2 Unix SMB/CIFS implementation.
4 Copyright (C) Andrew Tridgell 2001
5 Copyright (C) Jelmer Vernooij 2002
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with this program. If not, see <http://www.gnu.org/licenses/>.
21 /* This is a public header file that is installed as part of Samba.
22 * If you remove any functions or change their signature, update
23 * the so version number. */
30 /* this defines the charset types used in samba */
31 typedef enum {CH_UTF16LE=0, CH_UTF16=0, CH_UNIX, CH_DISPLAY, CH_DOS, CH_UTF8, CH_UTF16BE, CH_UTF16MUNGED} charset_t;
33 #define NUM_CHARSETS 7
36 * SMB UCS2 (16-bit unicode) internal type.
37 * smb_ucs2_t is *always* in little endian format.
40 typedef uint16_t smb_ucs2_t;
42 #ifdef WORDS_BIGENDIAN
48 /* turn a 7 bit character into a ucs2 character */
49 #define UCS2_CHAR(c) ((c) << UCS2_SHIFT)
52 * for each charset we have a function that pulls from that charset to
53 * a ucs2 buffer, and a function that pushes to a ucs2 buffer
56 struct charset_functions {
58 size_t (*pull)(void *, const char **inbuf, size_t *inbytesleft,
59 char **outbuf, size_t *outbytesleft);
60 size_t (*push)(void *, const char **inbuf, size_t *inbytesleft,
61 char **outbuf, size_t *outbytesleft);
62 struct charset_functions *prev, *next;
65 /* this type is used for manipulating unicode codepoints */
66 typedef uint32_t codepoint_t;
68 #define INVALID_CODEPOINT ((codepoint_t)-1)
71 * This is auxiliary struct used by source/script/gen-8-bit-gap.sh script
72 * during generation of an encoding table for charset module
75 struct charset_gap_table {
82 /* generic iconv conversion structure */
83 typedef struct smb_iconv_s {
84 size_t (*direct)(void *cd, const char **inbuf, size_t *inbytesleft,
85 char **outbuf, size_t *outbytesleft);
86 size_t (*pull)(void *cd, const char **inbuf, size_t *inbytesleft,
87 char **outbuf, size_t *outbytesleft);
88 size_t (*push)(void *cd, const char **inbuf, size_t *inbytesleft,
89 char **outbuf, size_t *outbytesleft);
90 void *cd_direct, *cd_pull, *cd_push;
91 char *from_name, *to_name;
94 /* string manipulation flags */
95 #define STR_TERMINATE 1
99 #define STR_NOALIGN 16
100 #define STR_NO_RANGE_CHECK 32
101 #define STR_LEN8BIT 64
102 #define STR_TERMINATE_ASCII 128 /* only terminate if ascii */
103 #define STR_LEN_NOTERM 256 /* the length field is the unterminated length */
105 struct loadparm_context;
106 struct smb_iconv_handle;
108 char *strchr_m(const char *s, char c);
110 * Calculate the number of units (8 or 16-bit, depending on the
111 * destination charset), that would be needed to convert the input
112 * string which is expected to be in in src_charset encoding to the
113 * destination charset (which should be a unicode charset).
115 size_t strlen_m_ext_handle(struct smb_iconv_handle *ic,
116 const char *s, charset_t src_charset, charset_t dst_charset);
117 size_t strlen_m_ext(const char *s, charset_t src_charset, charset_t dst_charset);
118 size_t strlen_m_ext_term(const char *s, charset_t src_charset,
119 charset_t dst_charset);
120 size_t strlen_m_term(const char *s);
121 size_t strlen_m_term_null(const char *s);
122 size_t strlen_m(const char *s);
123 char *alpha_strcpy(char *dest, const char *src, const char *other_safe_chars, size_t maxlength);
124 void string_replace_m(char *s, char oldc, char newc);
125 bool strcsequal(const char *s1,const char *s2);
126 bool strequal_m(const char *s1, const char *s2);
127 int strncasecmp_m(const char *s1, const char *s2, size_t n);
128 int strncasecmp_m_handle(struct smb_iconv_handle *iconv_handle,
129 const char *s1, const char *s2, size_t n);
130 bool next_token(const char **ptr,char *buff, const char *sep, size_t bufsize);
131 int strcasecmp_m_handle(struct smb_iconv_handle *iconv_handle,
132 const char *s1, const char *s2);
133 int strcasecmp_m(const char *s1, const char *s2);
134 size_t count_chars_m(const char *s, char c);
135 char *strupper_talloc(TALLOC_CTX *ctx, const char *src);
136 char *talloc_strdup_upper(TALLOC_CTX *ctx, const char *src);
137 char *strupper_talloc_n_handle(struct smb_iconv_handle *iconv_handle,
138 TALLOC_CTX *ctx, const char *src, size_t n);
139 char *strupper_talloc_n(TALLOC_CTX *ctx, const char *src, size_t n);
140 char *strlower_talloc_handle(struct smb_iconv_handle *iconv_handle,
141 TALLOC_CTX *ctx, const char *src);
142 char *strlower_talloc(TALLOC_CTX *ctx, const char *src);
143 bool strhasupper(const char *string);
144 bool strhaslower_handle(struct smb_iconv_handle *ic,
146 bool strhaslower(const char *string);
147 bool strhasupper_handle(struct smb_iconv_handle *ic,
149 char *strrchr_m(const char *s, char c);
150 char *strchr_m(const char *s, char c);
151 char *strstr_m(const char *src, const char *findstr);
153 bool push_ascii_talloc(TALLOC_CTX *ctx, char **dest, const char *src, size_t *converted_size);
154 bool push_ucs2_talloc(TALLOC_CTX *ctx, smb_ucs2_t **dest, const char *src, size_t *converted_size);
155 bool push_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src, size_t *converted_size);
156 bool pull_ascii_talloc(TALLOC_CTX *ctx, char **dest, const char *src, size_t *converted_size);
157 bool pull_ucs2_talloc(TALLOC_CTX *ctx, char **dest, const smb_ucs2_t *src, size_t *converted_size);
158 bool pull_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src, size_t *converted_size);
159 ssize_t push_string(void *dest, const char *src, size_t dest_len, int flags);
160 ssize_t pull_string(char *dest, const void *src, size_t dest_len, size_t src_len, int flags);
162 bool convert_string_talloc(TALLOC_CTX *ctx,
163 charset_t from, charset_t to,
164 void const *src, size_t srclen,
165 void *dest, size_t *converted_size);
167 bool convert_string(charset_t from, charset_t to,
168 void const *src, size_t srclen,
169 void *dest, size_t destlen,
170 size_t *converted_size);
171 bool convert_string_error(charset_t from, charset_t to,
172 void const *src, size_t srclen,
173 void *dest, size_t destlen,
174 size_t *converted_size);
176 ssize_t iconv_talloc(TALLOC_CTX *mem_ctx,
178 void const *src, size_t srclen,
181 extern struct smb_iconv_handle *global_iconv_handle;
182 struct smb_iconv_handle *get_iconv_handle(void);
183 struct smb_iconv_handle *get_iconv_testing_handle(TALLOC_CTX *mem_ctx,
184 const char *dos_charset,
185 const char *unix_charset,
186 const char *display_charset);
187 smb_iconv_t get_conv_handle(struct smb_iconv_handle *ic,
188 charset_t from, charset_t to);
189 const char *charset_name(struct smb_iconv_handle *ic, charset_t ch);
191 codepoint_t next_codepoint_ext(const char *str, charset_t src_charset,
193 codepoint_t next_codepoint(const char *str, size_t *size);
194 ssize_t push_codepoint(char *str, codepoint_t c);
197 codepoint_t next_codepoint_handle_ext(struct smb_iconv_handle *ic,
198 const char *str, charset_t src_charset,
200 codepoint_t next_codepoint_handle(struct smb_iconv_handle *ic,
201 const char *str, size_t *size);
202 ssize_t push_codepoint_handle(struct smb_iconv_handle *ic,
203 char *str, codepoint_t c);
205 codepoint_t toupper_m(codepoint_t val);
206 codepoint_t tolower_m(codepoint_t val);
207 bool islower_m(codepoint_t val);
208 bool isupper_m(codepoint_t val);
209 int codepoint_cmpi(codepoint_t c1, codepoint_t c2);
211 /* Iconv convenience functions */
212 struct smb_iconv_handle *smb_iconv_handle_reinit(TALLOC_CTX *mem_ctx,
213 const char *dos_charset,
214 const char *unix_charset,
215 const char *display_charset,
217 struct smb_iconv_handle *old_ic);
219 bool convert_string_handle(struct smb_iconv_handle *ic,
220 charset_t from, charset_t to,
221 void const *src, size_t srclen,
222 void *dest, size_t destlen, size_t *converted_size);
223 bool convert_string_error_handle(struct smb_iconv_handle *ic,
224 charset_t from, charset_t to,
225 void const *src, size_t srclen,
226 void *dest, size_t destlen,
227 size_t *converted_size);
229 bool convert_string_talloc_handle(TALLOC_CTX *ctx,
230 struct smb_iconv_handle *ic,
231 charset_t from, charset_t to,
232 void const *src, size_t srclen,
233 void *dest, size_t *converted_size);
235 smb_iconv_t smb_iconv_open(const char *tocode, const char *fromcode);
236 int smb_iconv_close(smb_iconv_t cd);
237 size_t smb_iconv(smb_iconv_t cd,
238 const char **inbuf, size_t *inbytesleft,
239 char **outbuf, size_t *outbytesleft);
240 smb_iconv_t smb_iconv_open_ex(TALLOC_CTX *mem_ctx, const char *tocode,
241 const char *fromcode, bool native_iconv);
243 void load_case_tables(void);
244 void load_case_tables_library(void);
245 bool smb_register_charset(const struct charset_functions *funcs_in);
247 /* The following definitions come from util_unistr_w.c */
249 size_t strlen_w(const smb_ucs2_t *src);
250 size_t strnlen_w(const smb_ucs2_t *src, size_t max);
251 smb_ucs2_t *strchr_w(const smb_ucs2_t *s, smb_ucs2_t c);
252 smb_ucs2_t *strchr_wa(const smb_ucs2_t *s, char c);
253 smb_ucs2_t *strrchr_w(const smb_ucs2_t *s, smb_ucs2_t c);
254 smb_ucs2_t *strnrchr_w(const smb_ucs2_t *s, smb_ucs2_t c, unsigned int n);
255 smb_ucs2_t *strstr_w(const smb_ucs2_t *s, const smb_ucs2_t *ins);
256 bool strlower_w(smb_ucs2_t *s);
257 bool strupper_w(smb_ucs2_t *s);
258 int strcmp_w(const smb_ucs2_t *a, const smb_ucs2_t *b);
259 int strcasecmp_w(const smb_ucs2_t *a, const smb_ucs2_t *b);
260 int strncasecmp_w(const smb_ucs2_t *a, const smb_ucs2_t *b, size_t len);
261 int strcmp_wa(const smb_ucs2_t *a, const char *b);
262 int toupper_ascii(int c);
263 int tolower_ascii(int c);
264 int isupper_ascii(int c);
265 int islower_ascii(int c);
268 * Define stub for charset module which implements 8-bit encoding with gaps.
269 * Encoding tables for such module should be produced from glibc's CHARMAPs
270 * using script source/script/gen-8bit-gap.sh
271 * CHARSETNAME is CAPITALIZED charset name
274 #define SMB_GENERATE_CHARSET_MODULE_8_BIT_GAP(CHARSETNAME) \
275 static size_t CHARSETNAME ## _push(void *cd, const char **inbuf, size_t *inbytesleft, \
276 char **outbuf, size_t *outbytesleft) \
278 while (*inbytesleft >= 2 && *outbytesleft >= 1) { \
282 uint16 ch = SVAL(*inbuf,0); \
284 for (i=0; from_idx[i].start != 0xffff; i++) { \
285 if ((from_idx[i].start <= ch) && (from_idx[i].end >= ch)) { \
286 ((unsigned char*)(*outbuf))[0] = from_ucs2[from_idx[i].idx+ch]; \
287 (*inbytesleft) -= 2; \
288 (*outbytesleft) -= 1; \
302 if (*inbytesleft == 1) { \
307 if (*inbytesleft > 1) { \
315 static size_t CHARSETNAME ## _pull(void *cd, const char **inbuf, size_t *inbytesleft, \
316 char **outbuf, size_t *outbytesleft) \
318 while (*inbytesleft >= 1 && *outbytesleft >= 2) { \
319 SSVAL(*outbuf, 0, to_ucs2[((unsigned char*)(*inbuf))[0]]); \
320 (*inbytesleft) -= 1; \
321 (*outbytesleft) -= 2; \
326 if (*inbytesleft > 0) { \
334 struct charset_functions CHARSETNAME ## _functions = \
335 {#CHARSETNAME, CHARSETNAME ## _pull, CHARSETNAME ## _push}; \
337 NTSTATUS charset_ ## CHARSETNAME ## _init(void); \
338 NTSTATUS charset_ ## CHARSETNAME ## _init(void) \
340 if (!smb_register_charset(& CHARSETNAME ## _functions)) { \
341 return NT_STATUS_INTERNAL_ERROR; \
343 return NT_STATUS_OK; \
347 #endif /* __CHARSET_H__ */