#include "includes.h"
#include "system/locale.h"
-#include "dynconfig/dynconfig.h"
-#include "param/param.h"
-/**
- * @file
- * @brief Unicode string manipulation
- */
-
-/* these 2 tables define the unicode case handling. They are loaded
- at startup either via mmap() or read() from the lib directory */
-static void *upcase_table;
-static void *lowcase_table;
-
-
-/*******************************************************************
-load the case handling tables
-********************************************************************/
-void load_case_tables(void)
-{
- TALLOC_CTX *mem_ctx;
-
- mem_ctx = talloc_init("load_case_tables");
- if (!mem_ctx) {
- smb_panic("No memory for case_tables");
- }
- upcase_table = map_file(talloc_asprintf(mem_ctx, "%s/upcase.dat", dyn_DATADIR), 0x20000);
- lowcase_table = map_file(talloc_asprintf(mem_ctx, "%s/lowcase.dat", dyn_DATADIR), 0x20000);
- talloc_free(mem_ctx);
- if (upcase_table == NULL) {
- /* try also under codepages for testing purposes */
- upcase_table = map_file("codepages/upcase.dat", 0x20000);
- if (upcase_table == NULL) {
- upcase_table = (void *)-1;
- }
- }
- if (lowcase_table == NULL) {
- /* try also under codepages for testing purposes */
- lowcase_table = map_file("codepages/lowcase.dat", 0x20000);
- if (lowcase_table == NULL) {
- lowcase_table = (void *)-1;
- }
- }
-}
-
-/**
- Convert a codepoint_t to upper case.
-**/
-_PUBLIC_ codepoint_t toupper_w(codepoint_t val)
-{
- if (val < 128) {
- return toupper(val);
- }
- if (upcase_table == NULL) {
- load_case_tables();
- }
- if (upcase_table == (void *)-1) {
- return val;
- }
- if (val & 0xFFFF0000) {
- return val;
- }
- return SVAL(upcase_table, val*2);
-}
-
-/**
- Convert a codepoint_t to lower case.
-**/
-_PUBLIC_ codepoint_t tolower_w(codepoint_t val)
-{
- if (val < 128) {
- return tolower(val);
- }
- if (lowcase_table == NULL) {
- load_case_tables();
- }
- if (lowcase_table == (void *)-1) {
- return val;
- }
- if (val & 0xFFFF0000) {
- return val;
- }
- return SVAL(lowcase_table, val*2);
-}
+struct smb_iconv_convenience *global_iconv_convenience = NULL;
-/**
- compare two codepoints case insensitively
-*/
-_PUBLIC_ int codepoint_cmpi(codepoint_t c1, codepoint_t c2)
+static inline struct smb_iconv_convenience *get_iconv_convenience(void)
{
- if (c1 == c2 ||
- toupper_w(c1) == toupper_w(c2)) {
- return 0;
- }
- return c1 - c2;
+ if (global_iconv_convenience == NULL)
+ global_iconv_convenience = smb_iconv_convenience_init(talloc_autofree_context(), "ASCII", "UTF-8", true);
+ return global_iconv_convenience;
}
/**
{
codepoint_t c1=0, c2=0;
size_t size1, size2;
- struct smb_iconv_convenience *iconv_convenience = lp_iconv_convenience(global_loadparm);
+ struct smb_iconv_convenience *iconv_convenience = get_iconv_convenience();
/* handle null ptr comparisons to simplify the use in qsort */
if (s1 == s2) return 0;
if (s2 == NULL) return 1;
while (*s1 && *s2) {
- c1 = next_codepoint(iconv_convenience, s1, &size1);
- c2 = next_codepoint(iconv_convenience, s2, &size2);
+ c1 = next_codepoint_convenience(iconv_convenience, s1, &size1);
+ c2 = next_codepoint_convenience(iconv_convenience, s2, &size2);
s1 += size1;
s2 += size2;
return strcasecmp(s1, s2);
}
- if (toupper_w(c1) != toupper_w(c2)) {
+ if (toupper_m(c1) != toupper_m(c2)) {
return c1 - c2;
}
}
{
codepoint_t c1=0, c2=0;
size_t size1, size2;
- struct smb_iconv_convenience *iconv_convenience = lp_iconv_convenience(global_loadparm);
+ struct smb_iconv_convenience *iconv_convenience = get_iconv_convenience();
/* handle null ptr comparisons to simplify the use in qsort */
if (s1 == s2) return 0;
while (*s1 && *s2 && n) {
n--;
- c1 = next_codepoint(iconv_convenience, s1, &size1);
- c2 = next_codepoint(iconv_convenience, s2, &size2);
+ c1 = next_codepoint_convenience(iconv_convenience, s1, &size1);
+ c2 = next_codepoint_convenience(iconv_convenience, s2, &size2);
s1 += size1;
s2 += size2;
return strcasecmp(s1, s2);
}
- if (toupper_w(c1) != toupper_w(c2)) {
+ if (toupper_m(c1) != toupper_m(c2)) {
return c1 - c2;
}
}
*
* @note The comparison is case-insensitive.
**/
-_PUBLIC_ bool strequal_w(const char *s1, const char *s2)
+_PUBLIC_ bool strequal_m(const char *s1, const char *s2)
{
return strcasecmp_m(s1,s2) == 0;
}
/**
Compare 2 strings (case sensitive).
**/
-_PUBLIC_ bool strcsequal_w(const char *s1,const char *s2)
+_PUBLIC_ bool strcsequal_m(const char *s1,const char *s2)
{
if (s1 == s2)
return true;
**/
_PUBLIC_ void string_replace_m(char *s, char oldc, char newc)
{
+ struct smb_iconv_convenience *ic = get_iconv_convenience();
while (s && *s) {
size_t size;
- codepoint_t c = next_codepoint(lp_iconv_convenience(global_loadparm), s, &size);
+ codepoint_t c = next_codepoint_convenience(ic, s, &size);
if (c == oldc) {
*s = newc;
}
_PUBLIC_ size_t strlen_m(const char *s)
{
size_t count = 0;
+ struct smb_iconv_convenience *ic = get_iconv_convenience();
if (!s) {
return 0;
while (*s) {
size_t c_size;
- codepoint_t c = next_codepoint(lp_iconv_convenience(global_loadparm), s, &c_size);
+ codepoint_t c = next_codepoint_convenience(ic, s, &c_size);
if (c < 0x10000) {
count += 1;
} else {
**/
_PUBLIC_ char *strchr_m(const char *s, char c)
{
+ struct smb_iconv_convenience *ic = get_iconv_convenience();
if (s == NULL) {
return NULL;
}
while (*s) {
size_t size;
- codepoint_t c2 = next_codepoint(lp_iconv_convenience(global_loadparm), s, &size);
+ codepoint_t c2 = next_codepoint_convenience(ic, s, &size);
if (c2 == c) {
return discard_const_p(char, s);
}
*/
_PUBLIC_ char *strrchr_m(const char *s, char c)
{
+ struct smb_iconv_convenience *ic = get_iconv_convenience();
char *ret = NULL;
if (s == NULL) {
while (*s) {
size_t size;
- codepoint_t c2 = next_codepoint(lp_iconv_convenience(global_loadparm), s, &size);
+ codepoint_t c2 = next_codepoint_convenience(ic, s, &size);
if (c2 == c) {
ret = discard_const_p(char, s);
}
*/
_PUBLIC_ bool strhaslower(const char *string)
{
+ struct smb_iconv_convenience *ic = get_iconv_convenience();
while (*string) {
size_t c_size;
codepoint_t s;
codepoint_t t;
- s = next_codepoint(lp_iconv_convenience(global_loadparm), string, &c_size);
+ s = next_codepoint_convenience(ic, string, &c_size);
string += c_size;
- t = toupper_w(s);
+ t = toupper_m(s);
if (s != t) {
return true; /* that means it has lower case chars */
*/
_PUBLIC_ bool strhasupper(const char *string)
{
+ struct smb_iconv_convenience *ic = get_iconv_convenience();
while (*string) {
size_t c_size;
codepoint_t s;
codepoint_t t;
- s = next_codepoint(lp_iconv_convenience(global_loadparm), string, &c_size);
+ s = next_codepoint_convenience(ic, string, &c_size);
string += c_size;
- t = tolower_w(s);
+ t = tolower_m(s);
if (s != t) {
return true; /* that means it has upper case chars */
{
size_t size=0;
char *dest;
- struct smb_iconv_convenience *iconv_convenience = lp_iconv_convenience(global_loadparm);
+ struct smb_iconv_convenience *iconv_convenience = get_iconv_convenience();
/* this takes advantage of the fact that upper/lower can't
change the length of a character by more than 1 byte */
while (*src) {
size_t c_size;
- codepoint_t c = next_codepoint(iconv_convenience, src, &c_size);
+ codepoint_t c = next_codepoint_convenience(iconv_convenience, src, &c_size);
src += c_size;
- c = tolower_w(c);
+ c = tolower_m(c);
c_size = push_codepoint(iconv_convenience, dest+size, c);
if (c_size == -1) {
{
size_t size=0;
char *dest;
- struct smb_iconv_convenience *iconv_convenience = lp_iconv_convenience(global_loadparm);
+ struct smb_iconv_convenience *iconv_convenience = get_iconv_convenience();
if (!src) {
return NULL;
while (*src && n--) {
size_t c_size;
- codepoint_t c = next_codepoint(iconv_convenience, src, &c_size);
+ codepoint_t c = next_codepoint_convenience(iconv_convenience, src, &c_size);
src += c_size;
- c = toupper_w(c);
+ c = toupper_m(c);
c_size = push_codepoint(iconv_convenience, dest+size, c);
if (c_size == -1) {
if (!*s)
return;
- iconv_convenience = lp_iconv_convenience(global_loadparm);
+ iconv_convenience = get_iconv_convenience();
d = s;
while (*s) {
size_t c_size, c_size2;
- codepoint_t c = next_codepoint(iconv_convenience, s, &c_size);
- c_size2 = push_codepoint(iconv_convenience, d, tolower_w(c));
+ codepoint_t c = next_codepoint_convenience(iconv_convenience, s, &c_size);
+ c_size2 = push_codepoint(iconv_convenience, d, tolower_m(c));
if (c_size2 > c_size) {
DEBUG(0,("FATAL: codepoint 0x%x (0x%x) expanded from %d to %d bytes in strlower_m\n",
- c, tolower_w(c), (int)c_size, (int)c_size2));
+ c, tolower_m(c), (int)c_size, (int)c_size2));
smb_panic("codepoint expansion in strlower_m\n");
}
s += c_size;
if (!*s)
return;
- iconv_convenience = lp_iconv_convenience(global_loadparm);
+ iconv_convenience = get_iconv_convenience();
d = s;
while (*s) {
size_t c_size, c_size2;
- codepoint_t c = next_codepoint(iconv_convenience, s, &c_size);
- c_size2 = push_codepoint(iconv_convenience, d, toupper_w(c));
+ codepoint_t c = next_codepoint_convenience(iconv_convenience, s, &c_size);
+ c_size2 = push_codepoint(iconv_convenience, d, toupper_m(c));
if (c_size2 > c_size) {
DEBUG(0,("FATAL: codepoint 0x%x (0x%x) expanded from %d to %d bytes in strupper_m\n",
- c, toupper_w(c), (int)c_size, (int)c_size2));
+ c, toupper_m(c), (int)c_size, (int)c_size2));
smb_panic("codepoint expansion in strupper_m\n");
}
s += c_size;
/**
Find the number of 'c' chars in a string
**/
-_PUBLIC_ size_t count_chars_w(const char *s, char c)
+_PUBLIC_ size_t count_chars_m(const char *s, char c)
{
+ struct smb_iconv_convenience *ic = get_iconv_convenience();
size_t count = 0;
while (*s) {
size_t size;
- codepoint_t c2 = next_codepoint(lp_iconv_convenience(global_loadparm), s, &size);
+ codepoint_t c2 = next_codepoint_convenience(ic, s, &size);
if (c2 == c) count++;
s += size;
}
}
+/**
+ * Copy a string from a char* unix src to a dos codepage string destination.
+ *
+ * @return the number of bytes occupied by the string in the destination.
+ *
+ * @param flags can include
+ * <dl>
+ * <dt>STR_TERMINATE</dt> <dd>means include the null termination</dd>
+ * <dt>STR_UPPER</dt> <dd>means uppercase in the destination</dd>
+ * </dl>
+ *
+ * @param dest_len the maximum length in bytes allowed in the
+ * destination. If @p dest_len is -1 then no maximum is used.
+ **/
+static ssize_t push_ascii(void *dest, const char *src, size_t dest_len, int flags)
+{
+ size_t src_len;
+ ssize_t ret;
+
+ if (flags & STR_UPPER) {
+ char *tmpbuf = strupper_talloc(NULL, src);
+ if (tmpbuf == NULL) {
+ return -1;
+ }
+ ret = push_ascii(dest, tmpbuf, dest_len, flags & ~STR_UPPER);
+ talloc_free(tmpbuf);
+ return ret;
+ }
+
+ src_len = strlen(src);
+
+ if (flags & (STR_TERMINATE | STR_TERMINATE_ASCII))
+ src_len++;
+
+ return convert_string(CH_UNIX, CH_DOS, src, src_len, dest, dest_len);
+}
+
+/**
+ * Copy a string from a unix char* src to an ASCII 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
+ * or -1 in case of error.
+ **/
+_PUBLIC_ ssize_t push_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_UNIX, CH_DOS, src, src_len, (void **)dest);
+}
+
+
+/**
+ * Copy a string from a dos codepage source to a unix char* destination.
+ *
+ * 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 ssize_t pull_ascii(char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
+{
+ size_t ret;
+
+ if (flags & (STR_TERMINATE | STR_TERMINATE_ASCII)) {
+ 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;
+ }
+ }
+
+ ret = convert_string(CH_DOS, CH_UNIX, src, src_len, dest, dest_len);
+
+ if (dest_len)
+ dest[MIN(ret, dest_len-1)] = 0;
+
+ return src_len;
+}
+
+/**
+ * Copy a string from a char* src to a unicode destination.
+ *
+ * @returns the number of bytes occupied by the string in the destination.
+ *
+ * @param flags can have:
+ *
+ * <dl>
+ * <dt>STR_TERMINATE <dd>means include the null termination.
+ * <dt>STR_UPPER <dd>means uppercase in the destination.
+ * <dt>STR_NOALIGN <dd>means don't do alignment.
+ * </dl>
+ *
+ * @param dest_len is the maximum length allowed in the
+ * destination. If dest_len is -1 then no maxiumum is used.
+ **/
+static ssize_t push_ucs2(void *dest, const char *src, size_t dest_len, int flags)
+{
+ size_t len=0;
+ size_t src_len = strlen(src);
+ size_t ret;
+
+ if (flags & STR_UPPER) {
+ char *tmpbuf = strupper_talloc(NULL, src);
+ if (tmpbuf == NULL) {
+ return -1;
+ }
+ ret = push_ucs2(dest, tmpbuf, dest_len, flags & ~STR_UPPER);
+ talloc_free(tmpbuf);
+ return ret;
+ }
+
+ if (flags & STR_TERMINATE)
+ src_len++;
+
+ if (ucs2_align(NULL, dest, flags)) {
+ *(char *)dest = 0;
+ dest = (void *)((char *)dest + 1);
+ if (dest_len) dest_len--;
+ len++;
+ }
+
+ /* ucs2 is always a multiple of 2 bytes */
+ dest_len &= ~1;
+
+ ret = convert_string(CH_UNIX, CH_UTF16, src, src_len, dest, dest_len);
+ if (ret == (size_t)-1) {
+ return 0;
+ }
+
+ len += ret;
+
+ return len;
+}
+
+
+/**
+ * Copy a string from a unix char* src to a UCS2 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
+ * or -1 in case of error.
+ **/
+_PUBLIC_ ssize_t push_ucs2_talloc(TALLOC_CTX *ctx, void **dest, const char *src)
+{
+ size_t src_len = strlen(src)+1;
+ *dest = NULL;
+ return convert_string_talloc(ctx, CH_UNIX, CH_UTF16, src, src_len, dest);
+}
+
+
+/**
+ * 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
+ *
+ * @returns The number of bytes occupied by the string in the destination
+ **/
+
+_PUBLIC_ ssize_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);
+}
+
+/**
+ Copy a string from a ucs2 source to a unix char* destination.
+ 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.
+**/
+
+static size_t pull_ucs2(char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
+{
+ size_t ret;
+
+ if (ucs2_align(NULL, src, flags)) {
+ src = (const void *)((const char *)src + 1);
+ if (src_len > 0)
+ src_len--;
+ }
+
+ if (flags & STR_TERMINATE) {
+ if (src_len == (size_t)-1) {
+ src_len = utf16_len(src);
+ } else {
+ src_len = utf16_len_n(src, src_len);
+ }
+ }
+
+ /* ucs2 is always a multiple of 2 bytes */
+ if (src_len != (size_t)-1)
+ src_len &= ~1;
+
+ ret = convert_string(CH_UTF16, CH_UNIX, src, src_len, dest, dest_len);
+ if (dest_len)
+ dest[MIN(ret, dest_len-1)] = 0;
+
+ return src_len;
+}
+
+/**
+ * Copy a string from a ASCII 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
+ **/
+
+_PUBLIC_ ssize_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);
+}
+
+/**
+ * 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
+ *
+ * @returns The number of bytes occupied by the string in the destination
+ **/
+
+_PUBLIC_ ssize_t pull_ucs2_talloc(TALLOC_CTX *ctx, char **dest, const void *src)
+{
+ size_t src_len = utf16_len(src);
+ *dest = NULL;
+ return convert_string_talloc(ctx, CH_UTF16, CH_UNIX, src, src_len, (void **)dest);
+}
+
+/**
+ * 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
+ *
+ * @returns The number of bytes occupied by the string in the destination
+ **/
+
+_PUBLIC_ ssize_t pull_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_UTF8, CH_UNIX, src, src_len, (void **)dest);
+}
+
+/**
+ Copy a string from a char* src to a unicode or ascii
+ dos codepage destination choosing unicode or ascii based on the
+ flags in the SMB buffer starting at base_ptr.
+ 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.
+**/
+
+_PUBLIC_ ssize_t push_string(void *dest, const char *src, size_t dest_len, int flags)
+{
+ if (flags & STR_ASCII) {
+ return push_ascii(dest, src, dest_len, flags);
+ } else if (flags & STR_UNICODE) {
+ return push_ucs2(dest, src, dest_len, flags);
+ } else {
+ smb_panic("push_string requires either STR_ASCII or STR_UNICODE flag to be set");
+ return -1;
+ }
+}
+
+
+/**
+ Copy a string from a unicode or ascii source (depending on
+ the packet flags) to a char* destination.
+ 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.
+**/
+
+_PUBLIC_ ssize_t pull_string(char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
+{
+ if (flags & STR_ASCII) {
+ return pull_ascii(dest, src, dest_len, src_len, flags);
+ } else if (flags & STR_UNICODE) {
+ return pull_ucs2(dest, src, dest_len, src_len, flags);
+ } else {
+ smb_panic("pull_string requires either STR_ASCII or STR_UNICODE flag to be set");
+ return -1;
+ }
+}
+
+
+/**
+ * Convert string from one encoding to another, making error checking etc
+ *
+ * @param src pointer to source string (multibyte or singlebyte)
+ * @param srclen length of the source string in bytes
+ * @param dest pointer to destination string (multibyte or singlebyte)
+ * @param destlen maximal length allowed for string
+ * @returns the number of bytes occupied in the destination
+ **/
+_PUBLIC_ ssize_t convert_string(charset_t from, charset_t to,
+ void const *src, size_t srclen,
+ void *dest, size_t destlen)
+{
+ return convert_string_convenience(get_iconv_convenience(), from, to,
+ src, srclen,
+ dest, destlen);
+}
+
+/**
+ * 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.
+ **/
+
+_PUBLIC_ ssize_t convert_string_talloc(TALLOC_CTX *ctx,
+ charset_t from, charset_t to,
+ void const *src, size_t srclen,
+ void **dest)
+{
+ return convert_string_talloc_convenience(ctx, get_iconv_convenience(),
+ from, to, src, srclen, dest);
+}
+
+
+_PUBLIC_ codepoint_t next_codepoint(const char *str, size_t *size)
+{
+ return next_codepoint_convenience(get_iconv_convenience(), str, size);
+}