*/
-static smb_iconv_t conv_handles[NUM_CHARSETS][NUM_CHARSETS];
static bool conv_silent; /* Should we do a debug if the conversion fails ? */
-
-/**
- * Return the name of a charset to give to iconv().
- **/
-static const char *charset_name(charset_t ch)
-{
- const char *ret = NULL;
-
- 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();
- else if (ch == CH_UTF8) ret = "UTF8";
-
-#if defined(HAVE_NL_LANGINFO) && defined(CODESET)
- if (ret && !strcmp(ret, "LOCALE")) {
- const char *ln = NULL;
-
-#ifdef HAVE_SETLOCALE
- setlocale(LC_ALL, "");
-#endif
- ln = nl_langinfo(CODESET);
- if (ln) {
- /* Check whether the charset name is supported
- by iconv */
- smb_iconv_t handle = smb_iconv_open(ln,"UCS-2LE");
- if (handle == (smb_iconv_t) -1) {
- DEBUG(5,("Locale charset '%s' unsupported, using ASCII instead\n", ln));
- ln = NULL;
- } else {
- DEBUG(5,("Substituting charset '%s' for LOCALE\n", ln));
- smb_iconv_close(handle);
- }
- }
- ret = ln;
- }
-#endif
-
- if (!ret || !*ret) ret = "ASCII";
- return ret;
-}
+static bool initialized;
void lazy_initialize_conv(void)
{
- static int initialized = False;
-
if (!initialized) {
- initialized = True;
- load_case_tables();
+ load_case_tables_library();
init_iconv();
+ initialized = true;
}
}
**/
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;
- }
- }
- }
+ TALLOC_FREE(global_iconv_convenience);
+ initialized = false;
}
/**
**/
void init_iconv(void)
{
- int c1, c2;
- bool did_reload = False;
-
- /* so that charset_name() works we need to get the UNIX<->UCS2 going
- first */
- 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_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++) {
- const char *n1 = charset_name((charset_t)c1);
- const char *n2 = charset_name((charset_t)c2);
- if (conv_handles[c1][c2] &&
- strcmp(n1, conv_handles[c1][c2]->from_name) == 0 &&
- strcmp(n2, conv_handles[c1][c2]->to_name) == 0)
- continue;
-
- did_reload = True;
-
- if (conv_handles[c1][c2])
- smb_iconv_close(conv_handles[c1][c2]);
-
- conv_handles[c1][c2] = smb_iconv_open(n2,n1);
- 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_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");
- }
- }
- }
- }
-
- if (did_reload) {
- /* XXX: Does this really get called every time the dos
- * codepage changes? */
- /* XXX: Is the did_reload test too strict? */
- conv_silent = True;
- init_valid_table();
- conv_silent = False;
- }
+ global_iconv_convenience = smb_iconv_convenience_reinit(NULL, lp_dos_charset(),
+ lp_unix_charset(), lp_display_charset(),
+ true, global_iconv_convenience);
}
/**
const char* inbuf = (const char*)src;
char* outbuf = (char*)dest;
smb_iconv_t descriptor;
+ struct smb_iconv_convenience *ic;
lazy_initialize_conv();
-
- descriptor = conv_handles[from][to];
+ ic = get_iconv_convenience();
+ descriptor = get_conv_handle(ic, from, to);
if (srclen == (size_t)-1) {
if (from == CH_UTF16LE || from == CH_UTF16BE) {
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) {
if (from == CH_UNIX) {
DEBUG(3,("E2BIG: convert_string(%s,%s): srclen=%u destlen=%u - '%s'\n",
- charset_name(from), charset_name(to),
+ charset_name(ic, from), charset_name(ic, 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),
+ charset_name(ic, from), charset_name(ic, to),
(unsigned int)srclen, (unsigned int)destlen));
}
}
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); */
}
#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
}
}
#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
}
}
#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.
- * (this is a bad interface and needs fixing. JRA).
* @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 *dst, 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;
+ struct smb_iconv_convenience *ic;
*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;
+ }
- lazy_initialize_conv();
+ if (src == NULL || srclen == (size_t)-1) {
+ errno = EINVAL;
+ return false;
+ }
+ if (srclen == 0) {
+ ob = talloc_strdup(ctx, "");
+ if (ob == NULL) {
+ errno = ENOMEM;
+ return false;
+ }
+ *dest = ob;
+ *converted_size = 0;
+ return true;
+ }
- descriptor = conv_handles[from][to];
+ lazy_initialize_conv();
+ ic = get_iconv_convenience();
+ descriptor = get_conv_handle(ic, from, to);
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"));
- if (!ctx)
- SAFE_FREE(outbuf);
- return (size_t)-1;
+ DEBUG(0, ("convert_string_talloc: destlen wrapped !\n"));
+ TALLOC_FREE(outbuf);
+ 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"));
- return (size_t)-1;
+ DEBUG(0, ("convert_string_talloc: realloc failed!\n"));
+ errno = ENOMEM;
+ return false;
}
outbuf = ob;
i_len = srclen;
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 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); */
- if (ctx) {
- TALLOC_FREE(ob);
- } else {
- SAFE_FREE(ob);
- }
- return (size_t)-1;
+ TALLOC_FREE(ob);
+ return false;
}
out:
destlen = destlen - o_len;
- if (ctx) {
- ob = (char *)TALLOC_REALLOC(ctx,ob,destlen);
- } else {
- ob = (char *)SMB_REALLOC(ob,destlen);
+ /* 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_allocate: out of memory!\n"));
- return (size_t)-1;
+ DEBUG(0, ("convert_string_talloc: out of memory!\n"));
+ errno = ENOMEM;
+ return false;
}
*dest = ob;
- return destlen;
+
+ /* 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:
}
}
-/**
- * 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.
- **/
-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;
- 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) {
+ if (!push_ucs2_talloc(talloc_tos(), &buffer, src, &size)) {
return (size_t)-1;
}
+
if (!strupper_w(buffer) && (dest == src)) {
- free(buffer);
+ TALLOC_FREE(buffer);
return srclen;
}
size = convert_string(CH_UTF16LE, CH_UNIX, buffer, size, dest, destlen, True);
- free(buffer);
+ TALLOC_FREE(buffer);
return size;
}
-/**
- strdup() a unix string to upper case.
-**/
-
-char *strdup_upper(const char *s)
-{
- char *out_buffer = SMB_STRDUP(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 (*p) {
- if (*p & 0x80)
- break;
- *q++ = toupper_ascii(*p);
- p++;
- }
-
- if (*p) {
- /* MB case. */
- size_t size;
- smb_ucs2_t *buffer = NULL;
-
- SAFE_FREE(out_buffer);
- size = convert_string_allocate(NULL,
- CH_UNIX,
- CH_UTF16LE,
- s,
- strlen(s) + 1,
- (void **)(void *)&buffer,
- True);
- if (size == (size_t)-1) {
- return NULL;
- }
-
- strupper_w(buffer);
-
- size = convert_string_allocate(NULL,
- CH_UTF16LE,
- CH_UNIX,
- buffer,
- size,
- (void **)(void *)&out_buffer,
- True);
-
- /* Don't need the intermediate buffer
- * anymore.
- */
-
- TALLOC_FREE(buffer);
- if (size == (size_t)-1) {
- return NULL;
- }
- }
-
- return out_buffer;
-}
-
/**
talloc_strdup() a unix string to upper case.
**/
while (*p) {
if (*p & 0x80)
break;
- *q++ = toupper_ascii(*p);
+ *q++ = toupper_ascii_fast(*p);
p++;
}
if (*p) {
/* MB case. */
- size_t size;
+ size_t converted_size, converted_size2;
smb_ucs2_t *ubuf = NULL;
/* We're not using the ascii buffer above. */
TALLOC_FREE(out_buffer);
- size = convert_string_talloc(ctx, CH_UNIX, CH_UTF16LE,
- s, strlen(s)+1,
- (void *)&ubuf,
- True);
- if (size == (size_t)-1) {
+ if (!convert_string_talloc(ctx, CH_UNIX, CH_UTF16LE, s,
+ strlen(s)+1, (void *)&ubuf,
+ &converted_size, True))
+ {
return NULL;
}
strupper_w(ubuf);
- size = convert_string_talloc(ctx, CH_UTF16LE, CH_UNIX,
- ubuf, size,
- (void *)&out_buffer,
- True);
+ 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);
-
- if (size == (size_t)-1) {
- return NULL;
- }
}
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_UTF16LE, src, srclen,
- (void **)(void *)&buffer, True);
- if (size == (size_t)-1 || !buffer) {
+ if (!convert_string_talloc(talloc_tos(), 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_UTF16LE, CH_UNIX, buffer, size, dest, destlen, True);
- SAFE_FREE(buffer);
+ 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;
}
+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))
smb_ucs2_t *buffer;
conv_silent = True;
- buffer_len = push_ucs2_allocate(&buffer, src);
- if (buffer_len == (size_t)-1) {
+ if (!push_ucs2_talloc(talloc_tos(), &buffer, src, &buffer_len)) {
smb_panic("failed to create UCS2 buffer");
}
}
((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.
********************************************************************/
-size_t push_ascii_allocate(char **dest, const char *src)
+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_allocate(NULL, CH_UNIX, CH_DOS, src, src_len, (void **)dest, True);
+ 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.
- Talloc version.
- Uses malloc if TALLOC_CTX is NULL (this is a bad interface and
- needs fixing. JRA).
+ * Talloc version.
*
* The resulting string in "dest" is always null terminated.
*
**/
static size_t pull_ascii_base_talloc(TALLOC_CTX *ctx,
- char **ppdest,
- const void *src,
- size_t src_len,
- int flags)
+ char **ppdest,
+ const void *src,
+ size_t src_len,
+ int flags)
{
char *dest = NULL;
- size_t dest_len = 0;
-
-#ifdef DEVELOPER
- /* Ensure we never use the braindead "malloc" varient. */
- if (ctx == NULL) {
- smb_panic("NULL talloc CTX in pull_ascii_base_talloc\n");
- }
-#endif
+ 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;
(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;
+ }
}
- dest_len = convert_string_allocate(ctx,
- CH_DOS,
- CH_UNIX,
- src,
- src_len,
- &dest,
- True);
+ /* src_len != -1 here. */
- if (dest_len == (size_t)-1) {
+ 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) {
- 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;
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]);
+ smb_ucs2_t v = toupper_m(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_UTF16LE, 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_UTF16LE, 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
}
if (flags & STR_UPPER) {
- tmpbuf = strdup_upper(src);
+ tmpbuf = strupper_talloc(talloc_tos(), src);
if (!tmpbuf) {
return (size_t)-1;
}
}
ret = convert_string(CH_UNIX, CH_UTF8, src, src_len, dest, dest_len, True);
- SAFE_FREE(tmpbuf);
+ TALLOC_FREE(tmpbuf);
return ret;
}
* 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 pull_ucs2(const void *base_ptr, char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
{
size_t ret;
+ size_t ucs2_align_len = 0;
if (dest_len == (size_t)-1) {
/* No longer allow dest_len of -1. */
src = (const void *)((const char *)src + 1);
if (src_len != (size_t)-1)
src_len--;
+ ucs2_align_len = 1;
}
if (flags & STR_TERMINATE) {
dest[0] = 0;
}
- return src_len;
+ return src_len + ucs2_align_len;
}
/**
{
char *dest;
size_t dest_len;
+ size_t ucs2_align_len = 0;
*ppdest = NULL;
src = (const void *)((const char *)src + 1);
if (src_len != (size_t)-1)
src_len--;
+ ucs2_align_len = 1;
}
if (flags & STR_TERMINATE) {
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 */
- if (src_len != (size_t)-1) {
- src_len &= ~1;
- }
+ src_len &= ~1;
- dest_len = convert_string_talloc(ctx,
- CH_UTF16LE,
- CH_UNIX,
- src,
- src_len,
- (void *)&dest,
- True);
- if (dest_len == (size_t)-1) {
+ if (!convert_string_talloc(ctx, CH_UTF16LE, CH_UNIX, src, src_len,
+ (void *)&dest, &dest_len, True)) {
dest_len = 0;
}
- if (src_len == (size_t)-1)
- src_len = dest_len*2;
-
if (dest_len) {
/* Did we already process the terminating zero ? */
if (dest[dest_len-1] != 0) {
}
*ppdest = dest;
- return src_len;
+ return src_len + ucs2_align_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_UTF16LE, CH_UNIX, src, src_len, (void **)dest, True);
-}
-
-/**
- * Copy a string from a UCS2 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
- **/
-size_t pull_ucs2_allocate(char **dest, const smb_ucs2_t *src)
-{
- size_t src_len = (strlen_w(src)+1) * sizeof(smb_ucs2_t);
*dest = NULL;
- return convert_string_allocate(NULL, CH_UTF16LE, 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 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_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src)
+bool pull_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_talloc(ctx, CH_UTF8, 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
+ * 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_allocate(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_allocate(NULL, 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 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
- **/
+ 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_ascii_talloc(TALLOC_CTX *ctx, 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_talloc(ctx, CH_DOS, 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, uint16 flags2,
- 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
if (!(flags & STR_ASCII) && \
((flags & STR_UNICODE || \
(flags2 & FLAGS2_UNICODE_STRINGS)))) {
- return push_ucs2(base_ptr, dest, src, dest_len, flags);
+ 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
return 0;
}
-/*
- 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.
-*/
-
-codepoint_t next_codepoint(const char *str, size_t *size)
-{
- /* 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];
- }
-
- /* We assume that no multi-byte character can take
- more than 5 bytes. This is OK as we only
- support codepoints up to 1M */
-
- 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;
- }
-
- *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;
-}