Line data Source code
1 : /*
2 : Unix SMB/CIFS implementation.
3 : Samba utility functions
4 : Copyright (C) Andrew Tridgell 1992-2001
5 : Copyright (C) Simo Sorce 2001
6 :
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.
11 :
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.
16 :
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/>.
19 : */
20 :
21 : #include "replace.h"
22 : #include "system/locale.h"
23 : #include "charset.h"
24 : #include "lib/util/byteorder.h"
25 : #include "lib/util/fault.h"
26 :
27 : /**
28 : String replace.
29 : NOTE: oldc and newc must be 7 bit characters
30 : **/
31 5 : _PUBLIC_ void string_replace_m(char *s, char oldc, char newc)
32 : {
33 5 : struct smb_iconv_handle *ic = get_iconv_handle();
34 24 : while (s && *s) {
35 : size_t size;
36 14 : codepoint_t c = next_codepoint_handle(ic, s, &size);
37 14 : if (c == oldc) {
38 5 : *s = newc;
39 : }
40 14 : s += size;
41 : }
42 5 : }
43 :
44 : /**
45 : Convert a string to lower case, allocated with talloc
46 : **/
47 3848560 : _PUBLIC_ char *strlower_talloc_handle(struct smb_iconv_handle *iconv_handle,
48 : TALLOC_CTX *ctx, const char *src)
49 : {
50 3848560 : size_t size=0;
51 : char *dest;
52 :
53 3848560 : if(src == NULL) {
54 44 : return NULL;
55 : }
56 :
57 : /* this takes advantage of the fact that upper/lower can't
58 : change the length of a character by more than 1 byte */
59 3848516 : dest = talloc_array(ctx, char, 2*(strlen(src))+1);
60 3848516 : if (dest == NULL) {
61 0 : return NULL;
62 : }
63 :
64 79832461 : while (*src) {
65 : size_t c_size;
66 72435202 : codepoint_t c = next_codepoint_handle(iconv_handle, src, &c_size);
67 72435202 : src += c_size;
68 :
69 72435202 : c = tolower_m(c);
70 :
71 72435202 : c_size = push_codepoint_handle(iconv_handle, dest+size, c);
72 72435202 : if (c_size == -1) {
73 0 : talloc_free(dest);
74 0 : return NULL;
75 : }
76 72435202 : size += c_size;
77 : }
78 :
79 3848516 : dest[size] = 0;
80 :
81 : /* trim it so talloc_append_string() works */
82 3848516 : dest = talloc_realloc(ctx, dest, char, size+1);
83 :
84 3848516 : talloc_set_name_const(dest, dest);
85 :
86 3848516 : return dest;
87 : }
88 :
89 3848554 : _PUBLIC_ char *strlower_talloc(TALLOC_CTX *ctx, const char *src)
90 : {
91 3848554 : struct smb_iconv_handle *iconv_handle = get_iconv_handle();
92 3848554 : return strlower_talloc_handle(iconv_handle, ctx, src);
93 : }
94 :
95 : /**
96 : Convert a string to UPPER case, allocated with talloc
97 : source length limited to n bytes, iconv handle supplied
98 : **/
99 598501178 : _PUBLIC_ char *strupper_talloc_n_handle(struct smb_iconv_handle *iconv_handle,
100 : TALLOC_CTX *ctx, const char *src, size_t n)
101 : {
102 598501178 : size_t size=0;
103 : char *dest;
104 :
105 598501178 : if (!src) {
106 164748 : return NULL;
107 : }
108 :
109 : /* this takes advantage of the fact that upper/lower can't
110 : change the length of a character by more than 1 byte */
111 598298398 : dest = talloc_array(ctx, char, 2*(n+1));
112 598298398 : if (dest == NULL) {
113 0 : return NULL;
114 : }
115 :
116 8905813309 : while (n && *src) {
117 : size_t c_size;
118 7890396871 : codepoint_t c = next_codepoint_handle_ext(iconv_handle, src, n,
119 : CH_UNIX, &c_size);
120 7890396871 : src += c_size;
121 7890396871 : n -= c_size;
122 :
123 7890396871 : c = toupper_m(c);
124 :
125 7890396871 : c_size = push_codepoint_handle(iconv_handle, dest+size, c);
126 7890396871 : if (c_size == -1) {
127 2 : talloc_free(dest);
128 2 : return NULL;
129 : }
130 7890396869 : size += c_size;
131 : }
132 :
133 598298396 : dest[size] = 0;
134 :
135 : /* trim it so talloc_append_string() works */
136 598298396 : dest = talloc_realloc(ctx, dest, char, size+1);
137 :
138 598298396 : talloc_set_name_const(dest, dest);
139 :
140 598298396 : return dest;
141 : }
142 :
143 : /**
144 : Convert a string to UPPER case, allocated with talloc
145 : source length limited to n bytes
146 : **/
147 598501172 : _PUBLIC_ char *strupper_talloc_n(TALLOC_CTX *ctx, const char *src, size_t n)
148 : {
149 598501172 : struct smb_iconv_handle *iconv_handle = get_iconv_handle();
150 598501172 : return strupper_talloc_n_handle(iconv_handle, ctx, src, n);
151 : }
152 : /**
153 : Convert a string to UPPER case, allocated with talloc
154 : **/
155 5891093 : _PUBLIC_ char *strupper_talloc(TALLOC_CTX *ctx, const char *src)
156 : {
157 5891093 : return strupper_talloc_n(ctx, src, src?strlen(src):0);
158 : }
159 :
160 : /**
161 : talloc_strdup() a unix string to upper case.
162 : **/
163 3421773 : _PUBLIC_ char *talloc_strdup_upper(TALLOC_CTX *ctx, const char *src)
164 : {
165 3421773 : return strupper_talloc(ctx, src);
166 : }
167 :
168 : /**
169 : Find the number of 'c' chars in a string
170 : **/
171 4 : _PUBLIC_ size_t count_chars_m(const char *s, char c)
172 : {
173 4 : struct smb_iconv_handle *ic = get_iconv_handle();
174 4 : size_t count = 0;
175 :
176 17 : while (*s) {
177 : size_t size;
178 9 : codepoint_t c2 = next_codepoint_handle(ic, s, &size);
179 9 : if (c2 == c) count++;
180 9 : s += size;
181 : }
182 :
183 4 : return count;
184 : }
185 :
186 2077530 : size_t ucs2_align(const void *base_ptr, const void *p, int flags)
187 : {
188 2077530 : if (flags & (STR_NOALIGN|STR_ASCII)) {
189 127085 : return 0;
190 : }
191 1950352 : return PTR_DIFF(p, base_ptr) & 1;
192 : }
193 :
194 : /**
195 : return the number of bytes occupied by a buffer in CH_UTF16 format
196 : the result includes the null termination
197 : **/
198 0 : size_t utf16_len(const void *buf)
199 : {
200 : size_t len;
201 :
202 0 : for (len = 0; SVAL(buf,len); len += 2) ;
203 :
204 0 : return len + 2;
205 : }
206 :
207 : /**
208 : return the number of bytes occupied by a buffer in CH_UTF16 format
209 : the result includes the null termination
210 : limited by 'n' bytes
211 : **/
212 910171 : size_t utf16_len_n(const void *src, size_t n)
213 : {
214 : size_t len;
215 :
216 2429755 : for (len = 0; (len+2 < n) && SVAL(src, len); len += 2) ;
217 :
218 910171 : if (len+2 <= n) {
219 910022 : len += 2;
220 : }
221 :
222 910171 : return len;
223 : }
224 :
225 : /**
226 : * Copy a string from a char* unix src to a dos codepage string destination.
227 : *
228 : * @converted_size the number of bytes occupied by the string in the destination.
229 : * @return bool true if success.
230 : *
231 : * @param flags can include
232 : * <dl>
233 : * <dt>STR_TERMINATE</dt> <dd>means include the null termination</dd>
234 : * <dt>STR_UPPER</dt> <dd>means uppercase in the destination</dd>
235 : * </dl>
236 : *
237 : * @param dest_len the maximum length in bytes allowed in the
238 : * destination. If @p dest_len is -1 then no maximum is used.
239 : **/
240 4180 : static bool push_ascii_string(void *dest, const char *src, size_t dest_len, int flags, size_t *converted_size)
241 : {
242 : size_t src_len;
243 : bool ret;
244 :
245 4180 : if (flags & STR_UPPER) {
246 8 : char *tmpbuf = strupper_talloc(NULL, src);
247 8 : if (tmpbuf == NULL) {
248 0 : return false;
249 : }
250 8 : ret = push_ascii_string(dest, tmpbuf, dest_len, flags & ~STR_UPPER, converted_size);
251 8 : talloc_free(tmpbuf);
252 8 : return ret;
253 : }
254 :
255 4172 : src_len = strlen(src);
256 :
257 4172 : if (flags & (STR_TERMINATE | STR_TERMINATE_ASCII))
258 4134 : src_len++;
259 :
260 4172 : return convert_string(CH_UNIX, CH_DOS, src, src_len, dest, dest_len, converted_size);
261 : }
262 :
263 : /**
264 : * Copy a string from a dos codepage source to a unix char* destination.
265 : *
266 : * The resulting string in "dest" is always null terminated.
267 : *
268 : * @param flags can have:
269 : * <dl>
270 : * <dt>STR_TERMINATE</dt>
271 : * <dd>STR_TERMINATE means the string in @p src
272 : * is null terminated, and src_len is ignored.</dd>
273 : * </dl>
274 : *
275 : * @param src_len is the length of the source area in bytes.
276 : * @returns the number of bytes occupied by the string in @p src.
277 : **/
278 164 : static ssize_t pull_ascii_string(char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
279 : {
280 164 : size_t size = 0;
281 :
282 164 : if (flags & (STR_TERMINATE | STR_TERMINATE_ASCII)) {
283 0 : if (src_len == (size_t)-1) {
284 0 : src_len = strlen((const char *)src) + 1;
285 : } else {
286 0 : size_t len = strnlen((const char *)src, src_len);
287 0 : if (len < src_len)
288 0 : len++;
289 0 : src_len = len;
290 : }
291 : }
292 :
293 : /* We're ignoring the return here.. */
294 164 : (void)convert_string(CH_DOS, CH_UNIX, src, src_len, dest, dest_len, &size);
295 :
296 164 : if (dest_len)
297 164 : dest[MIN(size, dest_len-1)] = 0;
298 :
299 164 : return src_len;
300 : }
301 :
302 : /**
303 : * Copy a string from a char* src to a unicode destination.
304 : *
305 : * @returns the number of bytes occupied by the string in the destination.
306 : *
307 : * @param flags can have:
308 : *
309 : * <dl>
310 : * <dt>STR_TERMINATE <dd>means include the null termination.
311 : * <dt>STR_UPPER <dd>means uppercase in the destination.
312 : * <dt>STR_NOALIGN <dd>means don't do alignment.
313 : * </dl>
314 : *
315 : * @param dest_len is the maximum length allowed in the
316 : * destination. If dest_len is -1 then no maxiumum is used.
317 : **/
318 429876 : static ssize_t push_ucs2(void *dest, const char *src, size_t dest_len, int flags)
319 : {
320 429876 : size_t len=0;
321 429876 : size_t src_len = strlen(src);
322 429876 : size_t size = 0;
323 : bool ret;
324 :
325 429876 : if (flags & STR_UPPER) {
326 2973 : char *tmpbuf = strupper_talloc(NULL, src);
327 : ssize_t retval;
328 2973 : if (tmpbuf == NULL) {
329 0 : return -1;
330 : }
331 2973 : retval = push_ucs2(dest, tmpbuf, dest_len, flags & ~STR_UPPER);
332 2973 : talloc_free(tmpbuf);
333 2973 : return retval;
334 : }
335 :
336 426903 : if (flags & STR_TERMINATE)
337 294726 : src_len++;
338 :
339 426903 : if (ucs2_align(NULL, dest, flags)) {
340 147571 : *(char *)dest = 0;
341 147571 : dest = (void *)((char *)dest + 1);
342 147571 : if (dest_len) dest_len--;
343 142367 : len++;
344 : }
345 :
346 : /* ucs2 is always a multiple of 2 bytes */
347 426903 : dest_len &= ~1;
348 :
349 426903 : ret = convert_string(CH_UNIX, CH_UTF16, src, src_len, dest, dest_len, &size);
350 426903 : if (ret == false) {
351 0 : return 0;
352 : }
353 :
354 426903 : len += size;
355 :
356 426903 : return (ssize_t)len;
357 : }
358 :
359 :
360 : /**
361 : Copy a string from a ucs2 source to a unix char* destination.
362 : Flags can have:
363 : STR_TERMINATE means the string in src is null terminated.
364 : STR_NOALIGN means don't try to align.
365 : if STR_TERMINATE is set then src_len is ignored if it is -1.
366 : src_len is the length of the source area in bytes
367 : Return the number of bytes occupied by the string in src.
368 : The resulting string in "dest" is always null terminated.
369 : **/
370 :
371 0 : static size_t pull_ucs2(char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
372 : {
373 0 : size_t size = 0;
374 :
375 0 : if (ucs2_align(NULL, src, flags)) {
376 0 : src = (const void *)((const char *)src + 1);
377 0 : if (src_len > 0)
378 0 : src_len--;
379 : }
380 :
381 0 : if (flags & STR_TERMINATE) {
382 0 : if (src_len == (size_t)-1) {
383 0 : src_len = utf16_len(src);
384 : } else {
385 0 : src_len = utf16_len_n(src, src_len);
386 : }
387 : }
388 :
389 : /* ucs2 is always a multiple of 2 bytes */
390 0 : if (src_len != (size_t)-1)
391 0 : src_len &= ~1;
392 :
393 : /* We're ignoring the return here.. */
394 0 : (void)convert_string(CH_UTF16, CH_UNIX, src, src_len, dest, dest_len, &size);
395 0 : if (dest_len)
396 0 : dest[MIN(size, dest_len-1)] = 0;
397 :
398 0 : return src_len;
399 : }
400 :
401 : /**
402 : Copy a string from a char* src to a unicode or ascii
403 : dos codepage destination choosing unicode or ascii based on the
404 : flags in the SMB buffer starting at base_ptr.
405 : Return the number of bytes occupied by the string in the destination.
406 : flags can have:
407 : STR_TERMINATE means include the null termination.
408 : STR_UPPER means uppercase in the destination.
409 : STR_ASCII use ascii even with unicode packet.
410 : STR_NOALIGN means don't do alignment.
411 : dest_len is the maximum length allowed in the destination. If dest_len
412 : is -1 then no maxiumum is used.
413 : **/
414 :
415 431075 : _PUBLIC_ ssize_t push_string(void *dest, const char *src, size_t dest_len, int flags)
416 : {
417 431075 : if (flags & STR_ASCII) {
418 4172 : size_t size = 0;
419 4172 : if (push_ascii_string(dest, src, dest_len, flags, &size)) {
420 4172 : return (ssize_t)size;
421 : } else {
422 0 : return (ssize_t)-1;
423 : }
424 426903 : } else if (flags & STR_UNICODE) {
425 426903 : return push_ucs2(dest, src, dest_len, flags);
426 : } else {
427 0 : smb_panic("push_string requires either STR_ASCII or STR_UNICODE flag to be set");
428 : return -1;
429 : }
430 : }
431 :
432 :
433 : /**
434 : Copy a string from a unicode or ascii source (depending on
435 : the packet flags) to a char* destination.
436 : Flags can have:
437 : STR_TERMINATE means the string in src is null terminated.
438 : STR_UNICODE means to force as unicode.
439 : STR_ASCII use ascii even with unicode packet.
440 : STR_NOALIGN means don't do alignment.
441 : if STR_TERMINATE is set then src_len is ignored is it is -1
442 : src_len is the length of the source area in bytes.
443 : Return the number of bytes occupied by the string in src.
444 : The resulting string in "dest" is always null terminated.
445 : **/
446 :
447 164 : _PUBLIC_ ssize_t pull_string(char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
448 : {
449 164 : if (flags & STR_ASCII) {
450 164 : return pull_ascii_string(dest, src, dest_len, src_len, flags);
451 0 : } else if (flags & STR_UNICODE) {
452 0 : return pull_ucs2(dest, src, dest_len, src_len, flags);
453 : } else {
454 0 : smb_panic("pull_string requires either STR_ASCII or STR_UNICODE flag to be set");
455 : return -1;
456 : }
457 : }
|
