAFE_FREE(buffer);
+ size = convert_string(CH_UTF16LE, CH_UNIX, buffer, size, dest, destlen, True);
+ TALLOC_FREE(buffer);
return size;
}
-/**
- strdup() a unix string to lower case.
-**/
-char *strdup_lower(const char *s)
+char *talloc_strdup_lower(TALLOC_CTX *ctx, const char *s)
{
- size_t size;
+ size_t converted_size;
smb_ucs2_t *buffer = NULL;
char *out_buffer;
-
- size = push_ucs2_allocate(&buffer, s);
- if (size == -1 || !buffer) {
+
+ if (!push_ucs2_talloc(ctx, &buffer, s, &converted_size)) {
return NULL;
}
strlower_w(buffer);
-
- size = pull_ucs2_allocate(&out_buffer, buffer);
- SAFE_FREE(buffer);
- if (size == (size_t)-1) {
+ if (!pull_ucs2_talloc(ctx, &out_buffer, buffer, &converted_size)) {
+ TALLOC_FREE(buffer);
return NULL;
}
-
+
+ TALLOC_FREE(buffer);
+
return out_buffer;
}
-static size_t ucs2_align(const void *base_ptr, const void *p, int flags)
+char *strlower_talloc(TALLOC_CTX *ctx, const char *s) {
+ return talloc_strdup_lower(ctx, s);
+}
+
+size_t ucs2_align(const void *base_ptr, const void *p, int flags)
{
if (flags & (STR_NOALIGN|STR_ASCII))
return 0;
* </dl>
*
* @param dest_len the maximum length in bytes allowed in the
- * destination. If @p dest_len is -1 then no maximum is used.
+ * destination.
**/
size_t push_ascii(void *dest, const char *src, size_t dest_len, int flags)
{
size_t src_len = strlen(src);
- pstring tmpbuf;
+ char *tmpbuf = NULL;
+ size_t ret;
- /* treat a pstring as "unlimited" length */
- if (dest_len == (size_t)-1)
- dest_len = sizeof(pstring);
+ /* No longer allow a length of -1. */
+ if (dest_len == (size_t)-1) {
+ smb_panic("push_ascii - dest_len == -1");
+ }
if (flags & STR_UPPER) {
- pstrcpy(tmpbuf, src);
+ tmpbuf = SMB_STRDUP(src);
+ if (!tmpbuf) {
+ smb_panic("malloc fail");
+ }
strupper_m(tmpbuf);
src = tmpbuf;
}
- if (flags & (STR_TERMINATE | STR_TERMINATE_ASCII))
+ if (flags & (STR_TERMINATE | STR_TERMINATE_ASCII)) {
src_len++;
+ }
- return convert_string(CH_UNIX, CH_DOS, src, src_len, dest, dest_len, True);
+ ret = convert_string(CH_UNIX, CH_DOS, src, src_len, dest, dest_len, True);
+ if (ret == (size_t)-1 &&
+ (flags & (STR_TERMINATE | STR_TERMINATE_ASCII))
+ && dest_len > 0) {
+ ((char *)dest)[0] = '\0';
+ }
+ SAFE_FREE(tmpbuf);
+ return ret;
}
size_t push_ascii_fstring(void *dest, const char *src)
return push_ascii(dest, src, sizeof(fstring), STR_TERMINATE);
}
-size_t push_ascii_pstring(void *dest, const char *src)
-{
- return push_ascii(dest, src, sizeof(pstring), STR_TERMINATE);
-}
-
/********************************************************************
Push an nstring - ensure null terminated. Written by
moriyama@miraclelinux.com (MORIYAMA Masayuki).
smb_ucs2_t *buffer;
conv_silent = True;
- buffer_len = push_ucs2_allocate(&buffer, src);
- if (buffer_len == (size_t)-1) {
+ if (!push_ucs2_talloc(NULL, &buffer, src, &buffer_len)) {
smb_panic("failed to create UCS2 buffer");
}
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;
}
((char *)dest)[dest_len] = '\0';
- SAFE_FREE(buffer);
conv_silent = False;
+ TALLOC_FREE(buffer);
return dest_len;
}
+/********************************************************************
+ Push and malloc an ascii string. src and dest null terminated.
+********************************************************************/
+
+bool push_ascii_talloc(TALLOC_CTX *mem_ctx, char **dest, const char *src, size_t *converted_size)
+{
+ size_t src_len = strlen(src)+1;
+
+ *dest = NULL;
+ return convert_string_talloc(mem_ctx, CH_UNIX, CH_DOS, src, src_len,
+ (void **)dest, converted_size, True);
+}
+
/**
* Copy a string from a dos codepage source to a unix char* destination.
*
{
size_t ret;
- if (dest_len == (size_t)-1)
- dest_len = sizeof(pstring);
+ if (dest_len == (size_t)-1) {
+ /* No longer allow dest_len of -1. */
+ smb_panic("pull_ascii - invalid dest_len of -1");
+ }
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_pstring(char *dest, const void *src)
+/**
+ * Copy a string from a dos codepage source to a unix char* destination.
+ Talloc version.
+ Uses malloc if TALLOC_CTX is NULL (this is a bad interface and
+ needs fixing. JRA).
+ *
+ * The resulting string in "dest" is always null terminated.
+ *
+ * @param flags can have:
+ * <dl>
+ * <dt>STR_TERMINATE</dt>
+ * <dd>STR_TERMINATE means the string in @p src
+ * is null terminated, and src_len is ignored.</dd>
+ * </dl>
+ *
+ * @param src_len is the length of the source area in bytes.
+ * @returns the number of bytes occupied by the string in @p src.
+ **/
+
+static size_t pull_ascii_base_talloc(TALLOC_CTX *ctx,
+ char **ppdest,
+ const void *src,
+ size_t src_len,
+ int flags)
{
- return pull_ascii(dest, src, sizeof(pstring), -1, STR_TERMINATE);
+ char *dest = NULL;
+ size_t dest_len;
+
+ *ppdest = NULL;
+
+ if (!src_len) {
+ return 0;
+ }
+
+ if (flags & STR_TERMINATE) {
+ if (src_len == (size_t)-1) {
+ src_len = strlen((const char *)src) + 1;
+ } else {
+ size_t len = strnlen((const char *)src, src_len);
+ if (len < src_len)
+ len++;
+ src_len = len;
+ }
+ /* Ensure we don't use an insane length from the client. */
+ if (src_len >= 1024*1024) {
+ char *msg = talloc_asprintf(ctx,
+ "Bad src length (%u) in "
+ "pull_ascii_base_talloc",
+ (unsigned int)src_len);
+ smb_panic(msg);
+ }
+ } else {
+ /* Can't have an unlimited length
+ * non STR_TERMINATE'd.
+ */
+ if (src_len == (size_t)-1) {
+ errno = EINVAL;
+ return 0;
+ }
+ }
+
+ /* src_len != -1 here. */
+
+ if (!convert_string_talloc(ctx, CH_DOS, CH_UNIX, src, src_len, &dest,
+ &dest_len, True)) {
+ dest_len = 0;
+ }
+
+ if (dest_len && dest) {
+ /* Did we already process the terminating zero ? */
+ if (dest[dest_len-1] != 0) {
+ size_t size = talloc_get_size(dest);
+ /* Have we got space to append the '\0' ? */
+ if (size <= dest_len) {
+ /* No, realloc. */
+ dest = TALLOC_REALLOC_ARRAY(ctx, dest, char,
+ dest_len+1);
+ if (!dest) {
+ /* talloc fail. */
+ dest_len = (size_t)-1;
+ return 0;
+ }
+ }
+ /* Yay - space ! */
+ dest[dest_len] = '\0';
+ dest_len++;
+ }
+ } else if (dest) {
+ dest[0] = 0;
+ }
+
+ *ppdest = dest;
+ return src_len;
}
size_t pull_ascii_fstring(char *dest, const void *src)
* </dl>
*
* @param dest_len is the maximum length allowed in the
- * destination. If dest_len is -1 then no maxiumum is used.
+ * destination.
**/
size_t push_ucs2(const void *base_ptr, void *dest, const char *src, size_t dest_len, int flags)
size_t src_len;
size_t ret;
- /* treat a pstring as "unlimited" length */
- if (dest_len == (size_t)-1)
- dest_len = sizeof(pstring);
+ if (dest_len == (size_t)-1) {
+ /* No longer allow dest_len of -1. */
+ smb_panic("push_ucs2 - invalid dest_len of -1");
+ }
if (flags & STR_TERMINATE)
src_len = (size_t)-1;
/* 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;
+ if ((flags & STR_TERMINATE) &&
+ dest &&
+ dest_len) {
+ *(char *)dest = 0;
+ }
+ return len;
}
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;
* allocating a buffer using talloc().
*
* @param dest always set at least to NULL
+ * @parm converted_size set to the number of bytes occupied by the string in
+ * the destination on success.
*
- * @returns The number of bytes occupied by the string in the destination
- * or -1 in case of error.
+ * @return true if new buffer was correctly allocated, and string was
+ * converted.
**/
-size_t push_ucs2_talloc(TALLOC_CTX *ctx, smb_ucs2_t **dest, const char *src)
+bool push_ucs2_talloc(TALLOC_CTX *ctx, smb_ucs2_t **dest, const char *src,
+ size_t *converted_size)
{
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, converted_size, True);
}
-/**
- * Copy a string from a unix char* src to a UCS2 destination, allocating a buffer
- *
- * @param dest always set at least to NULL
- *
- * @returns The number of bytes occupied by the string in the destination
- * or -1 in case of error.
- **/
-
-size_t push_ucs2_allocate(smb_ucs2_t **dest, const char *src)
-{
- size_t src_len = strlen(src)+1;
-
- *dest = NULL;
- return convert_string_allocate(NULL, CH_UNIX, CH_UCS2, src, src_len, (void **)dest, True);
-}
-
/**
Copy a string from a char* src to a UTF-8 destination.
Return the number of bytes occupied by the string in the destination
static size_t push_utf8(void *dest, const char *src, size_t dest_len, int flags)
{
- size_t src_len = strlen(src);
- pstring tmpbuf;
+ size_t src_len = 0;
+ size_t ret;
+ char *tmpbuf = NULL;
- /* treat a pstring as "unlimited" length */
- if (dest_len == (size_t)-1)
- dest_len = sizeof(pstring);
+ if (dest_len == (size_t)-1) {
+ /* No longer allow dest_len of -1. */
+ smb_panic("push_utf8 - invalid dest_len of -1");
+ }
if (flags & STR_UPPER) {
- pstrcpy(tmpbuf, src);
- strupper_m(tmpbuf);
+ tmpbuf = strupper_talloc(NULL, src);
+ if (!tmpbuf) {
+ return (size_t)-1;
+ }
src = tmpbuf;
+ src_len = strlen(src);
}
- if (flags & STR_TERMINATE)
+ src_len = strlen(src);
+ if (flags & STR_TERMINATE) {
src_len++;
+ }
- return convert_string(CH_UNIX, CH_UTF8, src, src_len, dest, dest_len, True);
+ ret = convert_string(CH_UNIX, CH_UTF8, src, src_len, dest, dest_len, True);
+ TALLOC_FREE(tmpbuf);
+ return ret;
}
size_t push_utf8_fstring(void *dest, const char *src)
* Copy a string from a unix char* src to a UTF-8 destination, allocating a buffer using talloc
*
* @param dest always set at least to NULL
+ * @parm converted_size set to the number of bytes occupied by the string in
+ * the destination on success.
*
- * @returns The number of bytes occupied by the string in the destination
- **/
-
-size_t push_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src)
-{
- size_t src_len = strlen(src)+1;
-
- *dest = NULL;
- return convert_string_talloc(ctx, CH_UNIX, CH_UTF8, src, src_len, (void**)dest, True);
-}
-
-/**
- * Copy a string from a unix char* src to a UTF-8 destination, allocating a buffer
- *
- * @param dest always set at least to NULL
- *
- * @returns The number of bytes occupied by the string in the destination
+ * @return true if new buffer was correctly allocated, and string was
+ * converted.
**/
-size_t push_utf8_allocate(char **dest, const char *src)
+bool push_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src,
+ size_t *converted_size)
{
size_t src_len = strlen(src)+1;
*dest = NULL;
- return convert_string_allocate(NULL, CH_UNIX, CH_UTF8, src, src_len, (void **)dest, True);
+ return convert_string_talloc(ctx, CH_UNIX, CH_UTF8, src, src_len,
+ (void**)dest, converted_size, True);
}
/**
{
size_t ret;
- if (dest_len == (size_t)-1)
- dest_len = sizeof(pstring);
+ if (dest_len == (size_t)-1) {
+ /* No longer allow dest_len of -1. */
+ smb_panic("pull_ucs2 - invalid dest_len of -1");
+ }
+
+ if (!src_len) {
+ if (dest && dest_len > 0) {
+ dest[0] = '\0';
+ }
+ return 0;
+ }
if (ucs2_align(base_ptr, src, flags)) {
src = (const void *)((const char *)src + 1);
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;
/* ucs2 is always a multiple of 2 bytes */
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;
+ ret = 0;
+ dest_len = 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 pull_ucs2_pstring(char *dest, const void *src)
+/**
+ Copy a string from a ucs2 source to a unix char* destination.
+ Talloc version with a base pointer.
+ Uses malloc if TALLOC_CTX is NULL (this is a bad interface and
+ needs fixing. JRA).
+ Flags can have:
+ STR_TERMINATE means the string in src is null terminated.
+ STR_NOALIGN means don't try to align.
+ if STR_TERMINATE is set then src_len is ignored if it is -1.
+ src_len is the length of the source area in bytes
+ Return the number of bytes occupied by the string in src.
+ The resulting string in "dest" is always null terminated.
+**/
+
+size_t pull_ucs2_base_talloc(TALLOC_CTX *ctx,
+ const void *base_ptr,
+ char **ppdest,
+ const void *src,
+ size_t src_len,
+ int flags)
{
- return pull_ucs2(NULL, dest, src, sizeof(pstring), -1, STR_TERMINATE);
+ char *dest;
+ size_t dest_len;
+
+ *ppdest = NULL;
+
+#ifdef DEVELOPER
+ /* Ensure we never use the braindead "malloc" varient. */
+ if (ctx == NULL) {
+ smb_panic("NULL talloc CTX in pull_ucs2_base_talloc\n");
+ }
+#endif
+
+ if (!src_len) {
+ return 0;
+ }
+
+ if (ucs2_align(base_ptr, src, flags)) {
+ src = (const void *)((const char *)src + 1);
+ if (src_len != (size_t)-1)
+ src_len--;
+ }
+
+ 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((const smb_ucs2_t *)src,
+ src_len/2);
+ if (len < src_len/2)
+ len++;
+ src_len = len*2;
+ } else {
+ /*
+ * src_len == -1 - alloc interface won't take this
+ * so we must calculate.
+ */
+ src_len = (strlen_w((const smb_ucs2_t *)src)+1)*sizeof(smb_ucs2_t);
+ }
+ /* Ensure we don't use an insane length from the client. */
+ if (src_len >= 1024*1024) {
+ smb_panic("Bad src length in pull_ucs2_base_talloc\n");
+ }
+ } else {
+ /* Can't have an unlimited length
+ * non STR_TERMINATE'd.
+ */
+ if (src_len == (size_t)-1) {
+ errno = EINVAL;
+ return 0;
+ }
+ }
+
+ /* src_len != -1 here. */
+
+ /* ucs2 is always a multiple of 2 bytes */
+ src_len &= ~1;
+
+ if (!convert_string_talloc(ctx, CH_UTF16LE, CH_UNIX, src, src_len,
+ (void *)&dest, &dest_len, True)) {
+ dest_len = 0;
+ }
+
+ if (dest_len) {
+ /* Did we already process the terminating zero ? */
+ if (dest[dest_len-1] != 0) {
+ size_t size = talloc_get_size(dest);
+ /* Have we got space to append the '\0' ? */
+ if (size <= dest_len) {
+ /* No, realloc. */
+ dest = TALLOC_REALLOC_ARRAY(ctx, dest, char,
+ dest_len+1);
+ if (!dest) {
+ /* talloc fail. */
+ dest_len = (size_t)-1;
+ return 0;
+ }
+ }
+ /* Yay - space ! */
+ dest[dest_len] = '\0';
+ dest_len++;
+ }
+ } else if (dest) {
+ dest[0] = 0;
+ }
+
+ *ppdest = dest;
+ return src_len;
}
size_t pull_ucs2_fstring(char *dest, const void *src)
* Copy a string from a UCS2 src to a unix char * destination, allocating a buffer using talloc
*
* @param dest always set at least to NULL
+ * @parm converted_size set to the number of bytes occupied by the string in
+ * the destination on success.
*
- * @returns The number of bytes occupied by the string in the destination
+ * @return true if new buffer was correctly allocated, and string was
+ * converted.
**/
-size_t pull_ucs2_talloc(TALLOC_CTX *ctx, char **dest, const smb_ucs2_t *src)
+bool pull_ucs2_talloc(TALLOC_CTX *ctx, char **dest, const smb_ucs2_t *src,
+ size_t *converted_size)
{
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, converted_size, True);
}
/**
- * Copy a string from a UCS2 src to a unix char * destination, allocating a buffer
+ * Copy a string from a UTF-8 src to a unix char * destination, allocating a buffer using talloc
*
* @param dest always set at least to NULL
+ * @parm converted_size set to the number of bytes occupied by the string in
+ * the destination on success.
*
- * @returns The number of bytes occupied by the string in the destination
+ * @return true if new buffer was correctly allocated, and string was
+ * converted.
**/
-size_t pull_ucs2_allocate(char **dest, const smb_ucs2_t *src)
+bool pull_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src,
+ size_t *converted_size)
{
- size_t src_len = (strlen_w(src)+1) * sizeof(smb_ucs2_t);
+ size_t src_len = strlen(src)+1;
+
*dest = NULL;
- return convert_string_allocate(NULL, CH_UCS2, CH_UNIX, src, src_len, (void **)dest, True);
+ return convert_string_talloc(ctx, CH_UTF8, CH_UNIX, src, src_len,
+ (void **)dest, converted_size, True);
}
+
/**
- * Copy a string from a UTF-8 src to a unix char * destination, allocating a buffer using talloc
+ * 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
+ * @parm converted_size set to the number of bytes occupied by the string in
+ * the destination on success.
*
- * @returns The number of bytes occupied by the string in the destination
+ * @return true if new buffer was correctly allocated, and string was
+ * converted.
**/
-size_t pull_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src)
+bool pull_ascii_talloc(TALLOC_CTX *ctx, char **dest, const char *src,
+ size_t *converted_size)
{
size_t src_len = strlen(src)+1;
+
*dest = NULL;
- return convert_string_talloc(ctx, CH_UTF8, CH_UNIX, src, src_len, (void **)dest, True);
+ return convert_string_talloc(ctx, CH_DOS, CH_UNIX, src, src_len,
+ (void **)dest, converted_size, True);
}
/**
- * Copy a string from a UTF-8 src to a unix char * destination, allocating a buffer
- *
- * @param dest always set at least to NULL
- *
- * @returns The number of bytes occupied by the string in the destination
- **/
+ Copy a string from a char* src to a unicode or ascii
+ dos codepage destination choosing unicode or ascii based on the
+ flags supplied
+ Return the number of bytes occupied by the string in the destination.
+ flags can have:
+ STR_TERMINATE means include the null termination.
+ STR_UPPER means uppercase in the destination.
+ STR_ASCII use ascii even with unicode packet.
+ STR_NOALIGN means don't do alignment.
+ dest_len is the maximum length allowed in the destination. If dest_len
+ is -1 then no maxiumum is used.
+**/
-size_t pull_utf8_allocate(char **dest, const char *src)
+size_t push_string_check_fn(const char *function, unsigned int line,
+ void *dest, const char *src,
+ size_t dest_len, int flags)
{
- size_t src_len = strlen(src)+1;
- *dest = NULL;
- return convert_string_allocate(NULL, CH_UTF8, CH_UNIX, src, src_len, (void **)dest, True);
+#ifdef DEVELOPER
+ /* We really need to zero fill here, not clobber
+ * region, as we want to ensure that valgrind thinks
+ * all of the outgoing buffer has been written to
+ * so a send() or write() won't trap an error.
+ * JRA.
+ */
+#if 0
+ clobber_region(function, line, dest, dest_len);
+#else
+ memset(dest, '\0', dest_len);
+#endif
+#endif
+
+ if (!(flags & STR_ASCII) && (flags & STR_UNICODE)) {
+ return push_ucs2(NULL, dest, src, dest_len, flags);
+ }
+ return push_ascii(dest, src, dest_len, flags);
}
-
+
+
/**
Copy a string from a char* src to a unicode or ascii
dos codepage destination choosing unicode or ascii based on the
is -1 then no maxiumum is used.
**/
-size_t push_string_fn(const char *function, unsigned int line, const void *base_ptr, void *dest, const char *src, size_t dest_len, int flags)
+size_t push_string_base(const char *function, unsigned int line,
+ const char *base, uint16 flags2,
+ void *dest, const char *src,
+ size_t dest_len, int flags)
{
#ifdef DEVELOPER
/* We really need to zero fill here, not clobber
* JRA.
*/
#if 0
- if (dest_len != (size_t)-1)
- clobber_region(function, line, dest, dest_len);
+ clobber_region(function, line, dest, dest_len);
#else
- if (dest_len != (size_t)-1)
- memset(dest, '\0', dest_len);
+ memset(dest, '\0', dest_len);
#endif
#endif
if (!(flags & STR_ASCII) && \
((flags & STR_UNICODE || \
- (SVAL(base_ptr, smb_flg2) & FLAGS2_UNICODE_STRINGS)))) {
- return push_ucs2(base_ptr, dest, src, dest_len, flags);
+ (flags2 & FLAGS2_UNICODE_STRINGS)))) {
+ return push_ucs2(base, dest, src, dest_len, flags);
}
return push_ascii(dest, src, dest_len, flags);
}
+/**
+ Copy a string from a char* src to a unicode or ascii
+ dos codepage destination choosing unicode or ascii based on the
+ flags supplied
+ Return the number of bytes occupied by the string in the destination.
+ flags can have:
+ STR_TERMINATE means include the null termination.
+ STR_UPPER means uppercase in the destination.
+ STR_ASCII use ascii even with unicode packet.
+ STR_NOALIGN means don't do alignment.
+ dest_len is the maximum length allowed in the destination. If dest_len
+ is -1 then no maxiumum is used.
+**/
+
+ssize_t push_string(void *dest, const char *src, size_t dest_len, int flags)
+{
+ size_t ret;
+#ifdef DEVELOPER
+ /* We really need to zero fill here, not clobber
+ * region, as we want to ensure that valgrind thinks
+ * all of the outgoing buffer has been written to
+ * so a send() or write() won't trap an error.
+ * JRA.
+ */
+ memset(dest, '\0', dest_len);
+#endif
+
+ if (!(flags & STR_ASCII) && \
+ (flags & STR_UNICODE)) {
+ ret = push_ucs2(NULL, dest, src, dest_len, flags);
+ } else {
+ ret = push_ascii(dest, src, dest_len, flags);
+ }
+ if (ret == (size_t)-1) {
+ return -1;
+ }
+ return ret;
+}
/**
Copy a string from a unicode or ascii source (depending on
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);
+ 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);
}
+/**
+ Copy a string from a unicode or ascii source (depending on
+ the packet flags) to a char* destination.
+ Variant that uses talloc.
+ Flags can have:
+ STR_TERMINATE means the string in src is null terminated.
+ STR_UNICODE means to force as unicode.
+ STR_ASCII use ascii even with unicode packet.
+ STR_NOALIGN means don't do alignment.
+ if STR_TERMINATE is set then src_len is ignored is it is -1
+ src_len is the length of the source area in bytes.
+ Return the number of bytes occupied by the string in src.
+ The resulting string in "dest" is always null terminated.
+**/
+
+size_t pull_string_talloc_fn(const char *function,
+ unsigned int line,
+ TALLOC_CTX *ctx,
+ const void *base_ptr,
+ uint16 smb_flags2,
+ char **ppdest,
+ const void *src,
+ size_t src_len,
+ int flags)
+{
+ 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 || \
+ (smb_flags2 & FLAGS2_UNICODE_STRINGS)))) {
+ return pull_ucs2_base_talloc(ctx,
+ base_ptr,
+ ppdest,
+ src,
+ src_len,
+ flags);
+ }
+ return pull_ascii_base_talloc(ctx,
+ ppdest,
+ src,
+ src_len,
+ flags);
+}
+
+
size_t align_string(const void *base_ptr, const char *p, int flags)
{
if (!(flags & STR_ASCII) && \
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).
+
+ Return INVALID_CODEPOINT if the next character cannot be converted.
+*/
-size_t next_mb_char_size(const char *s)
+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;
+}
+
+/*
+ push a single codepoint into a CH_UNIX string the target string must
+ be able to hold the full character, which is guaranteed if it is at
+ least 5 bytes in size. The caller may pass less than 5 bytes if they
+ are sure the character will fit (for example, you can assume that
+ uppercase/lowercase of a character will not add more than 1 byte)
+
+ return the number of bytes occupied by the CH_UNIX character, or
+ -1 on failure
+*/
+_PUBLIC_ ssize_t push_codepoint(char *str, codepoint_t c)
+{
+ smb_iconv_t descriptor;
+ uint8_t buf[4];
+ size_t ilen, olen;
+ const char *inbuf;
+
+ if (c < 128) {
+ *str = c;
+ return 1;
+ }
+
+ lazy_initialize_conv();
+
+ descriptor = conv_handles[CH_UNIX][CH_UTF16LE];
+ if (descriptor == (smb_iconv_t)-1 || descriptor == (smb_iconv_t)0) {
+ return -1;
+ }
+
+ if (c < 0x10000) {
+ ilen = 2;
+ olen = 5;
+ inbuf = (char *)buf;
+ SSVAL(buf, 0, c);
+ smb_iconv(descriptor, &inbuf, &ilen, &str, &olen);
+ if (ilen != 0) {
+ return -1;
+ }
+ return 5 - olen;
+ }
+
+ c -= 0x10000;
+
+ buf[0] = (c>>10) & 0xFF;
+ buf[1] = (c>>18) | 0xd8;
+ buf[2] = c & 0xFF;
+ buf[3] = ((c>>8) & 0x3) | 0xdc;
+
+ ilen = 4;
+ olen = 5;
+ inbuf = (char *)buf;
+
+ smb_iconv(descriptor, &inbuf, &ilen, &str, &olen);
+ if (ilen != 0) {
+ return -1;
+ }
+ return 5 - olen;
}
+
+
git.samba.org / ira / wip.git / blobdiff